* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Preface
Survey
Document related concepts
Transcript
British Journal of Nutrition (1998), 80, Suppl. 1, S1 S1 Preface This concerted action ‘Functional Food Science in Europe’ (FUFOSE) has been funded within the FAIR RTD programme, which is part of the Commission’s Fourth Framework Programme for research and technological development. This programme aims at promoting trans-European research in the primary production sectors of agriculture, horticulture, forestry, fisheries, and aquaculture, linking these with the input and processing industries, particularly food processing and renewable biomaterials. The food area is important within this programme and is covered by the theme ‘Generic Science and Advanced Technologies for Nutritious Foods’. There is growing interest in Europe in the concept of ‘Functional Foods’ and this concerted action, bringing together Europe’s scientists and industry, is fundamental to establishing a science-based approach to such foods. Liam Breslin European Commission DG XII - FAIR Programme British Journal of Nutrition (1998), 80, Suppl. 1, S3–S4 S3 Foreword Introduction We stand today at the threshold of a new frontier in nutritional sciences. The concepts of food are changing from a past emphasis on survival, hunger satisfaction, absence of adverse effect on health, and health maintenance to an emphasis on the promising use of foods to promote better health and well-being, thus helping to reduce the risk of chronic illnesses such as cardiovascular disease, some cancers and obesity. These new concepts are of particular importance in view of the benefits for health, consumer demand, the demand of the elderly population for an improved quality of their late life, the continuous increase in life expectancy, the increasing cost of health care, technical advances in the food industry, and the changing regulatory environment. There is already a recognition that European research expertise must be at the forefront in understanding the role of food in the maintenance and improvement of human health and well-being, in the reduction of risk of major diseases and in improving the competitive position of the European food industry. The number of major research programmes designed to investigate and clarify the healthpromoting value of foods and food components is forecast to continue to grow, particularly where serious debilitating diseases are concerned, e.g. heart disease, cancers and osteoporosis. The most recent knowledge in biochemistry, cell biology and physiology, but also in pathology, supports the hypothesis that diet also controls and modulates various functions in the body, and, in doing so, participates in the maintenance of the state of good health necessary to reduce the risk of some of the diseases. It is such an hypothesis which is at the origin both of the concept of ‘functional food’ and the development of a new scientific discipline ‘functional food science’. Functional food science aims to (1) identify beneficial interactions between the presence or absence of a food component (whether a macronutrient, micronutrient or socalled non-nutrient) and a specific function or functions in the body, and (2) understand their mechanisms, so as to support hypotheses to be tested in protocols relevant for human studies. The demonstration, in human subjects, of a specific interaction with one or a limited number of functions in the body will support a specific, often well-defined, claim of functional effects or disease risk reduction. Functional food science is indeed part of nutrition science, where the objectives are to maintain health and improve wellbeing and to create the conditions for disease risk reduction, and it is, in this respect, quite distinct from the medical or pharmaceutical sciences, where the objectives are mainly to cure diseases. A food is said to be ‘functional’ if it contains ‘a food component (whether a nutrient or not) which affects one or more targeted functions in the body in a positive way’. It can also include foods in which a potentially harmful component has (or components have) been removed by technological means. European Commission objectives An important objective is to improve the understanding of the role of food in the general health and well-being of the European consumer. Food can play a major role in maintaining and improving human health and well-being and in reducing the risk of major diseases. This will also lead to the design of special or tailored foodstuffs and ingredients for specific population groups or for specific health benefits. This will be an expanding area for the food industry in the future, and European industry, building on the considerable European research expertise, must be at the forefront here. This will involve multidisciplinary research projects combining the expertise of scientific partners, such as biochemists, nutritionists, the medical professionals and process technologists. The food and drink industry ranks as a major European industry processing raw materials from agriculture, horticulture, fisheries and aqua-culture into the diverse range of quality foodstuffs which are produced throughout Europe. Research in this sector has the major objective to improve the competitive position of the food industry which is composed of leading multinationals and a wide range of small and medium-sized enterprises specializing in food throughout Europe. ILSI Europe’s role In response to these critical developments, ILSI Europe has elaborated a project proposal for a European Commission Concerted Action aimed at establishing a science-based approach for concepts in functional food science. The goal of this concerted action is to establish a multidisciplinary European network to (1) critically assess the science base required to provide evidence that specific nutrients positively affect functions, (2) examine the available science from a function-driven point of view rather than a nutrientdriven one, and (3) reach consensus on targeted modifications of food and food constituents, and options for their application. This approach aims to provide key actors from Europe’s food and agricultural industry, governmental and inter-governmental bodies and the scientific community with an opportunity to exchange ideas and interact on a neutral platform. The project The Functional Food Science in Europe (FUFOSE) project was submitted in March 1995, approved in November 1995 and was expected to attain its objectives over a period S4 Foreword of 3 years. Project management and coordination was especially provided by ILSI Europe. Overall guidance on scientific and organizational issues was ensured through a steering committee, comprising members from both industry and academia. To attain the project objectives, the steering committee established individual theme groups (ITG) and organized a series of plenary meetings. The project started with a first plenary meeting, Functional Food Science in Europe: State of the Art, held 2–4 April 1996 in Nice, France. Based on the results of this meeting, six areas in human physiology were identified to be reviewed by the ITG responsible for producing theme papers to critically review the science base of the concept. The final composition of the ITG included industry and non-industry scientists. A draft brief was prepared by the steering committee to be addressed by each ITG while reviewing the literature data: (1) characterize specific body systems, including state-ofthe-art; (2) critically assess methodologies to characterize and quantify specific related functions; (3) identify and critically assess nutritional options modulating these functions; (4) evaluate potential safety implications related to these nutritional options; (5) identify the role of food technology in nutritional and safety aspects; (6) critically assess the science base required for providing evidence that specific nutrients positively affect functions; (7) identify areas where further research is required. The resulting documents were scrutinized in a Second Plenary Meeting held in July 1997 in Helsinki, Finland, and revised by the ITG chairs to include the comments made. The final reports of the six ITG are published in this issue of the British Journal of Nutrition. The papers need to be considered in the context of the entire project. They are not individual contributions and they form, all together, the reference to the FUFOSE project. Some repetitions, overlaps and contradictions may still appear. Only by reviewing all six papers will the reader have a balanced overview of both primary and secondary effects of functional foods. (5) Research needs. (6) Communication of the health benefits of functional foods. (7) Conclusions. The expert group that undertook the elaboration of the text was composed of two ITG chairs and four members of the steering committee. The document is currently under review. The goal is to publish this consensus document also in the British Journal of Nutrition. Food technology A group of experts on food technology have also been selected to examine the impact and feasibility of food technology on functional food development. Relevant viewpoints are: safety, nutrition and consumer acceptance/ sensory quality. This expert group has identified areas to concentrate on and terms of reference to be followed. Once the report is completed, it will be reviewed by the participants in the third plenary meeting. As soon as the comments are all taken into consideration, the paper will be published in a scientific journal. Acknowledgements We wish to thank, especially, all of the individual contributors to this FUFOSE project for devoting their time and efforts within such a tight timeframe. Their commitment and dedication will be remembered as exceptional and highly enthusiastic. Authors and contributors can be assured of ILSI Europe’s recognition and they will be paid tribute, as often as possible. Through their collaborative work they have participated in the making of ILSI Europe’s history and the Institute is extremely grateful to all. Coordinator: Dr Berry Danse, ILSI Europe, 83 Avenue E. Mounier, Box 6, B-1200 Brussels, Belgium. Scientific coordinator: Prof. Marcel Roberfroid, Catholic University of Louvain, Ecole de Pharmacie, Tour Van Helmont, 73 Avenue E. Mounier, B-1200 Brussels, Belgium. EC responsible: Dr Liam Breslin, Agro-industrial Research, Food, Commission of the European Communities, Directorate-General XII, Science, Research and Development, 200 Rue de la Loi, B-1049 Brussels, Belgium. Project manager; Dr Laura Contor, ILSI Europe, 83 Avenue E. Mounier, Box 6, B-1200 Brussels, Belgium. Consensus document These ITG papers provided the building blocks for a more general consensus document Concepts in Functional Food Science and Options for their Application. The outline of this consensus document was prepared based on the recommendations of the ITG and the steering committee members. This outline was also reviewed by the participants in the second plenary meeting who provided comments to be taken into consideration. The topics that will be addressed in the consensus document include the following. (1) Introduction. (2) Scientific basis for functional food science. (3) Target functions in relation to health outcome. (4) Food technology. q Nutrition Society 1998 British Journal of Nutrition (1998), 80, Suppl. 1, S5–S45 S5 Growth, development and differentiation: a functional food science approach B. Koletzko1 *, P. J. Aggett2 , J. G. Bindels3 , P. Bung4 , P. Ferré5 , A. Gil6 , M. J. Lentze7 , M. Roberfroid8 and S. Strobel9 1 Kinderpoliklinik, Klinikum Innenstadt der Ludwig-Maximilians-Universität, Pettenkoferstr. 8a, D-80336 München, Germany 2 Institute of Food Research, Norwich Laboratory, Norwich Research Park, Colney, Norwich NR4 7UA, UK 3 Nutricia Research, Verenigde Bedrijven Nutricia NV, PO Box 1, NL-2700 MA Zoetermeer, The Netherlands 4 University of Bonn, Women’s Hospital, Sigmund-Freud-Strasse 25, D-53105 Bonn, Germany 5 INSERM, Unité 465, Centre Biomédical des Cordeliers, 15, rue de l’Ecole de Médecine, F-75270 Paris, France 6 University of Granada, School of Pharmacy, Department of Biochemistry and Molecular Biology, Campus of Cartuja, E-18071 Granada, Spain 7 University of Bonn, Children’s Hospital, Adenauerallee 119, D-53113 Bonn, Germany 8 UCL, Ecole de Pharmacie, Tour Van Helmont, Avenue E. Mounier, B-1200 Brussels, Belgium 9 Clinical Sub-Dean’s Office, Institute of Child Health, 30 Guilford Street, London WC1N 1EH, UK Contents 1. Introduction 2. Nutrient–gene interaction, genetic regulation 2.1. Introduction 2.2. Modulation of gene expression participates in the adaptations of energy metabolism 2.3. Examples of gene regulation by nutrients 2.3.1. Carbohydrates 2.3.2. Fatty acids 2.3.3. Cholesterol 2.3.4. Amino acids 2.4. Nutrients and cell differentiation 2.4.1. Fatty acids 2.4.2. Retinoic acid 2.5. Concluding remarks 3. An overview of programmed cell death (apoptosis) 4. Supply of food ingredients before and during pregnancy 4.1. Physiological aspects of nutritional requirements in pregnancy 4.1.1. Energy 4.1.2. Protein 4.1.3. Carbohydrates 4.1.4. Lipids 4.1.5. Vitamins, minerals and trace elements 5. Modulation of growth 5.1. Introduction 5.2. Methods for the determination of growth 5.3. Growth factors in human milk and their influence on infant growth 5.4. Potential roles of non-protein nitrogen compounds as growth modulators S7 S7 S7 S8 S8 S8 S9 S9 S9 S10 S10 S10 S10 S10 S12 S12 S12 S13 S13 S13 S14 S15 S15 S15 S16 S16 5.5. 5.6. 5.7. 5.8. Human milk oligosaccharides and growth Free amino acids and tissue growth Polyamines and tissue growth Dietary nucleotides and tissue growth 5.8.1. Nucleotides and small intestine growth 5.8.2. Nucleotides and liver growth 5.9. Long-chain polyunsaturated fatty acids and cell growth 5.10. Early growth and later obesity 6. Maturation of the gastrointestinal tract 6.1. Introduction 6.2. Development of sugar hydrolases and transporters 6.3. Biosynthesis of intestinal brush–border membrane hydrolases 6.4. Intestinal absorption of glucose and fructose 6.5. Oligosaccharides and mucins 6.6 Probiotic substances in milk or milk substitutes 6.7. Dietary regulation of xenobiotic metabolism 7. Development of the immune system 7.1. Introduction 7.1.1. Which constituents of the immune system to investigate? 7.1.2. Special considerations for the immune system of the developing child 7.2. Antioxidants and vitamins 7.2.1. In general 7.2.2. Vitamin A 7.2.3. Vitamin C 7.2.4. Vitamin B complex 7.2.5. Vitamin E 7.2.6. Vitamin D S16 S16 S17 S17 S17 S17 S18 S18 S18 S18 S18 S19 S20 S20 S21 S21 S21 S21 S22 S22 S22 S22 S22 S22 S22 S22 S22 Abbreviations: APRT, adenosine phosphoribosyltransferase; cDNA, complementary DNA; DEXA dual-energy X-ray absorptiometry; DHA, docosahexaenoic acid; EFA, essential fatty acids; EGF, epidermal growth factor; GLUT, glucose transporter; hGH, human growth hormone; HGPRT, hypoxanthine phosphoribosyltransferase; HPA, hyperphenylalaninaemia; IDDM, insulin-dependent diabetes mellitus; Ig, immunoglobulin; IGF, insulinlike growth factor; LPH, lactase–phlorizin-hydrolase; MUC1, high-molecular mass glycoprotein; NPN, non-protein nitrogen; PAH, phenylalanine hydroxylase; PCD, programmed cell death; PKU, phenylketonuria; PPAR, peroxisome proliferator activated receptors; PUFA, polyunsaturated fatty acids; SGLT1, sodium-dependent glucose transporter 1; SI, sucrase–isomaltase; SPA, single-photon absorptiometry; SREBP, sterol regulatory element binding protein; XME, xenobiotic-metabolizing enzymes. *Corresponding author: Professor B. Koletzko, fax +49 89 5160 3336, email: [email protected] S6 B. Koletzko et al. 7.3. 7.4. 7.5. 7.6. 7.7. Multiple micronutrient supplementation studies Fatty acids Arginine Nucleotides Maturation of the immune system in formula-fed v. breast-fed infants 7.7.1. Effects of antigen transfer via breast milk on the infant’s immunity 7.7.2. Maternal diet during pregnancy and effects on the infant’s immunity 7.7.3. Maternal diet during pregnancy and lactation 7.8. Role of the gut flora and probiotic bacteria in the infant’s immunity and gut defence 7.8.1. Immune exclusion and elimination 7.9. Effects of formulas with protein hydrolysates on the infant’s immune responses 7.10. Insulin-dependent type 1 diabetes mellitus and cow’s milk exposure in infancy 8. Bone growth and mineralization 8.1. Cell biology of bone growth 8.2. Methodological aspects in bone-mass-related studies 8.3. Peak bone mass and relative risk of osteoporosis S23 S23 S23 S23 S23 S24 S24 S24 S24 S24 S24 S25 S25 S25 S25 S26 8.4. Bone growth and mineralization in infants and young children 8.5. Calcium supplementation in children and adolescents and bone health 8.6. Nutrients other than calcium and environmental factors involved in bone growth 9. Nutrient effects on development of neural functions and behaviour 9.1. Introduction 9.2. Physiology of neural development 9.3. Nutrition and neural development 9.3.1. Protein 9.3.2. Iodine 9.3.3. Iron 9.3.4. Zinc 9.3.5. Polyunsaturated fatty acids 9.4. Early nutrition and development of taste preferences 9.5. Methodological aspects 10. Production of bioactive factors for inclusion into food products 11. Commentary on biomarkers 12. Conclusions Abstract Few other aspects of food supply and metabolism are of greater biological importance than the feeding of mothers during pregnancy and lactation, and of their infants and young children. Nutritional factors during early development not only have short-term effects on growth, body composition and body functions but also exert long-term effects on health, disease and mortality risks in adulthood, as well as development of neural functions and behaviour, a phenomenon called ‘metabolic programming’. The interaction of nutrients and gene expression may form the basis of many of these programming effects and needs to be investigated in more detail. The relation between availability of food ingredients and cell and tissue differentiation and its possible uses for promoting health and development requires further exploration. The course of pregnancy, childbirth and lactation as well as human milk composition and the short- and longterm outcome of the child are influenced by the intake of foods and particularly micronutrients, e.g. polyunsaturated fatty acids, Fe, Zn and I. Folic acid supplementation from before conception through the first weeks of pregnancy can markedly reduce the occurrence of severe embryonic malformations; other potential benefits of modulating nutrient supply on maternal and child health should be further evaluated. The evaluation of dietary effects on child growth requires epidemiological and field studies as well as evaluation of specific cell and tissue growth. Novel substrates, growth factors and conditionally essential nutrients (e.g. growth factors, amino acids, polyunsaturated fatty acids) may be potentially useful as ingredients in functional foods and need to be assessed carefully. Intestinal growth, maturation, and adaptation as well as long-term function may be influenced by food ingredients such as oligosaccharides, gangliosides, highmolecular-mass glycoproteins, bile salt-activated lipase, pre- and probiotics. There are indications for some beneficial effects of functional foods on the developing immune response, for example induced by antioxidant vitamins, trace elements, fatty acids, arginine, nucleotides, and altered antigen contents in infant foods. Peak bone mass at the end of adolescence can be increased by dietary means, which is expected to be of long-term importance for the prevention of osteoporosis at older ages. Future studies should be directed to the combined effects of Ca and other constituents of growing bone, such as P, Mg and Zn, as well as vitamins D and K, and the trace elements F and B. Pregnancy and the first postnatal months are critical time periods for the growth and development of the human nervous system, processes for which adequate substrate supplies are essential. Early diet seems to have long-term effects on sensory and cognitive abilities as well as behaviour. The potential beneficial effects of a balanced supply of nutrients such as I, Fe, Zn and polyunsaturated fatty acids should be further evaluated. Possible long-term effects of early exposure to tastes and flavours on later food choice preferences may have a major impact on public health and need to be further elucidated. The use of biotechnology and recombinant techniques may offer the opportunity to include various bioactive substances in special dietary products, such as human milk proteins, peptides, growth factors, which may have beneficial physiological effects, particularly in infancy and early childhood. Growth: Development: Differentiation S26 S27 S27 S28 S28 S28 S29 S29 S29 S29 S29 S29 S30 S30 S30 S31 S33 Growth, development and differentiation 1. Introduction There are few other aspects of food supply and the metabolism of food ingredients that are of greater biological importance than the feeding of mothers during pregnancy and lactation, and of their children. The rapid growth of fetuses, infants and children, which double their body weights within only 6 weeks in utero and 4–5 months after birth respectively, depends on the supply of very large amounts of nutrients per kg body weight through the placenta, human milk and children’s diets. Marginal nutrient supplies are usually more critical in a developing and growing organism than in steady-state situations during adulthood. The ability to effectively utilize and compensate for unbalanced supplies is severely limited in the fetus and in young children due to small endogenous stores of a number of relevant substrates, and in many cases also due to immature metabolic pathways (e.g. amino acid metabolism) and physiological functions (e.g. renal conservation of substrates). Nutritional factors during early development have important immediate and short-term effects on growth, body composition and body functions. In addition, accumulating data indicate long-term effects of nutritional and metabolic factors during critical time periods of development on later physiological and metabolic processes, a phenomenon referred to as ‘metabolic programming’ (Barker, 1994). For example, epidemiological studies have suggested long-term effects of intrauterine and postnatal nutrient supply on the prevalence of obesity in adulthood (Ravelli et al. 1976) and on the risk of developing diabetes, hypertension, hypercholesterolaemia, CHD, and other disorders during adult life (Barker, 1994). The rate of death from all cardiovascular disease and from CHD (Fig. 1) in adulthood was found to be significantly related to body weight at birth (Osmond et al. 1993; Barker, 1994). Standardized mortality from CHD was also closely related to weight at 1 year in males but not in females (Fig. 2). These findings indicate the potential of influencing long-term health and life expectancy by modulations of maternal diet in pregnancy and of postnatal infant feeding. Moreover, S7 the type of postnatal infant feeding has been related to longterm outcomes such as the later incidence of insulin-dependent diabetes mellitus (IDDM) (Virtanen et al. 1991; Gerstein, 1994) and cognitive development (Lucas et al. 1992; Lanting et al. 1994), and a lasting effect of Ca intake during childhood and adolescence on bone mineral density and the risk of fractures in old age has been proposed (Ribot et al. 1995). This programming of permanent effects of the physiology and function of the organism during critical time periods of early development appears to be of major importance in preventive health strategies. Improved knowledge on the cellular and molecular mechanisms of programming and of the complex physiological and nutritional factors relevant to the health and well-being during the periods of reproduction and childhood is required. A better definition of the optimal supply of relevant nutrients and other food ingredients is expected to help optimize dietary intakes during these critical periods of life which should improve the chances that infants and children have to utilize fully their genetic potential, as well as maintain maternal health. 2. Nutrient–gene interaction, genetic regulation 2.1. Introduction The perinatal period is attended by important modifications in energy metabolism (Girard et al. 1992). In utero, the fetus receives a continuous intravenous supply of substrates for growth and oxidative metabolism. Immediately after birth, the maternal supply of substrates ceases abruptly and the newborn has to withstand a brief period of starvation before being fed at intervals with milk, a high-fat and low-carbohydrate diet. The sucking–weaning transition is also characterized by profound changes of nutrition (Girard et al. 1992). Towards the end of the sucking period, the milk is progressively replaced by the solid food diet of the adult, the composition of which is usually lower in fat and higher in carbohydrates. The successful adaptation of neonates to these changes in nutrition requires important modifications Fig. 1. Standardized mortality ratios for CHD below age 65 years, showing a statistically significant relationship with birth weight in 5585 women and 10 141 men born between 1911 and 1930. (Data from Barker, 1994.) S8 B. Koletzko et al. Fig. 2. Standardized mortality ratios for CHD below age 65 years, showing a statistically significant relationship with weight at age 1 year in 10 141 men, but not in 5585 women born between 1911 and 1930. (Data from Barker, 1994.) of energy metabolism in the vast majority of organs, intestine, liver, muscle, adipose tissue and brain. Postnatal development is also associated with differentiation processes and a high growth rate, involving specific requirements of nutrients such as amino acids (protein synthesis) or fatty acids (e.g. brain growth). Thus, nutrients can be considered to cause energy metabolism modifications but also to play a major role in organ growth and/or functional differentiation. 2.2. Modulation of gene expression participates in the adaptations of energy metabolism The adaptations of energy metabolism to the nutritional environment imply the modulation and/or emergence of metabolic pathways. This can be achieved through changes in the efficiency of a given step (specific transporters, enzymes) by allosteric or phosphorylation–dephosphorylation mechanisms, or by translocation of a protein into a different cellular compartment. However, many of these adaptations also imply a change in the amount of a given protein. This phenomenon is usually related to a change of the transcription rate of the corresponding gene. A good example of such a mechanism is the appearance at birth in mammals of phosphoenolpyruvate carboxykinase (EC 4.1.1.49), an enzyme of the gluconeogenic pathway allowing the de novo production of glucose by the liver (Girard et al. 1992). This pathway is essential for the survival of the newborn mammal, which undergoes a brief period of starvation followed by the ingestion of a diet low in carbohydrates, the milk. The transcription rate of this gene is extremely low during the fetal period and increases abruptly in the first hours after birth, allowing the newborn to maintain glucose homeostasis. Another example is the lipogenic pathway in the rat species. This pathway allows synthesis of fatty acids from glucose when this substrate is ingested in excess of energy requirements. During the sucking period, the capacity of this pathway is kept very low because the small quantity of glucose ingested is essentially directed towards oxidative processes. When the rat is weaned onto the high-carbohydrate diet, this pathway is switched on and excess glucose will be converted into fatty acids, ultimately stored as triacylglycerols in the adipose tissue. The expression of one of the key enzymes of this pathway, fatty acid synthase (EC 2.3.1.85) is extremely low during the sucking period but increases when the rat is weaned onto the adult high-carbohydrate diet, but not if weaning occurs onto a high-fat diet, clearly underlining the importance of the nutritional environment in this process. This phenomenon is due to the activation of the gene transcription process at weaning (Foufelle et al. 1996). Thus, modulation of gene expression must be considered as an integral part of the adaptations occurring during development. Although it has been known for a long time that nutrients can regulate the expression of specific genes in prokaryotes (the lactose operon in Escherichia coli for instance) or in primitive eukaryotes such as yeasts, the demonstration that a similar phenomenon occurs in higher eukaryotes is recent. The regulation of specific gene expression in mammals in response to changes in nutrition has become a major aspect of modern nutrition, due to the emergence of molecular biology that has allowed the cloning of most of the genes involved in the regulation of energy metabolism. Recently, it has been demonstrated that major (glucose, fatty acids, amino acids) or minor (Fe, vitamins) dietary constituents participate, in concert with hormones, in the regulation of gene expression in response to nutritional changes (see for instance Clarke & Abraham, 1992). 2.3. Examples of gene regulation by nutrients 2.3.1. Carbohydrates. In the liver and adipose tissue, excess glucose, after its metabolism into pyruvate through glycolysis is converted into fatty acids by the lipogenic pathway. The expression of three enzymes of the combined glycolytic–lipogenic pathway has been shown to respond to an increased glucose concentration: L-pyruvate kinase (EC 2.7.1.40; liver), fatty acid synthase and acetyl-coA carboxylase (EC 6.4.1.2; liver and adipose tissue) (Foufelle et al. 1992; Vaulont & Kahn, 1994). In vivo, it has been shown in rats that high-carbohydrate diets induce the transcription Growth, development and differentiation of these genes, whereas transcription is inhibited by starvation or a high-fat diet. During the sucking period in rats, the expression of these enzymes is kept low and dramatically increases at weaning on to a high-carbohydrate diet. In vitro studies have shown that glucose is the primary inducer of the gene transcription. This effect requires that glucose is metabolized at least into glucose-6-phosphate which might be the signal metabolite. The response of transcription to high glucose is a very rapid phenomenon (less than 1 h). Glucose response elements, which bind specific transcription factors of the USF/MLTF family have been characterized on these genes although the mechanism linking glucose-6-phosphate to transcription factors is presently unknown (Foufelle et al. 1996). In the b-cells of the islets of Langerhans, glucose induces the transcription of the insulin gene on which glucose response elements have also been characterized (Docherty & Clark, 1994). This feature is obviously relevant to the diabetic syndrome. 2.3.2. Fatty acids. In vitro studies have shown that the transcription of a number of genes of adipocytes is increased in the presence of fatty acids. This is the case, for instance, for phosphoenolpyruvate carboxykinase, an enzyme involved in this tissue in the provision of a-glycerophosphate necessary for the esterification of fatty acids (Antras-Ferry et al. 1995). Similarly, the expression of the fatty acid binding protein ap2, which binds fatty acids into the cell, is strongly stimulated by fatty acids (Grimaldi et al. 1992). Fatty acid response elements have been characterized in the promoter of these genes. They bind a transcription factor called peroxisome proliferator activated receptors (PPAR) which can be activated by the binding of fatty acids or a metabolite of fatty acids (prostaglandin for instance) (Schoonjans et al. 1996). Fatty acids have also been shown to inhibit gene expression in rats. The addition of a small amount (20–30 g/kg) of polyunsaturated fatty acids (PUFA) of the n-3 or n-6 families to a high-carbohydrate fat-free diet decreases markedly the lipogenic capacity and the activity of lipogenic enzymes (Clarke, 1994). In contrast, monounsaturated and saturated fatty acids have no effects. Interestingly S9 enough, this effect seems to be specific to the liver since lipogenesis is not affected in the adipose tissue. The decrease induced by PUFA of the activity of lipogenic enzymes such as fatty acid synthase, acetyl-CoA carboxylase or glucose-6-phosphate dehydrogenase (EC 1.1.1.49), is clearly linked to an inhibition of gene transcription as shown in studies using primary cultures of rat hepatocytes (Clarke, 1994). At the present time the cellular and molecular mechanisms involved in the inhibitory effect of PUFA on gene transcription have not been elucidated. 2.2.3. Cholesterol. A very interesting series of studies has been performed by the group of Brown and Goldstein on the effect of cholesterol on gene expression (Wang et al. 1994). Cholesterol represses the expression of genes involved either in the synthesis of cholesterol (cytoplasmic hydroxymethylglutaryl-CoA synthase, EC 4.1.3.5) or in its uptake from external sources, the LDL receptor (LDL are lipoproteins rich in cholesterol). In the absence of cholesterol, the transcription of these genes is activated by a transcription factor called sterol regulatory element binding protein (SREBP). SREBP is usually hooked onto the endoplasmic reticulum where it can be cleaved by a protease. SREBP can then be transferred into the nucleus and activates the transcription of relevant genes. In the presence of cholesterol, the protease is inhibited and SREBP can no longer enter into the nucleus and stimulate gene transcription (see Fig. 3). 2.3.4. Amino acids. In yeast, amino acid starvation results in the activation of several genes involved in N metabolism and the mechanisms involved are now known (Kilberg et al. 1994). In mammalian cells, amino acid availability also modulates the expression of some genes as shown, for instance, for asparagine synthetase (EC 6.3.5.4) which is responsible for the biosynthesis of asparagine from aspartate and glutamine. When cultured cells are transferred to a medium lacking asparagine, the concentration of asparagine synthetase mRNA increases. It must be underlined, however, that the signalling mechanism exhibits a broad substrate specificity since availability of other amino acids controls the asparagine synthetase mRNA Fig. 3. Model for the cholesterol-dependent control of gene transcription. SREBP, sterol regulatory element binding protein. S10 B. Koletzko et al. concentration as well. The mechanisms for the increase in asparagine synthetase mRNA concentration could involve both cis-acting elements contained in the mRNA itself and affecting its stability, as well as in the genomic promoter sequence (Kilberg et al. 1994). The signal metabolite could be, in fact, the degree of occupancy of the transfer RNA as in yeast, although evidence is clearly lacking. 2.4. Nutrients and cell differentiation The fetal–neonatal period is also characterized by the differentiation of a number of organs. One important question, which is presently totally unaddressed, is whether nutrients could affect differentiation per se and thus modulate physiological functions on a long-term basis. Two examples are listed here. 2.4.1. Fatty acids. One example of such a mechanism stems from in vitro studies on pre-adipocyte cell lines. It has been shown that fatty acids are not only able to modulate the transcription rate of specific genes in differentiated adipocytes, but are also able to induce preadipocyte differentiation into adipocytes through their action on a specific isoform of a PPAR transcription factor (Schoonjans et al. 1996). Obviously, if this happens also in vivo, it would imply that perinatal nutrition could modulate the number of adipocytes and, thus, be a crucial determinant of a possible expansion of this tissue, in relation to the obesity syndrome. 2.4.2. Retinoic acid. Vitamin A or retinol can be oxidized to retinoic acid in cells. In addition to retinol, b-carotene may also be a source of retinoic acid. Retinoic acid and retinoid X receptors which belong to the family of steroid–thyroid hormone receptors have been cloned (De Luca, 1991). They are, in fact, ligand-activated transcription factors (of the same family as the PPAR, see p.S9). It has been shown in numerous studies that retinoic acid is a potent morphogen and that it can affect fetal development (De Luca, 1991). Thus, vitamin A or b-carotene deficiency or overload might have major consequences on tissue differentiation and fetal development. 2.5. Concluding remarks These examples underline the idea that gene regulation by nutrients, a process we induce each time we eat a food of specific composition, includes a very wide range of mechanisms involved in the regulation of crucial pathways as well as in cell differentiation. Since the fetal–neonatal period is concomitant with marked changes in the nutritional environment, it represents a period in which these mechanisms are particularly important. As for other components of cell functions, it is very likely that gene regulation by nutrients is subject to variations linked to genetic polymorphisms among individuals, of which some could ultimately lead to pathologies. This obviously opens new fields for genetic studies on nutrientrelated pathologies (obesity, for instance). 3. An overview of programmed cell death (apoptosis) A type of cell death which does not involve primary cell membrane disintegration and tissue inflammatory responses was identified some 70 years ago, and the term apoptosis was applied by analogy with the deciduous loss of leaves by trees. The process was identified by characteristic histomorphological changes in the cell nuclei (Kroemer et al. 1995). Although apoptosis is sometimes used as a synonym for programmed cell death (PCD), it has become apparent that apoptosis can be caused by noxious or toxic events which might affect several cell components and not just primarily the nucleus. PCD applies to the entire process and phenomenology of this form of cell death and thus embraces the intrinsic mechanisms and regulatory processes involved. PCD might not necessarily be induced by external toxic stimuli, but rather by signals which are a fundamental component of cellular and tissue physiology. The distinction between the two terms PCD and apoptosis is subtle, and not invariably respected. Clearly it is misleading to characterize PCD on the basis of nuclear morphological changes since these are not always the primary events in PCD (Kroemer et al. 1995; Hale et al. 1996; Vaux & Strasser, 1996; Nagata, 1997) but one should not be distracted overmuch by the distinction: rather interest should focus on the processes involved in the induction, mediation, regulation and process of the cellular events, and the ways in which dietary components might affect these mechanisms which are fundamental to tissue differentiation, development and function. PCD is a more appropriate description of a variety of intrinsic cellular events, involving cytoplasmic as well as nuclear processes which precede loss of intracellular and cellular membrane integrity, should these occur at all. However, non-physiological stimuli can effect mechanisms involved with PCD, and the process or its dysfunction can be an integral part of oncogenesis and autoimmune disorders (Hale et al. 1996). All multicellular organisms use PCD to remove superfluous and damaged cells in tissues and organs, particularly, but not exclusively, in proliferating tissues. There is a strong evolutionary conservation of the mechanisms of PCD and the nematode Caenorhabditis elegans is a valuable model for the system in higher species, including man, in whom several genes and effectors of PCD are homologous to cell death genes and products in C. elegans (e.g. proto-oncogene bcl-2 is homologous to the nematode ‘cell-suicide gene’ ced3) (Kroemer et al. 1995; Vaux & Strasser, 1996). The induction mechanisms of PCD are more complex in higher animals in whom co-operativity between the different tissues of organs plays an important part in determining PCD. For example, the extracellular matrix exerts some control over differentiation and morphogenesis of organs by specific effects on constituent cells, tissue-specific gene expression, and cell death (Roskelley et al. 1995). This is important for the differentiation of tissues and organs (e.g. the gastrointestinal tract) throughout life as well as during embryogenesis. PCD can be envisaged to comprise three stages: induction, the effector stage, and degradation (Fig. 4) (Kroemer et al. 1995). There are at least two principal routes of inducing PCD. One involves genotoxic events damaging DNA; the other involves receptor-mediated stimuli such as specific death signals, the absence of rescue signals such as growth factors, and contradictory or conflicting signals (Kroemer et al. 1995; Nagata, 1997). Growth, development and differentiation S11 Fig. 4. The stages of programmed cell death and apoptosis. (From Kroemer et al : 1995.) The origin of signals can be systemic or local (e.g. extracellular matrix) (Hakomori & Igarashi, 1995) humoral cytokines (Nagata, 1997), or cellular such as cytotoxic T cells. The signal and transduction systems stimulating PCD include steroids (Evans-Storms & Cidlowski, 1995), cytokines such as tumour necrosis factor and nerve growth factor and interleukins, all of which operate through their cognate receptors (Cosman, 1994; Nagata, 1997). The responsiveness to such stimuli may also be determined by the nature of the cell surface receptors which might change during ontogenesis (and, for that matter, oncogenesis). The surface receptors include glycosphingolipids, lectins and sphingosines and their respective roles have not been totally clarified. It is noteworthy that inherited defects of cell-surface CD40 cause a human X-linked immunodeficiency syndrome associated with defective PCD, and analogous syndromes in mice are associated with congenital abnormalities of Fas receptors and ligands (Cosman, 1994; Nagata, 1997). Changes in structural components of cellular membranes (e.g. ceramide formation) can induce signal transduction cascades (Ballou et al. 1996) leading to cell death. The heterogeneity of signals is matched by the variety of effector mechanisms which are stimulated. These involve tyrosine kinase, protein kinase C, and other kinases, Ca permeability routes and intracellular proteases. These initiate an amplification cascade affecting a number of genes, several of which are known proto-oncogenes (e.g. c-myc) or tumour suppressors (e.g. p53) and which are common both to PCD and to the regulation of normal cell differentiation, function and intermediate metabolism (Kroemer et al. 1995; Hale et al. 1996; Nagata, 1997). This highlights the central conundrum of PCD, namely what is the ultimate determinant of whether a cell undergoes PCD or not? Is there a threshold event? It has been suggested that one basic determinant is the stage of the cell cycle when the cell receives the relevant signals (Meikrantz & Schlegel, 1995) or the availability of an appropriate nutrient supply (that is, do some stressors such as a specific nutrient excess or deficiency induce cell death by more direct means than, say, oxidant damage or simple starvation?). Intrinsic effector systems for cell death programmes involve cytoplasmic and nuclear metabolic and functional disintegration with ultimate morphological damage. Within the mitochondria there is a decrease in transmembrane potential, energy uncoupling of the respiratory chain, and increased production of reactive O species. These increase oxidative damage and increase leakage of mitochondrial Ca. In the cytoplasm there is a loss of anabolic activities with a corresponding activation of proteases, disruption of the cytoskeleton and of the endoplasmic reticulum: similarly in the nucleus the endonucleases are activated and there is nucleolysis with some activation of enzymes usually associated with repair activity. The mechanisms and regulation of PCD and apoptosis in abnormal cell proliferation have focused on their role in cancer and the possibility that external events or compounds might induce, suppress, regress, or protect against cancer by affecting these processes. The potential role of dietary components influencing these processes is being investigated actively in relation to functional food science. In other physiological areas there might exist other opportunities to explore and exploit a better understanding of these processes in cellular proliferation and differentiation. These are outlined in the following paragraphs. PCD of uterine cells enables blastocyst implantation and placentation (Welsh, 1993). Prostaglandins, leukotrienes, platelet-activating factor, and transforming growth factor are thought to induce this process. In embryogenesis there is extensive mesodermal PCD to eliminate undifferentiated S12 B. Koletzko et al. cells (Sanders & Wride, 1995; Wride & Sanders, 1995). The extracellular matrix and the regulation of cell adhesion molecules and integrins, perhaps involving tumour necrosis factor-a as a cellular growth and differentiation factor, have been shown in the regulatory processes in the morphogenesis of the limb bud (Hurle et al. 1995). Again extracellular matrix and specified cell adhesion molecules control endothelial cell position and behaviour and enable the definition of mechanisms directing endothelial cell differentiation, commitment, migration and organization into a tube as the fundamental components of vasculogenesis (Baldwin, 1996). PCD has also been shown to be integral to the development of the conduction processes and to be involved in the pathogenesis of congenital cardiac structural and electrical conduction abnormalities as well as acquired cardiovascular disease (James, 1993). The role of PCD in intestinal mucosal development and differentiation is discussed elsewhere: a similar dependence on PCD has been demonstrated in renal organogenesis (Igarashi, 1994). It is in haematopoiesis and in the development of the immune system that there might be some particular aspects of relevance to functional food science. Haematopoiesis and PCD involve an extensive multigene cytokine network which has positive regulators such as colony-stimulating factors and interleukins and negative regulators such as transforming growth factor-b and tumour necrosis factor (Krammer et al. 1994; Cidlowski et al. 1996; Sachs, 1996). Apoptosis in T and B lymphocytes is involved in all fundamental processes in the immune system and PCD appears to be vital in selecting or favouring development of specific lymphocytes and eliminating cells which are sensitized to autoantigens. Factors which would interfere with this process would be undesirable and there exists a possibility that similar selective processes might underlie the acquisition of immunotolerance to ingested antigens in foods or the development of adverse immune reactions to food constituents. In the central nervous system neurones and glia synthesize and secrete cytokines which affect the differentiation and function of nerve cells (Dragunow & Preston, 1995; Kawata, 1995; Sei et al. 1995). Disturbance of these cytokinemediated interactions may lead to neuronal dysfunction and/or cell death and contribute to the pathogenesis of central nervous system diseases. The functional and cellular differentiation underlying sexual dimorphism in the central nervous system might depend on the effects of steroids on PCD in the relevant regions of the brain. Many of the phenomena regarded as inevitable consequences of ageing, including nerve cell death, involve processes of PCD and apoptosis, and a better understanding of these events might in due course lead to a more sophisticated and rational scientific basis on which to base functional food science and its application to avoid the depredations of ageing (Zakeri & Lockshin, 1994). 4. Supply of food ingredients before and during pregnancy The nutritional status of the mother before and during pregnancy may influence fertility, the course of pregnancy and the incidence and severity of complications during gestation and birth, lactation (Rasmussen 1992), and the short- and long-term health and development of the baby. During recent years scientific understanding of the physiology of energy metabolism, weight gain and particularly the effects of micronutrient status and their clinical consequences have considerably improved. The dietary supply of certain nutrients may have beneficial or preventive aspects both for the course of gestation and for the offspring. 4.1. Physiological aspects of nutritional requirements in pregnancy The range of weight changes that occurs in pregnant women is wide and ranges from a loss of weight to a gain of 25 kg or more. A normal weight gain during pregnancy for most women with normal prepregnancy weight is in the order of 11–15 kg (Institute of Medicine, 1990). In addition to food intake, weight gain is also influenced by medical disorders such as development of oedema associated with preeclampsia. The weight increase and substrate deposition in maternal and fetal tissues is not only due to increased dietary intake during pregnancy, but also supported by a variety of complex adaptations of gastrointestinal, endocrine and metabolic functions. The effects of food ingredients on these adaptations as well as on placental function and the efficacy of materno-fetal substrate transfer are not well understood and require further exploration. Substrate needs during pregnancy are not only met by the food consumed during this period of time, but food intake before or between pregnancies may be of major importance for providing adequate availability of a number of nutrients, e.g. Ca and Fe. Also, it is increasingly appreciated that substrate needs during pregnancy may show marked interindividual variability, partly due to genetic heterogeneity in a population as is the case for metabolic pathways involving folic acid. 4.1.1. Energy. Based on a factorial approach calculating the additional energy requirements for the deposition of protein and fat as well as the energetic cost of tissue synthesis and enhanced maintenance metabolism, the total energy cost of pregnancy is assumed to be in the order of 334.72 MJ (80 000 kcal) (Food and Agriculture Organization/World Health Organization/United Nations University (FAO/WHO/UNU), 1985). In women who maintain their previous physical activity, this results in an additional energy allowance of about 837 kJ/d (200 kcal/d) equivalent to an increase of 9 % over the energy needs of 9.2 MJ (2200 kcal) of a non-pregnant, moderately active woman, or if adapted to an estimated increase of energy needs during the course of pregnancy, about 628 kJ/d (150 kcal/d) during the first and about 1464 kJ/d (350 kcal/d) during the second and third trimesters of pregnancy (FAO/WHO/UNU, 1985). However, energy needs may be lower with reduced physical activity, particularly towards the end of pregnancy. While moderately reduced energy intakes have little effect on pregnancy outcomes, severe energy deprivation during pregnancy results in reduced infantile birth weight. During the 6-month period of the Dutch famine with an available energy supply below 4184 kJ/d (1000 kcal/d) and a protein intake of no more than 30–40 g/d, average birth Growth, development and differentiation weight fell by 200 g (Ravelli et al. 1976). The effect was most marked if the hunger period coincided with the latter part of pregnancy. Later studies have confirmed these observations. In underweight women, a close relationship between maternal and neonatal body weights has been observed, while the correlation is much less close in pregnant women with a normal body weight (Luke & Petrie, 1980). Not only total body weight, but also the organ weights of liver, spleen, heart, adrenal gland, kidneys, skeleton and thymus are reduced in newborn infants of mothers who are underweight (Naeye et al. 1969), and placental size and number of cells by 15–20 % in pregnancy with intrauterine infantile growth failure or even up to 50 % in severe maternal malnutrition (Winnick, 1970). In malnourished South American women, average placental weights were reduced by about 15 %, and there was a marked reduction of the placental peripheral villi surface, the site of materno-fetal nutrient transfer, which may aggravate the restricted nutrient supply to the fetus in malnourished mothers. These findings are of concern, since neonatal birth weight is the strongest predictor of infant morbidity and mortality, and epidemiological studies have found strong associations of low birth weights with increased risks for cardiovascular disease and total mortality rates in adulthood (Barker, 1994). Several studies on energy supplementation during pregnancy in malnourished women found an increase in average birth weight and a reduction of the rate of low-birth-weight infants (Lechtig et al. 1975; Prentice et al. 1987), while intervention studies with an increased energy intake in apparently healthy women in Europe did not have significant effects on neonatal weight. However, dietary supplementation in Canadian women with twin pregnancies resulted in an increase of average birth weight by 80 g, and reductions of preterm deliveries and the rate of verylow-birth-weight infants by 30 and 50 % respectively (Dubois et al. 1991). In conclusion, energy supplements during pregnancy appear to be beneficial in selected populations at risk of very low intakes or increased demands. 4.1.2. Protein. With a factorial approach again based on calculated materno–fetal accretion, the additional protein requirements for the four successive 10-week periods of a normal pregnancy have been estimated as 0.6, 1.8, 4.8 and 6.1 g/d, respectively (Hytten & Leitch, 1971), and an increase of recommended protein intakes by about 6 g/d (expressed as milk or egg protein) over non-pregnancy values has been recommended (FAO/WHO/UNU, 1985). Even though this represents an increase of about 30 % over recommended intakes in non-pregnant women, and hence the recommended relative increment of protein intake is clearly higher than that of energy, the protein intakes of most pregnant women in Europe tend to exceed minimal requirements by far, with the possible exception of small subgroups consuming low-protein diets. It has been proposed that protein-enriched diets during pregnancy may be beneficial for the prevention of pregnancy-induced hypertension and pre-eclampsia, but conclusive evidence to support these hypotheses is missing (Roberts et al. 1974; Williams et al. 1981). 4.1.3. Carbohydrates. Glucose serves as the major energy source for the fetus, comprising about 90 % of the S13 energy supply. Hence, maternal carbohydrate metabolism during gestation is of potential relevance to the optimal supply for the fetus. Little is known about the effects of dietary habits, particularly the total amount and relative composition of sugars and starches on gestation, and potential implications for clinically relevant outcome variables such as macrosomia, postpartal hypoglycaemia, or increased risk of developing glucose intolerance later in life. 4.1.4. Lipids. Although there is a tendency in some pregnant women to aim at a low consumption of dietary fat (Hachey, 1994), there are no conclusive data to demonstrate the safety of maternal low-fat diets or immediate benefits for pregnant women and the fetus and infant. On the other hand, pregnant women appear to have high requirements for lipidsoluble vitamins and PUFA. During pregnancy, the concentrations of blood lipids and their constituent fatty acids increase considerably. Amounts (mg/l) of plasma phospholipid-associated essential fatty acids (EFA) were reported to increase during the course of pregnancy by about 40 % and those of the long-chain PUFA arachidonic acid (20:4n-6) and docosahexaenoic acid (DHA, 22:6n-3) by about 23 and 52 % respectively (Al et al. 1995). It has been proposed that a high supply of long-chain n-3 fatty acids may be beneficial for fetal development because of the importance of these compounds for neural tissue development (Koletzko, 1992), and that they may improve some obstetric complications, particularly lessen the severity of pregnancy-induced hypertension (Secher & Olsen, 1990). Moreover, observations in the population of the Faroe islands which consumes a diet rich in fish suggested that this high intake of n-3 fatty acids may increase average birth weight by prolonging gestation (Olsen et al. 1986). However, in this population there were also apparent adverse effects of n-3 fatty acids, including higher rates of blood loss on delivery, which may be explained by fish-oilinduced suppression of platelet aggregation, and a higher perinatal mortality. An intervention with supplementation of n-3 long-chain PUFA was reported to prolong pregnancy without any detrimental effects on growth of the fetus or the course of delivery (Olsen et al. 1992). Supplementation of fish oil with vitamins and minerals has been considered to reduce the frequency of pre-eclampsia, but a double-blind placebo-controlled trial did not find any effect of fish-oil supplementation on the occurrence of pregnancy-induced hypertension (Onwude et al. 1995). Pregnancy may be associated with DHA mobilization from maternal stores. Under the present dietary conditions, pregnancy is associated with a reduction of the EFA status and particularly of the DHA status. After delivery normalization takes place, but recovery of the DHA status appears to be still incomplete after 6 months (Al et al. 1995). Throughout pregnancy, the DHA content of plasma phospholipids of primigravida is significantly higher than that of multigravida, and a negative relationship was observed between DHA content and gravida number, which is reflected in neonatal DHA status and may have functional consequences for infant growth and development. Since increased supplies of selected PUFA during pregnancy may well have some benefits, the possible benefits as well as the potential risks of such strategies should be carefully evaluated. S14 B. Koletzko et al. 4.1.5. Vitamins, minerals and trace elements. Relative to the increased energy needs (about 10 %), the recommended relative increase of some other nutrients is markedly higher, for example, for folate the reference intake increases by 100 % in pregnancy (Scientific Committee for Food, 1993). Based on these considerations, the potential for an inadequate intake of one or more of these specific nutrients during pregnancy is greater than for total energy. Calcium. During the latter part of pregnancy, the fetus has a high rate of Ca accretion in the order of 250–300 mg/ d. This appears to be partly accounted for by extensive adjustments in Ca metabolism during pregnancy, particularly the inhibitory effects of placental oestrogen on maternal bone resorption that result in an enhanced release of parathyroid hormone and, hence, increased maternal Ca absorption, decreased urinary excretion and enhanced Ca retention (Cole et al. 1987). Although low dietary Ca intakes of malnourished pregnant women have been associated with reduced bone mineral densities of the newborn infants (Ramam et al. 1978), in well-nourished mothers no enhancement of neonatal Ca accretion by increased maternal intakes has been demonstrated. However, at relatively low intakes, maternal body stores of Ca may be utilized and depleted to meet fetal needs (Duggin et al. 1974). In women with multiple pregnancies, in particular, low Ca intakes may increase the risk of osteomalacia at a later age. Although dental caries is relatively common during pregnancy, there is no evidence for a causal link of its rate of occurrence with dietary Ca intakes. Low Ca intakes during pregnancy have also been associated with the occurrence of EPH-gestosis (oedema, proteinuria and hypertension) as well as eclampsia (Burke et al. 1943). In a controlled study, supplementation of the diets of pregnant women with Ca was associated with a reduced systolic blood pressure at term and a lower incidence of pregnancy-induced hypertension (Villar et al. 1987). In contrast, another study found no effect of a Ca supply on the incidence of pregnancy-induced hypertension, although mean blood pressure was lowered. Magnesium. Mg status is related to neuromuscular excitability, and it has been proposed that an additional Mg supply may contribute to prevention of eclampsia and other complications of pregnancy. However, controlled trials have not provided conclusive evidence for an improved course of pregnancy or delivery (Spatling & Spatling, 1988; Sibai et al. 1989). Iron. The total Fe needs for pregnancy have been estimated to be in the order of 300–850 mg (Bothwell et al. 1979; Hallberg, 1988). Although the efficacy of Fe absorption is markedly enhanced in pregnancy, the incidence of Fe deficiency and Fe-deficient anaemia during pregnancy remains a sizeable problem that may increase maternal and fetal morbidity and mortality (Llewellyn-Jones, 1965), cardiovascular stress associated with increased complication rates before and at birth (Banks & Beutler, 1988) and an elevated risk of delivering lowbirth-weight and premature infants (Scholl & Hediger, 1994). At particular risk of poor Fe status in pregnancy are women with vegetarian or predominantly vegetarian diets, because of the relatively low absorption of non-haem Fe. Thus, foods that promote net Fe absorption may be beneficial for some pregnant women. Zinc. Maternal Zn deficiency is highly teratogenic in rodents, and in monkeys it induces abnormal fetal brain development (Worthington-Roberts 1985). Studies in human pregnancies found that maternal leucocyte Zn levels during pregnancy were correlated with later infantile birth weight (Wells et al. 1987) and inversely related to the rate of low-birth-weight infants (Neggers et al. 1991). Moreover, a better maternal Zn status was associated with lower rates of pregnancy-induced hypertension, protracted delivery and maternal complications at birth and the rate of premature rupture of membranes (Sikorski et al. 1990). In contrast, another controlled study evaluating Zn supplementation in the second and third trimesters of pregnancy in Europe did not find any effect. To what extent these differing results were influenced by variations in Zn status of the respective populations studied, and may be reproduced in other populations, remains to be clarified. Iodine. In parts of Europe, the I status of many pregnant women is inadequate with a sizeable prevalence of increased thyroid size in mothers and newborn infants (World Health Organization, 1993). Poor I status is associated with an increased risk of miscarriage in early gestation and preterm delivery as well as compromised mental development of the infant (Xue-Yi et al. 1994). Fluoride. Since postnatal F ¹ supply has a strong preventive effect on the incidence of dental caries, the question was raised whether an increased F ¹ supply to pregnant women may contribute to protecting pre-eruptive teeth of the unborn child, but conclusive evidence to answer this question is not available. Folic acid. Folic acid is essential for the synthesis of pyrimidines, purines and hence of DNA and RNA, as well as amino acids and neurotransmitters; thus, an adequate folate status is of importance to allow undisturbed cell multiplication and growth during pregnancy. Following observations of a relation between folic acid status and the occurrence of neural tube defects in the offspring, several controlled clinical trials have demonstrated that folic acid supplied before conception and during the first weeks of pregnancy can markedly reduce the incidence of neural tube defects, such as spina bifida and anencephaly, by 40–70 % (Butterworth & Bendich, 1996). This important preventive effect is apparently associated with a folic acid-responsive derangement in homocysteine metabolism in a genetically determined subgroup of the population (SteegersTheunissen et al. 1991). Poor folic acid status of pregnant women has also been associated with miscarriages, repeated abortions, pregnancy length and neonatal outcome as well as unexplained sterility (Pietrzik et al. 1992; Bung et al. 1993, 1995). It has been recommended that all pregnant women should be supplemented with a daily dose of 0.4 mg folic acid from before conception to at least four completed weeks after conception (Scientific Committee for Food, 1993). In view of the postulated key role of folic acid status in the aetiology of atherosclerosis by means of modulating homocysteine metabolism, the question of a potential programming effect of intrauterine folic acid supply on the later risk of cardiovascular diseases has been raised (Pietrzik et al. 1992). Growth, development and differentiation 5. Modulation of growth 5.1. Introduction The term ‘growth’ expresses the increase in number and size of cells of a particular species and it refers to changes in body dimensions. Growth is a phenomenon usually associated with increase in length and weight whereas the term ‘development’ is a physiological concept indicating the progressive differentiation of tissues and organs with acquisition of their specific functions. All mammals start life as a single cell; during the early part of the gestation the fertilized ovum divides many times and different kinds of cells develop during the process of differentiation and arrange themselves to form part of the various organs of the body. The general principles of growth apply to all species but the rate of cell division is genetically determined and depends on nutrient supply and utilization. Regardless of the exact time that differentiation occurs it always results in the transformation of the parental cell into a large number of morphologically different progeny cell types. The rapidity of physical growth is regulated during the life cycle and is modulated by genetics, a variety of growth factors that interact with target cells, as well as environment and diet (Hernández & Argente, 1992; Philips, 1995). Human growth hormone (hGH) or somatotropin is essential for normal postnatal growth. It is released from the anterior pituitary gland on stimulation by growth hormonereleasing hormone or somatocrinin, a factor produced by the hypothalamic region of the brain. Like other pituitary hormones, hGH acts on target tissues, primarily the liver, to cause synthesis and release of a second hormone mediator, insulin-like growth factor (IGF)-I, also called somatomedin C, into the systemic circulation. IGF-I is a growth-accelerating peptide that acts directly on cartilage to promote bone growth. Deficiency of hGH production causes metabolic alteration and growth failure (Philips, 1995). Different requirements for growth of different cell types have been established (Sato et al. 1982) but despite considerable progress in tissue culture systems we are still a long way from identifying all the stimulatory and inhibitory macromolecules that regulate the growth of all the human cell types in vivo. Most growth factors are polypeptides or small proteins with molecular masses that vary from 1 to 40 kDa and bind to specific cell surface receptors, with pleiotropic effects on cells including changes in gene expression (Watson et al. 1987). By binding to their receptors, growth factors modify the activity of the membrane-bound enzyme adenylate cyclase (EC 4.6.1.1), using the GTP-binding protein as an intermediate so that the level of cAMP in the cell is altered. This, in turn affects the activity of cAMP-dependent protein kinases which phosphorylate specific target proteins, regulating their activity. Other possible second messengers to carry signals from growth receptors to the cell include Ca, inositol triphosphate and diacylglycerol (Lodish et al. 1995). In addition to genetic factors, neurohormonal and tissuespecific growth factors, growth is also affected by a number of metabolic and environmental factors which include the availability of nutrients. Recommended daily nutrient intakes have been established for all periods of life and both sexes to support an adequate growth of the human S15 being (National Research Council, 1989). However, there is a lack of information about how semi-essential nutrients can affect growth in specific periods of life and in particular situations including disease states. The rapidity of physical growth in the normal infant during the first months of life is remarkable and unmatched during later times of life. Moreover, physiological and developmental changes during infancy are as notable as the speed of physical growth. Changes in the rates of physical growth and in the allocation of dietary intake of energy and protein for growth and maintenance occur as a continuum rather than in discrete stages, but the progression of changes during the early months of life is very rapid. Human milk provides all nutrients necessary to support adequate growth of the term infant during the first 4–6 months of life. Furthermore, in addition to universally recognized nutrients, human milk contains a number of semi-essential nutrients, enzymes, hormones and growth factors which appear to have a role in supporting infantile growth (Koldovsky & Strbak, 1995). However, there is a lack of information about how those nutrients and factors interact with the growth process and how they affect specific tissue growth. 5.2. Methods for the determination of growth Anthropometric measurements such as weight, length, head circumference, skinfold thickness, limb circumference, mid-arm cross-sectional area, BMI etc. at various ages during infancy have been published in many developed countries and are useful in evaluating the size of an infant in relation to the size of his or her peers (Fomon & Nelson, 1994). In the infant the Quetelet index is difficult to interpret since the body mass increases rapidly from birth to 4 months of age and the percentage of body weight contributed by body fat also increases rapidly during that period. Reference data for increments in size and other indices of nutritional status are more sensitive in children beyond infancy, particularly in alerting to the possibility of illness or nutritional inadequacy (Fomon, 1991). Data on body composition at various ages are one of the bases for the factorial approach in estimating the requirements for various nutrients for growth. Chemical composition of animals is determined by direct whole-body analysis. The most extensive data on the composition of the term infant are those published by Widdowson (1982), but data beyond infancy are scanty. The most useful indirect methods of estimating various aspects of body composition in the infant are determination of total body water from the concentration of a suitable tracer in body fluid, i.e. heavy water, extracellular water, and determination of natural abundance of 40K in the whole body (Forbes, 1987). Measurements of both bioelectric impedance and of total body electrical conductivity can estimate non-invasively the fat-free body mass of infants and children (Mayfield et al. 1991; Houtkooper et al. 1992). Many new techniques are available for measurement of total body fat, although only a few can be used in general practice or in epidemiological research (Deurenberg, 1992). The urinary excretion rates of endogenous creatinine and 3-methyl-histidine have been proposed as indices of muscle S16 B. Koletzko et al. mass, and the urinary excretion rates of endogenous hydroxyproline and type I and type III procollagen propeptides as indices of growth rates (Trivedi et al. 1991). Measuring cell kinetics represents a new approach in evaluating cell and tissue growth. Cell kinetics can be used to measure the proliferation rate of cells, the phases of the cell cycle and the percentage of cells in cycle, to plot the position of dividing cells and determine the size of the proliferation compartment and to follow the movements of labelled cells. Cell kinetic studies received a boost when [3H]thymidine was introduced in the 1950s. Recently, several new techniques have been used in cell kinetics, i.e. incorporation of bromodesoxyuridine and labelling of cell antibodies (Kember, 1993). There is a need for new methods to evaluate the specific organ and tissue growth applicable in a wide range of conditions. 5.3. Growth factors in human milk and their influence on infant growth A large number of hormones and growth factors are present in human and bovine milks (Koldovsky & Thornburg, 1987; Strbak, 1991). Non-peptide hormones (thyroid hormones, cortisol, progesterone, pregnanediol, oestrogens and artificial contraceptives) and peptide hormones and growth factors (erythropoietin, hGH, growth hormone releasing factor, gonadotropin-releasing hormone, epidermal growth factor (EGF), insulin, IGF-I, nerve growth factor, gastrointestinal regulatory peptides and thyroid–parathyroid hormones) have been isolated and quantitated in human milk (Koldovsky & Strbak, 1995). The orogastric effects of hormones and growth factors on infant growth require further elucidation. However, there is some evidence that a number of hormones in human milk may contribute to the intestinal maturation of sucking infants. For example, oral administration of EGF to 10-d-old sucking rats resulted in changes in protein and DNA content of colonic mucosa (Koldovsky & Thornburg, 1987). 5.4. Potential roles of non-protein nitrogen compounds as growth modulators Non-protein N (NPN) compounds are present in most foods mainly as nucleic acids, free amino acids, small peptides and other minor compounds. The concentration of nucleic acids in foods depends on the number of cells of the original biological tissue. Thus, meat, fish and vegetal seeds are rich in nucleic acids whereas fruits have a low content (Gil & Uauy, 1995b). NPN accounts for 18–30 % of the total N in human milk. Some of this N, namely urea, contributes to the pool available for synthesis of non-essential amino acids in infants. Other NPN components may have particular roles in tissue growth. Among the NPN known to have specialized roles in the ontogeny of the human newborn are the growth factors, namely EGF, amino sugar oligosaccharides, free amino acids like taurine, arginine and glutamine, amino alcohols of phospholipids i.e. choline, nucleotides and nucleic acids and polyamines. 5.5. Human milk oligosaccharides and growth In addition to lactose, the carbohydrates of human milk include nucleotide sugars, glycolipids, glycoproteins, and oligosaccharides. Viverge et al. (1990) have isolated three oligosaccharide fractions representing 13–18 g/l; the concentration varied with the mother’s genetic ability to synthesize specific fucosyl linkages. Approximately eighty neutral and sialic acidic oligosaccharides have been isolated and identified (Newburg & Neubauer, 1995). Human milk oligosaccharides appear to be synthesized by some of the same glycosyltransferases that participate in the synthesis of glycoprotein and glycolipid cell surface components. Thus, it is reasonable to postulate that some of those compounds can act as analogues to host cell surface receptors for pathogens. Anderson et al. (1986) reported that specific oligosaccharides can inhibit binding of Streptococcus pneumoniae and Hemophilus influenzae to their receptors and Cravioto et al. (1991) described an oligosaccharide that inhibits adherence of enteropathogenic E. coli to their receptors. Other authors have reported that specific fucosylated oligosaccharides inhibit binding of invasive strains of Campylobacter jejuni (Ruiz-Palacios et al. 1992) and the toxicity of E.coli in vivo (Newburg et al. 1990). Gangliosides are glycosphingolipids that contain sialic acid (N-acetylneuraminic acid) as part of their carbohydrate moiety. GM1, a milk ganglioside present in human milk, binds to E. coli and Vibrio cholerae toxins and may contribute to infant protection against infection by those enteropathogens (Laegrid et al. 1986). Since lactating mothers differ genetically in their ability to produce various oligosaccharides, this variability might influence the susceptibility of breast-fed infants to enteric disease. The influence of supplementing infant milk formulas with oligosaccharides on the susceptibility of infants to gastrointestinal diseases, namely acute diarrhoea, is one of the current fields of intense investigation. 5.6. Free amino acids and tissue growth The free amino acid pool of human milk is small compared with the total amount of milk amino acids in protein. Free amino acids contribute only 10 % of the total NPN in human milk. Glutamine and taurine are the free amino acids found in higher concentrations in human milk. While clinical effects of low dietary intakes of taurine have not been demonstrated, there is a concern about the possibility of subclinical deficiency, particularly in the premature infant since its ability to synthestize taurine may be limited (Gaull et al. 1977). Free glutamine accounts for about 20 % of the total glutamine pool in human milk. Glutamine is currently extensively investigated because of its importance in cell and tissue culture and because it serves as a preferred respiratory fuel for rapidly proliferating and growing cells, such as enterocytes and lymphocytes. Moreover, it is a regulator of acid–base balance through the production of urinary NH 3, a carrier of N between tissues and an important precursor of nucleotides, amino sugars and proteins (Lacey & Wilmore, 1990). There is increasing evidence that glutamine may become a conditionally essential amino acid in critically ill patients, and glutamine appears to be important for the maintenance of small-intestinal structure and functionality (Newsholme & Carrié, 1994). Arginine and ornithine are also present in human milk as free amino acids, although the first is found in higher Growth, development and differentiation concentrations in milk proteins. Arginine has multiple biological properties, including the ability to stimulate anabolic hormone secretion: intravenous and enteral administration of arginine increases both insulin and hGH secretion (Cynober, 1994). Several studies show that arginine given to patients, as well as in various experimental stress models, acts by improving N balance, accelerating wound healing, and restoring depressed immunity. Dietary supplements of arginine have been shown to inhibit tumour growth in animals, probably by activating the immune system. However, in cancer patients arginine stimulates tumour protein synthesis, suggesting that arginine might have separate stimulatory effects on the tumour and on the immune system, the outcome depending on which effect predominates (Garlick & McNurlan, 1994). Arginine has also been shown to enhance the growth-hormone-releasinghormone-induced hGH rise in patients with anorexia nervosa (Ghigo et al. 1994). Moreover, oral administration of arginine enhances the hGH response to growth hormone releasing hormone in short children (Loche et al. 1993). Ornithine shares with arginine the ability to stimulate hGH secretion. In addition, ornithine as its a-ketoglutarate salt generates various molecules, i.e. glutamine. Ornithine ketoglutarate has been shown to improve N balance in various acute and chronic malnutrition states. It increases muscle protein anabolism in moderate catabolic states and reduces protein catabolism in hypercatabolic states (Cynober, 1994). Arginine and ornithine are precursors of NO and polyamines respectively. These metabolites participate intimately in permeability and adaptive responses of the gut. Recent animal studies showed improved morphology after ornithine ketoglutarate administration, acting perhaps through increased polyamine synthesis (Cynober, 1994). It is controversial whether exogenous arginine can be a relevant precursor of polyamines. 5.7. Polyamines and tissue growth Polyamines are detectable in relatively high quantities in human and rat milk. Artificial infant formulas do not contain appreciable amounts of polyamines, specifically putrescine and spermidine, and spermine is undetectable. Thus, formula-fed infants are not exposed to polyamines nor to any potential effects that these compounds may have on the developing intestine. It is noteworthy that food contains polyamines and that polyamines are produced by the gastrointestinal microflora. Thus, the direct uptake by enterocytes of preformed polyamines could contribute to the polyamine cellular pool. Indeed, putrescine and spermidine uptake has been shown in isolated rat enterocytes (Cynober, 1994). 5.8. Dietary nucleotides and tissue growth Human milk is the exclusive source of dietary nucleotides for infants during the first months of life, and its nucleotide profile (Gil & Sánchez-Medina, 1982; Gil & Uauy, 1995a) differs markedly from that of cow’s milk and most infant formulas (Gil & Sánchez-Medina, 1981; Gil & Uauy, 1989, 1995a). Preformed nucleotides may be of importance for the S17 growth of tissues with a rapid turnover (Van Buren et al. 1985; Nuñez et al. 1990; Uauy et al. 1990; Gil & Uauy, 1995b), particularly bone marrow, leucocytes and the intestinal mucosa which preferentially use the nucleotide salvage pathway to fulfil their purine and pyrimidine nucleotide requirements (Mackinnon & Deller, 1973; Savaiano & Clifford, 1981; LeLeiko et al. 1983; Cohen et al. 1984). Dietary nucleotides may modulate lipoprotein and fatty acid metabolism in human early life (Gil et al. 1986b, 1988; De-Lucchi et al. 1987; Pita et al. 1988; Morillas et al. 1994; Sánchez-Pozo et al. 1994), and they may enhance the growth of bifidobacteria and limit that of enterobacteria in the gut of newborn infants (Gil et al. 1986a; Gil & Uauy, 1995). Dietary nucleotides may affect small-intestinal growth in experimental animals (Gil & Uauy, 1995) and may have a role in the maintenance of the immune response both in animals (Van Buren et al. 1983, 1985; Pizzini et al. 1990; Kulkarni et al. 1992) and in human subjects (Carver et al. 1991; Gil & Uauy, 1995). 5.8.1. Nucleotides and small intestine growth. A number of factors are involved in the regulation of the renewal of the absorptive epithelium and in the repair of the epithelium under pathological conditions (Shiner et al. 1990). N-containing nutrients appear to be important for gut growth. At weaning, protein modulates the ontogenic changes in tissue DNA synthesis and plays a role in completing the growth of the rat’s gastrointestinal tract (Buts & Nyakabasa, 1985). Dietary nucleotides have been shown to influence gut development and repair after injury. Uauy et al. (1990) reported that intestinal disaccharidase activities are increased in rats during development by dietary nucleotides, and DNA, lactase (EC 3.2.1.23), sucrase (EC 3.2.1.48) and maltase (EC 3.2.1.20) activities increase with a nucleotide-supplemented diet in animals after chronic diarrhoea (Nuñez et al. 1990; Bueno et al. 1994). Dietary nucleotides promote the enterocyte growth in tissue culture (He et al. 1993; Sanderson & He, 1994) and improve the intestinal repair in an animal model of radiation injury (Uauy et al. 1994). Recent studies on the potential roles of exogenous nucleotides on proliferation, differentiation and apoptosis of human small-intestinal epithelium have shown that AMP may have an important role in controlling the dynamic balance of cellular turnover in the developing human small intestine (Tanaka et al. 1996). Moreover, dietary nucleotides influence the gene transcription in the intestine (LeLeiko et al. 1995). Animals receiving a purine and/or pyrimidine-free diet have a decreased protein synthesis and RNA throughout the intestine and specific mRNA for the enzymes hypoxanthine phosphoribosyltransferase (HGPRT) and adenosine phosphoribosyltransferase (APRT) (LeLeiko et al. 1987). A 35-base pair region identified in the HGPRT promoter is necessary to confer sensitivity to exogenous purines as a site for binding to trans-acting regulatory proteins (Walsh et al. 1990, 1992). 5.8.2. Nucleotides and liver growth. Extracellular nucleotides and nucleosides have been reported to modulate hepatocyte growth (Ohyanagi, 1989; Gil & Uauy, 1995b) and regeneration and to play an important role in the synthesis of glycogen (Buxton et al. 1986). Ogoshi et al. (1985, 1988) reported that parenterally administered S18 B. Koletzko et al. nucleotides improved the hepatic function and promoted earlier restoration of the N balance after liver injury or partial hepatectomy. Moreover, it has been observed that adenosine administration partially prevents cirrhosis induced by CCl 4 in rats due to a stimulation of total hepatic collagenase activity and is able to counteract the drastic decrease in adenine nucleotides (Hernández-Muñoz et al. 1990). Deprivation of dietary nucleotides results in a transient decrease in acid-soluble nucleotides and RNA content in rat liver as well as in a decreased protein synthesis rate (LópezNavarro et al. 1995). Dietary nucleotides have improved liver structural recovery and binuclearity in experimental cirrhosis induced by thioacetamide (Torres et al. 1996). In that model dietary nucleotides led to a lower number of stellate cells and to a lower collagen deposition. 5.9. Long-chain polyunsaturated fatty acids and cell growth Despite recent advances in neonatal care, low-birth-weight infants do not achieve first year growth equivalent to that of infants born at term. It has not been clarified how the administration of long-chain fatty acids to infants may affect growth and specific tissue growth and differentiation. Direct evidence that normalized growth might relate to arachidonic acid status came from the observation that formula supplemented with marine oil but no arachidonic acid decreased the concentrations of plasma phosphatidylcholine arachidonic acid (Carlson et al. 1991) and reduced weights compared with standard formula without marine oil (Carlson et al. 1992). Phosphatidylcholine arachidonic acid declined in preterm infants fed on non-supplemented formulas, and weight fell progressively beginning at 2 months of age. The nadir of plasma phosphatidylcholine arachidonic acid and growth was further reduced by formula containing marine oil compared with the non-supplemented formulas (Carlson et al. 1993). Koletzko & Braun (1991) have investigated whether birth weight correlates with the postnatal EFA status in premature infants. A significant and positive correlation between body weight and plasma triacylglycerol content of arachidonic acid and total n-6 long-chain-PUFA was found as well as a negative correlation with a-linolenic acid. There is only limited information about how intakes of n-6 and n-3 long-chain-PUFA may affect specific tissue growth. Animal studies using diets supplemented with both types of fatty acids have shown that in addition to plasma and erythrocyte cell membranes, small intestine, liver, kidney, lung and heart are affected in their fatty acid composition. Depending on the composition of the diet, susceptibility to oxidation may be affected, which might influence tissue growth and function (Suarez et al. 1996a,b). 5.10. Early growth and later obesity In animal studies, early overfeeding may have lasting effects on nutrient utilization and body composition (Davis et al. 1973; Lewis et al. 1986). In male infants of pregnant women who suffered from starvation, obesity in young adulthood was more prevalent if the mothers were exposed to the famine during the first half of gestation, while the incidence of later obesity was reduced if starvation occurred during late gestation and the early postnatal period (Ravelli et al. 1976). Also, food composition, and particularly protein intake, in early childhood has been suggested as a predictor of later risk of obesity. In view of the high prevalence of obesity in Europe and its major importance for public health and health-care costs, the potential modulation of later obesity by early food choice needs to be further explored. 6. Maturation of the gastrointestinal tract 6.1. Introduction Digestion and hydrolysis of macro- and micronutrients by the gastrointestinal tract are essential prerequisites for longterm survival of mammals including man. Proteins, fats and carbohydrates are digested and hydrolysed by a variety of potent excretory glands and by the brush-border enzymes of the small intestine as well as by bacterial breakdown within the large intestine. As for carbohydrates, a cascade of hydrolytic events finally leads to the presence of monosaccharides within the lumen of the gastrointestinal tract which are transported across the microvillus membrane by highly specialized transporters. Carbohydrates with high molecular mass in the form of amylose and amylopectin are hydrolysed by a-amylase (EC 3.2.1.1) of the saliva and the pancreas. a-Amylase can hydrolyse 1-4-a-glycosidic bonds which are present in both amylose and amylopectins. The branches of 1-6-a-glycosidic side chains in amylopectins remain after the action of a-amylase as a-limitdextrins, and are further hydrolysed by sucrase–isomaltase (SI) (Gray, 1967). Hydrolysis of starches is dependent on the age of the infant. In the first 6 months of life, activity of a-amylase is low and reaches full activity at the end of the first year of life (Lentze, 1986). Defects in sugar digestion occur because of disturbances within the combined action of pancreatic a-amylase and that of intestinal brush-border enzymes. Decreased digestion and hydrolysis of carbohydrates will induce either osmotic diarrhoea and/or bacterial overgrowth within the small intestine as well as bacterial breakdown of carbohydrates within the colon. Malnutrition is very often combined with chronic diarrhoea and damage of the gastrointestinal mucosa as a consequence of lack of protein and energy. Key factors in this devastating cascade are the brush–border membrane of the small intestine and its hydrolysing and absorptive capacity. Important observations with regard to the hydrolytic capacity of intestinal disaccharidases, which are responsible for sugar hydrolysis, have come through the study of their intracellular pathways and processing of enzyme molecules in normal and altered human as well as animal mucosa using various techniques of molecular biology. The knowledge of these investigations has considerably increased our understanding as to how carbohydrates are hydrolysed and absorbed from the intestinal epithelial cell. 6.2. Development of sugar hydrolases and transporters The morphological development of the small intestine starts in the 9th week of gestation from the proximal to the distal part of the gut (Hauri, 1986). Small villi develop over a Growth, development and differentiation stratified epithelium of several layers. The first crypts are seen at the age of 10–11 weeks gestation within the duodenum and jejunum, and at 11–12 weeks in the ileum and colon. The changes into a columnar epithelium occur together with the appearance of secondary lumina visible by electronmicroscopy and parallel invagination of mesenchymal cells and extrusion of surface cells (Naim et al. 1988). The development of a brush-border membrane is seen together with crypt development in the 10–12th week of gestation. At the same time brush-border membrane hydrolases start to appear. Lactase–phlorizin-hydrolase (LPH), SI and maltase–glucoamylase are first detectable at the 10th week of gestation (Dahlqvist & Lindberg, 1965). Their enzymic activities increase during gestation. As SI and maltase–glucoamylase reach their full activities by the 25th week of gestation (Jirsova et al. 1965–6), the activity of LPH remains low until the 28th week of gestation and increases slowly between the 32nd and 34th weeks of gestation (Dahlqvist & Lindberg, 1966). For the nutrition of very immature premature babies with very low birth weight between the 26th and 28th weeks of gestation this could play a role in the digestion and hydrolysis of lactose given in breast milk or infant formula. After introduction of lactose-containing milk the activity of LPH matures quickly to normal enzymic activities. The glucose transporters in the small intestine develop during gestation at about the same time as the sugar hydrolases. Sodium-dependent glucose transporter 1 (SGLT 1), glucose transporter (GLUT) 5 and GLUT 2 appear at the 11th week of gestation as seen by the expression of specific mRNA in human fetal intestine (Davidson et al. 1992). 6.3. Biosynthesis of intestinal brush-border membrane hydrolases Mature intestinal epithelial cells are highly polarized and are composed of two main membranous regions: the apical cell membrane, with its unique feature of a brush border, and the basolateral membrane. The microvillar membrane is characterized by a network of microvilli which contain the important glycoproteins responsible for the hydrolysis and absorption of micronutrients and minerals. For the degradation of various sugar and peptide molecules of different composition and chain length, the intestinal disaccharidases SI, maltase–glucoamylase, LPH, trehalase and a variety of peptide hydrolases are present within the microvillar region of the columnar epithelia in order to digest carbohydrate molecules and oligopeptides from nutritional intakes. The disaccharidases are the best studied brush-border hydrolases. Their enzymic activities and their distribution throughout the gastrointestinal tract as well as their age dependency have been investigated by many groups of researchers. The biogenesis of the disaccharidases produced and processed by the mature enterocyte has been elucidated in mammals as well as in man, demonstrating common pathways within the translational and post-translational routes. SI and LPH are the best studied brush-border membrane hydrolases in all species including man. The data accumulated from these studies have led to a general understanding as to how these hydrolases are synthesized and processed S19 within the small-intestinal enterocyte. After transcription a single-chain precursor (pro-SIh) rich in mannose (high mannose precursor) is produced in the rough endoplasmic reticulum. This contains carbohydrate residues which are N-glycosylated and has an apparent molecular mass of 210 kDa in human subjects (Hauri et al. 1980; Ghersa et al. 1986). From the rough endoplasmic reticulum the proSIh is transported to the Golgi apparatus where trimming of the mannose residues and addition of complex carbohydrates occur to yield pro-SIc (molecular mass 245 kDa) (Hauri et al. 1982). The complete primary structure of the pro-SI from rabbit is composed of 1827 amino acid residues containing the two active catalytic subunits isomaltase (140 kDa) and sucrase (120 kDa) which are associated by oncovalent, ionic interactions (Sjöström et al. 1980). After complex glycosylation in the Golgi the pro-SIc is translocated and inserted into the microvillus membrane by vesicular transport directly into the apical microvillus membrane. The exact route for the transportation of glycoproteins from the Golgi to the microvillar membrane remains to be established. The time course of transport of the pro-SIc from the Golgi into the brush-border membrane in a human colon carcinoma cell line (CaCo-2 cells) is rather slow (Herskovics et al. 1981). Similar transport kinetics were also obtained when the biosynthesis of SI was investigated in the organ culture of human intestinal explants (Naim et al. 1988). Insertion of pro-SIc into the microvillus membranes is obtained by anchoring a single hydrophobic segment of the molecule which is located at the N-terminus of isomaltase (Hauri et al. 1986). After insertion into the microvillar membrane pro-SIc is cleaved into sucrase and isomaltase by pancreatic proteases (Naim et al. 1988). The mature catalytic enzymes sucrase and isomaltase cleave various substrates including sucrose, isomaltose, maltose, maltotriose and amylose as well as a-limit dextrins which are derived from the hydrolysis of amylopectins. SI, together with maltase–glucoamylase, plays a major role in starch digestion during the first month of life as a-amylase in human infants is not developed during the first 6 months of life (Danielsen et al. 1981). Striking structural and functional similarities suggest that intestinal SI, human lysosomal a-glucosidase and Schwanniomyces occidentalis glucoamylase are derived from a common ancestral gene (Naim et al. 1991). LPH, as the only b-glycosidase of the brush-border membrane, has been reported in earlier work to be synthesized also as a single-chain precursor with a molecular mass of 150 kDa (Sjöström et al. 1983). However, conflicting results were obtained on the structure and identification of the precursor molecules. In the pig a precursor protein of 200 kDa was observed (Danielsen et al. 1984). The same group reported in a more recent publication that the precursor molecule of LPH in the pig small intestine is a membrane-bound polypeptide of 225 kDa which is intracellularly cleaved after complex glycosylation (Mantei et al. 1988). Similar data were obtained in CaCo-2 cells (Herskovics et al. 1981). In human intestinal epithelial cells a high-mannose precursor of 215 kDa was demonstrated in intestinal explants maintained in organ culture (Hauri, 1986). Here the intracellular cleavage of the highmannose precursor occurs during the translocation of the S20 B. Koletzko et al. molecule across the Golgi before complex glycosylation takes place. The mature form of LPH is then inserted into the membrane with a molecular mass of 160 kDa. The primary structure of the human lactase molecule is known and comprises 1927 amino acids in man and 1926 amino acids in the rabbit (Messer & Kerry 1967). The place at which the mature form of lactase is cleaved from its precursor is position 866 of the whole molecule. In contrast to most other brush-border membrane hydrolases the mature lactase is anchored within the lipid bilayer from its carboxyl end of the protein chain (Messer & Kerry, 1967). Maltase–glucoamylase hydrolyses 1-4-a-glycosidiclinked glucose polymers including maltose and maltotriose (Naim et al. 1989). The enzyme is developed early in gestation and contributes to the digestion of starch after birth. The biosynthesis of maltase–glucoamylase is similar to that of SI as a precursor molecule of 225 and 245 kDa. The former represents the high mannose and the latter the complex glycosylated precursor of maltase–glucoamylase in the pig small intestine (Danielsen et al. 1984). The biosynthesis and processing of maltase–glucoamylase in human intestinal biopsy specimens does not involve intracellular or extracellular proteolytic modifications, in contrast to SI and LPH (Pfeffer & Rothman, 1987). 6.4. Intestinal absorption of glucose and fructose Our current understanding of glucose (galactose) and fructose is that the monosaccharides are transported by different methods into the intestinal absorptive epithelial cells. Whereas SGLT 1 is responsible for the active transport of glucose or galactose with equimolar amounts of Na against a concentration gradient into the cytoplasm of the enterocyte (Crane 1975), fructose undergoes facilitated transport by the GLUT 5 transporter which is also located on the brush-border membrane (Davidson et al. 1992). Once taken up into the enterocyte, Na þ is exchanged with K þ by the Na þ, K þ-ATPase (EC 3.6.1.3) which is located in the basolateral membrane and glucose is pumped into the intracellular space by another glucose transporter protein, GLUT 2. GLUT 2 has also been shown to be localized within the basolateral membrane. The function of SGLT 1 is essential for survival of a given species such as man. When absent or deficient, as in congenital glucose–galactose malabsorption, the malfunction of SGLT 1 is a lethal factor. The absorption of these simple monosaccharides when present in the intestinal lumen is dependent on a variety of factors contributing to the rate of absorption. It is dependent on age, composition of food and species. In mice a relationship between the type of diet and sugar uptake was demonstrated. When fed on a high-carbohydrate, low-protein chow, the uptake of glucose remained high, and dropped considerably when the mice were put on a low-carbohydrate, high-protein diet (Karasov et al. 1983). Moreover, it was shown that a strong correlation exists between glucose uptake and the type of natural diet within various vertebrate species. The more herbivore the species is, the more glucose is absorbed; the more carnivore the species is, the less glucose is taken up (Riby et al. 1993). Fructose absorption depends strongly on the presence of other carbohydrates within the intestinal lumen. Fructose given together with glucose, galactose, sucrose or starch is better absorbed than fructose alone (Fujisawa et al. 1991) whereas the presence of sorbitol or dextrin leads to fructose absorption as with fructose alone. Species differences have a considerable influence on fructose absorption which depends entirely on the composition of the natural diet. Whereas the carnivores (cats) have a low fructose absorption after weaning, the rat and the rabbit increase their fructose absorption considerably after weaning (Buddington & Diamond, 1989). The complete mechanism for fructose absorption from the human intestine remains to be elucidated. A model of fructose absorption can be deduced from animal studies as well as from human studies. The rate of fructose absorption is influenced by glucose, but also by the amino acid glycine. Two mechanisms have been proposed to explain this effect. Fujisawa et al. (1991) speculate that a mechanism in the brush-border membrane exists which they call the disaccharidase-related transport system. This speculation is based on their findings that the specific inhibitor of SI, acarbazone, decreases fructose absorption when given with sucrose, whereas absorption increases without the inhibitor. Another explanation could well account for this effect: glucose absorption as well as glycine absorption increases the water flow from the lumen into the intercellular space by osmosis. Therefore, a solvent drag occurs leading to an enhanced uptake of fructose. The same mechanism applies when starches or sucrose are present at the same time as fructose because of the rapid hydrolysis of these carbohydrates to glucose and/or fructose by the action of maltase– glucoamylase and SI. In order to elucidate this effect fructose absorption should be studied in individuals with SI deficiency which would be the equivalent model to the rat intestinal fructose uptake studies with acarbazone. Except for LPH, the activities of other sugar hydrolases as well as SGLT 1 can be influenced by substrates (Buddington et al. 1991; Quan & Gray, 1993). The close presence of sugar hydrolases and sugar transporters within the microvillar membrane of the small-intestinal enterocyte as well as the transporters for amino acids is the guarantee of a steady uptake of sugars derived from various sources. 6.5. Oligosaccharides and mucins Besides the most abundant sugar lactose, human milk contains more than 130 different oligosaccharides and is unique among all mammalian species for its content of higher oligosaccharides, i.e. larger than lactose. For a long time the oligosaccharide fraction in human milk has been overlooked although it is the third largest solute (up to 18.5 g/l) and present in higher amounts than protein (Egge et al. 1983). As oligosaccharides escape the hydrolysis in the small intestine, two possible functions are discussed. One function would be the intact absorption of these components serving as substrates for organ maturation such as the brain, where rapid synthesis of sialoglycoproteins and gangliosides occurs (Sabharwal et al. 1991). Their role in the large bowel as ‘dietary fibre’ and their fermentation would be of significant value in nutrition. The oligosaccharides in human milk are based on five monosaccharide residues: sialic acid, N-acetylglucosamine, fucose, glucose Growth, development and differentiation and galactose. All oligosaccharides possess a lactose moiety at their reducing end, with sialic acid (when present) and fucose at the non-reducing end. The chain length varies between three and eleven units. The oligosaccharide composition of human milk shows temporal and individual variations (Miller et al. 1994). At the same stage of lactation the variation in oligosaccharide content was shown to be fourfold. Recent findings on the chemical structure of oligosaccharides in milk have demonstrated structural homologies to carbohydrates carried by glycoproteins and glycolipids on cell surfaces. Such oligosaccharides are very antigenic and were targets of monoclonal antibodies in the search for specific binding to human cancer cells. Similar novel oligosaccharides which can inhibit antigen– antibody reactions have been detected in human milk (Fievre et al. 1991; Kitagawa et al. 1991). They are useful hapten inhibitors to study the binding specificities of anticarbohydrate antibodies produced as mucins on cancer cells. This observation has great clinical implications with respect to breast cancer. High-molecular-mass glycoproteins (MUC1) in milk and lactating tissue have been found to contain up to 80 % of carbohydrates. The low level of expression of MUC1 in healthy, undifferentiated (nonlactating) breast tissue, and its presence in many, particularly metastasizing, breast tumours has established a very useful marker in breast cancer screening. The exact nature of MUC1, its biological role and expression, has been studied in milk and mammary tissue (Patton et al. 1995). MUC1 as expressed in tumours activates B- and T-lymphocytes. These epitopes represent underglycosylated forms of MUC1 characteristic of breast and pancreatic cancer. It is also responsible for keeping ducts and lumens, such as the mammary ducts, open. Here it binds also to L-selectin which is expressed on the surface of leucocytes. By this action leucocytes are bound to the lumen and excreted into the milk (Welply et al. 1994). It also escapes digestion and is excreted in the stools of breast-fed infants. In the colon it binds to micro-organisms, particularly to the fimbrins of E. coli, contributing to the host defence of the breast-fed infant (Cravioto et al. 1991; Schroten et al. 1993). 6.6. Probiotic substances in milk or milk substitutes There are indications that certain ingested micro-organisms added to milk or milk products may exert some physiological effects and promote health in human infants. Such an example is the use of Lactobacillus GG added to the milk formula for premature infants. After administration of these bacteria to premature infants it was noted that the bowel was colonized by Lactobacillus GG, and no clinical side-effects were seen. However, no clinical benefit was noted either (Millar et al. 1993). As far as potential effects on fermentation are concerned, Lactobacillus GG given to premature infants had no effect on production of short-chain fatty acids in stools. The observed small increase in ethanol excretion is unlikely to have any clinical significance (Stansbridge et al. 1993). Whether or not pre- and probiotics modulate gut maturation and have relevant health benefits in infancy remains to be elucidated. S21 6.7. Dietary regulation of xenobiotic metabolism The nutritional status or specific nutrients may influence the metabolic capacity of the liver, but also of other organs (intestinal mucosa). The exact mechanism of drug–nutrient interaction remains unknown. Moreover, the health effects of such interactions have to be explored. Diet could modulate the metabolism either through an action on substrate availability or by modifying key enzymes of metabolism. The influence of starvation on conjugation and glucuronidation illustrates this point (Mandl et al. 1995). One important discovery of recent years is that some components of food (antioxidants like butylated hydroxytoluene) are able to modulate the activity of key enzymes of phase 1 or 2 metabolism, by modifying gene expression (Kashfi et al. 1994). This is a promising area that requires further work. In fact the modulation of gene expression by nutrients is well described in the context of the influence of food on carbohydrate or lipid metabolism but the effect of macro- or micronutrients on xenobiotic-metabolizing enzymes (XME) remains unelucidated. Diet may also influence XME by inducing specific pathology e.g. steatosis. The part played by nutrients and morphological alteration in the modification of XME has to be established (Leclercq et al. 1996). Specific components of food (n-3 PUFA), even at very low concentrations (piperine from black pepper, naringenine from grapefruit) may modify the activity of specific isoforms of XME (e.g. glucuronosyltransferases) (Speck et al. 1991). This could lead to interesting developments in several fields. (1) Fundamental research: it will help in studying this metabolic reaction in detail by using those nutrients as activators or specific inhibitors. (2) Those nutrients which can be used at low dose could, thus, be considered as ‘toxico-modulators’. (3) Such compounds have been proposed as therapeutic adjuvants, allowing reduction of the dose of expensive drugs, or drugs with a low safety-therapeutic index. The discovery of new nutrients with new targets could constitute a very promising area. Finally, the validation of experimental models (and particularly in vitro models) allowing the study of drug–nutrient interactions would give a new input in this area. 7. Development of the immune system 7.1. Introduction Positive effects of particular foods or food ingredients on the human immune system (e.g. inhibition of CHD and cancer development etc.) could conceivably be related to early nutritional events or may only be seen after decades of intake or lifestyle changes. An important unresolved issue is whether there are critical time periods during which a provision may be especially beneficial to the immune system. In view of these difficulties, in vitro surrogate markers for study end-points are frequently used. A surrogate end-point can be defined as a laboratory measurement or a physical sign used as a substitute for a meaningful endpoint that measures directly how a patient feels, functions or survives. Changes to a surrogate end-point induced by a dietary intervention are expected to reflect changes in a S22 B. Koletzko et al. biologically meaningful end-point. However, surrogate endpoints do not always reflect the true clinical outcome and can be misleading or even meaningless (Fleming & DeMets, 1996). Other important limitations of studies reporting effects on the development of immunity are: short duration; lack of standardized tests; lack of correlation of in vitro and/or in vivo findings with immune protection or immune suppression; lack of demonstration of health-enhancing effects in a developing normal, presumably non-deficient population of infants, but with changing nutritional requirements. 7.1.1. Which constituents of the immune system to investigate? The immune system is a highly complex regulatory cellular and humoral system of protection and stimulation directed to avoid danger to the host. On this basis a possible role of diet in cancer prevention could be taken as summary evidence for beneficial effects of the diet on the immune system. This extrapolation, however, seems highly conjectural since it would argue that a number of cancers are due to deficient immune surveillance mechanisms. This may or may not be the case and a number of other protective mechanisms could be equally plausible. The present report focuses separately on published evidence mostly in infants and children and, where appropriate, in normal individuals and animals. Parenteral micronutrient supplementation in disease states or after surgery, low level toxicity and multiple chemical sensitivities will not be addressed. 7.1.2. Special considerations for the immune system of the developing child. There is a dearth of reliable information of the effects of vitamins, saturated and unsaturated fatty acids, trace minerals and other normal food constituents on the infant’s developing immune system. The majority of reports examine the effects of corrections of severe or moderate deficiencies on the immune responses in infants and children (or animals). Very little is known of the effects of supplementations, either in line with recommendations or above, in a non-deficient population. As indicated earlier, most measurements have been related to surrogate endpoints and the relationships between administration of test substances and their effects on the immune system are not strongly causal. 7.2. Antioxidants and vitamins 7.2.1. In general. Antioxidant vitamins generally enhance different aspects of cellular and non-cellular immunity. The antioxidant function of these micronutrients could, at least in part, enhance immunity by maintaining the functional and structural integrity of important immune cells. Multiple effects attributed to antioxidants include risk reduction of a variety of chronic diseases (Messina & Messina, 1996), anti-(retro)-viral activity (Formica & Regelson, 1995), immune enhancement in animals (Forni et al. 1986; Gebhard et al. 1990, ), in man (Chavance et al. 1989; Penn et al. 1991; Rall & Meydani, 1993), and evidence that certain vitamins alone or in combination and other micronutrients given in levels above the current recommendations have critical, beneficial effects on human immune responses (Bendich, 1995). 7.2.2. Vitamin A. The effects of vitamin A supplementation on measles in deficient and non-deficient children have been the subject of several recent reports (Coutsoudis et al. 1992, 1995; Rosales & Kjolhede, 1994; Semba et al. 1995; Stabell et al. 1995; Keusch, 1996). In general, in the last decade epidemiological, immunological, and molecular studies have yielded substantial evidence for a central role. The recent discovery of retinoic acid and retinoid X receptors has provided a molecular basis for the action of vitamin A and its metabolites at the level of gene activation. b-Carotene supplementation enhances the expression of functionally associated molecules on human monocytes (Hughes et al. 1996) and also enhances immune responses to poor immunogens, which may be relevant to infants receiving vaccines which are characterized by low seroconversion rates (Semba, 1996), however under certain conditions seroconversion rates may be negatively affected too (Semba et al. 1995). No difference was found (using whole-blood culture techniques) in the in vitro proliferative responsiveness of T-cells to concanavalin-A and tetanus toxoid of children with normal or low–normal concentrations of vitamin A or Zn (Kramer, 1996). 7.2.3. Vitamin C. Vitamin C has gained great scientific and media attention through the promotions of its effects on the common cold. Early reports of systemic conditioning of infants following an increased intake of vitamin C during development seem to have been unfounded (Gerster & Moser, 1988). Studies which failed to identify a positive effect on the symptoms of the common cold have recently been critically reviewed and it now seems likely that there is a reduction in clinical symptoms associated with intake of 2–3 g ascorbic acid/d at the onset of the cold (Hemila, 1996). 7.2.4. Vitamin B complex. Few studies have addressed the effects of supplementation with individual B vitamins and/or the coenzyme CQ10. When studied, positive enhancing effects on cell-mediated immunity, CD4 : CD8 ratios, delayed type hypersensitivity and antibody production have been reported (Gebhard et al. 1990; Miller & Kerkvliet, 1990; Folkers et al. 1993). Tumour inhibition by high dietary pyridoxine may be mediated by immunological mechanisms that are lacking in the genetically immunodeficient (athymic) mice in which these studies have been carried out (Gebhard et al. 1990). 7.2.5. Vitamin E. Vitamin E, in its role as a potent antioxidant and immunostimulant, has received a great deal of attention (Tengerdy, 1990; Shor Posner et al. 1995; Liang et al. 1995; Finch & Turner, 1996; Liang & Watson, 1996). Vitamin E supplementation enhances humoral and cellmediated immunity, and augments the efficiency of phagocytosis in laboratory animals and human subjects. Vitamin E deficiency has been suggested to contribute to alterations of the neonatal neutrophil function, and supplementation with 120 mg/kg over the first 14 d of life of healthy premature infants has been shown to increase phagocytosis (Chirico et al. 1983). 7.2.6. Vitamin D. There is now increasing evidence that the hormonal form of vitamin D, 1,25-dihydroxycholecalciferol (1,25(OH) 2D 3), is involved in the regulation of the Growth, development and differentiation immune system. 1,25(OH) 2D 3 exerts most of its actions after it has bound to its specific receptors which are present in monocytes and activated lymphocytes. The hormone inhibits lymphocyte proliferation and immunoglobulin production in a dose-dependent fashion. It interferes with T-helper cell function, reducing T-helper cell-induction of immunoglobulin production by B-cells and inhibits the passive transfer of cellular immunity by T-helper cells in vivo. Expression of Class II antigen by lymphocytes and monocytes is also affected. In experimental in vivo studies 1,25(OH) 2D 3 is particularly effective in preventing autoimmune diseases (Schwartz, 1992; Mathieu et al. 1994; Thomasset, 1994). There is no information on the effects of vitamin D or its metabolites on the developing human immune system. 7.3. Multiple micronutrient supplementation studies Trace elements perform important functions in growth and development. However, little information exists about dietary requirements of them during the demanding period of infancy. Although several factors influence the dietary needs of these essential elements, the basis for establishing dietary needs in infants is hindered by the dearth of studies that have assessed their bioavailability and effects on the immunity in this age group. Thus, until it has been conclusively shown otherwise, the physiological response to human milk is used as the standard for infant feeding practices (Milner, 1990). Key questions such as the risks to human health of altered environmental distribution of Zn, assessment of Zn status in man, effects of Zn status in relation to other essential metals on immune function, reproduction, neurological function and others remain (Sherman, 1992; Walsh et al. 1994; Prasad, 1995; Ripa & Ripa, 1995; Sazawal et al. 1996). Zn supplementation has recently been shown to reduce persistent diarrhoea in children (Sazawal et al. 1996). In vitro and in vivo studies show that antioxidants generally enhance different aspects of cellular and non-cellular immunity. The antioxidant function of these micronutrients could, at least in part, enhance immunity by maintaining the functional and structural integrity of important immune cells (Chew, 1995). Effects of trace minerals on the outcome of pregnancy have been reviewed. About 30 % of pregnant women suffer from Fe deficiency worldwide, and while its effects on neonatal Fe status are not severe, adverse sequelae include impaired neonatal immune status (Allen, 1986; for review, see Bryan & Stone, 1993). Low maternal intakes of Cu, Mn, and Se have not been associated with adverse outcomes of pregnancy. Se deficiency, however, appears to result in immunosuppression affecting neutrophil function, antibody production, cytotoxicity and lymphocyte proliferation (Kiremidjian Schumacher & Stotzky, 1987). 7.4. Fatty acids Several lines of evidence support the role of dietary lipids as regulators of the immune system. This is demonstrated by studies examining lipid alteration of the immune response to S23 allergens, malignancy, autoimmune disease, sepsis, trauma, and transplantation. Both the quantity and quality of lipid are important in immunoregulation. Both cell-mediated and humoral immunity are affected by dietary lipids. Multiple mechanisms probably contribute to the overall effects of lipids, including alteration of arachidonic acid metabolism, changes in cell membranes, production of inflammatory cytokines, and impairment of the reticuloendothelial system (Perez & Alexander, 1988; Yetiv, 1988; Fernandes et al. 1990; Melnik et al. 1991; Watanabe et al. 1994; Endres, 1996; Hellerstein et al. 1996). 7.5. Arginine Many of the known roles of arginine (e.g. in immune function, wound healing, and protection against NH 3 intoxication) are mediated by a metabolic pathway synthesizing NO in the liver. Both stimulatory and suppressive functions have been identified with a prominent role of the macrophage (Barbul, 1990; Rodeberg et al. 1995; Suzuki et al. 1995; Krenger et al. 1996; Marcinkiewicz et al. 1996). Particular effects on the developing human immune system are uncertain. Oral administration may be less effective than parenteral administration (Torre et al. 1993). A number of reports investigate the potential benefits of arginine supplementation during surgical and other periods of stress in human subjects and in experimental models. Beneficial effects on wound healing, reduction of postoperative septicaemias etc. cannot always be attributed entirely to arginine, since it was often given with other dietary modulations (Daly et al. 1990; Cerra et al. 1991; Seidman et al. 1991; Kemen et al. 1995; Senkal et al. 1995; Braga et al. 1996; Kudsk et al. 1996; Marcinkiewicz et al. 1996). 7.6. Nucleotides Dietary sources of preformed purines and pyrimidines seem to be important for optimal function of the cellular immune response. It was previously assumed that nucleotides were not needed for normal growth and development, but the results described in the present review demonstrate a need for nucleotides in the response to immunological challenges. The need for sources of preformed nucleotides in defined formulas such as parenteral and enteral formulas and infant formulas is suggested in some studies (Carver, 1994; Kulkarni et al. 1994; Rudolph, 1994). An exogenous source of nucleotides from the diet may optimize the function of rapidly dividing tissues when growth is rapid and the diet is low in nucleotides. Studies performed in human infants are, at most, inconclusive (Carver, 1994; Kulkarni et al. 1994) and further studies are required to assess in infants the interesting findings about dietary nucleotides reported in experimental models (Gil & Uauy, 1989; Sanchez Pozo et al. 1994, 1995; Ortega et al. 1995; Lopez-Navarro et al. 1996; Navarro et al. 1996). 7.7. Maturation of the immune system in formula-fed v. breast-fed infants In a small study of systemic and secretory immunity of breast-milk-fed v. formula-fed infants it was shown that S24 B. Koletzko et al. there is a general stimulation of responsiveness by cytokines in milk and a reduction of specific responses by antigen exclusion. Bottle-fed infants reached a similar level of immunological maturity by 3 months of age, exhibiting raised levels of serum antibodies against gut organisms and milk proteins and demonstrating increased non-specific activation of lymphoid cells (Stephens, 1986; Stephens et al. 1986a,b) 7.7.1. Effects of antigen transfer via breast milk on the infant’s immunity. Dietary antigen excretion into breast milk seems to be a general phenomenon and has been reported for milk, egg, wheat proteins and parasite antigens (Kilshaw & Cant, 1984; Troncone et al. 1987; Petralanda et al. 1988). Excreted amounts are in the range of mg/l. The immunological significance of transfer of dietary antigens during breast-feeding is still unclear. It is generally accepted that breast-feeding reduces the risk of food allergic reactions and also of atopy in a population at risk (uni- or biparental history of atopy) but sensitizing effects in infants have also been described (Warner, 1980; Gerrard & Shenassa, 1983; Savilahti et al. 1987; Lindfors & Enocksson, 1988). Studies by Chandra et al. (1986, 1989a), Zeiger et al. (1992) and others (Halken et al. 1993a; Vandenplas et al. 1995) suggest that elimination of (significant) dietary antigen transfer via breast milk for 6 months (amongst other preventive measures) in a population at risk reduces the probability of a food-specific sensitization (and possibly atopic symptoms) for up to 48 months, an effect which persists even after the diet of the infant has been liberalized. The following points merit special consideration. 7.7.2. Maternal diet during pregnancy and effects on the infant’s immunity. There are no reports which demonstrate any preventive (or sensitization) effects in infants with a parental history of atopy (Fälth-Magnusson et al. 1987; Lilja et al. 1989; Fälth-Magnusson & Kjellman, 1992). 7.7.3. Maternal diet during pregnancy and lactation. In view of the complexity of the studies and the number of confounding variables, it is not entirely surprising that a number of studies have come to different conclusions. Maternal diet, while continuing breast-feeding, has been shown to be of moderate benefit in the reduction of atopic manifestations (mainly eczema) and milk allergy in infants or children (Chandra et al. 1989a,b; Sigurs et al. 1992; Zeiger et al. 1992; Zeiger & Heller, 1995). Other studies have failed to show this effect (Lilja et al. 1989). 7.8. Role of the gut flora and probiotic bacteria in the infant’s immunity and gut defence Little is known about the immunomodulating capacity of the first bacteria colonizing the gut. Breast-fed infants develop a typical intestinal flora and this has been linked to a certain resistance to enteric infections. However, breast milk contains a host of other immunomodulating factors and it is difficult to claim any causality in these studies. Infants given breast-milk substitutes with various strains of lactic acid-producing bacteria may also exhibit some resistance to infections. In a small clinical study of thirty-nine children, protective effects of the supplementation of an infant formula with oligosaccharides, fermented milk or lactic acidproducing bacteria on the reduction of the incidence of acute diarrhoea were reported. A randomized controlled feeding trial with Bifidobacterium breve in ninety-one very-lowbirth-weight infants demonstrated effective colonization, fewer abdominal signs and better weight gain (Kitajima et al. 1997). An enhancement of the circulating antibodysecreting cell response was observed in infants with rotavirus diarrhoea supplemented with a strain of Lactobacillus casei, compared with a placebo group (Kaila et al. 1992). The duration of this response and other protective or longerterm effects are unknown. Other small studies reported an enhancement in the phagocytic activity of granulocyte populations in the blood of human volunteers after consumption of fermented milk with Lactobacillus acidophilus and B. bifidum (Schiffrin et al. 1995). It is unresolved whether studies in a larger number of unselected infants under different conditions in different countries would yield similar encouraging results. The gut microflora is an important constituent in the intestine’s defence barrier. Probiotic bacteria have been suggested to affect different aspects of gut defence: immune exclusion, immune elimination and immune regulation. 7.8.1. Immune exclusion and elimination. Although many clinical benefits have been ascribed to consumption of candidate probiotic strains in gastrointestinal disease (Isolauri et al. 1991; Saavedra et al. 1994), only a few human studies have assessed the effects of these bacteria on gut defence mechanisms. Early reports associated clinical observations with the effects on the intestinal microflora (Niv et al. 1963). Oral bacteriotherapy affected microbial imbalances shown during rotavirus infections in infants (Isolauri et al. 1994). Oral introduction of probiotic microorganisms has been associated specifically with reduction of intestinal inflammation (Majamaa & Isolauri, 1997) and an increase in circulating antibody-secreting cells in the serum as an indicator of the intestine’s immunological barrier function (Kaila et al. 1992). In children with rotavirus diarrhoea, probiotic bacteria administered during the diarrhoeal phase of the infection promoted clinical recovery and enhanced intestinal immunoglobulin A (IgA) responses (Kaila et al. 1992). 7.9. Effects of formulas with protein hydrolysates on the infant’s immune responses Extensively hydrolysed casein formula has been used in the treatment of children with cow’s milk protein allergy and/or intolerance. Recently, ultrafiltrated whey hydrolysates have been also been used therapeutically (Halken et al. 1993b). In an attempt to prevent and/or modulate the risk of developing food-allergic and atopic manifestations in infants and children, less extensive (partial whey hydrolysates) (Chandra, 1991; Vandenplas et al. 1992, 1995) and extensively hydrolysed formulas (Chandra et al. 1989a; Zeiger et al. 1989; Halken et al. 1993a; Zeiger & Heller, 1995; Oldaeus et al. 1997) have been used. Although these studies have been performed in infants of different (atopic) family and ethnic background, with different nutritional habits and different measures of control for confounding variables and a lack of standardized diagnostic protocols, they could be summarized as follows. (1) Infants with a high-risk family Growth, development and differentiation background of atopic disease are likely to benefit from exclusive breast-feeding for 4–6 months with some added benefit if the mother avoids certain foods such as milk, eggs, fish and possibly nuts (including peanuts) during lactation. The benefits include a reduction in the incidence of cow’s milk and food allergies and atopic eczema for up to 4 years. (2) If exclusive breast-feeding for 4–6 months cannot be sustained, the use of a hydrolysed infant formula may help reduce the overall incidence of atopic manifestations in the child at risk. (3) Preventive effects of hydrolysed formulas in infants with a normal risk of developing atopic manifestations and the respective benefits of extensively v. less extensively hydrolysed infant formulas need to be further evaluated. 7.10. Insulin-dependent type 1 diabetes mellitus and cow’s milk exposure in infancy Insulin-dependent diabetes mellitus (IDDM) is considered to be a chronic autoimmune disease characterized by gradual b-cell destruction mediated by autoreactive T-lymphocytes during an asymptomatic prediabetic phase of varying duration (Knip, 1992). In a Finnish study, associations of infant feeding patterns and milk consumption with cow’s milk protein antibody titres were studied in newlydiagnosed diabetic children, sibling-control children and birth-date- and sex-matched population-based control children. Inverse correlations were observed between the duration of breast-feeding, or age at introduction of dairy products, and antibody titres. High IgA antibody titres to cow’s-milk formula were associated with a greater risk of IDDM both among diabetic-population-control and diabetic-sibling-control pairs. The results suggested that young age at introduction of dairy products and high milk consumption during childhood increase the levels of cow’s milk antibodies and that high IgA antibodies to cow’s milk formula are independently associated with increased risk of IDDM (Vaarala et al. 1996). Similar associations between bovine serum albumin antibodies and onset of IDDM have also been found in a low-incidence French population (Levy Marchal et al. 1995). Human T-lymphocyte cultures of cells taken from affected children allowed the detection of bovine serum albumin-specific T-cells which were mapped to the ABBOS peptide (pre-bovine serum albumin position 152–169) previously identified as a possible immunological mimicry epitope which could explain the cross-reactivity with pancreatic islet cell antigens (Cheung et al. 1994). A currently ongoing prospective dietary intervention trial in children genetically at risk will be able to address the causality of this highly intriguing association and, it is hoped, open the way for a primary nutritional prevention strategy. 8. Bone growth and mineralization 8.1. Cell biology of bone growth Bone growth and mineralization is an ongoing process during human fetal and postnatal development, stabilizing at about 21 years of age. Skeletal Ca content increases from 30 g in the neonate to 1200 g in the adult, and skeletal P S25 from 17 to 700 g. Bone tissue possesses a series of enzymic mechanisms that permit mineralization of its extracellular matrix. This matrix is composed of collagen, proteoglycans and other non-collagen proteins in which insoluble mineral salts of hydroxyapatite and small amounts of other salts of Mg, sodium carbonate and citrate are deposited, converting it into a structure capable of supporting the organism. The two most important bone cell types are osteoblasts and osteoclasts. Osteoblasts are responsible for the formation and organization of the extracellular matrix and its subsequent mineralization. Osteoclasts are large motile multinucleated cells, located on bone surfaces, responsible for the resorption of bone matrix. Bone growth or bone modelling is the result of two processes: first formation of new bone and then resorption to maintain the same structural form with, as a net result, acquisition of bone mass. At the age of about 18 years, both male and female adolescents have reached 95–99 % of their individual peak bone mass. After adolescence the processes of bone resorption and formation become quantitatively in balance and this situation is referred to as bone remodelling. After 35–40 years of age the processes of bone resorption and formation become uncoupled and net bone loss will occur, eventually leading to osteoporosis (Price et al. 1994; Anderson, 1996a,b). The obvious strategies to prevent, or at least delay, the onset of osteoporosis include: (a) optimizing the attainment of peak bone mass in adolescents, and (b) preventing bone loss in later life. 8.2. Methodological aspects in bone-mass-related studies In the interpretation of the bone-mass-related results of the various studies one needs to be aware of the actual technique used. The first non-invasive methods for measuring bone mass were based on quantitative evaluation of standard radiographs. During the last decade gamma- or X-ray techniques were developed based on the variable effect of matter on the passage of radiation. Most studies were performed using either single photon absorptiometry (SPA) or dual-energy X-ray absorptiometry (DEXA). SPA is a relatively simple technique using 125I as the radioactive source. Its application is limited to the peripheral skeleton, particularly the radius. DEXA is becoming more and more the preferred technique to assess bone density at various sites of the skeleton with only minimal radiation exposure. Although calibration of DEXA instruments seems to be a rather trivial exercise, it relies strongly on highly specific software, with the results that measurements on one patient made with instruments of different brands do not necessarily result in similar bone mineral density data (Slosman et al. 1995). An additional feature of DEXA is its ability to measure whole-body mineral content as well as body composition. Appropriate DEXA software for infants has been developed recently and reference values from the first studies are now beginning to become available in the literature (Rigo et al. 1996). The fact that SPA and DEXA data cannot be compared with each other is illustrated by the fact that SPA produces a measurement of bone mineral content (g/cm) and assumes that the site of measurement is a small cylinder of constant width, whereas DEXA produces a measurement of bone mineral density (g/cm2) by correcting S26 B. Koletzko et al. Table 1. Additional increment in bone mineral density (BMD) (as a percentage) following supplementation of the diet with calcium or dairy products, in children and adolescents from five different studies (From Kerstetter, 1995) Johnston et al. (1992)* Prepubertal Subject no., total Sex Entry age (years) Intervention duration (months) Baseline Ca intake (mg/d) (placebo) Total Ca intake (mg/d) (diet þ supplement) Supplemental Ca source BMD determination Change in BMD‡ Midshaft radius Distal radius Lumbar spine Femoral neck Ward’s triangle Greater trochanter Total body 22 twin pairs Boys and girls 6.9 (SD 1.4) Pubertal Lloyd et al. (1993) Lee et al. (1994) 248 Girls 48 Girls 11.9 (SD 0.5) 18 935 162 Boys and girls 7.2 (SD 0.2) 18 280 11.4 (SD 0.8) 6 888 11.1 (SD 1.0) 12 728 1612 1370 580 1315 1618 1437 Ca citrate malate DEXA Ca citrate malate DEXA CaCO 3 SPA Ca citrate malate DEXA Dairy foods DEXA +5.1 +3.8 +2.8 +1.2 +2.9 +3.5 23 twin pairs Boys and girls 10.6 (SD 2.0) 36 908 94 Girls Andon et al. 1994)† Chan et al. (1995) −0.1 +2.9 −1.0 −0.4 −0.4 +0.2 +2.9 +3.1 +1.3 NS +9.9 NS +1.0 +2.2 +6.6 DEXA, dual-energy X-ray absorptiometry; SPA, single photon absorptiometry. p Two age groups were studied: prepubescent (6.9 (SD1.4) years) and pubescent (10.6 (SD2.0) years). † Two levels of dietary Ca (1315 and 1618 mg/d) were studied. ‡ Change in BMD = percentage increase in supplemented group minus increase in unsupplemented group. the bone mineral content for the projected area of bone (Slosman et al. 1995). A shortcoming of the usual expression of bone mineral density obtained by DEXA (g/cm2) is that this areal bone mineral density does not take the agerelated increase in bone thickness into account. Therefore Cowell et al. (1995) have developed a measure of true bone mineral density, volumetric bone mineral density (g/cm3), and demonstrated its usefulness in assesssing patients with phenylketonurea (PKU), chronic renal failure and chronic asthma. 8.3. Peak bone mass and relative risk of osteoporosis The relative importance of peak bone mass on the subsequent risk of osteoporosis has recently been reviewed by Ribot et al. (1995). They started with the available in vitro evidence on the relationship between low bone mass and the risk of osteoporosis relating to the mechanical properties of bone. The relevant in vivo studies include both crosssectional surveys and at least eleven prospective studies. From these studies it can be deduced that the relative risk of osteoporosis for each 1 SD decrease in bone mineral density is increased by a factor of between 1.7 and 2.7. On comparing this value with the relative risk of CHD for a 1 SD rise in serum cholesterol or that of stroke for 1 SD increase in blood pressure being 2.1 and 1.3 respectively, it is clear that low peak bone mass is a very strong risk factor for later osteoporosis. Another appealing value relating to the relevance of optimizing or increasing peak bone mass can be deduced from the study by Gilsanz et al. (1991) on comparing the development in peak bone mass in white and black girls. They found a 10–20 % higher bone density in black girls relative to white girls which is likely to correspond to an additional 10–20 years of protection against the decline in skeletal mass, and might explain the relatively low prevalence of osteoporosis in black women. About 80 % of the variance in bone mineral density is accounted for by genetic factors (Pocock et al. 1987), thus leaving only 20 % of the variance to be influenced by environmental factors such as diet. For this reason the investigation of twin pairs is very attractive because the genetic bias may thus be minimized (Johnston et al. 1992). From a study with postmenopausal twins in Britain, Spencer et al. (1995) reported a genetic linkage between the vitamin D receptor genotypes and bone mineral density. The degree to which this might explain the genetic factors is as yet unclear and a recent study from Denmark failed to find any significant association between common allelic variations at the vitamin D receptor locus and bone mineral density (Jørgensen et al. 1996). 8.4. Bone growth and mineralization in infants and young children The literature on the effects of different diets during infancy on bone mineral content at 2 or 5 years of age is not yet consistent. Exclusive feeding of breast milk during the first 6 months of life supports a bone growth considered adequate even though breast milk contents of Ca and vitamin D are relatively low. Infant formulas in general, and formulas for premature infants in particular, contain higher levels of these nutrients to meet the dietary requirements and to provide a safety margin to correct for a likely lower bioavailability. Surprisingly, Bishop et al. (1995) reported a strong positive association between the amount of human milk consumed and bone mineral content at the age of 5 years in their multi-centre cohort of prematurely born infants. The authors raised two possible hypotheses to interpret their finding. Bone mineral depletion in preterm infants fed on unsupplemented human milk might ‘programme’ these infants to be conservative with bone mineral and to reduce the overall rate of growth so that Growth, development and differentiation ‘over-mineralization’ occurs at a later stage when the intake of bone mineral substrates is normal. A second possibility is that one or more of the human milk growth factors might survive breast milk pasteurization and the immature digestion system and end up via the circulation at the target organ. Until one of these hypothetical mechanisms is further substantiated, the general goal in the nutrition of premature infants remains to provide enough mineral supplementation to allow attainment of bone mineral content comparable to that accrued in utero and to support catch-up growth in the first year of life. Several studies comparing bone mineralization in term-born breast-fed infants with that in infants fed on formulas containing moderate or high Ca content conclude that the bone mineral content of formula-fed infants is higher at the ages of 2 and 5 years (Demirini & Tsang, 1995). Whether this effect of cumulative Ca intake during the first 2 years of life will be retained until adolescence is still unclear. A negative impact on bone growth and mineralization has been reported for a number of chronic conditions during infancy and early childhood like cystic fibrosis, IDDM, cerebral palsy, leukaemia, renal disease, growth hormone deficiency and anorexia nervosa (Shaw & Bishop, 1995). In a group of fifty-five children with milk allergy showing a broad distribution of daily Ca intake (quartiles: 409, 663, 950 and 1437 mg Ca/d), a clear correlation was found between Ca intake and bone mineral density, thus underlining the vulnerability of this group and illustrating the efficacy of dietary measures or supplementation of Ca to achieve normal bone growth (Henderson & Hayes, 1994). 8.5. Calcium supplementation in children and adolescents and bone health Despite the relatively poor contribution diet is supposed to make to the variance of peak bone density, a substantial number of studies have been published in recent years to address the effects of nutrition, particularly of Ca, on bone density in children and adolescents. These studies have been reviewed by Kerstetter (1995) who concluded that the cross-sectional and correlation studies have yielded rather mixed results. Surprisingly, more consistent findings were obtained from the five recent prospective Ca or dairy supplementation studies in children and adolescents published between 1992 and 1995. In Table 1 the basic data from these studies (Johnston et al. 1992; Lloyd et al. 1993; Andon et al. 1994; Lee et al. 1994; Chan et al. 1995) are compared. In all studies the baseline Ca intake was less than 1000 mg/d and the amount of supplemented Ca ranged from 300 to 700 mg/d. The percentage increase in bone mineral density in all the supplemented groups amounted to 1–10 % and was significant in all studies. Of those intervention studies, the one by Chan et al. (1995) is most appealing because in this study a near doubling of the Ca intake (from 728 to 1437 mg/d) was reached only by supplementation with dairy products and the increase in total body bone mineral density amounted to 7 %. However, the conclusion which one might draw on superficial reading of the review by Kerstetter (1995) that dairy products outperform other Ca supplements cannot be substantiated yet. Nevertheless, it is remarkable that in the study by Chan et al. (1995) the S27 supplemented dairy products also contained some vitamin D and extra P, which can at least be interpreted as showing that P has no negative effect on peak bone mass accretion in adolescents under these conditions. In conclusion, the results from the Ca and dairy-product supplementation studies summarized have clearly demonstrate that it is possible to increase peak bone mass at the end of adolescence simply by dietary means. 8.6. Nutrients other than calcium and environmental factors involved in bone growth The crystal salt of bone resembles hydroxyapatite (Ca10(PO4)6(OH)2) which contains Ca and P in the proportion 2.15:1 (w/w); in addition approximately 60 % of the body Mg and 30 % of the body Zn are present in the skeleton. It is, therefore, obvious that P, Mg and Zn are also important nutrients in the process of bone mineralization. Although there is little information about conditions in developed countries where Mg or Zn would represent single limiting factors causing impaired bone growth, it seems reasonable and prudent to include these elements proportionally in supplements, possibly also taking their different rates of absorption into account. The case for P is more delicate and, except for premature infants where P intake can be a limiting factor for bone growth, much more concern has been shown about excessive dietary intakes of P (Calvo & Park, 1996). High P intake results in an increased serum P concentration which initiates several hormonal respones, of which an increase in parathyroid hormone level effects the balance between bone mineralization and bone resorption (Anderson, 1996a,b). It is for this reason that in several supplementation studies no phosphate was given in order to shift the Ca : P ratio as much as possible in the direction of Ca. As already mentioned, the results from the study by Chan et al. (1995) give support to a concept where the absolute amount of Ca is more important than the Ca : P ratio. A nutrient which is most important in bone metabolism is vitamin D. The major role of dietary vitamin D is to function as precursor for 25-hydroxy- and 1,25-dihydroxycholecalciferol which maintain the plasma Ca concentration within very narrow limits. This is accomplished by varying the proportion of dietary Ca absorbed and excreted. As the body becomes vitamin D depleted, the efficiency of Ca absorption decreases from 30–50 % to no more than 15 %. In addition, 1,25-dihydroxycholecalciferol has a direct effect on osteoblast production and thus on bone formation and mineralization (Anderson, 1996a,b). The major source for vitamin D in human subjects is exposure to sunlight which enables cutaneous synthesis of vitamin D. However, several groups ranging from premature infants to institutionalized elderly people may not be able to receive sufficient exposure to sunlight and thus require dietary vitamin D. The involvement of vitamin K in bone metabolism is through its action on maturation of osteoblastic bone proteins by carboxylation of their glutamate residues (Vermeer et al. 1996). Vitamin K may be generated by the intestinal microflora or obtained from dietary sources such as green vegetables and meat. Except for the fact that it depends on the quantity of fluoride ingested and the time of exposure, S28 B. Koletzko et al. the role of F in bone health is still poorly understood. It has been used as a therapeutic agent in bone pathology including osteoporosis, but also cases of skeletal fluorosis are reported with radiologically demonstrable abnormal bone densification (Boivin et al. 1993). Of the minerals for which a role and/or essentiality in man is still unclear, B has been connected with the mechanical properties of bone (Mastrmatteo & Sullivan, 1994). Because the effects of B supplementation are not striking, and plausible mechanistic explanations still need to be presented, it is uncertain whether B will become an important issue in bone health research. In a cross-sectional study in about 500 children and adolescents (aged 8–17 years) bone mineral density was found to be related to diet, weight-bearing exercise and daylight hours spent outdoors (Gunnes & Lehman, 1995). Next to an effect in accordance with the literature with respect to dietary Ca it is surprising that the authors found a positive correlation between bone mineral density and saturated fat, fibre and vitamin C. Any interpretation of these correlations can only be speculative. The positive association between fat intake and bone mineral density might be attributed to an increased intake of vitamin D with the fat or might possibly be mediated by a higher cholesterol level. The positive association between bone mineral density and fibre intake is even more puzzling because one might expect exactly the opposite based on impairment of Ca absorption by dietary fibre. As a possible explanation for the association between vitamin C and bone mineral density, the authors indicate the fact that vitamin C is a cofactor in collagen synthesis. Moreover, they point to a possible link between high intakes of Ca, fibre and vitamin C as constituents of a more wholesome diet. Considering the importance of a high Ca intake in achieving maximal bone growth and peak bone mass, it seems logical to prevent any negative dietary interactions which might interfere with maximal Ca absorption. Of the dietary interactions affecting Ca absorption reviewed by Licata (1993), the possible negative effect of caffeine is mentioned because it increases urinary Ca excretion. However, this effect is considered to be fairly unimportant relative to other factors and is also not consistently found in all the studies. A clear negative effect on Ca absorption is represented by dietary oxalate-exemplified by the poor Ca absorption (only 5 %) from oxalaterich spinach. Although some authors report a stimulating effect of lactose on Ca absorption in humans, others fail to find any significant effect. Essentially the same holds true for the effect of protein on Ca absorption: if there is any stimulating effect, it will be rather small. In view of the marginal gain on fractional Ca absorption which can possibly be obtained by optimizing the nutritional matrix, selecting Ca salts with a high bioavailability may be of particular relevance. In this respect an increase in fractional Ca absorption of 0.26 to 0.36 which can be achieved by replacing CaCO 3 with calcium citrate–malate is illustrative (Peacock, 1991). 9. Nutrient effects on development of neural functions and behaviour 9.1. Introduction Pregnancy and the first postnatal months are critical time periods for the growth, development and differentiation of the human nervous system. There is good evidence that availability of nutrients during these critical time periods affects brain growth and development and can have longterm programming effects on an individual’s central nervous functions. In contrast to some other mammalian species, in man the peak growth rate of the brain (‘brain growth spurt’) occurs both pre- and postnatally, with continued relatively rapid growth well into the second year of life. Between the 24th week of gestation and the time of term birth, brain weight increases more than fivefold. The disproportionate speed of brain growth, compared with total body growth, is apparent from the fact that at age 2 years the weight of the human brain has already reached 80 % of adult weight, while whole body weight at this age is less than 18 % of adult weight. An adequate substrate supply is essential for a physiological brain composition and differentiation during this rapid perinatal growth. 9.2. Physiology of neural development Development of neuronal tissues is characterized by the sequential occurrence of mitosis, cell migration, differentiation, synaptogenesis, apoptosis and synaptic reorganization. These consecutive steps begin during pregnancy at different gestational ages and occur over a certain time period in different brain areas; therefore, there is an overlap in time of the various brain development steps in different areas of the brain (Reisbick, 1996). Mitosis of germinal cells located along the neural tube and the brain ventricles, which form neuronal and glia cells, begins in the sixth embryonic week, peaks in the second trimester and is almost complete by the end of the second trimester, even though a small number of cells may form during the last trimester of gestation (Jacobson, 1991; Rakic, 1995). Under the influence of hormones, cell adhesion molecules and other local factors, the created cells migrate to the brain nuclei and cortex and the ganglia of the peripheral nervous system (Jacobson, 1991; Kandel et al. 1991). Although largely occurring during pregnancy, nervous cell migration particularly in the superficial layers of the cerebral and cerebellar cortices continues to occur after term birth (Rakic, 1995). Cell differentiation into particular neurons may begin during migration but generally is only completed at its final destination (Kandel et al. 1995). The differentiation of neuronal cells is modulated by induction through surrounding cells, the kind of cell innervated, growth factors and hormones such as glucocorticoids, and nutrient intakes. For example, pre- and postnatal protein– energy malnutrition has marked effects on brain cell differentiation (Cravioto & Cravioto, 1996), and long-term depletion of the n-3 PUFA DHA induced altered brain cell levels of monoamines and monoamine receptors in rats. When neurons have migrated to their destination and reached their differentiation, they grow and form synaptic connections (Kandel et al. 1995). Most cell growth in the human cerebellar cortex occurs between 32 weeks of gestation and 11 months after term birth (Rakic, 1995). Synaptogenesis in primates peaks in the first 2–3 months of life (Rakic, 1995), and the number of synapses in the human visual cortex increases twofold between the ages of 2 and 8 months after birth, which is paralleled by a marked increase Growth, development and differentiation in the number of neurotransmitter receptors (Huttenlocher et al. 1982). The number of synapses formed is far greater than the number of synapses later found in adult brains. The early rapid synaptogenesis depends on an extensive formation of highly fluid cell membranes and, hence, on sufficient availability of substrates required for this membrane formation. Brain structure and function is also greatly influenced by apoptosis, since cell death affects between 20 and 80 % of the neuronal cells formed during gestation before the time of birth (Rosenzweig et al. 1996), which appears to be regulated by genetically programmed apoptosis, protecting neurotrophic factors, synaptogenesis, hormones and eicosanoids. In the surviving cells, the synaptic connections are reduced and rearranged at a rapid rate initially, but continuing throughout the life of the organism in relation to stimulation and learning. This synaptic rearrangement appears to be regulated by many different modulators, including Ca channels, second messengers such as phosphatidyl inositol, protein kinases, membrane lipids and eicosanoid hormones (Reisbick, 1996; Wainwright, 1997). 9.3. Nutrition and neural development In Europe and other developed countries, an associated secular increase both of adult height and adult intelligence quotients has been observed over the last few decades, and the hypothesis has been raised that both these effects are related to the early nutrient supply. Several studies have documented that severe general malnutrition in infancy or early childhood, which is characterized by a combined deficiency of energy, protein and many other substrates, is associated with a marked reduction of cognitive ability (Cravioto & Cravioto, 1996; Kretchmer et al. 1996). However, the adverse effects on neural development cannot be attributed entirely to nutrient depletion, but may also be modulated by other factors typically associated with severe childhood malnutrition, including infections, poverty, psychological depression and lack of adequate stimulation. Evidence of organic effects of malnutrition on brain growth and maturation was provided by the demonstration of cerebral atrophy and lasting organic brain damage (Ambrosius, 1966; Stoch et al. 1982; Houscham & Devilliers, 1987) and by electrophysiological evidence of impaired information processing, such as altered auditory evoked brain stem potentials (Barnett et al. 1978; Bartel et al. 1986). 9.3.1. Protein. Randomized studies in premature infants demonstrated that a low protein intake during the early postnatal period resulted in poorer results of orientation, habituation and stability clusters when tested with the neonatal behaviour assessment scale (Bhatia et al. 1991), and in a markedly reduced mental and psychomotor development when assessed at an age of 18 months postterm with the Bayley scales of infant development (Morley & Lucas, 1993). 9.3.2. Iodine. Today I deficiency is considered the most common cause of nongenetic inborn neurological damage on a worldwide basis, and causes cretinism with severe mental retardation (Stanbury, 1994). In I-deficient populations, an early I supply beginning before or at the time of conception can prevent neural damage of the infant, S29 whereas a later start of I supply in the second or third trimester of pregnancy or after birth is associated with a smaller preventive effect (Xue-Yi et al. 1994). 9.3.3. Iron. Fe uptake into the brain is mediated by transferrin receptors on the endothelial surface of brain microvasculature, reaches its maximum during the period of rapid brain growth and peak myelinogenesis (Taylor & Morgan, 1990) and continues throughout life. Analysis of Fe distribution in the human brain during childhood with magnetic resonance imaging showed the highest concentrations in the globus pallidus, caudate nucleus, putamen and substancia nigra, while the cortex and cerebellum had substantially lower contents. Fe serves as an essential cofactor in a variety of cellular and metabolic functions, including the synthesis of dopamine, serotonin, catecholamines and possibly g-aminobutyric acid as well as myelin formation (Kretchmer et al. 1996), while Fe overload has toxic effects. In young rats deprived of Fe in early postnatal life, total brain Fe content was severely depleted to 27 % of that of controls and was resistant to later restoration despite aggressive treatment (Dallman & Spirito, 1977). Sustained early Fe deficiency in rats causes persistent behavioural and learning deficits, leading to the hypothesis that Fe sufficiency throughout the critical phases of early brain development is crucial to the achievement of normal brain Fe content, function and behaviour. Several studies in children have clearly documented that severe Fe depletion resulting in Fe-deficiency anaemia results in a poor attention span, poor performance in the Bayley mental development index, low intelligence scores, some degree of perceptual disturbance and altered affective behaviour (Lozoff & Brittenham, 1986; Beard et al. 1993; Pollitt, 1993; Sheard, 1994; Kretchmer et al. 1996). Although children with Fe depletion without anaemia also showed behavioural abnormalities in some studies, these abnormalities appear to have been influenced by poor environmental conditions that were associated with the occurrence of Fe depletion (Lozoff et al. 1996). At this time there is no conclusive evidence that poor Fe status without anaemia has adverse effects on neural development in children. 9.3.4. Zinc. In animal experiments, Zn deprivation was shown to adversely affect brain growth, learning ability, memory and activity (Smart, 1974; Halas & Sanstead, 1975, 1980; Peters, 1979). Low-birth-weight infants showed an improvement of motor development when supplemented with Zn (Friel et al. 1993). In a recent study in Indian children aged 6–35 months who were not malnourished, daily supplementation with 10 mg elemental Zn (as Zn gluconate) for a period of 6 months resulted in a significant increase of observed activity (Sazawal et al. 1996). 9.3.5. Polyunsaturated fatty acids. Some 50–60 % of the structural matter in the central nervous system is composed of lipids, which almost entirely serve structural functions in cell membranes and myelin. Much of the rapid lipid accretion during brain growth comprises lipids that can be synthesized de novo in the fetus and infant, e.g. cholesterol. In addition, the rapid brain growth occurring perinatally and the extensive synaptogenesis occurring during the first months of life require the incorporation of relatively large amounts of essential PUFA, primarily the highly unsaturated long-chain PUFA DHA and arachidonic S30 B. Koletzko et al. acid (Koletzko, 1992). In experimental studies, the addition of DHA and arachidonic acid to fetal mouse brain cultures increased the number, diversity and complexity of synaptic contacts (Tixier-Vidal et al. 1986). In addition to cell growth and synaptogenesis, long-chain PUFA as well as eicosanoids formed from long-chain PUFA may influence neural cell apoptosis (Finstad et al. 1994; Wainwright, 1997). Experimental studies in rodents and in non-human primates indicated that the degree of long-chain PUFA incorporation into the developing brain influenced reflex development, memory, discrimination learning, retinal function and visual acuity (Neuringer, 1993; Wainwright, 1993), i.e. functions related to the efficacy of information processing. The human fetal supply with preformed long-chain PUFA by a materno-fetal placental transfer (Koletzko & Müller, 1990) may be influenced by the maternal dietary long-chain PUFA intake. Also, the postnatal long-chain PUFA supply in human milk is affected by maternal diet and can be altered by dietary supplements provided to lactating women (Harris et al. 1984; Koletzko et al. 1992). In contrast to human milk, most infant formulas do not contain preformed DHA and arachidonic acid. Although newborn infants have the ability to synthesize long-chain PUFA from EFA precursors, the rate of synthesis appears to be low (Demmelmair et al. 1995; Sauerwald et al. 1997) and long-chain PUFA levels in blood lipids (Decsi & Koletzko, 1995) and brain (Farquharson et al. 1992; Makrides et al. 1994) of formula-fed infants are significantly lower than those found in breast-fed infants. Some randomized intervention studies comparing diets without and with a supply of preformed long-chain PUFA both in premature infants (Uauy et al. 1990; Carlson & Werkman, 1996) and in healthy term infants (Makrides et al. 1993, 1995; Agostoni et al. 1995) found indications of improved retinal and visual function and of cognitive development in infants receiving long-chain PUFA, while other studies found no appreciable advantage (Innis et al. 1996). Effects of supplementing additional long-chain PUFA during pregnancy or lactation on functional development of the infant have not been reported. 9.4. Early nutrition and development of taste preferences Preferences of taste and smell are critical factors in selecting foods and drinks throughout life. There are some indications that sensory perception of tastes and flavours may be modulated by early exposure. Observations of fetal swallowing following the ingestion of sweet- and bitter-tasting substances into the amniotic fluid suggest that the fetus is sensitive to sweet- and bittertasting substances (de Snoo, 1937; Liley, 1972). It has been concluded that both the olfactory apparatus and taste perception are fully developed in utero (Beauchamp et al. 1991). Molecules carrying flavours and aromas may cross the placenta, and it has been proposed that early sensory exposure may modulate the acquisition of later flavour preferences. In animal studies, such effects of intrauterine exposure to flavour-rich foods on postnatal food choices have been documented. Premature infants tested between 33 and 40 weeks postconception and newborns during the first hours after birth show a clear and reproducible preference for sweet tastes (Beauchamp & Mennella, 1980). In blinded controlled studies, the sucking behaviour of breast-fed infants is altered by supplementing their mothers with encapsulated garlic compared with placebo, and repeated consumption of garlic modifies the infantile response to this taste (Menella & Beauchamp, 1993). Also the volatile flavours of vanilla and alcohol modulate infantile sucking behaviour (Beauchamp & Mennella, 1980; Mennella & Beauchamp, 1991). The question of to what extent food choices during childhood and adult life are modulated by pre- and postnatal experience is of major importance and needs to be further explored. 9.5. Methodological aspects An optimization of the quality of nutrient intakes during pregnancy, lactation and infancy may well have the potential of improving developmental outcomes in the recipient infants. Since even small improvements of the population means by such potential effects are of major public health significance, they need to be carefully examined. If plausible hypotheses can be raised and supported by data from experimental models, epidemiological studies or pilot intervention trials, they should be subjected to rigorous testing with adequate scientific methodology in double-blind placebo-controlled randomized trials. While some electrophysiological measures of human neural function can be assessed with sufficient precision to justify interventions in small groups, behavioural methods tend to have a greater degree of variation and, therefore, effects on such outcome variables require large studies, particularly if one considers that dietary factors must be expected to have relatively small effects compared with genetic and other environmental influences. The results of such studies can be greatly influenced by the study design and factors such as the time points chosen for testing effects as well as the adequacy of the method chosen for testing the targeted effect. Large sample sizes may be required, for example, to show an effect of a dietary supplement, resulting in a comparably small mean difference of developmental scores due to the many other variables that influence such end-points. Therefore, the realization of adequate trials to test for long-term developmental effects of early food choices will usually require a large budget. In view of the major public health importance of the questions addressed by such trials, it appears justified that well-designed trials addressing relevant and pertinent questions are supported by public funds. 10. Production of bioactive factors for inclusion into food products Interest in the production of human milk proteins, peptides, growth factors and other bioactive substances with the use of biotechnology and recombinant techniques is growing. The inclusion of such substances into dietary products may have beneficial physiological effects particularly in infancy and early childhood, for example defence against infectious agents, the optimization of nutrient uptake from the diet as well as the differentiation and growth of cells and tissues. Micro-organisms and transgenic animals can now be used Growth, development and differentiation for the production of bioactive proteins (Lönnerdal, 1996). However, the benefits and safety of each substance must be evaluated in adequate studies in cells, animal models and clinical studies before routinely adding them to products for infants to improve their nutrition, health or development. Proper manufacturing conditions must be developed for introducing such substances into foods. The importance of post-translational modifications must also be taken into account. Some proteins may require proper glycosylation or phosphorylation for physiological activity. Several human milk proteins have been cloned and sequenced. The majority were cloned at the level of complementary DNA (cDNA). In a few cases the entire gene has been characterized. One of the first milk proteins which was cloned from a mammary gland library was alactalbumin (Hall et al. 1987). Other human milk proteins that have been cloned include lysozyme, lactoferrin, human b-casein and k-casein. For the production of human milk proteins and bioactive factors, several expression systems can be used, such as bacterial expression of recombinant proteins. The expression vector can be constructed so that the protein is made available in the supernatant fraction or in the bacteria. Bacterial expression will lead to the production of proteins which are not phosphorylated or glycosylated. If this property is needed for biological functions, yeast or fungi can be used as expression systems. Saccharomyces cerevisiae was used to produce human lactoferrin (Liang & Richardson, 1993) and human b-caseins. Aspergillus nidulans and Aspergillus oryzae were also used to produce human lactoferrin (Ward et al. 1992a,b). Alternatively, living cells such as baby hamster kidney cell can be used to produce human lactoferrin (Ward et al. 1992a). As such expression systems will still provide proteins with different phosphorylation and glycosylation, physiological function may be altered. Tissue-specific expression of human milk proteins in transgenic animals should result in recombinant protein glycosylations and phosphorylation similar to the native human milk proteins. The architecture of the transgene DNA that is introduced into the germline of animals by microinjection plays an important role in the level of expression of the transgene. DNA that is introduced by microinjection into the pronucleus is usually inserted randomly into the genome as head-to-tail concatemers. Because the eukaryotic genome is organized into topologically constrained domains, random integration can lead to position effects in which the transgene expression is influenced by the surrounding chromosomal sequences. Thus in many cases the level of transgene expression will vary over several logarithms, depending on the site of integration, and expression may be observed in <50 % of the positive transgenic mice. Because of the expense and time required to generate transgenic livestock, this presents a major problem (Stowell et al. 1991). To date, human lactoferrin (Rosen et al. 1996) and human lysozyme (Kim et al. 1994) have been expressed in transgenic mice with reasonable expression in the milk. Human lactoferrin was recently produced in transgenic cows (Maga et al. 1994). Whether the glycosylation pattern of human lactoferrin within the cow’s mammary cells will alter the biological function remains to be elucidated. S31 Similar to the gene farming of human milk proteins and other important functional proteins in transgenic animals, such bioreactors are used to produce human hormones and growth factors. In transgenic rabbits, hGH was produced in milk in suitable amounts without being affected by the transgene expression (Limonta et al. 1995). Bovine growth hormone was also placed by a recombinant technique into the milk of transgenic mice (Thépot et al. 1995). In addition, human IGF1 was expressed in the milk of transgenic mice (Hadsell et al. 1996) and rabbits (Brem et al. 1994). Some oligosaccharides which may play an important role in promoting growth and differentiation of intestinal epithelial cells have successfully been expressed in the milk of transgenic mice (Prieto et al. 1995). The enormous potential of these methodologies for improving food products needs to be further explored. 11. Commentary on biomarkers Drawing on experience from child health and development it is possible to use experience of inborn errors of metabolism to demonstrate the evolution and use of biomarkers to detect and predict a disease arising from an imbalance between systemic homeostasis and essential and non-essential dietary components. A good example is hyperphenylalaninaemia (HPA) which arises from an inborn error of metabolism of the essential amino acid phenylalanine (Scriver & Clow, 1980; Güttler, 1984). First of all, however, it is helpful to consider some general aspects of the strategy of biomarkers. Biomarkers can be defined as indicators of actual or possible changes of systemic, organ, tissue, cellular and sub-cellular structural and functional integrity which can be used, either singly or in batteries, to monitor health and exposure to compounds in populations and individuals: as such they are the essence of chemical pathology or clinical biochemistry. Thus although the term ‘biomarker’ might itself be relatively novel, the concept is far from being so: the use of biochemical biomarkers dates at least from the discriminatory use of the sweet taste of glycosuria to diagnose and name diabetes mellitus. This intolerance of glucose illustrates a disturbance of the usually tightly controlled internal milieu as conceived by Claude Bernard, and of the processes involved in maintaining this state, i.e. homeostasis. These mechanisms lie at the core of the regulatory adaptations which maintain a steady state in the face of changes in environment, including our interaction with desirable and non-desirable components of the diet, which, of course, is one of our most intimate interactions with the environment. Homeostasis is regulated by cellular and systemic mechanisms which, in turn, are dependent on gene-mediated responses. Genetic heterogeneity amongst people underlies the intrapopulation variability of the interaction between nurture and nature. These generalizations apply equally to exposure to abnormal compounds and to excessive or inadequate exposure to normal components in the diet or environment. Homeostasis comprises several processes, some of which act synchronously, but many of which act in a specific sequence, each being triggered by the relative efficiency of the preceding process. Some of these mechanisms are S32 B. Koletzko et al. specific for individual compounds. This applies to essential nutrients or generic groups of nutrients, but for non-nutrients and xenobiotics there appears to be a limited number of common protective mechanisms. In the chain arising from threatened or actual toxic exposure there would first be compensatory adaptations such as metabolic biotransformations (phase I: oxidation, hydrolysis, reduction; phase II: conjugations), excretion, and sequestration of compounds (for example in adipose tissue, bone, hair, skin); when these are inadequate biochemical and histopathological features follow with structural cellular damage and functional derangement. These lead to tissue damage with clinical manifestations (e.g. neurological and muscular toxicity, retinal toxicity, immunotoxicity, hepatotoxicity, nephrotoxicity, bone marrow damage, teratogenicity, cytotoxicity, cancer, methaemoglobinaemia, haemolysis) and overt disease states many of which are major health issues (e.g. cancers, cardiovascular disease, obesity, osteoporosis, adverse reactions to foods and food additives and contaminants, atopic disease, behavioural abnormalities, fetotoxic effects). This chain shows the series of genotypic (genetic), and phenotypic (biochemical and clinical) biomarkers. The underlying assumption here is that all diet-related disease arises from an inappropriate interaction between diet and systemic homeostasis and that the fundamental basis of this imbalance lies in the relative imbalance of dietary exposure and the genetic controlled response. At this level the genome and its product would serve as an ideal biomarker: unfortunately the processes involved might not have been identified or might not be accessible to practical and ethical sampling techniques. Alternative and less immediate biomarkers have to be used. However, the more remote the biomarker is from the primary event the more it is attenuated and subject to confounding factors. It becomes less specific. On the other hand the more immediate it is to the basic interaction the more quantitatively related and predictive it becomes: biomarkers represent Garrod’s concept that genetic factors determine the nature of chemical metabolism and human biochemical heterogeneity. More immediate biomarkers are not only potentially more predictive, they might also provide more immediate outcomes which could be used to assess interventions in a reasonable timescale and, in turn, replace the temporally and aetiologically remote outcomes which are so often the foci of epidemiological studies of diet and health: for example the diseases mentioned earlier are probably both metabolically and temporally remote from their aetiology. Insight into the metabolic processes involved in the particular issue being investigated can inform the choice of tissue or fluid to be sampled and the phenotypic marker to be measured. This applies also to genotypic biomarkers designed to assess adaptive phenomena but it is possible to measure some predictive genotypic markers in tissues other than those in which the gene product is expressed. Whatever the situation the selection of biomarker(s) should be dictated by the problem being considered rather than by the accessibility of tissue or fluid to be measured or the ease of an assay. These general points are represented by the disease PKU which epitomizes the interaction between nature and nurture and the benefits which can be attained by an appropriate manipulation of the diet. PKU is a manifestation of HPA which results from the altered activity of the principally hepatic enzyme phenylalanine hydroxylase (EC 1.14.16.1; PAH), functional defects of which arise from intrinsic defects in the apoenzyme or from defective synthesis of its cofactor (tetrahydrobiopterin) (Scriver & Clow, 1980; Güttler, 1984; Scriver, 1991; Scriver et al. 1996; Güttler & Guldberg, 1996). The condition was recognized in 1933 when a mother with a mentally retarded child sought help on the peculiar smell of her child’s urine. She saw several doctors, one of whom even referred her to a psychiatrist for help with her delusion. In the end one physician, Dr Asbjorn Folling confirmed the peculiar smell. He attributed it to the presence of phenylpyruvic acid in the urine and he called the condition ‘imbecillitas phenylpyruvica’ thereby incorporating both the initial clinical and biochemical biomarkers into the condition’s nomenclature (Güttler, 1984; Scriver, 1991). He developed the ‘FeCl 3’ colorimetric test of urine to detect the excess metabolite and used it to screen mentally retarded children and adults for the condition (Scriver & Clow, 1980). In 1950 Horst Bickel appreciated the implication of raised blood phenylalanine concentrations in PKU patients and explored the use of a phenylalanine-free mixture of amino acids to feed affected infants. It was found that blood phenylalanine levels fell and that the urine abnormality decreased. There was also some clinical and behavioural improvement, but the intervention was too late to affect the established developmental delay and mental damage. Nonetheless this progress in the phenotypic biomarkers demonstrated the potential clinical and economic benefits of screening and early diagnosis in the hope that early dietary intervention could prevent the neurological damage. Early dietary management was effective, but many cases went undetected because the FeCl 3 biomarker was relatively insensitive and non-specific (Scriver & Clow, 1980). Phenylpyruvic acid is one of a number of normal metabolites of phenylalanine which are found in excessive amounts in the urine in PKU as a consequence of the increased activity of alternative pathways for the metabolism of phenylalanine. Other variables affected the production of these metabolites and it was appreciated that blood phenylalanine concentrations would be a better biomarker, but the analytical techniques (paper chromatography) available at that time were costly and time consuming. In the late 1950s the abnormal gene product, i.e. defective PAH, was identified, but since this activity was in the liver this did not provide an effective biomarker. The finding of this functional defect reinforced the opinion that screening should be targeted at the HPA. Guthrie solved the problem in 1961 by establishing a semiquantitative bioassay based on the inhibition of bacterial growth by high concentrations of phenylalanine. The test was done on blood drops collected on filter card from infants at 4–7 d of age by which time the babies would have fed and, in contrast to in utero, would have had the opportunity to challenge the activity of their endogenous PAH with a phenylalanine load. The biomarker for definitive diagnosis and monitoring of management was direct quantitative measurement of blood phenylalanine concentrations. This enabled an appreciation that these concentrations could correlate with the efficiency of management and with behavioural and other clinical outcomes. However, not all Growth, development and differentiation patients responded similarly or predictably. Amongst this heterogeneity was an atypical group with HPA which was, in the mid 1970s, found to be secondary to a defect in the synthesis of the tetrahydrobiopterin cofactor. Thus, in the midst of the phenotypic and molecular heterogeneity which was becoming increasingly obvious, it was realized that PKU or HPA was not the result of a single gene defect. The allele for PAH has since been found on chromosome 12q, and that for the affected stage in tetrahydrobiopterin synthesis is on chromosome 4 (Güttler & Guldberg, 1996; Scriver et al. 1996). HPA is, thus, evident as a complex entity with many clinical and metabolic phenotypes. Consistent with this heterogeneity and with the variation in dietary tolerance for phenylalanine, almost 300 mutations in the PAH gene have been identified. Some of these genetic mutations have been specifically correlated with the activity of their enzyme products and with the severity of the clinical disease (Scriver, 1991; Güttler & Guldberg, 1996; Scriver et al. 1996). In the context of biomarkers, HPA represents a continuum of genotypic, biochemical (PAH, tetrahydrobiopterin defect), metabolic (HPA), and clinical (the PKU syndrome) phenotypic biomarkers of an abnormal interaction between a dietary component and an individual’s ability to achieve effective homeostasis at customary dietary exposure. The genotypic biomarker may be used for definitive diagnosis and family screening and possibly for prognostication, but the metabolic biomarker is most useful for monitoring and tailoring the dietary reduction of phenylalanine. The urinary biomarker (PKU) is no longer of much use, and the clinical phenotypic biomarker is one of a medical tragedy. PKU and HPA demonstrate the movement of biomarkers from remote and non-specific outcomes (many other conditions including other inborn errors of metabolism cause epilepsy and developmental delay) to more specific and informative outcomes, and ultimately to the basic genomic mutation offering the opportunity of appropriately designed diets. It is feasible that in due course a similar heterogeneity will be found in the metabolism of other nutrients, and that this will explain many of the conflicting phenomena found in epidemiological studies of the interaction between diet and health and in apparent heterogeneity of nutritional requirements. Such biomarkers should also provide more definitive indicators of response to interventions, thereby enabling shorter and more definitive epidemiological studies. Whether or not people would accept the corollary of this, namely specific diets tailored to their genotypic or surrogate metabolic characteristics, depends probably on the nature of the diets involved and on the people’s motivation in the context of the much less dramatic and immediate effects of an inappropriate diet compared with that experienced with PKU. The integration of markers of susceptibility with appropriate follow up of long-term outcomes by epidemiological studies of characterized populations will demonstrate the relevance or otherwise of formal interventions. 12. Conclusions Food supply and the metabolism of food ingredients in women during pregnancy and lactation and in their children S33 have implications for long-term health and child development. Epidemiological evidence and studies performed in infants have highlighted the fact that maternal and intrauterine influences are of special importance during the development of the infant and child. Early nutrition modulates growth and functional development of the organism and appears to exert life-long programming effects that modulate health, disease and mortality risks in adulthood, neural function and behaviour, and quality of life. The field of nutrient–gene interaction is in a phase of rapid expansion and development. There are several areas where dietary modulation of gene expression could exert beneficial effects, for example with respect to lipid metabolism and risk of cardiovascular disease. Applied research should further elucidate the interaction of nutrients and gene expression. The genes affected by specific nutrients, e.g. amino acids, must be characterized in animal models as well as the underlying cellular and molecular mechanisms. Since the vast majority of studies have been performed on animal models or animal cell lines, strategies must be defined to approach these questions in human subjects, for example by using human cell lines responding to nutrients. Once the mechanisms and relevance to man have been clarified, then development and testing of existing or new ‘functional foods’ could be performed both in animals and human subjects. The relation between nutrients and differentiation needs to be further explored by in vivo models, studying food effects on cell differentiation and later performance. Concomitantly, the effects of nutrients on cell differentiation in in vitro studies should be strongly encouraged. The course of pregnancy, childbirth and lactation as well as human milk composition and the short- and long-term outcome of the child may be influenced by the intake of foods and particularly micronutrients, e.g. PUFA, Fe, Zn and I. Folic acid supplementation from before conception through the first weeks of pregnancy was reported to markedly reduce the occurrence rates of severe embryonic malformations, including anencephaly and spina bifida. The potential of exerting beneficial effects for mother and child by modulating maternal nutrient supply should be further explored. The evaluation of dietary effects on child growth requires epidemiological and field studies as well as evaluation of specific cell and tissue growth. Novel substrates, growth factors and conditionally essential nutrients (e.g. growth factors, amino acids, unsaturated fatty acids) may be potentially useful as ingredients in functional foods and need to be assessed carefully with respect to their potential effects on growth, maturation and development of specific cell types, tissues and organs under different physiological conditions. In particular, the potential modulation of later obesity by early food choices needs to be further explored. Intestinal growth, maturation, intestinal adaptation and regulation and long-term function may be influenced by food ingredients. The roles of compounds such as dietary oligosaccharides, gangliosides, high-molecular-mass glycoproteins, bile salt-activated lipase, pre- and probiotics as to their physiological functions in the developing organism need to be further explored. The interaction of the appropriate genes of intestinal substrate transporters and their S34 B. Koletzko et al. substrates in early childhood is not well understood and needs to be clarified. There are indications for some beneficial effects of functional foods on the developing immune response in vitro and in vivo, for example induced by antioxidant vitamins, trace elements, fatty acids, arginine, nucleotides, and altered antigen content in infant foods. A general unresolved issue is related to the lack of reliable surrogate outcome markers when investigating the effects of nutrition on the developing immune response and possible long-term benefits. Peak bone mass at the end of adolescence can be increased by dietary means, which is expected to be of long-term importance for the prevention of osteoporosis at older ages. Future studies should be directed to the combined effects of Ca and other constituents of growing bone, such as P, Mg and Zn, as well as vitamins D and K, and trace elements F and B. In addition to observational and intervention studies, in vitro and animal model studies might provide a basis for new concepts on the interaction and optimal relative proportions of the several macro- and micronutrients involved in bone growth and mineralization. Pregnancy and the first postnatal months are critical time periods for the growth and development of the human nervous system, processes for which adequate substrate supplies are essential. Early diet may have long-term effects on the structure and function of the nervous system, sensory and cognitive abilities as well as behaviour. The potential beneficial effects of a balanced supply of nutrients such as I, Fe, Zn and PUFA need to be explored in further detail. The question of a possible relationship between early exposure to tastes and flavours and later food choice preferences may have a major impact on public health and needs to be further elucidated. Bioactive factors such as human milk proteins, peptides, growth factors and other substances may be produced for use in food products with the use of biotechnology and recombinant techniques. The inclusion of such substances into dietary products may have beneficial physiological effects particularly in infancy and early childhood, for example defence against infectious agents, the optimization of nutrient uptake from the diet as well as the differentiation and growth of cells and tissues. The enormous potential of these methodologies for improving food products needs to be further explored. References Agostoni C, Trojan S, Bellu R, Riva E & Giovannini M (1995) Neurodevelopmental quotient of healthy term infants at 4 months and feeding practice: the role of long-chain polyunsaturated fatty acids. Pediatric Research 38, 262–266. Al MDM, van Houwelingen AC, Kester ADM, Hasaart THM, de Jong AEP & Hornstra G (1995) Maternal essential fatty acid patterns during normal pregnancy and their relationship with the neonatal essential fatty acid status. British Journal of Nutrition 74, 55–68. Allen LH (1986) Trace minerals and outcome of human pregnancy. Clinical Nutrition 5, 72–77. Ambrosius KD (1966) Comportamiento del peso de algunos organos en niñoscon desnutición de tercer grado. Boletin Medico – Hospital Infantil de Mexico 28, 47. Anderson B, Porras O, Hanson LA, Lagergård T & Svanborg-Edén C (1986) Inhibition of attachment of Streptococcus pneumoniae and Hemophilus influenzae by human milk and receptor oligosaccharides. Journal of Infectious Diseases 153, 232–237. Anderson JJB (1996a) Introduction (symposium: nutritional advances in human bone metabolism). Journal of Nutrition 126, 1150S–1152S. Anderson JJB (1996b) Calcium, phosporus and human body development. Journal of Nutrition 126, 1153S–1158S. Andon MB, Lloyd T & Matkovic V (1994) Supplementation trials with calcium citrate malate: evidence in favor of increasing the calcium RDA during childhood and adolescence. Journal of Nutrition 124, 1412S–1417S. Antras-Ferry J, Robin P, Robin D & Forest C (1995) Fatty acids and fibrates are potent inducers of transcription of the phosphoenolpyruvate carboxykinase gene in adipocytes. European Journal of Biochemistry 234, 390–396. Baldwin HS (1996) Early embryonic vascular development. Cardiovascular Research 31, E34–E45. Ballou LR, Laulederkind SJ, Rosloniec EF & Raghow R (1996) Ceramide signalling and the immune response. Biochimica et Biophysica Acta 1301, 273–287. Banks VF & Beutler E (1988) Iron. In Modern Nutrition in Health and Disease, 7th ed., pp.193–226 [ME Shills and VR Young, editors]. Philadelphia, PA: Lea & Febiger. Barbul, A. (1990) Arginine and immune function. Nutrition 6, 53–58. Barker DJP (1994) Mothers, Babies and Disease in Later Life. London: BMJ Publishing. Barnett AA, Weis AP, Sotillo MV, Skurovich M & Cravioto J (1978) Abnormal auditory evoked potentials in early human malnutrition. Science 201, 450–452. Bartel PR, Robinson E, Conradie JM & Prinsloo JG (1986) Brainstem auditory evoked potentials in severely malnourished children with kwashiorkor. Neuropediatrics 17, 178–182. Beard JL, Connor JR & Jones BC (1993) Iron in the brain. Nutrition Reviews 51, 157–170. Beauchamp GK, Cowart BJ & Schmidt HJ (1991) Development of chemosensory sensitivity and preference. In Smell and Taste in Health and Disease, pp. 405–416 [TV Getchell, RL Doty, LM Bartoshuk and JB Sow, editors]. New York, NY: Raven Press. Beauchamp GK & Mennella JA (1980) Early feeding and the acquisition of flavor preferences. American Journal of Clinical Nutrition 33, 2202. Bendich A (1995) Immunology: functions to assess nutrient requirements. Journal of Nutritional Immunology 3, 47–56. Bhatia J, Rassin DK, Cerreto MC & Bee DE (1991) Effect of protein–energy ratio on growth and behaviour of premature infants: preliminary findings. Journal of Pediatrics 119, 103–110. Bishop NJ, Dahlenburg SL, Fewtrell MS, Morley R & Lucas A (1995) Early diet of preterm infants and bone mineralization at age five years. Acta Pediatricia 85, 230–236. Boivin G, Dupuis J & Meunier PJ (1993) Fluoride and osteoporosis. World Review of Nutrition and Dietetics 73, 80–103. Bothwell TH, Charlton RW & Cook JD (1979) Iron Metabolism in Man. Oxford: Blackwell. Braga M, Vignali A, Gianotti L, Cestari A, Profili M & Carlo VD (1996) Immune and nutritional effects of early enteral nutrition after major abdominal operations. European Journal of Surgery 162, 105–112. Brem G, Hartl P, Besenfelder U, Wolf E, Zinovieva N & Pfaller R (1994) Expression of synthetic cDNA sequences encoding human insulin-like growth factor-q (IGF1) in the mammary gland of transgenic rabbits. Gene 149, 351–355. Bryan CF and Stone MJ (1993) In Nutrient Modulation of the Immune Response, pp.105–126 [S Cunningham-Ruddles, editor]. New York, NY: Marcel Dekker. Growth, development and differentiation Buddington RK, Chen JW & Diamond JM (1991) Intestinal regulation of intestinal brush-border sugar and amino acid transport in carnivores. American Journal of Physiology 261, R793–R801. Buddington RK & Diamond JM (1989) Ontogenetic development of intestinal nutrient transporters. Annual Review of Physiology 51, 601–619. Bueno JD, Torres M, Almendros A, Carmona R, Núñez MC, Rı́os A & Gil A (1994) Effect of dietary nucleotides on smallintestinal repair after diarrhoea. Histological and ultrastructural changes. Gut 35, 926–933. Bung P, Stein C, Prinz R, Pietrzik K, Schlebusch H, Bauer O & Krebs D (1993) Folsäureversorgung in der Schwangerschaft– Ergebnisse einer prospektiven Longitudinalstudie (Folic acid provision in pregnancy: results of a prospective longitudinal study). Geburtsh u Frauenheilk 53, 92–99. Bung P, Prietl G, Willrodt A, Krebs D, Prinz-Langenohl R & Pietrzik K (1995) Idiopathische Infertilität und Sterilität–spielt Folsäure eine Schlüsselrolle? (Idiopathic infertility and sterility: does folic acid play a key role?). Journal für Fertilität und Reproduktion 2, 21–24. Burke BS, Harding VV & Stuart HC (1943) Nutrition upon the condition of the infant at birth. Journal of Nutrition 26, 569. Buts JP & Nyakabasa M (1985) Role of dietary protein to adaptation at weaning in the development of the rat gastrointestinal tract. Pediatric Research 19, 857–862. Butterworth CE & Bendich A (1996) Folic acid and the prevention of birth defects. Annual Review of Nutrition 16, 73–97. Buxton DB, Robertson SM & Olson MS (1986) Stimulation of glycogenolysis by adenine nucleotides in the perfused rat liver. Biochemical Journal 237, 773–780. Calvo MS & Park YK (1996) Changing phosphorus content of the US diet: potential for adverse effects on bone. Journal of Nutrition 126, 1168S–1180S. Carlson SE, Cooke RJ, Rhodes PG, Werkman SH, Peeples JM & Tolley EA (1991) Long term feeding of formulas high in linolenic acid and marine oil to very low birth weight infants: phospholipids fatty acids. Pediatric Researvh 30, 404–412. Carlson SE, Cooke RJ, Werkman SH & Peeples JM (1992) First year growth of preterm infants fed standard compared to marine oil n-3 supplemented formula. Lipids 27, 901–907. Carlson SE & Werkman SH (1996) A randomized trial of visual attention of preterm infants fed docosahexaenoic acid until two months. Lipids 31, 85–90. Carlson SE, Werkman SH, Peeples JM, Cooke RJ & Tolley EA (1993) Arachidonic acid status correlates with first year growth in preterm infants. Proceedings of the National Academy of Sciences USA 90, 1073–1077. Carver JD (1994) Dietary nucleotides: cellular immune, intestinal and hepatic system effects. Journal of Nutrition 124, 144S–148S. Carver JD, Pimentel B, Cox W & Barness L (1991) Dietary nucleotides effects upon immune function in infants. Pediatrics 88, 359–363. Cerra FB, Lehmann S, Konstantinides N, Dzik J, Fish J, Konstantinides F, LiCari JJ & Holman RT (1991) Improvement in immune function in ICU patients by enteral nutrition supplemented with arginine, RNA, and menhaden oil is independent of nitrogen balance. Nutrition 7, 193–199. Chan GM, Hoffman K & McMurry M (1995) Effects of dairy products on bone and body composition in pubertal girls. Journal of Pediatrics 126, 551–556. Chandra RK (1991) Interactions between early nutrition and the immune system. Ciba Foundation Symposium 156, 77–89. Chandra RK, Puri S & Hamed A (1989a) Influence of maternal diet during lactation and use of formula feeds on development of atopic eczema in high risk infants. British Medical Journal 299, S35 228–230 (published erratum appears in British Medical Journal 299, 896). Chandra RK, Puri S, Suraiya C & Cheema PS (1986) Influence of maternal food antigen avoidance during pregnancy and lactation on incidence of atopic eczema in infants. Clinical Allergy 16, 563–569. Chandra RK, Singh G & Shridhara B (1989b) Effect of feeding whey hydrolysate, soy and conventional cow milk formulas on incidence of atopic disease in high risk infants. Annals of Allergy 63, 102–106. Chavance M, Herbeth B, Fournier C, Janot C & Vernhes G (1989) Vitamin status, immunity and infections in an elderly population. European Journal of Clinical Nutrition 43, 827–835. Cheung R, Karjalainen J, Vandermeulen J, Singal DP & Dosch HM (1994) T cells from children with IDDM are sensitized to bovine serum albumin. Scandinavian Journal of Immunology 40, 623–628. Chew BP (1995) Antioxidant vitamins affect food animal immunity and health. Journal of Nutrition 125, 1804S–1808S. Chirico G, Marconi M, Colombo A, Chiara A, Rondini G & Ugazio AG (1983) Deficiency of neutrophil phagocytosis in premature infants: effects of vitamin E supplementation. Acta Paediatrica Scandinavia 72, 521–524. Cidlowski JA, King KL, Evans-Storms RB, Montague JW, Bortner CD & Hughes FM (1996) The biochemistry and molecular biology of glucocorticoid-induced apoptosis in the immune system. Recent Progress in Hormone Research 51, 457–490. Clarke SD (1994) Dietary polyunsaturated fatty acid regulation of gene transcription. Annual Review of Nutrition 14, 83–98. Clarke SD & Abraham S (1992) Gene expression: nutrient control of pre- and post-transcriptional events. FASEB Journal 6, 3146–3152. Cohen A, Barankiewicz J & Gelfand EW (1984) Roles of alternative synthetic and catabolic purine pathways in T lymphocyte differentiation. Proceedings of the National Academy of Sciences USA 81, 26–33. Cole DE, Gundberg CM, Stirk LJ, Atkinson SA, Hanley DA, Ayer LM & Baldwin LS (1987) Changing osteocalcin concentration during pregnancy and lactation: implications for maternal mineral metabolism. Journal of Clinical Endocrinology and Metabolism 65, 290. Cosman D (1994) A family of ligands for the TNF receptor superfamily. Stem Cells 12, 440–455. Coutsoudis A, Bobat RA, Coovadia HM, Kuhn L, Tsai WY & Stein ZA (1995) The effects of vitamin A supplementation on the morbidity of children born to HIV-infected women. American Journal of Public Health 85, 1076–1081. Coutsoudis A, Kiepiela P, Coovadia HM & Broughton M (1992) Vitamin A supplementation enhances specific IgG antibody levels and total lymphocyte numbers while improving morbidity in measles. Pediatric Infectious Disease Journal 11, 203–209. Cowell CT, Lu PW, Lloyd-Jones SA, Briody JN, Allen JR, Humphries IR, Reed E, Knight J, Howman-Giles R & Gaskin K (1995) Volumetric bone mineral density – a potential role in paediatrics. Acta Paediatrica 411, Suppl., 12–16. Crane R (1975) The physiology of intestinal absorption of sugars. In Physiological Effects of Food Carbohydrates, pp. 1–19 [A James and J Hodges, editors]. Washington, DC: American Chemical Society. Cravioto A, Tello A, Villafán H, Ruiz J, Del Vedovo S & Neeser JR (1991) Inhibition of localized adhesion of enteropathogenic Escherichia coli to Hep-2 cells by immunoglobulin and oligosaccharide fractions of human colostrum and breast milk. Journal of Infectious Diseases 163, 1247–1255. Cravioto J & Cravioto P (1996) Mental development and malnutrition. In Long-term Consequences of Early Feeding, pp. 35–56 S36 B. Koletzko et al. [J Boulton, Z Laron and J Rey, editors]. Philadelphia, PA: Lippincott-Raven. Cynober L (1994) Can arginine and ornithine support gut functions? Gut 35, Suppl. 1, S42–S45. Dahlqvist A & Lindberg T (1965) Fetal development of the smallintestinal disaccharidase and alkaline phosphatase in the human. Biology of the Neonate 9, 24–32. Dahlqvist A & Lindberg T (1966) Development of the intestinal disaccharidase and alkaline phosphatase activities in the human fetus. Clinical Science 30, 517–528. Dallman PR & Spirito RA (1977) Brain iron in the rat: extremely slow turnover in normal rat may explain the long-lasting effects of early iron deficiency. Journal of Nutrition 107, 1075–1081. Daly JM, Reynolds J, Sigal RK, Shou J & Liberman MD (1990) Effect of dietary protein and amino acids on immune function. Critical Care Medicine 18, S86–S93. Danielsen EM, Skovberg H, Norén O & Sjöström H (1984) Biosynthesis of intestinal microvillar proteins. Intracellular processing of lactase–phlorizin hydrolase. Biochemical and Biophysical Research Communications 122, 82–90. Danielsen EM, Skovbjerg H, Norén O & Sjöström H (1981) Biosynthesis of intestinal proteins. Nature of precursor forms of microvillar enzymes from Ca 2+-precipitated enterocyte membranes. FEBS Letters 132, 197–200. Davidson NO, Hausman AML, Ifkovits CA, Buse JB, Gould GW, Burant CF & Bell GI (1992) Human intestinal transporter expression and localization of GLUT 5. American Journal of Physiology 262, C795–C800. Davis RL, Hargen SM, Yeomans FM & Chow BF (1973) Long term effects of alterations of maternal diet in mice. Nutrition Reports International 7, 463–473. De Luca LM (1991) Retinoids and their receptors in differentiation, embryogenesis and neoplasia. FASEB Journal 5, 2924–2933. de Snoo K (1937) Das trinkende Kind im Uterus (The drinking child in utero). Monatsschrift für Geburtshilfe und Gynäkologie 105, 88–97. Decsi T & Koletzko B (1995) Growth, fatty acid composition of plasma lipid classes, and plasma retinol and a-tocopherol concentrations in full-term infants fed formula enriched with q-6 and q-3 long-chain polyunsaturated fatty acids. Acta Paediatrica 84, 725–732. De-Lucchi C, Pita ML, Faus MJ, Molina JA, Uauy R & Gil A (1987) Effects of dietary nucleotides on the fatty acid composition of erythrocyte membrane lipids in term infants. Journal of Pediatric Gastroenterology and Nutrition 6, 568–574. Demirini S & Tsang RC (1995) Calcium and phosphorus nutrition in infants. Journal of Pediatric Gastroenterology and Nutrition 21, 367–373. Demmelmair H, von Schenck U, Behrendt E, Sauerwald T & Koletzko B. (1995) Estimation of arachidonic acid synthesis in fullterm neonates using natural variation of 13C-abundance. Journal of Pediatric Gastroenterology and Nutrition 21, 31–36. Deurenberg P (1992) Methods for determining fat mass and fat distribution. Acta Paediatrica Suppl. 383, 53–57. Docherty K & Clark AR (1994) Nutrient regulation of insulin gene expression. FASEB Journal 8, 20–27. Dragunow M & Preston K (1995) The role of inducible transcription factors in apoptotic nerve cell death. Brain Research Reviews 21, 1–28. Dubois S, Dougherty C, DuQuetette M, Hanley JA & Montgin JM (1991) Twin pregnancy: the impact of the Higgins Nutrition Intervention Program on maternal and neonatal outcomes. American Journal of Clinical Nutrition 53, 1397. Duggin GG, Dale NE, Lyneham RC, Evans RA & Tiller DJ (1974) Calcium balance in pregnancy. Lancet 2, 926–927. Egge H, Dell A & von Nicolai H (1983) Fucose containing oligosaccharides from human milk. Archives of Biochemistry and Biophysics 224, 235–253. Endres S (1996) n-3 Polyunsaturated fatty acids and human cytokine synthesis. Lipids 31, S239–S242. Evans-Storms RB & Cidlowski JA (1995) Regulation of apoptosis by steroid hormones. Journal of Steroid Biochemistry and Molecular Biology 53, 1–8. Fälth-Magnusson K & Kjellman NI (1992) Allergy prevention by maternal elimination diet during late pregnancy – a 5-year follow-up of a randomized study. Journal of Allergy and Clinical Immunology 89, 709–713. Fälth-Magnusson K, Oman H & Kjellman NI (1987) Maternal abstention from cow milk and egg in allergy risk pregnancies. Effect on antibody production in the mother and the newborn. Allergy 42, 64–73. Farquharson J, Cockburn F, Patrick WA, Jamieson EC & Logan RW (1992) Infant cerebral cortex phospholipid fatty-acid composition and diet. Lancet 340, 810–813. Fernandes G, Venkatraman J, Khare A, Horbach GJ & Friedrichs W (1990) Modulation of gene expression in autoimmune disease and aging by food restriction and dietary lipids. Proceedings of the Society for Experimental Biology and Medicine 193, 16–22. Fievre S, Wieruszeski JM, Michalski JC, Lemoine J, Montreuil J & Strecker G (1991) Primary structure of trisialylated oligosaccharide from human milk. Biochemical and Biophysical Research Communications 177, 720–724. Finch JM & Turner RJ (1996) Effects of selenium and vitamin E on the immune responses of domestic animals. Research in Veterinary Science 60, 97–106. Finstad HS, Kolset SO, Holme JA, Wiger R, Farrants AK, Blomhoff R & Drevon CA (1994) Effect of n-3 and n-6 fatty acids on proliferation and differentiation of promyelocytic leukemic HL-60 cells. Blood 84, 3799–3809. Fleming TR & DeMets DL (1996) Surrogate endpoints in clinical trials: are we being misled? Annals of Internal Medicine 125, 605–613. Folkers K, Morita M & McRee J Jr (1993) The activities of coenzyme Q10 and vitamin B6 for immune responses. Biochemical and Biophysical Research Communications 193, 88–92. Fomon SJ & Nelson SE (1994) Size and growth. In Nutrition of the Infant, pp. 36–83 [SJ Fomon, editor]. St Louis, MO: Mosby. Fomon SJ (1991) Reference data for assessing growth of infants. Journal of Pediatrics 119, 415–416. Food and Agriculture Organization/World Health Organization/ United Nations University (1985) Energy and Protein Requirements. WHO Technical Report Series no. 724. Geneva: WHO. Forbes GB (1987) Human Body Composition, Growth, Aging, Nutrition and Activity. New York: Springer-Verlag. Formica JV & Regelson W (1995) Review of the biology of quercetin and related bioflavonoids. Food and Chemical Toxicology 33, 1061–1080. Forni G, Cerruti Sola S, Giovarelli M, Santoni A, Martinetto P & Vietti D (1986) Effect of prolonged administration of low doses of dietary retinoids on cell-mediated immunity and the growth of transplantable tumors in mice. Journal of the National Cancer Institute 76, 527–533. Foufelle F, Girard J & Ferré P (1996) Regulation of lipogenic enzyme expression by glucose in liver and adipose tissue: a review of the potential cellular and molecular mechanisms. Advances in Enzyme Regulation 36, 199–226. Foufelle F, Gouhot B, Pégorier J-P, Perdereau D, Girard J & Ferré P (1992) Glucose stimulation of lipogenic enzyme gene expression in white adipose tissue. Journal of Biological Chemistry 267, 20543–20546. Friel JK, Andrews WL, Matthew JD, Long DR, Cornel AM, Cox M, McKim E & Zerbe GO (1993) Zinc supplementation in very low birth weight infants. Journal of Pediatric Gastroenterology and Nutrition 17, 97–104. Growth, development and differentiation Fujisawa T, Riby J & Kretchmer N (1991) Intestinal absorption of fructose in the rat. Gastroenterology 101, 360–367. Garlick PJ & McNurlan MA (1994) Protein metabolism in the cancer patient. Biochimie 76, 713–717. Gaull GE, Rassin DK, Raiha NCR & Heinonen K (1977) Milk protein quantity and quality in low-birth weight infants: III. Effects on sulfur amino acids in plasma and urine. Journal of Pediatrics 90, 348–355. Gebhard KJ, Gridley DS, Stickney DR & Shulz TD (1990) Enhancement of immune status by high levels of dietary vitamin B-6 without growth inhibition of human malignant melanoma in athymic nude mice. Nutrition and Cancer 14, 15–26. Gerrard JW & Shenassa M (1983) Sensitization to substances in breast milk: recognition, management and significance. Annals of Allergy 51, 300–302. Gerstein HC (1994) Cow’s milk exposure and type I diabetes mellitus. A critical overview of the clinical literature. Diabetes Care 17, 13–19. Gerster H & Moser U (1988) Is high-dose vitamin C intake associated with systemic conditioning? Nutrition Research 8, 1327–1332. Ghersa P, Huber P, Semenza G & Wacker H (1986) Cell-free synthesis, membrane integration and glycosylation of prosucrase-isomaltase. Journal of Biological Chemistry 261, 7969–7974. Ghigo E, Arvat E, Gianotti L, Nicolosi M, Valetto MR, Avagnina S, Bellitti D, Rolla M, Muller EE & Camanni F (1994) Arginine but not pyridostigmine, a cholinesterase inhibitor, enhances the GHRH-induced GH rise in patients with anorexia nervosa. Biological Psychiatry 36, 689–695. Gil A, Corral E, Martı́nez A & Molina JA (1986a) Effects of dietary nucleotides on the microbial pattern of feces of at term newborn infants. Journal of Clinical Nutrition and Gastroenterology 1, 34–38. Gil A & Lozano E, De-Lucchi C, Maldonado J, Molina JA & Pita M (1988) Changes in the fatty acid profiles of plasma lipid fractions induced by dietary nucleotides in infants born at term. European Journal of Clinical Nutrition 42, 473–481. Gil A, Pita M, Martı́nez A, Molina JA & Sánchez-Medina F (1986b) Effect of dietary nucleotides on the plasma fatty acids in at-term neonates. Human Nutrition: Clinical Nutrition 40 C, 185–195. Gil A & Sánchez-Medina F (1981) Acid-soluble nucleotides of cow’s, goat’s and sheep’s milk at different stages of lactation. Journal of Dairy Research 48, 35–44. Gil A & Sánchez-Medina F (1982) Acid-soluble nucleotides of human milk at different stages of lactation. Journal of Dairy Research 49, 301–307. Gil A & Uauy R (1989) Dietary nucleotides in infant nutrition. Journal of Clinical Nutrition and Gastroenterology 4, 145–153. Gil A & Uauy R (1995a) Nucleotides and related compounds in human and bovine milks. In Handbook of Milk Composition, pp. 436–464 [RG Jensen, editor]. San Diego, CA: Academic Press. Gil A & Uauy R (1995b) Nutritional and Biological Significance of Dietary Nucleotides and Nucleic Acids. Barcelona: Limpergraf. Gilsanz V, Roe TF, Mora S, Costin G & Goodman WG (1991) Changes in vertebral bone density in black girls and white girls during childhood and puberty. New England Journal of Medicine 325, 1597–1600. Girard J, Ferré P, Pégorier JP & Duée PH (1992) Adaptations of glucose and fatty acid metabolism during perinatal period and suckling–weaning transition. Physiological Reviews 72, 507–562. Gray GM (1967) Malabsorption of carbohydrates. Federation Proceedings 26, 1415–1419. Grimaldi PA, Knobel SM, Whitesell RR & Abumrad NA (1992) Induction of aP2 gene expression by nonmetabolized long-chain S37 fatty acids. Proceedings of the National Academy of Sciences USA 89, 10930–10934. Gunnes M & Lehman EH (1995) Dietary calcium, saturated fat, fiber and vitamin C as predictors of forearm cortical and trabecular bone density in healthy children and adolescents. Acta Paeditrica 84, 388–392. Güttler F (1984) Phenylketonuria: 50 years since Follings discovery and still expanding our clinical and biochemical knowledge. Acta Paediatrica Scandinavia 73, 705–716. Güttler F & Guldberg P (1996) The influence of mutations of enzyme activity and phenylalanine tolerance in phenylalanine hydroxylase deficiency. European Journal of Pediatrics 155, Suppl. 1, S6–S10. Hachey DL (1994) Benefits and risks of modifying maternal fat intake in pregnancy and lactation. American Journal of Clinical Nutrition 59, Suppl., 454S–464S. Hadsell DL, Greenberg NM, Fligger JM, Baumrucker CR & Rosen JM (1996) Targeted expression of des(1-3) human insulin-like growth factor I in transgenic mice influences mammary gland development and IGF-binding protein expression. Endocrinology 137, 321–330. Hakomori S & Igarashi Y (1995) Functional role of glycosphingolipids in cell recognition and signaling. Journal of Biochemistry, Tokyo 118, 1091–1103. Halas ES, Sanstead HH (1975) Some effects of prenatal zinc deficiency on behavior of the adult rat. Pediatric Research 9, 94–97. Halas ES & Sanstead HH (1980) Malnutrition and behavior: the performance versus learning problem revisited. Journal of Nutrition 110, 1858–1864. Hale AJ, Smith CA, Sutherland LC, Stoneman VE, Longthorne VL, Culhane AC & Williams GT (1996) Apoptosis: molecular regulation of cell death. European Journal of Biochemistry 236, 1–26. Halken S, Host A, Hansen LG & Osterballe O (1993a) Preventive effect of feeding high-risk infants a casein hydrolysate formula or an ultrafiltrated whey hydrolysate formula. A prospective, randomized, comparative clinical study. Pediatric Allergy and Immunology 4, 173–181. Halken S, Host A, Hansen LG & Osterballe O (1993b) Safety of a new, ultrafiltrated whey hydrolysate formula in children with cow milk allergy: a clinical investigation. Pediatric Allergy and Immunology 4, 53–59. Hall L, Emery DC, Davies MS, Parker D & Craig RK (1987) Organization and sequence of the human a-lactalbumin gene. Biochemical Journal 242, 735–742. Hallberg L (1988) Iron balance in pregnancy. In Vitamins and Minerals in Pregnancy and Lactation. Nestle Nutrition Workshop Series, pp. 115–127 [H Berger, editor]. New York, NY: Raven Press. Harris WS, Connor WE & Lindsey S. (1984) Will dietary q-3 fatty acids change the composition of human milk? American Journal of Clinical Nutrition 40, 780. Hauri HP (1986) Use of monoclonal antibodies to investigate the intracellular transport and biogenesis of intestinal brush border proteins. Biochemical Society Transactions 161, 161–163. Hauri HP, Quaroni A & Isselbacher KJ (1980) Monoclonal antibodies to sucrase–isomaltase: probes for the study of postnatal development and biogenesis of the intestinal microvillus membrane. Proceedings of the National Academy of Sciences USA 77, 6629–6633. Hauri HP, Sterchi EE, Bienz D, Fransen JAW & Marxer A (1986) Expression and intracellular transport of microvillus membrane hydrolases in human intestinal epithelial cell. Journal of Cellular Biology 101, 838–851. Hauri HP, Wacker H, Rickli EE, Bigler-Meier B, Quaroni A & Semenza G (1982) Biosynthesis of sucrase–isomaltase. S38 B. Koletzko et al. Purification and NH 2-terminal amino acid sequence of the rat sucrase–isomaltase precursor (pro-sucrase–isomaltase) from fetal intestinal transplants. Journal of Biological Chemistry 257, 4522–4528. He Y, Chu SW & Walker WA (1993) Nucleotide supplements alter proliferation and differentiation of cultured human (CaCo-2) and rat (IEC-6) intestinal epithelial cells. Journal of Nutrition 123, 1017–1027. Hellerstein MK, Wu K, McGrath M, Faix D, George D, Shackleton CH, Horn W, Hoh R & Neese RA (1996) Effects of dietary n-3 fatty acid supplementation in men with weight loss associated with the acquired immune deficiency syndrome: relation to indices of cytokine production. Journal of Acquired Immune Deficiency Syndromes: Human Retrovirology 11, 258–270. Hemila H (1996) Vitamin C and common cold incidence: a review of studies with subjects under heavy physical stress. International Journal of Sports Medicine 17, 379–383. Henderson RC & Hayes PRL (1994) Bone mineralisation in children and adolescents with a milk allergy. Bone and Mineral 27, 1–12. Hernández M & Argente J (1992) Human Growth: Basic and Clinical Aspects. Amsterdam: Excerpta Medica. Hernández-Muñoz R, Dı́az-Muñoz M, Suárez J & Chagoya de Sánchez V (1990) Adenosine partially prevents cirrhosis induced by carbon tetrachloride in rats. Hepatology 12, 242–248. Herskovics A, Quaroni A, Bugge B & Kirsch K (1981) Partial characterization of the carbohydrate units of rat intestinal sucrase–isomaltase. Biochemical Journal 197, 511–514. Houscham KC & Devilliers JFK (1987) Computed tomography in severe protein–energy malnutrition. Archives of Disease in Childhood 62, 589–592. Houtkooper LB, Going SB, Lohman TG, Roche AF & Van Loan M (1992) Bioelectrical impedance estimation of fat free body mass in children and youth: a cross-validation study. Journal of Applied Physiology 72, 366–373. Hughes DA, Wright AJA, Finglas PM, Peerless ACJ, Bailey AL, Astley SB, Pinder AC & Southon S (1996) Beta-carotene supplementation enhances the expression of functionally associated molecules on human monocytes. Biochemical Society Transactions 24, 388S. Hurle JM, Ros MA, Garcia-Martinez V, Macias D & Ganan Y (1995) Cell death in the embryonic developing limb. Scanning Microscopy 9, 519–533. Huttenlocher PR, de Courten C, Garey LJ & van der Loos H (1982) Synaptogenesis in the human visual cortex–evidence for synapse elimination during normal development. Neuroscience Letters 33, 247–252. Hytten FE & Leitch I (1971) The Physiology of Human Pregnancy, 2nd ed. Oxford: Blackwell Scientific Publications. Igarashi P (1994) Transcription factors and apoptosis in kidney development. Current Opinion in Nephrology and Hypertension 3, 308–317. Innis SM, Nelson CM, Lwanga D, Rioux FM & Waslen P (1996) Formula feeding without arachidonic acid and docosahexaenoic acid has no effect on preferential looking acuity or recognition memory in healthy full-term infants at 9 mo of age. American Journal of Clinical Nutrition 64, 40–46. Institute of Medicine (1990) Nutrition During Pregnancy: Weight Gain and Nutrient Supplements. Washington, DC: National Academy Press. Isolauri E, Juntunen M, Rautanen T, Sillanaukee P & Koivula T (1991) A human Lactobacillus strain (Lactobacillus casei sp strain GG) promotes recovery from acute diarrhea in children. Pediatrics 88, 90–97. Isolauri E, Kaila M, Mykkanen H, Ling WH & Salminen S (1994) Oral bacteriotherapy for viral gastroenteritis. Digestive Diseases and Sciences 39, 2595–2600. Jacobson M (1991) Developmental Neurobiology. New York: Plenum Press. James TN (1993) Normal and abnormal consequences of apoptosis in the human heart: from postnatal morphogenesis to paroxysmal arrhythmias. Transactions – American Clinical and Climatological Association 105, 145–177. Jirsova V, Koldovsky O, Heringova A, Hoskova J, Jirasek J & Uher J (1965–66) The development of the functions of the small intestine of the human fetus. Biology of the Neonate 9, 44–49. Johnston CC, Miller JZ, Slemenda CW, Reister TK, Hui S, Christian JC & Peacock M (1992) Calcium supplementation and increase in bone mineral density in children. New England Journal of Medicine 327, 82–87. Jørgensen HL, Schøller J, Sand JC, Bjuring M, Hassager C & Christiansen C (1996) Relation of common allelic variation at vitamin D receptor locus to bone mineral density and postmenopausal bone loss: cross sectional and longitudinal population study. British Medical Journal 313, 586–590. Kaila M, Isolauri E, Soppi E, Virtanen E, Laine S & Arvilommi H (1992) Enhancement of the circulating antibody secreting cell response in human diarrhea by a human Lactobacillus strain. Pediatric Research 32, 141–144. Kandel ER, Schwartz JH & Jessell TM (1991) Principles of Neural Science. East Norwalk, CT: Appleton & Lange. Kandel ER, Schwartz JH & Jessell TM (1995) Essentials of Neural Science. East Norwalk, CT: Appleton & Lange. Karasov WH, Pond RS III, Solberg DH & Diamond JM (1983) Regulation of proline and glucose transport in mouse intestine by dietary substrate levels. Proceedings of the National Academy of Sciences USA 80, 7674–7678. Kashfi K, Yang E, Chowdbury J, Chowdbury N & Dannenberg A (1994) Regulation of UDPGT expression by phenolic antioxidants. Cancer Research 54, 5856–5859. Kawata M (1995) Roles of steroid hormones and their receptors in structural organization in the nervous system. Neuroscience Research 24, 1–46. Kember NF (1993) Cell kinetics and the control of bone growth. Acta Paeditrica Suppl. 391, 61–65. Kemen M, Senkal M, Homann HH, Mumme A, Dauphin AK, Baier J, Windeler J, Neumann H & Zumtobel V (1995) Early postoperative enteral nutrition with arginine- omega-3 fatty acids and ribonucleic acid-supplemented diet versus placebo in cancer patients: An immunologic evaluation of Impact(R). Critical Care Medicine 23, 652–659. Kerstetter JE (1995) Do dairy products improve bone density in adolescent girls? Nutrition Reviews 53, 328–332. Keusch GT (1996) Symposium on vitamin A and immunity: salient points and directions for future research. Journal of Nutritional Immunology 4, 133–143. Kilberg MS, Hutson RG & Laine RO (1994) Amino acid-regulated gene expression in eukaryotic cells. FASEB Journal 18, 13–19. Kilshaw PJ & Cant AJ (1984) The passage of maternal dietary proteins into human breast milk. International Archives of Allergy and Applied Immunology 75, 8–15. Kim SJ, Cho YK, Lee KW (1994) Expression of human lactoferrin in milk of transgenic mice using bovine b-casein/human lactoferrin cDNA fusion gene. Korean Society for Molecular Biology 4, 355–360. Kiremidjian Schumacher L & Stotzky G (1987) Selenium and immune responses. Environmental Research 42, 277–303. Kitagawa H, Nakada H, Fukui S, Funakoshi I, Kawasaki T & Yamashira I (1991) Novel oligosaccharides with the sialyl- Le structure in human milk. Biochemistry 30, 2869–2876. Kitajima H, Sumida Y, Tanaka R, Yuki N, Takayama H & Fujimura M (1997) Early administration of bifidobacterium breve to preterm infants: randomised controlled trial. Archives of Disease in Childhood 76, F101–F107. Growth, development and differentiation Knip M (1992) [Can type-1 diabetes in children be prevented?]. Nordisk Medicin, 107, 207–210. Koldovsky O & Strbak V (1995) Hormones and growth factors in human milk. In Handbook of Milk Composition, pp. 428–436 [RG Jensen, editor]. San Diego, CA: Academic Press. Koldovsky O & Thornburg W (1987) Hormones in human milk: a review. Journal of Pediatric Gastroenterology and Nutrition 6, 172–196. Koletzko B (1992) Fats for brains. European Journal of Clinical Nutrition 46, Suppl. 1, S51–S62. Koletzko B & Braun M (1991) Arachidonic acid and early human growth: is there a relation? Annals of Nutrition and Metabolism 35, 128–131. Koletzko B & Müller J (1990) Cis- and trans-isomeric fatty acids in plasma lipids of newborn infants and their mothers. Biology of the Neonate 57, 172–178. Koletzko B, Thiel I & Abiodun PO (1992) The fatty acid composition of human milk in Europe and Africa. Journal of Pediatrics 120, S62–S70. Kramer TR (1996) Relationship between vitamin A status and T-lymphocyte responsiveness. Journal of Nutritional Immunology 4, 77–85. Krammer PH, Behrmann I, Daniel P, Dhein J & Debatin KM (1994) Regulation of apoptosis in the immune system. Current Opinion in Immunology 6, 279–289. Krenger W, Falzarano G, Delmonte J Jr, Snyder KM, Byon JC & Ferrara JL (1996) Interferon-gamma suppresses T-cell proliferation to mitogen via the nitric oxide pathway during experimental acute graft-versus-host disease. Blood 88, 1113–1121. Kretchmer N, Beard JL & Carlson SE (1996) The role of nutrition in the development of normal cognition. American Journal of Clinical Nutrition 63, 997S–1001S. Kroemer G, Petit P, Zamzami N, Vayssiere JL & Mignotte B (1995) The biochemistry of programmed cell death. FASEB Journal 9, 1277–1287. Kudsk KA, Minard G, Croce MA, Brown RO, Lowrey TS, Pritchard FE, Dickerson RN & Fabian TC (1996) A randomized trial of isonitrogenous enteral diets after severe trauma. An immune-enhancing diet reduces septic complications. Annals of Surgery 531–540; discussion 540–543. Kulkarni AD, Fanslow WC, Rudolph FB &Van Buren CT (1992) Immunohemopoietic effects of dietary nucleotide restriction in mice. Transplantation 53, 467–472. Kulkarni AD, Rudolph FB & Van Buren CT (1994) The role of dietary sources of nucleotides in immune function: a review. Journal of Nutrition 124, 1442S–1446S. Lacey JM & Wilmore DW (1990) Is glutamine a conditionally essential amino acid? Nutrition Reviews 48, 297–309. Laegrid A, Otnaess ABK & Fugelsang J. (1986) Human and bovine milks: comparison of ganglioside compositions and enterotoxin inhibitory activity. Pediatric Research 20, 416–420. Lanting CI, Fidler V, Huisman M, Touwen BCL & Boersma ER (1994) Neurological differences between 9-year-old children fed breast-milk or formula-milk as babies. Lancet 344, 1319–1322. Lechting A, Habicht JP, Delgado H, Klein RE, Yarbrough C & Martorell R (1975) Effect of food supplementation during pregnancy on birth-weight. Pediatrics 56, 508. Leclercq I, Horsmans Y, Desager JP, Pauwels S & Geubel AP (1996) Influence of diet on liver volume and CYP activities in human non alcoholic fatty liver. Hepatology 24, 311A. Lee WTK, Leung SSF, Wang SH, Xu YC, Zeng WP, Lau J, Oppenheimer SJ & Cheng JC (1994) Double-blind controlled calcium supplementation and bone mineral accretion in children accustomed to a low-calcium diet. American Journal of Clinical Nutrition 60, 744–750. LeLeiko NS, Bronstein AD, Baliga S & Munro HN (1983) De novo purine nucleotide synthesis in the rat small and large intestine: S39 effect of dietary protein and purines. Journal of Pediatric Gastroenterology and Nutrition 2, 313–319. LeLeiko NS, Martin BA, Walsh MJ, Kazlow P, Rabinowitz S & Sterling K (1987) Tissue specific gene expression results from a purine and pyrimidine -free diet and 6-mercaptopurine in the rat small intestine and colon. Gastroenterology 93, 1014–1020. LeLeiko NS, Walsh MJ & Abraham S (1995) Gene expression in the intestine: the effect of dietary nucleotides. Advances in Pediatrics 42, 145–169. Lentze MJ (1986) Die Ernährung von Frühgeborenen unter 1500 genterale Voraussetzungen (Feeding of premature infants below 1500 g: enteral preconditions). Monatsschrift für Kinderheilkunde 134, 502–507. Levy Marchal C, Karjalainen J, Dubois F, Karges W, Czernichow P & Dosch HM (1995) Antibodies against bovine albumin and other diabetes markers in French children. Diabetes Care 18, 1089–1094. Lewis DS, Bertrand HA, McMahan CA, McGill HC, Carey KD & Masaro EJ (1986) Preweaning food intake influences the adiposity of young adult baboons. Journal of Clinical Investigation 78, 899–905. Liang B, Lane L & Watson RR (1995) Vitamin E stimulation of disease resistance and immune function. Expert Opinion on Investigational Drugs 4, 201–211. Liang B & Watson RR (1996) Vitamin E and immunomodulation for cancer and AIDS resistance. Expert Opinion on Investigational Drugs 5, 1221–1225. Liang Q & Richardson T (1993) Expression and characterization of human lactoferrin in yeast, Saccharomyces cerevisiae. Journal of Agriculture and Food Chemistry 41, 1800–1807. Licata AA (1993) Monitoring calcium intake and absorption. World Review of Nutrition and Dietetics 73, 27–52. Liley AW (1972) Disorders of the amniotic fluid. In Pathophysiology of Gestation: Fetal Placental Disorders, vol. 2 [NS Assali, editor]. New York, NY: Academic Press. Lilja G, Dannaeus A, Foucard T, Graff LV, Johansson SG & Oman H (1989) Effects of maternal diet during late pregnancy and lactation on the development of atopic diseases in infants up to 18 months of age – in-vivo results. Clinical and Experimental Allergy 19, 473–479. Limonta JM, Castro FO, Martinez R, Puentes P, Ramos B, Aguilar A, Lleonart RL & delaFuente J (1995) Transgenic rabbits as bioreactors for the production of human growth hormone. Journal of Biotechnology 40, 49–58. Lindfors A & Enocksson E (1988) Development of atopic disease after early administration of cow milk formula. Allergy 43, 11–16. Llewellyn-Jones D (1965) Severe anaemia in pregnancy. Australian and New Zealand Journal of Obstetrics and Gynaecology 5, 191–197. Lloyd T, Andon MB, Rollings N, Martel JK, Landis JR, Demers LM, Eggli DF, Kieselhorst K & Kulin HE (1993) Calcium supplementation and bone mineral density in adolescent girls. Journal of the American Medical Association 270, 841–844. Loche S, Carta D, Muntoni AC, Corda R & Pintor C (1993) Oral administration of arginine enhances the growth hormone response to growth hormone releasing hormone in short children. Acta Paediatrica 82, 883–884. Lodish H, Baltimore D, Berk A, Zipursky SL, Matsudaira P & Darnell J (1995) Molecular Cell Biology. New York, NY: Scientific American Books, WH Freeman. Lönnerdal B (1996) Recombinant human milk proteins – an opportunity and a challenge. American Journal of Clinical Nutrition 63, 622S–626S. López-Navarro AT, Gil A & Sánchez-Pozo A (1995) Deprivation of dietary nucleotides results in a transient decrease in acid soluble nucleotides and RNA content in rat liver. Journal of Nutrition 125, 2090–2095. S40 B. Koletzko et al. López-Navarro AT, Ortega MA, Peragon J, Bueno JD, Gil A & Sanchez-Pozo A (1996) Deprivation of dietary nucleotides decreases protein synthesis in the liver and small intestine in rats. Gastroenterology 110, 1760–1769. Lozoff B & Brittenham GM (1986) Behavioural aspects of iron deficiency. Progress in Hematology 14, 23–53. Lozoff B, Wolf AW & Jiminez E (1996) Iron-deficiency anemia and infant development: effects of extended oral iron therapy. Journal of Pediatrics 129, 382–389. Lucas A, Morley R, Cole TJ, Lister G & Leeson-Payne C (1992) Breast milk and subsequent intelligence quotient in children born preterm. Lancet 339, 261–264. Luke B & Petrie RH (1980) Intrauterine growth: correlation of infant birth weight and maternal postpartum weight. American Journal of Clinical Nutrition 33, 2311. Mackinnon AM & Deller DJ (1973) Purine nucleotides biosynthesis in gastrointestinal mucosa. Biochimica et Biophysica Acta 319, 1–4. Maga EA, Anderson GB, Mei CH & Murray JD (1994) Expression of human lysozyme mRNA in the mammary gland of transgenic mice. Transgenic Research 3, 36–42. Majamaa H & Isolauri E (1997) Probiotics: a novel approach in the management of food allergy. Journal of Allergy and Clinical Immunology 99, 179–185. Makrides M, Neumann MA, Byard RW, Simmer K & Gibson RA (1994) Fatty acid composition of brain, retina, and erythrocytes in breast- and formula-fed infants. American Journal of Clinical Nutrition 60, 189–194. Makrides M, Neumann M, Simmer K, Pater J & Gibson R (1995) Are long chain polyunsaturated fatty acids essential nutrients in infancy? Lancet 345, 1463–1468. Makrides M, Simmer K, Goggin M & Gibson RA (1993) Erythrocyte docosahexaenoic acid correlates with the visual response of healthy, term infants. Pediatric Research 34, 425–427. Mandl J, Banhegy G, Kapalos M, Garzo T (1995) Increased oxidation and decreased conjugation of drugs in the liver caused by starvation, altered metabolism of certain aromatic compounds and acetone. Chemicobiological Interactions 96, 87–101. Mantei N, Villa M, Enzler T, Wacker H, Boll W, James P, Hunziger W & Semenza G (1988) Complete primary structure of human and rabbit lactase-phlorizin hydrolase: implications for biosynthesis, membrane anchoring and evolution of the enzyme. EMBO Journal 7, 2705–2713. Marcinkiewicz J, Grabowska A & Chain BM (1996) Is there a role for nitric oxide in regulation of T cell secretion of IL-2? Journal of Immunology 156, 4617–4621. Mastrmatteo E & Sullivan F (1994) Summary: international symposium on the health effects of boron and its compounds. Environmental Health Perspectives 102, Suppl. 7, 139–141. Mathieu C, Waer M, Laureys J, Rutgeerts O & Bouillon R (1994) Prevention of autoimmune diabetes in NOD mice by 1,25 dihydroxyvitamin D3. Diabetologia 37, 552–558. Mayfield R, Yarbrough C & Lechtig A (1991) Body composition of low-birth weight infants determined by using bioelectrical resistance and reactance. American Journal of Clinical Nutrition 54, 296–303. Meikrantz W & Schlegel R (1995) Apoptosis and the cell cycle. Journal of Cell Biochemistry 58, 160–174. Melnik B, Plewig G & Tschung T (1991) Are deficiencies of prostaglandin-E-mediated immunoregulation involved in increased IgE synthesis of atopic mononuclear cells in vitro? Allergy 46, 502–506. Mennella JA & Beauchamp GK (1991) The transfer of alcohol to human milk: effects on flavor and the infant’s behavior. New England Journal of Medicine 325, 981–985. Mennella JA & Beauchamp GK (1993) The effects of repeated exposure to garlic-flavored milk on the nursling’s behavior. Pediatric Research 34, 805–808. Messer M & Kerry KR (1967) Intestinal digestion of maltotriose in man. Biochimica et Biophysica Acta 132, 432–443. Messina M & Messina V (1996) Nutritional implications of dietary phytochemicals. Advances in Experimental Medicine and Biology 401, 207–212. Millar MR, Bacon C, Smith SL, Walker V & Hall MA (1993) Enteral feeding of premature infants with Lactobacillus GG. Archives of Disease in Childhood 69, 483–487. Miller JB, Bull S, Miller J & McVeagh P (1994) The oligosaccharide composition of human milk: temporal and individual variations in the monosaccharide components. Journal of Pediatric Gastroenterology and Nutrition 19, 371–376. Miller LT & Kerkvliet NI (1990) Effect of vitamin B6 on immunocompetence in the elderly. Annals of the New York Academy of Sciences 587, 49–54. Milner JA (1990) Trace minerals in the nutrition of children. Journal of Pediatrics 117, S147–S155. Morillas J, Moltó L, Robles R, Gil A & Sánchez-Pozo A (1994) Lipoprotein changes in small-for-gestational-age infants fed nucleotide-supplemented milk formula. Acta Paediatrica 83, 481–485. Morley R & Lucas A (1993) Early diet and outcome in prematurely born children. Clinical Nutrition 12, Suppl., 6–11. Naeye RL, Diener MM & Dellinger WS (1989) Urban poverty: effects on prenatal nutritition. Science 166, 1026. Nagata S (1997) Apoptosis by death factor. Cell 88, 355–365. Naim HY, Roth J, Sterchi EE, Lentze MJ, Milla PJ, Schmitz J & Hauri HP (1988) Sucrase–isomaltase deficiency in humans. Different mutations disrupt intracellular transport, processing and function of an intestinal brush border enzyme. Journal of Clinical Investigation 82, 667–679. Naim HY, Niermann T, Kleinhans U, Hollenberg CP & Strasser AWM (1991) Strikung structural and functional similarities suggest that intestinal sucrase–isomaltase, human lysosomal a-glucosidase, and Schwanniomyces occidentalis glucoamylase are derived from common ancestral gene. FEBS Letters 294, 109–112. Naim HY, Sterchi EE & Lentze MJ (1989) Structure, biosynthesis and glycosylation of human small intestinal maltase–glucoamylase. Journal of Biological Chemistry 263, 19709–19717. National Research Council (1989) Recommended Dietary Allowances, 10th ed. Washington, DC: National Academy Press. Navarro J, RuizBravo A, JimenezValera M & Gil A (1996) Modulation of antibody-forming cell and mitogen-driven lymphoproliferative responses by dietary nucleotides in mice. Immunology Letters 53, 141–145. Neggers YH, Cutter GR, Alvarez JO, Goldenberg RL, Acton R, Go RC & Roseman JM (1991) The relationship between maternal serum zinc levels during pregnancy and birthweight. Early Human Development 25, 75–85. Neuringer M (1993) The relationship of fatty acid composition to function in the retina and visual system. In Lipids, Learning and the Brain: Fats in Infant Formulas. Report of the 103rd Ross Conference on Pediatric Research, pp. 69–101 [J Dobbing, editor]. Columbus, OH: Ross Laboratories. Newburg DS & Neubauer SH (1995) Carbohydrates in milks: analysis, quantities and significance. In Handbook of Milk Composition, pp. 273–349 [RG Jensen, editor]. San Diego, CA: Academic Press. Newburg DS, Pickering LK, McCluer RH & Cleary TG (1990) Fucosylated oligosaccharides of human milk protect suckling mice from heat stable enterotoxin of Escherichia coli. Journal of Infections and Diseases 162, 1075–1080. Newsholme EA & Carrié AL (1994) Quantitative aspects of Growth, development and differentiation glucose and glutamine metabolism by intestinal cells. Gut 35, Suppl. 1, S13–S17. Niv M, Levy W & Greenstein NM (1963) Yogurt in the treatment of infantile diarrhoea. Clinical Pediatrics 2, 407–411. Núñez MC, Ayudarte MV, Morales D, Suárez MD & Gil A (1990) Effect of dietary nucleotides on intestinal repair in rats with experimental chronic diarrhea. Journal of Parenteral and Enteral Nutrition 14, 598–604. Ogoshi S, Iwasa M, Kitagawa S, Ohmori Y, Mizobuchi S, Iwasa Y & Tamiya T (1988) Effects of total parenteral nutrition with nucleoside and nucleotide mixture on d -galactosamine-induced liver injury in rats. Journal of Parenteral and Enteral Nutrition 12, 53–57. Ogoshi S, Iwasa M, Yonezawa T & Tamiya T (1985) Effect of nucleotide and nucleoside mixture on rats given total parenteral nutrition after 70 % hepatectomy. Journal of Parenteral and Enteral Nutrition 9, 339–342. Ohyanagi H, Nishimatsu S, Kanbara Y, Usami M & Saitoh Y (1989) Effects of nucleosides and a nucleotide on DNA and RNA syntheses by the salvage and de novo pathway in primary monolayer cultures of hepatocytes and hepatoma cells. Journal of Parenteral and Enteral Nutrition 13, 51–58. Oldaeus G, Anjou K, Björksten B, Moran JR & Kjellmann NIM (1997) Extensively and partially hydrolysed infant formulas for allergy prophylaxis. Archives of Disease in Childhood 77, 4–10. Olsen SF, Soerensen JD, Secher NJ, Hedegaard M, Henriksen TB, Hansen HS & Grant A (1992) Randomized controlled trial of the effect of fish-oil supplementation on pregnancy duration. Lancet 339, 1003–1007. Olsen SF, Hansen HS, Sorensen TI, Jensen B, Secher NJ, Sommer S & Knudsen LB (1986) Intake of marine fat, rich in (n-3) polyunsaturated fatty acids, may increase birth-weight by prolonging gestation. Lancet 2, 367–369. Onwude JL, Lilford RJ, Hjartardottir H, Staines A & Tuffnell D (1995) A randomised double blind placebo controlled trial of fish oil in high risk pregnancy. British Journal of Obstetrics and Gynaecology 102, 95–100. Ortega MA, Núñez MC, Gil A & Sánchez-Pozo A (1995) Dietary nucleotides accelerate intestinal recovery after food deprivation in old rats. Journal of Nutrition, 125, 1413–1418. Osmond C, Barker DJP, Winter PD, Fall CHD & Simmonds SJ (1993) Early growth and death from cardiovascular disease in women. British Medical Journal 307, 1519–1524. Patton S, Gendler S & Spicer AP (1995) The epithelial mucin, MUC1, of milk mammary gland and other tissues. Biochimica et Biophysica Acta 1241, 407–424. Peacock M (1991) Calcium absorption efficiency and calcium requirements in children and adolescents. American Journal of Clinical Nutrition 54, 261S–265S. Penn ND, Purkins L, Kelleher J, Heatley RV, MascieTaylor BH & Belfield PW (1991) The effect of dietary supplementation with vitamins A, C and E on cell-mediated immune function in elderly long-stay patients: a randomized controlled trial. Age and Ageing 20, 169–174. Perez RV & Alexander JW (1988) Immune regulation by lipids. Transplantation Proceedings 20, 1162–1165. Peters DP (1979) Effects of prenatal malnutrition on learning and motivation in rats. Physiology and Behaviour 22, 1067–1070. Petralanda I, Yarzabal L & Piessens WF (1988) Parasite antigens are present in breast milk of women infected with Onchocerca volvulus. American Journal of Tropical Medicine and Hygiene 38, 372–379. Pfeffer SR & Rothman JE (1987) Biosynthetic protein transport and sorting by the endoplasmic reticulum and Golgi. Annual Review of Biochemistry 56, 829–852. Philips JA III (1995) Inherited defects in growth hormone synthesis S41 and action. In The Metabolic and Molecular Bases of Inherited Disease, pp. 1965–1983 [CR Scriver, AL Beaudet, WS Sly and D Valle, editors]. New York, NY: McGraw-Hill. Pietrzik K, Prinz R, Reusch K, Bung P, Mallmann P & Chronides A (1992) Folate status and pregnancy outcome. Annals of the New York Academy of Sciences 669, 371–373. Pita ML, Fernández MR, DeLucchi C, Medina A, Martı́nezValverde A, Uauy R & Gil A (1988) Changes in fatty acids pattern of red blood cell phospholipids induced by type of milk. Dietary nucleotide supplementation and postnatal age in preterm infants. Journal of Pediatric Gastroenterology and Nutrition 7, 740–747. Pizzini RP, Kumar S, Kulkarni AD, Rudolph FB & Van Buren CT (1990) Dietary nucleotides reverse malnutrition and starvationinduced immunosuppression. Archives of Surgery 125, 86–90. Pocock NA, Eisman JA, Hopper JL, Yeates MG, Sambrook PN & Eberl S (1987) Genetic determinants of bone mass in adults: a twin study. Journal of Clinical Investigation 80, 706–710. Politt E (1993) Iron deficiency and cognitive function. Annual Review of Nutrition 13, 521–537. Prasad AS (1995) Zinc: an overview. Nutrition 11, 93–99. Prentice AM, Cole TJ, Foord FA, Lamb WH & Whitehead RG (1987) Increased birth weight after prenatal supplementation of rural African women. American Journal of Clinical Nutrition 46, 912–925. Price JS, Oyajobi BO & Russell RGG (1994) The cell biology of bone growth. European Journal of Clinical Nutrition 48, Suppl. 1, S131–S149. Prieto PA, Mukerji P, Kelder B, Erney R, Gonzales D, Yun JS, Smith DF, Moremen KW, Nardelli C, Pierce M, Li Y, Chen X, Wagner T, Cummings R & Kopchick JJ (1995) Remodeling of mouse milk glycoconjugates by transgenic expression of a human glycosyltransferase. Journal of Biological Chemistry 270, 29516–29519. Quan R & Gray GM (1993) Sucrase-a-dextrinase in the rat. Journal of Clinical Investigation 91, 2785–2790. Rakic P (1995) Corticogenesis in human and nonhuman primates. In The Cognitive Neurosciences, pp.127–145 [MS Gazzaniga, editor]. Cambridge MA: MIT Press. Rall LC & Meydani SN (1993) Vitamin B6 and immune competence. Nutrition Reviews 51, 217–225. Ramam L, Rajalakshmi K, Krishnamachari KAVR & Sastry KG (1978) Effect of calcium supplementation to undernourished mothers during pregnancy on the bone density of neonates. American Journal of Clinical Nutrition 21, 466–469. Rasmussen KM (1992) The influence of maternal nutrition on lactation. Annual Review of Nutrition 12, 103–117. Ravelli GP, Stein ZA & Susser MW (1976) Obesity in young men after famine exposure in utero and early infancy. New England Journal of Medicine 295, 349–353. Reisbick S (1996) Neural development. In Assessment of Infant Visual and Cognitive Function in Relation to Long-chain Polyunsaturated Fatty Acids, pp.10–18 [S Carlson, M Neuringer and S Reisbick, editors]. Basel: Editiones Roche. Ribot C, Trémolieres F & Pouilles JM (1995) Late consequences of a low peak bone mass. Acta Paediatrica Scandinavica 411, 31–35. Riby J, Fujisawa T & Kretchmer N (1993) Fructose absorption. American Journal of Clinical Nutrition 58, Suppl., 748S–753S. Rigo J, Nyamugabo K, De Curtis M & Senterre J (1996) Bone mineralization during the first year of life. In Recent Developments of Infant Nutrition, pp. 98–111 [JG Bindels, AC Goedhart and HKA Visser, editors]. Dordrecht, Boston and London: Kluwer Academic Publishers. Ripa S & Ripa R (1995) Zinc and immune function. Minerva Medica 86, 315–318. Rodeberg DA, Chaet MS, Bass RC, Arkovitz MS & Garcia VF S42 B. Koletzko et al. (1995) Nitric oxide: an overview. American Journal of Surgery 170, 292–303. Roberts JA, Hill GW & Propelle AJ (1974) Maternal protein deprivation and toxemia of pregnancy studies in the rhesus monkey (Macacca mulatta). American Journal of Obstetrics and Gynecology 118, 14. Rosales FJ & Kjolhede C (1994) A single 210-mumol oral dose of retinol does not enhance the immune response in children with measles. Journal of Nutrition 124, 1604–1614 . Rosen JM, Li S, Raught B & Hadsell D (1996) The mammary gland as a bioractor: factor regulating the efficient expression of milk protein-based transgenes. American Journal of Clinical Nutrition 63, 627S–632S. Rosenzweig MR, Leiman AL & Breedlove SM (1996) Biological Psychology. Sunderland, MA: Sinauer Association. Roskelley CD, Srebrow A & Bissell MJ (1995) A hierarchy of ECM-mediated signalling regulates tissue-specific gene expression. Current Opinion in Cell Biology 7, 736–747. Rudolph FB (1994) Symposium: dietary nucleotides: a recently demonstrated requirement for cellular development and immune function. Journal of Nutrition 124, 1431S–1432S. Ruiz-Palacios GM, Cervantes LE, Newburg DS, López-Vidal Y & Calva JJ (1992) In vitro models for studying campilobacter infections. In Campilobacter Jejuni. Current Status and Future Trends, pp.176–183 [I Nachamkin, MJ Blaser and LS Tomkins, editors]. Washington, DC: American Society of Microbiology. Saavedra JM, Bauman NA, Oung I, Perman JA & Yolken RH (1994) Feeding of Bifidobacterium bifidum and Streptococcus thermophilus to infants in hospital for prevention of diarrhoea and shedding of rotavirus. Lancet 344, 1046–1049. Sabharwal H, Sjoblad S & Lundblad A (1991) Sialylated oligosaccharides in human milk and feces of preterm, full-term and weaning infants. Journal of Pediatric Gastroenterology and Nutrition 12, 280–284. Sachs L (1996) The control of hematopoiesis and leukemia: from basic biology to the clinic. Proceedings of the National Academy of Sciences USA 93, 4742–4749. Sánchez-Pozo A, Morillas J, Moltó L, Robles R & Gil A (1994) Dietary nucleotides influence lipoprotein metabolism in newborn infants. Pediatric Research 35, 112–116. Sánchez-Pozo A, Ramirez M, Gil A, Maldonado J, Van Biervliet JP & Rosseneu M (1995) Dietary nucleotides enhance plasma lecithin cholesterol acyl transferase activity and apolipoprotein A-IV concentration in preterm newborn infants. Pediatric Research 37, 328–333. Sanders EJ & Wride MA (1995) Programmed cell death in development. International Review of Cytology 163, 105–173. Sanderson IR & He Y (1994) Nucleotide uptake and metabolism by intestinal epithelial cells. Journal of Nutrition 124, 131S–137S. Sato G, Pardee A & Sirbasku D (1982) Growth of cells in hormonally defined media. Cold Spring Harbor Conference on Cell Proliferation 9. New York: Cold Spring Harbor Laboratory. Sauerwald TU, Hachey DL, Jensen CL & Heird WC (1997) New insights into the metabolism of long-chain polyunsaturated fatty acids during infancy. European Journal of Medical Research 2, 88–92. Savaiano DA & Clifford AJ (1981) Adenine, the precursor of nucleic acids in intestinal cells, is unable to synthesize purines de novo. Journal of Nutrition 111, 1816–1822. Savilahti E, Tainio VM, Salmenperä L, Siimes MA & Perheentupa J (1987) Prolonged exclusive breast feeding and heredity as determinants in infantile atopy. Archives of Disease in Childhood 62, 269–273. Sazawal S, Bentley M, Black RE, Dhingra P, George S & Bhan MK (1996) Effect of zinc supplementation on observed activity in low socioeconomic Indian preschool children. Pediatrics 98, 1132–1137. Sazawal S, Black RE, Bhan MK, Jalla S, Bhandari N, Sinha A & Majumdar S (1996) Zinc supplementation reduces the incidence of persistent diarrhea and dysentery among low socioeconomic children in India. Journal of Nutrition, 126, 443–450. Schiffrin EJ, Rochat F, Link Amster H, Aeschlimann JM & Donnet Hughes A (1995) Immunomodulation of human blood cells following the ingestion of lactic acid bacteria. Journal of Dairy Science 78, 491–497. Scholl TO & Hediger ML (1994) Anemia and iron-deficiency anemia: compilation of data on pregnancy outcome. American Journal of Clinical Nutrition 59, Suppl., 492S–501S. Schoonjans K, Staels B & Auwerx J (1996) Role of the peroxisome proliferator-activated receptor (PPAR) in mediating the effects of fibrates and fatty acids on gene expression. Journal of Lipid Research 37, 907–925. Schroten H, Plogmann R, Hanisch FG, Hacker J, Nobis-Bosch R & Wahn V (1993) Inhibition of adhesion of S-fimbriated E. coli to buccal epithelial cells by human skim milk is predominantly mediated by mucins and depends on the period of lactation. Acta Paediatrica 82, 6–11. Schwartz GG (1992) Multiple sclerosis and prostate cancer: what do their similar geographies suggest? Neuroepidemiology 11, 244–254. Scientific Committee for Food with the assistance of Aggett PJ, Arnal PJ, Arnal M, Bender D, Buss D, Christiansen C, Garabedian M, Grossklaus R, Hallberg L, Hermus R, Koletzko B, Pitt G, Schriver J & Scott J (1993) Nutrient and Energy Intakes for the European Community. Reports of the Scientific Committee for Food (31st series). Luxembourg: Office of Official Publications of the European Communities. Scriver CR (1991) Phenylketonuria – genotypes and phenotypes. New England Journal of Medicine 324, 1280–1281. Scriver CR, Byck S, Prevost L & Hoang L (1996) The phenylalanine hydroxylase locus: a marker for the history of phenylketonuria and human genetic diversity. PAH Mutation Analysis Consortium. Ciba Foundation Symposium 197, 73–90. Scriver CR & Clow CL (1980) Phenylketonuria: epitome of human biochemical genetics. New England Journal of Medicine 303, 1336–1342. Secher NJ & Olsen SF (1990) Fish oil and preeclampsia. British Journal of Obstetrics and Gynaecology 97, 1077–1079. Sei Y, Vitkovic L & Yokoyama MM (1995) Cytokines in the central nervous system: regulatory roles in neuronal function, cell death and repair. Neuroimmunomodulation 2, 121–133. Seidman E, LeLeiko N, Ament M, Berman W, Caplan D, Evans J, Kocoshis S, Lake A, Motil K & Sutphen J (1991) Nutritional issues in pediatric inflammatory bowel disease. Journal of Pediatric Gastroenterology and Nutrition 12, 424–438. Semba RD (1996) Linking vitamin A and childhood immunizations. Journal of Nutritional Immunology 4, 87–109. Semba RD, Munasir Z, Beeler J, Akib A, Muhilal, Audet S & Sommer A (1995) Reduced seroconversion to measles in infants given vitamin A with measles vaccination [see comments]. Lancet 345, 1330–1332. Senkal M, Kemen M, Homann HH, Eickhoff U, Baier J & Zumtobel V (1995) Modulation of postoperative immune response by enteral nutrition with a diet enriched with arginine, RNA, and omega-3 fatty acids in patients with upper gastrointestinal cancer. European Journal of Surgery, Acta Chirurgica 161, 115–122. Shaw NJ & Bishop NJ (1995) Mineral accretion in growing bones – a framework for the future? Archives of Disease in Childhood 72, 177–179. Sheard NF (1994) Iron deficiency and infant development. Nutrition Reviews 52, 137–140. Sherman AR (1992) Zinc, copper, and iron nutriture and immunity. Journal of Nutrition 122, 604–609. Growth, development and differentiation Shiner M, Nichols VN, Barrish JP & Nichols BL (1990) Pathogenesis of small-intestinal mucosal lesions in chronic diarrhea of infancy: II. An electron microscopic study. Journal of Pediatric Gastroenterology and Nutrition 11, 464–480. Shor Posner G, Miguez Burbano MJ, Lu Y, Feaster D, Fletcher M, Sauberlich H & Baum MK (1995) Elevated IgE level in relationship to nutritional status and immune parameters in early human immunodeficiency virus-1 disease. Journal of Allergy and Clinical Immunology 95, 886–892. Sibai BM, Villar MA, Bray E (1989) Magnesium supplementation during pregnancy: a double blind randomized controlled clinical trial. American Journal of Obstetrics and Gynecology 161, 115. Sigurs N, Hattevig G & Kjellman B (1992) Maternal avoidance of eggs, cow’s milk, and fish during lactation: effect on allergic manifestations, skin-prick tests, and specific IgE antibodies in children at age 4 years. Pediatrics 89, 735–739. Sikorski R, Juszkiewicz T & Paszkowski T (1990) Zinc status in women with premature rupture of membranes at term. Obstetrics and Gynecology 76, 675. Sjöström H, Norén O, Christiansen L, Wacker H & Semenza G (1980) A fully active, two-active site, single-chain sucrase– isomaltase from pig small intestine. Implications for the biosynthesis of a mammalian integral stalked membrane protein. Journal of Biological Chemistry 255, 11332–11338. Sjöström H, Norén O, Danielsen EM & Skovbjerg H (1983) Structure of microvillar enzymes in different phases of their life cycles. In Brush Border Membranes, Ciba Foundation Symposium 95, pp. 50–72 [R Porter and GM Collins, editors]. London: Pitman Books. Slosman DO, Rizzoli R & Bonjour J-Ph (1995) Bone absorptiometry: a critical appraisal of various methods. Acta Paediatrica 411, Suppl., 9–11. Smart JL (1974) Activity and exploratory behavior of adult offspring of undernourished mother rats. Developmental Psychobiology 7, 315–321. Spatling L & Spatling G (1988) Magnesium supplementation in pregnancy: a double-blind study. British Journal of Obstetrics and Gynaecology 95, 120–125. Speck R & Lauterburg B (1991) Fish oil protects mice against acetaminophen hepatotoxicity in vivo. Hepatology 13, 557–561. Spencer TD, Keen RW, Arden NK, Morrison NA, Major PJ, Nguyen TV, Kelly PJ, Baker JR, Sambrook PN & Lanchbury JS (1995) Influence of vitamin D receptor genotype on bone mineral density in postmenopausal women: a twin study in Britain. British Medical Journal 310, 1357–1360. Stabell C, Bale C, Pedro Da Silva AP & Aaby P (1995) No evidence of fontanelle-bulging episodes after vitamin A supplementation of 6- and 9-month-old infants in Guinea Bissau. European Journal of Clinical Nutrition 49, 73–74. Stanbury JB (editor) (1994) The Damaged Brain of Iodine Deficiency: Cognitive, Behavioral, Neuromotor, and Educative Aspects. Elmsford, NY: Cogniznat Communication. Stansbridge EM, Walker V, Hall MA, Millar MR, Bacon C & Chen S (1993) Effects of feeding premature infants with Lactobacillus GG on gut fermentation. Archives of Disease in Childhood 69, 488–492. Steegers-Theunissen RPM, Boers GHJ, Trijbels FJM & Eskes TKAB (1991) Neural tube defects and derangement of homocysteine metabolism. New England Journal of Medicine 324, 199–200. Stephens S (1986) Development of secretory immunity in breast fed and bottle fed infants. Archives of Disease in Childhood 61, 263–269. Stephens S, Brenner MK, Duffy SW, Lakhani PK, Kennedy CR & Farrant J (1986a) The effect of breast-feeding on proliferation S43 by infant lymphocytes in vitro. Pediatric Research 20, 227– 231. Stephens S, Duffy SW & Page C (1986b) A longitudinal study of gamma-interferon production by peripheral blood mononuclear cells from breast- and bottle-fed infants. Clinical and Experimental Immunology 65, 396–400. Stoch MB, Smythe PM, Moodie AD & Bradshaw O (1982) Psychosocial outcome and CT findings after gross undernourishment during infancy: a 20-year developmental study. Developmental Medicine and Child Neurology 24, 419–436. Stowell KM, Rado TA, Funk WD & Tweedie JW (1991) Expression of cloned human lactoferrin in baby hamster kidney cells. Biochemical Journal 276, 349–355. Strbak V (1991) Hormones and bioactive substances in milk. Endocrine Regulation 25, 3–143. Suarez A, Faus MJ & Gil A (1996a) Dietary long-chain polyunsaturated fatty acids modify heart, kidney, and lung fatty acid composition in weanling rats. Lipids 31, 345–348. Suarez A, Faus MJ & Gil A (1996b) Dietary supplementation with long-chain polyunsaturated fatty acids increases susceptibility of weanling rat tissue lipids to in vitro lipid peroxidation. Journal of Nutritional Biochemistry 7, 252–260. Suzuki H, Menegazzi M, Carcereri de Prati A, Mariotto S & Armato U (1995) Nitric oxide in the liver: physiopathological roles. Advances in Neuroimmunology 5, 379–410. Tanaka M, Lee K, Martı́nez-Augustin O, He Y, Sanderson IA & Walker W (1996) Exogenous nucleotides alter the proliferation, differentiation and apoptosis of human small intestinal epithelium. Journal of Nutrition 126, 424–433. Taylor EM & Morgan EH (1990) Developmental changes in transferrin and iron uptake by the brain in the rat. Developmental Brain Research 55, 35–42. Tengerdy RP (1990) The role of vitamin E in immune response and disease resistance. Annals of the New York Academy of Sciences 587, 24–33. Thépot D, Devinoy E, Fontaine ML, Stinnakre MG, Massoud M, Kann G & Houdébine LM (1995) Rabbit whey acidic protein gene upstream region controls high-level expression of bovine growth hormone in the mammary gland of transgenic mice. Molecular Reproduction and Development 42, 261–267. Thomasset M (1994) Vitamine D et système immunitaire (Vitamin D and immune system). Pathologie Biologie 42, 163–172. Tixier-Vidal A, Picart R, Loudes C & Bauman AF (1986) Effects of polyunsaturated fatty acids and hormones on synaptogenesis in serum-free medium cultures of mouse fetal hypothalamic cells. Neuroscience 17, 115–132. Torre PM, Ronnenberg AG, Hartman WJ & Prior RL (1993) Oral arginine supplementation does not affect lymphocyte proliferation during endotoxin-induced inflammation in rats. Journal of Nutrition, 123, 481–488. Trivedi P, Risteli J, Risteli L, Hindmarsh PC, Brook CG & Mowat AP (1991) Serum concentration of the type I and III procollagen propeptides as biochemical markers of growth velocity in healthy infants and children and in children with growth disorders. Pediatric Research 30, 276–280. Troncone R, Scarcella A, Donatiello A, Cannataro P, Tarabuson A & Auricchio S (1987) Passage of gliadin into human breast milk. Acta Paediatrica Scandinavica 76, 453–456. Uauy R, Birch DG, Birch EE, Tyson JE & Hoffman DR (1990) Effect of dietary omega-3 fatty acids on retinal function of verylow-birth-weight neonates. Pediatric Research 28, 485–492. Uauy R, Quan R & Gil A (1994) Role of nucleotides in intestinal development and repair: implications for infant nutrition. Journal of Nutrition 124, 1436S–1441S. Uauy R, Stringel G, Thomas R & Quan R (1990) Effect of dietary S44 B. Koletzko et al. nucleosides on growth and maturation of the developing gut in the rat. Journal of Pediatric Gastroenterology and Nutrition 4, 497–503. Vaarala O, Klemetti P, Savilahti E, Reijonen H, Ilonen J & Akerblom HK (1996) Cellular immune response to cow’s milk beta-lactoglobulin in patients with newly diagnosed IDDM. Diabetes 45, 178–182. Van Buren CT, Kulkarni AD, Fanslow WC & Rudolph FB (1985) Dietary nucleotides: a requirement for helper/inducer T lymphocytes. Transplantation 40, 694–697. Van Buren CT, Kulkarni AD, Schandle VB & Rudolph FB (1983) The influence of dietary nucleotide on cell-mediated immunity. Transplantation 36, 350–352. Vandenplas Y, Hauser B, Van Den Borre C, Clybouw C, Mahler T, HachimiIdrissi S, Deraeve L, Malfroot A & Dab I (1995) The long-term effect of a partial whey hydrolysate formula on the prophylaxis of atopic disease. European Journal of Pediatrics 154, 488–494. Vandenplas Y, Hauser B, Van den Borre C, Sacre L & Dab I (1992) Effect of a whey hydrolysate on the prophylaxis of atopic disease. Annals of Allergy 68, 419–424. Vaulont S & Kahn A (1994) Transcriptional control of metabolic regulation of genes by carbohydrates. FASEB Journal 8, 28– 35. Vaux DL & Strasser A (1996) The molecular biology of apoptosis. Proceedings of the National Academy of Sciences USA 93, 2239–2244. Vermeer C, Gijsbers BL, Craciun AM, Groenen Van Dooren MM & Knapen MH (1996) Effects of vitamin K on bone mass and bone metabolism. Journal of Nutrition 126, 1187S– 1191S. Villar J, Repke J, Belizan JM & Pareja G (1987) Calcium supplementation reduces blood pressure during pregnancy: results of a randomized controlled clinical trial. Obstetrics and Gynecology 70, 317–322. Virtanen SM, Räsänen L, Aro A, Lindström J, Sippola H, Lounamaa R, Toivanen L, Tuomilehto J & Akerblom HK (1991) Infant feeding in Finnish children <7 yr of age with newly diagnosed IDDM. Diabetes Care 14, 415–417. Viverge D, Grimmonprez L, Cassanas G, Bardet L & Solere M (1990) Variations in oligosaccharides and lactose in human milk during the first week of lactation. Journal of Pediatric Gastroenterology and Nutrition 11, 361–364. Wainwright PE (1993) Lipids and behaviour: the evidence from animal models. In Lipids, Learning and the Brain: Fats in Infant Formulas. Report of the 103rd Ross Conference on Pediatric Research, pp. 69–101 [J Dobbing, editor]. Columbus OH: Ross Laboratories. Wainwright PE (1997) Essential fatty acids and behavior: is there a role for the eicosanoids. In Handbook of Essential Fatty Acid Biology. Biochemistry, Physiology and Behavioral Neurobiology, pp. 299–341 [S Yehuda and DI Mostofsky, editors]. Totowa, NJ: Humana Press. Walsh MJ, Sánchez-Pozo A & LeLeiko NS (1990) A regulatory element is characterized by purine-mediated and cell type specific gene transcription. Molecular Cell Biology 10, 4356–4364. Walsh MJ, Tsao KL & LeLeiko NS (1992) Characterization of DNA-protein interactions within a distal regulatory element upstream of a mammalian housekeeping gene promoter. Journal of Biological Chemistry 267, 7026–7035. Walsh CT, Sandstead HH, Prasad AS, Newberne PM & Fraker PJ (1994) Zinc: health effects and research priorities for the 1990s. Environmental Health Perspectives 2, 5–46. Wang X, Sato R, Brown MS, Hua X & Goldstein JL (1994) SREBP-1, a membrane bound transcription factor released by sterol-regulated proteolysis. Cell 77, 53–62. Ward PP, Lo JY, Duke M, May GS, Headon DR & Conneely OM (1992b) Production of biologically active recombinant human lactoferrin in Aspergillus oryzae. BioTechnology 10, 784–789. Ward PP, May GS, Headon DR & Conneely OM (1992b) An inducible expression system for the production of human lactoferrin in Aspergillus nidulans. Gene 122, 219–223. Warner JO (1980) Food allergy in fully breast fed infants. Clinical Allergy 10, 133–136. Watanabe S, Sakai N, Yasui Y, Kimura Y, Kobayashi T, Mizutani T & Okuyama H (1994) A high alpha-linolenate diet suppresses antigen-induced immunoglobulin E response and anaphylactic shock in mice. Journal of Nutrition 124, 1566– 1573. Watson JD, Hopkins NH, Roberts JW, Steitz JA & Weiner AM (1987) Molecular Biology of the Gene, 4th ed. Menlo Park, CA: The Benjamin/Cummings Publishing Co. Wells JL, James DK, Luxton R & Pennock CA (1987) Maternal leukocyte zinc deficiency at start of third trimester as a predictor of fetal growth retardation. British Medical Journal 294, 1054. Welply JK, Keene JL, Schmuke JJ & Howard SC (1994) Selectins as potential targets of therapeutic intervention in inflammatory diseases. Biochimica et Biophysica Acta 1197, 215–226. Welsh AO (1993) Uterine cell death during implantation and early placentation. Microscopy Research and Technique 25, 223–245. Widdowson EM (1982) Importance of nutrition in development with special reference to feeding low-birth weight infants. In Meeting Nutritional Goals for Low Birth Weight Infants. Proceedings of the 2nd Ross Clinical Research Conference, pp. 4–11 [HS Sauls, WL Bachuber and LA Lewis, editors]. Columbus, OH: Ross Laboratories. Williams C, Highley W, Ma EH, Lewis J, Tolbert B, Woullard D, Kirmani S & Chung RA (1981) Protein, amino acid and caloric intakes of selected pregnant women. Journal of the American Dietetics Association 78, 28–35. Winick M (1970) Fetal malnutrition. Clinical Obstetrics and Gynecology 13, 526. World Health Organization (1993) Global Prevalence of Iodine Deficiency Disorders. MDIS Working Paper no. 1. Geneva: WHO. Worthington-Roberts B (1985) Nutritional deficiencies and excesses: impact on pregnancy, part 2. Journal of Perinatology 5, 12. Wride MA & Sanders EJ (1995) Potential roles for tumour necrosis factor alpha during embryonic development. Anatomy and Embryology, Berlin 191, 1–10. Xue-Yi C, Xin-Min J, Zhi-Hong D, Rakeman MA, Ming-Li Z, O’Donnell K, Tai M, Amette K, DeLong N & DeLong R (1994) Timing of vulnerability of the brain to iodine deficiency. New England Journal of Medicine 331, 1739–1744. Yetiv JZ (1988) Clinical applications of fish oils [see comments]. Journal of the American Medical Association 260, 665–670. Zakeri Z & Lockshin RA (1994) Physiological cell death during development and its relationship to aging. Annals of the New York Academy of Sciences 719, 212–229. Zeiger RS & Heller S (1995) The development and prediction of atopy in high-risk children: follow-up at age seven years in a prospective randomized study of combined maternal and infant food allergen avoidance. Journal of Allergy Clinical Immunology 95, 1179–1190. Zeiger RS, Heller S, Mellon MH, Forsythe AB, O0 Connor RD, Hamburger RN & Schatz M (1989) Effect of combined maternal and infant food-allergen avoidance on development of atopy in early infancy: a randomized study [published Growth, development and differentiation erratum appears in Journal of Allergy and Clinical Immunology (1989) 84 (5 Pt 1), 677]. Journal of Allergy and Clinical Immunology 84, 72–89. Zeiger RS, Heller S, Mellon MH, Halsey JF, Hamburger RN & S45 Sampson HA (1992) Genetic and environmental factors affecting the development of atopy through age 4 in children of atopic parents: a prospective, randomized study of food allergen avoidance. Pediatric Allergy and Immunology 3, 110–127. q Nutrition Society 1998 British Journal of Nutrition (1998), 80, Suppl. 1, S47–S75 S47 Functional food science and substrate metabolism W. H. M. Saris1¬ , N. G. L. Asp2 , I. Björck2 , E. Blaak1 , F. Bornet3 , F. Brouns4 , K. N. Frayn5 , P. Fürst6 , G. Riccardi7 , M. Roberfroid8 and M. Vogel9 1 Nutrition Research Centre, Department of Human Biology, Maastricht University, PO Box 616, NL-6200 MD Maastricht, The Netherlands 2 Chemical Centre, Applied Nutrition and Food Chemistry, University of Lund, PO Box 124, S-221 00, Lund, Sweden 3 Vilvoorde Research and Development Centre, Nutrition and Health Service, Havenstraat 84, B-1800 Vilvoorde, Belgium 4 Novartis Nutrition Research Unit, PO Box 1350, NL-6201 BJ Maastricht, The Netherlands 5 Oxford Lipid Metabolism Group, Sheikh Rashid Laboratory, Radcliffe Infirmary, Oxford OX2 6HE, UK 6 Institute of Biological Chemistry and Nutrition, University of Hohenheim, PO Box 700 562, Garbenstrasse 30, D-70593 Stuttgart, Germany 7 Dipartimento di Medicina Clinica e Sperimentale, Universitá degli Studi di Napoli – Federico II, Via Sergio Pansini 5, I-80131 Napoli, Italy 8 UCL, Ecole de Pharmacie, Tour Van Helmont, Avenue E. Mounier, B-1200 Brussels, Belgium 9 Dipl. Biochemiker–Biochemisches Lab, Südzucker AG Mannheim/Ochsenfurt, Zafes, Wormser Strasse 11, D-67190 Obrigheim, Pfalz, Germany Contents 1. Introduction 2. Chronic diseases related to energy balance and substrate regulation 2.1. Obesity 2.1.1. Genetic contribution to obesity 2.2.2. Regulation of energy balance 2.2.3. Costs of obesity 2.2. Insulin resistance syndrome 2.2.1. Features predisposing to the insulin resistance syndrome 2.3. Diabetes 2.3.1. Aetiology 2.4. Undernutrition 2.4.1. Definition 2.4.2. Present position 2.4.3. Pathophysiology and adaptive responses 2.4.4. Body composition 2.4.5. Geriatric undernutrition 2.5. Conclusions and further research 3. Metabolic conditions related to these chronic diseases 3.1. Body-weight control 3.1.1. Energy balance, macronutrient balance and body-weight regulation 3.1.2. Type of carbohydrates 3.1.3. Type of fat 3.1.4. Alcohol S48 S48 S48 S48 S49 S49 S49 S50 S50 S51 S51 S51 S51 S51 S51 S52 S52 3.2. 3.3. S53 S53 3.4. S53 S54 S54 S55 3.1.5. Macronutrient replacement 3.1.6. Dietary components stimulating thermogenesis 3.1.7. Physiological and metabolic consequences of undernutrition 3.1.8. Conclusions and further research Insulin resistance/sensitivity 3.2.1. Introduction 3.2.2. Dietary carbohydrates 3.2.3. Dietary fat 3.2.4. Niacin and insulin sensitivity 3.2.5. Minerals 3.2.6. Conclusions and further research Blood glucose control 3.3.1. Introduction 3.3.2. Nutritional influence on fasting and postprandial blood glucose levels 3.3.3. Food properties determining the glycaemic index 3.3.4. Indigestible carbohydrates and glucose metabolism: possible mechanisms of action 3.3.5. Conclusions and further research Plasma triacylglycerols 3.4.1. Introduction 3.4.2. Dietary carbohydrates 3.4.3. Dietary fat 3.4.4. Conclusions and further research S55 S57 S57 S58 S59 S59 S59 S59 S60 S60 S60 S60 S60 S61 S63 S64 S65 S66 S66 S66 S66 S66 Abbreviations: ALP, atherogenic lipoprotein phenotype; GI, glycaemic index; IDDM, insulin-dependent diabetes mellitus; IRS, insulin resistance sydnrome; MCT, medium-chain triacylglycerol; NEFA, non-esterified fatty acids; NIDDM, non-insulin-dependent diabetes mellitus; P : S ratio, polyunsaturated : saturated fatty acid ratio; PUFA, polyunsaturated fatty acids; SCFA, short-chain fatty acids. *Corresponding author: Dr W. H. M. Saris, fax þ31 43 367 0976, email [email protected] S48 W. H. M. Saris et al. 4. Nutrition, substrate metabolism and physical performance 4.1. Introduction 4.2. Carbohydrates 4.3. Fat 4.4. Protein S67 S67 S67 S68 S68 4.5. 4.6. 4.7. 4.8. 4.9 4.10. Fluid and electrolytes Minerals Trace elements Vitamins Ergogenic supplements Conclusions and further research S69 S69 S69 S69 S69 S70 Abstract The present review addresses the role of food constituents in the aetiology of metabolic conditions and chronic diseases, mostly related to energy metabolism and substrate regulation, such as obesity and non-insulin-dependent diabetes mellitus. Second, attention is paid to malnutrition, a major cause of mortality and morbidity in developing countries, which may be a cause of concern in Europe because of the increasing number of elderly people in the population. Finally, the role of diet during exercise, a condition of enormous substrate demands, is evaluated. Based on a critical evaluation of the existing knowledge in the literature, implications for future research in relation to functional foods are discussed. Functional foods: Substrate metabolism: Energy balance: Physical activity 1. Introduction It is clear that with modern food and physical activity habits, the human metabolic regulatory mechanisms are challenged. There is a growing imbalance between foodrelated energy intake and physical activity-related energy expenditure. Also, the balance in macronutrient intake has changed dramatically over the past few decades. Therefore, a number of specific body functions, mostly related to energy metabolism and substrate regulation, are at risk. A number of chronic diseases such as obesity, non-insulindependent diabetes mellitus (NIDDM) and osteoporosis can be identified as being directly related to changes in food intake and physical activity-related behaviour. At the same time, however, malnutrition thought of as a major cause of mortality and morbidity only in the developing countries, is a cause for concern in Europe, mainly because of the increasing number of elderly people in the population. There are also specific groups in whom undernutrition poses a particular problem, especially those with concurrent chronic or severe illness. This chapter will briefly evaluate the metabolic changes, and the underlying specific metabolic conditions as well as the role of specific dietary components in the aetiology of the previously mentioned chronic diseases. Of special interest is the relation between nutrition and physical performance. During physical stress such as exercise, the substrate demands are enormous. Nutrition plays a crucial role in this process. Therefore, a balanced diet with a carefully planned mix of food ingredients can play an important role in the level of performance. In this way well-designed and effective sports foods have been proven to be clear examples of functional foods. 2. Chronic diseases related to energy balance and substrate regulation 2.1. Obesity Obesity is defined as an excessive accumulation of body fat. Its prevalence may vary in different populations between 5 and 50 %, also depending on the definition of overweight (mild v. moderate v. severe obesity). A BMI (body weight (kg)/height (m)2 ) between 25 and 30 kg/m2 is defined as mild obesity (in USA: BMI > 27 kg/m2 ), a BMI in the range of >30–35 kg/m2 is moderate obesity, and a BMI > 35 kg/m2 is severe obesity (Björntorp & Brodoff, 1992). The high incidence of obesity in affluent societies is recognized as a major health problem (VanItallie, 1992). Obesity is associated with an increased risk of developing hypertension, insulin resistance, diabetes and cardiovascular disease. Apart from these major health hazards, obesity is thought to cause a great variety of health problems, some of which are listed in Table 1. 2.1.1. Genetic contribution to obesity. It has been recognized that there is an important genetic contribution to obesity. The fact that a sedentary lifestyle and high-fat Table 1. Health disorders and other problems thought to be caused or exacerbated by obesity (Based on VanItallie, 1992 and Frayn, 1996) Problems Cardiovascular disease Hypertension Non-insulin-dependent diabetes mellitus Gallstones Impaired reproductive and sexual function Reduced fertility (males) Polycystic ovary syndrome (females) Increased risk of breast and endometrial cancer Osteoarthritis in weight-bearing joints Impaired self-image, depression, suicides Accidents Possible cause Elevated LDL-cholesterol, decreased HDL-cholesterol, hypertriacylglycerolaemia, hypertension May result indirectly from insulin resistance Insulin resistance Increased cholesterol flux into bile (possibly related to insulin resistance and high insulin concentrations) Decreased androgens, increased oestrogen production in adipose tissue Increased oestrogen production in adipose tissue* Excess weight bearing Psychological aspects of poor body image Loss of mobility * Increased oestrogen production occurs because adipose tissue contains the enzyme aromatase, which converts androgens (e.g. testosterone) into oestrogens. Functional foods and substrate metabolism diet may predispose towards obesity, independent of genetics, has been equally well recognized and appreciated (Prentice & Jebb, 1995). The interaction of genetic predisposition and environmental factors is the most commonly accepted model for the genetics of human obesity (Bouchard & Perusse, 1993). The relative importance of these two factors in different populations may vary across studies. The heritability of obesity on the basis of studies of monozygotic twins raised together or apart has ranged from 20 to 60 %. The best current estimate of heritability of body fat is 35 % (Bouchard & Perusse, 1993). If another 15 % is allocated for subjects who are predisposed to obesity when a ‘Western high-fat’ diet is consumed, then genetic inheritance can account for about 50 % of the variation in percentage body fat (Campfield et al. 1997). 2.2.2. Regulation of energy balance. Obesity develops under conditions where energy intake exceeds energy expenditure. Much of the research in the obesity field has focused on the possible deviations in the different components of the energy balance equation. There is some evidence that a low resting energy expenditure may contribute to the development of obesity and that the metabolic rate during rest is, to some extent, genetically determined (Ravussin et al. 1988). Also, a blunted dietinduced thermogenesis response has been reported in obesity (Segal et al. 1990), but this has not been found consistently (D’Allesio et al. 1992). Thus, the relative importance of a low rate of energy expenditure in the development of obesity and the extent of individual differences in susceptibility to obesity remain controversial. A major difference in energy balance between lean and obese subjects is the increased energy expenditure in obese subjects, proportional to their increased fat-free mass, and as a consequence the proportionally increased energy intake (Prentice et al. 1986). Within the past decade, attention has been shifted to the investigation of the balances of the individual macronutrients, proteins, carbohydrates and fat, in the development of obesity. There are indications that subjects predisposed to obesity, or obese subjects, may have a diminished ability to oxidize fat, which would make them more susceptible to a positive fat balance on a high-fat diet (Zurlo et al. 1990; Blaak et al. 1994). These topics will be dealt with further in section 3.1. Extensive attention to the control of food intake is given in another paper in the present supplement (Bellisle et al. 1998). 2.2.3. Costs of obesity. The direct economic costs of obesity, defined as the costs related to diversion of resources to diagnosis and treatment of obesity as well as the treatment of obesity itself, have been estimated to vary between 2 and 5 % of total health-care costs of various countries. In this calculation, only the costs of moderate and severe obesity (BMI > 30 kg/m2 ) are taken into account. The costs associated with mild obesity (BMI between 25 and 30 kg/m2 ) may also be substantial, since a large proportion of the adult population is involved. These costs comprise costs of health services (visits to general practitioners, consultation with medical specialists, medication; Seidell, 1997). In addition societal costs (loss of productivity, disability pensions, premature death), and personal costs (job discrimination, higher premiums to life S49 insurance companies) contribute substantially to health-care costs, but many issues of these costs are too fragmentary to allow calculation (Seidell, 1997). Thus, there is much direct information to show that obesity contributes significantly to health-care costs. 2.2. Insulin resistance syndrome It has been recognized for many years that a number of adverse metabolic changes tend to cluster within individuals (Table 2). The underlying change appears to be a decrease in insulin sensitivity (insulin resistance) although cause and effect have never been properly disentangled. Nevertheless, plausible mechanisms have been described whereby development of insulin resistance can lead to the other aspects of the syndrome. Each of these changes has been shown independently to relate to risk of cardiovascular disease, and insulin resistance is a strong marker of risk of development of NIDDM (Reaven, 1995). The terms insulin resistance syndrome (IRS) or metabolic syndrome are now preferred to the original term syndrome X. This syndrome may be very prevalent. Some studies suggest that about 25 % of non-diabetic adults manifest this syndrome (Reaven, 1995). The concept of the IRS has been enormously helpful in understanding current patterns of chronic disease, especially CHD and NIDDM. As a risk factor for CHD it is probably greater in absolute terms than elevated cholesterol concentrations (Després, 1993). The characteristic dyslipidaemia of insulin resistance includes elevated fasting plasma triacylglycerol and depressed HDL-cholesterol concentrations. These changes are associated with a preponderance of small, dense LDL particles. This dyslipidaemia has been called the atherogenic lipoprotein phenotype (ALP) (Griffin & Zampelas, 1995). Postprandial triacylglycerol concentrations may be even more important than fasting plasma triacylglycerol concentrations. It has been suggested that the ALP results from exaggerated postprandial lipaemia associated with insulin resistance (Frayn, 1993). Insulin sensitivity is usually measured as the ability of insulin to stimulate glucose disposal, usually by means of the hyperinsulinaemic, euglycaemic clamp procedure. Table 2. Metabolic changes associated with insulin resistance syndrome Glucose metabolism Hyperinsulinaemia Glucose intolerance Lipid metabolism Hypertriacylglycerolaemia Exaggerated postprandial lipaemia Decreased HDL-cholesterol concentrations Preponderance of small, dense LDL particles Other Hypertension Increased coagulation Clinical correlates Cardiovascular disease NIDDM Gout Breast cancer NIDDM, non-insulin-dependent diabetes mellitus. S50 W. H. M. Saris et al. Insulin is infused at a rate chosen to elevate the plasma insulin concentration by a predetermined amount, and glucose is infused at a variable rate, as necessary to maintain a constant plasma glucose concentration (hence the glucose concentration is ‘clamped’). Under these conditions of constant glycaemia, glucose disposal must equal the rate of glucose entry into the plasma, which is known since it is being infused. (A correction can be made for endogenous glucose production by infusion of a glucose tracer.) Thus, the rate of glucose disposal from plasma at a predetermined insulin concentration is known, and may be used to compare different subjects. An alternative procedure involves mathematical modelling of the disappearance of glucose following its intravenous injection (Ader & Bergman, 1987). However, these specialized procedures are unsuitable for large-scale epidemiological studies. A simple ‘proxy’ may be measurement of the fasting plasma insulin and glucose concentrations. These may be referred to a simple mathematical model which interprets them in terms of insulin secretion and insulin sensitivity (Matthews et al. 1985). Even the fasting plasma insulin concentration alone provides useful information, and elevated fasting plasma insulin concentrations have been shown to be associated (for instance) with development of CHD independently of other lipid markers (Després et al. 1996). Although insulin sensitivity is measured in terms of glucose disposal, insulin resistance appears to affect many metabolic processes other than glucose disposal. Insulin resistance of multiple aspects of lipid metabolism may be the explanation for the characteristic dyslipidaemia of insulin resistance (Frayn, 1993). Thus, there must be some common change leading to multiple aspects of insulin resistance. The effects of insulin on metabolism are mediated through binding to specific receptors on cell surfaces. The intracellular domain of the insulin receptor possesses an intrinsic tyrosine kinase (EC 2.7.1.112) activity which is activated upon insulin binding, and this initiates the intracellular signalling chains leading ultimately to changes in enzyme expression and activity. In principle, insulin sensitivity might affect any one of the steps in these signal chains, which are themselves divergent for different aspects of insulin action. However, because insulin resistance, as commonly observed, affects so many diverse metabolic functions, it is likely that an early and common step is mainly affected. Attention has focused on events around the insulin receptor. One possible mechanism for widespread modulation of sensitivity to insulin might be a change in membrane fluidity caused by the incorporation into the membrane of different amounts of cholesterol or of phospholipids containing different fatty acids. 2.2.1. Features predisposing to the insulin resistance syndrome. The IRS may exist in apparently healthy individuals of normal body weight. There is evidence for a genetic component (Austin et al. 1990; Mitchell et al. 1996). However, it is most commonly associated with obesity, particularly obesity involving upper body fat distribution (often called visceral obesity). Since obesity and fat distribution are themselves in part genetically determined (Bouchard, 1992), there may be a considerable genetic component to the IRS. Nevertheless, it must also be strongly influenced by environmental factors including diet and activity level, because it is considered to have markedly increased in prevalence in recent years in Western countries. Development of the IRS may be the reason for the high incidence of CHD and NIDDM in immigrant people from the Indian sub-continent (McKeigue et al. 1991). This highlights both the possible genetic predisposition of some groups, and the effect of a change in lifestyle imposed on a susceptible genetic make-up. Many of the features of the IRS shown in Table 2 could also be described as the effects of a sedentary lifestyle. Therefore, physical inactivity, development of obesity and genetic make-up probably interact with dietary factors to explain the high prevalence of the IRS in Western societies. A number of studies have indicated that about 50 % of variation in insulin sensitivity amongst individuals is accounted for by variations in body fat content together with some measure of aerobic capacity (Bogardus et al. 1985). 2.3. Diabetes Diabetes mellitus is a disease characterized by increased plasma glucose concentrations due to a reduced insulin action at its target tissues (insulin resistance) and/or an impaired insulin secretion. High glucose levels sustained for several years induce structural abnormalities in the walls of small arteries at the level of the retina and kidney and in the peripheral nerves; moreover, they contribute (directly and/or indirectly) to acceleration of the atherosclerotic process in the large arteries at the level of heart, brain, and lower limbs (Table 3). Diabetes mellitus is classified into two large groups in relation to its clinical manifestations and aetiology (American Diabetes Association, 1997a). Type 1 or insulindependent diabetes mellitus (IDDM) usually develops in young lean individuals; it has an abrupt onset and requires insulin treatment to prevent the occurrence of a severe metabolic derangement that leads to death within a few days. This type of diabetes is due to an almost complete destruction of pancreatic b cells which represents a consequence of an autoimmune process. Therefore, IDDM is characterized by plasma insulin levels that are very low or even unmeasurable. Type II diabetes (NIDDM) usually develops in elderly individuals who are very often overweight; it has a very slow onset (it may be asymptomatic for several years) and does not necessarily require insulin treatment. It develops as a consequence of two metabolic derangements occurring together: (1) insulin resistance at the level of liver, muscle, adipose tissue; (2) impaired insulin secretion (which nevertheless is never completely Table 3. Long-term diabetic complications and their possible health implications Complications Retinopathy Nephropathy Neuropathy Macroangiopathy Health implications Blindness Dialysis Foot ulcers, impotence Stroke, myocardial infarction, foot gangrene Functional foods and substrate metabolism suppressed). In relation to the balance between these two metabolic derangements, plasma insulin concentrations can be increased, normal or slightly decreased in comparison with non-diabetic individuals. Nutrition has very important implications for diabetes in relation to aetiology, management and prevention of complications (Riccardi, 1994). 2.3.1. Aetiology. Different nutritional factors have been implicated in the aetiology of IDDM and NIDDM (WHO Study Group, 1994). For IDDM, two nutritional factors have been described in association with increased susceptibility to the disease. The first one is represented by a short duration of breast-feeding. It is not clear, so far, whether it is breast-feeding per se which has a protective effect on IDDM development or if an early introduction of dairy products and/or solid foods in the infant diet might have a predisposing effect. However, since duration of breastfeeding and age at introduction of dairy products and/or solid foods are highly correlated, it is difficult to dissect the independent effect of each of them on the risk of IDDM. The second nutritional factor linked with IDDM is represented by consumption of N-nitroso compounds (chemically related to streptozotocin, a well-known b-cell toxin). Studies in laboratory animals and ecological surveys in man have shown an increased risk of developing IDDM when N-nitroso compounds are consumed both by the parents at the time of conception and by the progeny during infancy. Further studies are needed, particularly intervention trials in human subjects, before the link between these dietary factors and IDDM is fully established (Virtanen & Aro, 1994). In relation to the aetiology of NIDDM a large body of evidence indicates obesity, particularly the visceral type, as an important cause of NIDDM (WHO Study Group, 1994). It is now generally accepted that both degree and duration of obesity are associated with an increased risk of NIDDM. Moreover, visceral adiposity represents a risk factor for the development of NIDDM independently of the degree of overweight. As a matter of fact, obesity, particularly visceral obesity, induces resistance to the action of insulin at its target tissues (probably due to high non-esterified fatty acid (NEFA) concentrations) thus deteriorating glucose tolerance in the presence of an impaired insulin secretion (this might also be secondary to high NEFA concentrations). Weight reduction can lower the risk of NIDDM and, in persons already affected by the disease, it can ameliorate insulin sensitivity and glucose tolerance. Dietary factors for treating diabetes should be able to limit glycaemic excursions, which are difficult to keep under control, even with the use of the available glucose lowering medications, and which negatively affect the overall blood glucose control. In addition, the diabetes diet should be able to improve the cardiovascular risk profile by reducing plasma lipid levels and blood pressure values, which are often elevated in diabetic patients. This aim can be achieved by dietary measures that are in principle identical to those recommended for the non-diabetic population (Riccardi & Rivellese, 1991; Diabetes and Nutrition Study Group EASD, 1995). Nutritional factors that might possibly play a role in the aetiology of NIDDM are: (1) a low consumption of fibre-rich foods; (2) a low S51 consumption of fish; (3) a high consumption of fat, particularly the saturated type. Two recent epidemiological studies indicate that low-glycaemic-index (GI) diets (low glycaemic load) might protect against the development of NIDDM (Salmerón et al. 1997a, b). However, so far none of these nutritional factors has been unequivocally proven as being implicated in the aetiology of NIDDM (Virtanen & Aro, 1994; Storlien et al. 1996). 2.4. Undernutrition 2.4.1. Definition. A condition of undernutrition or malnutrition occurs due to consumption of a diet deficient in one or more food constituents (specific undernutrition), or insufficient consumption of an otherwise adequate diet (general undernutrition). This statement is not a tautology; it makes the point that the adequacy of a diet, the extent to which it meets the requirements, can only be defined in terms of functions of the consumer: growth, health, activity etc. (Waterlow, 1992). Nutritional deficiency results from an imbalance between the body’s requirements for nutrients and energy and the supply of these substrates of metabolism. Specific undernutrition is usually associated with biochemical changes preceding the clinical sign, thus allowing the early diagnosis of subclinical or impending deficiency. A general undernutrition is reflected by failure to grow or by loss of weight. It results usually from a quantitatively inadequate intake. The two conditions may or may not coexist. 2.4.2. Present position. Despite the extensive understanding of human nutritional requirements, malnutrition (undernutrition) is one of the main causes of morbidity and mortality in developing regions of the world, especially in young children. In technologically advanced societies undernutrition no longer constitutes a major hazard to health, but occurs in especially vulnerable groups in various ways. Some of these are related to the introduction and widespread use in recent years of techniques such as parenteral feeding and renal dialysis, or other treatments in the frame of artificial nutrition. Ageing, chronic alcoholism, drug abuse or even the medically supervised use of drugs, and food faddism may lead to deficiency disease states. 2.4.3. Pathophysiology and adaptive responses. Undernutrition may develop gradually (days or months). This process allows a series of metabolic and behavioural adjustments that result in decreased nutrient demand. If the supply of nutrients becomes persistently lower than that to which the body can adapt, a critical situation supervenes. Metabolic equilibrium can also be disrupted in aged individuals during the progression of a disease or as a result of inadequate therapeutic measures (Torún & Viteri, 1988). 2.4.4. Body composition. An undernourished individual differs in two ways from a normal subject: in the relative proportions of the various organs and tissues, and in the chemical composition of the body. In malnutrition there is a preferential loss of muscle tissue which in the resting state has a low metabolic activity, while organs with high rates of activity are relatively well preserved. It is generally accepted that in undernutrition the proportion of extracellular water is increased. This is in S52 W. H. M. Saris et al. contrast to the condition of severe malnutrition in which excess extracellular fluid is only measurable in about 50 % of the cases. A decreased share of intracellular water is consistently found in malnutrition and undernutrition. This is important considering the recent hypothesis concerning the regulatory role of the cellular hydration state (Häussinger et al. 1993). Therefore, the measurement of the distribution of body water is fundamental to understanding the changes in body composition that occur in undernutrition. It is not inconceivable that undernutrition is associated with intracellular K or Mg depletion, while moderate decreases of these cations might be observed in serum. Ca and P have received very little attention. It seems reasonable to assume that Ca deficiency may well occur in undernutrition. The intracellular Ca concentration is about 1000 times lower than the extracellular; the constancy of the cytoplasmic Ca concentration is of great importance, since it plays a fundamental role in the integrated control of membrane permeability, the cellular response to stimulation and intracellular signalling. There seems to be a clear case for studies on the possibility of P deficiency in undernutrition. Nevertheless, in the case of muscle wasting, total body K might be decreased. The low insulin action and diminished intracellular energy substrates reduce the availability of ATP and phosphocreatine. This process probably alters the cellular exchange of Na and K, leading to K loss and increased intracellular Na. Water accompanies the Na influx, and although total body intracellular water, as mentioned, is decreased because of losses in lean body mass, there may occasionally be intracellular overhydration. These alterations in cell electrolytes may explain, at least in part, the increased fatiguability and reduced strength of skeletal muscle. 2.4.5. Geriatric undernutrition. At present, about 12.5 % of the population are over 65 years of age; and in the next 25 years the number of 80-year-old individuals may grow to over 20 %. Nutrition and ageing research yields several unique challenges. Undernutrition may be common among older persons (morbid consequences) due Table 4. Prevalence of risk factors for protein–(energy) undernutrition in the population of 65 years and older (Adapted from Miller et al. 1995) Risk factor Social factors poverty isolation Psychological factors depression dementia widowed Physical factors impaired mobility needs help with shopping, meal preparation, feeding visual deficit poor dentition chewing difficulty Functional impairment impaired instrumental ADL impaired basic ADL ADL, activities of daily living. Prevalence (%) to under-consumption of macro- and micronutrients, protein malnutrition being frequently observed, and deficiencies in the amino acids, glutamine and arginine, should be considered. Changes in nutrient digestion, absorption or metabolism may contribute to undernutrition though the nutrients are within the recommended limits for the general public. Ageing is associated with impaired immune responses and increased infection-related morbidity. Protein undernutrition impairs, especially, cell-mediated immunity. Previous studies substantiate the hypothesis that an optimum intake of nutrients (protein) leads to a striking reduction of proteindeficiency-induced cardiac failure and decreases the risk of infection. These findings are of considerable clinical and public health importance (Chandra, 1992; Miller et al. 1995). Several factors common among older adults have been shown to increase the risk of undernutrition (Table 4). The prevalence of risk factors for protein (energy) undernutrition in the elderly population ranges between 2 and 69 %. Despite the common occurrence of protein–energy undernutrition in older persons, its presence is rarely recognized and even when it is observed, it is often not treated (Move & Bohmer, 1991). In Table 5 the major conditions that appear to be secondary to the development of undernutrition are listed. 2.5. Conclusions and further research Obesity. Obesity is a major health hazard, associated with a great variety of health problems. The research with respect to body-weight regulation has shifted from the study of possible abnormalities in the energy balance equation towards the study of the individual macronutrients. These concepts, conclusions and recommendations for further research will be further discussed in section 3. Insulin resistance syndrome. The IRS may be an important conceptual link between obesity, NIDDM and cardiovascular disease. Insulin resistance of severity comparable with that of NIDDM is thought to be present in 25 % of non-diabetic individuals. The IRS is associated with a characteristic dyslipidaemia, the ALP. Impaired lipid metabolism in the postprandial period (exaggerated postprandial lipaemia) may be the link between insulin resistance and the ALP. The present knowledge base is lacking in several important areas. 15 30 What is the basis for the IRS in people of normal body weight? (This will require epidemiological and genetic studies.) 3–6 5 35 Table 5. Effects of protein undernutrition in ageing (Adapted from Morley, 1995) 8 20 8 20 35 4 8 Decubitus ulcers Infection Immune dysfunction (decreased CD4þ, CD8þ) Falls Euthyroid sick syndrome Anaemia (decreased maximum breathing capacity) Decreased bone mass Decreased glomerular filtration rate Weight loss Functional foods and substrate metabolism How prevalent is the ALP in non-obese, non-diabetic subjects? The ALP has been described as a genetic condition, and yet it also appears to be a secondary consequence of the IRS. Can the genetic and environmental components be dissected? Diabetes. Diabetes can be classified on the basis of its clinical manifestations and aetiology into type 1 or IDDM, due to almost complete destruction of the pancreatic b-cells as a consequence of an auto-immune process, and type 2 or NIDDM, associated with insulin resistance and impaired insulin secretion. Two nutritional factors linked with IDDM are the duration of breast-feeding and the consumption of N-nitroso-compounds during infancy and by the parents at the time of conception. Further studies are required, in particular intervention trials in human subjects, before the link between these dietary factors and IDDM is fully established. Type 2 diabetes is associated with ageing and obesity, particularly visceral obesity. Further research is necessary to elucidate the mechanisms behind the impaired insulin resistance, in particular in relation to obesity and intermediary carbohydrate and fat metabolism. Undernutrition. Undernutrition is an important risk factor, being the main cause of morbidity and mortality in developing regions, especially in children. In the Western world, undernutrition occurs in some vulnerable groups especially in the elderly as outlined in section 2.4. The physiology of undernutrition related to various organ functions and its influence on immunity as well as recommendations for further research are outlined in section 3.1.8. Protein undernutrition, in particular, impairs immunity. Furthermore, with the development of undernutrition, changes occur in the extra- and/or intracellular concentrations of K, Mg and Ca, leading to a shift in extra- and/or intracellular fluid and control of membrane permeability and intracellular signalling. Muscle wasting is frequently observed and the observed shift in muscle electrolytes may explain the reported fatiguability and reduced strength. 3. Metabolic conditions related to these chronic diseases 3.1. Body-weight control 3.1.1. Energy balance, macronutrient balance and bodyweight regulation. According to the classical energy balance equation, obesity develops when the equilibrium between energy intake and energy expenditure shifts towards a positive balance. The excess energy is stored in the form of triacylglycerol. More recently, evidence has accumulated that energy balance can only be achieved in the case of macronutrient balance and interest has shifted to an investigation of the balances of the different macronutrients, carbohydrates, fats and proteins, in the aetiology of obesity (Blaak & Saris, 1995; Frayn, 1995; Hill & Prentice, 1995; Flatt, 1996). Achievement of macronutrient balance requires that the net oxidation of each nutrient equals the average amount of the same macronutrient in the diet. Oxidation of the different macronutrients appears to take place in a hierarchical manner, some substrates being more readily S53 oxidized than others. Oxidation of carbohydrates and proteins tends to vary in response to the recent intake of each fuel in an autoregulatory manner. Adjustments in carbohydrate oxidation are capable of efficiently maintaining carbohydrate balance in the face of large changes in carbohydrate intake. Many experiments confirm this carbohydrate-driven autoregulation (Hill & Prentice, 1995; Flatt, 1996). In contrast, in the shorter term fat intake does not promote its own oxidation when carbohydrate content is constant. It has been demonstrated that over a 9 h period the same amounts of fat, carbohydrate and protein are oxidized whether or not the test meal is supplemented with extra fat (Schutz & Jéquier, 1989). The underlying mechanisms between the differences in macronutrients may be related to the characteristics of the macronutrient stores. This theory fits in the framework described by Flatt (1996) and is based on the fact that the carbohydrate stores of the body are limited (illustrated in Table 6) and are only capable of covering oxidation for a few days, which requires a tight metabolic control of carbohydrate balance. In contrast, the capacity for fat storage is enormous (Table 6), which implies that rapid adjustment of fat oxidation to intake is not necessary. When a diet high in fat is consumed over longer periods of time, the inability to acutely adjust fat oxidation will result in a positive energy balance and expanding fat stores and obesity. This will in turn result in an increased NEFA release from the expanded fat stores and an increased fat oxidation, until a situation of fat balance is again reached. Recent studies indicate that subjects predisposed to obesity, or obese subjects, may have a diminished ability to oxidize fatty acids (Zurlo et al. 1990; Blaak et al. 1994), which would make them even more susceptible to positive fat balance on a high dietary fat intake. However, further long-term well-controlled studies have to be performed in different types of subjects to obtain more information on the long-term effect of carbohydrate–fat exchange on energy expenditure and substrate utilization. The size of the protein pool, in relation to daily intake, may lead one to predict that the regulation of protein balance would resemble more the regulation of fat than carbohydrate balance, whereas the opposite is true. In this respect, it may be more realistic to think in terms of the body’s free amino acid pool since most of the body’s protein pool is in a relatively inert form (Frayn, 1995). Also, in several studies it has been shown that the carbohydrate : fat ratio of the diet may affect food intake. It has been shown that over a 2-week period compensation is less accurate when the diet has been diluted by removal of fat than by removal of carbohydrates (Lissner et al. 1987). There is now substantial evidence which shows that high-fat diets undermine the body’s ability to regulate energy intake Table 6. Macronutrient reserves and daily and annual macronutrient intake (Data from Flatt, 1996) Carbohydrate Intake (kg/year) Body content (kg) Intake (kg/d) (% body content) 100 0.5 0.3 60 Fat Protein 35 20 0.1 0.5 35 12 0.1 1.0 S54 W. H. M. Saris et al. in line with requirements and that they induce high-fat overfeeding. Several reports on the relationship between diet and the prevalence of obesity show that a higher fat intake is associated with a higher BMI (Gibney et al. 1987; Miller et al. 1990; Lovejoy & DiGirolamo, 1992). There remains a debate as to whether high-fat overfeeding is a passive overconsumption simply due to an energy density effect or whether carbohydrates play a specific role in inhibiting food intake (macronutrient effect; Flatt, 1996). As mentioned earlier, Flatt’s glycogenostatic hypothesis argues that the body’s different storage capacities for carbohydrates and fats result in mechanisms which give priority to the maintenance of stable glycogen levels, and that changes in these glycogen stores provide the primary feedback signal to regulate appetite (Flatt, 1996). In contrast, studies by Prentice and colleagues (Prentice & Poppit, 1996) showed that high-fat hyperphagia was eliminated when diets were manipulated to have the same energy density. Further studies are necessary to elucidate this issue. In the whole debate on the interaction of the carbohydrate : fat ratio of the diet and energy metabolism, less attention has been paid to possible differences of various types of carbohydrates and fats in the regulation of energy and macronutrient balance. 3.1.2. Type of carbohydrates. Monosaccharides, disaccharides and starch are defined as ‘digestible’ carbohydrates because they are digested and absorbed in the human small intestine. A second category of carbohydrates, defined as ‘non-digestible’, (i.e. dietary fibre) cannot be digested by intestinal enzymes. This latter category of carbohydrates, including a fraction of starch called resistant starch, may still have an important nutritional role because of their inhibitory effects on food intake (Blundell & Burley, 1987) and their possible role in weight management (Blundell & Burley, 1987; Hamilton & Anderson, 1992). This paragraph will further focus on the metabolic effects of digestible carbohydrates, also called glycaemic carbohydrates (Food and Agriculture Organization/World Health Organization, 1998). The ingestion of different types of digestible carbohydrates may lead to varying metabolic postprandial responses, which implies the possibility that different types of carbohydrates have varying effects on thermogenesis and substrate utilization. Indeed, differences in postprandial thermogenesis among various types of digestible carbohydrate have been reported, with sucrose and fructose being more thermogenic than glucose and readilydigestible starch (Tappy et al. 1986; Blaak & Saris, 1996). Additionally, carbohydrate oxidation, glycogen formation and the decrement in lipid oxidation have been reported to be higher after fructose than glucose ingestion. The higher thermogenic response with sucrose and fructose ingestion may be ascribed to differences in post-ingestive substrate utilization and is probably related to particularities of fructose metabolism in the liver. The higher increase in post-ingestive carbohydrate oxidation with fructose and sucrose compared with glucose may be explained by the finding that fructokinase (EC 2.7.1.4), responsible for fructose phosphorylation, is about ten times as active as the combined activities of glucokinase (EC 2.7.1.2) and hexokinase (EC 2.7.1.1), required for glucose phosphorylation (Blaak & Saris, 1996). Despite an increased carbohydrate oxidation with fructose, leaving less fructose available for storage, it has previously been reported that fructose is also a better substrate for hepatic glycogen synthesis (Nilsson & Hultman, 1974). Moreover, glycogen formation from fructose requires more energy than glycogen formation from glucose and starch, and may therefore be of importance in the higher thermogenic response with fructose and sucrose as compared with glucose and starch (Blaak & Saris, 1996). Furthermore, acute thermogenesis studies showed no differences in thermogenesis after ingestion of an easilydigestible maize starch and glucose (Blaak & Saris, 1996) or cooked manioc starch and glucose (Ritz et al. 1991), indicating that there are no differences in the thermogenic efficiency with which the body handles available carbohydrates with varying chain lengths. All the previously mentioned studies are acute experimental studies. Evidence for differences in the effects of digestible carbohydrates on long-term energy expenditure and substrate utilization is lacking. As indicated earlier the carbohydrate : fat ratio of the diet has a distinct effect on appetite regulation. In the current literature there are no indications that various types of digestible carbohydrates may differ in their effects on food intake or energy intake. However, it is a popular belief that starches and mono- and disaccharides may have varying effects on energy intake and that the attractiveness of ‘simple’-sugar-rich foods may promote overeating and may thereby contribute to the development of obesity (Yudkin, 1988). In man, taste preferences for both fat and sugar have been investigated. Studies examining sweetness have not revealed any difference in sensory functioning between normal-weight and obese individuals (Grinker, 1978). In addition, several studies indicated a negative relationship between preference for sweet taste and degree of body fatness and a strong positive relationship between body fatness and preferences for fatty foods (Drewnowski et al. 1985, 1992). This seems consistent with reports on the relationship between diet and the prevalence of obesity, showing that a higher fat intake is associated with a lower carbohydrate and sugar intake, which is in turn associated with a higher BMI (Gibney et al. 1987; Miller et al. 1990; Lovejoy & DiGirolamo, 1992; Bolton-Smith & Woodward, 1994). In summary, it can be said that the available literature indicates a closer regulation of carbohydrate compared with fat balance and that there is no conclusive evidence indicating differences in the effects of different types of digestible carbohydrates on long-term energy and substrate balance. 3.1.3. Type of fat. As for the type of carbohydrate, the impact of type of fat on substrate metabolism and balance is often ignored. Modifications in dietary fat profile may affect body weight and adiposity through changes in partitioning between oxidation and storage and/or alterations in membrane structure (Pan et al. 1994). Significant differences were observed in oxidation of stearate, oleate and linoleate with the oxidation rates for oleate sixteen times as high as for stearate (Jones et al. 1985). Furthermore, a low polyunsaturated : saturated fatty acid (P : S) ratio of the diet was associated with an increased basal fat oxidation, and a lower contribution of fat to the Functional foods and substrate metabolism thermic effect of feeding than a diet with a high P : S ratio in man, suggesting that the long-chain fatty acid composition of dietary fat modulates the oxidation of fat and carbohydrate after chronic feeding and after meal feeding (Jones & Schoeller, 1988). Another study showed that obese subjects consuming low P : S ratio diets exhibited a reduced contribution of fat oxidation to the thermogenic response, compared with lean individuals consuming low or high P : S ratio diets (Jones et al. 1992). Postprandial thermogenesis has been reported to be higher after a medium-chaintriacylglycerol meal than after a long-chain-triacylglycerol meal, indicating that besides the long-chain fatty acid composition of the diet the chain length of the fatty acid may be important in determining its metabolic effect (Seaton et al. 1986). Although these results indicate that the type of fatty acid may affect energy expenditure and substrate utilization, more (long-term) research is needed for these effects to result in nutritional implications. 3.1.4. Alcohol. The role of alcohol in human energy metabolism and human obesity is still a matter of debate. The consumption of excessive amounts of alcohol is usually discouraged because this may be a causal factor in the development and maintenance of obesity. The issues that receive most attention in the literature are whether alcohol is as efficiently used as an energy source as other macronutrients, i.e. carbohydrates and fats, and whether alcohol is added to, or substituted for, non-alcoholic energy in the diet. Several studies suggest that alcohol consumption may increase the risk of a positive energy balance and overweight (MacDonald et al. 1993; Tremblay et al. 1995). However, other epidemiological surveys show a negative association between alcohol consumption and adiposity and body weight (Colditz et al. 1991; MacDonald et al. 1993). Numerous studies have shown that the efficiency of use of alcohol for the maintenance of metabolizable energy is the same as for carbohydrates (Westrate et al. 1990; MacDonald et al. 1993; Sonko et al. 1994), whereas others have shown that ingested ethanol is less efficiently used as an energy source (Suter et al. 1992; MacDonald et al. 1993; Klesges et al. 1994). Other studies have shown that alcohol may have a fat-sparing effect (Suter et al. 1992), indicating that when consumed in excess this may promote fat gain, especially upper body fat. Although results remain controversial, alcohol does not seem to be a major determinant of body weight when consumed in moderate amounts, suggesting that replacing a moderate amount of energy from carbohydrates and fats with alcohol in non-alcoholics is not expected to decrease energy retention significantly or to be useful as an adjuvant in weight-reducing regimens. 3.1.5. Macronutrient replacement. Macronutrient substitutes are ingredients which are added to foods to replace macronutrients such as various digestible carbohydrates and fats in volume and in their technological functions (Finley & Leveille, 1996). The replacement may help the consumer to achieve dietary objectives, e.g.: reduced energy intake for reduction of body weight or maintenance of a normal body weight; reduced and modified fat intake (reduced cholesterol and saturated fat intake); S55 reduction of the risk of developing dental caries by replacing sugars with polyols. Carbohydrate replacement. Carbohydrates such as sucrose and glucose are bulking agents in foods such as chocolate, candies, cookies and cakes. Their replacement can be achieved by using a carbohydrate of similar or lower sweetness and similar taste but different physiological properties (e.g. reduced absorption and digestibility in the small intestine). In this respect polyols (sugar alcohols) have to be mentioned first (Bär et al. 1994). Complex polymeric carbohydrates e.g. polydextrose or inulin in combination with intensive sweeteners are the second group of bulking agents which can replace sugars in special cases (Finley & Leveille, 1996). Metabolic studies of polyols have been reviewed elsewhere (Schiweck & Ziesenitz, 1996) and can be summarized as follows: monosaccharide-derived polyols are more slowly absorbed from the small intestine than glucose. The absorption is thought to take place by means of passive diffusion along a concentration gradient. The rate of absorption differs among the polyols. Erythritol is well but not totally absorbed, whereas mannitol, sorbitol and xylitol are absorbed only slowly and incompletely. The absorbed part of the polyol is either excreted unchanged, mainly in the urine as in the case of erythritol and mannitol, or is converted to fructose by specific dehydrogenases, as found for sorbitol and, partly, mannitol. Xylitol is oxidized to xylulose by a polyol dehydrogenase and then enters the normal pentose pathway after phosphorylation by d-xylulokinase. The disaccharide alcohols isomalt, lactitol and maltitol are hydrolysed at varying rates from slow and partial hydrolysis to almost none by the various glycosidases located in the small intestine. Hydrogenated starch hydrolysates also contain di- and oligomeric polyols which are hydrolysed to glucose, sorbitol and maltitol. Most polyols have practically no impact on blood glucose concentrations and only a moderate influence, if any, on postprandial serum insulin profile. In the case of maltitol and hydrogenated starch hydrolysates substantially higher glycaemic responses were found (Felber et al. 1987). Significant amounts of ingested polyols reach the large intestine and the colon, where they are readily fermented by micro-organisms to H2 , CO2 , CH4 and short-chain fatty acids (SCFA). The latter lower the pH of the gut content and may, thereby, influence the composition of the colonic microbial flora. The most recent assessment of the energy value of polyols was that of the Expert Scientific Panel of the Life Sciences Research Office of the Federation of American Societies for Experimental Biology (1994). According to that evaluation the following energy values were given: mannitol 6·7 kJ/g, lactitol 6·7–9·2 kJ/g, isomalt approximately 8·4 kJ/g, sorbitol 7·5–13·8 kJ/g, xylitol approximately 10 kJ/g, maltitol and hydrogenated starch hydrolysates 11·7–13·4 kJ/g. Fructose as a replacer for glucose and sucrose should be mentioned in this context as a sweetener for diabetics, due to its low GI and low insulin response. Nutritional aspects are reviewed elsewhere (Bowman & Forbes, 1993). Whilst attention has been paid mainly to the use of polyols as sugar S56 W. H. M. Saris et al. Table 7. Polysaccharides as fat replacers (mimic fat) Starches Further types unmodified modified maltodextrins cellulose inulin b-glucan pectin resistant starch gums (xanthan, guar etc.) replacers, a number of carbohydrates (including polyols) may also act to partly or totally mimic fat in food (as stabilizers and thickeners), i.e. as fat replacers. Carbohydrates relevant in this aspect are summarized in Table 7. The degree of digestion and absorption of these carbohydrates in the small intestine varies according to the type of carbohydrate. All poorly digestible carbohydrates have in common the fact that the part which is not absorbed in the small intestine reaches the large intestine and is there fermented, to a greater or lesser extent, by the colonic microflora to SCFA and gases (H2 , CO2 , CH4 ). Due to the increase in substrate reaching the large intestine and the production of SCFA there is also an increase in the osmotic load in the large intestine leading to an increase of water within the gut content. These circumstances might be noticeable for the host by e.g. an increase in flatulence (gas) or soft-to-watery stools. Certain advantages with respect to bacterial composition (e.g. bifidogenic effects) or the properties of some SCFA (e.g. butyric acid and its possible preventive effects with respect to colon cancer) should be taken into account as well (Scheppach, 1994), but to clarify this, or other positive effects on health, further research work is required. Fat replacement. Generally, fat replacers are ingredients that are designed to replace all or part of the fat which is normally in a food without influencing the taste and texture quality of the food products (Jones, 1996). There are three categories of such fat replacers as indicated in Table 8. As fat mimics are carbohydrate- or protein-based ingredients, at least 20 kJ/g less energy is provided compared with fat. Some of the carbohydrate replacers mentioned provide even less than 16 kJ/g, making them an important tool in recipes for energy-reduced foods. Their metabolism has already been described. Fat substitutes are physically similar to fats and oils and therefore heat-stable. In the case of sucrose polyesters there are up to eight fatty acids (8–22 C atoms, saturated or Table 8. Categories of fat replacers Fat mimics Fat substitutes Low-energy fats microparticulated proteins 6 carbohydrates e.g. cellulose modified starches gums inulin polyols polydextrose sucrose polyesters triacylglycerols unsaturated) per sucrose molecule (Food and Drug Administration, 1996). The resulting molecules are too big to be split by lipases and are therefore neither digested nor appreciably absorbed. The small amount of material that is absorbed is metabolized to sucrose and fatty acids that are further metabolized normally in the body. These sucrose polyesters pass intact through the colon and are not used as a substrate for the bacterial microflora. The hydrophobic properties of these sucrose polyesters have influence on other hydrophobic substances in the gastrointestinal tract. For example fat-soluble vitamins are not well absorbed and cholesterol is removed in the faeces in the same way. Due to the fact that these sucrose polyesters pass through the small intestine undigested, gastrointestinal effects such as flatulence and soft stools occur in some people. Low-energy fats are true triacylglycerols with a new lipid structure. Triacylglycerols may be composed of mixtures of long-chain saturated fatty acids and SCFA esterified on a glycerol backbone. It is stated that their available energy is approximately 20 kJ/g (Finley et al. 1994). This reduced energy content is based on the poor absorption of stearic acid and the lower energy value of SCFA compared with long-chain fatty acids which normally occur in fats. SCFA are rapidly absorbed and converted to CO2 . Stearic acid in the 1- and 3-positions of the triacylglycerol would be hydrolysed by lipases. Free stearic acid would be poorly absorbed and stearic acid in the 2-position would be likely to remain on the glycerol and absorbed as monoacylglycerol, and further be converted to oleic acid (50 % of the absorbed stearic acid). Another example of a low-energy fat is a triacylglycerol prepared by the esterification of glycerol with capric, caprylic, and behenic acids resulting in caproacrylobehenin (Life Sciences Research Office/Federation of American Societies for Experimental Biology, 1991). Due to the limited intestinal absorption of behenic acid and the lower energy yield from capric and caprylic acids, an energy value of 20 kJ/g has been stated. This triacylglycerol is digested, absorbed, and metabolized by the usual pathways of triacylglycerol metabolism. Its medium-chain fatty acid component is readily absorbed. The long-chain component, behenic acid, is absorbed more slowly and less completely. An alternative to fat replacement is fat-binding. Dietary fat might be sequestered by binding to an appropriate nonabsorbed material. Derivatives of chitin such as chitosan have been shown to reduce plasma cholesterol concentrations in animals (Sugano et al. 1988), probably by binding of bile acids and interruption of the entero-hepatic circulation, but not to affect body weight. These derivatives of chitin and similar products are being widely marketed for human weight reduction, but we are not aware of scientific evidence of their efficacy. Their effectiveness might be estimated by comparison with the pharmacological agent tetrahydrolipstatin which is an inhibitor of pancreatic lipase (EC 3.1.1.3). This agent can lead to useful weight loss (typically a few kg greater than placebo) in controlled studies, although its usefulness is limited by unwanted side-effects of fat malabsorption and by reduction in absorption of fat-soluble vitamins (Drent & Van der Veen, 1993). The idea of preventing absorption of dietary fat or fat replacement is appealing because it could, in principle, Functional foods and substrate metabolism mitigate many adverse features of metabolic diseases. First, energy intake would be reduced, helping with body-weight control. The limited number of controlled experiments that have been conducted with low-fat foods indicate that they may possibly help to reduce fat intake (Jones, 1996), although it also has been suggested that the ‘fat-free foods’ may actually cause consumers to increase consumption (Rolls & Shide, 1992). Thus, more information on population-based consumption is needed to assess the impact of such foods in reaching the goal of fat reduction. Second, there would be a direct reduction of postprandial lipaemia, and given the evidence reviewed earlier (section 2.2) linking postprandial lipaemia with the ALP, this is likely to be beneficial in terms of cardiovascular risk. 3.1.6. Dietary components stimulating thermogenesis. The most important mechanism controlling thermogenesis is the activity of the sympathetic nervous system. The pharmacological approach to enhance metabolic rate has centred on the development of novel ß3 adreno-receptor agonists (Stock, 1989). However, ‘natural’ ingredients of food can also interact with the adrenergic system for thermogenic stimulation. Minor food constituents such as caffeine and associated methylxanthines in coffee and tea have a profound effect on metabolic rate. Also other minor constituents, such as spices, have thermogenic properties. Inclusion of these types of ‘natural’ food ingredients into food products could be a viable approach in stimulating energy expenditure to keep energy balance and thus body weight within acceptable limits. Methylxanthines: caffeine, theophylline and theobromine. Caffeine and other methylxanthines are alkaloids derived from at least sixty-three species of plants, including the familiar coffee bean, the tea leaf and the cocoa bean. Most human societies use caffeine regularly, most often in beverages, for its stimulant effect and flavour. Caffeine contents of beverages vary, depending on the plants they were made from and the food technological methods applied. The caffeine in cola soft drinks is added, using the purified compound that is obtained from the decaffeination of coffee beans. The Food and Drug Administration lists caffeine as a multipurpose generally recognized as safe (GRAS) substance that may be added to foods and beverages. It has been known since 1915 that ingestion of caffeine provokes an increase in the metabolic rate and subsequent investigations have confirmed this original observation (Acheson et al. 1980). Current theories attempting to explain the diverse pharmacological actions of dietary methylxanthines, favour their actions as antagonists of adenosineinhibitory effects on noradrenaline-induced cyclic AMP formation. The net result is an elevated cellular level of cyclic AMP, a critical intracellular mediator for the actions of catecholamines on thermogenesis. In a study by Dulloo et al. (1989), the effect of normal caffeine consumption on thermogenesis was studied. Single-dose oral administration of 100 mg caffeine (equal to a small cup of coffee) increased the metabolic rate by 3–4 % over 150 min. Measurements of energy expenditure in a room respiration chamber indicated that repeated caffeine administration (100 mg) at 2 h intervals over a 12 h daytime period, increased energy expenditure by about 10 %. Comparable results have been found by other groups. Acheson et al. (1980) S57 also observed an increase in fat oxidation, mediated by an increased lipolysis leading to higher blood NEFA levels. In a double-blind placebo-controlled study in moderate habitual coffee drinkers, Astrup et al. (1989) found increases in energy expenditure of 38.5, 30.1 and 136 kJ/h with 100, 200, 400 mg caffeine ingestion respectively. These effects were positively correlated to plasma caffeine response. In contrast, no significant correlations were found for plasma responses of theophylline and theobromine. It is suggested that most people develop caffeine tolerance due to a decrease in the inhibitory effect on adenosine. However, in the study of Dulloo et al. (1989) habitual caffeine intake of the subjects was 250–500 mg/d. In the study of Astrup et al. (1989) habitual intake of caffeine was 100–200 mg/d. Although a certain degree of tolerance to the thermogenic effect of caffeine may have been developed, these results suggest that a substantial effect remains during moderate daily caffeine consumption. Pungent ingredients of ginger and spices. Ginger is extensively used as a flavouring additive in foods, beverages and confectionery. Ginger is known for its apparent ability to subjectively warm the body. The pungent principles of ginger are present as two phenylalanine-derived homologous series: the gingerols and shogaols (Eldershaw et al. 1992). In a number of spices such as hot chillies, the compounds capsaicin and dihydrocapsaicin have been isolated and found to have thermogenic effects in isolated perfused rat hindlimb (Cameron-Smith et al. 1990). From the ginger components it turned out that gingerol especially induces thermogenesis. Interestingly, this effect was not inhibited by a- or b-adrenergic antagonists, suggesting that neither adrenergic receptors nor secondary catecholamine release was responsible for the observed effects. Gingerols, shogaols and capsaicinoids have some similarities in terms of both structure and function. All contain the 4-methoxy, 3-hydroxy phenylvanillyl moiety as well as a carbonyl-containing allyl side-chain. Each group of homologues is responsible for the pungent taste of the parent plant. The only human study on the thermogenic effects of spices was performed by Henry & Emery (1980) with chilli (component: capsaicin) and mustard (component: allyl isothiocyanate). Subjects were given test meals with or without 3 g mustard sauce and 3 g chilli sauce. Diet-induced thermogenesis over a 3 h period was 25 % higher after the spiced meal. This is a substantial increase in thermogenesis compared with other thermogenic substances. Further research is needed to investigate the role of spicy ingredients in human nutrition and the metabolic origins of their effect on diet-induced thermogenesis. The thermogenic properties are substantial. However, detailed human research on tolerance and identification of active compounds and metabolic interactions is lacking. Recently, also the catechin teoline, purified from tea leaves, showed thermogenic effects that are synergistic with caffeine (Dulloo et al. 1996). 3.1.7. Physiological and metabolic consequences of undernutrition. In undernutrition certain functions are affected and some nutrient reserves decrease, making S58 W. H. M. Saris et al. the undernourished individual more susceptible to injuries that a well-nourished individual can withstand with little repercussion. Cardiovascular and renal functions. In severe undernutrition cardiac work decreases, as does functional reserve, and central circulation takes precedence over peripheral circulation. Cardiovascular reflexes are altered, leading to postural hypotension and diminished venous return. Haemodynamic compensation occurs primarily from tachycardia rather than from increased stroke volume. Renal plasma flow and glomerular filtration rates may be reduced as a consequence of the decreased cardiac output, but water clearance and the ability to concentrate and acidify urine appear to be unimpaired. Gastrointestinal functions. Impaired intestinal absorption of lipids and disaccharides and a decreased rate of glucose absorption occur only in severe protein deficiency. The greater the protein deficit, the greater the functional impairment. Although the average protein requirement may not differ with advancing age, at least certain categories of elderly people have difficulties maintaining N balance when consuming the recommended daily amount (0.8 g/kg per d) (Young, 1992). To assess more accurately the needs of the elderly, they are usually evaluated in two age groups (65–75 years and 76 years and older). Distinctions are also made between healthy elderly people and those with chronic disease (Durnin, 1992; Morley, 1995). Due to the diminished efficiency of protein utilization in the elderly, the prudent dietary recommendations should ensure a minimum intake of 0.9 g/kg. A decrease in gastric, pancreatic, and bile production is also observed, with normal to low enzyme and conjugated bile acid concentrations. These alterations further impair the absorptive functions. Nevertheless, the ingestion of nutrients in high, therapeutic amounts usually allows for their uptake in sufficient quantity to permit nutritional recovery. Undernourished elderly people are prone to have diarrhoea probably due to irregular intestinal motility and gastrointestinal bacterial overgrowth. The immune system. The major defects are seen only in severe undernutrition. This seems to involve T-lymphocytes and the complement system. A marked depletion of lymphocytes from the thymus and atrophy of the gland occur. In addition, cells from the T-lymphocyte regions of the spleen and lymph nodes are depleted, probably owing to decreases in thymic factors. The production of several complement components, the functional activity of the complement system assessed by both the classic and alternative pathways, and the opsonic activity of serum are depressed. These deficiencies may explain the high susceptibility of severely undernourished patients to Gramnegative bacterial sepsis. Phagocytosis, chemotaxis, and intracellular killing are also impaired, partly due to the defects in opsonic and complement functional activities. The B-lymphocyte areas of spleen and lymph nodes and the circulating levels of B-cells and immunoglobulins are relatively normal, but there may be defects in antibody production, such as secretory immunoglobulin A. The overall consequences of all these alterations in severe undernutrition are a greater predisposition to infections and complications of otherwise less important infectious diseases. The defects in immune functions disappear with nutritional rehabilitation (Chandra, 1992). 3.1.8. Conclusions and further research. There seems to be a consensus view that a high-fat diet, resulting in a positive energy and fat balance, is an important risk factor in the aetiology of obesity. However, more information is necessary to elucidate whether (moderate) manipulation of the macronutrient content of the diet may affect body weight. In theory, epidemiological studies would be the best way to study these relationships. However, in this type of study there are too many variables that cannot be controlled. For this reason, long-term controlled intervention studies where the carbohydrate : fat ratio and types of carbohydrate and fat are manipulated would be more suitable. These studies have to cover at least 6 months, since changes in body weight are likely to be small. In this type of study acute experimental mechanistic studies can be included at several time points since more mechanistic research is necessary to elucidate the regulation of carbohydrate and fat balance within the body. In the latter studies, attention has to be focused on the issue of whether the different storage capacities for carbohydrate and fat within the body give rise to a specific role of carbohydrate (stores) in the regulation of appetite (Flatt’s theory) or whether the obesity-promoting effect of a high-fat diet is simply a passive overconsumption effect due to the high energy density of the diet. In mechanistic studies regarding the regulation of carbohydrate and fat balance, the type of fat or carbohydrate also has to be taken into account, since mechanisms behind the relationship of type of fat or carbohydrate and energy metabolism or substrate utilization are largely unknown. More specifically, mechanisms behind the increased sucrose-induced thermogenesis and decrement in fat oxidation and mechanisms behind the impact of the P : S ratio of the diet on fat oxidation require further study, since these metabolic effects may have important consequences for energy and macronutrient balance. Mechanistic studies have to include techniques for studying intermediary and tissue metabolism in man (stableisotope techniques, tissue balance studies, tissue biopsies, microdialysis), since only these types of studies will add information to the existing knowledge of body-weight regulation in man. With respect to alcohol, more mechanistic research is necessary to elucidate whether alcohol is handled in the body according to the law of thermodynamics (Macdonald et al. 1993). Additionally, long-term controlled experimental studies are necessary to study the relationship between alcohol consumption and body weight per se without confounding variables. However, it does not seem realistic to regard addition or substitution of alcohol in the diet in the concept of future functional foods. With respect to their digestion, most of the macronutrient replacers discussed herein do have in common that they are not, or are only partially, hydrolysed and absorbed in the small intestine. They are partially or completely fermented in the large intestine by the colonic flora. As a result, these macronutrients provide less energy to the body than completely absorbed and metabolized substrates. This makes them an important tool in the development of Functional foods and substrate metabolism energy-reduced foods and an important tool to achieve a balanced energy and fat intake or a reduction in energy and fat intake. More research has to be performed to investigate the long-term effect of these macronutrient replacers, in particular the fat replacers, on energy and fat balance and on body-weight control. Attention has to be paid to the suggestion that consumption of fat- or energy-reduced foods may be compensated by an increased total food intake, resulting in similar energy intake. Due to the very limited absorption of the carbohydrateand fat-replacers in the small intestine, the substances are fermented mostly in the large intestine thereby increasing the amount of SCFA produced. The energy content provided by the different fermented carbohydrates as well as the impact of SCFA like butyric and propionic acids on health may be a field of further research. A number of plant ingredients can be identified which elevate the diet-induced thermogenesis after ingestion. This elevation is mainly mediated by a prolonged activation of the sympathetic nervous system leading to an increased catecholamine release from the sympathetic nerve endings or an inhibitory effect on the action of adenosine. Both lead to an increase in cyclic AMP essential for the cellular increase in metabolism. However, other unknown mechanisms must also be involved since adrenergic blockade does not counteract the thermogenic properties of some ingredients. From the group of methylxanthines caffeine seems to be the most potent in thermogenic response after ingestion of a normal dose of 100 mg (þ 3–4 %). Another interesting group is the pungent ingredients from spices, such as ginger, chilli and mustard. With the increased interest in the non-nutrient compounds such as polyphenols in plants, it is clear that more ingredients that can increase thermogenesis will be discovered. Some of the known ingredients have already shown efficacy in the treatment of overweight subjects. Undernutrition affects cardiovascular, renal and gastrointestinal functions. Impaired intestinal absorption of nutrients occurs only in severe conditions, yet in certain categories of elderly people, maintenance of N balance is difficult with the recommended daily amount. The immune system is apparently depressed in (severe) undernutrition. The defects in immune functions disappear with adequate nutritional rehabilitation. In the field of undernutrition, the following points may need further research: (1) development of analytical methods (bioassay) and identification of biomarkers may help in making estimates of nutritional status and of nutrient requirements; (2) suitable animal and in vitro cell (cell culture) models would facilitate the investigation of metabolic handling of essential substrates (absorption, transport, receptor sites); (3) nutrient interactions are an essential focus of future studies because of alterations caused by the ageing process; (4) nutrient mechanisms that have an impact on genetic expression or immunological function should be examined with modern methods of molecular biology and immunology. Special interest should be directed to appraise post-translational modifications related to various proteins, diets or indispensible substrates. S59 3.2. Insulin resistance/sensitivity 3.2.1. Introduction. As reviewed earlier, the strongest factor predisposing to insulin resistance is obesity, particularly of the upper body. Therefore nutritional influences on body weight will also have a profound effect on insulin sensitivity. Since abdominal obesity is an even stronger predisposing factor, specific nutritional influences on body fat distribution are important. However, they are not clearly understood. Although the concept of the ‘beer belly’ is widespread, studies of the effect of alcohol on fat distribution have been conflicting (reviewed by Macdonald et al. 1993). More research is needed on this point. There is some evidence from animal studies that saturated fats may lead to intra-abdominal fat accumulation and that n-3 polyunsaturated fatty acids (PUFA) may protect against this (Hill et al. 1993), but no data on this point in human subjects are available. There is indirect evidence that a high dietary fat intake is associated with visceral obesity in women (Nicklas et al. 1995). On the whole, the most consistent explanation for the accumulation of intra-abdominal fat is an interaction between lifestyle, including stress factors, and excessive energy intake leading to obesity (Björntorp, 1991a). Thus, again, relevant dietary factors are those predisposing to obesity. 3.2.2. Dietary carbohydrates. In rats, feeding a fructoseenriched diet induces insulin resistance (reviewed in Frayn & Kingman, 1995). There is no clear evidence for this effect in human subjects at realistic levels of fructose intake. High-carbohydrate, low-fat diets are consistently found, at least in the short term, to raise plasma triacylglycerol concentrations (see p. S66) and also plasma insulin concentrations (Hollenbeck & Coulston, 1991). This is not surprising in view of the potentiation of insulin secretion by carbohydrates, but it has been interpreted as an adverse change indicative of insulin resistance (Hollenbeck & Coulston, 1991). This may be an over-interpretation: there is no prospective evidence that such diets have adverse consequences on insulin sensitivity or CHD. Their welldocumented beneficial effect on body-weight regulation (reviewed earlier) is likely to outweigh any possible direct adverse effect on insulin sensitivity. A reduction in the GI of the diet may improve insulin sensitivity. This has been shown indirectly by an improvement in glycaemic control observed in many studies of lowGI diets in NIDDM (Brand Miller, 1994), and directly in patients with CHD (Frost et al. 1996). 3.2.3. Dietary fat. Clearly excessive intake of fat will lead to obesity, and possibly specifically visceral obesity, and thus to development of the IRS. Of more interest are specific effects of the quality of dietary fat. There is considerable evidence in experimental animals that saturated fat in the diet may lead to insulin resistance (Vessby, 1995). The effects of saturated fat may be reversed by the addition of n-3 PUFA (Vessby, 1995; Storlien et al. 1996). The mechanism is likely to relate to a change in membrane fluidity affecting processes around the insulin receptor and recruitment of glucose transporters to the membrane. In man there is indirect evidence for the same effect. The phospholipid-fatty acids of skeletal muscle biopsies show a relationship to insulin sensitivity measured in vivo: a more S60 W. H. M. Saris et al. saturated fatty acid pattern is associated with insulin resistance and a high prevalence of PUFA is associated with increased insulin sensitivity (Storlien et al. 1996). Similar findings have been made with respect to plasma cholesterylester fatty acids (Vessby et al. 1994). Since these fatty acid patterns reflect long-term dietary intake, a link between saturated fat in the diet and insulin resistance is reasonably firmly established (Fig. 1). Prospective studies of dietary fat change and insulin sensitivity in human subjects have not been conclusive, perhaps because periods of study have not been long enough (Storlien et al. 1996). Thus, it would be premature at this stage to make any functional claim for an effect of dietary PUFA on sensitivity to insulin. 3.2.4. Niacin and insulin sensitivity. A common metabolic mechanism in insulin resistance may be an elevated concentration of NEFA in the plasma (Frayn et al. 1996). Nicotinic acid (one form of niacin: the other is nicotinamide) exerts a powerful suppressive effect on adipocyte lipolysis and thus NEFA release, and has been used as an effective hypolipidaemic agent (Farmer & Gotto, 1995). However, it has a number of side-effects (Farmer & Gotto, 1995). Over-the-counter niacin preparations are popular in the USA. Their long-term safety and efficacy have not been properly evaluated. It might be predicted that a short-term reduction in NEFA concentrations could improve insulin Fig. 1. Relationships between insulin sensitivity and the fatty acid composition of skeletal muscle phospholipids in normal men. The insulin sensitivity index was derived from a glucose clamp study; units are mg/m2 per min. Redrawn with permission from Borkman et al. (1993). sensitivity, but in the longer term consistent entrapment of fatty acids in adipocytes might lead to obesity with adverse effects on insulin sensitivity. This may not happen in practice because a well-known aspect of nicotinic acid action is a marked ‘rebound’ of plasma NEFA levels between doses. 3.2.5. Minerals. Some minerals have been associated with insulin sensitivity. Chromium. Cr may form a complex with nicotinic acid in plasma which has been called the glucose tolerance factor and may be associated with improved glucose tolerance (Mertz, 1993). It has been suggested that Cr deficiency might underlie the IRS (Mertz, 1993). Vanadium. Inorganic and organic compounds containing V, such as vanadyl sulfate and vanadate, have an insulin-like effect both in vitro and in vivo (Shechter, 1990). V is available as an over-the-counter preparation in some Southern American countries. As in the case of Cr supplementation, there are no prospective human data on safety or efficacy. Magnesium. There is a large body of evidence showing low plasma and intracellular Mg concentrations in diabetes. It is not clear whether these low levels represent Mg deficiency although it has been suggested that supplemental dietary Mg may be beneficial both in improving glycaemic control and in reducing the complications of diabetes (White & Campbell, 1993). In non-diabetic subjects low dietary Mg intake has been linked with a number of aspects of cardiovascular disease including insulin resistance (Ma et al. 1995). 3.2.6. Conclusions and further research. Insulin sensitivity is closely related to body fat content and body fat distribution. Thus, factors leading to obesity will increase insulin resistance, and those leading to upper-body or visceral obesity will have a greater influence. There are direct and indirect pieces of evidence that diets based on low-GI foods may improve sensitivity to insulin. The evidence linking saturated fat in the diet and insulin resistance appears to be reasonably firm, based on crosssectional studies, although it has not been proven in longterm intervention studies. Critical areas for future research are as follows. (1) Long-term prospective studies of the effects of high-carbohydrate, low-fat diets on insulin sensitivity and on plasma triacylglycerol concentrations (see p. S20) are needed in view of suggestions that such diets might be harmful in these respects. These studies need to be conducted in subjects with a range of sensitivities to insulin (e.g. relatives of those with diabetes) and from different ethnic backgrounds, to illuminate potential gene– nutrient interactions. The potential of low-GI foods to improve insulin sensitivity needs further investigation. (2) Long-term prospective studies of the effects of manipulation of the quality of dietary fatty acids on insulin sensitivity are also required to investigate whether insulin resistance can be ameliorated by dietary means. Properly controlled studies of micronutrient and mineral supplementation (niacin, Cr, V and Mg) in insulin resistance are required. 3.3. Blood glucose control 3.3.1. Introduction. One of the characteristics of untreated diabetes is hyperglycaemia. This is not only the Functional foods and substrate metabolism cause of a large proportion of the symptoms that heavily affect the quality of life of these patients but represents also the primary cause of the long-term specific complications of diabetes (retinopathy, nephropathy, neuropathy) and is an important contributor to the excess risk of cardiovascular disease (The Diabetes Control and Complications Trial Research Group, 1993). High blood glucose concentrations (but below the diagnostic level of diabetes) may also represent a cardiovascular risk factor in the general population. This has been particularly substantiated in individuals with impaired glucose tolerance (characterized by mildly elevated postprandial and normal fasting blood glucose concentrations), but there are some indications that even within the normal population blood glucose values in the upper part of the normal range might represent a cardiovascular risk factor (Gerstein & Yusuf, 1996). A recent contribution has shown that near optimal blood glucose control is able to prevent most cases of microvascular complications occurring in diabetes (Stamler et al. 1993), indicating that preventive strategies to control blood glucose level may reduce the occurrence of chronic complications in diabetes. In order to evaluate the effects of diet on blood glucose control, it is not sufficient to measure blood glucose concentrations in the fasting state since large fluctuations can occur throughout the day, particularly in patients with IDDM. Therefore, reliable information on blood glucose control needs to be based not only on fasting but also on postprandial evaluation of blood glucose concentrations; pre- and postprandial measurements can also be repeated at each meal. An additional useful marker of glucose control is the measurement of glycated haemoglobin, as it represents an integrated measure of blood glucose control during the preceding 2–3 months or, alternatively, fructosamine which reflects blood glucose levels during the previous 2–3 weeks. Relevant indices of blood glucose control are listed in Table 9, and any evaluation of functional foods in facilitating normoglycaemia needs to focus on these variables (American Diabetes Association, 1997b). Measurements of plasma insulin concentrations and insulin sensitivity are also relevant to evaluate the effects of foods on glucose metabolism but they have been considered earlier. The most powerful measure to improve blood glucose control in overweight diabetic patients is weight reduction. This is particularly effective in NIDDM patients who are very often (70–80 %) overweight, a condition which has a major impact on insulin resistance and, consequently, on Table 9. Relevant measures of blood glucose control Fasting blood glucose Postprandial blood glucose Daily blood glucose profile* Glycated haemoglobin† Fructosamine† Oral glucose tolerance test‡ * Several blood glucose measurements performed throughout the day either by conventional laboratory methods or by strips read on a reflectometer. † Only in diabetic patients. ‡ Only in non-diabetic individuals or in patients with a milder form of diabetes. S61 plasma glucose levels. However, in IDDM patients, also, the presence of overweight (occurring in about 30 % of cases) impairs the hypoglycaemic effect of exogenously administered insulin, thus hampering the achievement of optimal blood glucose control (Diabetes and Nutrition Study Group EASD, 1995; American Diabetes Association, 1996). However, dietary composition can also affect blood glucose control, particularly in the postprandial period, as discussed in the next paragraphs. 3.3.2. Nutritional influence on fasting and postprandial blood glucose levels. Although closely related, fasting and postprandial blood glucose levels are regulated by mechanisms that are, to some extent, different; in fact, while postprandial blood glucose concentrations are largely dependent on meal composition, fasting values are only minimally influenced by the amount and/or rate of glucose absorption during the previous meal, and reflect the rate of glucose production in the liver (the two key processes being glycogenolysis and gluconeogenesis). Among the various dietary constituents, the one with the strongest influence on blood glucose levels in the postprandial period is the amount of digestible carbohydrate in the diet. Digestible carbohydrates include monosaccharides (glucose, fructose), disaccharides (sucrose, lactose) and polysaccharides (starch), which are digested and absorbed in the human intestine, thus contributing to the glucose inflow to the bloodstream (Table 10) (Asp, 1996). In diabetic patients, postprandial blood glucose levels are directly related to the amount of digestible carbohydrate in the diet and although they may be regulated by appropriate pharmacological treatment this is not always feasible or fully successful (Perrotti et al. 1984). Therefore, diets with a very high content of digestible carbohydrate are not without problems in the treatment of diabetic patients, particularly those with IDDM who have a severely impaired endogenous insulin secretion and are therefore more susceptible to exogenous influences on blood glucose metabolism. On the other hand, a drastic reduction in the intake of digestible carbohydrates is not feasible since in a weightmaintaining diet, this reduction should be compensated by an increase in protein or fat intake. Very high intakes of both protein and fat are not recommended because of their possible untoward effects on the development of chronic diabetic complications (Diabetes and Nutrition Study Group EASD, 1995). Therefore, it is important to identify food characteristics able to reduce the impact of digestible carbohydrates on postprandial blood glucose levels. The glycaemic index. The GI is defined as the incremental blood glucose area after the test product has been ingested, expressed as a percentage of the corresponding area after a carbohydrate-equivalent amount of white bread (Jenkins et al. 1981). The ratio between mono-, di- and polysaccharides is no longer regarded as important in relation to the effects on postprandial blood glucose since amylase and disaccharidase activities in the human duodenum are sufficient to hydrolyse starch and disaccharides within minutes. The meal content of protein and fat, although able to influence postprandial glucose values, has limited clinical significance because the magnitude of these effects is rather small. More important are all dietary factors able to delay the S62 W. H. M. Saris et al. Table 10. Main food carbohydrates and their digestibility in the small intestine (From Asp, 1996) Monomers Digestibility Monosaccharides Glucose Fructose Galactose þ þ* þ Disaccharides Sucrose Lactose glucose, fructose glucose, galactose þ† þ† galactose, glucose, fructose fructose, glucose glucose ¹ ¹ þ Oligosaccharides a-Galactosides, e.g. raffinose, etc. Fructooligosaccharides Maltooligosaccharides Polysaccharides Starch amylose amylopectin modified starches NSP cellulose hemicelluloses pectins ß-glucans gums mucilages algal polysaccharides uronic acids fructans glucose glucose galactose glucose mannose arabinose xylose rhamnose þ (¹)‡ ¹ fructose ‘New’ carbohydrate food ingredients Inulin Polydextrose Polyols Pyrodextrins fructose glucose various sugar alcohols ¹ ¹ (+) glucose (¹) * Limited in some individuals when ingested without glucose. † Except in disaccharidase deficiency. ‡ Resistant starch is indigestible. process of digestion and/or absorption of carbohydrates in the intestine, thus reducing the glycaemic response to carbohydrate-rich foods (Table 11). Due to the variety of factors that influence the glycaemic impact of a meal, it is not possible to predict the glycaemic response of each food on the basis of its physical and chemical characteristics (Parillo & Riccardi, 1985). Therefore, it is necessary to test carbohydrate-rich foods in vivo to evaluate their glycaemic response, thus allowing the selection of low-GI foods. Although the procedure used for GI determination varies in different laboratories and needs to be standardized (Food and Agriculture Organization/World Health Organization, 1998), it allows the classification of carbohydrate-rich foods into broad categories (high, medium and low glycaemic responses) (Riccardi & Rivellese, 1987). In vitro methods for the prediction of the glycaemic response to starchy foods seem promising (Granfeldt et al. 1992); however, no methods are at hand to predict the glycaemic impact of foods containing lowmolecular-mass carbohydrates. It is now well documented that a diet preferentially containing low-GI foods improves the metabolic control in diabetic patients and has a number of other possible metabolic benefits. In fact, such a diet has been reported to lower the day-long blood glucose profile, reduce glycated haemoglobin or fructosamine, and improve glucose tolerance. In addition, in some studies, fasting blood glucose levels were decreased in diabetic subjects (Brand-Miller, 1994). Moreover, low-GI diets have also been shown to have beneficial effects on blood lipid metabolism and other cardiovascular risk factors (for review, see Björck, 1996; Food and Agriculture Organization/World Health Organization 1998). As to the mechanism of these beneficial effects, the slow rate of digestion and absorption per se, i.e. the GI features, may be important. In addition, low-GI foods may be more efficient in suppressing NEFA concentrations between meals, leading to an improved tissue uptake of glucose when the second meal is ingested 4 h later (Jenkins et al. 1982). Cumulative effects of low-GI foods, extending beyond the improved insulin economy in the acute prandial phase, may also stem from a more extensive colonic production of SCFA, since such foods are frequently richer sources of indigestible carbohydrates. Generation of SCFA through fermentation might, thus, explain improvements seen in fasting blood glucose and glucose tolerance at breakfast when preceded by a low-GI evening meal (Thorburn & Proietto, 1993). Effect on fasting glucose. The knowledge of nutritional factors influencing blood glucose metabolism in the fasting state is scarce. Liver glucose production, the major determinant of fasting glucose levels, is under the control of insulin and therefore not adequately suppressed when insulin resistance is present; thus nutritional factors influencing fasting plasma glucose concentrations are primarily those acting on insulin resistance. In this line it is well known that a diet with a very high fat and a very low carbohydrate content increases the production of ketone bodies, therefore impairing insulin sensitivity and elevating blood glucose levels, particularly in the fasting state. However, the relevance of this phenomenon in everyday life is questionable since such an extreme nutritional condition rarely occurs. In relation to different dietary fats there are some indications that saturated fat could impair insulin sensitivity, thus increasing blood glucose levels. More controversial are the effects of polyunsaturated fats; in fact it seems that while n-6 PUFA could decrease blood glucose concentrations, n-3 PUFA (if consumed in large amounts) could have a hyperglycaemic effect. However, the influence of dietary fat composition on blood glucose control has limited clinical significance because it is small and has still to be properly documented in randomized controlled intervention trials of sufficiently large sample size (Rivellese et al. 1996). Non-digestible carbohydrates (resistant starch, NSP, and oligosaccharides) escape digestion in the small intestine and are fermented by colonic bacteria in the large bowel generating SCFA (see p. S19). These metabolites could influence liver glucose production and, thus, the fasting glucose concentration. Functional foods and substrate metabolism S63 Table 11. Properties of carbohydrate foods that can be utilized to modify the postprandial blood glucose response (the glycaemic index) Chemical structure of the digestible carbohydrate Monomeric composition Amylose : amylopectin ratio Physical structure of carbohydrates Degree of gelatinization Starch–lipid and starch–protein interactions Retrogradation Viscous properties of dietary fibre Food form Botanical integrity (cells and/or tissue) Physical structure, e.g. pasta Other food components/supplements Viscous dietary fibre Organic acids Amylase inhibitors Finally, alcohol intake also has significant, although conflicting, clinical effects on plasma glucose levels. In fact, alcohol intake acutely suppresses hepatic glucose production thus lowering plasma glucose levels. Conversely, if habitually consumed in large amounts it impairs insulin sensitivity, thus impairing glucose tolerance and increasing plasma glucose levels. 3.3.3. Food properties determining the glycaemic index. As already mentioned, dietary carbohydrates represent the major dietary constituent influencing blood glucose control. However, the impact of dietary carbohydrates on glucose metabolism depends not only on the amount consumed, as believed in the past, but also on some specific food properties which can profoundly influence the metabolic effects (Parillo et al. 1985) (Table 12). These properties are a consequence not only of the ratio indigestible : digestible carbohydrates, but also of the food structure and of some specific physico-chemical characteristics of carbohydrates present in the food. This section will try to illustrate the importance of these properties in relation to the metabolic effects of different dietary carbohydrates. Table 12 lists some important properties of carbohydrate foods that can be utilized for the production of functional foods with a reduced postprandial blood glucose response, i.e. a low GI. It should be noted in this context that major sources of carbohydrates such as potatoes and bread are characterized by high GI values. Raw starch granules. Plant cells store starch in semicrystalline granules with variable size and structure. Raw starch granules are slowly digested by amylases. Cereal starches with an A pattern on X-ray diffraction analysis display a rather high digestibility in the small intestine, whereas B-type granules, e.g. from potatoes or highamylose maize starch, are mainly indigestible (Langkilde & Andersson, 1994). Formulas with raw maize starch have been utilized to provide extremely slow-release carbohydrates to children with glycogen storage disease, making it possible to avoid repeated tube-feeding during the night. The potential of raw starch in producing low-GI foods is limited to products produced below the gelatinization or melting temperature of starch granules. Gelatinization and retrogradation. Heating of starch in excess water results in swelling, leakage (especially of amylose), and eventually disintegration of the granule structure rendering the starch soluble or dispersible in water. This process is called gelatinization and occurs at different temperatures for different starches, usually between 60 and 808. Dispersed or soluble starch is highly susceptible to salivary and pancreatic amylase, which is present in excess in relation to the final hydrolysis and absorption at the brush-border level. When, for example, rolled cereals are produced under mild enough conditions the starch may be only partly gelatinized. A very low degree of gelatinization, however, is required to lower the GI of such products significantly (Granfeldt et al. 1995). Retrogradation is the process of recrystallization of starch from a solution. This occurs especially during slow cooling; amylose forms dense crystals resistant to amylase action, and staling of bread is related to retrogradation of amylopectin. Amylose retrogradation is a well-documented mechanism for resistant starch formation, but its effect on the GI is incompletely understood. Amylose : amylopectin ratio. Amylose, the virtually unbranched form of starch, forms double helixes with about six glucose residues per turn. The interior of the helixes can accommodate the hydrophobic end of polar lipids, forming inclusion complexes with reduced availability of the amylose for enzymic hydrolysis (Holm et al. 1986). This, and the propensity of retrogradation, are factors behind the usefulness of high-amylose starch varieties for production of foods with low GI and/or high resistant starch content. Cellular structure. Legumes such as beans have low GI and comparatively high resistant starch content. A main feature behind these properties is intact cell walls forming physical barriers to starch digestion even after boiling (Wursch et al. 1986). A rather high amylose content may contribute as well. Preservation of the intact cell structure is essential for keeping these properties in processed foods. Table 12. Properties of food carbohydrates modifying their effects on glucose metabolism Rate of absorption Type of absorbed monomers Extent of absorption Extent and rate of colonic fermentation Site and metabolites of colonic fermentation S64 W. H. M. Saris et al. Gross structure. The presence of intact grains in bread and other cereal products has been demonstrated to be a main determinant of the GI as well as a source of resistant starch (Jenkins et al. 1986; Björck et al. 1996). Pasta is another well-documented group of low-GI foods in which structural effects are important. In this case a protein network is responsible for the slow enzymic hydrolysis of the starch. At least in some studies, the maintenance of an intact botanical structure also reduces GI features of fruit products (Haber et al. 1977). Organic acids. Another factor influencing the glycaemic and insulinaemic impact of foods is the presence of organic acids produced on fermentation of foods e.g. sourdough baking. The acids appear to differ in gastrointestinal effect, and whereas some may reduce the rate of starch digestion, others reduce the gastric emptying rate (Liljeberg & Björck, 1996). Thus, fermentation processes represent one way of reducing the GI of carbohydrate foods. Amylase inhibitors. The presence of certain antinutrients such as phytate, polyphenols, and lectins has been discussed in relation to the low-GI features of legumes (Thompson et al. 1987). However, due to other effects such factors can hardly be used for optimizing starch properties in food products. Glucose : fructose : galactose ratio. Fructose occurs in free form in fruits, berries and honey, as part of sucrose and in high-fructose syrups. Its use as a sweetener is based on its high sweetness, at least under certain conditions. The low glycaemic response (GI about 20 % of glucose) has made fructose an alternative sweetener for use in diabetic diets (see also section 3.1.6.). The low-GI features of fructose probably explain the moderate GI range of certain fruits and the fact that sucrose has a moderate GI. The main concern about fructose as a sweetener has been the possibility that it induces elevated triacylglycerol concentrations, relevant in diabetic patients and individuals with the IRS (Truswell, 1994). The absorption capacity for fructose is limited when given alone (Truswell et al. 1988). Together with glucose, however, which is the normal situation in foods or diets, the absorption is improved and intolerance problems avoided. Galactose is part of lactose, but significant amounts of free galactose are found only in fermented milk products, due to the preferential utilization of glucose by the microorganisms. 3.3.4. Indigestible carbohydrates and glucose metabolism: possible mechanisms of action. The main types of indigestible carbohydrates are: dietary fibre (NSP), resistant starch and oligosaccharides. Although not the determinant of low-GI features per se, many low-GI foods are rich sources of these carbohydrates. The possible effects of indigestible carbohydrates on glucose metabolism may be related to different upper gastrointestinal events e.g. reduced motility and/or absorption of carbohydrates due to viscous properties of dietary fibre components or a reduced rate of carbohydrate digestion and/or absorption due to entrapment of the substrate within a fibre matrix at cell or tissue level. However, the SCFA produced on colonic fermentation of indigestible carbohydrates are increasingly being discussed in relation to systemic effects on glucose and lipid metabolism (Cummings & Macfarlane, 1991). Thus, carbohydrate fermentation has been reported to enhance suppression of hepatic glucose production and NEFA levels in man, leading to lowered fasting blood glucose and improved glucose tolerance (Thorburn & Proietto, 1993). Few studies are, however, available on this topic. NSP. Among the nutritional factors able to influence blood glucose levels, dietary fibre is certainly the one that has been most extensively studied. NSP constitute the main part of dietary fibre in most foods, the daily average amount in European diets being about 20 g/d (Cummings, 1993). Plant cell walls are the main source of naturally occurring NSP, such as cellulose, hemicellulose and pectic substances, although storage polysaccharides such as inulin and guar gum, as well as exudate gums, are also indigestible polysaccharides included in dietary fibre. A large body of evidence clearly shows that a diet consisting of a high consumption of fibre-rich foods of natural origin induces lower blood glucose levels particularly in the postprandial period in comparison with a diet containing the same amount of digestible carbohydrate but not rich in fibre (Riccardi & Rivellese, 1991). More controversial is the issue of whether dietary fibre represents the marker or the cause of a low glycaemic response (Nuttal, 1993). It is now clear that more important than the amount of fibre is the interaction between fibre and carbohydrates within foods. The presence of cells with intact walls composed of fibre polysaccharides is important as they are able to encompass carbohydrates, slowing their accessibility and thus their digestibility. As already mentioned, soluble viscous types of NSP are those affecting the postprandial glucose and insulin response after a meal. This has been demonstrated repeatedly for a large number of such polysaccharides when added to meals or incorporated into foods. The effect is related to viscosity, which inhibits mixing and diffusion in the intestinal tract and possibly delays gastric emptying, and which can be abolished by hydrolysis with loss of viscosity. A main obstacle in utilizing these properties to design functional foods lies in the organoleptic limitations to include enough viscous polysaccharides. The importance of viscosity in conditioning the GI of food naturally rich in soluble fibre, however, has been questioned. Even in products with high fibre content, e.g. flaked oats or bread with oat flour, structural properties seem more important in obtaining lowGI products. However, even if this is the case, the importance of fibre in conditioning the structural properties of foods (thus influencing the GI) cannot be neglected. In addition to their effects on the postprandial blood glucose response, NSP act as the fermentation substrate for colonic bacteria with production of SCFA which might influence liver glucose production. Resistant starch. Resistant starch, defined as starch and hydrolysis products thereof that are not absorbed in the small intestine, has emerged as a main substrate for the human intestinal microflora. Although the present intake in Europe seems low, about 4 g/d only (Dysseler & Hoffem, 1994), there is considerable potential for increasing it by providing food with elevated resistant starch content. Three main forms of resistant starch have been identified: (1) physically enclosed starch, (2) resistant B-type starch granules and (3) retrograded amylose (Englyst et al. 1992). Functional foods and substrate metabolism Chemically modified food starches and pyrodextrin are other forms that may contribute in processed foods. Methods for determination of resistant starch are designed to estimate the starch residue after treatment of the sample with enzymes simulating normal starch digestion in the small intestine. A critical step, not yet fully solved with any one of the suggested methods, is simulation of the normal disintegration of foods by chewing. This is essential to recover the physically enclosed fractions of resistant starch. The main interest in resistant starch stems from its properties as a fermentation substrate. Studies with human faecal flora in vitro have indicated a high yield of butyrate from resistant starch, which has been supported by some human in vivo data (Scheppach et al. 1988). The site, extent and SCFA pattern of fermentation of resistant starch from various sources need further study. There seems to be considerable potential for designing resistant-starch-containing foods for specific effects on colonic health (for review, see Asp & Björck, 1992; Asp et al. 1996). Oligosaccharides. The main types of indigestible oligosaccharides in foods and food ingredients are: (1) a-galactosides (raffinose, stacchyose, verbascose) found mainly in legumes; (2) fructans (fructo-oligosaccharides, inulin); (3) galacto-oligosaccharides (derived from lactose); and (4) pyrodextrins and cyclodextrins. The present interest in oligosaccharides stems from their properties as low-energy more-or-less sweet bulk substances (see section 3.1.6.) and their effects as fermentation substrates producing metabolites with local and/or systemic effects, as well as prebiotics promoting desirable intestinal micro-organisms (Mitsuoka et al. 1986). Fructans have been studied most extensively and shown to promote bifidobacteria with a concomitant inhibition of other species such as bacteroides and clostridia. In vitro studies have indicated a potential to inhibit pathogens. Other potentially beneficial effects include lower activities of hydrolytic and reductive enzymes thought to be involved in colonic carcinogenesis. A comparatively high yield of butyrate has been reported, which is also of interest in this context. Effects of fructans on lipid metabolism (decreased plasma triacylglycerol and cholesterol concentrations) have been demonstrated in rats. Human data are still scarce, however, and the potential of oligosaccharides to reduce plasma lipids in man needs further exploration. With respect to glucose metabolism, a daily supplementation with oligofructose at 8 g/d for 14 d significantly reduced fasting blood glucose in type II diabetics (Yamashita et al. 1984). Whereas fructo-oligosaccharides are being studied extensively, the potential of other indigestible oligosaccharides for prebiotic or systemic effects needs further exploration. Metabolic effects of short-chain fatty acids. An important mechanism by which food characteristics may influence glucose metabolism is represented by intestinal fermentation of carbohydrate in the large bowel which may play a role in modulating glucose metabolism through the production of SCFA. All carbohydrates (particularly fibre, oligosaccharides and resistant starch) escaping digestion S65 and absorption in the small intestine, pass to the large bowel for subsequent bacterial fermentation. There is some evidence, mainly from in vitro studies, that the proportion of the various SCFA (acetate, propionate, butyrate) formed during fermentation differs depending on the specific carbohydrate acting as substrate, with concomitant differences in physiological effects (Macfarlane & Cummings, 1991). Several studies show that dietary supplementation of propionic, acetic and lactic acids may diminish the postprandial glucose and insulin response (Brighenti et al. 1995; Liljeberg & Björck, 1996). Such effects are probably mediated by mechanisms in the upper gastrointestinal tract, such as inhibition of gastric emptying or inhibition of digestive enzymes. Furthermore, long-term propionate administration has been shown to lower the fasting glucose concentration (Venter et al. 1990), an effect possibly related to inhibition of glucose release from the liver. In spite of extensive fermentation, lactulose was demonstrated in one study (Jenkins et al. 1991) to increase plasma cholesterol levels in healthy subjects. The high yield of acetate obtained on fermentation of this substrate was suggested as an explanation. The gross energy values for SCFA range from 15 kJ/g for acetate to 25 kJ/g for butyrate. Of this, 75–85 % is metabolizable. The true energy value of fermented carbohydrate is dependent on the yield of SCFA as well as the fractions used for biomass and combustible gas production. Resistant starch, for example, has been estimated to provide about 8.4 kJ/g fermented substrate (Livesey, 1994). 3.3.5. Conclusions and further research. The recent progress in understanding mechanisms determining both rate and extent of carbohydrate absorption have provided tools for designing foods with specific nutritional effects. One such tool is represented by the GI concept, and accumulating data have demonstrated facilitated control of blood glucose concentrations in diabetes with diets characterized by low-GI foods. Some recent epidemiological data also suggest the hypothesis that low-GI foods could be protective against NIDDM. The importance of blood glucose levels for protein glycosylation implies a potential role in ageing. Further, low-GI foods may help to reduce plasma cholesterol concentrations and insulin resistance, and thus be of more general importance in defeating the metabolic syndrome. The emerging knowledge of properties important for the rate and site of fermentation of indigestible carbohydrates, as well as properties of various fermentation products, lays the ground for designing foods with optimal prebiotic effects, with the potential of influencing also metabolic variables through absorbed fermentation products. However, proper clinical testing (also long term) is required before the functional properties of foods active on glucose metabolism can be fully assessed. The European study group on diabetes (EASD) recommended in 1995 an increased use of low-GI foods in diabetes. Recently, a report (Food and Agriculture Organization/World Health Organization, 1998) recommended that the bulk of carbohydrate-containing foods consumed should be those rich in NSP and with low GI. The lack of variety of such S66 W. H. M. Saris et al. foods on the market, not least regarding cereal foods, makes the development of products with low GI a priority area for functional food development. 3.4. Plasma triacylglycerols 3.4.1. Introduction. An association between fasting plasma triacylglycerol concentrations and development of CHD has been recognized for decades. However, there has been some controversy over the nature of the link. In multivariate statistical analyses fasting plasma triacylglycerol concentrations have often been non-significant contributors compared with plasma HDL-cholesterol concentrations, with which they are strongly negatively related. More recently, however, meta-analysis has shown that elevation of the fasting plasma triacylglycerol concentration is a risk factor for development of CHD even when adjusted for plasma HDL-cholesterol (Hokanson & Austin, 1996). In fact, the combination of elevated plasma triacylglycerol and depressed HDL-cholesterol concentrations is a particularly strong risk marker for CHD. This is the typical dyslipidaemia associated with insulin resistance. In the fasting state the plasma triacylglycerol originates from the liver, which secretes VLDL particles containing triacylglycerol. The triacylglycerol in these particles is hydrolysed by the enzyme lipoprotein lipase (EC 3.1.1.34) in peripheral tissues which include adipose tissue, skeletal muscle and myocardium. As the particles shrink they may be taken up by receptors or they may remain in the circulation, becoming smaller and more cholesterol-enriched as they lose further triacylglycerol until they are classed as LDL particles. After a meal containing fat, dietary triacylglycerol is packaged within the enterocytes into chylomicron particles, which are released into the plasma via the lymphatics. The metabolism of chylomicron particles is similar to that of VLDL but they are a better substrate for lipoprotein lipase, and their triacylglycerol is rapidly removed before their remnants are taken up by receptors. Since the 1940s it has been realized that postprandial triacylglycerol concentrations may be related to CHD. It is now recognized that the magnitude and duration of elevated postprandial triacylglycerol concentrations (postprandial lipaemia) is a strong marker of CHD risk, far stronger, at least in some groups, than the fasting triacylglycerol concentration (Griffin & Zampelas, 1995). The inverse relationship between plasma triacylglycerol and HDL-cholesterol concentrations may be explained by events in the postprandial period (Frayn, 1993). Increased postprandial lipaemia, i.e. prolonged residence of triacylglycerol-rich particles in the circulation, leads to the exchange of their triacylglycerol for the cholesterol esters of HDL and LDL particles, leading to loss of HDL-cholesterol. Thus, low HDL-cholesterol concentrations may be a marker of impaired postprandial triacylglycerol metabolism. Loss of cholesterol esters from LDL particles makes them smaller and denser, thus accounting for the third component of the ALP (see section 2.2.). Dietary components may affect fasting triacylglycerol concentrations mainly through changes in the rate of hepatic VLDL-triacylglycerol secretion. The fasting triacylglycerol concentration is one determinant of the postprandial lipaemic response, perhaps because VLDL particles containing endogenous triacylglycerol compete with the chylomicrons for clearance by lipoprotein lipase. Therefore lowering of fasting triacylglycerol concentrations will usually also reduce postprandial lipaemia. Other dietary components may affect postprandial lipaemia more directly. 3.4.2. Dietary carbohydrates. When a low-fat, highcarbohydrate diet is introduced, a consistent change is an elevation of fasting (and in some studies postprandial) triacylglycerol concentrations (reviewed in Frayn & Kingman, 1995). This has led to the belief that such diets may be adverse in terms of their effects on blood lipids and thus CHD (Hollenbeck & Coulston, 1991). However, there is no evidence that this is so: rather, their beneficial effects on body weight are probably predominant. The effect on plasma triacylglycerol concentrations may be specifically to increase the proportion of large, buoyant VLDL particles and this might be a beneficial change since such particles are not major precursors of LDL. However, this is controversial and this area needs further work. Ingestion of fructose in large amounts (e.g. 50 g) also has a marked effect in potentiation of postprandial lipaemia. The mechanism is thought to be potentiation of hepatic lipogenesis and thus VLDL secretion, although impairment of plasma triacylglycerol clearance has also been suggested (reviewed in Frayn & Kingman, 1995). Again, further work is needed to elucidate the mechanism of this effect and its implications. 3.4.3. Dietary fat. Diets enriched in n-3 PUFA lead consistently to lower fasting plasma triacylglycerol concentrations (Nestel, 1990). Other than this the effects of dietary fat quality are more on plasma cholesterol than on fasting triacylglycerol concentrations (reviewed by Hornstra et al. 1998). Of more interest are dietary factors which may affect postprandial lipaemic responses, since this is an important determinant of the dyslipidaemia of the IRS. The greater the amount of fat eaten, the greater will be the lipaemic response. Again, different qualities of dietary fat affect this response, and in particular n-3 PUFA have a strong effect in diminishing postprandial lipaemia (Griffin & Zampelas, 1995). PUFA of the n-6 series lead to lesser postprandial lipaemia than saturated fatty acids, but the effects of monounsaturated fatty acids are not clear (Griffin & Zampelas, 1995): more research is needed in this area. 3.4.4. Conclusions and further research. Both fasting and postprandial plasma triacylglycerol concentrations are markers of risk of CHD. They are potentially modifiable by nutritional means. The short-term influence of a low-fat, high-carbohydrate diet is to increase plasma triacylglycerol concentrations. Whilst this had led to speculation that such diets may be disadvantageous in terms of CHD risk, it could also be argued that if such diets lead to weight loss they will be beneficial. The quality of dietary fat may be an important influence. Supplementation with long-chain n-3 PUFA reduces fasting and postprandial triacylglycerol concentrations, but the effects of other unsaturated fatty acids in this respect are not well documented. Critical areas for future research are as follows. It is critically important that long-term (at least 6 months) studies of the effects of low-fat, high-carbohydrate diets Functional foods and substrate metabolism on plasma lipid constituents are carried out. The confusion which presently reigns over the effects of such diets is preventing their widespread acceptance (or avoidance). Important questions to be answered by such studies are: (1) Is the elevation of plasma triacylglycerol concentration on a high-carbohydrate diet transient? (2) What is the nature of the elevation of plasma triacylglycerol concentrations (large, buoyant or small, denser VLDL particles) and what is its impact on other lipid constituents (e.g. the density distribution of LDL particles)? (3) Is it reasonable to apply conclusions from epidemiological studies to the effects of dietary manipulation within individuals (e.g. an elevated plasma triacylglycerol concentration may be a risk factor for CHD in epidemiological terms, but does elevation of plasma triacylglycerol by dietary means confer equivalent risk?). (This will require prospective studies lasting at least 5 years with hard end-points such as incidence of CHD.) (4) Do beneficial effects of such diets on body-weight regulation outweigh direct effects on plasma lipid constituents? Studies are needed of the effects of replacement of saturated fatty acids with monounsaturated fatty acids and n-6 PUFA, on plasma triacylglycerol concentrations, both fasting and postprandial. Any such effects must be related to changes in sensitivity to insulin and genetic factors (see section 3.2.). Research is needed into factors acutely affecting postprandial lipaemia, including dietary fructose or sucrose, and the effects of different types of dietary fat. 4. Nutrition, substrate metabolism and physical performance 4.1. Introduction The most important metabolic characteristic of physical exercise is the increased need for energy. Training or competition will increase the daily expenditure by 2– > 4 MJ/h depending on duration and intensity. Athletes must adapt their food consumption to meet the energy needs. This increased food intake should be well balanced, with respect to the macro- and micronutrients. However, this is not always simple. Many specific athletic events may be characterized by extremely high exercise intensities. Running a marathon, for example, costs about 10–12 MJ. Depending on the time needed to finish, this may induce an energy expenditure of approximately 3 MJ/h in a recreational athlete and 6 MJ/h in the elite athlete who finishes in approximately 2–2.5 h. A professional cycling race, such as the Tour de France, costs about 28 MJ/d, a value which will be increased to approximately 36 MJ/d when cycling over several mountain passes (Saris et al. 1989). Compensating for such high expenditures by ingesting normal solid meals will pose a problem to any athlete involved in such competitions, since digestion and absorption processes will be impaired during intensive physical S67 activity. These problems are not exclusively restricted to competition days. During intensive training days the values for energy expenditure are impressive as well. In such circumstances athletes tend to ingest a large number of ‘in-between meals’, often composed of energy-rich snacks, which can be low in protein and micronutrients. As such, their diet may become imbalanced. Especially adapted foods and fluids which are easily digestible and rapidly absorbable may solve this problem. During endurance sports activity the body will use its own energy stores (fat stored as adipose tissue-triacylglycerol and carbohydrate stored as glycogen in liver and muscle). Additionally, small amounts of functional proteins (in the liver, gastrointestinal tract and muscle) will be broken down due to mechanical and metabolic stresses. These losses have to be compensated by supply of the necessary nutrients. At the same time heat will be produced, which to a large extent will be eliminated by production and evaporation of sweat. As a result, fluids and electrolytes will be lost (Brouns et al. 1992). Large sweat losses may pose a risk to health by inducing severe dehydration, impaired blood circulation and heat transfer, leading to heat exhaustion and collapse (Sawka & Pandolf, 1990; Maughan & Noakes, 1991). Insufficient replacement of carbohydrate may lead to hypoglycaemia, central fatigue and exhaustion (Wagenmakers et al. 1991). Inadequate protein intake induces protein loss, especially of muscle and consequently a negative N balance and a reduced performance (Lemon, 1991). These observations show that increased needs for specific nutrients should be met according to the level of daily physical activity and exercise. These requirements depend on the type, intensity and duration of the physical effort. Depending on these factors, specific nutritional measures and dietary interventions can be taken, particularly in the phases of preparation, competition and recuperation. Some groups of athletes compete in sport events where a low body weight forms a prerequisite to perform well or to compete in a certain weight category. These athletes are on the one hand training frequently and intensively, but on the other hand they have to maintain a low body weight. The low energy intakes may in these situations lead to a low intake of essential nutrients such as protein, Fe, Ca and vitamins; the required carbohydrate intake to balance the carbohydrate used in training may also be marginal (Van Erp-Baart et al. 1989a,b). This aspect should receive special attention. In the next paragraphs sports nutritional aspects, specifically those related to the macro- and micronutrients which make up the daily nutrition of individuals involved in heavy physical work or exercise, will be described (for review, see Brouns, 1993). 4.2. Carbohydrates Carbohydrate is the most important nutrient for highintensity performance. Energy release from carbohydrate is up to three times as fast as from fat. However, carbohydrate stores in the form of liver and muscle glycogen in the body are small. This limits the duration of high-intensity exercise. Apart from decreasing performance, carbohydrate depletion S68 W. H. M. Saris et al. induces an increased utilization of protein for energy production (Wagenmakers et al. 1991). This results in the production of NH3 , which may enhance fatigue. Carbohydrate ingestion during exercise induces sparing of the body’s carbohydrate stores, reduction of protein utilization and NH3 production, and a delay of fatigue or improvement of performance (Costill, 1988; Coyle, 1991a,b; Wagenmakers et al. 1991). Adequate carbohydrate ingestion between training sessions or intense performance is of utmost importance to avoid progressive glycogen depletion and resulting fatigue development or overtraining. Carbohydrate sources to be used during exercise should be rapidly absorbable, i.e. have a high GI, and should be combined with sufficient fluid intake. Factors that determine whether foods are ‘fast or slow’ carbohydrate sources have been reviewed earlier. Food properties such as particle size, integrity of cellular structure, dietary fibre content, presence of organic acids, etc. determine the rate at which the carbohydrates are absorbed. The differences between starch-containing products disappear when the starch is extracted from the original source and is ingested as pure starch. Glucose, sucrose and maltodextrins (glucose polymers) solubilized in water are all absorbed at similar rates and lead to equal oxidation rates (Coyle, 1991b; Hawley et al. 1992). Exceptions are fructose and galactose which are absorbed relatively slowly and also have lower oxidation rates than the aforementioned carbohydrate sources. The effects of training and dietary factors in the modulation of muscle glycogen as well as substrate utilization have recently been reviewed (Hargreaves, 1995; Brouns, 1997a). 4.3. Fat Fat is a ‘slow’ energy source (Newsholme & Start, 1973). When using fat as a prime energy source, athletes can only work at 40–60 % of their maximal capacity. Nevertheless, increased fat utilization, as a result of training, reduces the use of carbohydrate from the stores in the body, and thus will influence carbohydrate availability and fatigue (Björntorp, 1991b). The idea that high-fat diets lead to adaptations which enhance fat oxidation during exercise in favour of glycogen sparing and performance capacity has been shown in animals. However, there is currently no evidence that this is also the case in human subjects. This topic has been reviewed by Brouns & van der Vusse (1998) and Jeukendrup et al. (1998). Recently a number of studies have focused on the effect of medium-chain triacylglycerol (MCT) ingestion on substrate oxidation and performance. The results show that MCT is rapidly absorbed and oxidized. However, MCT oxidation does not lead to an increase in total fat oxidation, nor to a sparing of muscle glycogen. Additionally it was shown that the amount of MCT which can be consumed without causing gastrointestinal upset is < 30 g. From these data is was concluded that MCT is a rapidly available energy source for athletes but that its consumption does not lead to measurable performance benefits (Jeukendrup et al. 1996). Daily fat intake in athletes should be relatively low, 20–30 % energy, allowing for an increase in the proportion of carbohydrate in the diet in favour of restoring tissue glycogen levels after the daily training or competition sessions (Björntorp, 1991b; Coyle, 1991a,b). From a general health point of view, saturated fat sources should be avoided and vegetable-, fish- and plant-oil-based foods should be promoted. 4.4. Protein Protein is needed for muscle growth, repair of tissues and enzymic adaptations. The protein requirement of athletes is increased and, according to present knowledge, amounts to approximately 1.2–1.8 g/kg body weight (Lemon, 1991a,b). The reason for this increase is an enhanced utilization of amino acids in oxidative energy production during physical exercise, a process which is known to be intensified at higher work levels and in a state of carbohydrate store depletion (Wagenmakers et al. 1991). Generally, however, the increased protein requirement is covered by an increased food intake to cover the daily energy needs. Studies carried out during the Tour de France, for example, have shown that the mean daily energy intake amounted to < 24 MJ (6000 kcal) (Saris et al. 1989; Van Erp-Baart et al. 1989a). Since the daily protein consumption (% of energy intake) remained the same, the cyclists ingested > 3 g protein/kg body weight per d, more than enough to cover the increased requirement. However, there are examples of athletes who may need to be protein-supplemented or need to increase the protein density of the diet. Athletes who ingest low-energy diets will generally have low protein intakes, which may not compensate for the net N loss from the body and will influence synthetic processes and training adaptations. To these categories belong, amongst others, body builders, athletes who have to fit into certain weight categories, gymnasts, dancers and female long-distance runners, and under certain circumstances vegetarian athletes (Van ErpBaart et al. 1989a). Protein supplements such as milk and vegetable-protein hydrolysates may be useful in these cases. However, one should keep in mind that protein intake and/or supplementation above levels normally required (1.2–1.8 g/kg body weight) will not enhance muscle growth or performance (Lemon, 1991a,b). The use of single amino acids, such as arginine, ornithine, tryptophan and branched-chain amino acids to influence metabolic pathways involved in fatigue development and hormone production, needs further research to make definite statements, especially because data on the safety aspects of high intakes of single amino acids in exercising athletes are generally lacking. Although amino acids, when supplied intravenously in high dosages, have been observed to enhance hormone release, there are no indications that this is also the case after oral ingestion of amounts which are normally present in supplements. Recently the ingestion of glutamine has been proposed as a means of supporting the immune competence of athletes involved in intense daily training. In addition, positive effects of branched-chain amino acid ingestion on the aetiology of central fatigue as well as on immune variables have recently been suggested. However, the currently available scientific data do not allow us (yet) to make Functional foods and substrate metabolism any claim in these respects (van Hall et al. 1995; Klarlund Pedersen & Rohde, 1997). 4.5. Fluid and electrolytes During prolonged physical exercise an adequate supply of fluid is of prime concern for the performance capacity as well as the health status of the athlete, especially when performing in the heat. Progressive fluid loss from the body, by means of sweating and breathing and in endurance events also by diarrhoea, is associated with a decreased blood flow through the extremities, a reduced plasma volume and central blood volume, a reduction in sweating and heat dissipation, and under circumstances of high intensity work in the heat, with heat stroke and collapse (Sawka & Pandolf, 1990; Maughan, 1991; Maughan & Noakes, 1991). Dehydration of > 1.5 litres is known to reduce the O2 transport capacity of the body and to induce fatigue. Appropriate rehydration is known to counter these effects and to delay fatigue (Sawka & Pandolf, 1990; Maughan, 1991; Maughan & Noakes, 1991). In contrast to plain water, the addition of Na and carbohydrate (up to 80 g/l) to rehydration drinks is known to stimulate water absorption (Maughan, 1991; Maughan & Noakes, 1991), as well as to supply energy. Addition of other electrolytes should generally not exceed the levels of loss with whole-body sweat (Brouns et al. 1992). Sport rehydration drinks should not be hypertonic because this reduces the rate of net water absorption (Maughan, 1991). Recently the importance of inclusion of a high amount of Na in post-exercise rehydration beverages has been underlined by Maughan et al. (1997). Aspects of dehydration and rehydration have recently been reviewed by Brouns (1997b). 4.6. Minerals Exercise is known to be associated with increased mineral losses, through sweating during exercise and through urine in the post-exercise phase (Costill, 1988). In general, athletes may develop an impaired mineral status in cases of poor selection of food items which may lead to an inappropriate intake of some minerals, compared with the daily losses. Impaired Fe, Zn and Mg status are known to induce malperformance and muscle weakness and are often associated with the occurrence of muscle cramp. The latter needs further research to validate the direct influence of mineral deficits. As with most nutrients, mineral intake depends on the quality of the diet and the amount of energy consumed (Van Erp-Baart et al. 1989b). Therefore, athletes consuming low-energy diets may be at risk of marginal mineral intake, especially of Zn, Fe and Mg (Clarkson, 1991). Vegetarian athletes are especially prone to Fe deficiency. In these cases the ingestion of a daily supplement may be recommended. 4.7. Trace elements The importance of an adequate trace element status for athletic training, performance and recovery has only received attention relatively recently (Kieffer, 1986). As S69 with minerals, trace elements are increasingly lost as a result of intensive physical training. Trace-element losses with sweat (Cu) and urine (Cr) may exceed the daily recommended intakes. The diet itself may strongly affect these losses. High carbohydrate intakes, especially of high-GI carbohydrates, are known to enhance losses of Cr, whereas diets rich in dietary fibre, often consumed by endurance athletes and vegetarians, are known to reduce trace-element absorption. Since science has become aware of the fact that exercise leads to enhanced tissue and cell damage, the importance of Se, which acts within the free-radical scavenging processes, has received attention. Much research is needed in this field, but it is considered that supplementation with amounts not exceeding the recommended safe daily intakes will contribute to adequate daily intakes in athletes. 4.8. Vitamins Vitamins belong among the nutrients which have received the most attention. Vitamins are essential co-factors in many enzymic reactions involved in energy production and in protein metabolism. Any shortage of a vitamin is therefore linked to sub-optimal metabolism, which in the long term will result in decreased performance or even illness (Van der Beek, 1985). Additionally, some of the vitamins act as antioxidant substances and therefore have a protective role for tissue and cell integrity, which may be threatened in the case of metabolic stress (Bendich, 1991). Vitamin supplementation has, thus, been shown to restore performance capacity in cases of a vitamin deficit and also to reduce tissue damage due to free radicals (Bendich, 1991; Brouns, 1993). Vitamin supplementation with quantities exceeding those needed for optimal blood levels has never been shown to improve performance (Van der Beek, 1985). As with minerals and trace elements, athletes involved in intensive training, but consuming low-energy diets, are the most prone to marginal vitamin intakes. In general it can be concluded that vitamin restoration of energy-dense processed foods for supplementation with preparations will not enhance performance but may, in athletic populations, contribute to adequate daily intakes. Daily intake of a low-dose vitamin preparation or nutrient preparations, supplying not more than the recommended daily intake or safe intake, may be advisable in periods of intensive training or in any situation where athletes abstain from a normal diet, such as during periods of limited food intake combined with intensive training (especially in females and in athletes who have to fit into certain weight categories). 4.9. Ergogenic supplements Substances such as creatine, caffeine, L-carnitine, aspartate, NaHCO3 , bee pollen, specific amino acids, etc. have recently received scientific attention due to their possible influence on performance, fatigue and recovery. In many cases such substances need further research to come to conclusive scientific evidence and recommendations (Brouns, 1993). In the case of L-carnitine, for example, it has long been suggested that the oral intake of this compound leads to S70 W. H. M. Saris et al. improved fat oxidation which may improve athletic performance and reduce body-fat levels (Wagenmakers, 1991). These suggestions were based on in vitro studies. However, more recent and well-controlled human studies showed no change in muscle carnitine content after L-carnitine supplementation, nor effects on fat oxidation and performance. The intake of bee-pollen preparations and of royal jelly has been said to initiate allergic responses in susceptible athletes, which may lead to anaphylactic shock and even death. Since there is no evidence for any performance effects of bee pollen its use by athletes should be discouraged. Recently a number of publications have appeared with respect to creatine. The use of creatine has so far been shown to be safe and to enhance physical performance in events lasting 30 s–3 min as well as in the case of repetitive sprinting. Creatine supplementation (four doses of 5 g/d for 4–6 d) has been shown to increase the muscle total creatine content in most but not all subjects. It is hypothesized that this effect causes a better ATP transfer from mitochondria to cytosol by the creatine shuttle. This results in a reduced formation of lactic acid and NH3 at a given work load and improved performance at a freely-chosen maximal workload (Balsom et al. 1994). Creatine is a clear example of a new functional food ingredient. Since there is a growing awareness of functional food ingredients which may affect athletic performance and health status positively, there will be a significant increase in the number of scientific studies needed to obtain evidence to support any functional claim. This development will be stimulated with the increasing knowledge about nutritional substances, which are involved in the different metabolic pathways, including brain metabolism. Judgement of the role of food-derived substances as being nutritional or pharmacological, requires research and optimal interaction between science and the food industry. 4.10. Conclusions and further research Oral rehydration products for athletes were one of the first categories in the segment of functional foods and drinks for which scientific evidence was obtained on all levels of functionality: rapid gastric emptying, fast intestinal absorption, improved water retention, improved thermal regulation and improved physical performance and delayed fatigue. Generally, research on nutrition for athletes has shown that the exercise-induced stress on the human body depends on the duration, intensity and type of exercise. Intense endurance exercise is characterized by changes in the functionality of the gastrointestinal system. This leads to the fact that nutrient and fluid supply during endurance exercise cannot be realized by ingestion of normal meals. Liquid food formulas, established to deliver fluid and energy-providing nutrients in a convenient way and easily digestible form, have been shown to be of benefit to athletes. Exercise-induced losses of N, minerals, vitamins and trace elements should be replenished by ingesting larger amounts of high-quality food with normal meals. However, this may be problematic in all cases where lowenergy diets are combined with intense training or in the case of multiple-day competitions such as cycling tours. The use of special meals or food products and micronutrient supplementation will help ensure adequate intakes under these conditions. The ever increasing amount of daily high-intensity training leads to a high stress on the metabolic machinery, the musculo-skeletal system and the hormonal system. There is a growing awareness that the supply of food ingredients or food-derived substances may interact with the biochemical and physiological systems involved with physical and mental performance, as well as with recovery from intensive training and hence with the physical well-being and health of the athlete. Therefore it is emphasized that future studies are directed to the following goals: (1) to define the safety or toxicity of promising functional food substances or formulas, to be taken shortly before, during or after exercise. This should be realized with respect to short-term, single dosage intake, as well as long-term chronic intake; (2) to define the functionality of these substances in terms of their influence on factors limiting performance and recovery; (3) to support any claim made with respect to such improvements. References Acheson KJ, Zahorska-Markiewicz B, Pittet Ph, Anantharaman K & Jéquier E (1980) Caffeine and coffee. Their influence on metabolic rate and substrate utilization in normal weight and obese individuals. American Journal of Clinical Nutrition 33, 989–997. Ader M & Bergman RN (1987) Insulin sensitivity in the intact organism. Ballière’s Clinical Endocrinology and Metabolism 1, 879–910. American Diabetes Association (1996) Nutrition recommendations and principles for people with diabetes mellitus. Diabetes Care 20, S14–S17. American Diabetes Association (1997a) Guide to diagnosis and classification of diabetes mellitus and other categories of glucose intolerance. Diabetes Care 20, S21–S24. American Diabetes Association (1997b) Clinical practice recommendations. Tests of glycaemia in diabetes. Diabetes Care 20, S18–S20. Asp NG (1996) Dietary carbohydrates: classification by chemistry and physiology. Food Chemistry 57, 9–14. Asp NG & Björck I (1992) Resistant starch. Trends in Food Science and Technology 3, 111–114. Asp NG, van Amelsvoort JMM & Hautvast JGAJ (1996) Nutritional implications of resistant starch. Nutrition Research Reviews 9, 1–31. Astrup A, Toubro S, Cannon S, Hein P, Breum L & Madsen J (1989) Caffeine: a double-blind placebo-controlled study of its thermogenic metabolic and cardiovascular effects in healthy volunteers. American Journal of Clinical Nutrition 51, 759–767. Austin MA, King M-C, Vranizan KM & Krauss RM (1990) Atherogenic lipoprotein phenotype. A proposed genetic marker for coronary heart disease. Circulation 82, 495–506. Balsom PD, Söderlund K & Ekblom B (1994) Creatine in humans with special reference to creatine supplementation. Sports Medicine 18, 268–280. Bär A, Kunz M, Schiweck H & Schwarz E (1994) In Ullmanns’s Encyclopedia of Industrial Chemistry, pp. 413–437. Weinheim: VCH Verlagsgesellschaft. Bellisle F, Blundell JE, Fantino M, Fern E, Fletcher RJ, Lambert J, Functional foods and substrate metabolism Roberfroid M, Specter S, Westenhöfer J & WesterterpPlantenga MS (1998) Functional food science and behaviour and psychological functions. British Journal of Nutrition 80, Suppl. 1, S173–S193. Bendich A (1991) Exercise and free radicals: effects of antioxidant vitamins. In Advances in Nutrition and Top Sport vol. 32, pp. 59–78 [F Brouns, E Newsholme and WHM Saris, editors]. Basel: Karger. Björck I (1996) Starch: nutritional aspects. In Carbohydrates in Food, pp. 505–554 [AC Eliasson, editor]. New York, NY and Basel: Marcel Dekker Inc. Björntorp P (1991a) Metabolic implications of body fat distribution. Diabetes Care 14, 1132–1143. Björntorp P (1991b) Importance of fat as a support nutrient for energy metabolism of athletes. Journal of Sports Sciences 9, 71–76. Björntorp P & Brodoff BN (editors) (1992) Obesity. Philadelphia, PA: J.B. Lippincott. Blaak EE & Saris WHM (1995) Health aspects of various digestible carbohydrates. Nutrition Research 15, 1547–1573. Blaak EE & Saris WHM (1996) Postprandial thermogenesis and substrate utilization after ingestion of different dietary carbohydrates. Metabolism 45, 1235–1242. Blaak EE, van Baak MA, Kemerink GJ, Pakbiers MTW, Heidendal GAK & Saris WHM (1994) Beta-adrenergic stimulation of whole body energy expenditure and skeletal muscle metabolism in relation to weight reduction in obese men. American Journal of Physiology 267, E316–E322. Blundell JE & Burley VJ (1987) Satiation, satiety and the action of fibre on food intake. International Journal of Obesity 11, Suppl. 1, 9–25. Bogardus C, Lillioja S, Mott DM, Hollenbeck C & Reaven G (1985) Relationship between degree of obesity and in vivo insulin action in man. American Journal of Physiology 248, E286–E291. Bolton-Smith C & Woodward M (1994) Dietary composition and fat to sugar ratios in relation to obesity. International Journal of Obesity 18, 820–828. Borkman M, Storlien LH, Pan DA, Jenkins AB, Chisholm DJ & Campbell LV (1993) The relation between insulin sensitivity and the fatty-acid composition of skeletal-muscle phospholipids. New England Journal of Medicine 328, 238–244. Bouchard C (1992) Genetic aspects of human obesity. In Obesity, pp. 343–351 [P Björntorp and BN Brodoff, editors]. Philadelphia, PA: J.B. Lippincott. Bouchard C & Perusse L (1993) Genetics of obesity. Annual Review of Nutrition 13, 337–354. Bowman BA & Forbes AL (1993) Health effects of dietary fructose. American Journal of Clinical Nutrition 58, 7219–8235. Brand Miller JC (1994) Importance of GI in diabetes. American Journal of Clinical Nutrition 59, Suppl., 747S–752S. Brighenti F, Castellani G, Benini L, Casiraghi M, Leopardi E, Crovetti R & Testolin G (1995) Effect of neutralized and native vinegar on blood glucose and acetate responses to a mixed meal in healthy subjects. European Journal of Clinical Nutrition 49, 242–247. Brouns F (1993) Nutritional Needs of Athletes. Chichester: John Wiley & Sons Ltd. Brouns F (1997a) The effect of athletic training and dietary factors on the modulation of muscle glycogen. In The Clinical Pharmacology of Sport and Exercise, pp. 181–193 [T Reilly and M Orne, editors]. Amsterdam: Elsevier Science B.V. Brouns F (1997b) Aspects of dehydration and rehydration in sport. In Nutrition and Fitness. Metabolic and Behavioral Aspects in Health and Disease, pp. 63–80 [AP Simopoulos and KN Pavlou, editors]. Basel: Karger. Brouns F, Saris WHM & Schneider H. (1992) Rationale for upper limits of electrolyte replacement during exercise. International Journal of Sport Nutrition 2, 229–238. S71 Brouns F & Van der Vusse GJ (1998) Utilization of lipids during exercise in humans: metabolic and dietary constraints. British Journal of Nutrition 79, 117–128. Cameron-Smith D, Colguhoun EQ, Hettiarachchi JM & Clark MG (1990) Capsaicin and dihydrocapsaicin stimulate oxygen consumption in the perfused rat hindlimb. International Journal of Obesity 14, 259–270. Campfield LA, Smith FJ & Burn P (1997) OB protein: a hormonal controller of central neural networks mediating behavorial, metabolic and neuroendocrine responses. Endocrinology and Metabolism 4, 81–102. Chandra RK (1992) Effect of vitamin and trace-element supplementation on immune responses and infection in elderly subjects. Lancet 340, 1124–1127. Clarkson PM (1991) Minerals: exercise performance and supplementation in athletes. Journal of Sports Sciences 9, 91–116. Colditz GA, Giovanucci E & Rimm ER (1991) Alcohol intake in relation to diet and obesity in women and men. American Journal of Clinical Nutrition 54, 49–55. Costill DL (1988) Carbohydrates for exercise: dietary demands for optimal performance. International Journal of Sports Medicine 9, 1–18. Coyle EF (1991a) Timing and method of increased carbohydrate intake to cope with heavy training, competition and recovery. Journal of Sports Sciences 9, 29–52. Coyle EF (1991b) Carbohydrate feedings: effects on metabolism, performance and recovery. In Advances in Nutrition and Top Sport, vol. 32, pp. 1–14 [F Brouns, WHM Saris and EA Newsholme, editors]. Basel: Karger. Cummings J (1993) Dietary fibre intakes in Europe. An overview. In Dietary Fibre Intakes in Europe. A Survey Conducted by Members of the Management Committee of COST 92: Metabolic and Physiological Aspects of Dietary Fibre in Food, pp. 11–19 [JH Cummings and W Frolich, editors]. Brussels: Commission of the European Communities, Directorate-General Science, Research and Development. Cummings JH & Macfarlane GT (1991) The control and consequences of bacterial fermentation in the human colon. Journal of Applied Bacteriology 70, 443–459. D’Alessio DA, Kavle EC, Mozzoli MA, Smalley KJ, Polansky M, Kendrick ZV, Owen LR, Bushman MC, Boden G & Owen OE (1992) Thermic effect of food in lean and obese men. Journal of Clinical Investigation 81, 1781–1789. Després J-P (1993) The insulin resistance–dyslipidemia syndrome: the most prevalent cause of coronary artery disease? Canadian Medical Association Journal 148, 1339–1340. Després J-P, Lamarche B, Mauriège P, Cantin B, Dagenais GR, Moorjani S & Lupien PJ (1996) Hyperinsulinemia as an independent risk factor for ischemic heart disease. New England Journal of Medicine 334, 952–957. Diabetes and Nutrition Study Group of the European Association for the Study of Diabetes (1995) Recommendations for the nutritional management of patients with diabetes mellitus. Diabetes, Nutrition and Metabolism 8, 1–4. Drent ML & Van der Veen EA (1993) Lipase inhibition: a novel concept in the treatment of obesity. International Journal of Obesity 17, 241–244. Drewnowski A, Brunzell JD, Sande K, Iverius PH & Greenwood MRC (1985) Sweet tooth reconsidered: taste responsiveness in human obesity. Physiology and Behavior 35, 617–622. Drewnowski A, Kurth C, Holden Wiltse J & Saari J (1992) Food preferences in human obesity. Carbohydrates vs fats. Appetite 18, 207–211. Dulloo AG, Geissler CA, Horton T, Collins A & Miller DS (1989) Normal caffeine consumption: influence on thermogenesis and daily energy expenditure in lean and obese human volunteers. American Journal of Clinical Nutrition 49, 44–50. S72 W. H. M. Saris et al. Dulloo AG, Seydoux J & Girardier L (1996) Tealine and thermogenesis: interactions between polyphenols, caffeine and sympathetic activity. International Journal of Obesity 20, Suppl. 4, 71 Abstr. Durnin J (1992) Energy metabolism in the elderly. In Nutrition in the Elderly. Nestlé Nutrition Workshop Series no. 29 [H Munro and G Schlierf, editors]. New York, NY: Raven Press. Dysseler P & Hoffem D (1994) Estimation of resistant starch intake in Europe. In Proceedings of the Concluding Plenary Meeting of EURESTA. European FLAIR Concerted Action no. 11 (COST 911), pp. 84–85 [N-G Asp, JMM Van Amelsvoort and JGAJ Hautvast, editors]. Wageningen: EURESTA. Eldershaw TPD, Colguhorn EQ, Dora KA, Peng ZC & Clark MG (1992) Pungent principles of ginger are thermogenic in the perfused rat hindlimb. International Journal of Obesity 16, 755–763. Englyst H, Kingman S & Cummings J (1992) Classification and measurement of nutritionally important starch fractions. European Journal of Clinical Nutrition 46, Suppl. 2, 33S–50S. Farmer JA & Gotto AM Jr (1995) Currently available hypolipidaemic drugs and future therapeutic developments. Baillière’s Clinical Endocrinology and Metabolism 9, 825–847. Felber JP, Jéquier E, Randin JP, Tappy L & Vaillamoz D (1987) Comparative study of maltitol and sucrose by means of continuous indirect calorimetry. Journal of Parenteral and Enteral Nutrition 11, 250–254. Finley JW & Leveille GA (1996) Macronutrient substitutes. In Present Knowledge in Nutrition, pp. 581–595 [EE Ziegler, editor]. Washington, DC: ILSI Press. Finley JW, Leveille GA & Smith ER (1994) Overview on salatrim, a family of low-calorie fats. Journal of Agricultural and Food Chemistry 42, 432–434. Flatt JP (1996) Glycogen levels and obesity. International Journal of Obesity 20, S1–S11. Food and Agriculture Organization/World Health Organization (1998) Joint FAO/WHO Expert Consultation on Carbohydrates in Human Nutrition. Food and Nutrition Paper no. 66, Rome: FAO. Food and Drug Administration (1996) In Food Additives Permitted for Direct Addition to Food for Human Consumption: Olestra. Federal register 1/30/1996;21 CRF Part 172, docket no. 87F– 0179. Frayn KN (1993) Insulin resistance and lipid metabolism. Current Opinion in Lipidology 4, 197–204. Frayn KN (1995) Physiological regulation of macronutrient balance. International Journal of Obesity 19, S4–S10. Frayn KN (1996) Metabolic Regulation: A Human Perspective, p. 243. Colchester: Portland Press. Frayn KN & Kingman SM (1995) Dietary sugars and lipid metabolism in humans. American Journal of Clinical Nutrition 62, Suppl., 250S–263S. Frayn KN, Williams CM & Arner P (1996) Are increased plasma non-esterfied fatty acid concentrations a risk marker for coronary heart disease and other chronic diseases? Clinical Science 90, 243–253. Frost G, Keogh B, Smith D, Akinsanya K & Leeds A (1996) The effect of low-glycaemic carbohydrate on insulin and glucose response in vivo and in vitro in patients with coronary heart disease. Metabolism 45, 669–672. Gerstein HC & Yusuf S (1996) Dysglycaemia and risk of cardiovascular disease. Lancet 347, 949–950. Gibney MJ, Maloney M & Shelley E (1987) Food and nutrient intakes in randomly selected healthy adults. British Journal of Nutrition 61, 129–137. Granfeldt Y, Björck I, Drews A & Tovar J (1992) An in vitro procedure based on chewing to predict metabolic response to starch in cereal and legume products. European Journal of Clinical Nutrition 46, 649–660. Granfeldt Y, Hagander B & Björck I (1995) Metabolic responses to starch in oat and wheat products. On the importance of food structure in complete gelatinization or presence of viscous fibre. European Journal of Clinical Nutrition 49, 189–199. Griffin BA & Zampelas A (1995) Influence of dietary fatty acids on the atherogenic lipoprotein phenotype. Nutrition Research Reviews 8, 1–26. Grinker JA (1978) Obesity and sweet taste. American Journal of Clinical Nutrition 31, 1078–1087. Haber GB, Heaton KW, Murphy D & Burroughs LF (1977). Depletion and disruption of dietary fibre: effects on satiety, plasma-glucose, and insulin. Lancet, 679–682. Hamilton CC & Anderson JW (1992) Fiber and weight management. Journal of Florida Medical Association 9, 379–381. Hargreaves M (1995) Skeletal muscle carbohydrate metabolism during exercise. In Exercise Metabolism, pp. 41–72 [M Hargreaves, editor]. Champaign, IL: Human Kinetics Publishers, Inc. Haussinger D, Roth E, Lang F & Gerok W (1993) Cellular hydration state: an important determinant of protein catabolism in health and disease. Lancet 341, 1330–1332. Hawley JA, Dennis SC & Noakes TD (1992) Oxidation of carbohydrate ingested during prolonged endurance exercise. Sports Medicine 14, 27–42. Henry CJK & Emery B (1986) Effect of spiced food on metabolic rate. Human Nutrition: Clinical Nutrition 40C, 165–168. Hill JO, Peters JC, Lin D, Yakuba F, Greene H & Swift L (1993) Lipid accumulation and body fat distribution is influenced by type of dietary fat fed to rats. International Journal of Obesity 17, 223–236. Hill JO & Prentice AM (1995) Sugar and body weight regulation. American Journal of Clinical Nutrition 62, Suppl., 2645–2745. Hokanson JE & Austin MA (1996) Plasma triglyceride level is a risk factor for cardiovascular disease independent of highdensity lipoprotein cholesterol level: a meta-analysis of population-based prospective studies. Journal of Cardiovascular Risk 3, 213–219. Hollenbeck CB & Coulston AM (1991) Effects of dietary carbohydrate and fat intake on glucose and lipoprotein metabolism in individuals with diabetes mellitus. Diabetes Care 14, 774–785. Holm J, Björk I, Drews A & Asp NG (1986) A rapid method for the analysis of starch. Stärke 38, 224–226. Hornstra G, Barth CA, Galli C, Mensink RP, Mutanen M, Riemersma RA, Roberfroid M, Salminen K, Vansant G & Verschuren PM (1998) Functional food science and the cardiovascular system. British Journal of Nutrition 80, Suppl. 1, S113–S146. Jenkins DJA, Wolever TMS, Jenkins A, Brighenti F, Vuksan V, Venketshwer Rao A, Cunnane S, Ocana A, Corey P, Vezina C, Connelly P, Buckley G & Patten R (1991) Specific types of colonic fermentation may raise low-density-lipoprotein cholesterol concentrations. American Journal of Clinical Nutrition 54, 141–147. Jenkins D, Wolever TMS, Jenkins A, Giordano C, Guidici S, Thompson L, Kalmusky J, Josse R & Wong G (1986) Low glycemic response to traditionally processed wheat and rye products: bulgur and pumpernickel bread. American Journal of Clinical Nutrition 43, 516–520. Jenkins DJA, Wolever TMS, Taylor RH, Ghafari H, Barker HM, Fielden H, Baldwin JM, Bowling AC, Newman HC, Jenkins AL & Goff DV (1981) Glycemic index of foods: a physiological basis for carbohydrate exchange. American Journal of Clinical Nutrition 34, 362–366. Jenkins DJA, Wolever TMS, Taylor RH, Griffiths C, Krzeminska K, James BA, Lawrie JA, Bennet CM, Goff DV, Sarson DL & Bloom SR (1982) Slow release dietary carbohydrate improves second meal tolerance. American Journal of Clinical Nutrition 35, 1339–1346. Functional foods and substrate metabolism Jeukendrup AE, Saris WHM, van Diessen R, Brouns F & Wagenmakers AJM (1996) Effect of endogenous carbohydrate availability on oral medium chain triglycerides oxidation during prolonged exercise. Journal of Applied Physiology 80, 949–954. Jeukendrup AE, Saris WHM & Wagenmakers AJM (1998) Fat metabolism during exercise: a review. International Journal of Sports Medicine (In the Press). Jones JM (1996) Fat substitutes: nutritional promise or potential disaster? Chemistry & Industry 3, 494–498. Jones PJH, Pencharz JB & Clandinin MB (1985) Whole body oxidation of dietary fatty acids: implications for energy utilization. American Journal of Clinical Nutrition 42, 769–777. Jones PJH, Ridgen JE, Phang PT & Birmingham CL (1992) Influence of dietary fat polyunsaturated to saturated ratio on energy substrate utilization in obesity. Metabolism 41, 396–401. Jones PJH & Schoeller DA (1988) Polyunsaturated : saturated ratio of diet fat influences energy substrate utilization in the human. Metabolism 37, 145–151. Kieffer F (1986) Trace elements: their importance for health and physical performance. Deutsche Zeitschrift für Sportmedizin 37, 118–123. Klarlund Pedersen B & Rohde T (1997) Exercise, glutamine and the immune system. In Exercise Immunology, pp. 75–88 [B Klarlund Pedersen, editor]. Heidelberg: Springer-Verlag. Klesges RC, Mealer CZ & Klesges LM (1994) Effect of alcohol intake on resting energy expenditure in young women social drinkers. American Journal of Clinical Nutrition 59, 805–809. Langkilde A-M & Andersson H (1994) In vivo quantification of resistant starch in EURESTA reference materials using the ileostomy model. In Proceedings of the Concluding Plenary Meeting of EURESTA. European FLAIR-Concerted Action no. 11 (COST 911), pp. 31–32 [N-G Asp, JMM Van Amelsvoort and JGAJ Hautvast, editors]. Wageningen: EURESTA. Lemon PWR (1991a) Effect of exercise on protein requirements. Journal of Sports Sciences 9, 53–70. Lemon PWR (1991b) Protein and amino acid needs of the strength athlete. International Journal of Sport Nutrition 1, 127–145. Life Sciences Research Office/Federation of American Societies for Experimental Biology (1991) Evaluation of the Health Aspects of Caprenin. Bethesda, MD: LSRO/FASEB. Life Sciences Research Office/Federation of American Societies for Experimental Biology (1994) The Evaluation of Energy of Certain Sugar Alcohols Used as Food Ingredients. Bethesda, MD: LSRO/FASEB. Liljeberg H & Björck I (1996) Delayed gastric emptying rate as a potential mechanism for lowered glycemia after eating sourdough bread: studies in humans and rats using test products with added organic acids or an organic salt. American Journal of Clinical Nutrition 64, 886–893. Lissner L, Levitsky DA, Strupp BJ, Kalkwarf HJ & Roe DA (1987) Dietary fat and the regulation of food intake in human subjects. American Journal of Clinical Nutrition 44, 772–778. Livesey G (1994) Energy value of resistant starch. In Proceedings of the Concluding Plenary Meeting of EURESTA. European FLAIR-Concerted Action no. 11 (COST 911), pp. 56–62 [N-G Asp, JMM Van Amelsvoort and JGAJ Hautvast, editors]. Wageningen: EURESTA. Lovejoy J & DiGirolamo M (1992) Habitual dietary intake and insulin sensitivity in lean and obese adults. American Journal of Clinical Nutrition 55, 1174–1179. Ma J, Folsom AR, Melnick SL, Eckfeldt JH, Sharrett AR, Nabulsi AA, Hutchinson RG & Metcalf PA (1995) Associations of serum and dietary magnesium with cardiovascular disease, hypertension, diabetes, insulin, and carotid arterial wall thickness: the ARIC study. Atherosclerosis Risk in Communities Study. Journal of Clinical Epidemiology 48, 927–940. Macdonald I, Debry G & Westerterp K (1993) Alcohol and S73 overweight. In Health Issues Related to Alcohol Consumption, pp. 263–280 [PM Verschuren, editor]. Washington, DC: ILSI Europe. Macfarlane GT & Cummings JH (1991) The colonic flora, fermentation, and large bowel digestive function. In The Large Intestine: Physiology, Pathophysiology and Disease, pp. 51–93 [SF Phillips, JH Pemberton and RG Shorter, editors]. New York, NY: Raven Press. McKeigue PM, Shah B & Marmot MG (1991) Relation of central obesity and insulin resistance with high diabetes prevalence and cardiovascular risk in South Asians. Lancet 337, 382–386. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF & Turner RC (1985). Homeostasis model assessment: insulin resistance and ß-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 28, 412–419. Maughan RJ (1991) Fluid and electrolyte loss and replacement in exercise. Journal of Sports Sciences 9, 117–142. Maughan RJ, Leiper JB & Shirreffs SM (1997) Factors influencing the restoration of fluid and electrolyte balance after exercise in the heat. British Journal of Sports Medicine 31, 175–182. Maughan RJ & Noakes TD (1991) Fluid replacement and exercise stress – a brief review of studies on fluid replacement and some guidelines for the athlete. Sports Medicine 12, 16–31. Mertz W (1993) Chromium in human nutrition: a review. Journal of Nutrition 123, 626–633. Miller DK, Morley JE & Rubenstein LZ (1995) An overview of international aging and nutrition. In Geriatric Nutrition, pp. 1–4 [JE Morley, Z Glick and LZ Rubenstein, editors]. New York, NY: Raven Press. Miller WC, Lindermann AK, Wallace J & Nierderpruem M (1990) Diet composition, energy intake and exercise in relation to body fat in men and women. American Journal of Clinical Nutrition 52, 426–430. Mitchell BD, Kammerer CM, Mahaney MC, Blangero J, Comuzzie AG, Atwood LD, Haffner SM, Stern MP & MacCluer JW (1996) Genetic analysis of the IRS. Pleiotropic effects of genes influencing insulin levels on lipoprotein and obesity measures. Arteriosclerosis, Thrombosis and Vascular Biology 16, 281–288. Mitsuoka T, Hata Y & Takahashi Y (1986) Effects of long term intake of Neosugar on intestinal flora and serum lipids. In Proceedings of the 3rd Neosugar Research Conference, topic 1–4. Tokyo. Morley JE (1995) Anorexia of aging and protein–energy-undernutrition. In Geriatric Nutrition, pp. 75–78 [JE Morley, Z Glick and LZ Rubenstein, editors]. New York, NY: Raven Press. Move M & Bohmer T (1991) Anorexia of aging. Journal of the American Geriatric Society 39, 1089–1092. Nestel PJ (1990) Effects of n-3 fatty acids on lipid metabolism. Annual Review of Nutrition 10, 149–167. Newsholme EA & Start C (1973) Adipose tissue and the regulation of fat metabolism. In Regulation in Metabolism, pp. 195–246 [EA Newsholme and C Start, editors]. Chichester: John Wiley and Sons. Nicklas BJ, Goldberg AP, Bunyard LB & Poehlman ET (1995) Visceral adiposity is associated with increased lipid oxidation in obese, postmenopausal women. American Journal of Clinical Nutrition 62, 918–922. Nilsson LH & Hultman E (1974) Liver and muscle glycogen in man after glucose and fructose ingestion. Scandinavian Journal of Clinical and Laboratory Investigation 33, 5–10. Nuttall FQ (1993) Dietary fiber in the management of diabetes. Diabetes 42, 503–508. Pan DA, Hulbert AJ & Storlien LH (1994) Dietary fats, membrane phospholipids, and obesity. Journal of Nutrition 124, 1444–1565. Parillo M, Giacco R, Riccardi G, Pacioni D & Rivellese A (1985) Different glucaemic responses to pasta, bread and potatoes in diabetic patients. Diabetic Medicine 2, 374–377. S74 W. H. M. Saris et al. Parillo M & Riccardi G (1985) Dietary carbohydrates and glucose metabolism in diabetic patients. Diabète et Metabolisme 21, 391–401. Perrotti N, Santoro D, Genovese S, Giacco A, Rivellese A & Riccardi G (1984) Effect of digestible carbohydrates of glucose control in insulin dependent diabetic patients. Diabetes Care 7, 354–359. Prentice AM, Black AE, Coward WA, Davies HL, Goldberg GR, Murgatroyd PR, Ashford J, Sawyer M & Whitehead RG (1986) High levels of energy expenditure in obese women. British Medical Journal 292, 983–987. Prentice AM & Jebb SA (1995) Obesity in Britain: gluttony or sloth? British Medical Journal 311, 437–439. Prentice AM & Poppitt SD (1996) Importance of energy density and macronutrients in the regulation of energy intake. International Journal of Obesity 20, S18–S23. Ravussin E, Lillioja S, Knowler WC, Cristin L, Freymond D, Abbot WGH, Boyce V, Howard BV & Bogardus C (1988) Reduced rate of energy expenditure as a risk factor for body weight gain. New England Journal of Medicine 318, 467–472. Reaven GM (1995) Pathophysiology of insulin resistance in human disease. Physiological Reviews 75, 473–486. Riccardi G (1994) Diet, blood glucose control and long term complications of diabetes: a lesson from the diabetes control and complication trial (DCCT). Nutrition, Metabolism and Cardiovascular Disease 4, 59–60. Riccardi G & Rivellese AA (1987) No indices for selection of carbohydrate foods in the diabetic diet: hopes and limitations. Diabetic Medicine 4, 140–143. Riccardi G & Rivellese AA (1991) Effects of dietary fiber and carbohydrate on glucose and lipoprotein metabolism, in diabetic patients. Diabetes Care 14, 115–125. Ritz P, Krempf M, Cloarec D, Champ M & Charbonnel B (1991) Comparative continuous indirect calorimetry study of two carbohydrates with different glycemic indices. American Journal of Clinical Nutrition 54, 855–859. Rivellese AA, Maffettone A, Lovine C, Di Marino L, Annuzzi G, Mancini M & Riccardi G (1996) Long term effects of fish oil on insulin resistance and plasma lipoproteins in NIDDM patients with hypertriglyceridemia. Diabetes Care 19, 1207–1213. Rolls RJ & Shide DJ (1992) The influence of dietary fat on food intake and body weight. Nutrition Reviews 50, 283–290. Salmerón J, Ascherio A, Rimm EB, Colditz GA, Spiegelman D, Jenkins DJ, Stampfer MJ, Wing AL & Willett WC (1997a) Dietary fiber, glycaemic load and risk of NIDDM in men. Diabetes Care 20, 545–550. Salmerón J, Stampfer MJ, Colditz GA, Manson JE, Wing AL & Willett WC (1997b) Dietary fiber, glycaemic load and risk of non-insulin-dependent diabetes mellitus in women. Journal of the American Medical Association 277, 472–477. Saris WHM, van Erp-Baart MA, Brouns F, Westerterp KR & ten Hoor F (1989) Study on food intake and energy expenditure during extreme sustained exercise: the Tour de France. International Journal of Sports Medicine 10, S26–S31. Sawka MN & Pandolf KB (1990) Effects of body water loss on physiological function and exercise performance. In Perspectives in Exercise Science and Sports Medicine. Fluid Homeostasis During Exercise, vol. 3, pp. 1–38 [CV Gisolfi and DR Lamb, editors]. Carmel: Benchmark Press. Scheppach W (1994) Effects of short-chain fatty acids on gut morphology and function. Gut 35, Suppl. 1, 535–538. Scheppach W, Fabian C, Sachs M & Kasper H (1988) The effect of starch malabsorption on faecal short-chain fatty acid excretion in man. Scandinavian Journal of Gastroenterology 23, 755–759. Schiweck H & Ziesenitz SC (1996) Physiological properties of polyols in comparison with metabolisable saccharides. In Advances in Sweeteners, pp. 56–84 [TH Grenby, editor]. Glasgow: Blackie. Schutz Y & Jéquier E (1989) Failure of dietary fat intake to promote fat oxidation: a factor favoring the development of obesity. American Journal of Clinical Nutrition 50, 307–314. Seaton TB, Welle SL, Warenko MK & Campbell RG (1986) Thermic effect of medium chain and long chain triglycerides in man. American Journal of Clinical Nutrition 44, 630–634. Segal KR, Edano A & Tomas MB (1990) Thermic effect of a meal over 3 and 6 hours in lean and obese men. Metabolism 39, 985–992. Seidell JC (1997) What is the cost of obesity? International Journal of Obesity 21, Suppl. 2, S1. Shechter Y (1990) Insulin-mimetic effects of vanadate. Possible implications for future treatment of diabetes. Diabetes 39, 1–5. Sonko BJ, Prentice AM, Murgatroyd PR, Goldberg GR, van de Ven MLHM & Coward WA (1994) Effect of alcohol on postmeal fat storage. American Journal of Clinical Nutrition 59, 619–625. Stamler J, Vaccaro O, Neaton J & Wentworth D (1993) Diabetes, other risk factors and 12 years cardiovascular mortality for men screened in the multiple risk factors intervention trial. Diabetes Care 16, 434–441. Stock MJ (1989) New approaches to the control of obesity in animals and their clinical potential. In Nutrition in the Prevention of Disease, pp. 32–37 [JC Somogyi, editor]. Basel: Kärger. Storlien LH, Baur LA, Kriketors AD, Pan DA, Cooney GJ, Jenkins AB, Calvert GD & Campbell LV (1996) Dietary fats and insulin action. Diabetologia 39, 621–631. Sugano M, Watanabe S, Kishi A, Izume M & Ohtakara A (1988) Hypocholesterolemic action of chitosans with different viscosity in rats. Lipids 23, 187–191. Suter PM, Schutz Y & Jéquier E (1992) The effect of ethanol on fat storage in healthy subjects. New England Journal of Medicine 326, 983–987. Tappy L, Randin JP, Felber JP, Chiolero R, Simonson DC, Jéquier E & DeFronzo RA (1986) Comparison of the thermogenic effect of fructose and glucose in normal humans. American Journal of Physiology 250, E718–E724. The Diabetes Control and Complications Trial Research Group (1993) The effect of intensive treatment of diabetes on the development and progression of long term complications in insulin dependent diabetes mellitus. New England Journal of Medicine 329, 977–986. Thompson L, Button C & Jenkins D (1987) Phytic acid and calcium affect the in vitro rate of navy bean starch digestion and blood glucose response in humans. American Journal of Clinical Nutrition 46, 467. Thorburn A & Proietto J (1993) Carbohydrate fermentation decreases hepatic glucose output in healthy subjects. Metabolism 42, 780–785. Torún B & Viteri FE (1988) Protein–energy malnutrition. In Modern Nutrition in Health and Disease, pp. 746–773 [ME Shils and VR Young, editors]. Philadelphia, PA: Lea & Febiger. Tremblay A, Wouters E, Wenker M, St-Pierre S, Bouchard C & Després JP (1995) Alcohol and high fat diet: a combination favoring overfeeding. American Journal of Clinical Nutrition 62, 639–644. Truswell A (1994) Food carbohydrates and plasma lipids – an update. American Journal of Clinical Nutrition 59, Suppl. 9, 710S–718S. Truswell A, Seach J & Thorburn A (1988) Incomplete absorption of pure fructose in healthy subjects and the facilitating effect of glucose. American Journal of Clinical Nutrition 48, 1424–1430. Van der Beek EJ (1985) Vitamins and endurance training. Food for running or faddisch claims. Sports Medicine 2, 175–197. Van Erp-Baart AMJ, Saris WHM, Binkhorst RA, Vos JA & Elvers JWH (1989a) Nationwide survey on nutritional habits in elite Functional foods and substrate metabolism athletes. Part I. Energy, carbohydrate, protein, and fat intake. International Journal of Sports Medicine 10, S3–S10. Van Erp-Baart AMJ, Saris WHM, Binkhorst RA, Vos JA & Elvers JWH (1989b) Nationwide survey on nutritional habits in elite athletes. Part II: Mineral and vitamin intake. International Journal of Sports Medicine 10, S11–S16. Van Hall G, Raaymakers J, Saris WHM & Wagenmakers A (1995) Ingestion of branched-chain amino acids and tryptophan during sustained exercise in man: failure to affect performance. Journal of Physiology 486, 789–794. VanItallie TB (1992) Body weight, morbidity and longevity. In Obesity, pp. 361–369 [P Björntorp and BN Brodoff, editors]. Philadelphia, PA: J.B. Lippincott Company. Venter C, Vorster H & Cummings J (1990). Effect of dietary propionate on carbohydrate and lipid metabolism in healthy subjects. American Journal of Clinical Nutrition 85, 549–553. Vessby B (1995) Nutrition, lipids and diabetes mellitus. Current Opinion in Lipidology 6, 3–7. Vessby B, Tengblad S & Lithell H (1994) Insulin sensitivity is related to the fatty acid composition of serum lipids and skeletal muscle phospholipids in 70-year-old men. Diabetologia 37, 1044–1050. Virtanen SM & Aro A (1994) Dietary factors in the aetiology of diabetes. Annals of Medicine 26, 460–478. Wagenmakers AJM (1991) L-Carnitine supplementation and performance in man. In Advances in Nutrition and Top Sport, vol. 32, pp. 110–127 [F Brouns, WHM Saris and EA Newsholme, editors]. Basel: S. Karger. Wagenmakers AJM, Beckers EJ, Brouns F, Kuipers H, Soeters PB & van der Vusse GJ (1991) Carbohydrate supplementation, S75 glycogen depletion, and amino acid metabolism during exercise. American Journal of Physiology 260, E883–E890. Waterlow JC (1992) Some general questions. In Protein–Energy Malnutrition, pp. 14–25 [JC Waterlow, editor]. London: Edward Arnold. Westrate JA, Wunnink I, Deurenberg P & Hautvast JGAJ (1990) Alcohol and its acute effects on resting metabolic rate and diet induced thermogenesis. British Journal of Nutrition 64, 413– 425. White JR Jr & Campbell RK (1993) Magnesium and diabetes: a review. Annals of Pharmacotherapy 27, 775–780. World Health Organization Study Group (1994) Prevention of Diabetes Mellitus. Geneva: World Health Organization. Wursch P, Vedovo S & Koellreutter B (1986) Cell structure and starch nature as key determinants of the digestion rate of starch in legumes. American Journal of Clinical Nutrition 43, 25–29. Yamashita K, Kawai K & Itakura M (1984) Effects of fructooligosaccharides on blood glucose and serum lipids in diabetic subjects. Nutrition Research 4, 961–966. Young VR (1992) Macronutrient needs in the elderly. Nutrition Reviews 50, 454–462. Yudkin J (1988) Sucrose, coronary heart disease, diabetes and obesity: do hormones provide a link? American Heart Journal 115, 493–498. Zurlo F, Lillioja S, Esposito-Del Punete A, Nyomba BL, Raz I, Saad MF, Swinburn BA, Knowler WC, Bogardus C & Ravussin E (1990) Low ratio of fat to carbohydrate oxidation as predictor of weight gain: study of 24-h RQ. American Journal of Physiology 259, E650–E657. q Nutrition Society 1998 British Journal of Nutrition (1998), 80, Suppl. 1, S77–S112 S77 Functional food science and defence against reactive oxidative species A. T. Diplock1 *, J.-L. Charleux2 , G. Crozier-Willi 3 , F. J. Kok4 , C. Rice-Evans1 , M. Roberfroid5 , W. Stahl6 and J. Viña-Ribes7 1 International Antioxidant Research Centre, UMDS, Guy’s Hospital, St Thomas Street, London SE1 9RT, UK F. Hoffmann-La Roche Ltd, Business Unit Carotenoids, Headoffice Kaiseraugst VM1 Building, CH04002 Basel, Switzerland 3 Nestec Ltd, Nestlé Research Center, Vers-Chez-Les-Blanc, PO Box 44, CH-1000 Lausanne 26, Switzerland 4 Division of Human Nutrition and Epidemiology, Wageningen Agricultural University, PO Box 8129, NL 6700 EV, Wageningen, The Netherlands 5 UCL, Ecole de Pharmacie, Tour Van Helmont, Avenue E. Mounier, B-1200 Brussels, Belgium 6 Heinrich-Heine-University Düsseldorf, Medizinische Einrichtungen, Institut für Physiologische Chemie I, Postfach 10 10 07, D-40001 Düsseldorf, Germany 7 Universidad de Valencia, Facultad de Medicina, Departamento de Fisiologia, Avenida Blasco Ibafiez 17, E-46010 Valencia, Spain 2 Contents 1. Introduction 2. Oxidative damage, antioxidant defence and the role of prooxidants in disease 2.1. Antioxidant defence system of the human organism 2.1.1 Origins and nature of free radicals and other oxidants 2.1.2. Enzymic and non-enzymic defence systems in vivo 2.1.3. Dietary antioxidants; nutrient and nonnutrient 2.2. Oxidative damage to bodily functions and its implications in disease 2.2.1. Coronary heart disease 2.2.2. Carcinogenesis 2.2.3. Cataract and age-related macular degeneration 2.2.4. Neuronal diseases 2.3. Conclusions 3. Available methodologies for evaluating and quantifying ex vivo damage to DNA, lipids and proteins by prooxidants in vivo 3.1. Oxidative damage to DNA 3.1.1. Measurement of guanine damage products in DNA by HPLC and gas chromatography– mass spectrometry (GC–MS) 3.2. Oxidative damage to lipids 3.2.1 Lipid peroxidation 3.2.2. Can some measure of ‘total’ peroxidation be obtained? 3.2.3. LDL oxidation 3.3. Oxidative damage to proteins 3.4. Measurement of antioxidant nutrients, carotenoids and flavonoids extracted from human plasma S79 S80 S80 S80 S80 S80 S82 S82 S83 S83 S83 S84 S84 S84 S84 S85 S85 S86 S86 S87 3.5. Conclusions 4. Nutritional options modulating oxidative damage and antioxidant defence systems 4.1. Introduction 4.2. Dietary antioxidants 4.2.1. Sources of dietary antioxidants 4.2.2. Antioxidant intake and status 4.2.3. Bioavailability of antioxidants 4.2.4. Fat intake and antioxidant status 4.3. Epidemiological studies on protective effects of antioxidants 4.3.1. Cardiovascular disease 4.3.2. Cancer 4.3.3. Other age-related diseases 4.4. Human intervention studies of antioxidants 4.4.1. Cardiovascular disease 4.4.2. Cancer 4.5. Conclusions 5. Potential safety implications related to antioxidant nutritional enhancement 5.1. Introduction 5.2. Vitamin C 5.3. Vitamin E 5.4. Carotenoids 5.5. Non-nutrient antioxidants (flavonoids and other related compounds) 5.5.1. Absorption 5.5.2. Possible adverse effects 6. Role of food technology in nutritional and safety aspects of antioxidants 6.1. Introduction 6.2. Physical processes S88 S88 S88 S89 S89 S89 S90 S90 S90 S90 S91 S91 S92 S92 S92 S93 S93 S93 S93 S94 S96 S96 S96 S97 S97 S97 S98 S87 Abbreviations: AMD, age-related macular degeneration; ATBC study, a-tocopherol b-carotene study; FOX method, ferrous oxidation in xylenol orange method; GC–MS, gas chromatography–mass spectrometry; 8-OHdG, 8-hydroxy-deoxyguanosine; 8-OHG, 8-hydroxyguanosine; 8-oxodG, 8-oxo-7,8dihydro-20 -deoxyguanine; PG, prostaglandin; PUFA, polyunsaturated fatty acids; RDA, recommended daily allowance; RNS, reactive nitrogen species; ROS, reactive oxygen species; TBA, thiobarbituric acid; TBARS, TBA-reactive substances; a-TE, a-tocopherol equivalents. * Corresponding author: Professor A. T. Diplock, fax +44 (0) 171 403 7195, tel +44 (0) 171 955 4521. S78 A. T. Diplock et al. 6.2.1. Structural integrity S98 6.2.2. Moisture content S98 6.2.3. Temperature S98 6.2.4. Minimizing oxygen S99 6.2.5. Protection from light S99 6.2.6. Irradiation S99 6.3. Chemical processes S99 6.3.1. Enzymes S99 6.3.2. Supplementation S99 6.4. Conclusions S100 7. Critical assessment of the science base and conclusions S100 7.1. Identification of criteria S100 7.2. Critical evaluation of the present knowledge base S101 7.2.1. Conclusions from section 2 S101 7.2.2. Conclusions from section 3 S101 7.2.3. Conclusions from section 4 S102 7.2.4. Conclusions from section 5 7.2.5. Conclusions from section 6 7.3. Evaluation of criteria 7.4. Final conclusions 8. Recommendations for future research 8.1. Introduction 8.2. Specific recommendations 8.2.1. Oxidative damage and antioxidant defence systems of the human organism 8.2.2. Ex vivo methodologies for quantitating and validating damage in vivo to biological macromolecules 8.2.3. Nutritional options modulating oxidative damage 8.2.4. Safety implications of nutritional enhancement of antioxidants 8.3. Priorities for the recommendations made Abstract This paper assesses critically the science base that underpins the argument that oxidative damage is a significant causative factor in the development of human diseases and that antioxidants are capable of preventing or ameliorating these disease processes. The assessment has been carried out under a number of headings, and some recommendations for future research are made based on the present day knowledge base. The knowledge database (1) Consideration of the basic science that underlies understanding of the role of free radicals in causing cellular pathologies, and the role of antioxidants in preventing this, shows that an imbalance of reactive oxygen species and antioxidant defence systems may lead to chemical modifications of biologically relevant macromolecules. This imbalance provides a logical pathobiochemical mechanism for the initiation and development of several disease states. Experimental data obtained in vivo provide evidence that antioxidants function in systems that scavenge reactive oxygen species and that these are relevant to what occurs in vivo. The relevance in vivo of these observations depends inter alia on knowledge of the uptake and distribution of the antioxidant within the human body, and on what tissue levels of the antioxidant may be expected in relation to dietary levels. (2) There is some way to go until validated precise methods are available for measuring biomarkers of oxidative damage in human subjects in vivo under minimally invasive conditions. With respect to oxidative damage to DNA, HPLC and GC–mass spectrophotometry methods have both merits and limitations. Lipid oxidation products in plasma are best measured as isoprostanes or as lipid hydroperoxides using specific HPLC techniques. Development of isoprostane measurement will advance specificity and precision. The measurement of oxidative damage to proteins has some potential but such methods have not been effectively exploited. (3) Epidemiological studies support the hypothesis that the major antioxidant nutrients vitamin E and vitamin C, and b-carotene (which may or may not be acting as an antioxidant in vivo), may play a beneficial role in prevention of several chronic disorders. More research is needed on the impact of other non-nutrient compounds, such as other carotenoids and flavonoids, on human health. In general, human intervention studies using hard end-points are the gold standard. Trials are restricted mainly to the major antioxidants and do not allow firm conclusions because of inconsistent findings, an insufficient number of studies and the use of varying doses. There is evidence that large doses of b-carotene may be deleterious to the health of certain subgroups of the population such as heavy habitual smokers. (4) With respect to the safety of administration of supplementary vitamins, vitamin C is safe at levels of supplementation up to 600 mg/d, and higher levels, up to 2000 mg/d, are without risk. Vitamin E has a very low human toxicity and an intake of 1000 mg/d is without risk; 3200 mg/d has been shown to be without any consistent risk. Large intakes of b-carotene must be viewed with caution because they have been shown to confer detriment to a population at high risk of lung cancer when administered after many years of high risk (smoking) behaviour. Until further work clarifies the situation in heavy smokers with respect to taking supplements, larger doses should be avoided by such individuals. There is little reliable information about the human toxicology of flavonoids and related non-nutrient antioxidant constituents of the diet. S102 S102 S102 S103 S104 S104 S104 S104 S104 S104 S105 S105 Defence against reactive oxidative species S79 (5) The food industry has long experience in the control of oxidative damage in foods and this experience can be used to advantage for the protection of food antioxidants which are beneficial. Some of these, such as vitamins C and E and b-carotene, are well known, and strategies for their protection in foods are already exploited by food technologies. Food technology strategies for the preservation of those antioxidants which have been shown to be beneficial to health can be applied in a cost-effective manner. Research needs (1) The review of the available scientific database enables the identification of areas where further research is required. Improvement in dietary antioxidant intake in human populations is expected to result in lowering of the risk of a number of degenerative diseases. While desirable as an ultimate objective per se, the impact on public health and the resultant decrease in health-care costs make it imperative that substantial sums of money should be spent on research in this important area. (2) Direct measurement of prooxidants in vivo is difficult or impossible. It is imperative to establish which are the critical free radical ‘hits’ that are the relevant ones in the aetiology of diseases. Are the processes examined really relevant to the disease causation? There is a need to identify which are the important antioxidants in terms of the maintenance of health, and what is their relationship to one another. Clarification is needed as to whether it is the antioxidant role of the substance that is important, or whether it is some other function, and the possible nonantioxidant effects of antioxidants, in particular with respect to modulation of gene expression, also need further research. The primary aim is to identify the active components in the overall system that promote health. (3) Ex vivo methodologies for quantitating and validating damage in vivo to biological macromolecules urgently need further attention before meaningful work can proceed on providing evidence of the level of antioxidants needed to maintain health and well-being. It is necessary to refine and validate methods that are already available for measurement of oxidative damage in human subjects in a non-invasive manner. Validation in many centres by measurement of oxidative damage in the same biological material needs to be undertaken using the same methodology. There is a somewhat longer term need for development of techniques to be used ex vivo as measures of protein oxidation in vivo. (4) Nutritional options modulating oxidative damage. For proper epidemiological research, as well as for human intervention studies, it would be desirable to put special emphasis on the following. (a) Chemical analysis of the antioxidant content of foods. (b) Studies of bioavailability of antioxidants from the diet, and the factors that influence the absorption, distribution and tissue uptake of the compounds and the likely impact of the antioxidants on metabolic processes. (c) Development and validation of biomarkers of intermediate end-points, both biological response markers and early disease markers, and emphasis on the relevance of the biomarker to the disease end-point as well as the disease process. (d) Application of the validated biomarkers of intermediate end-points in randomized controlled trials testing the efficacy of antioxidants in functional foods for the maintenance of health and well-being. (5) Safety implications of nutritional enhancement of antioxidants. The detailed evidence that is already available which demonstrates that vitamin C and vitamin E are safe at quite high levels of inclusion in the human diet, means that is unnecessary to recommend further work in this area. The safety of b-carotene was not questioned before the results of the Finnish and American intervention studies, which showed an apparent exacerbation in the incidence of lung cancer in heavy smokers who were given supplements of b-carotene. This observation needs urgent clarification. With respect to the flavonoids and other polyphenols, it is likely that their bioactivity will be explored, and the key question of their bioavailability clarified, in the near future. It will be necessary to examine the safety of such bioflavonoid compounds. (6) With regard to the role of food technology, there is no particular direction that research needs to take at the present time. Developments in food technology will be based on, and adapt to, nutritional recommendations resulting from biologically driven work. Prooxidants: Antioxidants: Oxidative damage: Diet and health ‘Good health is more than the mere absence of disease’. Mark Twain 1. Introduction This paper presents a comprehensive assessment of the literature to the end of December 1997 relevant to the role and importance of dietary antioxidants in human health. Section 2 reviews the basic science concerning free-radical damage and the ameliorating role at the cellular level of antioxidants, and introduces the concept of maintenance of health, and the prevention of some major human degenerative diseases. In order to assess the extent of free-radical damage in human subjects in vivo, and the modifying effects of antioxidants, it is necessary to have valid, precise biomarkers. Present literature on this topic is reviewed in S80 A. T. Diplock et al. section 3. An account is given in section 4 of the human epidemiological and interventional evidence that links a high intake of dietary antioxidants with a low risk of degenerative disease. If it should be proven with reasonable certainty that antioxidants do indeed lower the risk of human degenerative disease, it is essential to be sure that intervention with antioxidants in the diet, by fortification or supplementation of foods, is entirely free from harmful sideeffects; this important topic is addressed in section 5, and the possible contribution of the food industry and food technologists to this enhancement is discussed in section 6. The 7th and 8th sections assess the ‘state of the art’ with respect to the foregoing sections and make recommendations for research in the immediate future. 2. Oxidative damage, antioxidant defence and the role of prooxidants in disease 2.1. Antioxidant defence system of the human organism 2.1.1. Origins and nature of free-radicals and other oxidants. The development and existence of an organism in the presence of O2 is associated with the generation of reactive oxygen species (ROS), even under physiological conditions. ROS are responsible for the oxidative damage of biological macromolecules such as DNA, carbohydrates and proteins (Halliwell & Gutteridge, 1989; Sies, 1991; Halliwell, 1996). These processes are discussed as pathobiochemical mechanisms involved in the initiation or progression phase of various diseases (Diplock, 1994; Wiseman & Halliwell, 1996). Some of the most relevant ROS are: peroxyl radicals (ROO? ), the nitric oxide radical (NO? ), the superoxide anion radical 1 ¹ (O?¹ 2 ), singlet oxygen ( O2 ), peroxynitrite (ONOO ), and hydrogen peroxide (H2 O2 ). ROS are either radicals (molecules that contain at least one unpaired electron) or reactive nonradical compounds, capable of oxidizing biomolecules. Therefore, these intermediates are also called oxidants or prooxidants (Halliwell & Gutteridge, 1989; Sies, 1991). There are various sources for specific ROS in the human organism. However, the superoxide radical anion appears to play a central role, since other reactive intermediates are formed in reaction sequences starting with O2?¹ . It is generated by enzymic one-electron reduction of O from xanthine oxidase (EC 1.2.3.2), NADPH oxidase, or by leakage of the respiratory chain. It has been estimated that about 1–3 % of the O2 we utilize is converted to O?¹ 2 (Fridovich, 1986; Halliwell, 1996). H2 O2 is a non-radical reactive species and can easily diffuse between living cells. It is efficiently converted to water by the enzyme catalase (EC 1.11.1.6), a process which determines its half-life. Recent evidence suggests that H2 O2 is involved in signal transduction regulating the expression of genes through the nuclear factor kB and apoprotein-1 pathways (Schreck & Baeuerle, 1994; Sen & Packer, 1996). The most reactive species is the hydroxyl radical with an estimated half-life of about 10¹9 s. It might be formed in vivo on high-energy irradiation (e.g. X-rays) by homolytic cleavage of body water or from endogenous H2 O2 in metal-catalysed processes (Fenton reaction: Fe-catalysed Haber–Weiss reaction). u.v.-Light is insufficiently energetic to split water but it can cleave H2 O2 to yield two molecules of the hydroxyl radical. The high reactivity of this radical implies immediate reaction at the place where it is generated. The peroxyl radical (ROO? ) is relatively long lived (seconds) with a considerable diffusion pathlength in biological systems. It can be generated in the process of lipid peroxidation which is initiated by the abstraction of an H atom from polyunsaturated fatty acids (PUFA); the hydroxyl radical is capable of starting this reaction sequence (Esterbauer et al. 1992; Reaven & Witzum, 1996). Further products generated in lipid peroxidation are alkoxyl radicals (RO? ) and organic hydroperoxides (ROOH). The latter might rearrange to endoperoxide intermediates which are cleaved to yield aldehydes. The reaction of aldehydes with amine groups of proteins has been discussed as a mechanism involved in the modification of the protein part of lipoproteins. Singlet molecular oxygen (1 O2 ) is another non-radical ROS which is suggested to be formed in vivo in lightexposed tissue. Its half-life has been estimated to be 10¹6 s depending on the nature of the surrounding matrix. 1 O2 can interact with other molecules either by transferring its excitation energy or by combining chemically. Preferential targets for chemical reactions are double bonds; e.g. in PUFA or guanine in DNA bases (Kanofsky, 1989; Stahl & Sies, 1993; Cadet et al. 1994). An interesting ROS which has attracted attention within the past few years is the nitric oxide radical (NO? ). It is a signalling compound formed enzymically from arginine and relaxes smooth muscles in blood-vessel walls resulting in lowered blood pressure. It is also produced by activated macrophages contributing to the primary immune defence. An excess of NO? is cytotoxic. It might react directly with biomolecules or combine with O?¹ 2 to form peroxynitrite (ONOO¹ ). Peroxynitrite is capable of inducing lipid peroxidation in lipoproteins but might also interfere with cellular signalling by nitrating tyrosine residues in proteins (Beckman, 1996; Packer, 1996). The ROS described here, and also the biological pathways for their endogenous formation, are examples of a whole class of reactive intermediates and their ways of generation. It should further be noted that the organism is also exposed to ROS from external sources. With the diet many compounds of prooxidant nature, such as quinones capable of redox cycling, are delivered to the organism. Also an array of radicals are inhaled with cigarette smoke; ozone, of which increasing levels are reported due to air pollution, is an ROS which can oxidize lipids (Pryor et al. 1995). ROS are also produced in the organism as a part of the primary immune defence. Phagocytic cells such as neutrophils, monocytes, or macrophages defend against foreign organisms by synthesizing large amounts of O2?¹ or NO as a part of their killing mechanism. Several diseases are accompanied by excessive phagocyte activation resulting in tissue damage which is at least in part due to the activity of ROS. 2.1.2. Enzymic and non-enzymic defence systems in vivo. To counteract the prooxidant load a diversity of antioxidant defence systems are operative in biological systems including enzymic and non-enzymic antioxidants. An antioxidant has been defined as ‘any substance that, when present in low concentrations compared to that of an Defence against reactive oxidative species oxidizable substrate, significantly delays or inhibits the oxidation of that substrate’ (Halliwell & Gutteridge, 1989; Sies, 1993; Halliwell, 1995). The major enzymes directly involved in the detoxification of ROS are superoxide dismutase (EC 1.15.1.1), scavenging O?¹ 2 , as well as catalase and glutathione peroxidases (EC 1.11.1.9) which reduce H2 O2 and organic hydroperoxides respectively. Several subtypes of glutathione peroxidase are Se-dependent. In animal studies an elevated intake of Se was associated with protective effects against cancer. Its preventive effects in man are still under investigation (Levander & Burk, 1996). Indirect antioxidant functions are mediated by enzymes that restore endogenous antioxidant levels; e.g. GSH levels are replenished on reduction of GSSG by glutathione reductase (EC 1.6.4.1). Further, reactive intermediates produced in reactions of prooxidants and biological molecules (e.g. epoxides) are conjugated by phase II detoxification enzymes such as glutathione-Stransferases (EC 2.5.1.18) to favour their excretion. Another strategy to prevent the formation of ROS is the control of the levels of free Fe or Cu ions. Metal-binding proteins responsible for the transport of these ions bind them tightly, thus preventing the initiation of lipid peroxidation or DNA damage. Some of the most relevant metal-binding proteins are ferritin, transferrin, and caeruloplasmin. Various endogenous low-molecular-mass compounds are also involved in antioxidant defence. GSH, the major cytosolic thiol, serves as a cofactor for several detoxifying enzymes (glutathione peroxidases, glutathione-S-transferases), is involved in the reduction of protein disulfides and additionally scavenges ROS, being oxidized to GSSG. Other endogenous compounds such as ubiquinol-10, urate, or bilirubin also exhibit antioxidant activities (Jacob & Burri, 1996). 2.1.3. Dietary antioxidants; nutrient and nonnutrient. The human diet contains an array of different compounds that possess antioxidant activities or have been suggested to scavenge ROS based on their structural properties. The most prominent representatives of dietary antioxidants are ascorbate (vitamin C), tocopherols (vitamin E), carotenoids, and flavonoids. Apart from vitamin C, each group of these antioxidants consists of a number of structurally different compounds; e.g. more than 600 different carotenoids have been identified to date and about fifty of them might occur in the human diet (Sies & Stahl, 1995; Rice-Evans & Miller, 1996; Rock et al. 1996). In the diet, there may be synergistic effects of these various dietary compounds which are difficult to assess at present. Indeed, the diet may be considered as an orchestra where interactions between constituents may bring about effects which are not the necessary properties of the individual constituents. Vitamin C is considered to be one of the most powerful, least toxic natural antioxidants (Bendich et al. 1986; Weber et al. 1996). It is water-soluble and is found in high concentrations in many tissues; human plasma contains about 60 mmol ascorbate/l. On interaction with ROS it is oxidized to dehydro-ascorbate via the intermediate ascorbyl free radical. Dehydro-ascorbate is recycled back to ascorbic acid by the enzyme dehydro-ascorbate reductase. Thus, dehydro-ascorbate is found in only very low levels S81 compared with ascorbate. As a scavenger of ROS ascorbate has been shown to be effective against the superoxide radical anion, H2 O2 , the hydroxyl radical, and singlet oxygen. In aqueous solutions vitamin C also scavenges reactive nitrogen oxide species efficiently, preventing the nitrosation of target molecules. The major sources of ascorbate in the diet are fruits, especially citrus fruits, kiwi fruit, cherries and melons, and vegetables such as tomatoes, leafy greens, broccoli, cauliflower, Brussels sprouts, and cabbage; its content might exceed 100 mg ascorbate/100 g fresh weight. At low dose levels (100 mg) the bioavailability values for vitamin C from synthetic and food sources are very similar (Mangels et al. 1993a); the efficacy of absorption decreases with increasing dose levels (Levine et al. 1996). There is evidence from studies in vitro that vitamin C is capable of regenerating tocopherol from the tocopheroxyl radical which is formed on inhibition of lipid peroxidation by vitamin E (Niki et al. 1982, 1985). This process would allow for the transport of a radical load from a lipophilic compartment to an aqueous compartment where it is taken care of by efficient enzymic defence systems. It should be noted, however, that ascorbate might also act as a prooxidant in vivo. In the presence of free transition metal ions (Fe and Cu) and ascorbate the hydroxyl radical can be generated and initiation of lipid peroxidation may occur. However, the amounts of free transition metals in vivo are very small because they are efficiently bound to proteins. Vitamin C has additional well-established biological functions including cofactor-activity for several important enzymes (Levine et al. 1996). The term vitamin E is a generic description for all tocols and tocotrienol derivatives which exhibit the biological activity of a-tocopherol (Parker, 1989; Eldin & Appelqvist, 1996; Sokol, 1996; Traber & Sies, 1996). This group of compounds is highly lipophilic, and operative in membranes or lipoproteins. Their most important antioxidant function appears to be the inhibition of lipid peroxidation, scavenging lipid peroxyl radicals to yield lipid hydroperoxides and a tocopheroxyl radical. The latter is less reactive than peroxyl radicals towards neighbouring PUFA and acts as a chain-breaking antioxidant. The tocopheroxyl radical might be either reduced by ascorbate and GSH or further oxidized to the respective quinone. Since only small amounts of tocopheryl quinone are detectable in human blood and tissues, the regenerative pathway in vivo appears to be favoured. In comparison with other lipophilic antioxidants, a-tocopherol is probably the most efficient in the lipid phase (Niki, 1987). It contains shielding methyl groups adjacent to the phenolic hydroxyl group and it is optimally positioned in membranes by its phytyl side-chain, which is located in the hydrophobic region of the membrane structure. In addition to its peroxyl-radical scavenging properties, further interactions with ROS have been described, including quenching of singlet oxygen and interaction with peroxynitrite. The richest sources of vitamin E in the diet are vegetable oils (soyabean, maize, cottonseed, and safflowerseed), and products made from these oils such as margarine and mayonnaise. Further, wheat germ, nuts, and some green leafy vegetables contribute considerable amounts to the vitamin E supply (Parker, 1989). Vitamin E plasma levels in man are about 22 mmol/l; the compound is also found in tissues such S82 A. T. Diplock et al. as liver, kidney, fat and adrenals. In the liver the RRRisomer of a-tocopherol is preferentially incorporated into VLDL which are further catabolized in the circulation. Thus, RRR-a-tocopherol is the major form of vitamin E in LDL (Traber & Sies, 1996). Carotenoids are natural colourants with pronounced antioxidant activity (Stahl & Sies, 1993; Olson & Krinsky, 1995). Their chemical properties are closely related to the presence of an extended system of conjugated double bonds which is substituted with various endgroups. ROS which are efficiently scavenged by carotenoids are 1 O2 and peroxyl radicals (Palozza & Krinsky, 1992). Two different pathways are operative with respect to the deactivation of 1 O2 : physical and chemical quenching (Truscott, 1990). Physical quenching implies the deactivation of 1 O2 by energy transfer from the excited oxygen species to the carotenoid, yielding a triplet excited carotenoid. The energy of the excited carotenoid is dissipated through vibrational interactions with the solvent to recover ground state carotenoid. The carotenoid remains intact in this process and might undergo further cycles of deactivation. Chemical quenching contributes less than 0.05 % to total 1 O2 -quenching by carotenoids but is responsible for the eventual destruction of the molecule. Carotenoids are the most efficient naturally occurring quenchers for 1 O2 with quenching rate constants of about 5–12 × 109 /mol per s. Carotenoids were reported to scavenge peroxyl radicals by chemical interaction (Kennedy & Liebler, 1992). It is suggested that carotene radical intermediates are formed in this process which finally leads to the destruction of the molecule. Like vitamin E, carotenoids belong to the group of lipophilic antioxidants present in lipoproteins such as LDL and HDL. It has been shown that they are consumed when isolated LDL is exposed to the process of lipid peroxidation. Their contribution to the antioxidant defence system of LDL is not clear, since no regeneration pathways for oxidized carotenoids are known at present. A variety of structurally different carotenoids are present in fruits and vegetables. Some of the major sources are carrots (a-carotene, b-carotene), tomatoes (lycopene), citrus fruits (b-cryptoxanthin), spinach (lutein), and maize (zeaxanthin) (Mangels et al. 1993b). The absorption and transport processes of carotenoids are quite complex. Several factors influencing carotenoid bioavailability from food, such as co-ingestion of fat or fibre, cooking or food processing, have been identified (Erdman et al. 1993). Flavonoids are a large group of polyphenolic antioxidants that occur in several fruits, vegetables, and beverages such as tea, wine and beer mainly as O-glycosides. They are efficient antioxidants capable of scavenging radical species (peroxyl radicals, hydroxyl radical, O?¹ 2 ) forming a phenoxy radical (Rice-Evans et al. 1995; Rice-Evans & Miller, 1996). The term flavonoids summarizes a number of structurally different subgroups including flavanols (catechin, epicatechin), flavonols (quercetin, myricetin, kaempherol), flavanones (naringenin, taxifolin), flavones (apigenin, hesperetin), isoflavones (genestein), or anthocyanidins (cyanidin, malvidin). Several criteria for optimal radical scavenging properties of flavonoids have been postulated based on pulse radiolysis studies. These include the presence of the 30 ,40 -dihydroxy structure in ring B, the presence of the 2,3-double bond in conjugation with the 4oxo-group in ring C, and the presence of a 5-hydroxyl group in ring A with a 3-hydroxyl group and a 4-oxo function in the C-ring. The antioxidant properties of flavonoids have been investigated in various studies in vivo and in vitro. It should be mentioned, however, that the bioavailability of these compounds is rather poor. They are rapidly conjugated in phase II detoxification reactions and levels of free flavonoids in human plasma are very low. Further phenolic compounds with antioxidant activity are derivatives of cinnamic acid; e.g. caffeic acid, chlorogenic acid, and ferulic acid (Rice-Evans & Miller, 1996). In addition to the flavonoids, a number of other phenolic compounds of potential interest occur in foods. Thus, olive oil contains a number of phenolic substances, notably the odiphenol tyrosol, which may contribute to the antioxidant content of diets rich in olive oil (Kiritsakis, 1990). Similarly, plants of the Lamiaceae family, notably rosemary, oregano, sage, mint and thyme, contain a range of potential antioxidants such as carnosol, rosemanol and carvacrol, which can contribute to the antioxidant potential of the diet (Lagouri & Boskou, 1996). As with the flavonoids, however, little is known of the human absorption and tissue distribution of these compounds. Several other dietary constituents might also be involved in the antioxidant defence system either by direct action as antioxidants or by effects related to the induction of detoxifying enzymes. Enzymes such as glutathione peroxidase and superoxide dismutase, which require a dietary supply of Se, and of Cu and Zn respectively, contribute to the overall oxidative defence mechanism. Some endogenous substances such as urate also add to the antioxidant potential of living cells, although their significance is only speculative. Enhancement of dietary intake of the minerals identified may be beneficial when their content in the diet is low. 2.2. Oxidative damage to bodily functions and its implications in disease ROS are suggested, or known, to be involved in pathogenic processes of numerous diseases (Esterbauer et al. 1992; Luis & Navab, 1993; Diplock, 1994; Sies, 1997) such as cardiovascular disease, some forms of cancer, cataract, age-related macular degeneration, rheumatoid arthritis and a number of neurodegenerative diseases. Oxidative damage to important biomolecules is a deleterious pathway, but also influences of ROS on gene regulation or the immune system might impair bodily functions. There is increasing evidence from clinical and intervention studies, as well as from basic research that antioxidants might prevent or delay the development of disease states. There may also be particular population groups that will benefit from enhanced antioxidant intake, such as pregnant women, neonates and children, senior citizens and, perhaps, sportspeople. 2.2.1. Coronary heart disease. The primary cause for most cardiovascular diseases is thought to be arteriosclerosis, a multifactorial disease of the artery wall. It is suggested that in the early stages of arteriosclerosis lipid deposits, so-called fatty streaks, are formed in the subendothelial space. There is increasing evidence that Defence against reactive oxidative species oxidative stress, particularly oxidation of LDL, is a risk factor and plays a role in the pathogenic pathway (Berliner & Heinecke, 1996). LDL oxidation is due to a lipid peroxidation reaction initiated by free radicals. Separate investigations of the lipid and protein parts of oxidized LDL demonstrated that oxidative modifications of both contribute to the proatherogenic properties of oxidized LDL. Several biochemical mechanisms underlying this effect have been discussed. These include the formation of foam cells on the uptake of oxidized LDL via the scavenger receptor by macrophages resident in the subendothelial area, release of cytotoxic lipid peroxidation products from oxidized LDL, or chemoattractant properties of the oxidized lipoprotein. LDL oxidation is efficiently inhibited by lipophilic antioxidants of which a-tocopherol appears to be the most important. Epidemiological studies suggest preventive effects towards atherogenic lesions to be associated with an increased uptake of lipophilic antioxidants such as vitamin E or carotenoids (Rimm et al. 1993). Additional effects of RRR-a-tocopherol, independent of its antioxidant activity, have been related to the protective properties of this compound. An early event in the onset of arteriosclerosis is the migration of smooth-muscle cells from the media to the intima of the arterial wall followed by proliferation of these cells. There is increasing evidence that RRR-a-tocopherol acts as a negative regulator of smooth-muscle cell proliferation via modulation of protein kinase C activity. Protein kinase C is an important element in the signal transduction cascade mediated by growth factors such as platelet-derived growth factor which are involved in the control of cell proliferation. It should be noted that these effects are limited to RRR-a-tocopherol; RRR-b-tocopherol does not inhibit protein kinase C (Azzi et al. 1995; Özer et al. 1995). 2.2.2. Carcinogenesis. Carcinogenesis is a complex multistep process including initiation, promotion and progression. The generation of ROS is thought to be linked to tumourigenesis at different levels. Oxidative damage to DNA has been demonstrated in vitro and in vivo leading to DNA single or double strand breaks and DNA cross linking, as well as to chromosomal aberrations such as breakage or rearrangement. Modified DNA bases (e.g. hydroxythymidine, or hydroxyguanine) have been determined after exposure of cells to situations of oxidative stress. The modification of DNA bases might result in point mutations, deletions, or gene amplification as a first step of carcinogenesis. Further, ROS are capable of deactivating detoxifying enzymes responsible for the scavenging of potent carcinogens. Data from epidemiological studies support the idea that antioxidants are preventive in carcinogenesis by scavenging ROS (Flagg et al. 1995). Carotenoids exhibit further biological functions which are not related to their antioxidant activities but might be of importance with respect to their cancer preventive effects (Gerster, 1995). It has been shown that provitamin A and non-provitamin A carotenoids are capable of inhibiting the growth of transformed fibroblasts (Bertram & Bortkiewicz, 1995). There is increasing evidence that growth arrest is due to the stimulation of gap-junctional communication between transformed and surrounding normal cells. These findings suggest that carotenoids or carotenoid-derived S83 retinoids play a role in intercellular signalling involved in growth control. Inhibitory effects of b-carotene and lycopene on cell proliferation have also been described for several human cancer cell lines (Sharoni & Levy, 1996). 2.2.3. Cataract and age-related macular degeneration. Oxidative damage and impaired vision have been discussed in the context of two ophthalmological diseases of the elderly, cataract and age-related macular degeneration (AMD) (Schalch, 1992; Taylor, 1993). Senile cataract indicates the opacity of ocular lenses. Lens proteins are extremely longlived and often show oxidative damage. This is not surprising, since they are subjected to chronic exposure to light and O2 , which is likely to be responsible for the formation of ROS which might react with lens proteins. As a consequence, the damaged proteins may aggregate and precipitate, thus losing their regular function. Supplementation studies support the hypothesis that a higher intake of vitamins including vitamin C and vitamin E prevents or delays the development of cataracts (Seddon et al. 1994b). AMD is the major cause of visual impairment in Western countries and affects the anatomical region of the retina with the highest degree of visual activity. The macular pigment (yellow spot) represents a colour filter through which light must pass before detection. The carotenoids lutein and zeaxanthin are the predominant pigments in this area (Landrum, 1997). Carotenes (hydrocarbon carotenoids) are not present in the yellow spot. The function of the macular pigment has not been unequivocally identified but it might protect against photo-oxidation by blue light, mediated by excited triplet state molecules, 1 O2 , or superoxide. There are hints from food-frequency questionnaires that an increased consumption of food rich in lutein and zeaxanthin is associated with a diminished risk of AMD (Seddon et al. 1994a). Carotenoids are the most efficient natural compounds scavenging 1 O2 and excited triplet state molecules. 2.2.4. Neuronal diseases. Growing data from experimental models and human brain studies add evidence that oxidative stress might play a role in the development of neuronal degeneration related to diseases such as Parkinson’s disease, amyotrophic lateral sclerosis, and Alzheimer’s disease (Kondo, 1996; Simonian & Coyle, 1996). ROS are capable of inducing both necrosis and apoptosis. As a consequence of lipid peroxidation membrane rupture might occur or ion gradients, operative over compartments which are separated by membranes, might be disturbed. Neurons might undergo necrotic cell death as has been demonstrated in cell culture following depletion of intracellular GSH, the major endogenous antioxidant thiol. NO has been hypothesized to be an important mediator of neuronal death under pathological conditions. The ultimate species responsible for NO toxicity may be peroxynitrite which is formed by the reaction of the NO-radical with the superoxide radical. Beyond the classical aspects of oxidative damage to biologically relevant molecules as pathological mechanisms underlying several diseases, and the protective effects of antioxidants, new fields of research in this area are rapidly developing. This includes effects of prooxidants and antioxidants on immune functions (Bendich, 1990) and antioxidant and redox regulatory properties on gene expression. Both mechanisms may be involved in the development of S84 A. T. Diplock et al. disease states while protection might be provided by antioxidants via these pathways. 2.3. Conclusions An imbalance between ROS and antioxidant defence systems may lead to chemical modifications of biologically relevant macromolecules like DNA, proteins or lipids which are possible pathobiochemical mechanisms in the initiation or development of several disease states. Experimental data provide evidence that dietary antioxidants scavenge ROS and are useful in the prevention of these diseases. Epidemiological studies clearly show a correlation between the increased consumption of food rich in antioxidants and a decreased risk of several diseases. Thus, an increased intake of fruits and vegetables can be recommended. Data on antioxidant supplementation are contradictory. Further research is necessary to establish whether supplementation beyond dietary intake levels is of benefit. 3. Available methodologies for evaluating and quantifying ex vivo damage to DNA, lipids and proteins by prooxidants in vivo 3.1. Oxidative damage to DNA The most abundant base alteration induced in DNA by ROS is the formation of 8-oxo-7,8-dihydro-20 -deoxyguanine (8oxodG). In vivo this DNA base alteration is repaired by excision and the resulting product 8-oxodG is excreted unchanged, and independently of diet, into the urine. Thus, the rate of excretion of 8-oxodG (as given by the appearance of the metabolite in urine with time) serves as a biological marker of the integrated rate of oxidative DNA damage in the whole body. DNA damage is usually measured in lymphocytes isolated from blood or in urine. Baseline levels of DNA damage are considered to be important because repair may be incomplete (one damaged base per 106 bases); the actual measurement may therefore provide an estimate of the balance between damage and repair, so the time window is a crucial consideration here. From studies in vitro it is known that when DNA is exposed to the activated oxygen species the product is specific to the oxygen species involved, thus: 1 : O2 ; ROO? → guanine oxidized OH → multiplicity of changes to all four bases O?¹ 2 ; H2 O2 → no base changes ONOO → xanthine, hypoxanthine, 8-nitroguanine. ¹ There are two types of measurement of oxidative DNA damage. First, steady-state damage can be measured when DNA is isolated from human cells and tissues and analysed for base damage products: it presumably reflects the balance between damage and DNA repair. Hence a rise in steadystate oxidative DNA damage (e.g. as has been reported in some human cancerous tissues; Malins & Haimonot, 1991; Olinski et al. 1992) could be due to increased damage and/or decreased repair. Second, several DNA base damage products are excreted in human urine, including the nucleoside 8-hydroxy-deoxyguanosine (8-OHdG), 8-hydroxy-adenine and 7-methyl-8-hydroxyguanine (Ames, 1989; Stillwell et al. 1989) but the one most exploited is 8-OHdG, usually measured by a method involving HPLC with electrochemical detection (Ames, 1989; Shigenaga et al. 1994). The validity of these urinary measurements of oxidative DNA damage must be considered. The level of 8-OHdG in urine is presumably unaffected by the diet since nucleosides are not absorbed from the gut. The question of whether any 8-OHdG is metabolized to other products in man has not been rigorously addressed. Additionally, it is possible that some or all of the 8-OHdG excreted in urine may arise not from DNA, but from deoxyGTP in the DNA precursor pool of nucleotides. An enzyme has been described which hydrolyses deoxyGTP containing oxidized guanine to prevent its incorporation into DNA (Mo et al. 1992; Sakumi et al. 1993). These uncertainties require clarification. 3.1.1. Measurement of guanine damage products in DNA by HPLC and gas chromatography–mass spectrometry (GC–MS). As mentioned earlier, 8-hydroxyguanine (8OHG) and 8-OHdG are the products most frequently used as indicators of oxidative DNA damage. Analysis of 8-OHdG using HPLC coupled to electrochemical detection (Floyd et al. 1986), is a highly sensitive technique that is frequently used after release of 8-OHdG from DNA, usually by enzymic hydrolysis. GC–MS with selective ion monitoring has also been used to characterize oxidative DNA base damage by the identification of a spectrum of products (Dizdaroglu, 1993a), including 8-OHG, after formic acid hydrolysis of DNA and derivatization (often by trimethylsilylation) to generate volatile products. When GC–MS is used to measure modified DNA bases, a quantitative analysis of these bases in a DNA sample can be achieved by adding a suitable internal standard to that sample at an early stage of the analysis, such as before the hydrolysis of the DNA (Dizdaroglu, 1993b). Stable-isotope-labelled analogues of the modified bases can also be used as internal standards (Dizdaroglu, 1993b). One advantage of the GC–MS approach is that measurement of a wide range of base damage products allows more accurate quantification of DNA damage and can help to identify the ROS and reactive nitrogen species (RNS) that caused the damage (Malins & Haimonot, 1991); O2 selectively attacks guanine whereas OH? attacks all four DNA bases. However, the levels of 8-OHdG measured in DNA by HPLC with electrochemical detection are often (Halliwell & Dizdaroglu, 1992) (but not always; Lunec et al. 1994; Herbert et al. 1996) less than the levels of 8-OHG measured by GC–MS with selective ion monitoring. HPLC could underestimate the real amount of 8OHdG in DNA if the enzymic hydrolysis was incomplete; the action of the exonucleases and endonucleases used to hydrolyse the DNA may be affected by the modification of the bases (Halliwell & Dizdaroglu, 1992; Turk & Weitzman, 1995) and the acid pH often used for nuclease digestions might induce hydrolysis of 8-OHdG to 8-OHG, resulting in the loss of HPLC-detectable material. In contrast, GC–MS might overestimate 8-OHG (and perhaps other base damage products) as a result of their artifactual formation during the heating step involved in classical silylation-based derivatization procedures (Halliwell & Defence against reactive oxidative species Dizdaroglu, 1992; Ravanat et al. 1995). A ‘cold’ derivatization procedure has been developed that should avoid this problem (Hamberg & Zhang, 1995). The important factor is that any necessary heating stages should be done anoxically: heating DNA bases in the presence of O2 inevitably results in oxidation. Hence some of the claimed artifacts (Hamberg & Zhang, 1995; Ravanat et al. 1995), are possibly due to failure to remove O2 . However, it is difficult to remove O2 completely. Indeed, a major problem to be considered in all these techniques is the possibility that DNA is oxidatively damaged during its isolation from cells and tissues, particularly if phenol-based methods are used, since oxidizing phenols generate ROS (Claycamp, 1992; Finnegan et al. 1996). However, rigorous control of isolation procedures and avoidance of phenol in many laboratories (e.g. by studying isolated chromatin or by using different DNA isolation methods) does not abolish oxidative damage detected in isolated DNA (Halliwell & Dizdaroglu, 1992; Dizdaroglu, 1993a; Harris et al. 1994; Shigenaga et al. 1994; Finnegan et al. 1996), strongly supporting the view that there is a low steady-state DNA damage in vivo. Indeed the presence of a DNA repair enzyme system and the excretion of base damage products support the view that oxidative damage really does occur in vivo. As an alternative means of avoiding possible problems with derivatization an HPLC method with electrochemical detection has been developed that allows measurement of 8OHG and three of the other oxidized base products in acidhydrolysed DNA, thus avoiding the need for derivatization. Liquid chromatography–mass spectrometry techniques are under development in several laboratories: this is another approach to avoiding derivatization problems if sufficient sensitivity can be achieved. 3.2. Oxidative damage to lipids 3.2.1. Lipid peroxidation. There is a range of methods available for measurement of markers of lipid peroxidation and products of peroxidation in vivo that can be measured in blood and urine as indicators of oxidative stress. There are, however, a number of problems that need to be resolved before it is possible to be confident that measurements made are valid and reproducible, and inter-laboratory studies using the same reference material are urgently needed to resolve the remaining difficulties. These include variability of the standards used, and small differences in the technique employed which can have a marked effect on the result achieved; furthermore, no single method can, by itself, provide an unambiguous indicator of levels of lipid peroxidation, whether by measuring lipid hydroperoxides, or degradation products therefrom. Evidence for damage to lipids in vivo is derived from measurement of peroxides or isoprostanes in blood and urine. Such indicators of peroxides in vivo are important vis-à-vis, for example, the relationship of plasma peroxides to vessel wall oxidation of LDL in the context of atherosclerosis. The major problem which has yet to be addressed with some consolidated approach is the differentiation in identification of peroxides formed as a consequence of in vivo oxidative stress and those ingested from dietary sources. S85 Peroxide levels in cells and tissues present a balance between peroxide formation and peroxide metabolism or decomposition, i.e. they are essentially a ‘steady-state’ measurement. With respect to measurement of levels of lipid hydroperoxides, a number of methods are available. Ex vivo measurement of the lipid hydroperoxide products directly is best achieved by HPLC determination following partitioning of the hydroperoxide into a polar solvent, which achieves a primary separation between less polar triacylglycerol and cholesterol hydroperoxides and the more polar free fatty acids and phospholipid hydroperoxides (RiceEvans et al. 1991). Chemiluminescence-based detection has proved very satisfactory in providing a reliable assay procedure (Yamamoto, 1994); an alternative is luminolchemiluminescence. The steady-state levels of peroxides in human body fluids, such as blood plasma, appear very low, usually < 100 nmol/l. These data come from assays that measure ‘real’ lipid peroxides (Holley & Slater, 1991; Akasaka et al. 1995) viz by HPLC with chemiluminescence detection rather than notoriously-unspecific methods such as diene conjugation or the simple thiobarbituric acid (TBA) test (Halliwell & Chirico, 1993). HPLC-based TBA tests can, however, record comparably-low values, provided that butylated hydroxytoluene is added with the TBA reagents (Halliwell & Chirico, 1993). Several other different approaches have been used for measurements of the lipid hydroperoxide products of peroxidation of PUFA. The most simple method conceptually involves direct iodometric determination of lipid hydroperoxide. A further alternative is provided by the ferrous oxidation in xylenol orange (FOX) method in its two variants which provide methods for measuring low levels of soluble hydroperoxides in the aqueous phase, or lipid hydroperoxides derived from membranes of lipoproteins in the lipid phase (Wolff, 1994). Assays of human tissues and body fluids by simple ‘peroxide-determinations’ such as those involving xylenol orange (Jiang et al. 1992) or iodometric methods (Thomas et al. 1989) could measure protein peroxides; this could conceivably explain why levels of alleged ‘lipid peroxides’ measured by such techniques in human body fluids tend to be higher (often in the mM range) than those revealed by the more-specific techniques for measuring lipid peroxides that were discussed earlier (see section 3.2.3). A more recently introduced type of assay concerns measurement of isoprostanes which are derived from PUFA by a non-cyclooxygenase-mediated free-radicalcatalysed mechanism. Formation of the arachidonic acidderived compounds involves formation of four positional peroxyl radical isomers of the fatty acid which undergo endocyclization to prostaglandin (PG)G2 -like compounds that are then reduced to PGF2 -like compounds. Four F2 isoprostane isomers are formed, each of which can, in theory, comprise eight diastereoisomers. Quantification of F2 -isoprostanes represents a reliable and useful approach to assessment of lipid peroxidation and oxidant stress in vivo (Morrow & Roberts, 1994). Human body fluids also contain low levels of F2 -isoprostanes, compounds isomeric to prostaglandins that appear to S86 A. T. Diplock et al. arise by free-radical oxidation of phospholipids containing arachidonic acid (Morrow & Roberts, 1994; Morrow et al. 1995). Isoprostanes appear to exist in human plasma largely esterified to phospholipids rather than ‘free’, and sensitive assays to measure them have been described (Morrow & Roberts, 1994; Wang et al. 1995). Isoprostanes and their metabolites can be measured in human urine (Morrow & Roberts, 1994; Morrow et al. 1995) by GC–MS and this may prove a valuable assay of whole-body lipid peroxidation if a confounding effect of diet can be ruled out. These compounds are useful ‘markers’ of lipid peroxidation and can be measured in plasma (35 (SD 6) pg/ml) and urine (1600 (SD 600) pg/mg creatinine) of healthy volunteers, indicative of ongoing lipid peroxidation even in healthy human subjects (Halliwell, 1996). 3.2.2. Can some measure of ‘total’ peroxidation be obtained? Approaches to measurements of ‘total-body’ lipid peroxidation have been by measurements of urinary TBA-reactive substances (TBARS), using the HPLC TBA method (Chirico & Halliwell, 1994), measurements of hydrocarbon gas excretion and by measurements of F2 isoprostanes in urine. Urinary TBARS measurements have been found to be confounded by a multiplicity of urinary constituents that react with TBA, and this problem is further complicated by contributions from dietary constituents, particularly cooked meats. Most of the lipid-related TBARS appearing in urine seem to arise from lipid peroxides or aldehydes in ingested food, which are presumably largely generated during cooking (Dhanakoti & Draper, 1987; Brown et al. 1995). Hence urinary TBARS is not a suitable assay to assess whole-body lipid peroxidation in response to changes in dietary composition, although it could theoretically be used to look at effects of supplementary antioxidants in individuals on a standardized ‘fixed diet’ (Dhanakoti & Draper, 1987). In any case, HPLC must be used to separate the real (TBA)2 malondialdehyde adduct since the majority of the TBARS in urine are not even lipidderived (Gutteridge & Tickner, 1978) or derive from a wide variety of aldehydes other than malondialdehyde. Measurement of hydrocarbon gases (alkanes and alkenes), degradation products of lipid peroxidation in vivo (Burk & Ludden, 1989; Springfield & Levitt, 1994), can be confounded by interference from air pollutants and from the products of gastrointestinal bacterial metabolism. They are unreliable also because of the low level of the metabolites to be measured, which challenges the sensitivity of the assay, and the possibility of metabolism of the metabolite before excretion, so that the measured amount represents only a portion of the true level; co-elution of other metabolites which are indistinguishable from the products of interest is a further problem; the method is therefore not considered further here. There appears to be a general consensus that the most reliable assay procedures available are the chemiluminescence-linked HPLC determination of lipid hydroperoxides (Yamamoto, 1994) and the HPLC-linked TBA measurement (Yamamoto, 1994). The isoprostane assay is gaining momentum and has been excellently reviewed recently (Morrow & Roberts, 1996). 3.2.3. LDL oxidation. Particular techniques have been developed to determine the oxidation of LDL and a wide range of methods is now available (Esterbauer et al. 1992). A critical review of practical approaches to LDL oxidation has also appeared recently (Rice-Evans et al. 1996). The xylenol orange assay or FOX assay describes a sensitive spectrophotometric system for detecting authentic peroxides in LDL and has been successfully applied to the measurement of lipid hydroperoxides in LDL. Hydroperoxides oxidize ferrous to ferric ions in dilute acid and the resultant ferric ions are determined using ferric sensitive dyes as an indirect measure of hydroperoxide concentration. Xylenol orange [o-cresolsulfonephthalein 30 300 -bis (methylimino) diacetic acid sodium salt) binds ferric ions with high selectivity to produce a coloured (blue-purple) complex with an extinction coefficient of 1:5 × 104 /mol per cm and absorbance maximum of 560 nm. The method compares favourably with the iodometric assay, TBA assay and conjugated diene measurement. No extraction step is required for analysis of lipoprotein in the 900 ml/l methanol–25 mM-H2 SO4 environment in which the assay is performed. Triphenylphosphine is used as a specific reductant of hydroperoxides, converting them to the corresponding alcohol. This allows the measurement of authentic hydroperoxides reacting in the assay and removes any background signal generated. Each sample is therefore measured with and without triphenylphosphine, the difference between the two being the lipid hydroperoxides in the sample. Background values for plasma have been reported to be high. The mean value obtained for native LDL is reported to be 13.3 (SD 8.8) nmol/mg LDL (Rice-Evans et al. 1996). Lipid hydroperoxides can also be measured easily in LDL using the tri-iodide assay (El-Sadaani, 1989). The lipids of LDL are dispersed by the detergent used in an enzymic cholesterol assay kit and the hydroperoxides oxidize I¹ to I2 which is detected spectrophotometrically. The values given for native LDL for the iodometric method correspond to 25 nmol/mg LDL protein (El-Sadaani, 1989), 18.6 (SD 9.4) nmol/mg LDL protein (Esterbauer et al. 1992) and in the range of 10–20 nmol/mg LDL protein (O’Leary et al. 1992). This sort of range was deemed to be at the borderline of the detection limit for the iodometric assay (Esterbauer et al. 1992). The HPLC procedure for detecting lipid hydroperoxides has the advantage that the identity and mass of specific lipid peroxides may be determined down to very low concentrations, well below that available by the colorimetric methods. This sensitivity is dependent on the availability of chemiluminescence detection (Kritharides et al. 1994; Stocker et al. 1991). It has recently been reported (Kritharides et al. 1994) that LDL freshly isolated from healthy subjects was free from detectable amounts of cholesterol-ester hydroperoxides and phospholipid hydroperoxides as measured by HPLC with post-column chemiluminescence detection, suggesting that if lipid hydroperoxides are present at all, the levels must be below 1 nmol/mg LDL protein. This method of detection does require considerable dedication of technical resources and is unlikely ever to be suitable for any routine assays in large-scale clinical research. In some cases, u.v. detection can be used but with lower sensitivity. A comment should be made on the wide-ranging differences between the detected endogenous peroxide levels in LDL Defence against reactive oxidative species applying the HPLC–chemiluminescence method compared with the spectrophotometric assays. (A direct comparison of the methods has yet to be carried out in a single laboratory on the same LDL samples.) No one is sure whether the FOX and El-Sadaani (1989) methods are determining additional unidentified components (which might include protein hydroperoxides) or whether the HPLC assay is missing some contributing features. On the other hand, it may relate to the methods applied for isolating the LDL; for example, it has been reported that applying the FOX assay after rapid isolation procedures gives 3 nmol/mg LDL protein, compared with 13.3 nmol/mg LDL using the sequential isolation methods (Kritharides et al. 1995). 3.3. Oxidative damage to proteins Oxidative damage to proteins is of particular importance in vivo both in its own right (affecting the function of receptors, enzymes, transport proteins etc. and perhaps generating new antigens that provoke immune responses), and because it can contribute to secondary damage to other biomolecules, e.g. inactivation of DNA repair enzymes and loss of fidelity of DNA polymerases in replicating DNA. The chemical reactions resulting from attack of ROS or RNS on proteins are complex. Free-radical attack can generate protein peroxides, which can decompose in complex ways (Ambe & Tappel, 1961; Fu et al. 1995). Most use has been made of the carbonyl assay, a general assay of oxidative protein damage (Levine et al. 1995) to assess steady-state protein damage in human tissues and body fluids. The carbonyl assay is based on the fact that amino acid residues in proteins (particularly histidine, arginine, lysine and proline) are particularly susceptible to attack by ROS producing carbonyl functions. Such increased carbonyl content can be measured after reaction with 2,4-dinitrophenylhydrazine. The carbonyl assay has become widely used and many laboratories have developed individual protocols for it (Levine et al. 1994, 1995). Sometimes the assay procedures used in a particular laboratory are not precisely specified in published papers and even when they are, they often differ from those used originally by the group of Stadtman (Oliver et al. 1987; Levine et al. 1994, 1995). This point is important because there is a considerable variation in the ‘baseline’ levels of protein carbonyls in certain tissues, depending on how the assay is performed (Cao & Cutler, 1995; Lyras et al. 1996). Contrariwise, broadly comparable values for protein carbonyls in human plasma, of < 1 nmol/mg protein have been reported by most groups, so plasma protein carbonyls should be a useful marker of oxidative protein damage for nutritional studies. More work needs to be done to identify the molecular nature of the carbonyls, namely, which amino acid residues have been damaged and in which proteins they reside. Western-blotting assays based on the use of antidinitrophenylhydrazine antibodies have also been developed in an attempt to identify oxidatively-damaged proteins in tissues and body fluids (Keller et al. 1993; Levine et al. 1994). A cautionary note is the covalent binding of certain aldehyde end-products of lipid peroxidation to proteins, generating ‘carbonyls’. Indeed, many oxidized molecules contain carbonyls which will interfere in the protein carbonyl assay. S87 Several in vitro assays for damage to specific amino acid residues in proteins have been developed including assays of 3-hydroxy-L-tyrosine (L-DOPA) (produced by tyrosine hydroxylation) (Giseg et al. 1993), valine hydroxides derived from valine hydroperoxides (Giseg et al. 1993), ring-opening products of tryptophan oxidation (Griffiths et al. 1992; Maskos et al. 1992) 8-oxohistidine (Uchida & Kawakishi, 1993, 1994), dityrosine (Giulivi & Davies, 1993) and ortho and meta-tyrosines, products of attachment of OH: to phenylalanine (Karam et al. 1991; Wells-Knecht et al. 1993). The levels of any one (or, preferably, of more than one) of these products in proteins could, in principle, be used to assess the balance between oxidative protein damage and the repair or (more likely) hydrolytic removal of damaged proteins. The only products exploited to date have been the hydroxylated phenylalanines (Wells-Knecht et al. 1993). Attack of various RNS (ONOO¹ , NO?2 and possibly some other species) on tyrosine (both free and in proteins) leads to production of 3-nitrotyrosine, which can be measured immunologically or by HPLC or GC–MS techniques (reviewed by Wells-Knecht et al. 1993). Reduction of nitrotyrosine to aminotyrosine increases the sensitivity of measurement, since the latter compound can be measured using highly-sensitive electrochemical detection. Nitrotyrosine is also excreted in human urine (Oshima et al. 1990), although the possible confounding effect of dietary nitrotyrosine (if any) and of dietary nitrate and/or nitrite is yet to be evaluated. For measures of total ongoing protein damage, urinary nitrotyrosine (Oshima et al. 1990) might be useful as a generalized index of attack by RNS. Very little research has been carried out on the presence of oxidized amino acids and their metabolites in urine, except that bityrosine has been detected and can be measured by HPLC with fluorescence detection. More work needs to be done in this area, and the possible confounding effects of oxidized proteins and amino acids in the diet (e.g. in irradiated foods; Halliwell & Chirico, 1993) must be considered. 3.4. Measurement of antioxidant nutrients, carotenoids and flavonoids extracted from human plasma Several antioxidants are routinely measured in plasma: atocopherol (and g-tocopherol) whose antioxidant roles are well clarified in vivo; b-carotene, lycopene and other dietary carotenoids for which there is, as yet, little evidence of antioxidant activity in vivo; ascorbic acid, the most efficient reducing agent in vivo, its redox potential defining its central role as an aqueous phase antioxidant. There is, as yet, little information as to the importance of dietary flavonoids as antioxidants in vivo, nor evidence for such activity in vivo, although these polyphenols are highly efficacious freeradical scavengers in vitro. Furthermore, it is only recently that it has become possible to detect and identify flavonoids (and other glycosides) in human plasma, in non-supplemented subjects. Table 1 indicates the most favourable systems for detection, identification and quantification of the antioxidants by HPLC. Plasma samples can be stored at –708. S88 A. T. Diplock et al. Table 1. The most favourable systems for detection, identification and quantitation of antioxidants by HPLC Antioxidant Method a-Tocopherol and g-tocopherol HPLC with fluorescence detection (lex 296 lem 340) Carotenoids HPLC with diode array detection Flavonoids HPLC with diode array detection Conditions Solvent system: hexane–methyl-t-butyl ether (92 : 8, v/v) Column: Novapak Silica 150 × 4:6 mm (4 mm) Internal standard: d-tocopherol Detection limit: in plasma, 1.1 mmol/l Solvent system: acetonitrile–methanol (90 : 10, v/v) Column: Supelco PKB 100, 5 mm or Solvent system: acetonitrile–methanol–dichloromethane– hexane (gradient) Column: Merck lichrocart man-fix, 5 mm Detection at: 455 nm (b-cryptoxanthin) 468 nm (canthaxanthin) 474 nm (lycopene) 454 nm (b-carotene) Solvent system: 1 ml/l HCl in 20 % of aqueous methanol and acetonitrile (gradient) Column: Nova-Pak C18 250 × 4:6 mm (4 mm) Internal standard: salicylic acid Detection: 280 nm (selective) lex and lem , excitation and emission wavelengths. 3.5. Conclusions There is still a long way to go until validated, precise biomarkers of oxidative damage become routinely available. Concerning oxidative damage to DNA, both HPLC and GC–MS have their relative merits but also their limitations. Sight must not be lost of the potential for oxidative damage to DNA during its isolation from cells and tissues. Nevertheless, artifacts arising during derivatization are also problematical and HPLC with electrochemical detection is one proposed route which circumvents the need for derivatization. Liquid chromatography–mass spectrometry techniques are also under development. An initiative by the British Ministry of Agriculture Fisheries and Food (MAFF) is seeking to coordinate, within the MAFF Antioxidants in Foods Research Programme, validation work at a number of centres within Europe, concerned with biomarkers of oxidative damage to DNA. The agenda seeks to have validated agreed biomarkers available for a new generation of human studies within 2–3 years. With regard to ex vivo markers of lipid oxidation in vivo, the measurement of circulating isoprostanes and lipid hydroperoxides is the best approach for plasma, the latter applying HPLC with chemiluminescence detection. Where this detection system is not available, the HPLC–TBA method has been shown to be an alternative, although lacking the high precision of the former. While the principle of isoprostane analysis is highly promising, this is still in its early infancy as there is a dearth of information in the peerreviewed literature as yet, but it is rapidly gaining momentum. For urinary markers of lipid oxidation as an indicator of whole-body lipid peroxidation, the development of the isoprostane analytical techniques will be an important advance, since hydrocarbon gas exhalation has too many confounding variables to be applicable to studies on freeliving human subjects. The potential for the application of methods for ex vivo detection of in vivo protein oxidation is promising. The outcome, of course, is the balance between damage and repair but to date very few products of oxidative damage to proteins have been exploited in this context. There is also a lot of excitement and activity in the areas of evidence for the formation of RNS in vivo. A major stumbling block here in the accumulation of evidence for oxidative damage in vivo is the confounding effects of the presence of oxidized proteins and amino acids in the diet. A lot more work needs to be done in this area. 4. Nutritional options modulating oxidative damage and antioxidant defence systems 4.1. Introduction The body’s antioxidant defence system is capable of being altered by dietary means. A first strategy to balance oxidative damage and antioxidant defence of human cells and tissues would be to enhance the antioxidant capacity by optimizing the dietary intake of antioxidants. A second approach may be to neutralize oxidative compounds in the diet. Crucial to these strategies is knowledge of the required level of relevant antioxidants in the diet to provide protective effects. Another prerequisite is accurate information about food sources, content and bioavailability of antioxidants. Epidemiological studies are necessary to quantify the impact of antioxidants on disease aetiology. Intervention trials formally test the efficacy of enhancing intake of antioxidants. In evaluating these health benefits preferably hard end-points (disease incidence, or recurrence and mortality) should be used. Alternatively intermediate endpoints may be effective, provided that they are genuinely predictors of the disease of interest. In research on functional foods, the development and application of biomarkers is extremely important. In the causal pathway of disease occurrence one can distinguish biomarkers of exposure (dietary intake), biomarkers of biological response and of (subclinical) disease, and biomarkers of susceptibility. For antioxidants, all types of markers have clear relevance. For example, blood levels of vitamin E (an Defence against reactive oxidative species exposure marker) may be studied in relation to oxidation resistance of LDL (a biological response marker) or to carotid artery wall thickness (a disease marker), in subjects with familial hypercholesterolaemia, or specific genotype (both susceptibility markers). Although biomarkers have the potential for improving validity and reducing bias, several problems are encountered. Biomarkers of exposure should accurately reflect relevant dietary intake or body status and early disease markers should have predictive value for the hard end-point. Since chronic diseases have long latency periods requiring large initial numbers to evaluate health effects, biomarkers of intermediate end-points may, in certain circumstances, legitimately be used more efficiently. An effective nutritional strategy will require knowledge of the type of antioxidants in the diet, their food sources, bioavailability and required levels of intake for protective effects. Protective effects of antioxidants have been found in mechanistic studies in vitro and in vivo, and epidemiological studies and certain intervention studies have provided useful information. However, it is appropriate to consider the totality of the evidence from basic science, epidemiology and intervention studies, rather than to rely on the evidence from any one type of study. 4.2. Dietary antioxidants 4.2.1. Sources of dietary antioxidants. There are both nutrient and non-nutrient antioxidants. Non-nutrient antioxidants include flavonoids (found for example in tea, red wine, onions, and apples), polyphenols and terpenes. The focus here will be on nutrient antioxidants, in particular vitamin C, vitamin E and carotenoids, although possible effects of non-nutrient antioxidants must be borne in mind in reaching complete understanding. Thus, it is important to ascertain whether non-nutrient antioxidants are genuinely bioavailable, in the particular sense of whether they are delivered under normal circumstances to tissues where they might be expected to be effective. Their activity as antioxidants themselves, or whether they participate in a cyclical fashion with other antioxidants, also remains to be established. The flavonoid content of beverages has been of great interest recently (Hertog et al. 1993). There is a large variation in the quercetin and myricetin content of red wines, which appears to depend on the type of grape and the vineyard of origin. Quercetin, kaempferol and myricetin are present in black and green teas, so that, together with fruit juices (mainly quercetin), these beverages can provide substantial amounts of flavonoids in the human diet. Vitamin C is found in citrus fruits, peppers, potatoes and other fruits and vegetables. The principal sources of vitamin E are vegetable oils and wheat germ. Other sources are nuts, seeds and leafy green vegetables. Of the over 600 carotenoids, b-carotene has been the most extensively studied. Good sources of b-carotene are yellow or orange fruits and vegetables such as carrots, sweet potatoes, apricots, and mangoes, as well as dark green leafy vegetables such as spinach. There is now increasing interest in other major dietary carotenoids including lycopene (tomatoes), lutein (spinach, broccoli, maize), zeaxanthin (maize), a-carotene (carrots), and b-cryptoxanthin (citrus fruits). Chemical S89 analysis of food products is steadily improving and until recently, there was a lack of reliable data for the food content of carotenoids other than b-carotene. However, the recently released carotenoid food composition database of the US Department of Agriculture has included analysis of a-carotene, b-cryptoxanthin, lutein, zeaxanthin, and lycopene. A number of epidemiological studies of b-carotene and risk reduction of certain diseases are now being re-evaluated using information derived from this source on the other carotenoids. 4.2.2. Antioxidant intake and status. Several crosssectional surveys in a variety of population groups have been conducted. From these studies it can be concluded that according to the recommended daily allowances (RDA), the intake of antioxidants is adequate in healthy subjects. Lower levels have been observed in smokers, the elderly, and in patients with specific diseases or risk factors and several studies have demonstrated that intake of a number of antioxidants may be suboptimal in certain populations. In countries such as France and Italy antioxidant intake is largely adequate due to the abundant supply of fresh fruits and vegetables. However, in other countries where the selection of products is limited or more seasonal, a number of population groups are not able to meet the minimum requirements for vitamins E, C or b-carotene. At particular risk are the less affluent and the elderly. The RDA, which are not established for carotenoids, are defined to prevent nutrient deficiencies and do not take into account the reduction in risk of chronic diseases. The levels of antioxidant nutrients that are effective in the reduction in risk of chronic diseases generally lie higher than the RDA. Since antioxidants may play important roles in the prevention of chronic diseases, the question is what would be the optimal range of intake to recommend. Lachance (1996) distinguishes the following categories: ‘experimental protective intake’ ‘amount in optimal menus’, ‘desirable blood level’, ‘calculated intake necessary to achieve desirable blood levels’. For example, for vitamin E (current RDA: 10 mg men, 8 mg women) the following values have been proposed: experimental protective intake > 23–100 mg; amount in optimal menus 23 mg; desirable blood level 23 mmol/l; calculated intake necessary to achieve desirable blood levels 23 mg. For carotenoids the following values have been suggested: > 4 mg, 5.7 mg, 0.4 mmol/l, 3.2 mg respectively. Diplock (1994) reviewed epidemiological studies of antioxidants and disease and suggested that the following daily intakes were associated with a reduced risk of cancer and cardiovascular disease; 150 mg vitamin C, 40–60 mg vitamin E, and 9–12 mg b-carotene. Lachance (1996) has estimated optimal daily antioxidant intakes to be 145 mg vitamin C, 23 mg vitamin E, and 3.2 mg carotene. Biomarkers of antioxidant intake reflecting internal status (blood, adipose tissue, nails etc.) have been used successfully but need much further development, and these also need to be non-invasive. Conceptually it is important to know the biological relevance and exposure timeframe of the biomarker. It is not only the dietary intake but other exogenous factors such as smoking and alcohol intake, and endogenous factors, that affect antioxidant status measured by a biomarker. For many bio-active compounds intake data are hard to get and, thus, more reliance will have to be S90 A. T. Diplock et al. placed on biomarker data. It is thus imperative that the biomarkers that are used can be shown to be relevant to the outcome of the balance between oxidant and antioxidant in the majority of the population. 4.2.3. Bioavailability of antioxidants. Bioavailability of antioxidants depends on several food and host-related factors, as recently summarized by de Pee & West (1996) for carotenoids. A well-recognized food-related factor is the amount of antioxidant in a meal; for nutrients which are absorbed by a process of passive diffusion, the proportion of antioxidant absorbed decreases with increasing amounts in the food. The molecular forms of antioxidants in foods, for example, for which isomers or molecular linkages such as esters exist, are also important determinants of bioavailability. In addition, the food matrix in which antioxidants are located often influences availability of the nutrient. For example, b-carotene is organized in a pigment–protein complex in green vegetables, but found in lipid droplets in other vegetables and fruits. The b-carotene can be released more readily, and is thus more biovailable, from a fat droplet than from a protein complex. Host-related factors influencing bioavailability include genetic factors, nutrient status and absorption modifiers. Absorption modifiers for fat-soluble vitamins and carotenoids, are lipids in the diet. To ensure efficient absorption, sufficient fat must be present in the meal and diet. A recent epidemiological study on lycopene intake and prostate cancer illustrates the importance of bioavailability in functional food research. Intake of tomatoes, tomato sauce, and pizza were significantly related to lower risk of prostate cancer. Tomato juice was not associated with a protective effect (Giovannucci et al. 1995). The lycopene from tomato juice has a low bioavailability, but cooking tomatoes in an oil-based medium substantially enhances intestinal absorption. 4.2.4. Fat intake and antioxidant status. Consumption of reduced-fat products in order to reduce energy intake, or of products with enhanced PUFA content (Sarkkinen et al. 1993), may also affect the antioxidant status. It was shown that vitamin E intake was significantly lower among subjects who had an increased intake of reduced-fat products; however, no effect was observed for b-carotene or lycopene, or for the antioxidant enzyme activities (superoxide dismutase, catalase and glutathione peroxidase) (Velthuis-te Wierik et al. 1996). There is, however, evidence that plasma antioxidant levels may not be a reliable index of body status; it was found that erythrocyte levels were unaffected in a study in men who had significantly lowered plasma levels of vitamin E (Haddad & Blankenship, 1985). Inhibitors of dietary fat absorption, which may be used to counteract obesity, appear to have fatsoluble vitamin-lowering properties (Melia et al. 1996). Besides a reduced antioxidant intake, fat substitutes like the sucrose polyester-based products may also reduce the absorption of fat-soluble antioxidants, although this only occurs when the fat intake is below 20 % of total energy. However, published data do not suggest a major impact on the absorption of antioxidants when the level of inclusion of the fat substitute in the diet is low, such as would be the case when snack foods are prepared using the fat substitute. In general, diets low in PUFA do result in lower intake of vitamin E. On the other hand, PUFA-rich diets may affect the antioxidant–prooxidant balance requiring higher intake of antioxidants. 4.3. Epidemiological studies on protective effects of antioxidants 4.3.1. Cardiovascular disease. Several scientific reviews have addressed the role of antioxidants in cardiovascular disease (van Poppel et al. 1994). As discussed earlier, the primary role of antioxidants in reducing the risk of cardiovascular disease is through inhibition of peroxidation in LDL, although they may also influence other cardiovascular disease processes. The epidemiological evidence for a protective role of antioxidants in cardiovascular disease is strongest for vitamin E. Three large-scale epidemiological studies demonstrated a relation between vitamin E intake and CHD. The Nurses Health Study, conducted in 87 245 women, found a significant 34 % reduction in CHD in women who had consumed vitamin E supplements containing more than 67 a-tocopherol equivalents (a-TE) daily for more than 2 years (for a definition of a-tocopherol equivalents see p. S95). The Health Professionals Study of 39 910 men showed vitamin E to be associated with a 41 % reduction in risk of CHD. Again, the greatest risk reduction was found with intakes of supplements of more than 67 aTE for 2 years or more. Kushi et al. (1996) reported a significant 62 % reduction in mortality from CHD in women consuming foods containing more than 6.46 a-TE/d. Surprisingly, there was no risk reduction seen with use of vitamin E supplements. This may be due to the lack of information on duration of supplement use, since from the other studies it appears that benefits of supplements are evident only after 2 years. Several epidemiological studies have examined the association between vitamin C and cardiovascular disease. A European cross-cultural study found a significant inverse relationship between CHD mortality and serum vitamin C levels. The first National Health and Nutrition Examination Survey (NHANES I) results showed a 50 % reduction in cardiovascular mortality associated with a daily consumption of greater than 50 mg vitamin C. Gale et al. (1995) reported a significant protective effect against stroke of vitamin C intakes more than 45 mg. However, neither the Nurses’ Health Study nor the Health Professionals Study described earlier found a significant protective effect of vitamin C. A recent study found that use of both vitamin E and C supplements was significantly more protective against cardiovascular mortality than use of vitamin E supplements alone, or no use of supplements (Losonczy et al. 1996). The epidemiological evidence for a role of carotenoids in cardiovascular disease prevention has been reviewed by Kohlmeier & Hastings (1995). A number of studies consistently showed a decreased risk of CHD among subjects with high b-carotene intake or serum levels. In the Massachusetts Health Care Panel study of 1299 elderly people, Gaziano et al. (1995) found a 75 % reduction in risk of fatal myocardial infarction in subjects in the highest quartile of Defence against reactive oxidative species carotene intake. A study of 25 802 subjects showed a significant doubling of the risk of myocardial infarction in subjects with low serum b-carotene levels (Street et al. 1994). Several studies have, however, shown increased risk of CHD among subjects with low b-carotene status. The risk seems to be confined to current smokers. However, bcarotene intake contributes about 25 % of total carotenoid intake, thus carotenoids other than b-carotene as well as other components in fruits and vegetables may be responsible. The sparse data on individual carotenoids do not allow any firm conclusions. Several non-nutritive bioactive compounds may be of relevance in the aetiology of cardiovascular disease. Epidemiological studies on flavonoids are promising, but evidence of benefit is still fragmentary (Hertog, 1994; Muldoon & Kritchevsky, 1996). 4.3.2. Cancer. Numerous epidemiological studies have shown that individuals who regularly consume fruits and vegetables have a decreased risk of cancer. A protective effect of fruit and vegetable consumption was found in 128 of 156 dietary studies (Block et al. 1992). High intake of fruits and vegetables (in the upper one-fourth of the population) is associated with an approximately 50 % reduced risk of cancer, depending on the tumour site. The most consistent evidence of risk reduction associated with fruits and vegetables has been seen in the epithelial cancers of the respiratory and gastrointestinal tract (Steinmetz & Potter, 1991). Evidence is strong for lung cancer, with significant risk reduction found in twenty-four of twentyfive studies, as well as for stomach and pancreatic cancer with protection in twenty-six of thirty studies. For oesophageal, laryngeal and oral cancer, twenty-eight of twenty-nine studies showed a significantly reduced risk with fruit consumption. Cancers of the cervix, ovary and endometrium were associated with a significant risk reduction in eleven of thirteen studies. Although there is, therefore, indirect evidence that antioxidants (carotenoids, vitamin C, vitamin E and possibly non-nutritional antioxidants) may be beneficial (TaylorMayne, 1996), it has to be shown whether, and to what extent, this is true for individual antioxidants and which specific components are responsible. To illustrate the complexities: plasma b-carotene in the normal physiological range is inversely related to lung cancer incidence. Other carotenoids such as a-carotene, which is strongly correlated with b-carotene, may however be the relevant factor. In addition, other bioactive components in fruits and vegetables or other aspects of lifestyle, including diet, may play a causal role. Focusing, in epidemiological studies, on food items which mainly contribute to the daily intake of a specific bioactive compound, e.g. tomatoes and tomato products in the case of lycopene, may be a promising research strategy to identify the responsible factor. In attempts to identify the components in fruits and vegetables associated with risk reduction of cancer, epidemiological studies focusing on intake or serum levels of specified antioxidants have been undertaken. In epidemiological studies of the protective effect of vitamin E on various cancer sites, vitamin E was associated with slightly reduced risk of lung cancer in studies involving low exposure to tobacco smoke. Several studies of oral, pharyngeal S91 and cervical cancer have also found a relationship between vitamin E status and cancer risk. The evidence for stomach and pancreatic cancers has not been consistent, and no association with breast cancer has been found. Diets high in vitamin E intake have been less consistently shown to be associated with cancer protection. Moreover, observational studies of vitamins C and E consumed in supplements provide little support for a strong protective role against cancer. As discussed earlier, vitamin C inhibits the formation of carcinogenic nitrosamines, stimulates the immune system, protects against chromosomal breakage, and regenerates vitamin E as part of the antioxidant defence system. The epidemiological evidence for a risk-reducing role of vitamin C in cancer is not as strong as for fruits and vegetables. However, an extremely strong and consistent protective effect of vitamin C was found in seventeen of nineteen studies of stomach, oesophageal, oral and pharyngeal cancers (Block et al. 1992). Additional studies showed that subjects with low serum levels of vitamin C have a 50 % increased risk of gastric metaplasia or chronic gastritis, which are both precancerous lesions. The most consistent body of epidemiological evidence showing protective effects of an antioxidant nutrient on cancer is for b-carotene. The strongest results are seen for lung cancer. Of twenty-five studies which investigated bcarotene and lung cancer risk, twenty-four showed a significant reduction in risk with high b-carotene intakes or plasma levels (van Poppel & Goldbohm, 1995). For stomach cancer, of fifteen studies, eight found a significant risk reduction associated with b-carotene, and six showed a non-significant risk reduction. Only one study did not find a protective effect of b-carotene against stomach cancer. No consistent associations were found with colorectal, prostate or breast cancer. 4.3.3. Other age-related diseases. Cataract and AMD are eye disorders which show increasing incidence among the elderly. Cataracts result from glycosidation of lens proteins initiated by u.v. light, which leads to opacification of the lens. It has been shown that in vitro lens proteins may be protected against oxidative attack by carotenoids and vitamins C and E. Of ten epidemiological studies, nine showed strong inverse relationships with at least one antioxidant nutrient. Observations are most consistent for vitamins C and E. Macular degeneration is the leading cause of irreversible blindness among persons older than 65 years. The carotenoids lutein and zeaxanthin are present in the retina as pigments to protect against the damaging effects of light; strong inverse relationships were found between intakes of b-carotene, lutein and zeaxanthin and risk of AMD. Consumption of spinach, which is a good source of lutein and zeaxanthin, was also associated with significantly reduced risk of AMD. Free radicals have also been implicated in the development of neurodegenerative disorders such as Parkinson’s disease and Alzheimer’s disease, and in diabetes, rheumatoid arthritis, and chronic obstructive pulmonary diseases. These age-related diseases may, therefore, be beneficially influenced by antioxidant consumption. However, as yet few epidemiological data exist on the association of antioxidants with these disease risks. S92 A. T. Diplock et al. The question whether it is a specific antioxidant or another component, or mixture, of fruits and vegetables that exerts the protective effect on disease cannot be definitively answered by epidemiological studies. Additional information provided by intervention trials is needed before drawing any final conclusions. 4.4. Human intervention studies of antioxidants The gold standard for testing the effectiveness of specific antioxidants is a randomized placebo-controlled intervention trial, preferably with a hard end-point such as disease occurrence or cause-specific mortality. Since trials conducted with clinical end-points of disease incidence or mortality are exceedingly costly and time consuming, intermediate end-points or validated biomarkers are increasingly being used. The limitations of intervention trials are that they can often only be interpreted for the particular study population, and for the antioxidant dose provided during the trial. Most of the intervention trials have focused on cardiovascular disease and cancer; only a few studies have addressed other age-related disorders, i.e. cataract. 4.4.1. Cardiovascular disease. Intermediate endpoints. A number of studies have supplemented individuals with antioxidants and measured the reduction in lipid peroxidation. However, indices of lipid peroxidation have not yet been definitively associated with cardiovascular disease end-points. The only intervention studies which have evaluated the effect of antioxidants on cardiovascular disease using intermediate disease endpoints have been with vitamin E. Using arterial narrowing as a marker for cardiovascular disease progression, the effect of vitamin E was evaluated in 100 patients randomized to 804 a-TE or placebo following angioplasty. The group receiving vitamin E showed significantly less recurrence of stenosis over the 4-month supplementation period. Carotid artery wall thickness, measured by non-invasive arterial wall imaging, has also been used as a reliable intermediate end-point for atherosclerosis. A recent study (Azen et al. 1996) demonstrated that use of vitamin E supplements larger than 67 a-TE/d reduced the progression of atherosclerosis. Disease end-points. Although vitamin E shows promise in reducing cardiovascular disease, relatively few intervention trials have been conducted. The first of the trials to be reported was the a-tocopherol b-carotene (ATBC) study which found no effect on cardiovascular disease mortality among heavy chronic smokers receiving 50 mg vitamin E daily (Albanes et al. 1994). Linxian, China was the site of a study which gave a multivitamin–multimineral supplement plus 15 mg b-carotene and found a 38 % reduction in stroke incidence in subjects with oesophageal dysplasia. A non-significant 10 % reduction in stroke incidence was also observed in a trial involving 35 000 subjects from the general population of Linxian who received 15 mg b-carotene, 30 mg vitamin E and 50 mg Se (Blot et al. 1993). A trial conducted in 161 Japanese subjects, who consumed 3 or 100 mg a-tocopherol/d for 6 years, showed a significant reduction in the number of coronary disorders in the 100 mg/d group (Takamatsu et al. 1995). The recently published Cambridge Heart Antioxidant Study (CHAOS) found a significant reduction in the incidence of non-fatal myocardial infarctions in 2002 subjects randomized to 268 or 536 a-TE/d (Stephens et al. 1996). However, there was a non-significant excess of cardiovascular deaths in the vitamin E-supplemented group. b-Carotene supplementation resulted in increased cardiovascular disease incidence or mortality in two studies conducted in high-risk populations of smokers and asbestos-exposed workers, ATBC and the b-carotene and retinol efficacy trial (CARET) (Albanes et al. 1994; Omenn et al. 1996). In a trial of healthy male US doctors, the Physicians’ Health Study (Hennekens et al. 1996) showed no effect on cardiovascular disease following 12 years of supplementation with 50 mg b-carotene every second day. This was in contrast to earlier findings in a subgroup of the study which found a significant 50 % reduction in secondary coronary events in subjects randomized to b-carotene (Gaziano et al. 1995). It is known that in vitro vitamin C can regenerate vitamin E from the vitamin E radical which is formed during the inhibition of lipid peroxidation. Several large-scale trials are currently underway to study a combination of vitamins E and C, and b-carotene, in the prevention, or amelioration, of risk of cardiovascular disease. 4.4.2. Cancer. Intermediate end-points. Most of the intervention trials investigating the potential role of antioxidants in reducing the risk of cancer have focused on precancerous lesions. Relatively little work has been done on biomarkers for cancer, such as DNA damage. Recent reviews have summarized the use of antioxidants in oral leukoplakia, a precancerous lesion of the oral cavity (Kaugars et al. 1994; Garewal, 1995). Significant impact on lesion size or occurrence was found in seven intervention studies with a supplement of between 30 mg and 180 mg bcarotene. b-Carotene has also been shown to reduce cervical as well as gastric dysplasia, and there are a number of intervention studies currently in progress. Several promising biomarker candidates for cancer have been suggested. Ideal characteristics of these biomarkers include: they should appear earlier than, or more frequently before, tumour development; they should be directly associated with tumour progression; they should be shown to be reversible; they should have inexpensive, accurate and simple methods of detection; they should be validated. Biomarkers of cellular proliferation and differentiation e.g. colonic adenomatous polyps and premalignant lesions such as dysplasia have been used to evaluate the preventive role of nutrient antioxidants. In contrast to observations using microscopic measures of cellular proliferation, the major antioxidants vitamins E and C, and b-carotene seem to be ineffective in decreasing the recurrence of colonic adenomas. Clarification is needed of changes in biomarkers that have been shown to occur with progression of proliferation, and in subsequent steps in the carcinogenic process, before definite conclusions can be drawn from these intermediate biomarker data. The preventive potential of antioxidants has been tested with intermediate end-points for other tumour sites, such as Defence against reactive oxidative species lung, mouth, oral cavity and cervix. It is difficult to summarize the overall findings. However, until now, no really strong benefits have been identified. There is a need for a second generation of chemoprevention trials that include biomarkers, provided that these markers have fulfilled the criteria described earlier. Disease end-point. There have been a number of intervention trials which studied the potential for b-carotene to reduce the risk of cancer. The 5-year Linxian trial of 29 584 adults in China found significant reductions in mortality from total and stomach cancer of 13 % and 21 % respectively, in the group randomized to 15 mg b-carotene, 30 mg vitamin E and 50 mg Se (Blot et al. 1993). In 3318 adults with oesophageal dysplasia in the same community, there was a non-significant 8 % decrease in oesophageal cancer in subjects randomized to a multivitamin, multimineral supplement with 15 mg b-carotene. The Finnish ATBC study of 29 133 chronic heavy smokers tested the effects of 20 mg b-carotene, either alone or in combination with 33.5 a-TE for an average of 6 years. There was a significant increase of lung cancer incidence (16 %) in the groups which received b-carotene (ATBC Cancer Prevention Study Group, 1994). A more detailed analysis of the results revealed that the increased risk of lung cancer appeared to be restricted to participants who had smoked more than twenty cigarettes daily over an average period of 30 years (Albanes et al. 1996). The CARET study of 18 314 subjects at high risk of lung cancer (heavy smokers, and asbestos-exposed workers) evaluated the combination of 30 mg b-carotene and 25 000 IU vitamin A (equivalent to 7.5 mg retinol) over an average of 4 years. The intervention group had a significantly increased risk of lung cancer (relative risk 1.36; Omenn et al. 1996). A reduced risk of lung cancer (relative risk 0.80) was seen in subjects who were former smokers at the beginning of the study. Interestingly, participants with high initial serum b-carotene concentrations had a 31 % reduction in risk of lung cancer (P ¼ 0:003), regardless of which group they were randomized to. This effect was also seen in the ATBC study and Physicians’ Health Study which was conducted over 12 years in 22 071 male physicians who consumed 50 mg b-carotene every second day. There was no beneficial influence of b-carotene on cancer incidence. However, due to the long duration of the trial it is important to note that there were also no negative effects seen (TaylorMayne, 1996). 4.5. Conclusions Epidemiological studies support the hypothesis that the antioxidants vitamin E, vitamin C and b-carotene may play a beneficial role in reducing the risk of several chronic disorders. More research is needed on the impact of other non-nutrient compounds, such as other carotenoids and flavonoids, on human health. Human intervention trials testing the efficacy of antioxidants do not allow firm conclusions because of inconsistent findings, an insufficient number of studies and the use of varying doses. However, there is some evidence that large doses of b-carotene may be deleterious to the health of heavy habitual smokers. S93 In general, human intervention studies using hard endpoints should be regarded as the gold standard. However, for diseases with a long induction period, such as cancer and cardiovascular disease, these types of studies may not be very feasible because of high costs, and intermediate endpoints need to be sought to overcome this difficulty. The relationship between the disease and the nutritional factors that may have been involved at an early stage of its aetiology is, however, very complex and difficult to interpret. The development of biomarkers may, however, help in understanding the complexity of degenerative diseases at their different stages. 5. Potential safety implications related to antioxidant nutritional enhancement 5.1. Introduction The substances to be considered are the antioxidant nutrients vitamin C (ascorbic acid), vitamin E (a-tocopherol): the carotenoids (particularly b-carotene): and the non-nutrient antioxidants, mainly flavonoids, which occur in food and which may be significant in the overall antioxidative protection afforded by the diet. This benefit may be conferred in three main ways: (1) as antioxidants in food during storage and in the gastrointestinal tract; (2) as antioxidants in the human body in vivo; (3) by providing protection in food against oxidation as well as acting as true antioxidants in vivo. From the point of view of safety, the effect of antioxidants must strictly be concerned with their effects in vivo; however, the possibility that a given compound may be converted by chemical reaction, or bacterial action, into a toxic substance within the gastrointestinal tract, or during storage of the food that contains it, must also be borne in mind because such products may themselves be absorbed and exert their toxicity in vivo. 5.2. Vitamin C The tolerance and safety of ingested vitamin C in human subjects has been reviewed several times (Hanck, 1982; Rivers, 1989; Diplock, 1995). In their review Bendich & Langseth (1995) consider in detail the fourteen controlled clinical studies that have reported no side-effects of vitamin C dosage, consistent with uncontrolled anecdotal reports which have appeared. In a more recent detailed review (Bendich, 1997), twenty-two placebo-controlled doubleblind studies are reviewed that indicate no consistent detrimental side-effects of dosages of vitamin C up to daily doses of 6 g. This conclusion is supported by the findings of a further eight less-well-controlled studies. The fact that very large numbers of people regularly take large doses of vitamin C without reports of any adverse effects is anecdotal support for the view that vitamin C is very safe and free from any adverse side-effects. It is clear that overload with vitamin C cannot occur in man even at very high levels of dietary intake (Rivers, 1989). Absorption, tissue concentration, metabolic pathways in which ascorbate participates and renal elimination are all controlled by homeostatic mechanisms. The amount of a dose of vitamin C that is absorbed is inversely proportional to the size of the S94 A. T. Diplock et al. dose and saturation was achieved at a Km of 5.44 mM in a human study using intestinal perfusion. A consistent body pool size of ascorbate in man of about 20 mg/kg body weight, which appeared to change little irrespective of increases in intake to very high levels, has been reported (Kallner et al. 1979). However, a recent detailed study (Levine et al. 1996) provides more information; steadystate plasma concentrations were determined in normal subjects following administration of 30–2500 mg vitamin C daily. Steady-state plasma concentrations followed sigmoid kinetics, the steep portion of the curve occurring between 30 and 100 mg/d, and complete saturation did not occur until a daily intake of 1000 mg. Different kinetics were obtained in blood cells. Possible adverse effects on human health have been reported from time to time. However, an exhaustive search of the literature has failed to find confirmation of this, and in each case evidence exists which refutes the finding which has led to the suggestion. This is summarized as follows. (1) The formation of urinary oxalate stones in subjects ingesting large amounts of vitamin C over a long period proved to be without foundation. Although human subjects do metabolize some ascorbate to CO2 , they excrete considerable amounts of unchanged ascorbate, and a range of metabolites among which is a small amount of oxalate; intake of ascorbate at levels in excess of the level required to maintain plasma levels at about 10 mg/l results in excretion of the excess ascorbate unchanged (Kallner et al. 1979). Approximately 35–40 % of the daily excretion of oxalate is derived from ascorbate, but ingestion of large amounts of vitamin C results in a very small increase in the excretion of oxalate. It was shown clearly (Schmidt et al. 1981) that there is no dose–response relationship between administered vitamin C and excreted oxalate. Part of the explanation of the difference between this result and that of earlier workers is that in the early experiments the alkalinity of urine samples that arises on standing caused conversion of some ascorbate to oxalate; if steps are taken to avoid alkalinity then this conversion is minimal (Wandzilak et al. 1994). Recent work (Levine et al. 1996) indicates, however, that both oxalate and urate excretion are elevated beyond a daily intake of 1000 mg. (2) Similar anxieties were expressed with regard to urate excretion with the possibility that ascorbate might therefore indirectly exacerbate the effect of urate on gout. Two studies demonstrate that in healthy subjects ascorbate ingestion has no effect on the excretion of urate (Mitch et al. 1981; Schmidt et al. 1981). When high non-physiological plasma levels of ascorbate were induced by continuous infusion of ascorbate in gouty, as compared with normal, subjects (Berger et al. 1977), there was no effect on the clearance of urate indicating that it is highly improbable that high dietary intake of ascorbate has any effect on the urinary excretion of urate in subjects with gout. (3) Low plasma levels of vitamin B12 were reported (Herbert & Hacob, 1974) to occur in subjects taking large doses of ascorbic acid but this was shown to be explained by analytical error. Erroneously low levels of plasma vitamin B12 can occur if no cyanide is added to the assay to liberate protein-bound cobalamins and to stabilize the cobalamins so released (Newmark et al. 1976; Markus et al. 1980). (4) High ascorbic acid intake has been shown to have only a small effect on Fe absorption in healthy Fereplete subjects (see review by Bendich, 1997) which repudiates suggestions that Fe overload could be a consequence of high ascorbate intake (Cook et al. 1984). It appears that the regulation of body Fe stores is unaffected by any increased availability of Fe from the diet that might be caused by an effect of the excess ascorbate on Fe. (5) Early reports (Cochrane, 1965; Rhead & Schrauzer, 1971) of rebound scurvy in a small number of subjects following withdrawal of high vitamin C supplements were uncontrolled and have not been substantiated. Studies in guinea-pigs showed no evidence for these claims even when the study was designed to demonstrate a rebound effect. No increased catabolism of ascorbate was demonstrated during high vitamin C dosage nor was there any such increase when the vitamin C dosage was withdrawn (Norkus & Rosso, 1975, 1981). Although the human data available remain contradictory, evidence available at present leads one to conclude that the phenomenon, if it exists at all, does not constitute a significant health problem. (6) Ascorbic acid added to cells in vitro in culture increases the rate of mutagenesis (for review see Rivers, 1989). Detailed reports exist of increased DNA fragmentation, increased DNA repair and chromosome aberrations in cells cultured in media that include added ascorbate. However, these effects only occur in cultures that contain added Cu2þ or Fe3þ ions and when steps were taken to ensure very low levels of these metals in the culture medium, no detrimental effect on DNA was observed. It can be concluded that in any such system in vitro the mutagenic effect of ascorbate is probably due to an ascorbate–metal ion-driven generation of O-derived free radicals. There is no evidence of ascorbate-induced mutagenicity in vivo so that it is highly improbable that any effect that depends on metal ion-driven generation of free radicals caused by ascorbate has any significance. Efficient freeradical scavenging and repair systems protect DNA in vivo from such effects, and intracellular concentrations of ascorbate, and concentrations of metals, which are efficiently sequestered on binding proteins, are so low as to be unlikely to be harmful (Halliwell & Gutteridge, 1989). The conclusion from an exhaustive survey of the literature is that oral intake of high (up to 600 mg/d, i.e. six times the current RDA) levels of vitamin C are safe and entirely free from side-effects (Bendich, 1997). Very high levels (up to 2000 mg/d) have not been consistently reported to result in side-effects, although some reports of low reliability suggest that minor side-effects may occur. 5.3. Vitamin E There have been four reliable reviews of the toxicological safety of oral intake of vitamin E (Bendich & Machlin, 1988, 1993; Kappus & Diplock, 1992; Diplock, 1995). A problem in comparing studies of this nature has been the confusion that exists in the literature between different forms of vitamin E. In particular, IU are often quoted as a measure of quantity. The IU was abandoned by WHO in 1957 (cited by Diplock, 1985) and studies which cite this measure are often ambiguous because it is not possible to determine the precise amount of vitamin E that was administered. Defence against reactive oxidative species The international standard was based on 2-ambo-a-tocopheryl acetate, which was an early sample of vitamin E which contained RRR-a-tocopheryl acetate with an unknown content of ‘2-epi-a-tocopheryl acetate’ (i.e. S,R,R-a-tocopheryl acetate). Because of the uncertainty as to the precise composition of this standard, the IU was abandoned, and in 1983 WHO recommended the use of precise descriptions of pure compounds. Despite this, the IU has continued to be used for labelling purposes, particularly in the US and Canada. There is no description of the various stereoisomers present, which may have widely differing biological activity, and this causes serious confusion. Vitamin E activity is expressed in the present paper as mg RRR-a-tocopherol equivalents (a-TE) wherever this is possible. To estimate the total a-TE the number of mg of tocopherols present is multiplied by a factor as follows: RRR-a-tocopherol RRR-a-tocopheryl acetate RRR-a-tocopheryl succinate all rac-a-tocopherol all rac-a-tocopheryl acetate × 1 :0 × 0:91 × 0:81 × 0:74 × 0:67: Study of conventional aspects of the toxicity of vitamin E in animals was undertaken by many workers over a long period of time (Demole, 1939; Weissberger & Harris, 1943; Levander et al. 1973; Dysma & Park, 1975; Krasavage & Terhaar, 1977; Abdo et al. 1986). There is no evidence of any detrimental effect attributable to vitamin E, and similar conclusions are possible with respect to the teratogenicity and reproductive toxicity of vitamin E in animals at even large levels of intake of the vitamin (Hook et al. 1974; Krasavage & Terhaar, 1977). The possibility that vitamin E might have anticlastogenic and mutagenic effects has been studied extensively, and its potential mutagenicity was tested in several different ways (Shamberger et al. 1979; Beckman et al. 1982; Gebhart et al. 1985). Vitamin E was conclusively demonstrated to have no mutagenic properties, and indeed appears to have some effect in reversing the mutagenic effects of other compounds. Although early literature suggested that impure fractions containing vitamin E had tumour-promoting capability, there is a very large body of evidence that refutes this claim in studies where pure compounds were used (Yang & Desai, 1977; Weldon et al. 1983). Even at very high levels of inclusion in the feed (up to 25 000 IU/kg feed), vitamin E was not shown to have any carcinogenic activity. Many reports purport to deal with the toxicity in human subjects of vitamin E, but there are different levels of reliability that can be attributed to them. Published reports include those that describe single observations on one subject, and planned toxicological studies which may or may not include placebo groups, but which often include bias because they lack blinding. There are fewer highly reliable studies which have been planned and carried out with all necessary rigour, which include sufficient numbers of subjects to enable proper statistical evaluation of the results; they use placebo groups and careful double-blinding, so that the results obtained are valid and reliable. Sufficient such studies have been undertaken to enable an S95 authoritative view of the human toxicity, or lack of toxicity, of vitamin E. There are a few studies (Farrell & Bieri, 1975; Corrigan, 1982; Ernst & Matrai, 1985) which do not have adequate controls but are nevertheless of some interest. No consistent adverse effects of vitamin E emerge from these uncontrolled studies. However, the study of Corrigan (1982) indicates that vitamin E supplementation may aggravate vitamin K deficiency induced by warfarin anticoagulant therapy (see next paragraph). With regard to controlled double-blind studies of vitamin E toxicity in man, several reports show conclusively that vitamin E has very low toxicity in human subjects with no consistent adverse effects being reported in most subjects (Anderson & Reid, 1974; Gillian et al. 1977; Inagaki et al. 1978; Tsai et al. 1978; Stampfer et al. 1983; Bierenbaum et al. 1985; Kitagawa & Mino, 1989). However, some adverse effects were observed on prothrombin time, or on other factors associated with blood clotting. This question was reviewed carefully (Kappus & Diplock, 1992; Diplock, 1995); in several studies no effect was observed on blood clotting whereas in other studies there was a marked effect of vitamin E on some aspects of blood-clotting mechanistics. Following a study of all the reports available it was concluded that vitamin E at a high level of intake may affect the coagulation variables if a low level of vitamin K is also present. Therefore, it is clear that vitamin E cannot be recommended for administration under these conditions. Vitamin E does not, by itself, cause coagulation abnormalities in persons who have no pre-existing coagulation abnormalities, and in individuals who are the majority of the population, vitamin E supplementation is entirely free from these adverse effects on blood-clotting. This view has been challenged recently (Kim & White, 1996). In a doubleblind study in twenty-five patients given warfarin therapy, vitamin E (either 536 or 804 a-TE/d) was given for 4 weeks without any measurable effects on blood coagulation variables. However, in view of the short time span of this study and the small number of subjects (four per experimental group) it would be unwise to overturn the consensus view that vitamin E therapy is contra-indicated in subjects with coagulation abnormalities. The question has been raised (Steinberg, 1993) of the safety of long-term ingestion of amounts of vitamin E in doses of 100 mg/d which were found (Rimm et al. 1993; Stampfer et al. 1993) to confer significant protection against the risk of coronary artery disease. Higher doses than this (up to about 500 mg vitamin E daily) have also been recently shown to confer benefit in apparently causing improvement in subjects with angiographically proven cardiovascular disease (Stephens et al. 1996). Results (Takamatsu et al. 1995) of a trial in which 100 mg d-atocopheryl acetate was administered to human subjects for 6 years indicate that there were no adverse effects of the treatment during clinical follow-up. In another study about 250 mg vitamin E was given daily for 5 years without any side-effects (Greenberg & The Polyp Prevention Study Group, 1994). While it is not possible at present to state categorically that oral ingestion of large amounts of vitamin E is entirely safe for long periods, because no one has been able to test the toxicity of the vitamin for very long periods of time, application to vitamin E of the usual criteria of S96 A. T. Diplock et al. safety applied to any drug suggests that not only will longterm supplementation prove to be free from harmful sideeffects but that this may also convey considerable health benefit. The following conclusions, which were reached earlier (Kappus & Diplock, 1992; Diplock, 1995) with respect to the safety of oral intake of vitamin E by human subjects, can be endorsed here. (1) The toxicity of vitamin E is very low. (2) Animal studies show that vitamin E is not mutagenic, carcinogenic or teratogenic. (3) Reported increases in serum lipids in human subjects following high oral dosage are inconsistent and of little significance. (4) In double-blind human studies, oral dosage resulted in few side-effects, even at a dosage as high as 3.2 g/d. (5) Dosage up to 1000 mg/d is considered to be entirely safe and without side-effects. (6) Oral intake of high levels of vitamin E can exacerbate the blood coagulation defect of vitamin K deficiency: high vitamin E intake is contra-indicated in these subjects. 5.4. Carotenoids Present knowledge of the human toxicology of carotenoids derives almost exclusively from work on b-carotene; the assumption that other carotenoids have similar toxicology to b-carotene may not be justified. For instance the absorption, uptake and tissue distribution may differ among the different carotenoids, some of which are bioavailable and others of which are not. Reviews about the safety of b-carotene (Bendich, 1988; Diplock, 1995) are supported by a review (Wang, 1994) on its absorption and metabolism. Use of bcarotene as a food and cosmetic colourant and as a drug and nutrient has necessitated extensive reliable toxicity studies done using a range of techniques (Bagdon et al. 1960; Heywood et al. 1985). The Ames test revealed no mutagenicity, which was confirmed by studies using the mouse bone-marrow micronucleus test. Embryotoxicity was not found in rats and rabbits, and in a multiple-generation study in rats given up to 1 g/kg per d orally, reproductive function was normal, and there was no interference with embryonic morphology. A study conducted over a 2-year period in dogs revealed no tumourigenicity or chronic toxicity of any kind and in a mouse carcinogenicity study b-carotene was without any tumourigenicity. In several organs of dogs and mice given high doses of b-carotene vacuolated cells were seen due to the formation of fat storage cells; this was not dose-related and was thought to be harmless. These toxicity trials led to b-carotene being placed in the US Food and Drug Administration category of ‘foods generally recognized as safe’ for use as a food colourant, in drugs and cosmetics and as a dietary supplement and nutrient (Office of Life Sciences Research, 1979). In addition b-carotene has been used for 30 years to treat patients with genetically inherited photosensitivities; in this context the ingestion of large amounts of pure b-carotene has not produced toxic sideeffects (Matthews-Roth, 1986). Some individuals taking supplements of >30 mg/d may experience hypercarotenaemia but this disappears quickly after discontinuing the treatment, and it is a benign condition without permanent adverse effects. Anecdotal reports, linked to b-carotene, of leukopaenia, reproductive disorders, increased prostatic cancer incidence, retinopathy, and allergic reactions have not been substantiated in proper clinical trials. A short-term phase I toxicity trial of supplemental b-carotene in a small number of human volunteers (Xu et al. 1992) demonstrated a progressive statistically significant decrease in serum vitamin E concentration during supplementation for 9 months with 15, 30, 45 and 60 mg b-carotene/d. However, other studies have demonstrated no such interaction (Willett et al. 1983; Albanes et al. 1992; McLarty, 1992; Goodman et al. 1994; Nierenberg et al. 1994; Ribaya-Mercado et al. 1995). There is no satisfactory explanation available for the results of Xu et al. (1992), but the balance of probability is that there is no likelihood of any interaction between bcarotene and a-tocopherol that would alter the nutritional availability of vitamin E to human subjects. Careful monitoring of nutrient interactions should become part of all long-term intervention studies. It can be concluded that supplementation of normal individuals in the population with moderate amounts of bcarotene can be undertaken safely. The safety of this practice by heavy smokers, who are at high risk of developing lung cancer, has been put in question by two recent studies (ATBC Cancer Prevention Study Group, 1994; Omenn et al. 1996), details of which are given in section 4. The increase in incidence of lung cancer (18 % and 28 % respectively in the two studies) has no easy explanation. Both studies were conducted among subjects who had a high risk of developing lung cancer and may have been at an advanced precancerous state when the b-carotene administration was commenced. Observational epidemiological evidence suggests that subjects who are not in this state may benefit from b-carotene administration which is likely to exert a protective role at an early stage of the cancer process. Until further work clarifies the situation in heavy smokers with respect to taking supplements, larger doses should be avoided by such individuals. However, it should be stressed that for normal subjects who do not smoke, bcarotene supplementation is entirely safe. 5.5. Non-nutrient antioxidants (flavonoids and other related compounds) Few data exist on absorption, metabolism and possible adverse effect of flavonoids and flavonoid-related compounds, such as genistein, daidzein or phenolic acids. Since these compounds are present in our daily diet they are believed to carry no health problems. Obviously, reliable comprehensive safety data, practically non-existent today, will be required, should it appear that increased intakes of specific flavonoids and other phenolic compounds would confer significant health benefits. As concentrations of phenolic compounds in edible plants, fruits and vegetables vary considerably, estimates of human daily intakes range from approximately 100 mg to 1000 mg (Kuhnau, 1976), depending on the diet. 5.5.1. Absorption. It appears that two major metabolic routes operate in man in the absorption of dietary flavonoids. Intestinal micro-organisms hydrolyse the flavonoid glycosides (nearly all the flavonoids are present as glycosides) to their constituent aglycone and sugar. Most of the aglycones are subsequently metabolized by microorganisms. A minor portion is absorbed as aglycones Defence against reactive oxidative species (Hackett, 1986). A study on the absorption of quercetin in volunteers found no detectable amount of quercetin or of its metabolites in plasma or in urine (Gugler et al. 1975). In a more recent study less than 0.25 % of ingested quercetin was found to be excreted, either unchanged or as conjugates, in the urine of healthy ileostomy volunteers (Hollman et al. 1995). In contrast, 0.1–1.4 % 14 C-labelled catechin was excreted in urine in the form of unchanged compound. However, 55 % of the ingested dose was excreted as its methylated derivatives (Hackett et al. 1983). There appear to be no human studies available on the absorption and metabolism of anthocyanins. 5.5.2. Possible adverse effects. When tested in vitro in cell cultures, flavonols and especially quercetin have a high cytotoxicity (Babich et al. 1993). However, their toxicity in vivo was shown to be remarkably low when tested in rats (National Toxicological Program, 1992). Most of the flavonols including quercetin are also mutagenic in the Ames test and other short-term tests in vitro (Sugimura et al. 1977; International Agency for Research on Cancer, 1983). However, other tests in vitro have shown quercetin to be a potent inhibitor of mutagenic activity of food carcinogens, such as heterocyclic aromatic amines (Stavric et al. 1990). The same anti-mutagenic activity was found in the livers of mice when quercetin was added to their feed (Stavric, 1994; Stavric et al. 1990). Another test on mice showed the mutagenic effect of heterocylic aromatic amines to be enhanced by quercetin (Rowland, 1993). Studies on three flavonols, quercetin, myricetin and kaempferol carried out by the molecular toxicology branch of the US Food and Drug Administration point to the prooxidant properties of these compounds and ‘suggest a dual role for these flavonoids in mutagenesis and carcinogenesis’. The authors stress that these polyphenolic flavonoids are generally considered to be both antioxidants and anticarcinogens (Sahu, 1992; Sahu & Gray, 1993, 1994); these studies were performed in isolated rat liver nuclei under aerobic conditions. A study on quercetin conducted by the US National Toxicological Program (1992) concluded that there was ‘some evidence of carcinogenic activity’ in male rats receiving 40 000 mg quercetin/kg diet for 2 years; there was no evidence of carcinogenicity in the lower dose groups. It is important to note that the statistical evaluation of the National Toxicological Program study was not without controversy (Ito, 1992) and a number of previous long-term, well-controlled feeding experiments with rats, mice and hamsters did not show any carcinogenic activity of quercetin (Hirono et al. 1981; Ito et al. 1989). It should also be emphasized that quercetin is one of the most abundant flavonoids, being present in most common vegetables and fruits with an average daily intake estimated to be approximately 25 mg/person (National Toxicological Program, 1992). The consumption of quercetin from a regular diet appears not to induce health problems. The observed very low bioavailability of quercetin in human subjects may explain the apparent contradiction between the results of the tests carried out in vitro and its apparent beneficial effects in man, although the available evidence for beneficial effects of quercetin needs further confirmation. Whereas flavonoids are not mutagenic, anthocyanins have S97 not been tested regarding their potential mutagenic properties. No carcinogenic effects have been reported for other flavonoids. Tannins, which may exert anticarcinogenic effects by acting as free-radical scavengers, have also been reported to be associated with an increased incidence of oesophageal cancer (Mortin, 1989), but there is no definite evidence to support this assertion. The acute toxicity of tannins was studied in rats, mice and rabbits. Median lethal dose (LD50 ) values for a single dose of orally administered tannins range from 2.25 to 6.00 g/kg body weight (Singleton & Kratzer, 1989). These results are difficult to apply to the human diet since in food the chemical structures and the molecular masses of tannins vary considerably. Anthocyanin extracts are already widely used by the food industry. Anthocyanins obtained by extraction of vegetables and edible fruits have been approved worldwide as food colours (see, for Europe, European Parliament and Council Directive, 1994). They are most commonly based on the following anthocyanidins: peonidin, malvidin, delphinidin, petunidin, pelargonidin, cyanidin. There are two reviews (World Health Organization, 1982; Timberlake, 1988) of the limited toxicological data concerning anthocyanidins. These include data on mutagenicity, reproductive toxicity, and teratogenicity and conclude that anthocyanin-containing extracts are of a very low order of toxicity. The only negative effects were reduced organ and body weights associated with reduced energy intakes at the highest dose, probably reflecting reduced palatability (Clifford, 1996). An average daily intake of 2.5 mg/kg body weight per d was allocated for anthocyanin colour from grape skin extracts (World Health Organization, 1982), the composition of which may vary depending on grape variety and extraction process. Anthocyanin colours from grape skin extracts also contain other flavonoids and some tannins. Recently epigallocatechin gallate and green-tea extract have been shown to prevent gastrointestinal carcinogenesis in volunteers; no adverse effects were reported (Yamana et al. 1996). In conclusion, it is not possible to provide a definitive statement as to the toxicity of flavonoids and other nonnutrient antioxidants, because the literature is either controversial or lacking. A key question that needs to be resolved is the uptake and tissue distribution of non-nutrient substances, which has been hampered by the lack of suitable methodology. Recently a reliable and reproducible method has been published which shows that flavonoids are indeed absorbed in human subjects and enables the measurement in human plasma of flavonoids as their glycosides at concentrations of 0.5–1.6 mmol/l (Paganga & Rice-Evans, 1997). It seems advisable therefore to test rigorously well-defined flavonoids in toxicological programmes for possible adverse effects if new scientific evidence confirms potential beneficial effects on human health. 6. Role of food technology in nutritional and safety aspects of antioxidants 6.1. Introduction Our food supply is safer and offers more variety today than it ever has in the past and this can be largely attributed to the S98 A. T. Diplock et al. applications of food technology to food raw materials. Food processing preserves foods safely so that they can maintain high nutritional and organoleptic values during storage and achieve a wide distribution. After microbial spoilage, oxidation is the second most important cause of food spoilage, even in those products which might be considered low in oxidizable substrates, such as potato flakes. Important nutrients such as unsaturated lipids, vitamins and proteins can be lost through oxidation. This has been known for a long time and therefore attention to the control of oxidative processes in foods has been one of the highest priorities of the food industry. Until recently antioxidants have been viewed as tools in this fight against oxidation but now, because of growing recognition that antioxidants are themselves important for the maintenance and optimization of health, special efforts may be required to protect the antioxidant nutritional value of food. O2 is the enemy of antioxidants. This is evident but bears mentioning because the protection of foods from interaction with O2 is the basic principle on which antioxidant protective technologies are based. Many of these have been drawn from experience with lipid oxidation. These technologies can be conveniently broken down into two sections: physical and chemical. 6.2. Physical processes 6.2.1. Structural integrity. The structural integrity of the food plays an important role in protecting antioxidants from contact with O2 and thus potential oxidative destruction. In its whole food form, the antioxidant is encased in its protective liposome or cell membrane structures and is out of contact with both O2 and oxygenases. For example, intact oilseeds are quite stable but once they are crushed, extracted and heated in the refining process, resistance of the oils to oxidation decreases. Thus, one nutritional consequence of this loss of structural integrity is negative in that some antioxidants are lost to oxidation. However, the positive nutritional consequence is that antioxidants become more bioavailable. It has been known for some time that in human subjects the gastrointestinal absorption of carotenoids from vegetables is inversely proportional to particle size (Rodriguez & Irwin, 1972). Heat treatment also improves absorption: more lycopene appeared in plasma when human subjects drank heated tomato juice compared with the same dose of raw tomato juice (Stahl & Sies, 1992). In another recent study, ascorbic acid was more bioavailable from cooked broccoli than it was from the raw form (Mangels et al. 1993a). This appears to be a human attribute: studies in other species such as rats (Sweeney & Marsh, 1974) and preruminant cows (Poor et al. 1993) have shown, respectively, no differences or only small differences in absorbability of cooked v. uncooked vegetables. Therefore it may not be appropriate to extrapolate information obtained using animal models to man. 6.2.2. Moisture content. The moisture content of foods should be judiciously chosen and carefully controlled if the antioxidant content of the food is to be optimized since oxidation shows a U-shaped curve, i.e. rapid rates of oxidation when moisture content is low as well as when it becomes too high. Most dehydration processing will destroy lipoxidases but if water content falls below a level which permits the formation of a protective monolayer of water over the surface of the food increased auto-oxidation can result (Labuza, 1971). According to Labuza (1971), a higher level of water can act as a solvent which mobilizes catalysts and reactants to sites of oxidation. Water may also interact chemically or by H-bonding with other molecular species. 6.2.3. Temperature. Thermal treatment holds an important place in food processing because of the many benefits it brings to food preservation. It also has both a negative and a positive impact on antioxidants. The positive effects include inactivation of oxidases, and breakdown of food structures leading to improved bioavailability. These topics have been dealt with earlier. Degradation. Thermal treatment also has an important negative impact on antioxidants in foods. Carotenoids and anthocyanins are sensitive to heating: in many studies of foods and food models, these compounds progressively degrade as time and temperature increase. With radical treatments this effect can be radical too. For example at 1008 for 5 min, only 0.93 % b-carotene was lost in a glycerol model, but 97.8 % was lost when the thermal treatment was carried on for 4 h at 2108 (Onyewu, 1985). In palm oil thermal destruction of b-carotene doubles with every 208 rise in processing temperature (Jideani, 1992). However, technologies exist to protect b-carotene, tocopherols and tocotrienols. Using a modification of the refining method one Malaysian supplier guarantees specifications of 400 mg b-carotene/kg and 800 mg/kg minimum of total tocopherols and tocotrienols. This is quite a good level given that most crude palm oils contain on average only 500–750 mg/kg (Hood, 1995). Water-soluble antioxidants such as ascorbic acid are also sensitive to heating. Blanching of green beans, spinach, broccoli, or peas before freezing resulted in, respectively, losses of 0, 20, 20 and 33 % of this vitamin (Unilever, 1995). Isomerization. Another effect of heating is to increase isomerization of carotenoids in foods. In one study, thermal processing of guava juice increased cis-lycopene levels by 5-fold and decreased trans-lycopene, although in this food, b-carotene remained unchanged (Padula & Rodriguez, 1987). Supplemental b-carotene added to wholewheat breads and crackers before baking showed significant trans to cis isomerization (4–15 % for bread and 18–23 % for crackers) (Ranhota et al. 1995). More isomerization was observed in the crackers because of the more severe thermal treatment, and less water was left in the final product compared with the bread product. The processing associated with canning is sufficient to cause isomerization of carotenoids: 18–30 % of b-carotene is converted to cis isomers by usual treatment of typical fruits and vegetables (Quackenbush, 1987). The nutritional significance of this is that the cis isomer of b-carotene is markedly less well absorbed by the gastrointestinal tract and markedly less well transported in the body (Gaziano et al. 1995). Thus, cis and trans isomers of carotenoids are not biologically equivalent. Storage temperature. Losses of antioxidants from fresh fruits and vegetables during storage can be quite significant. Defence against reactive oxidative species At ambient temperature 90 % of vitamin C was lost from spinach leaves within 3 d. At refrigerator temperatures this loss was greater than 50 %. In one recent study, freezing effectively halted any decrease in vitamin C losses in peas, broccoli, green beans and spinach over 3 months (Unilever, 1995). 6.2.4. Minimizing oxygen. There are some processes which can eliminate or minimize the presence of O2 in the foods. This is quite important as, for example, the rate of degradation of L-ascorbic acid in orange juice (Kennedy et al. 1992) and in canned, sterilized green beans (Bloeck et al. 1986) depends on O2 present in the sample. Fruit juices containing entrapped air are deaerated by being sprayed into a vacuum deaerator which minimizes the potential for destructive changes to ascorbic acid and other oxidizable components due to O2 . Other products e.g. dried red peppers (Lee et al. 1992) can be prepared or packaged under a N2 atmosphere with successful retention of antioxidants. One of the newer approaches is the potential to use enzymes as antioxidative agents in food systems. Enzymes can act by removing O2 , ROS such as H2 O2 and superoxide radicals or by reducing lipid hydroperoxides. The enzymes presently under most study are glucose oxidase (EC 1.1.3.4), thiol oxidase, galactose oxidase (EC 1.1.3.9), pyranose oxidase and hexose oxidase in conjunction with catalase, superoxide dismutase and glutathione transferase and glutathione peroxidase (for review see Meyer & Isaksen, 1995). These systems are under study but are not currently of commercial significance in the food industry. 6.2.5. Protection from light. At wavelengths of less than 500 nm, light is an important generator of lipid oxidation and vitamin destruction. Improving packaging by adding colour (red, brown, black) or obscuring material is one of the most effective strategies for protecting foods from light. Some components of foods act as photosensitizers: riboflavin, chlorophyll, myoglobin, haemoglobin pigment, pheophytin and certain conjugated double-bond systems are examples. Sandmeier (1996) has recently shown that curcumin and curcumoid compounds, which are often present in curry mixes, have rather potent photosensitizing properties. Removing these photosensitizers or incorporating food-grade quenchers such as b-carotene, and atocopherol, together with chelators of trace metals, can also be effective means of controlling the damaging effects of light. 6.2.6. Irradiation. The effects of irradiation are somewhat variable depending on the food being irradiated, the nutrient under study and the strength and length of time of irradiation. The irradiation of peppers demonstrated no change in vitamin C content in doses up to 300 Gy (Mitchell et al. 1992). In fact, post-irradiation storage resulted in increases in total vitamin C in intact fruits and vegetables (Mitchell et al. 1992). In one study of chicken meat, b- and a-tocopherol decreased linearly with irradiation in direct relationship to dose. At 3 kGy, which is the maximum recommended by the Food and Drug Administration for chicken, 15 % b-tocopherol and 30 % a-tocopherol were lost (Lakritz & Thayer, 1992). However, a newer study by the same authors (Lakritz & S99 Thayer, 1994) showed different results: only a 6 % reduction in a-tocopherol and no significant change for b-tocopherol in chicken meat at 3 kGy and 28. Irradiation with a low dose (0.010 kGy) had no effect on the polyphenol content of herb teas (Katusin et al. 1988). 6.3. Chemical processes 6.3.1. Enzymes. The control of polyphenol oxidases has been a priority in food science for a long time because oxidized polyphenols are responsible for undesirable brown colours in cut fruits and vegetables, and for the development of off-colours and off-flavours in frozen foods over time. Inactivation of these enzymes can be achieved by rapid high-temperature heat treatment (blanching) while the fresh character of the food can still be retained. Rapid inactivation of the oxidases after pressing apples is a prerequisite for the long-term stability of apple juice (Anonymous, 1993). Extrusion processes, because of their conditions of high temperature and pressure, destroy oxidases and increase lipid binding to protein and carbohydrate elements in the food which acts to protect from lipid oxidation (Artz et al. 1992). Cloudy fruit juices are more stable than clear juices because of their content of polyphenols and their glycosides. These latter compounds are not substrates for polyphenol oxidase (EC 1.10.3.1) (Baruah & Swain, 1959). For the preparation of clear juices such as apple or raspberry, pectinases are often used. These enzymes hydrolyse the glycosidic side-chains of flavonoids which are then less oxidatively stable in the juice (Rommel & Wrolstad, 1993a,b). However, glycosides of flavonols and isoflavones are poorly absorbed by the small intestine compared with their aglycones (Brown, 1988). Therefore the nutritional impact of using pectinases is difficult to gauge: on the one hand the polyphenols become more bioavailable, and therefore potentially capable of having an impact on health, but on the other hand there may be fewer of them in the food, as being less stable, they are oxidized. 6.3.2. Supplementation. Supplemental antioxidant nutrients can be added to foods for technological reasons. Ascorbic acid is widely used to increase the resistance to oxidation of many foods. For example, it is used as a dip for cut fruit and vegetables as it is preferentially oxidized instead of the catechol–tannin compounds and effectively inhibits enzymic browning reactions (Potter & Hotchkiss, 1995). In hydrophobic food matrices, vitamin E is also frequently used to provide antioxidant protection. Since vitamin C recycles vitamin E, it is desirable to use both of these antioxidants together. However, since vitamin C is hydrophilic and vitamin E lipophilic, it is difficult to use them together effectively in food systems. One of the solutions is to use ascorbyl palmitate, a hydrophobic vitamin C which has been shown to be useful in food lipid systems (Klaui & Pongracz, 1981). A mixture of vitamin E, vitamin C and phospholipids is an effective antioxidant partly because the phospholipid acts as an emulsifier, allowing the two vitamins to be in contact with one another and also because the phospholipid itself actively participates in the antioxidant process (Loeliger et al. 1996). Both vitamins can be incorporated into liposomes S100 A. T. Diplock et al. which can be used as a delivery system in oil-in-water emulsions (Pothakamury & Barbaosa-Canovas, 1995). Sometimes higher levels of antioxidants can be incorporated indirectly into a foodstuff. Feeding chickens supplemental vitamin E increased a-tocopherol levels in the meat and this was associated with greater stability (less formation of volatiles) of the meat when it was later irradiated (Patterson & Stevenson, 1995). Supplementation with nutrient antioxidants is, thus, a reasonable strategy in many situations for the improvement of the stability and nutritional value of a processed food. Indiscriminate supplementation with antioxidants, as with any nutrient, would however be quite undesirable as there are many unknown factors concerning potential interrelationships in vivo between different antioxidants and eventual long-term effects. To illustrate this point, in a recent study, b-carotene doses resulted in less lycopene in LDL suggesting that the two related hydrocarbons compete for a similar absorption or transport mechanism (Gaziano et al. 1995). Non-nutrient antioxidants such as SO2 can be used to advantage to protect those with known or suspected nutritional benefit. For example, the phenolic and procyanidin composition of grape juice processed with SO2 is consistently higher than that of grape juice processed without SO2 addition (Spanos & Wrolstad, 1990), although there appears to be no effect on quercetin. SO2 will also prevent oxidation of carotenoids but causes bleaching of anthocyanins (Sian & Soleha, 1991). Supplemental antioxidants from natural sources such as herbs and spices are effective in conserving foods from oxidation. Many plant extracts have antioxidant activity: rosemary, sage, thyme, oregano, ginger, turmeric, cloves and bay leaves all are active in descending potency (Loliger et al. 1996). Different compounds have very different activities. For example three components of thyme extract, thymol, carvacrol and p-cymene-2,3-diol, were decreasingly resistant to thermal stress in a lard model (Ternes et al. 1995). Mixtures of antioxidants can be obtained from natural sources, for example, by mixing with oil and pressing (Aeschbach & Wille, 1993), or by using supercritical CO2 extraction of antioxidants. This latter technique has aroused interest because the solvent involved (CO2 ) is nontoxic, non-flammable and of low cost (Tsuda et al. 1995). O2 is almost eliminated from the system. Polyphenols can be recovered through membrane technologies, for example from the waste water from olive oil processing (Trägardh, 1995). It is presently poorly understood whether compounds from this or other fruit and/or vegetable sources have antioxidant ability in vivo similar to that they demonstrate in vitro and therefore could potentially contribute direct health benefits. 6.4. Conclusions The food industry has long experience in the control of oxidative damage in foods and this experience can be used to advantage for the protection of food antioxidants which are beneficial. Some of these, such as vitamins C and E and b-carotene are known, and strategies for their protection in foods are already exploited by many food technologies. However, there are many compounds with antioxidant activity which may, or equally may not, have biological activity. Since oxidation cannot be eliminated completely, efforts must be made to identify those antioxidants which are important to health, and in what form they are most useful. This requires that more information must be obtained in the much neglected field of their bioavailability as well as on their bioactivity. With this kind of information, food technology strategies for the preservation of those antioxidants which are beneficial to health could then be applied in a cost-effective manner. 7. Critical assessment of the science base and conclusions 7.1. Identification of criteria The foregoing sections have surveyed the science base that underpins the argument that oxidative damage is a significant causative factor in the development of human diseases; that antioxidants are capable of preventing or ameliorating these disease processes; that the administration of antioxidants to human subjects is safe; and that food technology can adapt to meet the needs for nutritional quality and safety in the use of antioxidants in foods. It is now necessary to evaluate clearly what the impact of this survey is on the question of whether an identified specific antioxidant nutrient can be said to positively affect function and have a riskreducing role. This must involve consideration of whether a cause–effect relationship can be established between a dietary antioxidant and a health benefit, and what is the optimum level of intake of the antioxidant in conferring that benefit. There are four questions the answers to which are considered to be fundamental to providing the required critical assessment. (1) What is needed to establish that free-radical events are involved in detriment to health which eventually can be associated with the pathology of an identified disease? (2) What is needed to establish that antioxidants have specific health benefits? (3) What is needed to make a claim about the functional, nutritional or health benefits of antioxidant nutrients? (4) What is needed to bring an antioxidant product to the market? The technology base is very relevant because it is important, in bringing a product to the market, that the active ingredients in the food are still active when the consumer consumes them. A further matter also needs careful consideration. There is a clear difference between, on the one hand optimizing health, by improving intake of food constituents, and on the other hand preventing disease, although the boundaries between these two objectives are not clear-cut. Furthermore, the difference between an effect of a nutrient as a food, and its pharmacological effect, usually at a higher level of intake, needs to be borne in mind. It is important to distinguish between healthy products which optimize well-being and health, and those which may be capable of preventing disease. A single functional food is unlikely by itself to prevent disease; if it were to do so it would be at a very high level of dietary Defence against reactive oxidative species intake and questions of toxicity might become important. The disease is the paradigm, the ‘experimental model’, by means of which it is possible to determine the importance to human subjects of the nutrient. Antioxidants appear to be useful for optimizing well-being, such as in ageing and in simple stress situations, not preventing disease. The future emphasis for functional foods that contain antioxidants must be in providing a suitable level of intake that enhances wellbeing and health. A physician examines a patient by a set of established signs, as a result of which the patient is pronounced ‘well’. There is, however, a category beyond this which may be referred to as ‘well-being’; this can be defined by objective criteria (McDonald & Newell, 1996; Spilker, 1996). The optimal level of intake of an antioxidant might be demonstrated to be higher than that which is capable of being delivered by conventional food, in which case the ‘functional food’ might have the level of the antioxidant enhanced during manufacture by fortification or some other means. The eventual goal of all this is to determine whether specified foods can be considered as ‘functional foods’ in that they may be able to confer health benefit following their consumption. For the purposes of this discussion, a functional food is considered to be a food which delivers a physiological benefit, and description of it should convey unambiguous information that is without deception to the consumer about physiological or health benefit. In order that a claim about health benefit can be made, in an ideal world information should be available, which can be critically substantiated from the literature, which establishes the following criteria, each of which, although inter-related with the others, should be individually satisfied. Criterion 1. A plausible and validated basic science rationale for: (a) the involvement of free radicals in those biochemical and cellular processes which have been shown to lead to detriment to well-being and health, and to specified human diseases; and (b) the involvement of dietary antioxidants in the prevention of these free-radical-involving biochemical processes. This should include results from studies carried out in vitro, in cells in culture, in in vitro/ex vivo models, and in animal models. Criterion 2. The existence of human population epidemiological data which demonstrate a statistically validated inverse relationship between intake (or preferably serum concentration), of individually specified antioxidants and the risk of, or mortality from, particular diseases. Criterion 3. The existence of prospective statistically validated epidemiological evidence that links intake (or serum concentration), of identified antioxidants at an early stage of the disease process, with risk of human disease which may develop some time after the exposure to the antioxidants. In such studies both intermediate end-points that have been clearly shown to predict subsequent disease, as well as final end-points, may be used. Criterion 4. The existence of biomarkers for evaluating free-radical events in human subjects, and the modulating effect on them of antioxidants, which are of interest in maintaining well-being as well as the balance between health and disease. This will include validation of biomarkers, which must preferably be methodologically uncomplicated, by inter-laboratory studies of the same material S101 which has been shown, when investigated in different laboratories, to give the same answers, preferably by more than one method. A further requirement is that the chosen biomarker must be directly relevant, that is it must have functional significance; the biomarker must have been shown by unequivocal techniques to have some significance to both function and the health maintenance or disease risk to which it is linked. Criterion 5. The existence of statistically validated interventional human evidence in large groups of human subjects (a) which clearly shows that enhancement of intake of specified antioxidants is associated with improvement in a valid index of well-being and health, or a lowered risk of subsequent disease; (b) which demonstrates optimal levels of intake of antioxidants, derived from indices measured by chosen biomarkers. This would imply that such biomarkers have been established which demonstrate a valid relationship between intake of the antioxidant and the index evaluated by the chosen biomarker(s). Criterion 6. The existence of clear evidence that the intervention that is proposed with an antioxidant nutrient is safe. This will include evidence that a conclusion as to safety applies with equal force to all groups in the population, including those that are indulging in behaviour which might be expected to increase the risk of the disease concerned. These are clearly extremely stringent criteria. What is required in coming to a conclusion as to the health benefit of antioxidants is to evaluate the ‘state of the art’ with respect to them, and to come to a conclusion as to how and to what extent these criteria are satisfied at present. 7.2. Critical evaluation of the present knowledge base Each of the foregoing sections 2–6 has examined in depth one aspect of the science database that exists concerning the discussion here. The following conclusions are derived from those given in detail in the appropriate section. 7.2.1. Conclusions from section 2. An imbalance of ROS and antioxidant defence systems may lead to chemical modifications of biologically relevant macromolecules. This imbalance provides a logical pathobiochemical mechanism for the initiation and development of several disease states. Experimental data obtained in vitro provide evidence that antioxidants function in systems that scavenge ROS and that these are relevant to what occurs in vivo. The relevance in vivo of these observations depends inter alia on knowledge of the uptake and distribution of the antioxidant within the human body, and on what tissue levels of the antioxidant may be expected in relation to dietary levels. Epidemiological studies show a correlation between consumption of foods rich in antioxidants and a decreased risk of several diseases. In particular, diets rich in fruit and vegetables are associated with low risk of several diseases and antioxidants are among the responsible constituents of such a diet. Data on antioxidant supplementation are contradictory and further research is necessary to establish whether supplementation beyond normal dietary intake levels is of benefit. 7.2.2. Conclusions from section 3. There is some way to go until validated precise methods are available for measuring biomarkers of oxidative damage in human S102 A. T. Diplock et al. subjects in vivo under minimally invasive conditions. With respect to oxidative damage to DNA, HPLC and GC–MS methods have both merits and limitations. Oxidation artifacts also arise in sample preparation and in derivatization of semi-purified samples. Lipid oxidation products in plasma are best measured as isoprostanes or as lipid hydroperoxides using specific HPLC techniques. Development of isoprostane measurement will advance specificity and precision and will enable measurement of lipid peroxidation products in urine, which will give a measure of whole-body lipid peroxidation. Measurement of hydrocarbon exhalation is a technique fraught with artifacts and is unsuitable for human studies. The measurement of oxidative damage to proteins has some potential but such methods have not been effectively exploited. Measurement of metabolites derived from RNS has potential, but at present there are many confounding factors which restrict its use. 7.2.3. Conclusions from section 4. Epidemiological studies support the hypothesis that the major antioxidant nutrients vitamin E and vitamin C, and b-carotene, which may or may not be acting as an antioxidant in vivo, may play a beneficial role in prevention of several chronic disorders. More research is needed on the impact of other non-nutrient compounds, such as other carotenoids and flavonoids, on human health. In general, human intervention studies using hard end-points are the gold standard. Trials are restricted mainly to the major antioxidants and do not allow firm conclusions because of inconsistent findings, an insufficient number of studies and the use of varying doses. However, there is evidence that large doses of b-carotene may be deleterious to the health of certain subgroups of the population such as heavy habitual smokers. In functional food research, preventive trials with intermediate end-points may be of help for testing the efficacy of antioxidants. Bioavailability studies and dose-finding studies in combination with the development and application of biomarkers are required for a successful research strategy. 7.2.4. Conclusions from section 5. Vitamin C is safe at levels of supplementation up to 600 mg/d, and higher levels, up to 2000 mg/d, are without risk. Vitamin E has a very low human toxicity and an intake of 1000 mg/d is without risk; 3200 mg/d has been shown to be without any consistent risk. Large intakes of b-carotene must be viewed with caution because they have been shown to confer detriment to a population at high risk of lung cancer when administered after many years of high-risk (smoking) behaviour. Until further work clarifies the situation in heavy smokers with respect to taking supplements, larger doses should be avoided by such individuals. There is little reliable information about the human toxicology of flavonoids and related non-nutrient antioxidant constituents of the diet. A key question is whether these substances are taken up by human subjects and distributed to the tissues in quantities sufficient to confer biological effect. 7.2.5. Conclusions from section 6. The food industry has long experience in the control of oxidative damage in foods and this experience can be used to advantage for the protection of food antioxidants which are beneficial. Some of these, such as vitamins C and E and b-carotene, are well known, and strategies for their protection in foods are already exploited by food technologies. There are, however, many compounds with antioxidant activity which may, or equally may not, have biological activity. Since oxidation during manufacture cannot be eliminated completely, efforts must be made to identify those antioxidants which may be important to health and to discover in what form they are most useful. This requires that more information must be obtained in the much neglected field of their bioavailability as well as of their bioactivity. With this information, food technology strategies for the preservation of those antioxidants which have been shown to be beneficial to health can be applied in a cost-effective manner. 7.3. Evaluation of criteria By using what is revealed in section 7.2, it is now possible to evaluate the criteria set out in section 7.1. Criterion 1. There is excellent scientific evidence, derived from several kinds of investigation, for the involvement of free-radical events, either as initiating cause of, or at a later stage in, biochemical and cellular processes that lead to health detriment. Antioxidants that are consumed in the human diet are involved in modulating these free-radicalinvolving events and, thus, they promote health. Valid animal models exist for some forms of cancer and these have been used to demonstrate an apparent preventive effect of some antioxidants against the development of cancer, which gives general support to the contention of health benefit and cancer prevention by antioxidants. There is no acceptable animal model for arteriosclerotic cardiovascular disease. However, there are models which reproduce some aspects of atherosclerosis and limited studies with these support the view that vitamin E may have a role with other factors in delaying the arteriosclerotic process. Criterion 2. There is human epidemiological evidence for an inverse association between intake of antioxidant nutrients, or the fruits and vegetables that contain them, sometimes supported by measures of serum concentration, and risk of specified diseases in the population. Although there are some instances of lack of such correlations being reported, most studies support this conclusion. The validity of many studies would have been greatly strengthened by inclusion of measures of serum concentrations of the antioxidants under study, and the validity of the conclusion varies in strength for the individual antioxidants under consideration. Criterion 3. There is abundant prospective epidemiological evidence for a clear correlation between dietary intake of fresh fruit and vegetables and a lowered subsequent risk of cancer. Some data also show a similar correlation with vascular disease. This is a widely accepted view and a claim that increased fruit and vegetable intake is associated with lowered incidence of some forms of cancer and atheromatous cardiovascular disease is entirely valid. Although it is generally thought that this health benefit is conferred by the antioxidant nutrient and non-nutrient substances in fruit and vegetables, no conclusive proof of this contention exists and much confusion has been caused by statements to this effect; data that show an association between high serum levels of specified antioxidants and lowered disease incidence provide no answer to the question Defence against reactive oxidative species because the antioxidant may merely be a marker for some active agent derived from fruit and vegetables, or a lifestyle difference associated with high fruit and vegetable intake. The use of ‘cancer’ or ‘cardiovascular disease’ as biomarkers of free-radical causation, or antioxidant efficacy, are also unreliable because both diseases are multifactorial and develop over a long period of time with the possibility of multiple inputs into their aetiology, in which antioxidants may be only one relevant factor. There is some evidence, particularly with respect to cardiovascular disease, that links very high intakes of vitamin E, derived from food supplements, with a dramatically lowered incidence of disease. However, the likely level of intake in this case is beyond what might be feasibly obtained from the diet, and at normal dietary levels of intake there appears to be little health benefit in this respect. Such an effect could, however, be obtained by the consumption of a functional food fortified to contain a high level of vitamin E. Criterion 4. At the present time biomarkers exist for evaluating oxidative damage to DNA and PUFA; some biomarkers for oxidative damage to proteins also exist. However, there are many anomalies with respect to these biomarkers in that there are both internal inconsistencies in measurement of the same material by different methods, and external inconsistencies in that some laboratories report values that differ widely from those obtained elsewhere using apparently similar methods. There is a great need for validation studies in which methodology is tested within the same laboratory, but also most importantly between different laboratories which must be able to obtain measurements that are consistent with the results of others using the same material. The predictive value of the biomarker in evaluating its relevance to the aetiology of a particular disease also requires validation. Only when this kind of validation has been done and all experts agree as to methodology will it be possible to utilize the methods in a new generation of human studies, which will assess the quantitative relationship between antioxidant intake and health benefit in human subjects. Criterion 5. The results of intervention studies have been very mixed and at present the situation is confused. There are intervention studies in specified populations from which some conclusions can be drawn. However, it is not valid, and indeed may be highly dangerous, to extrapolate to other populations. For example, intervention studies in China have used end-points which relate to a particularly high natural incidence locally of a particular cancer; the observed data are of interest insofar as they refer to a particular cancer in a specific location. The fact that the incidence of that cancer elsewhere is very low implies that there are other factors involved which may prevent widening the conclusions from China to other locations. Second, the trials involving b-carotene supplementation in smokers should only be perceived as applying to heavy smokers and must not be applied to other sections of the population who do not smoke or who only smoke moderately. A further disadvantage of present interventional evidence is the fact that, because cost considerations have driven experimental design, individual antioxidants have often not been examined but instead a cocktail of several has been employed. There are no valid studies that demonstrate optimal levels of S103 antioxidant intake in human subjects, and no data exist concerning dose–response relationships of nutrient intake and biological efficacy. Thus, this criterion may prove to be too stringent, although it is important to identify some kind of ‘gold standard’ in this respect. In order to give informed advice as soon as possible about satisfaction of this criterion it is necessary to shorten the time that is required to conduct and evaluate human studies because those that rely on final end-points are very long and extremely costly to carry out. The development and use of intermediate end-points that give genuine information about the final end-point must, therefore, be an important consideration. Criterion 6. There is excellent evidence that antioxidant nutrients are safe and can be included in the food of human subjects up to quite high levels without discernible risk. It is necessary to cite some specific cautions, for example, in administering large levels of vitamin E to persons with blood coagulation disorders, and b-carotene to heavy smokers; such caveats are, however, rare, specific and easily identified. There is less certainty with respect to non-nutrient antioxidants and, if these substances are shown to have potential functional health benefit, then it will be necessary to devote attention to their safety. 7.4. Final conclusions (1) There is evidence that mechanisms that involve free radicals are implicated at some stage of the development of human diseases, and that the maintenance of well-being depends on the supply through the diet of antioxidant nutrients and b-carotene which modulate free-radical processes in vivo. If it is shown that non-nutrient antioxidants are biologically available then they too may contribute to the total antioxidant effects of the diet in vivo, as well as contributing antioxidant potential during processing, storage and in the gastrointestinal tract. (2) Present epidemiological evidence is incomplete but in general it supports the basic hypothesis that antioxidant nutrients contribute to well-being and health. There is no evidence to the contrary which would negate this conclusion. There is minimal similar evidence with respect to the non-nutrient antioxidants. (3) The development of biomarkers for use in human studies is well advanced but biomarkers that are becoming established need critical evaluation before they can be used in a new generation of human studies to examine and evaluate a quantitative cause-and-effect relationship between antioxidant intake and health benefit. (4) There are few reliable human intervention studies which establish cause–effect relationships, and there is no evidence which shows clearly the optimal amounts of nutrient and non-nutrient antioxidants that are needed in the human diet. The identification and development of intermediate end-points for evaluating the effect of intervention needs further careful work. (5) Increasing human dietary intake of antioxidants is safe and without undesirable side-effects except in rare welldefined instances. (6) There is a need to ensure that measures adopted by food technologists in the processing of foods maintain the antioxidant content of the food, and that the antioxidants are S104 A. T. Diplock et al. still active at a suitable level when the food is consumed by the public. If it can be shown with certainty that a certain level of antioxidants in the food is associated with health benefit it may be necessary for food producers to consider means of enhancing the content or form of the antioxidant up to or beyond that usually found in the natural food. 8. Recommendations for future research 8.1. Introduction The foregoing section 7 has compared the available scientific database with the evidence that is required to achieve scientific consensus as to what may be needed in order to make a specific claim about the role of antioxidants in maintenance of health and well-being. This enables the identification of areas where further research is required. Improvement in dietary antioxidant intake in human populations is expected to result in lowering of the risk of a number of degenerative diseases. While desirable as an ultimate objective per se, the impact on public health and the resultant decrease in health-care costs make it imperative that substantial sums of money should be spent on research in this important area. 8.2. Specific recommendations Specific recommendations for further research are warranted in each of the areas covered by the foregoing sections 2 to 5. With regard to section 6 which deals with the role of food technology, there is no particular direction that research needs to take at the present time. Developments in food technology will be based on, and adapt to, nutritional recommendations resulting from biologically driven work. For example, if it is found that the bioactivity of certain flavonoids is high, then steps will have to be taken to preserve, or even enhance, the amounts of these compounds in food products. Similarly, if glycosides of certain compounds prove to be more bioavailable than the free antioxidant then similar considerations will apply to the handling of glycosidic derivatives by food technologists to maximize their content in the food. 8.2.1. Oxidative damage and antioxidant defence systems of the human organism. It is axiomatic to the present considerations that prooxidants cause, or are implicated in the development of, human disease. Direct measurement of prooxidants in vivo is difficult or impossible. It is imperative to establish which are the critical free-radical ‘hits’ that are the relevant ones; for cancer, which of the multiple modifications of DNA caused by oxidative insult are those that lead to detriment; for atherogenesis, what is the relative relevance of protein and lipid modifications to LDL which contribute to the process, and what is their relationship to the overall multifactorial process of arteriosclerosis? Are the processes examined really relevant to the disease causation? For cancer, there is strong evidence that prooxidants are involved in the initiation stage of the complex process that leads to disease. But what is the relevance of the different forms of oxidative damage to the process? Investigation of this should clarify the role of different antioxidants in controlling carcinogenesis. We need simpler and more specific methods to track ROS both in the initiation and progression of the cancer process. Studies on free-radical involvement in atherosclerosis need to focus more clearly on the relevance of oxidative damage to other factors involved, and on the relative importance of each to the overall complex biochemical process. Noninvasive techniques need to be developed to study these phenomena in human subjects in vivo and this applies particularly to the field of eye diseases which do not readily lend themselves to direct investigation. We need to identify which are the important antioxidants in terms of the maintenance of health, and what is their relationship to one another. We need to clarify whether it is the antioxidant role of the substance that is important, or whether it is some other function, and the possible non-antioxidant effects of antioxidants, in particular with respect to modulation of gene expression, also need further research. We need to know whether there are combined effects of synergy or antagonism both with respect to the antioxidants themselves and to other food components. The primary aim is to identify the active components in the overall system that promotes health. With respect to single antioxidants there is a need for more information regarding regeneration of vitamin E from its radical under cellular conditions, and the role of b-carotene in apparently exacerbating the cancer process, when given at a late stage to heavy smokers, needs investigation. 8.2.2. Ex vivo methodologies for quantitating and validating damage in vivo to biological macromolecules. Before meaningful work can proceed on providing evidence of the level of antioxidants needed to maintain health and well-being, it is necessary to refine and validate methods that are already available for measurement of oxidative damage in human subjects in a non-invasive manner. This will enable real measurement of variables related to oxidative damage and its attenuation by antioxidants to be made. With respect to biomarkers for oxidative damage to DNA, methods available need to be developed further to prevent oxidation during workup, and formation of artifacts during derivatization of samples, and careful validation studies in a number of different laboratories need to be undertaken using identical test material and methodology. Studies of mitochondrial, as opposed to nuclear, DNA may prove to be of value because there is some evidence that suggests that mitochondrial DNA oxidation correlates with other variables of oxidative damage. Similarly, for biomarkers of lipid oxidation the relatively new isoprostane method needs development and comparison with older established, but less reliable, methods, and this needs to be examined both in plasma and urine as well as other accessible biological material. Validation in many centres by measurement of oxidative damage in the same biological material needs to be undertaken using the same methodology. There is a somewhat longer-term need for development of techniques to be used ex vivo as measures of protein oxidation in vivo, which will, however, probably be able only to reflect the difference at a steady state between damage and repair. It is possible that the present early indications of biomarkers of RNS may be capable of development into useful methodology. 8.2.3. Nutritional options modulating oxidative damage. For proper epidemiological research, as well as for human intervention studies, it would be desirable to put special Defence against reactive oxidative species emphasis on the following. (1) Chemical analysis of the antioxidant content of foods so that more realistic food composition tables can be compiled; ideally this should take account of agricultural practices, industrial processing and food preparation. (2) Studies of bioavailability of antioxidants from the diet, and the factors that influence the absorption, distribution and tissue uptake of the compounds and the likely impact of the antioxidants on metabolic processes. This will include studies of the metabolism of antioxidants and the possible metabolic interactions between them. (3) Development and validation of biomarkers of intermediate end-points, both biological response markers and early disease markers, and emphasis on the relevance of the biomarker to the disease end-point as well as the disease process. (4) Application of the validated biomarkers of intermediate end-points in randomized controlled trials testing the efficacy of antioxidants in functional foods for the maintenance of health and wellbeing. 8.2.4. Safety implications of nutritional enhancement of antioxidants. The detailed evidence that is already available which demonstrates that vitamin C and vitamin E are safe at quite high levels of inclusion in the human diet, means that it is unnecessary to recommend further work in this area. The safety of b-carotene was not questioned before the results of the Finnish and American intervention studies, which showed an apparent exacerbation in the incidence of lung cancer in heavy smokers who were given supplements of b-carotene. This observation needs urgent clarification, and this should also include work on the safety of other carotenoids which have been shown to have biological activity in human subjects. With respect to the flavonoids and other polyphenols, it is likely that their bioactivity will be explored, and the key question of their bioavailability clarified, in the near future. It will be necessary to examine the safety of such bioflavonoid, and other similar compounds, which may be shown to have bioactivity in human subjects, as information becomes available about their bioactivity. 8.3. Priorities for the recommendations made The most urgent requirement for further research is the validation of available biomarkers of oxidative damage. Much useful work has been done already but several anomalies remain and new questions have emerged. Before these biomarkers can be used, in particular as biomarkers of intermediate end-points, in a new generation of human studies, which especially need to assess the level of inclusion of antioxidants in the diet that is optimal for health and well-being, it is necessary to engage in a programme of validation of the biomarkers that are available. These validation studies will also necessitate the inclusion of studies on the analysis of antioxidants and their metabolites. Three types of validation are essential: first there is a need for comparison of results obtained in the same laboratory on identical material using different but complementary methodology so that any numerical differences in the results obtained can be eliminated or at least minimized. Different methods sometimes appear to give different results for what at first sight should be the S105 same measurement by different means. Where numerical differences remain following evaluation, there must be clear reasons given as to the scientific explanation of the difference; for example, the different methods may be measuring slightly different aspects of the same oxidative damage. Second, there is a need for comparison of results obtained in as many different laboratories as possible of identical material which is exchanged between participating laboratories, each of which should use as many agreed methods as possible to assess the degree of oxidative damage in the samples employed. Third, there is a need for a different kind of validation, which is that measurements made must be shown to be clearly linked to those phenomena which give rise to disease in human subjects. For example, is the DNA oxidative damage that is measured a real indicator of the involvement of such damage in mutagenesis and eventual carcinogenicity? Is the oxidation of LDL a reliable indicator of atherogenesis and eventual vascular disease? The development of a disease must be the ultimate paradigm by which the relevance of a biomarker is judged. As a logical second stage of the work described in the foregoing paragraph, the validated and accepted biomarkers will be used in a new generation of human studies. These need not be as lengthy or time-consuming, and thus not as expensive, as previous prospective epidemiological intervention studies, because the use of intermediate end-points should enable answers to key questions to be obtained considerably more quickly than in earlier studies whose end-points were disease phenomena occurring many years after the initiation of the study. These new studies will provide, for the first time, rigidly controlled evidence of the benefit to be gained from antioxidants in the human diet, and will enable quantitation of the optimal levels of intake of antioxidants. In this connection, care must be taken to ensure that the importance of the antioxidant contribution of the whole diet, as distinct from that of each individual antioxidant, will be evaluated. The proposals for further research set out in section 8.2 will need to be addressed as and when funds are available, but this should be done at the same time as the major programme of research elaborated in the first two paragraphs of this section. The proposed work set out in the first paragraph could be set in motion with minimal delay and it is anticipated that preliminary results could be available within 2 years of its inception, with reliable final results being available shortly thereafter. The interventional studies in human subjects that are proposed in the second paragraph could commence at the end of 3 years and results would begin to accumulate within 5 years of the initiation of the studies. Although it is possible to be quite specific with respect to the objectives and proposed methodology of the research that is envisaged, and the time-frame within which specified outcomes may be expected to be achieved, it is beyond the capabilities of those concerned with this report to attempt to evaluate the costs that might be involved. Research costs are a highly volatile quantity and there is considerable variation in cost that might be expected even within different institutions of a single Member State of the European Union. Cost comparisons between different Member States are even more difficult because labour costs, and cost of delivery of S106 A. T. Diplock et al. research, differ very widely. It would be invidious to attempt to identify locations where the research might be carried out more cheaply than in another, and it is the considered view of this group that money identified by the European Union for supporting this research should be made available on the usual competitive basis, with invitation of specific proposals from researchers concerned with their own specific areas of expertise. These can then be assessed both with respect to their scientific merit, as well as their cost, and appropriate awards of money made on this basis, which follows accepted practice in all Member States of the Union. References Abdo KM, Rao G & Montgomery CA (1986) Thirteen week toxicity study of D-alpha-tocopheryl acetate (vitamin E) in Fischer 344 rats. Food and Chemical Toxicology 24, 1043– 1050. Aeschbach R & Wille HJ (1993) Procédé d’extraction d’antioxydant d’origine végétale sous forme liquide (Procedure for extraction of antioxidants of vegetable origin in liquid form). European Patent 639336. Akasaka K, Ohata A, Ohrui H & Meguro H (1995) Automatic determination of hydroperoxides of phosphatidycholine and phosphatidylethanolamine in human plasma. Journal of Chromatography B 665, 37–43. Albanes D & 22 others (1996) Alpha-tocopherol and beta-carotene supplements and lung cancer incidence in the ATBC cancer prevention study: effects of baseline characteristics and study compliance. Journal of the National Cancer Institute 88, 1560– 1570. Albanes D, Virtamo J, Rautalahti M, Haukka J, Palmgren J, Gref CG & Heinonen OP (1992) Serum b-carotene before and after b-carotene supplementation. European Journal of Clinical Nutrition 46, 15–24 Ambe KS & Tappel AL (1961) Oxidative damage to amino acids peptides and proteins by radiation. Journal of Food Science and Technology 26, 448–451. Ames BN (1989) Endogenous oxidative DNA damage, ageing and cancer. Free Radical Research Communications 7, 121–128. Anderson TW & Reid DB (1974) A double-blind trial of vitamin E in angina pectoris. American Journal of Clinical Nutrition 27, 1174–1178. Anonymous (1993) Production of light colored cloudy juice. Food Marketing and Technology, December issue, 28–32. Artz WE, Rao SK & Sauer RM (1992) Lipid oxidation in extruded products during storage as affected by extrusion temperature and selected antioxidants. In Food Extrusion Science and Technology, pp. 55–97 [JL Kokini, CT Ho and MV Karwe, editors]. New York: Marcel Dekker Inc. ATBC: Alpha Tocopherol, Beta Carotene Cancer Prevention Study Group (1994) The effect of vitamin E and b-carotene on the incidence of lung cancer and other cancers in male smokers. New England Journal of Medicine 330, 1029–1035. Azen SP, Qian D, Mack WJ, Sevanlan A, Selzer RH, Liu Ch, Liu Ci & Hodis HN (1996) Effect of supplementary antioxidant vitamin intake on carotid arterial wall intima-media thickness in a controlled clinical trial of cholesterol lowering. Circulation 94, 2369–2372. Azzi A, Boscoboinik DO, Marilley D, Ozer NK, Stauble B & Tasinato A (1995) Vitamin E: a sensor and an information transducer of the cell oxidation state. American Journal of Clinical Nutrition 62, 1337S–1346S. Babich H, Borenfreund E & Stern A (1993) Comparative cytotoxicities of selected minor dietary non-nutrients with chemopreventive properties. Cancer Letters 73, 127–133. Bagdon RE, Zbinden G & Studer A (1960) Chronic toxicity studies of b-carotene. Toxicology and Applied Pharmacology 2, 225– 236. Baruah P & Swain T (1959) The action of potato phenolase on flavonoid compounds. Journal of the Science of Food and Agriculture 10, 125. Beckman JS (1996) Oxidative damage and tyrosine nitration from peroxynitrite. Chemical Research in Toxicology 9, 836–844. Beckmann C, Roy RM & Sproule A (1982) Modification of radiation-induced sex-linked recessive lethal mutation frequency by tocopherol. Mutation Research 105, 73–77. Bendich A (1988) The safety of b-carotene. Nutrition and Cancer 11, 207–214. Bendich A (1990) Antioxidant nutrients and immune functions – introduction. Advances in Experimental Medicine and Biology 262, 1–12. Bendich A (1997) Vitamin C safety in humans. In Vitamin C in Health and Disease, pp. 367–379 [L Packer and K Fuchs, editors]. New York, NY: Marcel Dekker. Bendich A & Langseth L (1995) The effects of vitamin C supplementation: a review. American Journal of Clinical Nutrition 14, 124–136. Bendich A & Machlin LJ (1988) Safety of oral intake of vitamin E. American Journal of Clinical Nutrition 48, 642–649. Bendich A & Machlin L (1993) The safety of oral intake of vitamin E: data from clinical studies from 1986 to 1991. In Vitamin E in Health and Disease, pp. 411–417 [L Packer and K Fuchs, editors]. New York, NY: Marcel Dekker. Bendich A, Machlin LJ, Scandurra O, Burton GW & Wayner DDM (1986) The antioxidant role of vitamin C. Free Radical Biology and Medicine 2, 419–444. Berger L, Gerson CD & Yu TY (1977) The effect of ascorbic acid on uric acid excretion with a commentary on the renal handling of ascorbic acid. American Journal of Medicine 62, 71–76. Berliner JA & Heinecke JW (1996) The role of oxidized lipoproteins in atherogenesis. Free Radical Biology and Medicine 20, 707–727. Bertram JS & Bortkiewicz H (1995) Dietary carotenoids inhibit neoplastic transformation and modulate gene expression in mouse and human cells. American Journal of Clinical Nutrition 62, 1322S–1326S. Bierenbaum ML, Noonan FN, Machlin LJ, Machlin S, Stier A, Watson PB & Naso AM (1985) The effect of supplemental vitamin E on serum parameters in diabetics, postcoronary and normal subjects. Nutrition Reports International 31, 1171– 1180. Block G, Patterson B & Subar A (1992) Fruit, vegetables and cancer prevention: a review of the epidemiological evidence. Nutrition and Cancer 18, 1–29. Bloeck M, Iseli-Winters S, Perren M, Escher F & Solms J (1986) Quality changes during storage of heat sterilized vegetables. In The Shelf Life of Beverages, pp. 393–411 [G Charalambous, editor]. Amsterdam: Elsevier. Blot WJ, Li JY, Taylor RR, Guo W, Dawsey S, Wang GQ, Yang CS, Zhrug SF, Gali M, Li GY, Liu BQ, Fangren J, Sum YH, Liu F, Fraument JF, Zhang YH & Li B (1993) Nutrition intervention trials in Linxian, China. Supplementation with specific vitamin and mineral combinations, cancer incidence and disease-specific mortality in the general population. Journal of the National Cancer Institute 85, 1483–1492. Brown ED, Morris VM, Rhodes DG, Sinha R & Levander OA (1995) Urinary excretion of malondialdehyde in subjects fed meat cooked at high or low temperatures. Lipids 30, 1053–1056. Brown JP (1988) Hydrolysis of glycosides and esters. In Role of the Defence against reactive oxidative species Gut Flora in Toxicity and Cancer, pp. 109–144 [IR Rowlands, editor]. San Diego, CA: Academic Press. Burk RF & Ludden TM (1989) Exhaled alkanes as indices of in vivo lipid peroxidation. Biochemistry and Pharmacology 38, 1029–1032. Cao G & Cutler RG (1995) Protein oxidation and ageing. Difficulties in measuring reactive protein carbonyls in tissues using 2, 4-dinitrophenylhydrazine. Archives of Biochemistry and Biophysics 320, 106–114. Cadet J, Ravanat J-L, Buchko GW, Yeo H & Ames BN (1994) Singlet oxygen DNA damage: chromatographic and mass spectrometric analysis of damage products. Methods in Enzymology 234, 79–88. Chirico S & Halliwell B (1994) High performance liquid chromatography-based TBA tests. Methods in Enzymology 233, 314–318. Chu YH & Lin JY (1993) Factors affecting the content of tocopherol in soyabean oil. Journal of the American Oil Chemists’ Society 70, 1263–1268. Claycamp HG (1992) Phenol sensitization of DNA to subsequent oxidative damage in 8-hydroxydeoxyguanosine assays. Carcinogenesis 13, 1289–1292. Clifford MN (1996) Anthocyanins in foods. In Proceedings of the Symposium of Polyphenols and Anthocyanins as Food Colorants and Antioxidants, pp. 1–19 [W Phannhauser and A Strigl, editors]. Vienna: University of Vienna. Cochrane HA (1965) Overnutrition in prenatal life: a problem? Canadian Medical Association Journal 93, 893–899. Cook JD, Watson SS, Simpson KH, Lipschitz DA & Skikne BS (1984) The effect of high ascorbic acid supplementation on body iron stores. Blood 64, 721–726. Corrigan JJ (1982) The effect of vitamin E on warfarin-induced vitamin K deficiency. Annals of the New York Academy of Sciences 393, 361–368. De Pee S & West CE (1996) Dietary carotenoids and their role in combating vitamin A deficiency: review of the literature. European Journal of Clinical Nutrition 50, Suppl. 2, 237–259. Demole V (1939) Pharmakologisches über Vitamin E (Pharmacology of vitamin E). Internationale Zeitschrift für Vitaminforschung 8, 338–341. Dhanakoti SN & Draper HH (1987) Response of urinary malondialdehyde to factors that stimulate lipid peroxidation in vivo. Lipids 22, 643–646. Diplock AT (1985) Fat Soluble Vitamins: Their Biochemistry and Applications. London: Heinemann. Diplock AT (1994) Antioxidants and disease prevention. In Molecular Aspects of Medicine, vol. 15, pp. 295–376 [H Baum, editor]. Oxford and New York: Pergamon Press. Diplock AT (1995) Safety of antioxidant vitamins and b-carotene. American Journal of Clinical Nutrition 62, Suppl. 6, 1510S– 1516S. Dizdaroglu M (1993a) Chemistry of free radical damage to DNA and nucleoproteins. In DNA and Free Radicals, pp. 19–39 [B Halliwell and OJ Aruoma, editors]. Chichester: Ellis Horwood. Dizdaroglu M (1993b) Quantitative determination of oxidative base damage to DNA by stable isotope-dilution mass spectrometry. FEBS Letters 315, 1–6. Dysma HA & Park J (1975) Excess dietary vitamin E in rats. FASEB Journal 34, 912. Eldin AK & Appelqvist LA (1996) The chemistry and antioxidant properties of tocopherols and tocotrienols. Lipids 32, 671–701. El-Sadaani M, Esterbauer H, El-Sayed M, Goher M, Nassar AY & Jurgens G (1989) A spectrophotometric assay for lipid peroxides in serum lipoproteins using a commercially available reagent. Journal of Lipid Research 30, 627–630. European Parliament and Council Directive (1994) European Parliament and Council Directive 94/36/EC on colours for use in foodstuffs. L239, 13–29. S107 Erdman JW, Bierer TL & Gugger ET (1993) Absorption and transport of carotenoids. Annals of the New York Academy of Sciences 691, 76–85. Ernst E & Matrai A (1985) Einfluss von a-Tocopherol (Vitamin E) auf die Fliesseigenschaft des Blutes (Influence of a-tocopherol (vitamin E) on blood flow characteristics). Therapiewoche 35, 5701–5702. Esterbauer H, Gebicki J, Puhl H & Jurgens G (1992) The role of lipid peroxidation and antioxidants in oxidative modification of LDL. Free Radicals in Biology and Medicine 13, 341–390. Farrell PM & Bieri JG (1975) Megavitamin E supplementation in man. American Journal of Clinical Nutrition 28, 1381–1386. Finnegan MTV, Herbert KE, Evans MD, Griffiths HR & Lunec J (1996) Evidence for sensitization of DNA to oxidative damage during isolation. Free Radicals in Biology and Medicine 20, 93– 98. Flagg EW, Coates RJ & Greenberg RS (1995) Epidemiologic studies of antioxidants and cancer in humans. Journal of the American College of Nutrition 14, 419–426. Floyd RA, Watson JJ, Wong PK, Altmiller DH & Rickard RC (1986) Hydroxy-free radical adduct of deoxyguanosine: sensitive detection and mechanisms of formation. Free Radical Research Communications 1, 163–172. Fridovich I (1986) Superoxide dismutases. Advances in Enzymology 58, 61–97. Fu S, Gebicki S, Jessup W, Gebicki J & Dean RT (1995) Biological fate of amino acid, peptide and protein hydroperoxides. Biochemical Journal 311, 821–827. Gale CR, Martyn CN, Winter PD & Cooper C (1995) Vitamin C and risk of death from stroke and coronary heart disease in a cohort of elderly people. British Medical Journal 310, 1563–1566. Garewal H (1995) Antioxidants in oral cancer prevention. American Journal of Clinical Nutrition 62, Suppl. 6, 1410S–1416S. Gaziano JM, Johnson EJ, Russell RM, Manson JE, Stampfer MJ, Ridker PM, Frei B, Henekens CH & Krinsky NI (1995) Discrimination in absorption or transport of b-carotene isomers after oral supplementation with either all-trans- or 9-cis-bcarotene. American Journal of Clinical Nutrition 61, 1248–1252. Gebhart E, Wagner H, Grziwok K & Behnsen M (1985) The actions of anticlastogens in human lymphocyte cultures and their modification by rat liver S9 mix. Chapter II. Studies with vitamins C and E. Mutation Research 149, 83–94. Gerster H (1995) b-Carotene, vitamin E and vitamin C in different stages of experimental carcinogenesis. European Journal of Clinical Nutrition 49, 155–168. Gillian RE, Mondell B & Warbasse JR (1977) Quantitative evaluation of vitamin E in the treatment of angina pectoris. American Heart Journal 93, 444–449. Giovanucci E, Ascherio A, Rimm EB, Stampfer MJ, Colditz GA & Willett WC (1995) Intake of carotenoids and retinol in relation to risk of prostate cancer. Journal of the National Cancer Institute 87, 1767–1776. Giseg SP, Simpson JA, Charlton TS, Duncan MW & Dean RT (1993) Protein-bound, 3,4-dihydroxyphenylalanine is a major reductant formed during hydroxyl radical damage to proteins. Biochemical Journal 32, 4780–4786. Giulivi C & Davies KJA (1993) Dityrosine and tyrosine oxidation products are endogenous markers for the selective proteolysis of oxidatively modified red blood cell hemoglobin by (the 19S) proteasome. Journal of Biological Chemistry 268, 8752–8759. Goodman GE, Metch BJ & Omenn GS (1994) The effect of longterm b-carotene and vitamin A administration on serum concentrations of a-tocopherol. Cancer Epidemiology Biomarkers Prevention 3, 429–432. Greenberg ER & The Polyp Prevention Study Group (1994) A clinical trial of antioxidant vitamins to prevent colorectal adenoma. New England Journal of Medicine 331, 141–147. S108 A. T. Diplock et al. Griffiths HR, Lunec J & Blake DR (1992) Oxygen radical induced fluorescence in proteins: identification of the fluorescent tryptophan metabolite, N-formylkynurenine, as a biological index of radical damage. Amino Acids 3, 183–194. Gugler RM, Leschik M & Dengler HJ (1975) Disposition of quercetin in man after single oral and intravenous doses. European Journal of Clinical Pharmacology 9, 229–234. Gutteridge JMC & Tickner TR (1978) The characterisation of thiobarbituric acid reactivity in human plasma and urine. Analytical Biochemistry 91, 250–257. Hackett AM (1986) The metabolism of flavonoid compounds in mammals. In Progress in Clinical and Biological Research, vol. 213, pp. 177–194 [VE Cody, JR Middleton and HB Harborne, editors]. New York, NY: A.R. Liss. Hackett AM, Griffiths LA, Broillet A & Wermeille M (1983) The metabolism and excretion of 1-(14 C)-cyanidanol-3 in man following oral administration. Xenobiotica 13, 279–286. Haddad E, Blankenship JW & Register UD (1985) Short-term effect of a low-fat diet on plasma retinol and a-tocopherol and red cell a-tocopherol levels in hyperlipidemic men. American Journal of Clinical Nutrition 41, 599–604. Halliwell B (1995) Antioxidant characterization; methodology and mechanism. Biochemical Pharmacology 49, 1341–1348. Halliwell B (1996) Oxidative stress, nutrition and health. Free Radical Research 25, 57–74. Halliwell B & Chirico S (1993) Lipid peroxidation: its mechanism, measurement and significance. American Journal of Clinical Nutrition 57, 715S–725S. Halliwell B & Dizdaroglu M (1992) The measurement of oxidative damage to DNA by HPLC and GC/MS techniques. Free Radical Research Communications 16, 75–87. Halliwell B & Gutteridge JMC (1989) Free Radicals in Biology and Medicine, 2nd ed. Oxford: Clarendon Press. Hamberg M & Zhang LY (1995) Quantitative determination of 8hydroxyguanine and guanine by isotope dilution mass spectrometry. Analytical Biochemistry 229, 336–344. Hanck A (1982) Tolerance and effects of high doses of ascorbic acid. Dosis facit venenum. International Journal of Vitamin and Nutrition Research, Suppl. 23, 221–238. Harris G, Bashir S & Winyard PG (1994) 7,8-Dihydro-8-oxo20 deoxyguanosine present in DNA is not simply an artefact of isolation. Carcinogenesis 15, 411–413. Hennekens CH, Buring JE, Manson JE, Stampfer M, Rosner B, Cook NR, Belanger C, LaMotte F, Gaziano JM, Ridker PM, Willett W & Peto R (1996) Lack of effect of long-term supplementation with b-carotene on the incidence of malignant neoplasms and cardiovascular disease. New England Journal of Medicine 334, 1145–1149. Herbert KE, Evans MD, Finnegan MTV, Griffiths HR & Lunec J (1996) A novel HPLC procedure for the analysis of 8-oxoguanine in DNA. Free Radicals in Biology and Medicine 3, 467–473. Herbert V & Hacob E (1974) Destruction of vitamin B12 by ascorbic acid. Journal of the American Medical Association 230, 241–242. Hertog MGL (1994) Flavonols and flavones in foods and their relation with cancer and coronary heart disease risk. PhD Thesis, Wageningen University, The Netherlands. Hertog MGL, Hollman PCH & van de Putte B (1993) Content of potentially anticarcinogenic flavonoids in tea infusions, wines and fruit juices. Journal of Agricultural and Food Chemistry 41, 1242–1246. Heywood R, Palmer AK, Gregson RL & Hummler H (1985) The toxicity of b-carotene. Toxicology 36, 91–100. Hirono I, Ueno I, Hosaka S, Takanashi H, Matshushima T, Sugimura T & Natori S (1981) Carcinogenicity examination of quercetin and nutin in ACI rats. Cancer Letters 13, 15–21. Holley AE & Slater TF (1991) Measurement of lipid hydroperoxides in normal human blood plasma using HPLCchemiluminescence linked to a diode array detector for measuring conjugated dienes. Free Radical Research Communications 15, 51–53. Hollman PC, de Vries JH, van Leeuwen SD, Mengelers MJ & Katan MB (1995) Absorption of dietary quercetin glycosides and quercetin in healthy ileostomy volunteers. American Journal of Clinical Nutrition 62, 1276–1282. Hood RL (1995) Tocotrienols and cholesterol metabolism. In Nutrition, Lipids, Health and Disease, pp. 133–137 [ASH Ong, E Niki and L Packer, editors]. Champaign, IL: AOCS Press. Hook EB, Healey KM, Niles AM & Shalko RC (1974) Vitamin E: teratogen or anti-teratogen? Lancet, 809. Inagaki Y, Kinoshita MO, Nakamura Y & Masuda YA (1978) Double-blind controlled study of the efficacy of DL-a-tocopherylnicotinate in patients with vascular disease. In Tocopherol, Oxygen and Biomembranes, pp. 339–349 [C DeDuve and O Hayashi, editors]. Amsterdam and New York: Elsevier– North Holland Biomedical Press. International Agency for Research on Cancer (1983) Quercetin. IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans no. 31, pp. 213–229. Lyon: IARC. Ito N (1992) Is quercetin carcinogenic? Japanese Journal of Cancer Research 83, 312–313. Ito N, Hagiwara A, Tamano S, Kagawa M, Shibata M, Kurata Y & Fukushima S (1989) Lack of carcinogenicity of quercetin in F344/ Du Crj rats. Japanese Journal of Cancer Research 80, 317–325. Jacob RA & Burri BJ (1996) Oxidative damage and defense. American Journal of Clinical Nutrition 63, 985S–990S. Jiang ZY, Woodland ACS & Wolff SP (1992) Lipid hydroperoxide measurement by oxidation of Fe2þ in the presence of xylenol orange. Comparison with the TBA assay and iodometric method. Lipids 26, 853–857. Jideani VAE (1992) Carotene retention in palm oil by mechanised and traditional processes. Journal of Food Science and Technology 29, 68–69. Kallner A, Hartmann D & Hornig D (1979) Steady-state turnover and body pool of ascorbic acid in man. American Journal of Clinical Nutrition 32, 530–539. Kanofsky JR (1989) Singlet oxygen production in biological systems. Chemical and Biological Interactions 70, 1–28. Kappus H & Diplock AT (1992) Tolerance and safety of vitamin E: a toxicological position report. Free Radicals in Biology and Medicine 13, 55–74. Karam LS, Bertgold DS & Simic MG (1991) Biomarkers of OH radical damage in vivo. Free Radical Research Communications 12/13, 11–16. Katusin RB, Matic S, Razem D & Mihokovic V (1988) Radiation decontamination of tea herbs. Journal of Food Science and Technology 53, 1120–1126. Kaugars GE, Silverman S, Lovas JG, Brandt RB, Riley WT, Dao C, Singh VN & Gallo J (1994) A clinical trial of antioxidant supplements in the treatment of oral leukoplakia. Oral Surgery, Oral Medicine, Oral Pathology 78, 462–468. Keller J, Halmes NC, Hinson JA & Humford NR (1993) Immunochemical detection of oxidised proteins. Chemical Research in Toxicology 6, 430–433. Kennedy JF, Rivera ZS, Lloyd LL, Warner FP & Jumel K (1992) Ascorbic acid stability in aseptically processed orange juice in TetraBrik cartons and the effect of oxygen. Food Chemistry 45, 327–331. Kennedy TA & Liebler DC (1992) Peroxyl radical scavenging by b-carotene in lipid bilayers. Journal of Biological Chemistry 267, 4658–4663. Kim JM & White RH (1996) Effect of vitamin E on the anticoagulant response to warfarin. American Journal of Cardiology 77, 545–546. Defence against reactive oxidative species Kiritsakis A (1990) Chemistry of olive oil. In Olive Oil, pp. 25–55. Champaign, IL: American Oil Chemists Society. Kitagawa M & Mino M (1989) Effects of elevated D-a-(RRR)tocopherol dosage in man. Journal of Nutritional Science and Vitaminology 35, 133–142. Klaui H & Pongracz G (1981) Ascorbic acid and its derivatives as antioxidants in oils and fats. In Vitamin C (Ascorbic Acid), pp. 139–166 [JN Counsell and DH Hornig, editors]. London and New Jersey: Applied Science Publishers. Kohlmeier L & Hastings SB (1995) Epidemiologic evidence of a role of carotenoids in cardiovascular disease prevention. American Journal of Clinical Nutrition 62, 1370S–1376S. Kondo T (1996) Parkinsons disease and free radicals. Annals of the New York Academy of Sciences 786, 206–216. Krasavage WJ & Terhaar CJ (1977) D-alpha-tocopheryl-polyethyleneglycol 1000 succinate. Acute toxicity, subchronic feeding, reproduction and teratologic studies in the rat. Journal of the Science of Food and Agriculture 25, 273–278. Kritharides L, Jessup W, Gifford J & Dean RT (1994) A method for defining the stages of oxidation by the separation of cholesterol and cholesterol ester oxidation products using HPLC. Analytical Biochemistry 213, 79–89. Kuhnau J (1976) The flavonoids. A class of semi-essential food components: their role in human nutrition. World Review of Nutrition and Dietetics 24, 117–191. Kushi LH, Fulsom AR, Prineas RJ, Mink PJ, Wu Y & Bostick RM (1996) Dietary antioxidant vitamins and death from coronary heart disease in postmenopausal women. New England Journal of Medicine 334, 1156–1162. Labuza TP (1971) Kinetics of lipid oxidation in foods. CRC Critical Review of Food Technology, 355–405. Lachance PA (1996) Future vitamin and antioxidant RDAs for health promotion. Preventive Medicine 25, 46–47. Lagouri V & Boskou D (1996) International Journal of Food Sciences and Nutrition 47, 493–497. Lakritz L & Thayer DW (1992) Effect of ionising radiation on unesterified tocopherols in fresh chicken breast muscle. Meat Science 32, 257–265. Lakritz L & Thayer DW (1994) Effect of gamma radiation on total tocopherols in fresh chicken breast muscle. Meat Science 37, 439–448. Landrum JT, Bone RA & Kilburn MD (1997) The macular pigment: a possible role in protection from age-related macular degeneration. Advances in Pharmacology 38, 537–556. Lee DS, Chung SK & Yam KL (1992) Carotenoid loss in dried red pepper products. International Journal of Food Science and Technology 27, 179–185. Levander OA & Burk RF (1996) Selenium. In Present Knowledge in Nutrition, pp. 320–328 [EE Ziegler and LJ Filer, editors]. Washington, DC: ILSI Press. Levander OA, Morris VC, Higgs DJ & Varma RN (1973) Nutritional relationships among vitamin E, selenium, antioxidants and ethyl alcohol in the rat. Journal of Nutrition 103, 536–542. Levine M, Conry-Cantilena C, Wang Y, Welch RW, Washko PW, Dhariwal KR, Park JB, Lazarev A, Graumlich JF, King J & Cantilena LR (1996) Vitamin C pharmacokinetics in healthy volunteers: evidence for a recommended dietary allowance. Proceedings of the National Academy of Sciences USA 93, 3704–3709. Levine RL, Garland D, Oliver CN, Amici A, Climent I, Lenz AG, Ahn BW, Shaltiel S & Stadtman ER (1995) Determination of carbonyl content in oxidatively modified protein. Methods in Enzymology 186, 464–487. Levine RL, Williams JA, Stadtman ER & Shacter E (1994) Carbonyl assay for determination of oxidatively modified proteins. Methods in Enzymology 233, 346–357. Loeliger J, Lambelet P, Aeschbach R & Prior E (1996) Natural S109 antioxidants: from radical mechanisms to food stabilisation. In Food Lipids and Health, pp. 68–77 [RE McDonald and DB Min, editors]. New York, NY: Marcel Dekker Inc. Losonczy KG, Harris TB & Havlik RJ (1996) Vitamin E and vitamin C supplement use and risk of all-cause and coronary heart disease mortality in older persons: the Established Populations for Epidemiologic Studies of the Elderly. American Journal of Clinical Nutrition 64, 190–196. Luis A & Navab M (1993) Lipoprotein oxidation and gene expression in the artery wall. Biochemical Pharmacology 46, 2119–2126. Lunec J, Herbert K, Blount S, Podmore ID & Griffiths HR (1994) 8-Hydroxydeoxyguanosine, marker of oxidative DNA damage in systemic lupus erythematosus. FEBS Letters 348, 131–138. Lyras L, Shaw PJ, Evans PJ & Halliwell B (1996) Oxidative damage and motor neurone disease. Difficulties in the measurement of protein carbonyls in human brain tissue. Free Radical Research 24, 399–406. McDonald I & Newell C (1996) Measuring Health. A Guide to Rating Scales and Questionnaires, 2nd ed. Oxford: Oxford University Press. McLarty JW (1992) An intervention trial in high-risk asbestosexposed persons. In The Biology and Prevention of Aerodigestive Tract Cancer, pp. 141–149 [GR Newell and WK Hongs, editors]. New York, NY: Plenum Press Inc. Malins DC & Haimonot R (1991) Major alterations in the nucleotide structure of DNA in cancer of the female breast. Cancer Research 51, 5430–5432. Mangels AR, Block G, Frey CM, Patterson BH, Taylor PR, Norkus EP & Levander OA (1993a) The bioavailability to humans of ascorbic acid from oranges, orange juice and cooked broccoli is similar to that of synthetic ascorbic acid. Journal of Nutrition 123, 1054–1061. Mangels AR, Holden JM, Beecher GR, Forman R & Lanza E (1993b) Carotenoid content of fruits and vegetables: an evaluation of analytic data. Journal of the American Dietetic Association 93, 284–296. Markus M, Prabhudesai M & Wassef S (1980) Stability of vitamin B12 in the presence of ascorbic acid in food and serum: restoration by cyanide of apparent loss. American Journal of Clinical Nutrition 33, 137–143. Maskos Z, Rush JD & Koppenol WH (1992) The hydroxylation of tryptophan. Archives of Biochemistry and Biophysics 296, 514– 520. Matthews-Roth MM (1986) b-Carotene therapy for erythropoietic protoporphyria and other photosensitivity diseases. Biochemie 68, 875–884. Melia AT, Kosstwardy SG & Zhi JG (1996) The effect of Orlistat, an inhibitor of dietary fat absorption on the absorption of vitamins A and E in healthy volunteers. Journal of Clinical Pharmacology 36, 647–653. Meyer AS & Isaksen A (1995) Application of enzymes as food antioxidants. Trends in Food Science and Technology 6, 300– 304. Mitch WE, Johnson MW, Kirshenbaum JM & Lopez RE (1981) Effect of large oral doses of ascorbic acid on uric acid excretion by normal subjects. Clinical Pharmacology and Therapeutics 29, 318–321. Mitchell GE, McLauchlin RL, Isaacs AR, Williams DJ & Nottingham SM (1992) Effect of low dose irradiation on composition of tropical fruits and vegetables. Journal of Food Composition and Analysis 5, 291–311. Mo JY, Maki H & Sekiguchi M (1992) Hydrolytic elimination of mutagenic nucleotide, 8-oxodGTP, by human 18-kilodalton protein, sanitization of nucleotide pool. Proceedings of the National Academy of Sciences USA 89, 11021–11025. Morrow JD, Frei B, Longmire AW, Gaziano JM, Lynch SM, S110 A. T. Diplock et al. Shys Y, Strauss WE, Oakes TA & Roberts LJ (1995) Increase in circulating products of lipid peroxidation (F2 isoprostanes) in smokers. New England Journal of Medicine 332, 1198–1203. Morrow JD & Roberts LJ II (1994) Mass spectrometry of prostanoids: F2 -isoprostanes produced by non-cyclooxygenase free radical-catalyzed mechanism. Methods in Enzymology 233, 163–174. Morrow JD & Roberts LJ (1996) The isoprostanes. Biochemical Pharmacology 51,1–9. Mortin JF (1989) Tannin as carcinogen in bush-tea: tea, mate and khat. In Chemistry and Significance of Condensed Tannins, pp. 403–421 [RW Hemingway and JJ Karchesy, editors]. New York, NY: Plenum Press. National Toxicological Program (1992) NTP Technical Report on the Toxicology and Carcinogenesis Studies of Quercetin in F344/N Rats. National Institutes of Health Publication no. 91. Washington, DC: National Institutes of Health. Muldoon M & Kritchevsky SB (1996) Flavonoids and heart disease. British Medical Journal 312, 458–459. Newmark HL, Scheiner J, Markus M & Prabhudesai M (1976) Stability of vitamin B12 in the presence of ascorbic acid. American Journal of Clinical Nutrition 29, 645–649. Nierenberg DW, Stukel TA, Mott LA & Greenberg ER (1994) Steady-state serum concentration of alpha tocopherol not altered by supplementation with oral beta carotene. The Polyp Prevention Study 1 Group. Journal of the National Cancer Institute 86, 117–120. Niki E (1987) Antioxidants in relation to lipid peroxidation. Chemistry and Physics of Lipids 44, 227–253. Niki E, Kawakami A, Yamamoto Y & Kamiya Y (1985) Oxidation of lipids: VIII. Synergistic inhibition of oxidation of phosphatidylcholine liposome in aqueous dispersion by vitamin E and vitamin C. Bulletin of the Chemical Society of Japan 58, 1971– 1975. Niki E, Tsuchiya J, Tanimura R & Kamiya Y (1982) Regeneration of vitamin E from a-chromanoxyl radical by glutathione and vitamin C. Chemistry Letters 27, 789–792. Norkus EP & Rosso P (1975) Changes in the ascorbic acid metabolism of the offspring following high maternal intake of this vitamin in the pregnant guinea pig. Annals of the New York Academy of Sciences 258, 401–409. Norkus EP & Rosso P (1981) Effects of maternal intake of ascorbic acid on the postnatal metabolism of this vitamin in the guinea pig. Journal of Nutrition 11, 624–630. Olson JA & Krinsky NI (1995) Introduction: the colorful fascinating world of carotenoids: important biological modulators. FEBS Letters 9, 1547–1550. O’Leary V, Darley-Usmar V, Rusell L & Stone D (1992) Prooxidant effects of lipoxygenase-derived peroxides and antioxidants on the copper-initiated oxidation of low density lipoprotein. Biochemical Journal 282, 631–634. Office Life Sciences Research (1979) Evaluation of the health aspects of carotene (b-carotene) as a food ingredient. Federation of European Biochemistry and Societies’ Letters Bethesda Md, USA Contract Number (FDA) 223-75-2004. Olinski R, Zastawny T, Bubzbon J, Foksincki M, Barecki A & Dizdaroglu M (1992) DNA base modifications in chromatin of human cancerous tissues. FEBS Letters 309, 193–198. Oliver CN, Ahn BA & Moerman EJ (1987) Age related changes in oxidised proteins. Journal of Biological Chemistry 262, 5488– 5491. Omenn GS, Goodman GE, Thornquist MD, Balmes J, Cullen MR, Glass A, Keogh JP, Meyskens FL, Valanis B, Williams JH, Barnhardt S & Hammar S (1996) Effects of combination of bcarotene and vitamin A on lung cancer and cardiovascular disease. New England Journal of Medicine 334, 1150–1155. Onyewu PN (1985) Thermal degradation of beta carotene under simulated time and temperature conditions of various food processes. Dissertation Abstracts International B 46, 716. Oshima H, Freisen M, Brouet I & Bartsch H (1990) Nitrotyrosine as a new marker for endogenous nitrosation and nitration of proteins. Food and Chemical Toxicology 28, 647–652. Özer NK, Boscoboinik DO & Azzi A (1995) New roles of lipoproteins and vitamin E in the pathogenesis of atherosclerosis. Biochemistry and Molecular Biology International 35, 117–124. Packer L (1996) Nitric oxide. Part A: sources and detection of NO; NO synthase. Methods in Enzymology 268, 331–340. Padula M & Rodriguez ADB (1987) Changes in individual carotenoids and vitamin C on processing and storage of guava juice. Acta Alimentaria 16, 209–216. Paganga G & Rice-Evans CA (1997) The identification of flavonoids as glycosides in human plasma. FEBS Letters 401, 78–82. Palozza P & Krinsky NI (1992) Antioxidant effects of carotenoids in vivo and in vitro: an overview. Methods in Enzymology 268, 127–136. Parker RS (1989) Dietary and biochemical aspects of vitamin E. Advances in Food Nutrition and Research 33, 157–232. Patterson RLS & Stevenson MH (1995) Irradiation-induced off odour in chicken and its possible control. British Journal of Poultry Science 36, 425–441. Poor CL, Bierer TL, Merchen NR, Fahey GC & Erdman J (1993) The accumulation of a and b carotene in serum and tissues of preruminant calves fed raw and steamed carrot slurries. Journal of Nutrition 123, 1296–1304. Pothakamury UR & Barbaosa-Canovas GV (1995) Fundamental aspects of controlled release in foods. Trends in Food Science and Technology 6, 397–406. Potter NN & Hotchkiss JH (1995) Food Science. New York, NY: Chapman and Hall. Pryor WA, Squadrito GL & Friedman M (1995) The cascade mechanism to explain ozone toxicity: the role of lipid ozonation products. Free Radicals in Biology and Medicine 19, 935–941. Quackenbush FW (1987) Reverse phase HPLC separation of cisand trans-carotenoids and its application to b-carotenes in food materials. Journal of Liquid Chromatography 10, 643–653. Ranhota GS, Gelroth J, Langemeier A & Rogers DE (1995) Stability and contribution of beta carotene added to whole wheat bread and crackers. Cereal Chemistry 72, 139–141. Ravanat JL, Turesky RJH, Gremaud E, Trudel LJ & Stadler RH (1995) Determination of 8-oxoguanine in DNA by gas chromatography–mass spectrometry and HPLC–electrochemical detection: overestimation of the background level of the oxidised base by the gas chromatography–mass spectrometry assay. Chemical Research in Toxicology 8, 1039–1045. Reaven PD & Witztum JL (1996) Oxidized low density lipoproteins in atherogenesis: role of dietary modification. Annual Reviews in Nutrition 16, 51–71. Rhead WA & Schrauzer GN (1971) Risks of long-term ascorbic acid overdosage. Nutrition Reviews 29, 262–263. Ribaya-Mercado JD, Ordovas JM & Russell RM (1995) Effect of b-carotene supplementation on the concentrations and distribution of carotenoids, vitamin E, vitamin A and cholesterol in plasma lipoprotein and non-lipoprotein fractions in healthy older women. Journal of the American College of Nutrition 14, 614– 620. Rice-Evans C, Diplock AT & Symons MCR (1991) Techniques in Free Radical Research, pp. 143–147. Amsterdam, London and New York: Elsevier. Rice-Evans C, Leake D, Bruckdorfer KR & Diplock AT (1996) Practical approaches to low density lipoprotein oxidation: whys, wherefores and pitfalls. Free Radical Research 25, 285–311. Rice-Evans C & Miller NJ (1996) Antioxidant activities of flavonoids as bioactive components of food. Biochemical Society Transactions 24, 790–795. Defence against reactive oxidative species Rice-Evans C, Miller NJ, Bolwell PG, Bramley PM & Pridham JB (1995) The relative antioxidant activity of plant-derived polyphenolic flavonoids. Free Radical Research 22, 375–383. Rimm EB, Stampfer MJ, Ascherio A, Giovannucci E, Colditz GA & Willett WC (1993) Vitamin E consumption and the risk of coronary heart disease in men. New England Journal of Medicine 328, 1450–1456. Rivers JM (1989) Safety of high-level vitamin C ingestion. International Journal of Vitamin and Nutrition Research 30, Suppl., 95–102. Rock CL, Jacob RA & Bowen PE (1996) Update on the biological characteristics of the antioxidant micronutrients: vitamin C, vitamin E, and the carotenoids. Journal of the American Dietetics Association 96, 693–702. Rodriguez MS & Irwin MI (1972) A conspectus of research on vitamin A requirements in man. Journal of Nutrition 102, 909–968. Rommel A & Wrolstad RE (1993a) Ellagic acid content of red raspberry juice as influenced by cultivar, processing and environmental factors. Journal of Agricultural and Food Chemistry 41, 1951–1960. Rommel A & Wrolstad RE (1993b) Composition of flavonols in red raspberry juice as influenced by cultivar, processing and environmental factors. Journal of Agricultural and Food Chemistry 41, 1941–1950. Rowland I (1993) Dietary modification of the activity of food mutagens. Mutagenesis 8, 476. Sahu SC (1992) Effect of ascorbic acid and curcumin on quercetin induced nuclear DNA damage, lipid peroxidation and protein degradation. Cancer Letters 83, 237–241. Sahu SC & Gray GC (1993) Interaction of flavonoids, trace metals and oxygen. Nuclear damage and lipid peroxidation induced by myricetin. Cancer Letters 70, 73–79. Sahu SC & Gray GC (1994) Kaempferol-induced nuclear DNA damage and lipid peroxidation. Cancer Letters 85, 159–164. Sakumi K, Furuichi M, Tsuzuki T & Kakuma T (1993) Cloning and expression of cDNA for a human enzyme that hydrolyzes 8oxo-dGTP, a mutagenic substrate for DNA synthesis. Journal of Biological Chemistry 268, 23524–23530. Sandmeier D (1996) Prooxidative Effekte von Gewürzen unter Lichteinfluss (Prooxidative effects on spices under the influence of light). Fett/Lipid 98, 199–202. Sarkkinen ES, Uusitupa MI, Nyyssonen K, Parviainen M, Pentitila I & Salonen J (1993) Effects of two low-fat diets, high and low in polyunsaturated fatty acids, on plasma lipid peroxides and serum vitamin E levels in free-living hypercholesterolaemic men. European Journal of Clinical Nutrition 47, 623–630. Schalch W (1992) Carotenoids in the retina – a review of their possible role in preventing or limiting change caused by light and oxygen. In Free Radicals and Ageing, pp. 280–289 [I Emerit and B Chance, editors]. Basel: Birkhauser Verlag. Schmidt KH, Hagmaier V, Hornig DH, Vuilleumier JP & Rutishauser G (1981) Urinary oxalate excretion after large intakes of ascorbic acid in man. American Journal of Clinical Nutrition 34, 305–311. Schreck R & Baeuerle PA (1994) Assessing oxygen radicals as mediators in activation of inducible eukariotic transcription factor NF-kB. Methods in Enzymology 234, 151–163. Seddon JM, Ajani UA, Sperduto RD, Hiller R, Blair N, Burton TC, Farber MD, Gragoudas ES, Haller J, Miller DT, Yannuuzzi LA & Willett WC (1994a) Dietary carotenoids, vitamins A, C, and E, and advanced age-related macular degeneration. Journal of the American Medical Association 272, 1413–1420. Seddon JM, Christen WG, Manson JE, LaMotte FS, Glynn RJ, Buring JE & Hennekens CH (1994b) The use of vitamin supplements and the risk of cataract among US male physicians. American Journal of Public Health 84, 788–792. S111 Sen CK & Packer L (1996) Antioxidant and redox regulation of gene transcription. FEBS Letters 10, 709–720. Shamberger RJ, Corlett CL, Beaman KD & Kasten BL (1979) Antioxidants reduce the mutagenic effect of malonaldehyde and b-propiolactone. Part IX. Antioxidants and cancer. Mutation Research 66, 349–355. Sharoni Y & Levi J (1996) Anticarcinogenic properties of lycopene. In Natural Antioxidants and Food Quality in Atherosclerosis and Cancer Prevention, pp. 378–385 [JT Kumpulainen and J Salonen, editors]. Cambridge: Royal Society of Chemistry. Shigenaga MK, Aboujaoude EN, Chen Q & Ames BN (1994) Assays of oxidative DNA damage biomarkers. Methods in Enzymology 234, 16–33. Sian NK & Soleha I (1991) Carotenoid and anthocyanin contents of papaya and pineapple: influence of blanching and predrying treatments. Food Chemistry 39, 175–185. Sies H (1991) Oxidative stress: from basic research to clinical application. American Journal of Medicine 91, 31S–39S. Sies H (1993) Strategies of antioxidant defence. European Journal of Biochemistry 215, 213–219. Sies H (1997) Antioxidants in Disease Mechanisms and Therapy. London: Academic Press. Sies H & Stahl W (1995) Vitamins E and C, b-carotene, and other carotenoids as antioxidants. American Journal of Clinical Nutrition 62, 1315S–1321S. Simonian NA & Coyle JT (1996) Oxidative stress in neurodegenerative diseases. Annual Reviews of Pharmacology and Toxicology 36, 83–106. Singleton VL & Kratzer FH (1989) Toxicity and related physiological activity of phenolic substances of plant origin. Journal of Agricultural and Food Chemistry 17, 497–512. Sokol RJ (1996) Vitamin E. In Present Knowledge in Nutrition, pp. 130–136 [EE Ziegler and LJ Filer, editors]. Washington, DC: ILSI Press. Spanos GA & Wrolstad RE (1990) Influence of processing and storage on the phenolic compositions of Thompson seedless grape juice. Journal of Agricultural and Food Chemistry 38, 565–571. Spilker B (1996) Quality of Life and Pharmacoeconomics in Clinical Trials, 2nd ed. London: Lippincott Raven. Springfield JR & Levitt MD (1994) Pitfalls in the use of breath pentane measurements to assess lipid peroxidation. Journal of Lipid Research 35, 1497–1504. Stahl W & Sies H (1992) Uptake of lycopene and its geometrical isomers is greater from heat processed than from unprocessed tomato juice in humans. Journal of Nutrition 122, 2161–2166. Stahl W & Sies H (1993) Physical quenching of singlet-oxygen and cis–trans isomerization of carotenoids. Annals of the New York Academy of Sciences 691, 10–19. Stampfer MJ, Hennekens CH, Manson JE, Colditz GA, Rosner B & Willett WC (1993) Vitamin E consumption and the risk of coronary disease in women. New England Journal of Medicine 328, 1444–1449. Stampfer MJ, Willett W, Castelli WP, Taylor JO, Fine J & Hennekens VCH (1983) Effect of vitamin E on lipids. American Journal of Clinical Nutrition 79, 714–716. Stavric B (1994) Quercetin in our diet: from potent mutagen to probable anticarcinogen. Clinical Biochemistry 27, 245–248. Stavric B, Mutula TI, Klassen R & Downie RH (1990) Inhibitory effect of flavonoids against mutagenic and carcinogenic xenobiotics in foods. In Flavonoids in Biology and Medicine. III Current Issues in Flavonoid Research, pp. 515–529 [NP Das, editor]. Singapore: National University of Singapore. Steinberg D (1993) Antioxidant vitamins and coronary heart disease (editorial; comment). New England Journal of Medicine 328, 1487–1489. S112 A. T. Diplock et al. Steinmetz A & Potter JD (1991) Review: vegetables, fruit, and cancer. 1. Epidemiology. Cancer Causes and Control 2, 325–357. Stephens NG, Parsons A, Schofield PM, Kelly F, Cheeseman K & Mitchinson MJ (1996) Randomised controlled trial of vitamin E in patients with coronary disease: Cambridge Heart Antioxidant Study (CHAOS). Lancet 347, 781–786. Stillwell WG, Xu HX, Adkins JA, Cramp FT & Jenski LJ (1989) Analysis of methylated and oxidised purines in urine by capillary gas chromatography–mass spectrometry. Chemical Research in Toxicology 2, 94–99. Stocker R, Bowry V & Frei B (1991) Ubiquinol-10 protects human low density lipoprotein more efficiently against lipid peroxidation than does alpha tocopherol. Proceedings of the National Academy of Sciences USA 88, 1646–1650. Street DA, Comstock GW, Salkeld RM, Schuep W & Klag MJ (1994) Serum antioxidants and myocardial infarction. Are low levels of carotenoids and a-tocopherol risk factors for myocardial infarction? Circulation 90, 1154–1161. Sugimura T, Nagao M, Matshushima T, Yahagi T, Selno Y, Shiral A, Sawamura M, Natori S, Yoshihira K, Fukuoka M & Kuroyanagi M (1977) Mutagenicity of flavone derivatives. Proceedings of the Japanese Academy of Science B 853, 194–197. Sweeney JP & Marsh AC (1974) Liver storage of vitamin A by rats fed carrots in various forms. Journal of Nutrition 104, 115–120. Takamatsu S, Takamatsu M, Satoh K, Imaizumi T, Yoshida H, Hiramoto M, Koyama M, Ohgushi Y & Mizuno S (1995) Effects on health of dietary supplementation with 100 mg d-alphatocopheryl acetate daily for 6 years. Journal of International Medical Research 23, 342–357. Taylor A (1993) Cataract: relationship between nutrition and oxidation. Journal of the American College of Nutrition 12, 138–146. Taylor-Mayne S (1996) b-Carotene, carotenoids, and disease prevention in humans. FASEB Journal 10, 690–701. Ternes W, Gronemeyer M & Schwartz K (1995) Determination of pcymene-2,3-diol, thymol, and carvacrol in different foodstuffs. Zeitschrift Lebensmittel Unterschrift Forschungen 201, 544–547. Thomas SM, Jessup W, Gebicki JM & Dean RT (1989) A continuous-flow automated assay of iodometric estimation of hydroperoxides. Analytical Biochemistry 176, 353–359. Timberlake CF (1988) The biological properties of anthocyanins. National Quarterly Information Bulletin, 4–15. Traber MG & Sies H (1996) Vitamin E in humans: demand and delivery. Annual Reviews of Nutrition 16, 321–347. Trägardh G (1995) New developments in membrane processing. In Food Processing: Recent Developments, pp. 231–237 [AG Gaonkar, editor]. Amsterdam: Elsevier Science BV. Truscott TG (1990) The photophysics and photochemistry of the carotenoids. Journal of Photochemistry and Photobiology 6, 359–371. Tsai AC, Kelly JJ, Peng B & Cook N (1978) Study on the effect of megavitamin supplementation in man. American Journal of Clinical Nutrition 31, 831–838. Tsuda T, Mizuno K & Ohshima K, Kawakishi S & Osawa T (1995) Supercritical carbon dioxide extraction of antioxidative components from tamarind (Tamarindus indica L) seed coat. Journal of Agricultural and Food Chemistry 43, 2803–2806. Turk PW & Weitzman SA (1995) Free radical DNA adduct, 8-hydroxydeoxyguanosine affects activity of HPAII and MSPI restriction endonucleases. Free Radical Research 23, 255–258. Uchida K & Kawakishi S (1993) 2-Oxo-histidine as a novel biological marker for oxidatively modified proteins. FEBS Letters 332, 208–210. Uchida K & Kawakishi S (1994) Identification of oxidised histidine generated at the active site of Cu, Zn-superoxide dismutase exposed to H2 O2 . Journal of Biological Chemistry 269, 2405–2410. Unilever (1995) Diet and Health News, vol. 1. Vlardingen: Unilever Research Laboratories. van Poppel G & Goldbohm RA (1995) Epidemiologic evidence for b-carotene and cancer prevention. American Journal of Clinical Nutrition 62, Suppl., 1393S–1402S. van Poppel G, Kardinaal A, Princen H & Kok FJ (1994) Antioxidants and coronary heart disease. Annals of Medicine 26, 429–434. Velthuis-te Wierik EJ, Van den Berg H, Weststrate JA, Van het Hof KH & deGraaf C (1996) Consumption of reduced-fat products: effects on parameters of antioxidative capacity. European Journal of Clinical Nutrition 50, 214–219. Wandzilak TR, D’Andre SD, Davis PA & Williams HE (1994) Effect of high dose vitamin C on urinary oxalate levels. Journal of Urology 151, 834–837. Wang XD (1994) Review: absorption and metabolism of b-carotene. Journal of the American College of Nutrition 13, 314–325. Wang Z, Ciabattoni G & Creminon C (1995) Immunological characterization of urinary 8-epi-prostaglandin F2a excretion in man. Journal of Pharmacology and Experimental Therapeutics 275, 94–100. Weber P, Bendich A & Schalch W (1996) Vitamin C and human health – a review of recent data relevant to human requirements. International Journal of Vitamin and Nutrition Research 66, 19–30. Weissberger LH & Harris PL (1943) Effect of tocopherols on phosphorus metabolism. Journal of Biological Chemistry 151, 543–551. Weldon GH, Bhatt A, Keller P & Hummler H (1983) dl-aTocopheryl acetate (vitamin E): a long-term toxicity and carcinogenicity study in rats. International Journal of Vitamin and Nutrition Research 53, 287–296. Wells-Knecht MC, Huggins TG, Dyer DG, Thorpe SR & Baynes JW (1993) Oxidised amino acids in lens proteins with age. Measurement of o-tyrosine and dityrosine in the ageing human lens. Journal of Biological Chemistry 268, 12348–12352. Willett WC, Stampfer MJ, Underwood BA, Taylor JO & Hennekens CH (1983) Vitamins A, E and carotene: effect of supplementation on their plasma levels. American Journal of Clinical Nutrition 38, 559–566. Wiseman H & Halliwell B (1996) Damage to DNA by reactive oxygen and nitrogen species: role in inflammatory disease and progression to cancer. Biochemical Journal 313, 17–29. Wolff SP (1994) Ferrous ion oxidation in the presence of the ferric ion indicator xylenol orange for measurement of hydroperoxides. Methods in Enzymology 233, 182–189. World Health Organization (1982) Anthocyanins. Toxicological Evaluation of Certain Food Additives. Technical Report Series no. 17, pp. 42–49. Geneva: WHO. Xu MJ, Plezia PM, Alberts DS, Emerson SS, Peng YM, Sayers SM, Liu Y, Ritenbaugh C & Gensler HL (1992) Reduction in plasma or skin a-tocopherol concentration with long-term oral administration of b-carotene in humans and mice. Journal of the National Cancer Institute 84, 1559–1565. Yamamoto Y (1994) Chemiluminescence-based high performance liquid chromatography assay of lipid hydroperoxides. Methods in Enzymology 233, 319–324. Yamana T, Nakatani H, Kiknoka B, Matsumoto H, Iwata Y, Kitao Y, Oya K & Takahashi T (1996) Inhibitory effects and toxicity of green tea polyphenols for gastrointestinal carcinogenesis. Cancer 8, 1662–1667. Yang NY & Desai ID (1977) Effect of high levels of dietary vitamin E on hematological indices and biochemical parameters in rats. Journal of Nutrition 107, 1410–1417. # Nutrition Society 1998 British Journal of Nutrition (1998), 80, Suppl. 1, S113–S146 S113 Functional food science and the cardiovascular system G. Hornstra1*, C. A. Barth2, C. Galli3, R. P. Mensink4, M. Mutanen5, R. A. Riemersma6, M. Roberfroid7, K. Salminen8, G. Vansant9 and P. M. Verschuren10 1 Department of Human Biology, Maastricht University, PO Box 616, NL-6200 MD, Maastricht, The Netherlands 2 German Institute for Human Nutrition, Stiftung des Öffentlichen Rechts, Arthur-Scheunert-Allee 114-116, D-14558 Bergholz-Rehbrücke, Germany 3 Institute of Pharmacological Sciences, University of Milano, Via Balzaretti 9, I-20133 Milan, Italy 4 Nutrition Research Centre, Department of Human Biology, Maastricht University, PO Box 616, NL-6200 MD, Maastricht, The Netherlands 5 Department of Applied Chemistry and Microbiology, University of Helsinki, PO Box 27, SF-00014 Helsinki, Finland 6 Cardiovascular Research Unit, Hugh Robinson Building, University of Edinburgh, George Square, Edinburgh EW8 9XF, UK 7 UCL, Ecole de Pharmacie, Tour Van Helmont, Avenue E. Mounier, B-1200 Brussels, Belgium 8 Research and Development, Valio Ltd, PB 390, SF-00101 Helsinki, Finland 9 Laboratory voor Experimentele geneeskunde endocrinologie (LEGENDO), Katholieke Universiteit Leuven, Gasthuisberg, B-3000 Leuven, Belgium 10 Unilever Research Laboratory, Olivier van Noortlaan 120, NL-3133 AT Vlaardingen, The Netherlands Contents 1. Some aspects of coronary heart disease (CHD) aetiology 1.1. Lipoprotein metabolism 1.2. Arterial thrombosis 1.3. Immunological interactions 1.4. Hypertension 1.5. Insulin resistance 1.6. Hyperhomocysteinaemia 2. Dietary components and serum lipoproteins 2.1. Effects of dietary components on fasting lipid and lipoprotein concentrations 2.1.1. Fatty acids 2.1.2. Fat replacers 2.1.3. Soyabean protein preparations 2.1.4. Mono- and disaccharides 2.1.5. Resistant starch 2.1.6. Ethanol 2.1.7. Dietary cholesterol 2.1.8. Fibre 2.1.9. Phytosterols 2.1.10. Tocopherols and tocotrienols 2.1.11. Garlic 2.1.12. Other components 2.2. Postprandial effects 2.3. Gene–diet interaction 2.4. Possible mechanisms of dietary fats 2.4.1. Concept of Spady and colleagues 2.4.2. Concept of Hayes and colleagues 3. Some diet effects on arterial thrombotic processes: platelet and endothelial cell functions, blood coagulation and fibrinolysis 3.1. S115 S115 S115 S115 S115 S115 S115 S116 S116 S116 S117 S117 S117 S117 S117 S117 S118 S118 S118 S118 S118 S118 S118 S119 S119 S119 S119 Arterial thrombosis and cardiovascular disease 3.2. Platelet function as a marker for CHD 3.3. Some diet effects on arterial thrombogenesis and platelet function 3.3.1. Fatty acids 3.3.2. Antioxidants and platelet function 3.4. Endothelial cell function 3.4.1. Dietary fatty acids and endothelial cell function 3.5. Coagulation and fibrinolysis 3.5.1. Effect of dietary factors on coagulation and fibrinolysis 4. Immune-mediated processes underlying CHD 4.1. Immunocompetent cells involved in the atherosclerotic lesion 4.1.1. Endothelial cells 4.1.2. Smooth-muscle cells 4.1.3. Immunocompetent leucocytes 4.1.4. Mechanisms for the recruitment of blood cells in arterial lesions 4.2. The immune system response modulates atherosclerosis progression 4.3. Nutrition and the immunological aspects of atherosclerosis 4.3.1. Effects of n-3 fatty acids on cellular immune response and inflammatory events in atherogenesis 4.3.2. Antioxidants 5. Diet, hypertension and heart function 5.1. Aetiology of hypertension S119 S119 S120 S120 S121 S121 S121 S122 S122 S123 S123 S123 S124 S124 S124 S124 S124 S125 S125 S126 S126 Abbreviations: ALA, a-linolenic acid; apo, apolipoprotein; AT-III, antithrombin-III; DHA, docosahexaenoic acid; EPA, eicosapentaenoic acid; FVIIc, factor VII coagulant activity; ICAM, intracellular adhesion molecule; Lp(a), lipoprotein(a); MI, myocardial infarction; MTHFR, 5,10-methylenetetrahydrofolate reductase; NIDDM, non-insulin-dependent diabetes mellitus; oxLDL, oxidized LDL; PAI-1, plasminogen activator inhibitor-1; PG, prostaglandin; P : S, polyunsaturated : saturated fatty acid ratio; SMC, smooth-muscle cells; TGF-b, transforming growth factor-b; t-PA, plasminogen activator; Tx, thromboxane; VCAM, vascular cell adhesion molecule; vWf, von-Willebrand factor. *Corresponding author: Dr G. Hornstra, fax +31 43 367 0976, email G. [email protected] S114 G. Hornstra et al. 5.1.1. Membrane function 5.1.2. Role of humoral mediators 5.1.3. Insulin resistance 5.1.4. Environmental factors 5.2. Strategies to reduce CHD by lowering blood pressure 5.2.1. Intervention trials 5.2.2. Individual v. population approach 5.3. Dietary fatty acid composition and hypertension 5.4. Heart function 5.4.1. Effect of diet 5.5. Function of the ischaemic heart 5.5.1. Dietary fatty acids and arrhythmia 6. Insulin resistance, obesity and non-insulin-dependent diabetes mellitus 6.1. Insulin resistance, cardiovascular disease and cardiovascular risk factors 6.1.1. Lipid abnormalities and insulin resistance 6.1.2. Hypertension and insulin resistance 6.2. Nutritional aspects 7. Hyperhomocysteinaemia and cardiovascular risk 7.1. Causes of hyperhomocysteinaemia 7.2. Athero-thrombotic mechanisms of hyperhomocysteinaemia 7.2.1. Interaction with lipoproteins 7.2.2. Smooth muscle cell proliferation 7.2.3. Endothelial functions S126 S126 S126 S126 S127 S127 S127 S127 S127 S128 S128 S128 S128 S129 S129 S130 S130 S131 S131 S131 S131 S131 S131 7.2.4. Function of blood platelets 7.2.5. Coagulation and natural anticoagulants 7.2.6. Fibrinolysis 7.2.7. Altered gene expression 7.3. Hyperhomocysteinaemia or reduced B-vitamin status of primary importance in cardiovascular risk? 7.4. Dietary B-vitamins lower plasma homocysteine 8. Critical assessment of the science base 8.1. Identification of criteria 8.2. Evaluation of the present knowledge base with respect to food functionality 8.2.1. Plasma lipoproteins 8.2.2. Arterial thrombosis 8.2.3. Immunological interactions 8.2.4. Hypertension 8.2.5. Insulin resistance 8.2.6. Hyperhomocysteinaemia 9. Conclusions and recommendations for further research 9.1. Plasma lipoproteins 9.2. Arterial thrombosis 9.3. Immunological interactions 9.4. Hypertension 9.5. Insulin resistance 9.6. Hyperhomocysteinaemia Abstract Cardiovascular disease has a multifactorial aetiology, as is illustrated by the existence of numerous risk indicators, many of which can be influenced by dietary means. It should be recalled, however, that only after a cause-and-effect relationship has been established between the disease and a given risk indicator (called a risk factor in that case), can modifying this factor be expected to affect disease morbidity and mortality. In this paper, effects of diet on cardiovascular risk are reviewed, with special emphasis on modification of the plasma lipoprotein profile and of hypertension. In addition, dietary influences on arterial thrombotic processes, immunological interactions, insulin resistance and hyperhomocysteinaemia are discussed. Dietary lipids are able to affect lipoprotein metabolism in a significant way, thereby modifying the risk of cardiovascular disease. However, more research is required concerning the possible interactions between the various dietary fatty acids, and between fatty acids and dietary cholesterol. In addition, more studies are needed with respect to the possible importance of the postprandial state. Although in the aetiology of hypertension the genetic component is definitely stronger than environmental factors, some benefit in terms of the development and coronary complications of atherosclerosis in hypertensive patients can be expected from fatty acids such as a-linolenic acid, eicosapentaenoic acid and docosahexaenoic acid. This particularly holds for those subjects where the hypertensive mechanism involves the formation of thromboxane A2 and/or a1-adrenergic activities. However, large-scale trials are required to test this contention. Certain aspects of blood platelet function, blood coagulability, and fibrinolytic activity are associated with cardiovascular risk, but causality has been insufficiently proven. Nonetheless, well-designed intervention studies should be initiated to further evaluate such promising dietary components as the various n-3 and n-6 fatty acids and their combination, antioxidants, fibre, etc. for their effect on processes participating in arterial thrombus formation. Long-chain polyenes of the n-3 family and antioxidants can modify the activity of immunocompetent cells, but we are at an early stage of examining the role of immune function on the development of atherosclerotic plaques. Actually, there is little, if any, evidence that dietary modulation of immune system responses of cells participating in atherogenesis exerts beneficial effects. Although it seems feasible to modulate insulin sensitivity and subsequent cardiovascular risk factors by decreasing the total amount of dietary fat and increasing the proportion of polyunsaturated fatty acids, additional studies on the efficacy of specific fatty acids, dietary fibre, and low-energy diets, as well as on the mechanisms involved are required to understand the real function of these dietary components. Finally, dietary supplements containing folate and vitamins B6 and/or B12 should be tested for their potential to reduce cardiovascular risk by lowering the plasma level of homocysteine. S131 S132 S132 S132 S132 S133 S133 S133 S133 S133 S134 S134 S134 S134 S135 S135 S135 S135 S135 S136 S136 S137 Cardiovascular system 1. Some aspects of coronary heart disease (CHD) aetiology Major health risks with respect to the cardiovascular system are CHD and hypertension. In addition, cardiovascular complications of diabetes mellitus are important in this respect. The main aetiological processes involved comprise disturbances in lipoprotein metabolism, a prothrombotic shift in the arterial thrombogenic balance, derangements of the immunological system, insulin resistance and hyperhomocysteinaemia. In the present paper, the potential effects of nutritive and non-nutritive food components on these processes will be reviewed. Epidemiological studies have provided important information on the factors involved in the aetiology of CHD, which has been used as a basis for preventive strategies. Nethertheless, only approximately 50 % of the incidence of cardiovascular disease can be explained by the major risk factors, leaving space for substantial, largely unexplored, contributing factors. S115 mostly present in an inactive form and become activated as a result of vascular injury. Platelets undergoing a series of biochemical and morphological changes express proteins and cell receptors, adhere and form aggregates, and bind to neutrophils and monocytes. Similarly, endothelial cells express intercellular adhesion molecules after stimulation. 1.3. Immunological interactions The atherosclerotic lesion is associated with multiple interactions between immuno-competent cells in the blood (monocytes, T-lymphocytes and platelets) together with the two major cell types in the artery wall, endothelial cells and smooth-muscle cells (SMC). Thus, circulating blood monocytes and T-lymphocytes interact with ‘damaged’ endothelium, enter the subendothelial space of the artery wall and release bioactive molecules. SMC can produce and secrete proteoglycans and facilitate the formation of connective tissue and, thereby, contribute to the formation of advanced atherosclerotic lesions which are also promoted by a thrombotic process (McGill, 1984). 1.1. Lipoprotein metabolism The pathological changes in the coronary arteries that lead to the development of atherosclerotic plaques are now better understood. One of the earliest changes may be endothelial dysfunction (Healy, 1990) followed by the development of fatty streaks due to the formation and accumulation of oxidized lipoprotein particles in the subendothelial space (Steinberg et al. 1989). A critical role for antioxidant vitamins, such as a-tocopherol, ascorbic acid and (perhaps) b-carotene in the prevention of endothelial dysfunction and/ or LDL oxidation has been hypothesized (Gey et al. 1993). Many studies have found an association between serum lipoprotein concentrations and the risk of CHD. Associations, however, do not necessarily reflect a causal relationship; such a relationship can only be established by wellcontrolled intervention studies. Formally, causality has only been proven for the positive relationship between LDLcholesterol levels and the risk of CHD (Frick et al. 1987). However, strong evidence also exists that a high concentration of HDL-cholesterol or a low LDL- (or total) : HDLcholesterol ratio protects against CHD (Shaten et al. 1991; Castelli et al. 1992). Further, raised fasting triacylglycerol (Hokanson & Austin, 1996) and lipoprotein(a) (Lp(a)) concentrations (Bostom et al. 1996), as well as the presence of small LDL particles (Austin, 1992) and postprandial lipidaemia (Karpe et al. 1994) may be positively associated with an increased CHD risk. 1.2. Arterial thrombosis Arterial thrombosis starts within seconds after vascular damage and involves the participation of blood platelets and leucocytes, and of coagulation and fibrinolysis. The process results in the formation of mural, embolizing and, ultimately, occlusive thrombi, thereby promoting the progress of atherosclerotic disease, tissue and organ infarction and sudden death (Fuster et al. 1990). Under normal physiological conditions, the cellular components and proteins involved (e.g. proenzymes and procofactors) are 1.4. Hypertension CHD is strongly related to both systolic and diastolic blood pressure in a graded fashion as demonstrated in an analysis of nine large prospective studies (MacMahon et al. 1990). There are multiple causes for primary hypertension with a strong genetic component. Treatment of hypertension and isolated hypertension (without an increased diastolic blood pressure) results in a reduction in coronary disease-related events (Collins et al. 1990). Increased blood pressure per se appears to increase atherosclerosis, presumably by promoting the entry of LDL into the subendothelial space (Curmi et al. 1990). Haemodynamic factors appear to play a role and it is a wellknown fact that certain arteries and sites near vessel bifurcations have a predilection to develop atherosclerosis. 1.5. Insulin resistance Although the phenomenon of insulin resistance has been known for a long time (Himsworth, 1936), the link with atherogenesis was first hypothesized by Stout & VallanceOwen (1969). This theory was revived by Reaven (1988) when he proposed the existence of a syndrome characterized by obesity, hypertension, dyslipidaemia and glucose intolerance, in which insulin resistance was the common link (metabolic syndrome X). Since then, several excellent reviews regarding the link between insulin resistance, metabolic abnormalities and diseases have been published (Ferrannini et al. 1991; Elliott & Viberti, 1993; Laws & Reaven, 1993; Desprès & Marette, 1994). 1.6. Hyperhomocysteinaemia Compelling evidence is now available suggesting that homocyst(e)ine is implicated in cardiovascular disease. This view is based on a large number of epidemiological studies, recently summarized by Boushey et al. (1995) and by Malinow (1996) and supporting the hypothesis that the S116 G. Hornstra et al. plasma homocysteine concentration is an independent graded risk indicator for arteriosclerotic vascular diseases (coronary, cerebral, and peripheral arterial occlusive diseases, as well as carotid thickening). From the meta-analysis by Boushey et al. (1995) a total of 10 % of the population’s coronary artery disease risk was suggested to be attributable to homocysteine. Recent studies by Tonstad et al. (1996) demonstrate that a modest elevation in plasma homocysteine level in children is related to premature cardiovascular death in their male relatives and may partly account for the contribution of family history to risk of cardiovascular disease. 2. Dietary components and serum lipoproteins The association between serum lipoprotein concentrations and the risk of CHD is generally acknowledged. As diet plays an important role in the modulation of lipoprotein metabolism, the purpose of this section is to briefly summarize effects of various dietary components on lipoprotein metabolism. Attention will also be given to the role of the genetic background of individuals in modulating these dietary effects. 2.1. Effects of dietary components on fasting lipid and lipoprotein concentrations 2.1.1. Fatty acids. For the purpose of this discussion, the fatty acids are categorized into four classes: saturated fatty acids, monounsaturated fatty acids (mainly oleic acid, 18 : 1n-9), polyunsaturated fatty acids (mainly linoleic acid, 18 : 2n-6), and trans fatty acids (mainly 18 : 1trans). In a meta-analysis of twenty-seven well-controlled dietary studies (Mensink & Katan, 1992), it was found that, relative to an isoenergetic amount of carbohydrates, a mixture of saturated fatty acids increases LDL-cholesterol concentrations. Polyunsaturated fatty acids, however, lower LDL-cholesterol, but to a lesser extent, as estimated by Keys et al. (1965b). The effect of oleic acid was between that of carbohydrates and polyunsaturated fatty acids. Further, it was demonstrated that all fatty acids increase HDL-cholesterol, but this effect appeared to diminish with increasing unsaturation of the fatty acid. Therefore, it was concluded that under isoenergetic metabolic-ward conditions, the most favourable lipoprotein risk profile for CHD was achieved if saturated fatty acids were replaced by unsaturated fatty acids. However, no distinction was made between the effects of the individual saturated fatty acids. As already indicated by the studies of Keys et al. (1965b), the cholesterolaemic effects of the various saturated fatty acids are not equal. It was therefore suggested that saturated fatty acids should be divided into those with less than twelve C atoms, those with twelve to sixteen C atoms (lauric acid, 12 : 0, myristic acid, 14 : 0, and palmitic acid, 16 : 0) and those with eighteen C atoms (stearic acid, 18 : 0). For statistical reasons, the data from the meta-analysis do not allow estimation of the impact of all the various saturated fatty acids, but it is possible to calculate the separate effects of palmitic and stearic acids, the two most abundant saturated fatty acids in the diet. In agreement with the findings of others, these analyses clearly show that, in contrast with other saturated fatty acids, stearic acid affects neither serum LDL- nor HDL-cholesterol levels (Fig. 1). Other studies suggest that lauric and myristic acids are more cholesterolaemic than palmitic acid, due to increases in both LDL- and HDL-cholesterol (Zock et al. 1994; Temme et al. 1996). In a recent study, it was demonstrated that a mixture of caprylic acid (8 : 0) and capric acid (10 : 0), two mediumchain fatty acids (which are fatty acids with a chain length between four and ten C atoms) slightly increased LDLcholesterol relative to oleic acid, and had no effect on HDLcholesterol (Cater et al. 1997). Taking these studies together, it appears that, despite different effects on HDLand LDL-cholesterol, all saturated fatty acids, including stearic acid, increase the LDL : HDL ratio to a comparable degree. Fig. 1. Relative effects of palmitic acid (f), stearic acid (,), cis-monounsaturated fatty acids (*) and cis-polyunsaturated fatty acids (j) on fasting serum lipid and lipoprotein concentrations in human subjects. (From Mensink & Katan, 1992.) Cardiovascular system Results from various studies have shown that trans monoenoic acids increase LDL- and decrease HDLcholesterol relative to oleic acid (Katan et al. 1995; Aro et al. 1997). The effect of trans polyunsaturated fatty acids, which can be formed on treatments as mild as deodourization of vegetable oils, has not been properly examined as yet. A few recent studies do suggest that in normolipidaemic subjects palmitic acid is not always an LDL-cholesterolraising saturated fatty acid (Ng et al. 1992; Choudhury et al. 1995). This finding, of course, would be of great practical significance if it proved to be correct that under certain conditions palmitic acid can replace oleic acid without affecting LDL-cholesterol levels. Therefore, these results need to be confirmed under various experimental conditions, before any solid conclusions can be drawn. Although linoleic acid is the most abundant polyunsaturated fatty acid in the diet, a small part of the dietary polyunsaturates is provided by a-linolenic acid (ALA; 18 : 3n-3) and by the very-long-chain fatty acids eicosapentaenoic acid (EPA; 20 : 5n-3) and docosahexaenoic acid (DHA; 22 : 6n-3) from fish oils. Effects on plasma lipoproteins seem comparable between ALA and linoleic acid (Chan et al. 1991). Fish oils, however, have a hypotriacylglycerolaemic effect and have, in normolipidaemic subjects, no effects on LDL and HDL levels. In hyperlipidaemic subjects, fish oils also lower triacylglycerols, but they can increase LDL- and HDL-cholesterol concentrations (Harris, 1997). Controversy still exists with respect to the question of whether EPA is the major triacylglycerol-lowering component of fish oil (Rambjor et al. 1996; Frøyland et al. 1997), or whether DHA has the same property (Ågren et al. 1996; Davidson et al. 1997; Grimsgaard et al. 1997). The structure of dietary triacylglycerols may also affect serum lipid levels. Each natural triacylglycerol has a unique distribution of the three fatty acids over the glycerol molecule. However, the fatty acid configuration of dietary triacylglycerols is sometimes modified to produce fats that have features desired by food manufacturers and consumers. As human lipases preferentially remove fatty acids from the 1 and 3 positions of triacylglycerols, it is possible that changing the positional distribution of fatty acids could have an effect on serum lipoprotein concentrations. However, results from different studies (Grande et al. 1970; Nestel et al. 1995; Zock et al. 1995) have demonstrated that the position of dietary stearic acid or palmitic acid on the glycerol molecule is not an important determinant of the fasting serum lipoprotein profile. Nevertheless, animal studies carried out by Kritchevsky et al. (1973) suggested that randomization of peanut oil may prevent the promoting effect of peanut oil on cholesterol-induced atherosclerosis. These effects could not be explained by differences in absorption or transport of dietary cholesterol (Tso et al. 1984). Randomized butter or lard, however, did not protect against atherosclerosis (Kritchevsky & Tepper, 1977) and the interpretation and relevance of these studies for the human situation are not clear and need further investigation. 2.1.2. Fat replacers. Cholesterol absorption is decreased when human subjects consume diets containing nonabsorbable fat replacers (Jandacek et al. 1990) and therefore these compounds do lower serum LDL-cholesterol S117 concentrations (Mellies et al. 1983). When fat intake is decreased due to replacement by these fat substitutes, HDL-cholesterol concentrations may also be lowered (Widhalm et al. 1994). 2.1.3. Soyabean protein preparations. Anderson et al. (1995) have recently published the results of a meta-analysis concerning the effects of soyabean protein on serum lipid concentrations in human subjects. It was estimated that a daily intake of 47 g soyabean protein would lower serum total cholesterol concentrations by 0.60 mmol/l, which was mainly explained by a decrease of 0.56 mmol/l in LDLcholesterol. The estimated reduction in serum total cholesterol concentrations in subjects with total cholesterol below 6.5 mmol/l was about 4 % (0.20 mmol/l) and about 20 % (1.85 mmol/l) in subjects with cholesterol levels above 8.7 mmol/l. Triacylglycerol levels decreased by 0.15 mmol/l, while no significant changes were seen in HDL-cholesterol levels. No difference in effect could be demonstrated between isolated soyabean protein and/or textured soyabean protein. However, only the individual results from seven out of thirtyone studies reached statistical significance. In those studies, all carried out in Italy by four different groups, subjects were hyperlipidaemic, and textured soyabean protein was used as the source of soyabean protein. Thus, it cannot be excluded that some special, unknown, characteristic of the textured soyabean protein explains the findings, and results cannot be extrapolated to all types of subjects per se. Further, it remains to be determined whether possible beneficial effects of soyabean protein are due to soyabean protein per se or to, for example, phyto-oestrogens, as suggested by the authors. 2.1.4. Mono- and disaccharides. Blaak & Saris (1995) have published a comprehensive review on health aspects of various digestible carbohydrates. It was concluded that, in the majority of studies with normolipidaemic, hypertriacylglycerolaemic or diabetic subjects, monoand disaccharides had similar effects on the serum lipoprotein profile to those of starch, when consumed in amounts found in Western diets. 2.1.5. Resistant starch. Resistant starch is not, at least not entirely, degraded in the small intestine, and reaches the large intestine. Here, it is metabolized by the action of certain bacteria. Although it has been suggested that the metabolic products favourably affect cholesterol metabolism, neither raw nor retrograded starches appear to have a beneficial effect on the serum lipoprotein profile (Heijnen et al. 1996). 2.1.6. Ethanol. A moderate alcohol consumption is negatively related to the risk of CHD. This association, which may be partly explained by the ability of alcohol to increase HDL-cholesterol (Choudhury et al. 1994), appears not to be due to a specific alcoholic drink in particular, but rather to alcohol per se (Rimm et al. 1996). 2.1.7. Dietary cholesterol. Keys et al. (1965a) suggested that the serum total cholesterol concentration is a function of the square root of cholesterol intake. A nonlinear relationship between dietary cholesterol intake and serum total cholesterol concentrations was also proposed by Hopkins (1992), but Hegsted et al. (1965) suggested that serum total cholesterol concentrations are linearly related to the absolute dietary cholesterol intake. Whatever the exact relationship is between dietary and serum cholesterol concentrations, lowering dietary cholesterol intake will S118 G. Hornstra et al. lower serum total cholesterol concentrations, although this effect may diminish when saturated-fat intake is low (Bronsgeest-Schoute et al. 1979). About 75–85 % of this effect is due to an increase in LDL- and about 15–25 % to an increase in HDL-cholesterol (Katan et al. 1986; Clarke et al. 1997). 2.1.8. Fibre. Based on a meta-analysis of ten trials, Ripsin et al. (1992) concluded that the daily consumption of approximately 3 g soluble fibre from oat products lowers serum total cholesterol concentrations by about 0.15 mmol/l. This effect was positively related to the initial serum cholesterol concentration. Other water-soluble fibres have also been reported to reduce total cholesterol concentrations, mainly by lowering LDL-cholesterol (Stasse-Wolthuis et al. 1980). Insoluble fibres have a lesser impact on serum total cholesterol levels (Glore et al. 1994). 2.1.9. Phytosterols. The estimated daily intake of phytosterols in Western countries is about 160–360 mg, of which campesterol, sitosterol and stigmasterol are the most common. These compounds are structurally related to cholesterol, lower cholesterol absorption, and have long been recognized as LDL-cholesterol-lowering agents (Miettinen et al. 1995). Saturated phytosterols are more efficient in reducing serum LDL-cholesterol concentration than unsaturated phytosterols (Ling & Jones, 1995). Esterification of sitostanols, the saturated equivalent of sitosterols, to rapeseed oil fatty acids further increases the LDLcholesterol-lowering efficacy of phytosterols. A daily intake of about 2.0–2.5 g esterified sitostanol lowers serum LDL-cholesterol concentrations by about 10 % in hypercholesterolaemic subjects (Miettinen et al. 1995). 2.1.10. Tocopherols and tocotrienols. Tocopherols and tocotrienols are components with vitamin E activity. Tocopherols are present in most vegetable oils and are more common in the diet than the tocotrienols, which are found at relatively high concentrations in palm and rice-bran oils. Tocopherols do not have any effect on serum lipoprotein concentrations (Kesaniemi & Grundy, 1982). Some studies have suggested that tocotrienols lower LDL-cholesterol concentrations (Qureshi et al. 1995), but other studies have not found any improvement of the serum lipoprotein profile after tocotrienol supplementation (Wahlqvist et al. 1992; RP Mensink, AC van Houwelingen, D Kromhout and G Hornstra, unpublished results). To explain these differences in findings, it was postulated that effective tocotrienol preparations should contain less than 150– 200 mg a-tocopherol/g and 450 mg g- plus d-tocotrienol/g (Qureshi et al. 1996). This suggestion, however, awaits further confirmation. Also, the very recently reported potent LDL-cholesterol-lowering effect of a novel tocotrienol-enriched fraction (TRF25) from rice bran deserves attention in future studies (Qureshi et al. 1997). 2.1.11. Garlic. Two recent meta-analyses found that garlic preparations, in amounts approximately equivalent to half to one clove per day, decreased serum total and LDLcholesterol levels by about 10 % in subjects with elevated plasma cholesterol concentrations (Warshafsky et al. 1993; Silagy & Neil, 1994). It was, however, noticed that many of the studies used had methodological shortcomings, which were accounted for in two very recent studies (Simons et al. 1995; Adler & Holub, 1997). However, results of these two studies were conflicting, despite the fact that the same garlic powder was used and in the same amount. Thus, although evidence exists that garlic may lower LDL-cholesterol concentrations, some questions still remain to be resolved. In addition, it should be emphasized that the cholesterollowering effect may be confined to certain fractions of garlic only. This emphasizes that each substance or preparation should be evaluated properly in well-controlled studies at more than one location, before any firm conclusions can be drawn. 2.1.12. Other components. Despite promising results in the past, very recent well-controlled studies have demonstrated that fermented milk products (Richelsen et al. 1996) and inulin (Pedersen et al. 1997) are not very likely to have beneficial effects on the fasting serum lipoprotein profile. Also, the claimed lipid-lowering effects of oligofructose in man have not yet been properly demonstrated. 2.2. Postprandial effects So far, only the effects of dietary components on fasting lipid and lipoprotein concentrations have been discussed. However, lipoprotein remnant particles, which circulate in the blood after a meal, are also atherogenic. As chylomicrons, precursors of the remnant particles, mainly transport dietary triacylglycerols (and cholesterol), it is not surprising that postprandial triacylglycerol concentrations are more pronounced on high-fat diets, even if fasting triacylglycerol levels are lower. The fatty acid composition of the habitual diet might also be an important determinant of the postprandial triacylglycerol response, as this response appears to decrease when the diet contains highly unsaturated fatty acids from fish oils (Harris, 1997). Hayford et al. (1979) have also reported that sucrose-containing diets induced a higher triacylglycerol response than diets containing maize-syrup. Finally, components that interfere with dietary cholesterol absorption may affect the composition of the chylomicron and its remnant particles. However, the extent and importance of these effects are difficult to quantify and the postprandial effects of diets is certainly an area that should be investigated more thoroughly in the very near future. 2.3. Gene–diet interaction Apolipoprotein (apo) E, an apolipoprotein associated with chylomicrons, VLDL, HDL and remnant particles, is a ligand for both the remnant receptor and the LDL-receptor. There are three common alleles in the population, which are, in decreasing frequency: E3, E4, and E2. As apoE2 has a lower affinity for the remnant receptor than the other two isoforms, subjects with the apoE2 isoform exhibit a delayed clearance of chylomicrons and chylomicron remnant particles after a fat load. Subjects with the apoE4 isoform, however, appear to be more responsive to a reduced cholesterol and saturated fat intake than other subjects (Ordovas et al. 1995), which might partly be explained by the higher fractional intestinal cholesterol absorption in apoE4-subjects (Miettinen, 1991). This would also explain why apoE4-carriers benefit more from sitostanol ester intake than non-apoE4 carriers (Vanhanen et al. Cardiovascular system 1993). Interestingly, Dreon et al. (1995) have also reported that reducing fat intake caused a shift from large to smaller LDL particles, which was most pronounced in apoE4-subjects. Although less extensively studied, it has also been suggested that common polymorphisms for apoA-I (LopezMiranda et al. 1994) and apoA-IV (Mata et al. 1994) may explain a part of the inter-individual response when dietary fat and/or cholesterol intake is modified. Associations between certain polymorphisms of apoB, apoC-III and lipoprotein lipase (EC 3.1.1.34) and lipid responses to dietary changes have also been reported, but are not very consistent (Ordovas et al. 1995). More research is needed to further assess the genetic impact on diet–lipoprotein interactions. 2.4. Possible mechanisms of dietary fats Theoretically, diet may modify cholesterol metabolism in several ways at different levels: (1) cholesterol and/or fat absorption; (2) faecal sterol excretion; (3) cholesterol and/or apolipoprotein synthesis and excretion; (4) receptordependent and -independent lipoprotein uptake; (5) lipoprotein composition and catabolism; (6) changes in enzymes and/or proteins, like lipoprotein lipase, cholesterol-ester transfer protein, and lecithin-cholesterol acyl transferase. It should be realized, however, that these mechanisms do not operate in isolation. For example, it can be imagined that endogenous cholesterol synthesis is increased in order to compensate for a decreased cholesterol absorption. To date, most of the studies have focused on the effects of dietary fatty acids on LDL metabolism and two competing theories will be summarized briefly. 2.4.1. Concept of Spady and colleagues. A detailed model to delineate the effects of dietary fatty acids on lipoprotein metabolism in the hamster has been described by Spady et al. (1993). According to their hypothesis, dietary fatty acids change LDL-receptor activity, but not the production of apoB-100 by the liver or whole-body cholesterol synthesis. Thus, if the activity of the LDLreceptor is reduced, LDL-cholesterol levels are also increased because of increased conversion of intermediate density lipoproteins to LDL. Although the results from this hamster model are very consistent, it is not known whether this concept can be extrapolated to man. 2.4.2. Concept of Hayes and colleagues. Hayes and Khosla (1992) and Hayes et al. (1992) have postulated that the cholesterol-raising saturated fatty acids increase the production of apoB-100 by the liver and have no effect on the activity of the hepatic LDL-receptor. This will result in an increased VLDL-output, the effect being the strongest for lauric and myristic acids; palmitic acid has a smaller effect. As the activity of the LDL-receptor is not increased, LDLcholesterol levels must rise because of increased conversion of VLDL into intermediate-density lipoproteins and subsequently LDL. Linoleic acid lowers the LDL-cholesterol concentration, because, according to Hayes’ hypothesis, it up-regulates the LDL-receptor. This effect of linoleic acid is maximal already at an intake of 6–7 % of daily energy. It is now postulated that, at adequate intakes of linoleic acid, the up-regulated LDL-receptor can counterbalance the relatively small effect of palmitic acid on VLDL production. S119 Under these conditions, palmitic acid and oleic acid have similar effects on LDL-cholesterol concentrations. However, in situations that down-regulate the LDL receptors (for example, dietary cholesterol intake above 300 mg/d or total serum cholesterol above 6.5 mmol/l), linoleic acid cannot fully neutralize the effect of palmitic acid on apoB-100 production. 3. Some diet effects on arterial thrombotic processes: platelet and endothelial cell functions, blood coagulation and fibrinolysis 3.1. Arterial thrombosis and cardiovascular disease Evidence that arterial thrombosis contributes to genesis and complications of cardiovascular disease is mainly based on pathological and epidemiological studies, showing significant associations between the various thrombotic processes, or the levels of factors involved in these processes, and disease morbidity or risks (Fuster et al. 1992; Davies, 1997). Evidence that modulation of these processes or factors affects disease risk seems strong for platelet function (Antiplatelet Trialist’s Collaboration, 1994), is reasonable for coagulation (Chalmers et al. 1977; Loeliger, 1984; Smith et al. 1990), and substantial for fibrinolysis (Fibrinolytic Trialists’ Collaborative Group, 1994), but is lacking for endothelial function. The protective effect of aspirin on cardiovascular morbidity and mortality has long been considered the major evidence for a causal relationship between arterial thrombosis (and platelet function in particular) and cardiovascular disease (Steering Committee of the Physicians’ Health Study Research Group, 1989), but recent evidence indicates that the beneficial effect of aspirin may be secondary to its anti-inflammatory properties (Ridker et al. 1997). So, for the time being, thrombotic processes and factors should be considered risk markers for cardiovascular disease, not risk factors. 3.2. Platelet function as a marker for CHD Although platelet activation is instrumental in arterial thrombogenesis, too little is known about the predictive value of platelet function (adhesion, release and aggregation) on the incidence of CHD. It has been clearly shown that the suppression of platelet activation, either by drugs (Antiplatelet Trialist’s Collaboration, 1994) or by blocking the platelet fibrinogen receptor, glycoprotein IIb/IIIa (EPIC Investigators, 1994) offers protection against myocardial infarction (MI) and other ischaemic events respectively. However, rather little is known about the association between increased platelet activation and CHD morbidity and mortality. So far, increased platelet aggregation induced by ADP or thrombin has been shown to be associated with past MI and electrocardiographic evidence of ischaemia respectively (Elwood et al. 1990, 1991), and prospective evidence has been presented of an association between increased platelet count and ADP-induced platelet aggregability, and long-term incidence of fatal CHD (Thaulow et al. 1991). However, a nearly twofold difference in the CHD rate of two Finnish cohorts was not associated with S120 G. Hornstra et al. differences in platelet aggregation induced by different agonists (Salo et al. 1985). Two studies have demonstrated that platelet activation, as measured by the urinary excretion of the platelet specific protein b-thromboglobulin (b-TG), is significantly associated with the risk of CHD (Ghaddar et al. 1995; Gorgels et al. 1995). Platelet volume is increased at the time of acute MI (Bath & Butterworth, 1996) and there is compelling evidence that changes in platelet volume are associated with myocardial risk (Martin et al. 1991; Brown & Martin, 1994). The main method to assess platelet function in dietary studies has been the platelet aggregation test in vitro, with the help of which it has repeatedly been shown that changes in dietary fatty acids can modulate the platelet aggregation pattern. However, the results are far from consistent and their interpretation in terms of thrombosis tendency is difficult, if not impossible, since direct comparisons between platelet aggregation in vitro and arterial thrombosis in vivo have not been made in man. In the rat, diet-induced changes in platelet aggregation in vitro appeared to be negatively related to changes in arterial thrombosis in vivo (Hornstra et al. 1993). 3.3. Some diet effects on arterial thrombogenesis and platelet function 3.3.1. Fatty acids. The intake of linoleic acid (18 : 2n-6) is strongly correlated with the linoleic acid content of plasma phospholipids, cholesterol esters and triacylglycerols. In addition, platelet total linoleic acid, ALA (18 : 3n-3), arachidonic acid (20 : 4n-6), EPA (20 : 5n-3) and DHA (22 : 6n-3) are significantly correlated with the concentrations of these fatty acids in plasma triacylglycerols, plasma phospholipids and/or adipose tissue. However, the concentrations of unsaturated fatty acids in e.g. adipose tissues do not predict risk for thrombosis (Kardinaal et al. 1995). In two Finnish cohorts, platelet aggregation induced by ADP showed a significant positive correlation with the contents of linoleic acid in adipose tissue and plasma triacylglycerols, but not with linoleic acid in platelets (Salo et al. 1985). There may be a specific preventive influence of ALA on CHD, since the intake of ALA (range 0.8–1.5 g/d) was inversely associated with the risk of MI in the health professionals follow-up study (Ascherio et al. 1996b). Also in the lifestyle intervention of de Lorgeril et al. (1994), among several other dietary changes, about 2 g ALA/d was estimated to be protective. However, in the large Multiple Risk Factor Intervention Trial, no effect with an ALA intake of 1.7 g/d was found (Dolecek & Grandits, 1991). In this study, a positive association was observed between the dietary linoleic acid : ALA ratio and cardiovascular mortality. Serum ALA concentration appeared to be negatively associated with the risk of stroke (Simon et al. 1995) and in a prospective study (Miettinen et al. 1982), serum ALA, EPA, and DHA were all low in MI patients. Observational studies in Norway during the Second World War and cohort studies among Greenland Eskimos and populations in Japan point to a protective effect against CHD of long-chain n-3 fatty acids from fish or marine mammals (Hornstra, 1989). Although later observational studies gave conflicting results (Hornstra, 1989), there are now five large-scale prospective studies demonstrating a negative association between fish consumption and cardiovascular mortality (Kromhout et al. 1985; Norell et al. 1985; Shekelle et al. 1985; Dolecek & Grandits, 1991; Daviglus et al. 1997). In five other studies of about similar size and design, however, no significant relationship was found (Curb & Reed, 1985; Vollset et al. 1985; Lapidus et al. 1986; Morris et al. 1992; Ascherio et al. 1995). This has been suggested to be due to the rather high habitual fish consumption in the ‘low-fish’ group in these latter studies (Kromhout, 1985). However, since all processes that are thought to be involved in the cardio-protective effects of fish (oil) show clear dose–response relationships over a wide range of fish (oil) intakes, this is a rather unlikely explanation. Evidence has also been reported for a positive association between fish consumption and cardiovascular risk. Thus, in two cohort studies a higher mortality from CHD was observed in areas with a relatively high fish consumption as compared with low-fish regions (Simonsen et al. 1987; Hunter et al. 1988). In two large-scale prospective studies in Finland, fish consumption was also positively related to cardiovascular mortality (Salonen et al. 1995; Piettinen et al. 1997). In the study of Salonen et al. (1995) it was suggested that the high intake of Hg from the freshwater fish may have caused increased cardiovascular risk by promoting lipid peroxidation. In conclusion, results of epidemiological studies with respect to the importance of dietary fish (oil) for the prevention of IHD are equivocal and not conclusive. Moreover, it should be reiterated that epidemiological studies can only indicate associations between two phenomena; they can never discriminate between causal and casual relationships. The final proof for the effectiveness of a fish (oil)enriched diet for the prevention of cardiovascular disease has to be obtained via long-term, well-controlled, prospective primary intervention trials, which have not yet been reported. So far, only one secondary intervention study has been published (Burr et al. 1989), demonstrating that subjects who were advised to eat fatty fish at least twice a week had a 29 % reduction in 2-year cardiovascular mortality as compared with volunteers whose diet advice did not include fish. The effect of cis unsaturated fatty acids on platelet function including in vitro aggregation data has recently been reviewed (Mutanen, 1997). From this review it appears that the results are very inconsistent which, at least in part, may have methodological reasons. A promising approach to assess platelet activation in vivo is the measurement of thromboxane (Tx) metabolites (2,3dinor-TxB2 and 11-dehydro-TxB2 ) in urine, or of the concentration of the platelet-specific protein b-thromboglobulin, released from a-granules. Dietary fish oil or long-chain n-3 fatty acids lower high basal Tx excretion rate, while only a modest effect is found at a low basal excretion rate. Results concerning the effects of other unsaturated fatty acids on urinary Tx metabolites are almost totally lacking. Preliminary results indicate that two diets with the same saturated fat content but differing in their linoleic acid contents (5 and 12 % energy) similarly increased 2,3dinor-TxB2 in urine; Turpeinen et al. 1997), which indicates Cardiovascular system enhanced platelet activation in vivo. High stearic acid and trans fatty acid diets also stimulated 2,3-dinor-TxB2 excretion (Turpeinen et al. 1998). Furthermore, the results from these studies indicate that platelet b-thromboglobulin release in vivo was not affected by changes in dietary fatty acids. A general shortcoming in most of the studies to explain the effects of dietary fatty acids has been a lack of information on the fatty acid composition of individual platelet phospholipids. In addition, little is known about the role of the baseline diet with respect to the incorporation of fatty acids into platelets. Platelet membrane fatty acid composition can be changed by dietary means to some extent. The total amount of a given fatty acid in the platelet is probably less important than the factors regulating free fatty acid levels and types in the membrane and in the platelet interior. Platelet receptor responsiveness to physiological stimuli and subsequent signal transduction and fatty acid liberation for eicosanoid synthesis are probably highly dependent on membrane fatty acid composition. However, one can only speculate as to the precise underlying mechanisms. A potentially important second messenger during platelet activation is protein kinase C, the activation of which can be modulated by cis unsaturated fatty acids, while saturated and trans unsaturated fatty acids are inactive (Khan et al. 1995). Six isoenzymes of protein kinase C have now been identified in human platelets, and these may be involved in various aspects of platelet activation. Other mechanisms by which fatty acids, especially n-3 fatty acids, might regulate platelet function involve changes in TxA2 /prostaglandin(PG)H2 receptor affinity following changes in membrane phospholipid composition (Bayon et al. 1995). An alteration of the platelet redox state and the resulting modulation of the expression of certain enzymes could also be involved (M Lagarde, F Achard, M Gilbert, C Bénistant, D Lemaitre and E Véricel, unpublished results). The enhanced sensitivity of platelets from hypercholesterolaemic patients indicates that LDL may also activate platelets. In vitro mildly oxidized LDL (ox-LDL) has been shown to activate platelets significantly while purified apoE seems to inhibit this (Weidtman et al. 1995; Zhao, 1996). In addition, activated platelets release substances, e.g. platelet-derived growth factor, which can modify LDL and enhance the macrophage uptake of ox-LDL (Aviram, 1995). 3.3.2. Antioxidants and platelet function. On the basis of epidemiological studies, dietary antioxidants (tocopherols, carotenoids, flavonoids and Se) have repeatedly been suggested to reduce CHD risk, but the results of intervention studies are more equivocal (Öhrval et al. 1996; van de Vijver, 1997). Platelet function in vitro, and platelet adhesion in particular, has been shown to be inhibited by high levels of a-tocopherol which cannot be obtained from dietary sources alone (Steiner et al. 1995). This mechanism may partly explain the beneficial efficiency of pharmacological amounts of a-tocopherol to prevent MI in the Cambridge Heart Antioxidant Study (CHAOS) (Stephens et al. 1996). Se supplementation in human subjects with a low Se status decreases platelet aggregation in vitro, but has no effect on S121 platelet activation in human subjects with a normal Se status. No experimental data are available as to the effects of carotenoids on platelet function in man. Data on the effects of flavonoids are from in vitro experiments only. The results indicate an inhibition of platelet eicosanoid synthesis and platelet aggregation (Goldberg, 1996). 3.4. Endothelial cell function Damage of the endothelium leads to endothelial dysfunction which is characterized by enhanced expression of cytokines, cell adhesion molecules, von Willebrand factor (vWf), platelet activating factor, and endothelin, and decreased synthesis of PGI2 (prostacyclin) and transforming growth factor-b (TGF-b). The level of vWf has been related to the risk of MI and sudden death in patients with angina pectoris (Thompson et al. 1995). The soluble form of the vascular cell adhesion molecule (VCAM) and vWf were both shown to be raised in stroke patients, while the intracellular adhesion molecule (ICAM) was raised in patients at risk of stroke only (Blann et al. 1996). Latent TGF-b in human vascular SMC is activated by plasmin which is produced from plasminogen by plasminogen activator (t-PA). In vitro Lp(a) impairs this activation. Active TGF-b inhibits SMC migration, proliferation and activation. Suppression of TGF-b led to increased in vitro expression of ICAM-1 in endothelial cells incubated with Lp(a) (Grainger & Metcalfe, 1995). The question of whether cell adhesion molecules are regulators of platelet function is far from clear at the moment and requires further study. 3.4.1. Dietary fatty acids and endothelial cell function. Dietary fatty acids are able to regulate prostacyclin production to some extent. Studies in Greenland Eskimos (Fischer et al. 1986) and in a Japanese fishing village (Hamazaki et al. 1989), as well as various intervention studies with fish or fish oil (for reviews, see Hornstra, 1989; Hornstra et al. 1990) have led to the conclusion that n-3 fatty acids of marine origin increase both PGI2 and PGI3 production in man. However, the methods used (measurement of major PGI metabolites in urine) are very complicated and this is probably the reason why data from various other studies with respect to PGI2 are not consistent (Knapp et al. 1986) and effects of other dietary fatty acids have not been reported. The experimental evidence with respect to the effect of dietary fatty acids on NO synthase regulation is not clear at present. In two cohort studies, negative associations were found between dietary consumption of n-3 fatty acids and plasma levels of vWf (Shahar et al. 1993). In an intervention study, a low-fat (28 % energy), low-saturated fatty acid (9 % energy), and low-cholesterol (215 mg/d) diet for 3 years resulted in significantly lower plasma vWf levels than the control diet. Moreover, a negative correlation between plasma vWf and dietary n-3 and n-6 fatty acids was found (Blann et al. 1995). In patients suffering from non-insulindependent diabetes mellitus (NIDDM), a diet enriched in monoenoic fatty acids (30 % energy) decreased plasma vWf when compared with a diet high in carbohydrate (11 % energy as monoenes) (Rasmussen et al. 1994). In human endothelial cell cultures, both DHA and EPA attenuated the induction of ICAM-1, VCAM-1 or E-selectin S122 G. Hornstra et al. in interleukin-1b-activated cells (Collie-Duguid & Wahle, 1996). On the other hand, DHA, but not EPA or arachidonic acid, was shown to inhibit the cytokine-induced expression of VCAM-1 (Weber et al. 1995) by blocking the activation of nuclear factor kB, an inducible transcription factor which specifically activates transcription of cell adhesion molecules. Activation of nuclear factor-kB is significantly enhanced in vitro by linoleic acid (Henning et al. 1996) and recent results with rabbits suggest that monounsaturated fatty acids might inhibit VCAM-1 expression in vivo (De Caterina et al. 1995b). Availability of arachidonic acid is an important determinant of PGI2 synthesis by endothelial cells but recent results with respect to the effects of DHA and EPA on PGI2 production suggest that alteration of the expression of the enzymes responsible for formation of PGI2 may also be crucial (M Lagarde, F Achard, M Gilbert, C Bénistant, D Lemaitre and E Véricel, unpublished results). The regulation of the expression of cell adhesion molecules probably includes oxidant–antioxidant sensitive mechanisms, since in vitro VCAM-1 gene expression can be inhibited by synthetic antioxidants. Conversely, LysoPC, a component in oxidized LDL, has been shown to upregulate VCAM-1 and ICAM-1 expression in endothelial cells and rapidly induces P-selectin expression in both platelets and endothelial cells (Ochi et al. 1995; Murohara et al. 1996). This latter effect is probably the basis of leucocyte deposition. 3.5. Coagulation and fibrinolysis In the circulation, coagulation and fibrinolysis factors balance each other. The main markers used to evaluate blood coagulability are fibrinogen, factor VII (and other coagulation factors), antithrombin III (AT-III), fibrinopeptide A released from fibrinogen by thrombin, and prothrombin fragment F1þ2 . Today, prothrombin F1þ2 is considered a sensitive marker of clotting activation. Fibrinolytic potential is assessed by measuring plasminogen, tPA, its inhibitor plasminogen activator inhibitor-1 (PAI-1), and cross-linked fibrinogen degradation products (D-dimers). The latter indicator reflects both coagulation and fibrinolysis. The early results from the Northwick Park Heart Study (Meade et al. 1986) indicated strong independent associations between baseline plasma fibrinogen and factor VII coagulant (FVIIc) activity levels and the risk of CHD. In the same population, a U-shaped association between ATIII and the risk for CHD was found. Low fibrinolytic activity predicted a higher risk for CHD in a later analysis (Meade et al. 1993). Fibrinogen is also generally accepted as an independent risk factor for CHD, while the predictive value of PAI-1 and tPA levels as risk factors is still contradictory (Hamsten, 1995; Ridker & Vaughan, 1995). In patients with angina pectoris, the levels of fibrinogen and tPA antigen have been shown to be independent predictors of subsequent MI or sudden death (Thompson et al. 1995). Elevated plasma levels of D-dimers have been shown to be associated with early atherosclerosis (Salomaa et al. 1995) and increased risk of future MI (Ridker et al. 1994), although in the latter study the D-dimer level did not appear to be an independent predictor. Recent results from 2952 men clinically free from CHD show that six markers of the hypercoagulable state (FVIIc, FVII antigen, activated factor VIII and factor IX, prothrombin fragment F1þ2 , and fibrinopeptide A) are all positively associated with CHD risk (Miller et al. 1996). 3.5.1. Effect of dietary factors on coagulation and fibrinolysis. According to several studies, dietary fatty acids hardly influence plasma fibrinogen. There is one study from Denmark (Bladbjerg et al. 1995) showing increased plasma fibrinogen level after an extremely high stearic acid (about 15 % energy) diet when compared with a diet high in myristic and lauric acids. In a recent study, plasma fibrinogen concentration increased slightly during the stearic acid (9.3 % energy) diet, but the biological significance of this is questionable (Mutanen & Aro, 1997). In the population-based cross-sectional atherosclerosis risk in communities (ARIC) study, a negative association between the intake of long-chain n-3 fatty acids and plasma fibrinogen levels was found (Shahar et al. 1993). However, intervention studies with long-chain n-3 fatty acids have given very inconsistent results (Hornstra, 1992). Current knowledge about diet and factor VII (FVIIc activity or FVII antigen levels) indicates that fasting FVIIc can be reduced by low-fat diets. The fatty acid composition of the diet, i.e. saturated, monounsaturated or n-3 and/or n-6 polyunsaturated fatty acid contents, have not been found to be important in short-term experiments (Mennen et al. 1996). Habitual high-fat diets seem to increase both FVIIc and FVII antigen. Increased postprandial responses of FVIIc are seen after high-fat test meals regardless of the type of fat. It seems that the change in fasting FVIIc is part of a general change in concentrations of vitamin K-dependent proteins, while changes in non-fasting FVIIc activities are primarily mediated by activation of the factor VII zymogen (Bladbjerg et al. 1995). The activation of factor VII has been suggested to be related to free fatty acid production during lipolysis of triacylglycerol-rich lipoproteins (Silveira et al. 1994). There are only a few reports about the effects of diet on AT-III. Early studies indicate that n-6 polyunsaturates might increase plasma AT-III, while long-chain n-3 have either no effect or may increase it. A recent study comparing ALA with EPA+DHA indicated that ALA might have a beneficial effect on plasma AT-III levels (Freese & Mutanen, 1997). Supplementation for 16 weeks with long-chain n-3 fatty acids of patients with chronic atherosclerotic disease induced a significant increase in plasma levels of tissue factor pathway inhibitor, indicating down-regulation of the extrinsic pathway of blood coagulation (Berrettini et al. 1996). In an earlier study, a shorter supplementation period did not produce such an effect (Hansen et al. 1994). In the study of Berrettini et al. (1996) a significant reduction of F1þ2 plasma levels was found also. A slight but significant decrease in F1þ2 has been reported after a high-stearic-acid diet when compared with a high-myristic and -lauric acid diet (Bladbjerg et al. 1995). Circulating amounts of F1þ2 were not different between low-fat and high-monoene diets (Lopez-Seguara et al. 1996). In a long-term study with a low-fat (26 % energy) highfibre diet, tPA activity increased significantly in healthy subjects, while no change in tPA antigen was found (Marckmann et al. 1993). However, the reduction of the Cardiovascular system total dietary fat content alone (from 39 to 31 % energy) had no effect. Several studies have found no effect on plasma tPA activity of dietary long-chain n-3 fatty acids (Eritsland et al. 1994), olive oil (Lopez-Segura et al. 1996), maize oil (Hellsten et al. 1993) or stearic acid or trans fatty acids (Mutanen & Aro, 1997). Effects of the dietary composition on PAI-1, either antigen or activity, are not consistent (Hornstra, 1992; Hellsten et al. 1993). Long-chain n-3 fatty acid supplementation mainly increases PAI-1 antigen and either increases, or has no effect on PAI-1 activity. In a recent study, PAI-1 activity increased similarly with either ALA or EPA+DHA supplementation (Freese & Mutanen, 1997). There are only a few studies addressing the effects of other dietary fatty acids on PAI-1. A high-oleic-acid diet (fat 38 % energy, monoenes 24 % energy) decreased both PAI-1 activity and antigen when compared with a high-carbohydrate diet (fat 27 % energy, monoenes 13 % energy). The decrease was accompanied by a parallel decrease in plasma insulin levels (Lopez-Segura et al. 1996). Maize oil supplementation resulted in decreased PAI-1 activity (Hellsten et al. 1993), but in another study olive oil did not have an effect (Oosthuizen et al. 1994). Changes in total fat and fibre intake did not affect PAI-1 either (Marckmann et al. 1993). No changes in D-dimer concentrations have been detected in some recent studies with long-chain n-3 fatty acids (Eritsland et al. 1994, 1995), trans fatty acids (Almendingen et al. 1996; Mutanen, 1997) or stearic acid (Mutanen & Aro, 1997). A decrease in D-dimer level was found in the study of Mutanen & Aro (1997) when the subjects changed from their habitual diet (polyunsaturated : saturated fatty acid ratio (P : S) 0.36) to more saturated type of diet (P : S 0.24). Data concerning the effects of other dietary factors on coagulation and fibrinolysis are scarce. The results from a low-fat, high-fibre experiment by Marckman et al. (1993), however, indicate that some components of dietary fibre may affect coagulation and fibrinolysis. Two other studies support this assumption (Nilsson et al. 1990; Sundell & Ranby, 1993). Recently, in a large Finnish cohort of middleaged men an inverse association was observed between the intake of dietary fibre and the risk of CHD. Adjustment for serum cholesterol did not change the results, indicating that in the mechanism lipoprotein metabolism is not involved (Pietinen et al. 1996). Platelets are important contributors to both coagulation and fibrinolysis. Although tissue factor present in monocytes and the blood vessel wall, in combination with activated factor VII (FVIIa), is the main initiator of coagulation, activated platelets, by exposing phosphatidyl serine at their surface, provide the preferred surface on which coagulation occurs. This platelet procoagulant activity, also called platelet factor 3, is closely related to platelet aggregation. AT-III can rapidly inhibit FVIIa that is bound to tissue factor, thus inhibiting the start of coagulation (Rapaport & Rao, 1995). How platelet membrane fatty acid composition affects the exposure of phosphatidyl serine, or how tightly the fatty acid composition of phosphatidyl serine is regulated is not known at present. The functional association between fatty acids and tissue factor presentation in tissue-factor-containing cells is not known either. S123 There is some evidence that long-chain saturated fatty acids might provide a contact surface for activation of clotting factors XII and IX (Mitropoulos, 1994). Activation of these factors can cause the activation of factor VII and thus increase FVIIc. Fibrinolysis also occurs at the platelet surface after direct binding of plasminogen, tPA and plasmin. Once bound, tPA manifests enhanced catalytic activity to convert plasminogen to plasmin, thereby enhancing thrombolysis. Formed plasmin also binds to the platelet surface and, at low concentrations, reduces fibrinogen binding which results in reduced platelet aggregation. At high concentrations, however, plasmin activates platelets (Loscalzo et al. 1995). How fatty acids would regulate either the production of plasminogen or tPA in the endothelium or their activation on the surface of platelet membranes is not clear as yet. In endothelial cells, both tPA and PAI-1 productions seem to be mediated by protein kinase C activation (Rydholm et al. 1995) and thus may be influenced by fatty acids. There are two recent reviews on lipoprotein metabolism and thrombosis (Mitropoulos, 1994; Miller, 1995). The current opinion is that fatty acid composition of lipoprotein particles may be important for the activation of the contact system of coagulation. Furthermore, high blood lipid levels may change platelet function by influencing platelet membrane composition and fluidity. 4. Immune-mediated processes underlying CHD Maintaining the vascular integrity and defending the circulatory system against pathogenic processes require regulatory interactions among blood cells and between blood cells and the vessel wall. The interacting cells are leucocytes (monocytes and T-lymphocytes) and platelets in the circulation, and endothelial cells and SMC in the vessel wall (Ross, 1995). These processes are controlled through activation of adhesion receptors already present on resting blood cells and endothelium, or through the expression of new receptors on the cell surface (Frenette & Wagner, 1996). Cell activation, production of chemoattractants and cell growth factors are key components in these events, which are involved in repair and defence systems, but also, under certain conditions, in tissue injury and disruption in the cardiovascular compartment. Long-term processes also trigger the participation of key components in cellular immunity, such as T-lymphocytes and macrophages (Lodish et al. 1995). These cells are recognized to play a role in inflammatory and immune-mediated processes in atherosclerosis, since T-lymphocytes are present in the arterial plaque (Jonasson et al. 1986), and antibody responses to plaque constituents have been detected (Palinski et al. 1989). 4.1. Immunocompetent cells involved in the atherosclerotic lesion 4.1.1. Endothelial cells. The endothelium, which lines vessel walls and acts as a permeability barrier controlling the exchange of nutrients and fluids, is a dynamic component of the artery. It provides a non-adherent surface for leucocytes and platelets, maintains the vascular tone S124 G. Hornstra et al. by releasing vasoactive molecules such as NO, PGI2 , endothelin and angiotensin II, and produces and secretes growth factors and cytokines. The endothelium also forms and maintains the connective tissue matrix, has the capacity to modify plasma lipoproteins, and provides anti- and procoagulant activities. When these functions are altered, in the initiation of the atherosclerotic lesions, leucocytes adhere to the vessel wall, following the formation of cell adhesion proteins (ICAM-1, VCAM-1 etc.) (Springer, 1990; Poston et al. 1992). There is formation of oxidativelymodified particles and an accumulation of lipoproteins in the subendothelial space (Simionescu et al. 1986). Associated modifications take place, such as altered vascular tone, the inability to regenerate wound sites and to prevent platelet adhesion, thrombosis and coagulation. In addition, growth factors and cytokines are released after cell stimulation. 4.1.2. Smooth-muscle cells. The second major type of cell in the arterial wall is the SMC. During the formation of arterial lesions, SMC, monocyte-derived macrophages and T-lymphocytes accumulate in the lesion, and this process is associated with deposition of connective tissue matrix and lipid. SMC, which are activated to migrate from the media into the intima and to proliferate there, produce a variety of growth factors, and the genes for these molecules and for cytokines (e.g. interleukin-1, tumour necrosis factor-a) are induced by various agents (Stemme & Hansson, 1994). SMC are present in two phenotypic states: the contractile and the synthetic. In the contractile state, they respond to vasoactive agents, whereas in the synthetic state they express genes for growth factors and cytokines and also produce various forms of connective tissue matrix. 4.1.3. Immunocompetent leucocytes. In addition to the constitutive cells of the vessel wall, all forms of lesions contain elements of specialized chronic inflammation, e.g. monocyte-derived macrophages and T-lymphocytes. The macrophages, in addition to acting as scavengers and as antigen-presenting cells, produce growth-regulatory proteins and could contribute to lipoxygenase-mediated generation of oxLDL. Macrophages are the main source of foam cells, since they take up oxLDL through scavenger receptors and a putative oxLDL receptor (Stemme & Hansson, 1994). They can also produce growth factors and chemotactic molecules for other monocytes, endothelial cells and SMC. T-lymphocytes represent the second type of cells derived from the circulation and found in common atherosclerotic lesions (Jonasson et al. 1986). These cells appear to be in a low degree of activation, and have a low proliferation rate (Gordon et al. 1990). Large numbers of T-lymphocytes are generally found in lesions associated with risk factors, e.g. hyperlipidaemia, diabetes, and hypertension. 4.1.4. Mechanisms for the recruitment of blood cells in arterial lesions. Recruitment of lymphocytes and monocytes, their binding to the endothelium, and adhesion of activated platelets to monocytes and endothelium, all these processes are mediated by cell–cell adhesion molecules. Movement of leucocytes from the blood into tissues, contributing to tissue oedema and necrosis following ischaemia, involves additional adhesion molecules, such as ICAM-1 and VCAM-1. Growth factors and cytokines participate in cell interactions and in the development of the arterial lesions. They have been detected in atherosclerotic plaques in vivo, by in situ detection methods. Evidence for the activation of cell–cell interactions in atherosclerotic disease is now obtained from the assessment of plasma levels of cell adhesion molecules in atherosclerotic patients (Blann & McCollum, 1994). These levels were found to differ depending on the type of dyslipidaemia (Hackman et al. 1996). 4.2. The immune system response modulates atherosclerosis progression One of the products of activated T-cells, interferon-g inhibits SMC proliferation in vitro and in vivo. Therefore, reduced plaque growth would be expected following increased interferon-g production by cells in the plaque. It has been found, in fact, that T-cell depletion leads to increased lesion size after experimental arterial injury (Hansson et al. 1991), and that cyclosporin A, an inhibitor of T-cell functions, accelerates atherosclerosis in hypercholesterolaemic mice (Roselaar et al. 1995). Interferon-g also down-regulates the expression of the scavenger receptor by human macrophages, inhibiting foam-cell formation in vitro (Geng & Hansson, 1992). The following additional observations indicate protective effects of the immune system with respect to the progression of atherosclerosis: in LDL-receptor-deficient rabbits hyperimmunized with homologous oxLDL, there is a substantial reduction in the progression of the lesions (Palinski et al. 1995); elimination of T-lymphocytes with monoclonal antibodies results in larger proliferative lesions in balloon-catheterized rat aortas (Hansson et al. 1991), and mice lacking cytotoxic T-cells develop much larger lesions in the aorta (Fyfe et al. 1994). On the other side, early studies have shown that immunization of rabbits with HSP/65 induces an inflammatory type of lesion (Xu et al. 1992). It is too early, at this point, to conclude what would be the net effect on atherogenesis of a local immune response in the plaque. The same holds for the effect of systemic immune responses, although the systemic antibody response to the plaque autoantigens against oxLDL tends to correlate with aggravation of the disease. 4.3. Nutrition and the immunological aspects of atherosclerosis An array of major and minor components of the diet is able to modulate some functional factor of various types of cells, including those that participate in the formation of the arterial plaque and involve the immune system in atherosclerotic disease. Among the major components of the diet, polyunsaturated fatty acids play an important role in atherogenesis. Among the minor components of the diet, antioxidants have been reported to affect the atherosclerotic process. This heterogeneous class of compounds includes antioxidant vitamins and a large number of molecules, e.g. flavonoids and polyphenols, present in several foods. While some of these effects may be attributed to typical antioxidant Cardiovascular system activities, such as reduced production of reactive O species and of the compounds generated by them (e.g. oxLDL), other effects appear to be mediated by effects on cellular functions. We will consider only the activities of antioxidants on cell-mediated processes. Those concerning the direct effects on the production of reactive O species in biological systems are discussed by Diplock et al. (1998). 4.3.1. Effects of n-3 fatty acids on cellular immune response and inflammatory events in atherogenesis. As recently reviewed by Calder (1996), polyunsaturated fatty acids such as arachidonic acid (20 : 4n-6), EPA (20 : 5n-3) and DHA (22 : 6n-3) affect functional variables in various types of cells, including those involved in inflammation and immunity. The compounds of the n-3 series have been shown to be particularly potent and, therefore, their effects will be more specifically discussed. In vitro effects. ALA (18 : 3n-3), EPA and DHA have been shown to reduce the proliferation of human lymphocytes (Kelly & Parker, 1979; Santoli et al. 1990). They also inhibit the response to antigens and the ability of antigenpresenting cells to present antigen (Fujikawa et al. 1992), and suppress the production of interleukin-2 (Calder & Newsholme, 1992), a major stimulator of the proliferation of lymphocytes and regulator of cytotoxic T-lymphocytes, natural killer cells and B-cells. This type of action suggests that n-3 polyunsaturated fatty acids may play a role in controlling cellular immune processes in atherogenesis. EPA appears to be more active, but ALA has also been shown to exert some of the effects mentioned. However, the experimental conditions in these studies are often far from physiological and, consequently, the relevance of these studies is questioned. Ex vivo effects. Feeding fish oils, often in large amounts, to animals suppresses the response of spleen, thymus, lymph node and peripheral blood lymphocytes to mitogenic stimuli (Kelley et al. 1988) and reduces the proportion of spleen lymphocytes bearing the interleukin-2 receptor (Yaqoob & Calder, 1993). Results concerning the effects on the phagocytic activity of macrophages are conflicting, but reduction of macrophage function has consistently been reported in various animal models and in human subjects. Finally, chemotaxis of blood neutrophils and monocytes towards a variety of chemoattractants is reduced after n-3 administration (Sperling et al. 1993). In feeding studies, n-3 fatty acids appeared to directly affect cellular adhesion processes, as shown by the reduction in the expression of adhesion molecules in T-lymphocytes (Sanderson et al. 1995). These effects have been further investigated in in vitro systems, where n-3 fatty acids appeared to reduce the expression of various adhesion molecules, as well as the binding between monocytes and endothelial cells (De Caterina et al. 1995a). A reduced production of various cytokines by peripheral blood monocytes after n-3 fatty acid supplementation has been reported in human subjects (Endres et al. 1993, 1995b), but later studies by others have not always given similar results. Studies with respect to lymphocyte-derived cytokines are limited and the results are somewhat contradictory too. Clinical aspects of immunomodulation by n-3 fatty acids. Clinical studies aiming to assess the effects on S125 functional variables of the immune system in human subjects have shown that diets rich in n-3 long-chain polyunsaturates v. diets rich in n-6 fatty acids (mainly linoleic acid) decrease the ex vivo synthesis of the cytokines interleukin-1b and tumour necrosis factor-a and reduce T-cell proliferation (Meydani et al. 1993). Suppression of ex vivo interleukin-2 production and of mononuclear cell proliferation was observed in a cohort of the same study (Endres et al. 1993). These effects on cytokine synthesis may take place at the level of transcription, as is suggested from the observed reduction of mRNA (Kaminski et al. 1993). The potential benefits of n-3 fatty acids in controlling cellular events in atherogenesis are indirectly supported by in vitro studies showing reduction of the expression of cytokine-induced pro-atherogenic and pro-inflammatory proteins in human endothelial cells by DHA (De Caterina et al. 1994). These studies concerning the effects of n-3 fatty acids on the function of cells involved in the remodelling of vessel walls and in the atherosclerotic process under pathogenic conditions suggest that these fatty acids exert a protective action with respect to the arterial wall. It should be kept in mind, however, that the in vitro results should be carefully evaluated in the context of the experimental conditions in order to make comparisons with the in vivo situation. Moreover, the in vivo or ex vivo effects have been obtained in general with high doses of the n-3 fatty acids, for relatively short periods of treatment, whereas few studies have been made mimicking the dietary situation (relatively low intakes of long duration). Mechanisms of action. Long-chain polyunsaturates of the n-3 family may act at the cellular level through various mechanisms which can be summarized as follows. (a) Modulation of eicosanoid production by cells of the immune system, especially reduction of proinflammatory PGE2 and leukotriene B4 . (b) Modulation of membrane fluidity. (c) Modulation of signal transduction pathways, especially those involving lipid mediators, protein kinase C and Ca2þ mobilization. (d) Modulation of the expression of genes involved in cytokine production or in peroxisomal proliferation, fatty acid oxidation and lipoprotein assembly. Safety. Although the issue of safety of n-3 fatty acids has not been specifically addressed, there are some reports of alterations of liver function in rodents, at high levels of intake. On the other hand, there is no evidence of negative effects even at relatively high levels of intake, in population groups. It is recommended however, on the basis of some reports of enhanced susceptibility of LDL enriched in n-3 fatty acids to oxidation in vitro (a marker which has a somewhat disputable significance), to increase the intake of antioxidants, such as vitamin E, as a preventive measure. 4.3.2. Antioxidants. Uncontrolled oxidative stress in the cardiovascular system is considered to promote the progression of arterial wall lesions through various mechanisms, the major ones being the enhanced oxidation of lipoproteins (resulting in greater atherogenicity and enhanced accumulation in the cells of the vascular wall) and the activation of cells involved in the pathogenesis of S126 G. Hornstra et al. atherosclerosis (monocytes, endothelium, SMC, platelets), as a consequence of enhanced formation of activators (oxLDL, cytokines, eicosanoids, etc.). In various types of cells, reactive O species may act at the transcriptional level through the activation by cytokines of the transcription factor nuclear factor-kB (Schreck et al. 1992). The major antioxidants in the diet, found also in plasma, are tocopherols, mainly a, but also b and g, b-carotene and other carotenoids, ubiquinone (coenzyme Q10 ), flavonoids, and other plant polyphenols (all lipidsoluble), and vitamin C, which is water-soluble. In vitro and ex vivo studies have shown effects of natural antioxidants on immune competent cells (Middleton & Kandaswami, 1992; Faruqi et al. 1994) and cells in the cardiovascular system. In addition, potent synthetic antioxidants have been demonstrated to inhibit the expression of genes coding for cytokines (DeForge et al. 1992) and to reduce VCAM-1 gene expression in human vascular endothelial cells (Marui et al. 1993). 5. Diet, hypertension and heart function CHD is strongly related to both systolic and diastolic blood pressure in a graded fashion (MacMahon et al. 1990) and treatment of hypertension results in a reduction in coronary disease-related events (Collins et al. 1990). Hypertension due to known factors or diseases (i.e. secondary forms of hypertension) is distinguished from primary hypertension where no known clinical cause for the persistent elevated blood pressure can be identified (essential hypertension). In approximately 90–95 % of hypertensives the causes are unknown. 5.1. Aetiology of hypertension 5.1.1. Membrane function. The role of ion transport across cell membranes in the development of hypertension has been studied extensively. Some cellular activities, such as Naþ –Liþ counter transport, ouabain binding sites (Naþ , Kþ -ATPase) or activity, Liþ –Kþ co-transport, or Liþ leakage, are considered to reflect ion channel function and there is a strong genetic influence on the association between these markers of ion channel activity and hypertension. Depending on the specific marker, the genetic make-up can ‘explain’ 20–60 % of the occurrence of hypertension (Williams et al. 1991). In contrast, environmental influences (including dietary factors) are much less important and explain between 0 and 16 %. This is important when considering the effect of dietary fatty acids that may change cellular membrane fatty acid composition (see section 5.3). The alterations in ionic channel activity are assumed to lead to increased intracellular Ca2þ content and activity, contraction of the arterial smooth muscles and, ultimately, to vasoconstriction. The gene effect related to decreased ouabain-binding sites which are associated with increased intracellular Naþ levels (and presumably with increased intracellular Ca2þ activity due to Naþ –Ca2þ exchange) may occur in 8 % of hypertensives. Other ion-channel-related gene effects mentioned are less common: 2–3 %. In contrast, some ‘ion channels’ are associated with obesity and the combined gene effect occurs in 10 % of the hypertensive population. 5.1.2. Role of humoral mediators. A number of vasoactive substances (angiotensin, endothelin) have been implicated in the development of hypertension. Endothelin1 is a potent vasoconstrictor produced by endothelial cells, but most endothelin is not secreted luminally and hence plasma levels may not adequately reflect local production. Inhibitors of endothelin synthesis or blockade of receptors can reduce blood pressure in genetic hypertensive animal models. Interestingly, noxious stimuli (including oxLDL and cytokines) cause endothelial cells to synthesize endothelin-1 (Boulanger et al. 1992). Thus oxLDL could lead to exaggerated endothelin-1 production in atherosclerotic vessels. The renin–angiotensin system plays an important role in the homeostasis of salt and water. The conversion of angiotensinogen to angiotensin I by renin is rate limiting. Angiotensin II is produced by an angiotensin converting enzyme (EC 3.4.15.1). Angiotensinogen levels have been related to hypertension and a gene resulting in higher angiotensinogen levels has been described. This complex system with positive and negative feedback mechanisms has many aspects that may be important in the aetiology of hypertension. Obesity, excess energy intake and increased angiotensinogen levels have also been linked. A deletion polymorphism of the angiotensin converting enzyme gene has been thought to be connected with the development of CHD (Cambien et al. 1992) and increased pressor responsiveness to angiotensin I in normotensive men (Ueda et al. 1995). However, no linkage with essential hypertension was observed (Jeunemaitre et al. 1992) and apparent linkage with CHD is now strongly contested. 5.1.3. Insulin resistance. Insulin resistance is closely related to hypertension and hyperlipoproteinaemia. Much interest is focused on the metabolic syndrome X, not to be confused with the cardiac syndrome X (chest pain with normal coronary arteries on angiography). The metabolic syndrome is associated with central obesity (although not always), hyperinsulinaemia, hypertriacylglycerolaemia, maturity onset diabetes and hypertension and will be discussed in more detail in section 6. Hyperinsulinaemia has been shown to be related to the development of hypertension in prospective studies (Skarfors et al. 1991; Lissner et al. 1992). Since insulin also affects ion transport and acts as a growth factor, it is thought that these mechanisms may lead to hypertension (Stout, 1990). 5.1.4. Environmental factors. Environmental factors also play a role. Dietary intake of Na will increase blood pressure although some human subjects are more saltsensitive than others. Yet there is a range of responses and it would be an oversimplification to dichotomize the population into those who are either salt-sensitive or not (Weinberger, 1990). On the other hand an inverse association between dietary Kþ intake and blood pressure is recognized. It may be by this mechanism that fruit and vegetable consumption (rich in Kþ ) helps to prevent hypertension, although the lower blood pressure in women consuming a fruit- and vegetable-rich diet and in vegetarians may be independent of K (Ascherio et al. 1996a; Beilin & Burke, 1995). In addition, fruit and vegetables may help to lower the dietary fat intake and the development of obesity. Smoking also increases blood pressure acutely for Cardiovascular system up to 30 min and when it is considered that many smokers may smoke twenty cigarettes per day or more, blood pressure would be raised for long periods of time (Groppelli et al. 1992). However, earlier studies in which blood pressures were compared between smokers and nonsmokers before smoking failed to document this (Groppelli et al. 1992). The diet of smokers differs from that of nonsmokers and in the UK smokers consume more salt and fewer essential fatty acids, in particular linoleic acid (Fulton et al. 1988). Epidemiological studies have generally found positive associations between alcohol consumption and blood pressure (World Hypertension League, 1991). Alcohol in large quantities will contribute significantly to energy intake and may lead to obesity and hypertriacylglycerolaemia. Alcohol intake is associated with other unfavourable lifestyle factors such as smoking and low physical activity. 5.2. Strategies to reduce CHD by lowering blood pressure 5.2.1. Intervention trials. Many of the early clinical trials had insufficient statistical power to evaluate the benefits of blood pressure reduction in hypertensives in terms of CHD (and stroke). Recently, despite a wide range of treatments with drugs, inclusion criteria, etc. these trials have been subjected to meta analysis. CHD was reduced by 10–14 % by hypertensive treatment. (The reduction in haemorrhagic stroke was 40 %.) The reduction in CHD events is less than was predicted from observational studies and this has led to concern about possible adverse side-effects of drugs (MacMahon et al. 1990). b-Blockers adversely affect lipid and glucose metabolism, whilst high doses of diuretics share some of these metabolic effects. Neither Ca nor Mg supplements reduce blood pressure in subjects that are not deficient in these minerals. Algorithms have been developed for the treatment of essential hypertension. After it has been established that blood pressure is really elevated (several readings after reasonable rest on several occasions) a nonpharmacological approach to lower blood pressure is commenced (see section 5.2.2). Only thereafter is the need for pharmacological treatment considered as part of a multiplerisk-factor approach to management. For reasons given earlier, traditional therapy with diuretics and b-blockers is often replaced by angiotensin converting enzyme inhibitors, Caantagonists or a1-adrenoceptor blockers. 5.2.2. Individual v. population approach. There are more hypertensive subjects in populations with a high median blood pressure level. Most CHD events are seen in the many patients with mildly elevated blood pressure, although their risk is less than that of the small group of severe hypertensive subjects. These arguments form the basis for the population strategy to prevent hypertension and CHD by non-pharmacological measures: reduction in salt and high alcohol intake, avoidance of obesity and increasing the dietary Kþ : Naþ ratio. It is estimated that the average systolic blood pressure would be lowered by some 8 mmHg. The magnitude of this blood pressure lowering effect would be difficult to discern in an individual. Nevertheless, at the population level such a reduction in systolic blood pressure in middle-aged men has the potential to reduce CHD and stroke mortality by 16 and 23 % respectively. S127 5.3. Dietary fatty acid composition and hypertension There are some epidemiological observations that suggest that dietary polyunsaturated fatty acids, whether n-3 or n-6, may reduce blood pressure. However, dietary intervention studies are contradictory. Diets supplemented with n-6 polyunsaturates (mainly linoleic acid) do not consistently reduce blood pressure. There have been many supplementation studies with mixtures of the fish-oil-derived n-3 longchain polyenes EPA (20 : 5n-3) and DHA (22 : 6n-3). Many of these studies, in normotensive, hypertensive, hypercholesterolaemic and CHD patients were primarily designed to examine the effect of n-3 long-chain polyenes on plasma lipids and generally the number of subjects in the study was too low. The two largest studies, examining the effect in 350 healthy or in 156 hypertensive men and women are still very small in comparison with current cardiovascular trials (Bønaa et al. 1990; Trials of Hypertension Collaborative Research Group, 1992). Nevertheless, they documented no benefit in healthy subjects and a reduction in hypertensives. A recent meta analysis has been carried out. There are significant differences between the studies in terms of blood pressure recording (single or repeated, automatic device or random zero sphygmomanometer, blinding of patients and observers, choice of olive or n-6 polyunsaturated oil as a placebo, etc.). The results suggest that n-3 long-chain polyenes (mainly EPA) reduce blood pressure in a dosedependent manner in hypertensives, but have little or no effect in healthy volunteers (Morris et al. 1993). The mechanism is not clear, but is assumed to be by a reduction in the production of the vasoconstrictor TxA2 . It is widely held that linoleic acid and n-3 long-chain polyenes could affect blood pressure because they can change membrane fatty acid composition and/or membrane fluidity and thereby alter ion-channel activity and prostaglandin synthesis. The fatty acid composition of membrane phospholipids is hardly changed by diets rich in linoleic acid. Long-chain polyenes of the n-3 family, on the other hand, markedly reduce the level of arachidonic acid in phospholipids. A shift from TxA2 to TxA3 , devoid of powerful vasoconstricting properties, is now accepted. Fish oil may not reduce arachidonic acid in phosphatidyl inositol, a phospholipid central to a1-adrenoceptor-mediated inositol pathway signal transduction (MacLeod et al. 1994). However, little is known about the effect of fish oil on the fatty acid composition of small arterial resistance vessels (MacLeod et al. 1994). In view of these uncertainties it is clear that a meta analysis is no substitute for a properly conducted trial. What is needed is a large double-blind controlled trial in hypertensive patients, in whom TxA2 production or a1adrenergic mechanisms have been implicated in their hypertension. 5.4. Heart function The maintenance of blood pressure and perfusion of organs is the main function of the heart. The energy for this is derived from oxidative phosphorylation in mitochondria. As myocardial energy reserves last but a few seconds, there is a constant need of O2 supply. Cardiac output can vary enormously, due to large changes in the emotional and S128 G. Hornstra et al. environmental influences, and changes in O2 consumption will follow in its tract. The function of the heart is controlled by the autonomic nervous system. Increasing O2 requirements are met by increased O2 extraction and blood flow (vasodilatation). The lumen of coronary vessels is controlled by a tonic vasoconstriction mediated by a1adrenergic receptors balanced by a vasodilatation under the influence of adenosine, prostaglandins and endothelialderived relaxing factor (NO). The heart uses a variety of substrates (non-esterified fatty acids, glucose, lactate, etc.) depending on the nutritional state. Non-esterified fatty acids form the main substrate, but the importance of glucose increases during the fed state. 5.4.1. Effect of diet. The contractile function of the heart is under autonomic control and there is no evidence that myocardial function is sensitive to dietary factors under conditions of adequate perfusion and O2 supply. However, when O2 supply is limited due to atherosclerotic lesions in coronary vessels or microvessel disease, the function may become compromised. Oxidative metabolism of glucose requires less O2 than that of non-esterified fatty acids to produce ATP and may help to improve myocardial function under those conditions. Relatively few experimental studies have examined whether myocardial function is sensitive to changes in the dietary fatty acid composition. Early studies in rats have limited value as control diets were deficient in essential fatty acids (Semafuko et al. 1984; Wince et al. 1984). However, a linoleic acid-rich diet may increase contractile force in isolated papillary muscles and improve the relationship between left ventricular work and filling pressure in the isolated perfused rat heart (de Deckere & ten Hoor, 1979). It should be noted that these preparations may have a limited O2 supply and some of these effects may not apply to the well-oxygenated heart. A reduction in basal and a1-adrenoceptor-mediated peak left ventricular pressure by a large intake of fish oil has been reported, but no effect was observed in pigs. Yet, endothelium-derived relaxing factor, which was reduced in some studies by feeding fish oil, can reduce cardiac contractility (Mohan et al. 1994). Thus the potential of n-3 and perhaps also of n-6 fatty acids to influence cardiac contractility under conditions of limited O2 supply or at high work loads can be envisaged. However, without a clear mechanism and understanding of conflicting reports it is too early to speculate. 5.5. Function of the ischaemic heart Acute myocardial ischaemia occurs when coronary flow can no longer meet the O2 requirements of the heart. Ischaemia may be precipitated by an increased O2 demand due to severe exercise or stress and/or acute vasoconstriction or thrombus in the coronary vessel. Prolonged ischaemia results in acute MI and necrosis of the muscle that is inadequately or not perfused. The occurrence of serious ventricular arrhythmias very early is the principal cause of sudden cardiac death. In a patient who has survived this critical period, the loss of ventricular mass may result in heart failure. Early restoration of blood flow (dissolution of the thrombus) may prevent necrosis; however, reperfusion may also stimulate the production of free radicals, which may reduce the function and induce serious ventricular arrhythmias (see Diplock et al. 1998). 5.5.1. Dietary fatty acids and arrhythmia. Diets rich in linoleic acid protect against experimental ischaemiainduced arrhythmias (Leprán et al. 1981; McLennan et al. 1985; Riemersma et al. 1988). There is no consensus about the mechanism that underlies the anti-arrhythmic effect of such diets (Riemersma et al. 1988; Charnock, 1994). It is assumed to be mediated by prostaglandins as a result of changing phospholipid fatty acid composition. However, the protective effect may not be abolished by non-steroidal anti-inflammatory agents (Leprán et al. 1981; Sargent, 1990). Some, but not all, studies have also shown a protection against the so-called ischaemic reperfusioninduced arrhythmias (in contrast to ischaemia-induced arrhythmias discussed earlier) for reasons that are not clear. It is important to point out that diets rich in linoleic acid are low in saturated fatty acids. Thus, it is possible that saturated fatty acids are pro-arrhythmic (Riemersma et al. 1988). However, fewer arrhythmias were also observed when the control diet was rich in monoenoic fatty acids (McLennan, 1993). Large amounts of fish oil, in quantities that would be difficult to consume in the context of a Western diet, protect against the development of serious ventricular arrhythmias in animal studies (McLennan et al. 1989, 1990; Sargent & Riemersma, 1990). Conflicting reports have appeared as to whether this protection extends itself to reperfusion-induced arrhythmias. The underlying mechanism is assumed to be due to increased PGI3 and reduced TxA2 production. Alternatively, it may be due to a direct inhibition of the voltage-sensitive Naþ channel by non-esterified EPA (Weylandt et al. 1996). It is worth emphasizing again that diets rich in polyunsaturated fatty acids usually reduce the intake of saturated fat. Saturated fat intake is not reduced when animals receive small fish-oil supplements (0.4 % energy). Long-term supplementation with small amounts of fish oil was not antiarrhythmic, yet the well-documented biochemical effects (reduction of phospholipid arachidonate content and an increase in the amounts of EPA and DHA) were observed (Riemersma, 1995). Epidemiological and clinical data suggest that small amounts of fish oil may prevent CHD mortality, reduce the risk of lethal events following MI (Burr et al. 1989), and lower the chance of primary cardiac arrest (related to serious ventricular arrhythmias) in the community. It has been suggested that ALA as part of a Mediterranean diet may reduce events after an acute MI (De Lorgeril et al. 1994). This is an intriguing possibility (McLennan & Dallimore, 1995), but requires confirmation in a single-factor study. The convergence between experimental and human studies suggests that dietary fatty acid composition may indeed reduce the risk of atherosclerosis (via haemostatic and/or immunological effects) and lethal coronary events by prevention of serious ventricular arrhythmias. 6. Insulin resistance, obesity and non-insulin-dependent diabetes mellitus Both experimental data and epidemiological studies suggest that abnormalities in lipid- and lipoprotein metabolism, as well as the presence of hypertension, are associated with an Cardiovascular system increased cardiovascular risk. In addition, several other risk indicators, for example insulin resistance, obesity and NIDDM interfere with lipid metabolism and hypertension. The relationship between insulin resistance and subsequent CHD morbidity and mortality is evident from several epidemiological and a few prospective studies. In the Helsinki Police Officers study (Pyörälä, 1979), for instance, 982 men with an age range of 30–59 years were followed over a period of 10 years. In this study, high 1 and 2 h postglucose insulin levels were independent predictors of CHD end-points, where fasting insulin was not an independent contributor. The Paris Prospective Study (Fontbonne et al. 1991) examined the incidence of fatal and non-fatal CHD in 7038 males aged 43–54 years for a period of 15 years. Analysis at 15 years showed that the 2 h post-load insulin level was a significant, independent predictor of death from CHD. 6.1. Insulin resistance, cardiovascular disease and cardiovascular risk factors Insulin resistance may be defined as a diminution of the biological response to a given concentration of insulin. It is a heterogeneous syndrome with both genetic and environmental factors playing a determinant role in the development. Several factors proposed in this respect have been reviewed by Desprès & Marette (1994). These include genetic factors, such as an excessive accumulation of visceral fat, and circulating factors, such as free fatty acids, sex-steroid hormones, tumour necrosis factor-a, hyperinsulinaemia and hyperglycaemia, with main target tissues muscle, adipose tissue and liver. In addition, the morphology of skeletal muscle may itself contribute to the insulin resistance syndrome, since it has been demonstrated that the proportion of oxidative fibres (type I) and capillary density are positively correlated with insulin action in vivo, as determined by the hyperinsulinaemic euglycaemic clamp technique (Lillioja et al. 1987). Environmental factors associated with the insulin resistance syndrome include lack of physical activity and intake of dietary fat (Desprès & Marette, 1994). Large-scale, prospective epidemiological studies revealed a clear association between insulin resistance and CHD (Pyörälä, 1979; Fontbonne et al. 1991). Factors which may contribute to this association are certainly the abnormalities in lipid metabolism and hypertension. 6.1.1. Lipid abnormalities and insulin resistance. There is now abundant evidence for a relation between hyperinsulinaemia and/or insulin resistance and various lipid abnormalities which are known risk factors for CHD and other macrovascular complications. Results of the Helsinki Heart Study (Manninen et al. 1992) and of the PROCAM study (Assman & Schulte, 1992) both demonstrate that the characteristic dyslipidaemia associated with an insulinresistant hyperinsulinaemic state is associated with a marked increase in CHD risk. The alterations in lipid metabolism commonly associated with insulin resistance have been reviewed by Frayn (1993). The insulin resistant state is characterized by elevated plasma triacylglycerol concentrations (particularly elevation of VLDL-triacylglycerol and VLDL-apoB), a decreased plasma HDL-cholesterol S129 concentration (especially HDL2 -cholesterol) and the presence of small, dense LDL particles. The presence of this dense LDL phenotype may result from an overproduction of apoB, induced by an increased availability of free fatty acids (Sniderman et al. 1992). Although cause and effect have never been properly elucidated, it seems that hyperinsulinaemia causes dyslipidaemia, because correction of the dyslipidaemic state leaves the insulin resistance unchanged, whereas correction of the insulin resistance by weight reduction, increased physical activity, and a low-fat diet is immediately followed by the normalization of dyslipidaemia. In most studies concerning the relationship between lipoprotein metabolism and the risk of CHD, only fasting blood lipids and lipoproteins were considered. However, as proposed by Zilversmit (1979) almost 20 years ago and more recently by Patsch (1987), postprandial hyperlipaemia may be particularly atherogenic, especially since a major part of lifetime is spent in the period between food ingestion and 6–8 h thereafter. The magnitude of postprandial lipaemia differs substantially among individuals, including those considered to be normolipidaemic on the basis of fasting blood lipid values (Patsch, 1987). Factors which have been reported to influence postprandial lipaemia include basal triacylglycerol and HDL-cholesterol concentrations (O’Meara et al. 1992), obesity (Lewis et al. 1990; Potts et al. 1995), diabetes mellitus (Stinson et al. 1993) and insulin resistance (Frayn, 1993). Lewis et al. (1990) demonstrated that even normolipidaemic obese subjects have greater postprandial lipaemia and triacylglycerol enrichment of HDL after ingestion of a highfat meal. Potts et al. (1995) concluded that a disturbed triacylglycerol clearance in subcutaneous adipose tissue is related to elevated plasma triacylglycerol concentrations and reduced HDL-cholesterol levels. Roust & Jensen (1993) investigated postprandial free fatty acid kinetics in obese persons and concluded that impaired suppression of adipose tissue lipolysis is a potentially important abnormality present in upper body obesity. Since obesity, and particularly abdominal obesity, is frequently associated with insulin resistance, it could be expected that insulin resistance also might interfere with the magnitude of postprandial lipaemia. The effects of insulin on lipid metabolism occur both in adipose tissue and the liver. Insulin mediates the activation of lipoprotein lipase in adipose tissue; the consequence of a disruption at this level will result directly in an impaired postprandial triacylglycerol clearance (Frayn, 1993). Other actions of insulin comprise the suppression of non-esterified fatty acid release in adipose tissue by inactivation of the hormone-sensitive lipase (EC 3.1.1.3) and increased re-esterification, and the suppression of hepatic secretion of VLDL-triacylglycerol in the liver. The latter may lead to inappropriate postprandial VLDL-triacylglycerol secretion and the presence of large triacylglycerol-enriched VLDL in the postprandial period. As a consequence, neutral lipid exchange with LDL may lead to small, dense LDL particle formation (Frayn, 1993). These observations show a clear relationship between insulin resistance and certain disturbances in lipoprotein metabolism which contribute to the development of S130 G. Hornstra et al. cardiovascular diseases. However, to ascertain a causal relationship, future research has to concentrate on intervention studies which may elucidate the sequence of events. 6.1.2. Hypertension and insulin resistance. Arterial hypertension is an established risk factor for CHD. Several large studies have already reported on the relationship between insulin resistance and hypertension. Some authors described a positive relationship between insulin concentrations and blood pressure (Welborn et al. 1966; Lucas et al. 1985; Modan et al. 1985) but others were unable to demonstrate such a relationship (Cambien, 1987; Asch et al. 1991). The reason for this discrepancy may be found in ethnicity (Saad et al. 1991) and in the presence of obesity (Cambien, 1987). On the basis of published reports, at least half of the patients with hypertension can be considered to have insulin resistance and hyperinsulinaemia (Reaven et al. 1996). The possible mechanisms underlying a relationship between insulin and blood pressure are complex and can be divided into direct and indirect effects. Hypertension may result directly from insulin resistance through the stimulatory effect of high insulin concentrations on vascular smooth muscle proliferation (Banskota et al. 1989). Insulin also enhances renal Na retention directly via its effects on the proximal tubuli (DeFronzo, 1981) and indirectly through stimulation of the sympathetic nervous system and augmentation of angiotensin II-induced aldosterone secretion (Rocchini et al. 1990). Stimulation of the sympathetic nervous system by insulin may also have a direct hypertensive effect (Landsberg & Krieger, 1989). Hypertension frequently occurs in combination with other metabolic alterations such as disturbances in lipid metabolism, obesity and NIDDM. Since insulin resistance seems to be the common link between these factors, the nonpharmacological treatment approach should focus on the increase of insulin sensitivity. Effective tools are weight reduction, increased physical activity, low-fat diet and perhaps consumption of foods that reduce the insulinaemic response. This strategy probably results in a lowering of the blood pressure as well. 6.2. Nutritional aspects Although insulin resistance also occurs in persons with normal body weight, it is a common feature in obese patients with or without impaired glucose tolerance or NIDDM. In this specific patient group, diet and exercise are two common, non-pharmacological approaches for treatment. Considering the aim of the present review, only relevant dietary factors will be discussed. In obese, insulin-resistant patients, several studies have examined the effect of overall weight loss, by diet or a diet þ exercise combination, on cardiovascular risk factors such as lipid and lipoprotein abnormalities and hypertension. From the results, it is clear that weight loss is accompanied by improved insulin sensitivity and a subsequent better metabolic profile (Colman et al. 1995). Whether specific dietary components may influence the status of insulin resistance in obese and non-obese persons is not fully understood. The relationship between dietary factors and physical activity with hyperinsulinaemia was examined in 389 nondiabetic men, 70–89 years of age, who participated in the Zutphen Elderly Study (Feskens et al. 1994). A significant, negative association was observed between insulin levels (during an oral glucose tolerance test) and the intake of dietary fibre and polyunsaturated fatty acids, which could not be accounted for by energy intake, BMI, physical activity, prescribed diets or the presence of CHD. In contrast, insulin levels increased with the increasing intake of saturated fatty acids and alcohol. Apart from overweight, physical activity and dietary factors such as the intake of fatty acids, fibre, carbohydrates and alcohol, were independently associated with hyperinsulinaemia and insulin resistance. In a study with 544 non-diabetic women (aged 30–84 years), the habitual intake of total dietary fat was positively related to fasting insulin concentrations, particularly among sedentary women. The positive relation of dietary fat content with the percentage of body fat accounted for a substantial proportion (630 %) of the association of dietary fat with insulin concentrations (Mayer et al. 1993). In a group of male Swedish elite athletes, diet modification during 1 year resulted in decreased insulin levels in conjunction with a decreased relative fat energy content. Insulin levels returned to baseline amounts when the relative fat energy content increased again (Tegelman et al. 1996). Information about the effect of specific fatty acids on insulin metabolism is scarce. The incorporation of n-3 fatty acids, and of DHA (22 : 6n-3) in particular, into phospholipids, prevents the expected insulin resistance in rats fed on a high-fat diet (Storlien et al. 1991). In human subjects, decreased insulin sensitivity is associated with decreased concentrations of certain long-chain polyunsaturated fatty acids (20 : 4n-6, 22 :4n-6, 22 : 5n-6, and 22 : 5n-3) in skeletal muscle phospholipids (Borkman et al. 1993). Specifically, decreases in C20–C22 long-chain polyunsaturated fatty acids were associated with increased insulin resistance. This raises the possibility that changes in the fatty acid composition of muscle modulate the action of insulin. The results of the study of Borkman et al. (1993) demonstrate that in patients with coronary artery disease, linoleic acid (18 : 2n-6) correlated directly with hyperinsulinaemia, but this was not the case in normal controls. Since insulin has an effect on D6 -desaturation, the conversion of linoleic acid to g-linolenic acid (18 : 3n-6) may be impaired. In addition, Pan et al. (1995) demonstrated that an impaired insulin action and obesity are independently associated with reduced D5 -desaturase activity. In these circumstances, the direct supply of g-linolenic acid (impairment of D6 desaturase) or arachidonic acid (impairment of D5 -desaturase) may be of value (Horrobin, 1993). Obesity was also found to be associated with reduced elongase activity and higher D9 -desaturase activity (Pan et al. 1995). Trans fatty acids interfere with desaturation and elongation of 18 : 2n-6 and 18 : 3n-3 (ALA), thereby further contributing to decreases in C20–C22 long-chain polyunsaturated fatty acids. A decrease in C20–C22 polyunsaturates leads to increased fatty acid synthesis, lipogenesis, insulin resistance and hyperinsulinaemia, with the subsequent development of obesity, hypertension, NIDDM and CHD (OstlundLindqvist et al. 1985; Simopoulos, 1994). Cardiovascular system Further investigations are needed to evaluate if the essential fatty acids linoleic acid and ALA influence insulin resistance and, if so, whether this effect requires desaturation and elongation of these fatty acids. As was already demonstrated in a few studies, postprandial triacylglycerol concentrations correlate with the degree of hyperinsulinaemia and/or insulin resistance, at least in obese persons. The influence of dietary factors on postprandial lipaemia was investigated by Jeppesen et al. (1995). The acute effects of varying amounts of fat and fructose were studied in eleven healthy, non-diabetic subjects with a wide range of plasma triacylglycerol concentrations. Increasing the dietary intake of fat from 5 to 40 to 80 g led to a significant increase in postprandial concentrations of both triacylglycerol and retinyl palmitate. Furthermore, adding 50 g fructose to 5 g fat also led to a significant increase in postprandial concentrations of triacylglycerol and retinyl palmitate. Chronic intake of fish oil (64 mg n-3 fatty acids/kg body weight per d) reduced postprandial lipaemia in eight normolipidaemic volunteers. This effect was not due to increased chylomicron clearance but more probably to reduced chylomicron production or secretion (Harris & Muzio, 1993). Further research is needed to elucidate whether these effects are reflected in changes in insulin sensitivity in these persons. The results suggest that it may be possible to modulate insulin sensitivity and subsequent cardiovascular risk factors by diet. However, further research on mechanisms is unavoidable to determine the real functional component in the diet. 7. Hyperhomocysteinaemia and cardiovascular risk 7.1. Causes of hyperhomocysteinaemia Although there is some evidence that increased plasma homocysteine levels are caused by a massive export of homocysteine from tissues into plasma (Guttormsen et al. 1996), it is usually thought to result from impaired elimination of plasma homocysteine, either by a defective methylation to methionine or by a reduced trans-sulfuration to cystathionine and cystathione. Betaine (an oxidation product of choline) and folic acid (in the form of 5-methyltetrahydrofolate) are the methyl donors for the transmethylation of homocysteine. 5-Methyltetrahydrofolate is obtained from the reduction of 5,10-methylenetetrahydrofolate which is catalysed by the enzyme 5,10methylenetetrahydrofolate reductase (EC 1.5.1.20; MTHFR). In this reaction, methylcobalamin, derived from vitamin B12 , serves as a cofactor. A deficiency or reduced activity of MTHFR results in increased plasma homocysteine levels. Such a reduced MTHFR activity has been shown to result from a series of mutations in the gene coding for this enzyme (Frosst et al. 1995; Goyette et al. 1995). Transfer of a methyl group from betaine to homocysteine requires the active enzyme betaine : homocysteine methyltransferase (EC 2.1.1.5), whereas pyridoxal-50 -phosphate, a form of vitamin B6 , is required as a cofactor. A reduction in this pathway of homocysteine methylation has not been reported so far (Dudman et al. 1996). Because folic acid is an important methyl donor for the S131 methylation of homocysteine, methylcobalamin (derived from vitamin B12 ) is the necessary coenzyme in this reaction and B6 -derived pyridoxal-50 -phosphate is required for homocysteine removal by trans-sulfuration, folate deficiency and/or a poor status of the vitamins B6 or B12 may be a nutritional reason for hyperhomocysteinaemia. 7.2. Athero-thrombotic mechanisms of hyperhomocysteinaemia 7.2.1. Interaction with lipoproteins. In the chemical pathology of atherosclerosis, homocysteine is thought to play an important role, because the free amino groups of LDL can be thiolated by homocysteine thiolactone, causing aggregation and increased uptake of LDL by macrophages, explaining lipid deposition in atheromas. Homocysteine thiolactone, released from homocysteinylated LDL within the vascular wall, promotes intimal injury, oxidation of cholesterol and unsaturated lipids, platelet aggregation, thrombogenic factors, myointimal hyperplasia, deposition of sulfated glycosaminoglycans, fibrosis and calcification of atherosclerotic plaques (McCully, 1993). It has also been suggested (Harpel & Borth, 1992) that homocysteine increases the atherogenic and antifibrinolytic potential of Lp(a). 7.2.2. Smooth-muscle cell proliferation. Homocysteine has been shown to stimulate vascular SMC proliferation, a hallmark of arteriosclerosis, possibly by increasing the transcription rate of cyclin A (Tsai et al. 1996). 7.2.3. Endothelial functions. In cell culture studies, it was demonstrated that homocysteine significantly lowers endothelial cell growth (Tsai et al. 1994). Moreover, as demonstrated by Dudman et al. (1991) in in-vitro studies, homocysteine causes endothelial detachment, but since fibronectin greatly diminished this process, it was considered of limited relevance to atherogenesis in hyperhomocysteinaemia. It has been suggested that high plasma homocysteine levels cause endothelial injury, largely as a consequence of facilitating the generation of H2 O2 from O2 (Stamler & Loscalzo, 1992; Jones et al. 1994; Hultberg et al. 1995), although the evidence is not unanimous (Clarke et al. 1992). H2 O2 , in turn, is presumed to induce dysfunction and damage to the endothelial cell resulting in platelet activation, coagulation and reduced fibrinolysis. Further studies by Stamler et al. (1993) indicate that normal endothelium modulates the potential adverse effects of homocysteine by releasing NO and forming the adduct S-NO-homocysteine. The adverse vascular properties of homocysteine may result from an inability to sustain S-NO-homocysteine formation owing to an imbalance between the production of NO by progressively dysfunctional endothelial cells and the levels of homocysteine. There is no evidence that homocysteine inhibits the formation of endothelial prostacyclin (Wang et al. 1993). 7.2.4. Functions of blood platelets. As summarized by Stamler & Slivka (1996), the effect of homocysteine on platelet function and survival is controversial. Thus, the shortened survival as measured by Harker et al. (1974) in homocysteinuria patients could not be reproduced by others (Uhleman et al. 1976; Hill-Zobel et al. 1982). S132 G. Hornstra et al. Homocysteine has been shown to inhibit the ecto-ADPase activity of human umbilical vein endothelial cells. Because ADP is a potent platelet aggregatory agent, this action of homocysteine may enhance platelet aggregability (Harpel et al. 1996). 7.2.5. Coagulation and natural anticoagulants. Hyperhomocysteinaemia has been associated with a consumption coagulopathy, resulting in reduced amounts of clotting factor VIIc and AT-III. However, there is some evidence that deficient synthesis of these substances is involved, which is normalized on treatment with pyridoxine plus folate (Schienle et al. 1994). In male CHD patients, homocysteine levels were significantly correlated with fibrinogen content and plasma viscosity (von Eckardstein et al. 1994). Within the patient group of this study, both fibrinogen and homocysteine contents significantly increased in parallel with the number of stenosed coronary vessels. Homocysteine was shown to induce tissue factor procoagulant activity in cultured human endothelial cells in a time- and concentration-dependent manner by tissue factor gene transcription (Fryer et al. 1993). In monkeys, Lentz et al. (1996) demonstrated that dietinduced hyperhomocysteinaemia is associated with significantly decreased thrombomodulin anticoagulant activity. In contrast, van den Berg et al. (1995) noted increased plasma levels of thrombomodulin in young patients with peripheral arterial occlusive disease and hyperhomocysteinaemia after methionine loading. Since the thrombomodulin levels decreased on treatment with pyridoxine plus folic acid, in this patient group hyperhomocysteinaemia appears to be associated with enhanced thrombomodulin levels. Together with the increased levels of vWf, this is considered by the authors to be a marker of endothelial dysfunction. Homocysteine inhibits the expression and activity of endothelial cell surface thrombomodulin, the thrombin cofactor responsible for activation of a natural anticoagulant, protein C (Harpel et al. 1996). Although Rodgers & Conn (1990) demonstrated homocysteine to reduce protein C activation by endothelial cells in vivo, this finding could not be confirmed by others (Bienvenu et al. 1993; Aronson et al. 1994). Homocysteine also lowers expression of the natural anticoagulant heparan sulfate proteoglycan on the surface of porcine aortic endothelial cells in culture, as reflected by the reduced binding of another natural anticoagulant, AT-III (Nishinaga et al. 1993). Increased coagulation and fibrinolysis in hyperhomocysteinaemia in vivo has recently been substantiated by increased concentrations of thrombin–antithrombin complexes and D-dimers (Hamano et al. 1996). 7.2.6. Fibrinolysis. In stroke patients, plasma homocysteine levels appeared significantly related to the concentrations of tPA, but not to PAI-1 (Lindgren et al. 1996). Similar results had also been found by Bienvenue et al. (1993) in fifty patients with arterial and venous thrombosis. These authors failed to demonstrate a significant relationship between plasma homocysteine and plasminogen levels. Van den Berg et al. (1995) found normal tPA levels in young patients with peripheral arterial occlusive disease and hyperhomocysteinaemia after methionine loading. Since homocysteine has been shown to inhibit the binding of tPA to endothelial cells in culture (Hajjar, 1993), it may interfere with the fibrinolytic properties of the endothelial surface. 7.2.7. Altered gene expression. By using a modified non-radioactive differential display analysis to evaluate gene expression in cultured human umbilical vein endothelial cells, Kokame et al. (1996) demonstrated that homocysteine can alter the expressivity of multiple genes, including a stress protein, which may contribute to atherogenesis. 7.3. Hyperhomocysteinaemia or B-vitamin status of primary importance in cardiovascular risk? In many studies relating hyperhomocysteinaemia to cardiovascular risk, increased plasma levels of homocysteine are associated with reduced amounts of folic acid, vitamin B12 (cobalamin), vitamin B6 and/or pyridoxal-50 -phosphate (Ubbink et al. 1993; Jacobsen et al. 1994; Pancharuniti et al. 1994; Dalery et al. 1995; Robinson et al. 1995, 1996; Chasan-Taber et al. 1996; Petri, 1996). In addition, cobalamin-deficient patients usually have increased plasma levels of homocysteine (Stabler et al. 1990). This implies that a relative deficiency of these B-complex vitamins rather than the high homocysteine plasma levels may be the actual cardiovascular risk factor (Chasan-Taber et al. 1996). Findings by Schmitz et al. (1996) that homozygosity for the C677T mutation, associated with increased plasma homocysteine level, is not associated with increased risk of MI, irrespective of folate intake, support this contention. Comparable results were obtained by Ma et al. (1996), who suggest that a gene–environment interaction might increase the risk by further elevating plasma homocysteine, especially when folate intake is low. Further evidence for a primary role of folate deficiency in cardiovascular risk comes from studies by Selhub and co-workers (Selhub et al. 1995; Selhub, 1996) who demonstrated that plasma concentrations of folate and pyridoxal-50 -phosphate as well as folate intake were inversely related to extracranial carotid stenosis after adjustment for other known risk factors. In a group of 367 elderly patients undergoing coronary angiography, Herzlich et al. (1996) observed no significant trend in change in homocysteine as the extent of coronary artery disease increased. However, a low vitamin B12 status was shown to be associated with a lower left ventricular ejection fraction, suggesting a primary role for the cobalamin status in determining left ventricular function. Verhoef et al. (1996) also demonstrated that plasma levels of vitamin B6 and folate (but not of vitamin B12 ) were inversely associated with the risk of MI, independently of other potential risk factors. From their studies, Robinson et al. (1995) conclude that low pyridoxal-50 -phosphate confers an independent risk for coronary artery disease and Ellis & McCully (1995) observed that the treatment of patients with carpal tunnel syndrome and related disorders with vitamin B6 was associated with only 27 % of the risk of developing cardiac chest pain or MI compared with patients who had not taken vitamin B6 . Dalery et al. (1995), however, did not observe differences for folate, vitamin B12 or total vitamin B6 between CHD patients and controls. Cardiovascular system 7.4. Dietary B-vitamins lower plasma homocysteine In numerous studies, increased plasma homocysteine levels appear to be associated with reduced plasma folate concentrations (Verhoef et al. 1996) and since folate is the main methyl donor in the conversion of homocysteine to methionine, folate supplementation may be the preferred way to lower homocysteine-mediated cardiovascular risk. In their meta-analysis Boushey et al. (1995) calculated that an additional intake of 200 mg folate/d would reduce the plasma homocysteine content by about 4 mmol/l and that by increasing dietary folate in the USA 13 500–50 000 CHD deaths per year could be avoided. As suggested by Jacques et al. (1996), individuals carrying a gene mutation resulting in expression of a sub-normal activity of the homocysteine transmethylation enzyme MTHFR may have a higher folate requirement. Consequently, this population may certainly require folate supplementation to prevent hyperhomocysteinaemia. Van den Berg et al. (1995) demonstrated, in young patients suffering from peripheral arterial occlusive disease and hyperhomocysteinaemia after methionine loading, that treatment with pyridoxine plus folic acid resulted in normalization of homocysteine metabolism and ameliorated endothelial dysfunction as reflected by a change towards normal circulating levels of vWf and thrombomodulin. Schienle et al. (1994) reported a case study, demonstrating that in a patient with homocystinuria due to cystathionine-bsynthase deficiency and thromboembolic disease, treatment with pyridoxine plus folate not only led to normalization of amino acids in urine and plasma and of plasma levels of plasma coagulation and anti-coagulation factors, but also prevented further thromboembolic episodes. 8. Critical assessment of the science base 8.1. Identification of criteria In this section, the science base presented previously will be critically evaluated, using the following criteria of decreasing importance (except criterion 5). Criterion 1. Plausible and validated evidence does exist for the involvement of the various variables investigated as (anti)risk factors or -indicators in the aetiology of CHD. For risk factors a causal involvement in CHD aetiology should have been proven; for risk indicators such a causality is not required. Criterion 2. Well-designed human intervention studies have been published, demonstrating that higher intakes (or body levels) of the food items considered lower the levels of the identified risk factors or indicators. Criterion 3. Prospective, statistically validated epidemiological evidence is available that links higher intakes (or body levels) of these food items to reduced levels of the risk factors or indicators. Criterion 4. Retrospective epidemiological data are present which demonstrate a statistically validated association between intake or body levels of the food items investigated and the levels of the risk factors or indicators identified under criterion 1. Criterion 5. Clear evidence exists for the safety of the food items considered. S133 8.2. Evaluation of the present knowledge base with respect to food functionality 8.2.1. Plasma lipoproteins. Results from well-designed intervention trials clearly demonstrate that the plasma concentration of LDL-cholesterol is a causal risk factor for CHD. Most probably, plasma HDL-cholesterol concentration is an anti-risk factor, but confirmation still depends on results of intervention studies showing that an isolated increase in the plasma HDL-cholesterol concentration significantly lowers the risk of CHD. Evidence for plasma VLDL or triacylglycerol levels being associated with the risk of CHD is mainly based on epidemiological studies. Therefore, the plasma VLDL or triacylglycerol concentration can only be considered a risk marker for CHD. The same holds for the plasma concentration of Lp(a), which has been demonstrated to be a powerful risk marker in epidemiological studies. It should be mentioned that so far the lipoprotein profile has mainly been investigated in blood sampled under fasting conditions, whereas man is usually in a state of postprandial hyperlipidaemia for at least 8 of every 24 h. Since the importance of the postprandial lipoprotein profile for determining the risk of CHD has only been superficially investigated, it will not be emphasized here. Taking these considerations into account, dietary saturated fatty acids can be classified as CHD-risk-promoting nutrients because, as compared with carbohydrates, they increase plasma LDL-cholesterol concentrations more strongly than plasma HDL-cholesterol levels (chain lengths up to sixteen C atoms) or reduce the plasma HDLcholesterol concentration (stearic acid, 18 : 0), even if they seem to lower slightly the plasma Lp(a) concentration. Dietary trans-monounsaturated fatty acids increase LDLand reduce HDL-cholesterol levels in plasma; moreover they increase the plasma Lp(a) concentration. Foods low in saturated and trans fatty acids and high in linoleic acid and ALA can, therefore, be classified as functional with respect to lowering the lipoprotein-associated risk of CHD. The cis-unsaturated fatty acids oleic acid, linoleic acid and ALA reduce the plasma concentration of LDL-cholesterol, whereas they hardly affect plasma HDL-cholesterol and Lp(a) concentrations. Therefore, foods enriched in these unsaturated fatty acids can be classified as functional in reducing CHD risk. Oils rich in the highly unsaturated fatty acids EPA and DHA have consistently been shown to lower plasma VLDL concentrations and may, therefore, reduce CHD risk. However, in certain population groups they increase the plasma LDL-cholesterol concentration. So, with respect to lipoprotein effects, foods enriched in EPA and/or DHA cannot be classified as functional in reducing CHD risk by virtue of their effect on the plasma lipoprotein profile. Dietary soluble fibre and certain phytosterols can be classified as functional in lowering CHD risk, because they improve the plasma lipoprotein profile. Although ethanol and a number of fat replacers have similar effects, their side-effects may hamper their use as functional foods. Insufficient evidence is available with respect to soyabean protein preparations, garlic, inulin and oligofructose. Finally, the available evidence does not support the S134 G. Hornstra et al. classification of mono- and disaccharides, resistant starch, fermented milk products, tocopherols and tocotrienols as functional food components. 8.2.2. Arterial thrombosis. In Western societies with ageing populations the modulation of thrombosis tendency is likely to become an important approach to the prevention of CHD. The main problem today, however, is the lack of reliable variables to measure the prothrombotic state in human subjects. In addition, there is a considerable lack of indicators reliably reflecting thrombotic risk in man. So far it has not been shown that changes found in platelet function measured in vitro significantly predict changes in thrombosis tendency in vivo. The plasma levels of factors involved in coagulation and fibrinolysis do not necessarily reflect the degree to which these phenomena really occur. Similarly, the predictive value of endothelial cell function for CHD has been insufficiently evaluated. Diet, especially dietary fatty acids, has been shown to affect many of the previously mentioned variables, but the mechanisms involved are largely unknown. Consequently, increasing mechanistic knowledge about the influence of dietary factors on platelet, leucocyte and endothelial functions and on coagulation and fibrinolysis in vivo, is required for improving dietary strategies to control the prothrombotic state. According to current knowledge, long-chain n-3 and n-6 fatty acids are particularly able to modulate both endothelial cell and platelet functions. However, the optimal n-6 : n-3 fatty acid ratio and the effect of these fatty acids on the antioxidant status of the body is not clear. The same holds for the mechanisms by which platelet and/or leucocyte fatty acid composition affect coagulation and fibrinolysis in man. Also the role of dietary factors as regulators of the interaction between different cell types involved in thrombogenesis has been insufficiently studied so far. Because of all these uncertainties, there is no solid evidence for any food item to be considered ‘functional’ with respect to lowering platelet and endothelial functions, coagulation and fibrinolysis. 8.2.3. Immunological interactions. The immune system responses in the cardiovascular system cannot be considered risk factors with respect to the atherosclerotic process, because of a lack of evidence for the causal involvement of these responses in atherogenesis. Therefore, the term risk indicators should be used. Although various studies have shown that a high intake of n-3 fatty acid-rich foods (fish), or of n-3-rich preparations (fish oils) may exert antiatherosclerotic activities, there is no direct evidence that these effects are mediated by modifications of immune responses participating in the atherogenic process. Some indirect evidence may be provided by the results of some, but not all, studies showing favourable effects of n-3 intake on the rate of re-stenosis of dilated coronary arteries (for reviews, see Gapinski et al. 1993 and Cairns et al. 1996), a process which appears to involve cells of the immune system and the proliferation of cells of the arterial walls (Westerband et al. 1997). Additional studies are required to substantiate the effects of n-3 fatty acids on re-stenosis following angioplasty. However, these results may not disclose the mechanism(s) of n-3 activities. Diets rich in antioxidants have been shown to exert protective effects with respect to the atherogenic process. They have also been shown to affect the activities of immune competent cells and to inhibit the expression of genes coding for cell–cell adhesion molecules, which play a role in the development of the arterial lesions. As for the n-3 long-chain polyenes, however, there are no statistically validated epidemiological, prospective, or intervention data indicating that these effects may be mediated by modulation of immune system responses. 8.2.4. Hypertension. CHD is strongly related to both systolic and diastolic blood pressure in a graded fashion and treatment of hypertension results in a reduction in coronary disease-related events. Therefore, hypertension is a risk factor for coronary artery disease. Reports on the blood pressure-reducing effect of linoleic acid are inconsistent. With respect to n-3 long-chain polyenes meta-analyses suggest that these fatty acids may reduce blood pressure in hypertensive, but not in normotensive, subjects. Consequently, n-3 long-chain polyenes may be considered ‘functional’ with respect to reducing increased blood pressure. Since it is not known whether these fatty acids will prevent normotensive people from becoming hypertensive, blood pressure-related functionality of these fatty acids is restricted to hypertensive subjects. A diet rich in fruit and vegetables also helps to lower blood pressure; however, the mechanism involved has not yet been elucidated. The potential of n-3 and perhaps also of n-6 fatty acids to influence cardiac contractility under conditions of limited O2 supply or at high work loads can be envisaged, but evidence is available from in vitro and animal studies only, results are not consistent and mechanisms involved remain controversial. The same holds for the reported preventive or reducing effects of n-6 and n-3 fatty acids on arrhythmia: the data are largely based on animal studies and underlying mechanisms are hardly known. Therefore, insufficient evidence is available for the classification of unsaturated fatty acids as functional with respect to cardiac contractility and prevention of arrhythmia. 8.2.5. Insulin resistance. Although several excellent studies are available, demonstrating a link between insulin resistance, obesity, NIDDM, metabolic abnormalities and coronary artery disease, cause-and-effect relationships have not been proven by statistical means. Consequently, these conditions can only be regarded as risk indicators, not risk factors. From epidemiological studies it is suggested that the intake of dietary fibre (positively) and the intake of dietary fat (negatively) affect insulin sensitivity. However, welldesigned intervention trials of sufficient size and duration concerning the effect of either of these dietary components on insulin sensitivity have not yet been performed. The relatively short-term and mostly small studies that have been reported were largely carried out in obese subjects in which many physiological variables (e.g. general food habits, body weight, insulin sensitivity, blood pressure) are different from normal-weight subjects. There are some data from intervention studies about the effect of specific fatty acids on insulin metabolism. However, mechanisms underlying these associations have not yet been elucidated. Further intervention studies will be important in determining the sequence of events. Cardiovascular system In these studies, the use of stable isotopes will be instrumental. Taken together, there is insufficient evidence to classify any of the food items ‘functional’ with respect to insulin resistance and related conditions. 8.2.6. Hyperhomocysteinaemia. The view that the level of homocysteine is a risk factor for cardiovascular disease is exclusively based on epidemiological investigations, most of which were case–control studies. In-vitro studies with, mainly, endothelial cell cultures clearly demonstrate an endothelium-activating effect of homocysteine, possibly resulting in thrombogenic conditions. However, in vivo data to confirm this thrombogenic potential of plasma homocysteine are not available as yet. Because of the rather consistent inverse relationship between plasma levels of homocysteine and of folate, vitamin B12 and/or vitamin B6 , no final answer can be given to the question of whether hyperhomocysteinaemia or a reduced vitamin status is ultimately associated with an increased cardiovascular risk. Since no well-designed intervention studies have been reported showing that reducing hyperhomocysteinaemia or increasing the folate and/or B-vitamin status causes a reduction in cardiovascular risk, plasma homocysteine, folate, vitamin B6 and vitamin B12 levels can be considered (anti)risk indicators at best. Increasing the consumption of folate and/or vitamins B12 or B6 lowers plasma homocysteine quite consistently, but whether this will result in a reduced cardiovascular risk remains to be proven. In principle, improvement of the folic acid status by dietary folate supplementation may mask or even precipitate clinical manifestations related to vitamin B12 deficiency. However, extensive studies in more than 700 elderly participants in the Framingham Heart Study revealed that the benefit of folate fortification through projected decreases in homocysteine level and heart disease risk greatly outweigh this risk (Tucker et al. 1996). Moreover, concerns about masking cobalamin deficiency by folic acid supplementation could be lessened by adding cobalamin to folic acid supplements (Boushey et al. 1995). 9. Conclusions and recommendations for further research 9.1. Plasma lipoproteins Dietary lipids are able to affect lipoprotein metabolism in a significant way, thereby modifying the risk of cardiovascular disease. Although effects of the individual dietary fatty acids and dietary cholesterol on fasting serum lipids and lipoproteins have been studied extensively, possible interactions among fatty acids or with dietary cholesterol, as well as postprandial effects, are only poorly understood. This should be investigated more thoroughly in well-controlled dietary trials, using recently developed techniques. For example, stable-isotope methodology should be used to measure apoprotein metabolism, or to measure in mononuclear cells mRNA levels of the LDL receptor and of hydroxymethylglutaryl CoA reductase. Also, effects on other lipid variables like, for example, cholesterol ester transfer protein-activity and lipoprotein particle sizes, S135 should then be taken into account so as to increase our understanding of the dietary effects on lipoprotein metabolism. In addition, special attention should be paid to (potential) gene–diet interactions. These remarks, of course, also apply to other dietary components that interfere with cholesterol absorption. 9.2. Arterial thrombosis Platelet function may possibly affect cardiovascular risk and the relatively low platelet content of n-3 polyunsaturated fatty acids may present a risk for platelet hyperactivity. Insufficient evidence is available to reliably link endothelial cell function to cardiovascular risk. Increased blood coagulability and reduced fibrinolytic activity are associated with increased risk for cardiovascular disease, but causality has not been proven and, consequently, ‘functional foods’ cannot be identified. Further research is needed in the following areas. (1) Prospective validation studies should be performed to find out to what extent the presently available putative indicators of arterial thrombosis tendency (i.e. platelet aggregation in vitro, urinary excretion of thromboxaneand prostacyclin metabolites and of specific platelet proteins, plasma concentrations of soluble forms of cell adhesion molecules, activation fragments of clotting factors, and fibrin degradation products) reflect the risk for arterial thrombosis. (2) Depending on the results, it may be necessary to develop and validate new methods to measure in vivo arterial thrombosis tendency in human subjects and to search for and prospectively validate more specific in vivo activation markers for platelets, endothelial cells, leucocytes, clotting factors and the fibrinolytic process. (3) Well-designed intervention studies should be initiated to investigate the effect of selected dietary components (e.g. the various n-3 and n-6 fatty acids and their combination, antioxidants, fibre) on the processes participating in arterial thrombus formation. These studies should not only measure effects, but should also try and unravel the mechanisms involved. 9.3. Immunological interactions Long-chain polyenes of the n-3 family and antioxidants are examples of food components endowed with various biological activities, which can be assessed in in vitro and in ex vivo experiments. These activities include modification of immune system responses of cells participating in atherogenesis, which may thus be considered markers of an active state of this process. Certain foods are rich in n-3 fatty acids (e.g. fish rich in the n-3 long-chain polyenes and some vegetable oils, such as soyabean and low-erucic acid rapeseed, rich in ALA). Other foods (e.g. vegetables and vegetable oils, fruits) are rich in various types of antioxidants (vitamins, flavonoids, polyphenols, etc.). Diets based on high intakes of these foods are, therefore, expected to exert beneficial health effects on the atherosclerotic process, as shown by various studies. However, the variable contents, from both a quantitative and qualitative point of S136 G. Hornstra et al. view, of these bioactive components in these foods make it difficult to define them as ‘functional foods’. In addition, although beneficial effects, for instance on the cardiovascular system, have been shown in human studies, there is little evidence, from statistically validated epidemiological, prospective, or intervention studies, that these effects may be mediated by modulation of immune system responses. As to the safety of high intakes of foods rich in n-3 longchain polyenes, the absence of detrimental effects, except for possible minor intestinal dysfunctions, in the reported studies, indicates that they should be considered safe. The same holds for foods rich in antioxidants, since there is no evidence that a high consumption of these foods results in detrimental effects. We are at an early stage of examining the role of immune function on the development of atherosclerotic plaques and there is a great need to develop strategies for studying the effects of macro- and micronutrients on the function of the immune system. These should take place at different levels of complexity and biological organization, and using dietary investigations that are relevant to the diets consumed in the Western world. Strictly standardized in vitro experiments are required to obtain new information on the role of the cells involved in the onset of the arterial lesions, and main research areas are functional activities, and their controlling factors, of the main cell types participating in the formation of atherosclerotic plaques. These activities need to be tested either alone or during cell–cell interactions, and should involve the assessment of the factors responsible for these events (expression of cell-adhesion molecules, cytokines and growth factors). As to the underlying mechanisms, both short-term effects, mediated by fast cell-signalling pathways, and long-term processes, generally mediated by gene activation and transcription, need to be studied in detail. In this context, special attention should be paid to the interplay between functionally specialized cells in the vessel wall, e.g. endothelial cells and SMC, and inflammatory and immune cells. Recruitment of these latter cells from the circulation into the vessel wall is a major factor in controlling locally the progression of the lesions, whereas systemic immune responses may differently modulate the process. Clearly, studies of the in vivo effects of nutrients on these steps represent the first approach in the identification of active components and they will also shed some light on potential and most promising mechanisms of action. While animal studies allow the assessment of pathological events at the organ and tissue level and of the effects of treatments on these processes, this can obviously not be done in human subjects. The most important aspects of research on cell-mediated processes in atherogenesis, i.e. human studies, and on the effects of drugs and nutrients, are therefore also the most difficult ones and completely rely on specific markers of the disease state. Therefore, a most important area of research is the assessment of clear relationships between different stages and forms of the disease and selected markers of cellular and immune activation, to be detected in the circulation and possibly in urine. More specifically, measurements of soluble adhesion molecules in plasma and, possibly, of cleavage products in urine, may improve the diagnosis and the evaluation of prognosis of the disease. In addition this may help to establish and follow the impact of nutrient supplementation. 9.4. Hypertension There are many different reasons for hypertension in individuals. In the aetiology of hypertension, the genetic component is definitely stronger than environmental factors, including diet. Future work should consider the multiple reasons that may lead to hypertension. The effect of dietary fatty acids on blood pressure should be examined in patients in whom TxA2 production and/or a1-adrenergic mechanisms are implicated in hypertension. The effect of individual fatty acids, such as ALA, EPA and DHA, on the development of atherosclerosis (via haemostatic or immunological effects) and lethal coronary events should be examined in large well-designed trials in man. 9.5. Insulin resistance Several studies indicate an existing relationship between insulin resistance and cardiovascular disease. Factors which may contribute are fasting and postprandial lipoprotein levels in plasma, as well as hypertension. Environmental factors include the lack of physical activity and the intake of dietary fat. It may be possible to modulate insulin sensitivity and subsequent cardiovascular risk factors by diet, more specifically by decreasing the total amount of dietary fat and increasing the proportion of polyunsaturated fatty acids. However, additional studies on the mechanisms involved are required to understand the real function of these dietary components. Further research should also focus on intervention studies, not only to test the efficacy of specific fatty acids, dietary fibre, low-energy diets, etc., but also to try and explain the mechanisms underlying the observed changes. Moreover, these studies will be helpful in determining the sequence of events. Further investigations are also needed to evaluate whether the essential fatty acids linoleic acid and ALA ameliorate insulin resistance and, if so, whether this effect requires desaturation and elongation of these fatty acids. For these studies, the use of stable isotopes is instrumental. 9.6. Hyperhomocysteinaemia Compelling evidence is now available for the association between the plasma level of homocysteine and the risk of cardiovascular disease, although further studies are needed to substantiate the causality of this relationship. In addition, well-designed intervention trials are required to prove the beneficial role of dietary supplements containing folate and vitamins B6 and B12 in reducing the risk of CHD. References Adler AJ & Holub BJ (1997) Effect of garlic and fish-oil supplementation on serum lipid and lipoprotein concentrations in hypercholesterolemic men. American Journal of Clinical Nutrition 65, 445–450. Cardiovascular system Ågren JJ, Hanninen O, Julkunen A, Fogelholm L, Vidgren H, Schwab U, Pynnonen O & Unsitupa M (1996) Fish diet, fish oil and docosahexaenoic acid rich oil lower fasting and postprandial plasma lipid levels. European Journal of Clinical Nutrition 50, 765–771. Almendingen K, Seljeflot I, Sandstad B & Pedersen JI (1996) Effects of partially hydrogenated fish oil, partially hydrogenated soybean oil, and butter on hemostatic variables in men. Arteriosclerosis, Thrombosis and Vascular Biology 16, 375–380. Anderson JW, Johnstone BM & Cook-Newell ME (1995) Metaanalysis of the effects of soy protein intake on serum lipids. New England Journal of Medicine 333, 276–282. Antiplatelet Trialist’s Collaboration (1994) Collaborative overview of randomised trials of antiplatelet therapy. 1: Prevention of death, myocardial infarction, and stroke by prolonged antiplatelet therapy in various categories of patients. British Medical Journal 308, 81–106. Aro A, Jauhiainen M, Partanen R, Salminen I & Mutanen M (1997) Stearic acid, trans fatty acids, and dairy fat: effects on serum and lipoprotein lipids, apolipoproteins, lipoprotein(a), and lipid transfer proteins of healthy subjects. American Journal of Clinical Nutrition 65, 1419–1426. Aronson DC, Onkenhout W, Raben AM, Oudenhoven LF, Brommer EJ & van Bockel JH (1994) Impaired homocysteine metabolism: a risk factor in young adults with atherosclerotic arterial occlusive disease of the leg. British Journal of Surgery 81, 1114–1118. Asch S, Wingard DL & Barrett-Connor EL (1991) Are insulin and hypertension independently related? Annals of Epidemiology 1, 234–244. Ascherio A, Hennekens C, Willett WC, Sachs F, Rosner B, Manson J, Witteman J & Stampfer MJ (1996a) Protective study of nutritional factors, blood pressure, and hypertension among US women. Hypertension 27, 1065–1072. Ascherio A, Rimm EB, Giovannucci EL, Spiegelman D, Stampfer M & Willett WC (1996b) Dietary fat and risk of coronary heart disease in men: cohort follow up study in the United States. British Medical Journal 313, 84–90. Ascherio A, Rimm EB, Stampfer MJ, Giovannucci EL & Willett WC (1995) Dietary intake of marine n-3 fatty acids, fish intake, and the risk of coronary disease among men. New England Journal of Medicine 332, 977–982. Assman G & Schulte H (1992) Relation of high-density lipoprotein cholesterol and triglycerides to incidence of atherosclerotic coronary heart disease (The PROCAM Experience). American Journal of Cardiology 70, 733–737. Austin MA (1992) Genetic epidemiology of low-density lipoprotein subclass phenotypes. Annals of Internal Medicine 24, 477– 481. Aviram M (1995) LDL-platelet interaction under oxidative stress induces macrophage foam cell formation. Thrombosis and Haemostasis 74, 560–564. Banskota NK, Taub R, Zellner K, Olsen P & King GL (1989) Characterisation of induction of protooncogene c-myc and cellular growth in human vascular smooth muscle cells by insulin and IGD-I. Diabetes 38, 123–129. Bath PMW & Butterworth RJ (1996) Platelet size: measurement, physiology and vascular disease. Blood Coagulation and Fibrinolysis 7, 157–161. Bayon Y, Croset M, Daveloose D, Guerbette F, Chirouze V, Viret J, Kader JC & Lagarde M (1995) Effect of specific phospholipid molecular species incorporated in human platelet membranes on thromboxane A2/prostaglandin H2 receptors. Journal of Lipid Research 36, 47–56. Beilin LJ & Burke V (1995) Vegetarian diet components, protein and blood pressure: which nutrients are important? Journal of S137 Clinical, Experimental and Pharmacological Physiology 22, 195–198. Berrettini M, Parise P, Ricotta S, Iorio A, Peirone C & Nenci GG (1996) Increased plasma levels of tissue factor pathway inhibitor (TFPI) after n-3 polyunsaturated fatty acids supplementation in patients with chronic atherosclerotic disease. Thrombosis and Haemostasis 75, 395–400. Bienvenu T, Ankri A, Chadefaux B, Montalescot G & Kamoun P (1993) Elevated total plasma homocysteine, a risk factor for thrombosis. Relation to coagulation and fibrinolytic parameters. Thrombosis Research 70, 123–129. Blaak EE & Saris WHM (1995) Health aspects of various digestible carbohydrates. Nutrition Research 15, 1547–1573. Bladbjerg EM, Tholstrup T, Marckmann P, Sandström B & Jespersen J (1995) Dietary changes in fasting levels of factor VII coagulant activity (FVII:C) are accompanied by changes in factor VII protein and other vitamin K-dependent proteins. Thrombosis and Haemostasis 73, 239–242. Blann AD, Jackson P, Bath PMW & Watts GF (1995) Von Willebrand factor, a possible indicator of endothelial cell damage, decreases during long-term compliance with a lipidlowering diet. Journal of Internal Medicine 237, 557–561. Blann A, Kumar P, Krupsinski J, Lip G & Beevers D (1996) Endothelial damage and adhesion molecules ICAM, VCAM, and Eselectin in acute stroke. Blood Coagulation and Fibrinolysis 7, 377. Blann AD & McCollum CN (1994) Circulating endothelial cell/ leukocyte adhesion molecules in atherosclerosis. Thrombosis and Haemostasis 72, 151–154. Bønaa K, Bjerve K, Straume B, Gram I & Thelle D (1990) Effect of eicosapentaenoic and docosahexaenoic acids on blood pressure in hypertension. New England Journal of Medicine 322, 795–801. Borkman M, Storlien LH, Pan DA, Jenkins AD, Chisholm DJ & Campbell LV (1993) The relation between insulin sensitivity and the fatty-acid composition of skeletal muscle phospholipids. New England Journal of Medicine 328, 238–244. Bostom AG, Cupples LA, Jenner JL, Ordovas JM, Seman LJ, Wilson PW, Schaefer EJ & Castelli WP (1996) Elevated plasma lipoprotein(a) and coronary heart disease in men aged 55 years and younger. A prospective study. Journal of the American Medical Association 276, 544–548. Boulanger CM, Tanner FC, Hahn A, Wa A, Werner A & Lüscher TF (1992) Oxidised low-density lipoproteins induce mRNA expression and release of endothelin from human and porcine endothelium. Circulation Research 70, 1191–1197. Boushey CJ, Beresford SA, Omenn GS & Motulsky AG (1995) A quantitative assessment of plasma homocysteine as a risk factor for vascular disease. Probable benefits of increasing folic acid intakes [see comments]. Journal of the American Medical Association 274, 1049–1057. Bronsgeest-Schoute DC, Hautvast JG & Hermus RJ (1979) Dependence of the effects of dietary cholesterol and experimental conditions on serum lipids in man. I. Effects of dietary cholesterol in a linoleic acid-rich diet. American Journal of Clinical Nutrition 32, 2183–2187. Brown AS & Martin JF (1994) The megakaryocyte platelet system and vascular disease. European Journal of Clinical Investigation 24, Suppl. 1, 9–15. Burr ML, Fehily AM, Gilbert JF, Rogers S, Holliday RM, Sweetnam PM & Deadman NM (1989) Effect of changes in fat, fish and fibre intakes on death and myocardial infarction: diet and reinfarction trial (DART). Lancet ii, 757–761. Cairns JA, Gill J, Morton B, Roberts R, Gent M, Hirsch J, Holder D, Finnie K, Marquis JF, Naqvi S & Cohen E (1996) Fish oils and low molecular-weight heparin for the reduction of restenosis after percutaneous transluminal coronary angioplasty. The EMPAR study. Circulation 94, 1553–1560. S138 G. Hornstra et al. Calder PC (1996) Immunomodulatory and anti-inflammatory effects of n-3 polyunsaturated fatty acids. Proceedings of the Nutrition Society 55, 737–774. Calder PC & Newsholme EA (1992) Polyunsaturated fatty acids suppress human peripheral blood lymphocyte proliferation and interleukin-2 production. Clinical Science 82, 695–700. Cambien F (1987) Is insulin the key factor to explain the associations between body mass, blood pressure and glucose? Diabète et Métabolisme 13, 973–978. Cambien F, Poirier O, Lecerf L, Evans A, Cambou JP, Arveiler D, Luc G, Bard JM, Bara L & Ricard S (1992) Deletion polymorphism in the gene for angiotensin converting enzyme is a potent risk factor for myocardial infarction. Nature 359, 641–644. Castelli WP, Anderson K, Wilson PW & Levy D (1992) Lipids and risk of coronary heart disease. The Framingham Study. Annals of Epidemiology 2, 23–28. Cater NB, Heller HJ & Denke MA (1997) Comparison of the effects of medium chain triacylglycerols, palm oil, and high oleic acid sunflower oil on plasma triacylglycerol fatty acids and lipid and lipoprotein concentrations in humans. American Journal of Clinical Nutrition 65, 41–45. Chalmers TC, Matta RJ, Smith HR Jr & Kunzler AM (1977) Evidence favoring the use of anticoagulants in the hospital phase of acute myocardial infarction. New England Journal of Medicine 297, 1091–1096. Chan JK, Bruce VM & McDonald BE (1991) Dietary a-linolenic acid is as effective as oleic acid and linoleic acid in lowering blood cholesterol in normolipidemic men. American Journal of Clinical Nutrition 53, 1230–1234. Charnock JS (1994) Lipids and cardiac arrhythmias. Progress in Lipid Research 33, 355–385. Chasan-Taber L, Selhub J, Rosenberg IH, Malinow MR, Terry P, Tishler PV, Willett W, Hennekens CH & Stampfer MJ (1996) A prospective study of folate and vitamin B6 and risk of myocardial infarction in US physicians. Journal of the American College of Nutrition 15, 136–143. Choudhury N, Tan L & Truswell AS (1995) Comparison of palmolein and olive oil: effects on plasma lipids and vitamin E in young adults. American Journal of Clinical Nutrition 61, 1043–1051. Choudhury SR, Ueshima H, Kita Y, Kobayashi KM, Okayama A, Yamakawa M, Hirao Y, Ishikawa M & Miyoshi Y (1994) Alcohol intake and serum lipids in a Japanese population. International Journal of Epidemiology 23, 940–947. Clarke R, Frost C, Collins R, Appleby P & Peto R (1997) Dietary lipids and blood cholesterol: quantitative meta-analysis of metabolic ward studies. British Medical Journal 314, 112–117. Clarke R, Naughten E, Cahalane S, Sullivan KO, Mathias P, McCall T & Graham I (1992) The role of free radicals as mediators of endothelial cell injury in hyperhomocysteinaemia. Irish Journal of Medical Science 161, 561–564. Collie-Duguid ESR & Wahle KWJ (1996) Inhibitory effect of fish oil n-3 polyunsaturated fatty acids on the expression of endothelial cell adhesion molecules. Biochemical and Biophysical Research Communications 220, 969–974. Collins R, Peto R, MacMahon S, Hebert P, Fiebach NH, Eberlein KA, Godwin J, Qizilbash N, Taylor JO & Hennekens CH (1990) Blood pressure, stroke, and coronary heart disease. Part 2. Lancet 335, 827–838. Colman E, Katzel LI, Rogus E, Coon P, Muller D & Goldberg AP (1995) Weight loss reduces abdominal fat and improves insulin action in middle-aged and older men with impaired glucose tolerance. Metabolism 44, 1502–1508. Curb JD & Reed DM (1985) Fish consumption and mortality from coronary heart disease. New England Journal of Medicine 313, 821–822. Curmi PA, Juan L & Tedgui A (1990) Effect of transmural pressure on low density lipoprotein and albumin transport and distribution across the intact arterial wall. Circulation Research 66, 1692–1702. Dalery K, Lussier CS, Selhub J, Davignon J, Latour Y & Genest JJ (1995) Homocysteine and coronary artery disease in French Canadian subjects: relation with vitamins B12, B6, pyridoxal phosphate, and folate. American Journal of Cardiology 75, 1107–1111. Davidson MH, Maki KC, Kalkowski J, Schaefer EJ, Torri SA & Drennan KB (1997) Effects of docosahexaenoic acid on serum lipoproteins in patients with combined hyperlipidemia. A randomized, double-blind, placebo-controlled trial. Journal of the American College of Nutrition 16, 236–243. Davies MJ (1997) The composition of coronary artery plaques. New England Journal of Medicine 336, 1312–1314. Daviglus ML, Stamler J, Orencia AJ, Dyer AR, Liu K, Greenland P, Walsh MK, Morris D & Shekelle RB (1997) Fish consumption and the 30-year risk of fatal myocardial infarction. New England Journal of Medicine 336, 1046–1053. De Caterina R, Cybulski MI, Clinton SK, Gimbrone MA Jr & Libby P (1994) The omega-3 fatty acid docosahexaenoate reduces cytokine-induced expression of proatherogenic and proinflammatory proteins in human endothelial cells. Arteriosclerosis and Thrombosis 14, 1829–1836. De Caterina R, Cybulski MA, Clinton SK, Gimbrone MA & Libby P (1995a) Omega-3 fatty acids and endothelial leukocyte adhesion molecules. Prostaglandins, Leukotrienes and Essential Fatty Acids 52, 191–195. De Caterina R, Tanaka H, Nakagawa T, Hauptman PJ & Libby P (1995b) The direct effect of injectable cyclosporine and its vehicle, cremophor, on endothelial vascular cell adhesion molecule-1 expression. Transplantation 60, 270–275. De Deckere EAM & ten Hoor F (1979) Effects of dietary fats on the coronary flow rate and the left ventricular function of the isolated rat heart. Nutrition and Metabolism 23, 88–97. DeForge LE, Fantone JC, Kenney JS & Remick DG (1992) Oxygen radical scavengers selectively inhibit interleukin 8 production in human whole blood. Journal of Clinical Investigation 90, 2123–2129. DeFronzo RA (1981) The effect of insulin on renal sodium metabolism: a review with clinical implications. Diabetologia 21, 165–171. de Lorgeril M, Renaud S, Mamelle N, Salen P, Martin J-L, Monjaud I, Guidollet J, Touboul P & Delaye J (1994) Mediterranean alpha-linolenic acid-rich diet in secondary prevention of coronary heart disease. Lancet 343, 1454–1459. Desprès JP & Marette A (1994) Relation of components of insulin resistance syndrome to coronary disease risk. Current Opinion in Lipidology 5, 274–289. Diplock AT, Charleux J-L, Crozier-Willi G, Kok FJ, Rice-Evans C, Roberfroid M, Stahl W & Viña-Ribes J (1998) Functional food science and defence against reactive oxygen species. British Journal of Nutrition 80, Suppl. 1, S77–S112. Dolecek TA & Grandits G (1991) Dietary polyunsaturated fatty acids and mortality in the Multiple Risk Factor Intervention Trial (MRFIT). World Review of Nutrition and Dietetics 66, 205–216. Dreon DM, Fernstrom HA, Miller B & Krauss RM (1995) Apolipoprotein E isoform phenotype and LDL subclass response to a reduced-fat diet. Arteriosclerosis, Thrombosis and Vascular Biology 15, 105–111. Dudman NP, Guo XW, Gordon RB, Dawson PA & Wilcken DE (1996) Human homocysteine catabolism: three major pathways and their relevance to development of arterial occlusive disease. Journal of Nutrition 126, 1295S-1300S. Dudman NP, Hicks C, Wang J & Wilcken DE (1991) Human endothelial cell detachment in vitro: its promotion by homocysteine and cysteine. Atherosclerosis 91, 77–83. Cardiovascular system Elliott TG & Viberti G (1993) Relationship between insulin resistance and coronary heart disease in diabetes mellitus and the general population: a critical appraisal. Baillière’s Clinical Endocrinology and Metabolism 7, 1079–1103. Ellis JM & McCully KS (1995) Prevention of myocardial infarction by vitamin B6. Research Communications in Molecular Pathology and Pharmacology 89, 208–220. Elwood PC, Beswich AD, Sharp DS, Yarnell JW, Rogers S & Renaud S (1990) Whole blood impedance platelet aggregometry and ischaemic heart disease. The Caerphilly Collaborative Heart Disease Study. Arteriosclerosis 10, 1032–1036. Elwood PC, Renaud S, Sharp DS, Beswick AD, O’Brien JR & Yarnell JW (1991) Ischemic heart disease and platelet aggregation. The Caerphilly Collaborative Heart Disease Study. Circulation 83, 38–44. Endres S, De Caterina R, Schmidt EB & Kristensen SD (1995a) N-3 polyunsaturated fatty acids: update 1995. European Journal of Clinical Investigation 25, 629–638. Endres S, Eisenhut T & Sinha B (1995b) N-3 polyunsaturated fatty acids in the regulation of human cytokine synthesis. Biochemical Society Transactions 23, 277–281. Endres S, Meydani SN, Ghorbani R, Schindler R & Dinarello CA (1993) Dietary supplementation with n-3 fatty acids suppresses interleukin-2 production and mononuclear cell proliferation. Journal of Leukocyte Biology 54, 599–603. EPIC Investigators (1994) Use of monoclonal antibody directed against the platelet glycoprotein IIb/IIIa receptor in high-risk coronary angioplasty. New England Journal of Medicine 330, 956–961. Eritsland J, Seljeflot I, Abdelnoor M & Arnesen H (1994) Longterm influence of omega-3 fatty acids on fibrinolysis, fibrinogen, and serum lipids. Fibrinolysis 8, 120–125. Eritsland J, Arnesen H, Seljeflot I & Kierulf P (1995) Long-term effects of n-3 polyunsaturated fatty acids on haemostatic variables and bleeding episodes in patients with coronary artery disease. Blood Coagulation and Fibrinolysis 6, 17–22. Faruqi R, de la Motte C & Di Corleto P (1994) a-Tocopherol inhibits agonist-induced monocytic cell adhesion to cultured human endothelial cells. Journal of Clinical Investigation 94, 592–600. Ferrannini E, Haffner SL, Mitchell BD & Sterm MP (1991) Hyperinsulinemia: the key feature of a cardiovascular and metabolic syndrome. Diabetologia 3, 416–422. Feskens EJM, Loeber JG & Kromhout D (1994) Diet and physical activity as determinants of hyperinsulinemia: the Zutphen Elderly Study. American Journal of Epidemiology 140, 350– 360. Fibrinolytic Trialists’ Collaborative Group (1994) Indications for fibrinolytic therapy in suspected acute myocardial infarction: collaborative overview of early mortality and major morbidity events from the randomized trials of more than 1000 patients. Lancet 343, 311–322. Fischer S, Weber PC & Dyerberg J (1986) The prostacyclin/ thromboxane balance is favourably shifted in Greenland Eskimos. Prostaglandins 32, 235–240. Fontbonne A, Charles MS & Thibult N, Richard JL, Claude JR, Warnet JM, Rosselin GE & Eschwege E (1991) Hyperinsulinaemia as a predictor of coronary heart disease mortality in a healthy population: the Paris Prospective Study 15-year followup. Diabetologia 34, 356–361. Frayn KN (1993) Insulin resistance and lipid metabolism. Current Opinion in Lipidology 4, 197–204. Freese R & Mutanen M (1997) Alpha-linolenic acid and marine n-3 fatty acids only slightly differ in their effects on hemostatic factors in healthy subjects. American Journal of Clinical Nutrition 66, 591–598. Frenette PS & Wagner DD (1996) Adhesion molecules – Part II: S139 Blood vessels and blood cells. New England Journal of Medicine 335, 43–45. Frick MH, Elo O, Haapa K, Heinonen OP, Heinsalmi P, Helo P, Huttunen JK, Kaitaniemi P, Koskinen P, Manninen V, Mäenpää H, Mälkönen M, Mänttäri M, Norola S, Pasternack A, Pikkarainen J, Romo M, Sjöblom T & Nikkilä EA (1987) Helsinki Heart Study: primary-prevention trial with gemfibrozil in middle-aged men with dyslipidemia. New England Journal of Medicine 317, 1237–1245. Frosst P, Blom HJ, Milos R, Goyette P, Sheppard CA, Matthews RG, Boers GJ, den Heyer M, Kluijtmans LA, van den Heuvel LP & Rozen R (1995) A candidate genetic risk factor for vascular disease: a common mutation in methylenetetrahydrofolate reductase (letter). Nature Genetics 10, 111–113. Frøyland L, Madsen L, Vaagenes H, Totland GK, Auwerx J, Kryvi H, Staels B & Berge RK (1997) Mitochondrion is the principal target for nutritional and pharmacological control of triglyceride metabolism. Journal of Lipid Research 38, 1851–1858. Fryer RH, Wilson BD, Gubler DB, Fitzgerald LA & Rodgers GM (1993) Homocysteine, a risk factor for premature vascular disease and thrombosis, induces tissue factor activity in endothelial cells. Arteriosclerosis and Thrombosis 13, 1327–1333. Fujikawa, M, Yamashita N, Yamazaki K, Sugiyama E, Suzuki H & Hamazaki T (1992) Eicosapentaenoic acid inhibits antigenpresenting cell function of murine splenocytes. Immunology 75, 330–335. Fulton M, Thomson M, Elton RA, Brown S, Wood DA & Oliver MF (1988) Cigarette smoking, social class and nutrient intake. Relevance to coronary heart disease. European Journal of Clinical Nutrition 42, 797–803. Fuster V, Badimon L, Badimon JJ & Chesebro JH (1992) The pathogenesis of coronary artery disease and the acute coronary syndromes. New England Journal of Medicine 326, 242–250. Fuster V, Stein B, Ambrose JA, Badimon JJ & Chesebro JH (1990) Atherosclerotic plaque rupture and thrombosis: evolving concepts. Circulation 82, Suppl. II, 47–59. Fyfe AI, Qiao JH & Lusis AJ (1994) Immune-deficient mice develop typical atherosclerotic fatty streaks when fed an atherogenic diet. Journal of Clinical Investigation 94, 2516–2520. Gapinski JP, van Ruiswyk JV, Heudebert GR & Schectman GS (1993) Preventing restenosis with fish oils following coronary angioplasty. A meta-analysis. Archives of Internal Medicine 153, 1595–1601. Geng Y-J & Hansson GK (1992) Interferon-g inhibits scavenger receptor expression and foam cell formation in human monocyte-derived macrophages. Journal of Clinical Investigation 89, 1322–1330. Gey KF, Moser UK, Jordan P, Stahelin HB, Eichholzer M & Ludin E (1993) Increased risk of cardiovascular disease at suboptimal plasma concentrations of essential antioxidants: an epidemiological update with special attention to carotene and vitamin C. American Journal of Clinical Nutrition 57, Suppl., 787–797. Ghaddar HM, Cortes J, Salomaa V, Kark JD, Davis CE, Folsom AR, Heiss G, Stinson V & Wu KK (1995). Correlation of specific platelet activation markers with carotid arterial wall thickness. Thrombosis and Haemostasis 74, 984–948. Glore SR, Van Treeck D, Knehans AW & Guild M (1994) Soluble fiber and serum lipids: a literature review. Journal of the American Dietetic Association 94, 425–436. Goldberg DM (1996) More on antioxidant activity of resveratrol in red wine (letter). Clinical Chemistry 42, 113–114. Gordon D, Reidy MA, Benditt EP & Schwartz SM (1990) Cell proliferation in human coronary arteries. Proceedings of the National Academy of Sciences USA 87, 4600–4604. Gorgels WJMJ, van der Graaf Y, Hjemdahl P, Kortlandt W, Collette HJA, Erkelens DW & Banga J-D (1995) Urinary excretions of high molecular weight b-thromboglobulin and S140 G. Hornstra et al. albumin are independently associated with coronary heart disease in women, a nested case–control study of middle-aged women in the diagnostisch onderzoek mammacarcinoom (DOM) cohort, Utrecht, Netherlands. American Journal of Epidemiology 142, 1157–1164. Goyette P, Frosst P, Rosenblatt DS & Rozen R (1995) Seven novel mutations in the methylenetetrahydrofolate reductase gene and genotype/phenotype correlations in severe methylenetetrahydrofolate reductase deficiency. American Journal of Human Genetics 56, 1052–1059. Grainger DJ & Metcalfe JC (1995) Transforming growth factorbeta: the key to understanding lipoprotein(a)? Current Opinion in Lipidology 6, 81–85. Grande F, Anderson JT & Keys A (1970) Comparison of effects of palmitic and stearic acids in the diet on serum cholesterol in man. American Journal of Clinical Nutrition 23, 1184–1193. Grimsgaard S, Bønaa KH, Hansen JB & Nordøy A (1997) Highly purified eicosapentaenoic and docosahexaenoic acid in humans have similar triacylglycerol-lowering effects but divergent effects on serum fatty acids. American Journal of Clinical Nutrition 66, 649–659. Groppelli A, Giorgi DMA, Omboni S, Parati G & Mancia G (1992) Persistent blood pressure increase induced by heavy smoking. Journal of Hypertension 10, 495–499. Guttormsen AB, Schneede J, Ueland PM & Refsum H (1996) Kinetics of total plasma homocysteine in subjects with hyperhomocysteinemia due to folate or cobalamin deficiency. American Journal of Clinical Nutrition 63, 194–202. Hackman A, Abe Y, Insull W Jr, Pownall H, Smith L, Dunn K, Gotto AM Jr & Ballantyne CM (1996) Levels of soluble cell adhesion molecules in patients with dyslipidemia. Circulation 93, 1334–1338. Hajjar KA (1993) Homocysteine induced modulation of tissue plasminogen activator binding to its endothelial cell membrane receptor. Journal of Clinical Investigation 91, 2873–2879. Hamano H, Nanba A, Nagayama M, Takizawa S & Shinohara Y (1996) [An adult case of homocystinuria due to methylenetetrahydrofolate-reductase deficiency–treatment with folic acid and the course of coagulation-fibrinolysis parameters.] RinshoShinkeigaku 36, 330–335. Hamazaki T, Fischer S, Urakaza M, Sawazaki S, Yano S & Kuwamori T (1989) Urinary excretion of PGI2/3-M and recent n-6/3 fatty acid intake. Prostaglandins 37, 417–424. Hamsten A (1995) Hemostatic function and coronary artery disease. New England Journal of Medicine 332, 677–678. Hansen JB, Huseby NE, Sandset PM, Svensson B, Lyngmo V & Nordøy A (1994) Tissue-factor pathway inhibitor and lipoproteins. Evidence for association with and regulation by LDL in human plasma. Arteriosclerosis and Thrombosis 14, 223–229. Hansson GK, Holm J, Holm S, Fotev Z, Hedric HJ & Fingerle J (1991) T-lymphocytes inhibit the vascular response to injury. Proceedings of the National Academy of Sciences USA 88, 10530–10534. Harker LA, Slichter SJ, Scott CR & Ross R (1974) Homocysteinemia, vascular injury and arterial thrombosis. New England Journal of Medicine 291, 537–543. Harpel PC & Borth W (1992) Identification of mechanisms that may modulate the role of lipoprotein(a) in thrombosis and atherogenesis. Annals of Epidemiology 2, 413–417. Harpel PC, Zhang X & Borth W (1996) Homocysteine and hemostasis: pathogenic mechanisms predisposing to thrombosis. Journal of Nutrition 126, 1285S–1289S. Harris WS (1997) n-3 fatty acids and serum lipoproteins: human studies. American Journal of Clinical Nutrition 65, 1645S– 1654S. Harris WS & Muzio F (1993) Fish oil reduces postprandial triglyceride concentrations without accelerating lipid-emulsion removal rates. American Journal of Clinical Nutrition 58, 68– 74. Hayes KC & Khosla P (1992) Dietary fatty acid thresholds and cholesterolemia. FASEB Journal 6, 2600–2607. Hayes KC, Khosla P, Pronczuk A & Lindsey S (1992) Re-examination of the dietary fatty acid–plasma cholesterol issue: is palmitic acid (16 : 0) neutral? In Cholesterol and Coronary Heart Disease – The Great Debate, pp. 189–206 [P Gold, S Grover and DAK Roncari, editors]. Carnforth: Parthenon Publishing Group. Hayford JT, Danney MM, Wiebe MCD, Roberts S & Thompson RG (1979) Triglyceride integrated concentrations: effect of variation of source and amount of dietary carbohydrate. American Journal of Clinical Nutrition 32, 1670–1678. Healy B (1990) Endothelial cell dysfunction: an emerging endocrinopathy linked to coronary artery disease. Journal of the American College of Cardiology 16, 357–358. Hegsted DM, McGandy RB, Myers ML & Stare FJ (1965) Quantitative effects of dietary fat on serum cholesterol in man. American Journal of Clinical Nutrition 17, 281–295. Heijnen ML, van Amelsvoort JM, Deurenberg P & Beynen AC (1996) Neither raw nor retrograded resistant starch lowers fasting serum cholesterol concentrations in healthy normolipidemic subjects. American Journal of Clinical Nutrition 64, 312– 318. Hellsten G, Boman K, Saarem K, Hallmans G & Nilsson T (1993) Effects on fibrinolytic activity of corn oil and a fish oil preparation enriched with omega-3 polyunsaturated fatty acids in a long-term study. Current Medical Research Opinion 13, 133– 139. Henning B, Toborek M, Joshi-Barve S, Barger SW, Barve S, Mattson MP & McClain CJ (1996) Linoleic acid activates nuclear transcription factor-kB (NF-kB) and induces NF-kBdependent transcription in cultured endothelial cells. American Journal of Clinical Nutrition 63, 322–328. Herzlich BC, Lichstein E, Schulhoff N, Weinstock M, Pagala M, Ravindran K, Namba T, Nieto FJ, Stabler SP & Allen RH (1996) Relationship among homocyst(e)ine, vitamin B-12 and cardiac disease in the elderly: association between vitamin B-12 deficiency and decreased left ventricular ejection fraction. Journal of Nutrition 126, 1249S–1253S. Hill-Zobel RL, Pyeritz RE, Scheffel E, Malpica O, Engin S, Camargo EE, Abott M, Guilarte TR, Hill J, McIntyre PA, Murphy EA & Tsan MF (1982) Kinetics and distribution of 111 indium-labeled platelets in patients with homocystinuria. New England Journal of Medicine 307, 781–786. Himsworth HP (1936) Diabetes mellitus: its differentiation into insulin-sensitive and insulin-insensitive types. Lancet i, 127– 130. Hokanson JE & Austin MA (1996) Plasma triglyceride level is a risk factor for cardiovascular disease independent of highdensity lipoprotein cholesterol level: a meta-analysis of population-based prospective studies. Journal of Cardiovascular Risk 3, 213–219. Hopkins PN (1992) Effects of dietary cholesterol on serum cholesterol: a meta-analysis and review. American Journal of Clinical Nutrition 55, 1060–1070. Hornstra G (1989) The significance of fish and fish-oil enriched food for prevention and therapy of ischaemic cardiovascular disease. In The Role of Fats in Human Nutrition II, pp. 151–235 [AJ Vergroesen and MA Crawford, editors]. New York, NY: Academic Press. Hornstra G (1992) The effect of the n-3 fatty acids on blood coagulation. In Fish Oil and Vascular Disease, pp. 65–72 [R DeCaterina, SD Kristensen and EB Schmidt, editors]. Berlin: Springer-Verlag. Hornstra G, Hennissen AAHM & Wierts JW (1993) The effect of Cardiovascular system dietary fatty acids on platelet function and arterial thrombosis. In Essential Fatty Acids and Eicosanoids, pp. 287–289 [A Sinclair and R Gibson, editors]. Champaign, IL: American Oil Chemist’s Society. Hornstra G, van Houwelingen AC, Kivits GAA, Fischer S & Uedelhoven W (1990) Influence of dietary fish on eicosanoid metabolism in man. Prostaglandins 40, 311–329. Horrobin DF (1993) Fatty acid metabolism in health and disease: the role of D6-desaturase. American Journal of Clinical Nutrition 57, Suppl., 732S–737S. Hultberg B, Andersson A & Isaksson A (1995) Metabolism of homocysteine, its relation to the other cellular thiols and its mechanism of cell damage in a cell culture line (human histiocytic cell line U-937). Biochimica et Biophysica Acta 1269, 6–12. Hunter DJ, Kazda T, Chokalingam A & Fodor JG (1988) Fish consumption and cardiovascular mortality in Canada: an interregional comparison. American Journal of Preventive Medicine 4, 5–10. Jacobsen DW, Gatautis VJ, Green R, Robinson K, Savon SR, Secic M, Ji J, Otto JM & Taylor LJ (1994) Rapid HPLC determination of total homocysteine and other thiols in serum and plasma: sex differences and correlation with cobalamin and folate concentrations in healthy subjects [see comments]. Clinical Chemistry 40, 873–881. Jacques PF, Bostom AG, Williams RR, Ellison RC, Eckfeldt JH, Rosenberg IH, Selhub J & Rozen R (1996) Relation between folate status, a common mutation in methylenetetrahydrofolate reductase, and plasma homocysteine concentrations. Circulation 93, 7–9. Jandacek RJ, Ramirez MM & Crouse JR (1990) Effects of partial replacement of dietary fat by olestra on dietary cholesterol absorption in man. Metabolism 39, 848–852. Jeppesen J, Chen YDI, Zhou MY, Wang T & Reaven GM (1995) Effect of variations in oral fat and carbohydrate load on postprandial lipemia. American Journal of Clinical Nutrition 62, 1201–1205. Jeunemaitre X, Lifton RP, Hunt SC, Williams RR & Lalouel J-M (1992) Absence of linkage between angiotensin converting enzyme locus and human essential hypertension. Nature Genetics 1, 72–75. Jonasson L, Holm J, Skalli O, Bondjers G & Hansson GK (1986) Regional accumulation of T cells, macrophages, and smooth muscle cells in the human atherosclerotic plaque. Arteriosclerosis 6, 131–138. Jones BG, Rose FA & Tudball N (1994) Lipid peroxidation and homocysteine induced toxicity. Atherosclerosis 105, 165–170. Kaminski WE, Jendraschak E, Kiefl R & von Shacky C (1993) Dietary q-3 fatty acids lower levels of platelet-derived growth factor mRNA in human mononuclear cells. Blood 81, 1871– 1879. Kardinaal AFM, Aro A, Kark JD, Riemersma RA, van’t Veer P, Gomez-Aracena J, Kohlmeier L, Ringstad J, Martin BC, Mazaev VP, Delgado-Rodriguez M, Thamm M, Huttunen JK, MartinMoreno JM & Kok FJ (1995) Association between b-carotene and acute myocardial infarction depends on polyunsaturated fatty acid status. The Euramic Study. Arteriosclerosis, Thrombosis and Vascular Biology 15, 726–732. Karpe F, Steiner G, Uffelman K, Olivecrona T & Hamsten A (1994) Postprandial lipoproteins and progression of coronary atherosclerosis. Atherosclerosis 106, 83–97. Katan MB, Beynen AC, de Vries JHM & Nobels A (1986) Existence of consistent hypo- and hyperresponders to dietary cholesterol in man. American Journal of Epidemiology 123, 221–234. Katan MB, Zock PL & Mensink RP (1995) Trans fatty acids and their effects on lipoproteins in humans. Annual Reviews of Nutrition 15, 473–493. S141 Kelley DS, Nelson GJ, Serrato CM, Schmidt PC & Branch LB (1988) Effects of type of dietary fat on indices of immune status of rabbits. Journal of Nutrition 118, 1376–1384. Kelly JP & Parker CW (1979) Effect of arachidonic acid and other unsaturated fatty acids on mitogenesis in human lymphocytes. Journal of Immunology 122, 1556–1562. Kesaniemi YA & Grundy SM (1982) Lack of effect of tocopherol on plasma lipids and lipoproteins in man. American Journal of Clinical Nutrition 36, 224–228. Keys A, Anderson JT & Grande F (1965a) Serum cholesterol response to changes in the diet II. The effect of cholesterol in the diet. Metabolism 14, 759–765. Keys A, Anderson JT & Grande F (1965b) Serum cholesterol response to changes in the diet IV. Particular saturated fatty acids in the diet. Metabolism 14, 776–786. Khan WA, Blobe GC & Hannun YA (1995) Arachidonic acid and free fatty acids as second messengers and the role of protein kinase C. Cellular Signalling 7, 171–184. Knapp HR, Rielly IAG, Allessandrini P & FitzGerald GA (1986) In vivo indexes of platelet and vascular function during fish-oil administration in patients with atherosclerosis. New England Journal of Medicine 314, 937–942. Kokame K, Kato H & Miyata T (1996) Homocysteine-respondent genes in vascular endothelial cells identified by differential display analysis. GRP78/BiP and novel genes. Journal of Biological Chemistry 271, 29659–29665. Kritchevsky D & Tepper SA (1977) Cholesterol vehicle in experimental atherosclerosis. Part 15. Randomized butter and randomized lard. Atherosclerosis 27, 339–345. Kritchevsky D, Tepper SA, Vesselinovitch D & Wissler RW (1973) Cholesterol vehicle in experimental atherosclerosis. Randomized peanut oil. Atherosclerosis 17, 225–243. Kromhout D (1985) Fish consumption and mortality from coronary heart disease. New England Journal of Medicine 313, 824. Kromhout D, Bosschieter EB & de Lezenne Coulander C (1985) The inverse relation between fish consumption and 20-year mortality from coronary heart disease. New England Journal of Medicine 312, 1205–1209. Landsberg L & Krieger DR (1989) Obesity, metabolism, and the sympathetic nervous system. American Journal of Hypertension 2, 125S–132S. Lapidus L, Andersson H, Bengtsson C & Bosaeus I (1986) Dietary habits in relation to incidence of cardiovascular disease and death in women: a 12 year follow-up of participants in the population study of women in Gothenburg, Sweden. American Journal of Clinical Nutrition 44, 444–448. Laws A & Reaven GM (1993) Insulin resistance and the risk factors for coronary heart disease. Baillière’s Clinical Endocrinology and Metabolism 7, 1079–1103. Lentz SR, Sobey CG, Piegors DJ, Bhopatkar MY, Faraci FM, Malinow MR & Heistad DD (1996) Vascular dysfunction in monkeys with diet-induced hyperhomocyst(e)inemia [see comments]. Journal of Clinical Investigation 98, 24–29. Leprán I, Nemecz GY, Koltai M & Szekeres L (1981) Effect of linoleic acid rich diet on the acute phase of coronary occlusion in conscious rats: influence of indomethacin and aspirin. Journal of Cardiovascular Pharmacology 3, 847–853. Lewis GF, O’Meara NM, Soltys PA, Blackman JD, Iverius PH, Pugh WL, Getz GS & Polonsky KS (1990) Postprandial lipoprotein metabolism in normal and obese subjects: comparison after the vitamin A fat-loading test. Journal of Clinical Endocrinology and Metabolism 71, 1041–1050. Lillioja S, Young AA, Culter CA, Ivy JL, Abbott WG, Zawadzki JK, Yki-Jarvinen H, Christin L, Secomb TW & Bogardus C (1987) Skeletal muscle capillary density and fibre type are possible determinants of in vivo insulin resistance in man. Journal of Clinical Investigation 80, 415–424. S142 G. Hornstra et al. Lindgren A, Lindoff C, Norrving B, Astedt B & Johansson BB (1996) Tissue plasminogen activator and plasminogen activator inhibitor-1 in stroke patients. Stroke 27, 1066–1071. Ling WH & Jones PJ (1995) Dietary phytosterols: a review of metabolism, benefits and side effects. Life Science 5, 195–206. Lissner L, Bengtsson C, Lapidus L, Kristjansson K & Wedel H (1992) Fasting insulin in relation to subsequent blood pressure changes and hypertension in women. Hypertension 20, 797–801. Lodish H, Baltimore D, Berk A, Zipurski SL, Matsudavia M & Darnell J (1995) Chapter 27. Immunity. In Molecular Cell Biology, 3rd ed. New York, NY: Scientific American Books. Loeliger EA (1984) Oral anticoagulation in patients surviving myocardial infarction. A new approach to old data. European Journal of Clinical Pharmacology 26, 137–139. Lopez-Miranda J, Ordovas JM, Espino A, Marin C, Salas J, LopezSegura F, Jimenez-Pereperez J & Perez-Jimenez F (1994) Influence of mutation in human apolipoprotein A-1 gene promoter on plasma LDL cholesterol response to dietary fat. Lancet 343, 1246–1249. Lopez-Segura F, Velasco F, Lopez-Miranda J, Castro P, LopezPedrera R, Blanco A, Jimenez-Pereperez J, Torres A, Trujillo J, Ordovas JM & Perez-Jimenez F (1996) Monounsaturated fatty acid-enriched diet decreases plasma plasminogen activator inhibitor type 1. Arteriosclerosis, Thrombosis and Vascular Biology 16, 82–88. Loscalzo J, Pasche B, Ouimet H & Freedman JE (1995) Platelets and plasminogen activation. Thrombosis and Haemostasis 74, 291–293. Lucas CP, Estigarribia JA, Darga LL & Reaven GM (1985) Insulin and blood pressure in obesity. Hypertension 7, 702–706. Ma J, Stampfer MJ, Hennekens CH, Frosst P, Selhub J, Horsford J, Malinow MR, Willett WC & Rozen R (1996) Methylenetetrahydrofolate reductase polymorphism, plasma folate, homocysteine, and risk of myocardial infarction in US physicians. Circulation 94, 2410–2416. McCully KS (1993) Chemical pathology of homocysteine. I. Atherogenesis. Annals of Clinical and Laboratory Sciences 23, 477–493. McGill HC Jr (1984) Persistent problems in the pathogenesis of atherosclerosis: atherogenesis and inflammation. Arteriosclerosis 4, 443–451. McLennan PL (1993) Relative effects of dietary saturated, monounsaturated, and polyunsaturated fatty acids on cardiac arrhythmias in rats. American Journal of Clinical Nutrition 57, 207– 212. McLennan PL, Abeywardena MY & Charnock JS (1985) Influence of dietary lipid on arrhythmias and infarction after coronary ligation in rats. Canadian Journal of Physiology and Pharmacology 63, 1411–1417. McLennan PL, Abeywardena MY & Charnock JS (1989) The influence of age and dietary fat in an animal model of sudden cardiac death. Australian and New Zealand Journal of Medicine 19, 1–5. McLennan PL, Abeywardena MY & Charnock JS (1990) Reversal of the arrhythmogenic effects of long-term saturated fatty acid intake by dietary n-3 and n-6 polyunsaturated fatty acids. American Journal of Clinical Nutrition 51, 53–58. McLennan P & Dallimore JA (1995) Dietary canola oil modifies myocardial fatty acids and inhibits cardiac arrhythmias in rats. Journal of Nutrition 125, 1003–1009. MacLeod DC, Heagerty AM, Bund SJ, Lawal TS & Riemersma RA (1994) Effect of dietary polyunsaturated fatty acids on contraction and relaxation of rat femoral resistance arteries. Journal of Cardiovascular Pharmacology 23, 92–98. MacMahon S, Peto R, Cutler J, Collins R, Sorlie P, Neaton J, Abbott R, Godwin J, Dyer A & Stamler J (1990) Blood pressure, stroke, and coronary heart disease. Part 1. Lancet 335, 765–774. Malinow MR (1996) Plasma homocyst(e)ine: a risk factor for arterial occlusive diseases. Journal of Nutrition 126, 1238S– 1243S. Manninen V, Tenkanen L, Koskinen P, Huttunen JK, Manttari M, Heinonen OP & Frick MH (1992) Joint effects of serum triglyceride and HDL cholesterol concentrations on coronary heart disease risk in the Helsinki Heart Study. Implications for treatment. Circulation 85, 37–45. Marckmann P, Sandström B & Jespersen J (1993) Favorable longterm effect of a low-fat/high-fiber diet on human blood coagulation and fibrinolysis. Arteriosclerosis and Thrombosis 13, 505– 511. Martin JF, Bath PM & Burr ML (1991) Influence of platelet size on outcome after myocardial infarction. Lancet 338, 1409–1411. Marui N, Offermann MK, Swerlick R, Kunsch C, Rosen CA, Ahmad M, Aleander RW & Medford RM (1993) Vascular adhesion molecule-1 (VCAM) gene transcription and expression are regulated through an antioxidant-sensitive mechanism in human vascular endothelial cells. Journal of Clinical Investigation 92, 1866–1874. Mata P, Ordovas JM, Lopez-Miranda J, Lichtenstein AH, Clevidence B, Judd JT & Schaefer EJ (1994) ApoA-IV phenotype affects diet-induced plasma LDL cholesterol lowering. Arteriosclerosis and Thrombosis 14, 884–891. Mayer EJ, Newman B, Quesenberry CP & Selby JV (1993) Usual dietary fat intake and insulin concentrations in healthy women twins. Diabetes Care 16, 1459–1469. Meade TW, Brozovic M, Chakrabarti RR, Hannes AP, Imeson JD, Mellows S, Miller GJ, North WRS, Stirling Y & Thompson SG (1986) Haemostatic function and ischaemic heart disease: principal results of the Northwick Park heart study. Lancet ii, 533–537. Meade TW, Ruddock V, Stirling Y, Chakrabarti R & Miller GI (1993) Fibrinolytic activity, clotting factors, and long-term incidence of ischaemic heart disease in the Northwick Park Heart Study. Lancet 342, 1076–1079. Mellies MJ, Jandacek RJ, Taulbee JD, Tewkesbury MB, Lamkin G, Baehler L, King P, Boggs D, Goldman S, Gouge A, Tsang R & Glueck CJ (1983) A double-blind, placebo controlled study of sucrose polyester in hypercholesterolemic outpatients. American Journal of Clinical Nutrition 37, 339–346. Mennen LI, Schouten EG, Grobbee DE & Kluft C (1996) Coagulation factor VII, dietary fat and blood lipids: a review. Thrombosis and Haemostasis 76, 492–499. Mensink RP & Katan MB (1992) Effect of dietary fatty acids on serum lipids and lipoproteins. A meta-analysis of 27 trials. Arteriosclerosis and Thrombosis 12, 911–919. Meydani SN, Lichtenstein AH, Cornwall S, Meydani M, Goldin BR, Rasmussen H, Dinarello CA & Schaefer EJ (1993) Immunologic effects of national cholesterol education panel step-2 diets with and without fish-derived n-3 fatty acid enrichment. Journal of Clinical Investigation 92, 105–113. Middleton E & Kandaswami C (1992) Commentary. Effects of flavonoids on immune and inflammatory cell functions. Biochimica et Biophysica Acta 43, 1167–1179. Miettinen TA (1991) Impact of apoE phenotype on the regulation of cholesterol metabolism. Annals of Medicine 23, 181–186. Miettinen TA, Naukkarinen V, Huttunen JK, Mattila S & Kumlin T (1982) Fatty-acid composition of serum lipids predicts myocardial infarction. British Medical Journal 285, 993–996. Miettinen TA, Puska P, Gylling H, Vanhanen H & Vartiainen E (1995) Reduction of serum cholesterol with sitostanol-ester margarine in a mildly hypercholesterolemic population. New England Journal of Medicine 333, 1308–1312. Miller GJ (1995) Lipoproteins and thrombosis: effects of lipid lowering. Current Opinion in Lipidology 6, 38–42. Miller GJ, Bauer KA, Barzegar S, Cooper JA & Rosenberg RD (1996) Increased activation of the haemostatic system in men at Cardiovascular system high risk of fatal coronary heart disease. Thrombosis and Haemostasis 75, 767–771. Mitropoulos KA (1994) Lipoprotein metabolism and thrombosis. Current Opinion in Lipidology 5, 227–235. Modan M, Halkin H, Almog S, Lusky A, Eshkol A, Shefi M, Shitrit A & Fuchs Z (1985) Hyperinsulinaemia: a link between hypertension, obesity and glucose intolerance. Journal of Clinical Investigation 75, 809–817. Mohan P, Sys SU & Brutsaert DL (1994) Nitric oxide donors induce a positive inotropic effect mediated by cGMP in isolated cardiac muscle without endothelium. European Heart Journal 15, Suppl., 144. Morris MC, Manson JE, Rosner B, Buring JE, Willett WC & Hennekens CH (1992) A prospective study of fish consumption on cardiovascular disease. Circulation 86, Suppl. 1, 463. Morris MC, Sacks F & Rosner B (1993) Does fish oil lower blood pressure? A meta analysis of controlled trials. Circulation 88, 525–533. Mutanen M (1997) Cis-unsaturated fatty acids and platelet function. Prostaglandins Leukotrienes and Essential Fatty Acids 57, 403–410. Mutanen M & Aro A (1997) Coagulation and fibrinolysis factors in healthy subjects consuming high stearic or trans fatty acids. Thrombosis and Haemostasis 77, 99–104. Murohara T, Scalia R & Lefer AM (1996) Lysophosphatidylcholine promotes P-selectin expression in platelets and endothelial cells. Possible involvement of protein kinase C activation and its inhibition by nitric oxide donors. Circulation Research 78, 780– 789. Nestel PJ, Noakes M, Belling GB, McArthur R & Clifton PM (1995) Effect on plasma lipids of interesterifying a mix of edible oils. American Journal of Clinical Nutrition 62, 950–955. Ng TKW, Hayes KC, DeWitt GF, Jegathesan M, Satgunasingam N, Ong AS & Tan D (1992) Dietary palmitic acid and oleic acids exert similar effects on serum cholesterol and lipoprotein profiles in normocholesterolemic men and women. Journal of the American College of Nutrition 11, 383–390. Nilsson TK, Sundell IB, Hellsten G & Hallmans G (1990) Reduced plasminogen activator inhibitor in high consumers of fruits, vegetables and root vegetables. Journal of Internal Medicine 227, 267–271. Nishinaga M, Ozawa T & Shimada K (1993) Homocysteine, a thrombogenic agent, suppresses anticoagulant heparan sulfate expression in cultured porcine aortic endothelial cells. Journal of Clinical Investigation 92, 1381–1386. Novell SE, Ahlbom A, Feychting M & Pedersen NL (1986) Fish consumption and mortality from coronary heart disease. British Medical Journal 293, 426. Ochi H, Kume N, Nishi E & Kita T (1995) Elevated levels of cAMP inhibit protein kinase C-independent mechanisms of endothelial platelet-derived growth factor-B chain and intercellular adhesion molecule-1 gene induction by lysophosphatidylcholine. Circulation Research 77, 530–535. Öhrvall M, Sundlöf G & Vessby B (1996) Gamma, but not alpha, tocopherol levels in serum are reduced in coronary heart disease patients. Journal of Internal Medicine 239, 111–117. O’Meara NM, Lewis GF, Cabana VG, Iverius PH, Getz GS & Polonsky KS (1992) Role of basal triglyceride and high density lipoprotein in determination of postprandial lipid and lipoprotein response. Journal of Clinical Endocrinology and Metabolism 75, 465–471. Oosthuizen W, Vorster HH, Jerling JC, Barnard HC, Smuts CM, Silvis N, Kruger A & Venter CS (1994) Both fish oil and olive oil lowered plasma fibrinogen in women with high baseline fibrinogen levels. Thrombosis and Haemostasis 72, 557–562. Ordovas JM, Lopez-Miranda J, Mata P, Perez-Jiminez F, Lichtenstein AH & Schaefer EJ (1995) Gene–diet interaction in S143 determining plasma lipid response to dietary intervention. Atherosclerosis 118, Suppl., S11–S27. Ostlund-Lindqvist AM, Albanus L & Croon LB (1985) Effect of dietary fatty acids on microsomal enzymes and membranes. Lipids 20, 620–624. Palinski W, Miller E & Witzum JL (1995) Immunization of low density lipoprotein (LDL) receptor-deficient rabbits with homologous malondialdehyde-modified LDL reduces atherogenesis. Proceedings of the National Academy of Sciences USA 92, 821– 825. Palinski W, Rosenfeld ME, Ylä-Herttuala S, Gurtner GC, Socher SS, Butler SW, Parthasarathy S, Carew TE, Steinberg D & Witztum JL (1989) Low density lipoprotein undergoes oxidative modification in vivo. Proceedings of the National Academy of Sciences USA 86, 1372–1376. Pan DA, Lillioja S, Milner MR, Kriketos AD, Baur LA, Bogardus C & Storlien LH (1995) Skeletal muscle membrane lipid composition is related to adiposity and insulin action. Journal of Clinical Investigation 96, 2802–2808. Pancharuniti N, Lewis CA, Sauberlich HE, Perkins LL, Go RC, Alvarez JO, Macaluso M, Acton RT, Copeland RB & Cousins AL (1994) Plasma homocyst(e)ine, folate, and vitamin B-12 concentrations and risk for early-onset coronary artery disease. American Journal of Clinical Nutrition 59, 940–948. Patsch JR (1987) Postprandial lipaemia. Baillière’s Clinical Endocrinology and Metabolism 1, 551–580. Pedersen A, Sandstrom B & Van Amelsvoort JMM (1997) The effect of ingestion of inulin on blood lipids and gastrointestinal symptoms in healthy females. British Journal of Nutrition 78, 215–222. Petri M, Roubenoff R, Dallal GE, Nadeau MR, Selhub J & Rosenberg IH (1996) Plasma homocysteine as a risk factor for atherothrombotic events in systemic lupus erythematosus. Lancet 348, 1120–1124. Pietinen P, Ascherio A, Korhonen P, Hartman AM, Willett WC, Albanes D & Virtamo J (1997) Intake of fatty acids and risk of coronary heart disease in a cohort of Finnish men. The alphatocopherol, beta-carotene cancer prevention study. American Journal of Epidemiology 145, 876–887. Poston RN, Haskard DO, Coucher JR, Gall NP & Johnson-Tidey RR (1992) Expression of intercellular adhesion molecule-1 in atherosclerotic plaques. American Journal of Pathology 140, 665–673. Potts JL, Coppack SW, Fisher RM, Humphreys SM, Gibbons GF & Frayn KN (1995) Impaired postprandial clearance of triacylglycerol-rich lipoproteins in adipose tissue in obese subjects. American Journal of Physiology 268, E588–E594. Pyörälä K (1979) Relationship of glucose tolerance and plasma insulin to the incidence of coronary heart disease: results from two population studies in Finland. Diabetes Care 2, 131–141. Qureshi AA, Bradlow BA, Brace L, Manganello J, Peterson DM, Pearce BC, Wright JJ, Gapor A & Elson CE (1995) Response of hypercholesterolemic subjects to administration of tocotrienols. Lipids 30, 1171–1177. Qureshi AA, Bradlow BA, Salser WA & Brace LD (1997) Novel tocotrienols of rice bran modulate cardiovascular disease risk parameters of hypercholesterolemic humans. Journal of Nutritional Biochemistry 8, 290–298. Qureshi AA, Pearce BC, Nor RM, Gapor A, Peterson DM & Elson CE (1996) Dietary alpha-tocopherol attenuates the impact of gamma-tocotrienol on hepatic 3 hydroxy-3-methylglutaryl coenzyme A reductase activity in chickens. Journal of Nutrition 126, 389–394. Rambjor GS, Walen AI, Windsor SL & Harris WS (1996) Eicosapentaenoic acid is primarily responsible for hypotriglyceridemic effect of fish oil in humans. Lipids 31, Suppl., S45– S49. S144 G. Hornstra et al. Rapaport SI & Rao LVM (1995) The tissue factor pathway: how it has become a ‘‘prima ballerina’’. Thrombosis and Haemostasis 74, 7–17. Rasmussen O, Thomsen C, Ingerslev J & Hermansen K (1994) Decrease in von Willebrand factor levels after a high-monounsaturated-fat diet in non-insulin-dependent diabetic subjects. Metabolism 43, 1406–1409. Reaven GM (1988) Role of insulin resistance in human disease. Diabetes 37, 1595–1607. Reaven GM, Litthel H & Landsberg L (1996) Hypertension and associated metabolic abnormalities – the role of insulin resistance and the sympatho-adrenal system. New England Journal of Medicine 334, 374–381. Richelsen B, Kristensen K & Pedersen SB (1996) Long-term (6 months) effect of a new fermented milk product on the level of plasma lipoproteins – a placebo controlled and double blind study. European Journal of Clinical Nutrition 50, 811–815. Ridker PM, Cushman M, Stampfer MJ, Tracy RP & Hennekens CH (1997) Inflammation, aspirin, and the risk of cardiovascular disease in apparently healthy men. New England Journal of Medicine 336, 973–979. Ridker PM, Hennekens CH, Cerskus A & Stampfer MJ (1994) Plasma concentration of cross-linked fibrin degradation product (D-dimer) and the risk of future myocardial infarction among apparently healthy men. Circulation 90, 2236–2240. Ridker PM & Vaughan DE (1995) Comment on NEJM. Hemostatic factors and the risk of myocardial infarction. New England Journal of Medicine 333, 389. Riemersma RA (1995) n-3 Fatty acids and experimental arrhythmogenesis. In n-3 Fatty Acids: Prevention and Treatment in Vascular Disease, pp. 98–106 [SD Kristensen, EB Schmidt, R De Caterina and S Endres, editors]. Verona: Bi & Gi Publishers, and London: Springer-Verlag. Riemersma RA, Sargent CA, Saman S, Rebergen SA & Abraham R (1988) Dietary fatty acids and ischaemic arrhythmias (letter). Lancet ii, 285–286. Rimm EB, Klatsky A, Grobbee D & Stampfer MJ (1996) Review of moderate alcohol consumption and reduced risk of coronary heart disease: is the effect due to beer, wine, or spirits. British Medical Journal 312, 731–736. Ripsin CM, Keenan JM, Jacobs DR Jr, Elmer JJ, Welch RR, van Horn L, Liu K, Turnbull WH, Thye FW, Kestin M, Hegsted M, Davidson DM, Davidson MH, Dugan LD, Demark-Wahnefried W & Beling S (1992) Oat products and lipid lowering. A metaanalysis. Journal of the American Medical Association 267, 3317–3325. Robinson K, Gupta A, Dennis V, Arheart K, Chaudhary D, Green R, Vigo P, Mayer EL, Selhub J & Kutner M (1996) Hyperhomocysteinemia confers an independent increased risk of atherosclerosis in end-stage renal disease and is closely linked to plasma folate and pyridoxine concentrations. Circulation 94, 2743–2748. Robinson K, Mayer EL, Miller DP, Green R, van Lente F, Gupta A, Kottke MK, Savon SR, Selhub J & Nissen SE (1995) Hyperhomocysteinemia and low pyridoxal phosphate. Common and independent reversible risk factors for coronary artery disease. Circulation 92, 2825–2830. Rocchini AP, Moorehead C, DeRemer S, Goodfriend TL & Ball DL (1990) Hyperinsulinemia and the aldosterone and pressor responses to angiotensin II. Hypertension 15, 861–866. Rodgers GM & Conn MT (1990) Homocysteine, an atherogenic stimulus, reduces protein C activation by arterial and venous endothelial cells. Blood 74, 895–901. Roselaar SE, Schonfeld G & Daugherty A (1995) Enhanced development of atherosclerosis in cholesterol fed rabbits by suppression of cell-mediated immunity. Journal of Clinical Investigation 96, 1389–1394. Ross R (1995) Cell biology of atherosclerosis. Annual Reviews of Physiology 57, 791–804. Roust RR & Jensen MD (1993) Postprandial free fatty acid kinetics are abnormal in upper body obesity. Diabetes 42, 1567–1573. Rydholm H, Boström S, Eriksson E & Risberg B (1995) Complex intracellular signal transduction regulates tissue plasminogen activator (t-PA) and plasminogen activator inhibitor type-1 (PAI-1) synthesis in cultured human umbilical vein endothelium. Scandinavian Journal of Laboratory and Clinical Investigation 55, 323–330. Saad MF, Lillioja S, Nyomba BL, Castillo C, Ferraro R, De Gregorio M, Ravussin E, Knowler WC, Bennett PH, Howard BV & Bogardus C (1991) Racial differences in the relation between blood pressure and insulin resistance. New England Journal of Medicine 324, 733–739. Salo MK, Vartiainen E, Puska P & Nikkari T (1985) Platelet aggregation in Finnish men and its relation to fatty acids in platelets, plasma and adipose tissue. Thrombosis and Haemostasis 54, 563–569. Salomaa V, Stinson V, Kark JD, Folsom AR, Davis CE & Wu KK (1995) Association of fibrinolytic parameters with early atherosclerosis. The Aric Study. Circulation 91, 284–290. Salonen JT, Seppänen K, Nyyssönen K, Korpela H, Kauhanen J, Kantola M, Tuomilehto J, Esterbauer H, Tatzber F & Salonen R (1995) Intake of mercury from fish, lipid peroxidation, and the risk of myocardial infarction and coronary, cardiovascular, and any death in eastern Finnish men. Circulation 91, 645–655. Sanderson P, Yaqoob P & Calder PC (1995) Effects of dietary lipid manipulation upon rat spleen lymphocyte functions and the expression of lymphocyte surface molecules. Journal of Nutrition and Environmental Medicine 5, 119–132. Santoli D, Phillips PD, Colt TL & Zurier RB (1990) Suppression of interleukin-2-dependent human T cell growth in vitro by prostaglandin E (PGE) and their precursor fatty acids. Journal of Clinical Investigation 85, 424–432. Sargent CA (1990) Dietary fat and ischaemic arrhythmias. PhD Thesis, University of Edinburgh. Sargent C & Riemersma RA (1990) Polyunsaturated fatty acids and cardiac arrhythmias. Biochemical Society Transactions 18, 1077–1078. Schienle HW, Seitz R, Rohner I, Lerch L, Krumpholz B, Krauss G, Fowler B, Baumgartner R, Willenbockel U & Egbring R (1994) Coagulation factors and markers of activation of coagulation in homocystinuria (HOCY): a study in two siblings. Blood Coagulation and Fibrinolysis 5, 873–878. Schmitz C, Lindpaintner K, Verhoef P, Gaziano JM & Buring J (1996) Genetic polymorphism of methylenetetrahydrofolate reductase and myocardial infarction. A case–control study. Circulation 94, 1812–1814. Schreck R, Albermann K & Baeuerle PA (1992) Nuclear factor kappa B: an oxidative stress-responsive transcription factor of eukaryotic cells. Free Radical Research Communications 17, 221–237. Selhub J, Jacques PF, Bostom AG, D’Agostino RB, Wilson PW, Belanger AJ, O’Leary DH, Wolf PA, Schaefer EJ & Rosenberg IH (1995) Association between plasma homocysteine concentrations and extracranial carotid-artery stenosis [see comments]. New England Journal of Medicine 332, 286–291. Selhub J, Jacques PF, Bostom AG, D’Agostino RB, Wilson PW, Belanger AJ, O’Leary DH, Wolf PA, Rush D & Schaefer EJ (1996) Relationship between plasma homocysteine, vitamin status and extracranial carotid-artery stenosis in the Framingham Study population. Journal of Nutrition 126, 1258S–1265S. Semafuko WEB, Rutledge CO & Dixon WR (1984) Alterations in myocardial alpha-adrenoceptor function by modification of dietary lipids. Federation Proceedings 43, 414. Shahar E, Folsom AR, Wu KK, Dennis BH, Shimakawa T, Cardiovascular system Conlan MG, Davis CE & Williams OD (1993) Association of fish intake and dietary n-3 polyunsaturated fatty acids with a hypocoagulable profile. The Atherosclerosis Risk in Communities (ARIC) Study. Arteriosclerosis and Thrombosis 13, 1205–1212. Shaten BJ, Kuller LH & Neaton JD (1991) Association between baseline risk factors, cigarette smoking, and CHD mortality after 10.5 years. MRFIT Research Group. Preventive Medicine 20, 655–659. Shekelle RB, Missell LV, Paul O, Shyrock AM & Stamler J (1985) Fish consumption and mortality from coronary heart disease. New England Journal of Medicine 313, 820. Silagy C & Neil A (1994) Garlic as a lipid lowering agent – a metaanalysis. Journal of the Royal College of Physicians, London 28, 39–45. Silveira A, Karpe F, Blombäck M, Steiner G, Walldius G & Hamsten A (1994) Activation of coagulation factor VII during alimentary lipemia. Arteriosclerosis and Thrombosis 14, 60–69. Simionescu N, Vasile E, Lupu, Popescu G & Simionescu M (1986) Prelesional events in atherogenesis. Accumulation of extracellular cholesterol-rich liposomes in the arterial intima and cardiac valves of the hyperlipidemic rabbit. American Journal of Pathology 123, 109–125. Simon JA, Fong J, Bernert JT & Browner WS (1995) Serum fatty acids and the risk of stroke. Stroke 26, 778–782. Simons LA, Balasubramaniam S, von Konigsmark M, Parfitt A, Simons J & Peters W (1995) On the effect of garlic on plasma lipids and lipoproteins in mild hypercholesterolaemia. Atherosclerosis 113, 219–225. Simonsen T, Vårtun A, Lyngmo V & Nordøy A (1987) Coronary heart disease, serum lipids and dietary fish in two communities in Northern Norway. Acta Medica Scandinavica 222, 237–245. Simopoulos AP (1994) Is insulin resistance influenced by dietary linoleic acid and trans fatty acids? Free Radical Biology and Medicine 17, 367–372. Skarfors ET, Lithell HO & Selinns I (1991) Risk factors for the development of hypertension: a 10-year longitudinal study in middle-aged men. Journal of Hypertension 9, 217–223. Smith P, Arnesen H & Holme I (1990) The effect of warfarin on mortality and reinfarction after myocardial infarction. New England Journal of Medicine 323, 147–152. Sniderman A, Brown BG, Stewart BF & Cianflone K (1992) From familial combined hyperlipidaemia to hyperapo B: unravelling the overproduction of hepatic apolipoprotein B. Current Opinion in Lipidology 3, 137–142. Spady DK, Woollett LA & Dietschy JM (1993) Regulation of plasma LDL cholesterol levels by dietary cholesterol and fatty acids. Annual Review of Nutrition 13, 355–381. Sperling RI, Benincaso AI, Knoell CT, Larkin JK, Austen KF & Robinson DR (1993) Dietary q-3 polyunsaturated fatty acids inhibit phosphoinositide formation and chemotaxis in neutrophils. Journal of Clinical Investigation 91, 651–660. Springer TA (1990) Adhesion receptors of the immune system. Nature 346, 425–434. Stabler SP, Allen RH, Savage DG & Lindenbaum J (1990) Clinical spectrum and diagnosis of cobalamin deficiency [see comments]. Blood 76, 871–881. Stamler JS & Loscalzo J (1992) Endothelium-derived relaxing factor modulates the atherothrombogenic effects of homocysteine. Journal of Cardiovascular Pharmacology 20, Suppl. 12, S202–S204. Stamler JS, Osborne JA, Jaraki O, Rabbani LE, Mullins M, Singel D & Loscalzo J (1993) Adverse vascular effects of homocysteine are modulated by endothelium-derived relaxing factor and related oxides of nitrogen. Journal of Clinical Investigation 91, 308–318. Stamler JS & Slivka A (1996) Biological chemistry of thiols in the S145 vasculature and in vascular-related diseases. Nutrition Reviews 54, 1–30. Stasse-Wolthuis M, Albers HF, van Jeveren JG, Wil-de Jong J, Hautvast JGAJ, Hermus RJ, Katan MB, Brydon WG & Eastwood MA (1980) Influence of dietary fiber from vegetables and fruits, bran or citrus pectin on serum lipids, fecal lipids, and colonic function. American Journal of Clinical Nutrition 33, 1745–1756. Steering Committee of the Physicians’ Health Study Research Group (1989) Final report on the aspirin component of the ongoing Physicians’ Health Study. New England Journal of Medicine 321, 129–135. Steinberg D, Parthasarathy S, Carew TE, Khoo JC & Witzum JL (1989) Beyond cholesterol: modifications of low density lipoprotein that increase its atherogenicity. New England Journal of Medicine 32, 915–923. Steiner M, Glantz M & Lekos A (1995) Vitamin E plus aspirin compared with aspirin alone in patients with transient ischemic attacks. American Journal of Clinical Nutrition 62, Suppl., 1381S–1384S. Stemme S & Hansson GK (1994) Immune mechanisms in atherosclerosis. Coronary Artery Disease 5, 216–222. Stephens NG, Parsons A, Schofield PM, Kelly F, Cheeseman K, Mitchinson MJ & Brown MJ (1996) Randomised controlled trial of vitamin E in patients with coronary disease: Cambridge Heart Antioxidant Study (CHAOS). Lancet 347, 781–786. Stinson JC, Owens D, McBrinn S, Collins P, Johnson A & Tomkin GH (1993) The regulation of postprandial cellular cholesterol metabolism in type 2 diabetic and non-diabetic subjects. Diabetic Medicine 10, 420–426. Storlien LH, Jenkins AB, Chisholm DJ, Pascal WS, Khouri S & Kraeger EW (1991) Influence of dietary fat composition on development of insulin resistance in rats: relationship to muscle triglyceride and omega 3 fatty acids in muscle phospholipids. Diabetes 40, 280–289. Stout RW (1990) Insulin and atheroma. 20-Yr perspective. Diabetes Care 13, 631–654. Stout RW & Vallance-Owen J (1969) Insulin and atheroma. Lancet i, 1078–1080. Sundell IB & Rånby M (1993) Oat husk fiber decreases plasminogen activator inhibitor type 1 activity. Haemostasis 23, 45–50. Tegelman R, Aberg T, Eklöf R, Pousette A, Carlstrom K & Berglund I (1996) Influence of a diet regimen on glucose homeostasis and serum lipid levels in male elite athletes. Metabolism 4, 435–441. Temme EHM, Mensink RP & Hornstra G (1996) Comparison of the effect of diets enriched in lauric, palmitic, or oleic acids on serum lipids and lipoproteins in healthy women and men. American Journal of Clinical Nutrition 63, 897–903. Thaulow E, Erikssen J, Sandvik L, Stormorken H & Cohn PF (1991) Blood platelet count and function are related to total and cardiovascular death in apparently healthy men. Circulation 84, 613–617. Thompson SG, Kienast J, Pyke SDM, Haverkate F & van de Loo JC (1995) Hemostatic factors and risk of myocardial infarction or sudden death in patients with angina pectoris. New England Journal of Medicine 332, 635–641. Tonstad S, Refsum H, Sivertsen M, Christophersen B, Ose L & Ueland PM (1996) Relation of total homocysteine and lipid levels in children to premature cardiovascular death in male relatives. Pediatric Research 40, 47–52. Trials of Hypertension Collaborative Research Group (1992) The effects of nonpharmacologic interventions on blood pressure of persons with high normal levels: results of the Trials of Hypertension Prevention (Phase I). Journal of the American Medical Association 267, 1213–1220. Tsai JC, Perrella MA, Yoshizumi M, Hsieh CM, Haber E, Schlegel R & Lee M (1994) Promotion of vascular muscle S146 G. Hornstra et al. cell growth by homocysteine: a link to atherosclerosis. Proceedings of the National Academy of Sciences USA 91, 6369–6373. Tsai JC, Wang H, Perrella MA, Yoshizumi M, Sibinga NE, Tan LC, Haber E, Chang TH, Schlegel R & Lee ME (1996) Induction of cyclin A gene expression by homocysteine in vascular smooth muscle cells. Journal of Clinical Investigation 97, 146–153. Tso P, Pinkston G, Klurfeld DM & Kritchevsky D (1984) The absorption and transport of dietary cholesterol in the presence of peanut oil or randomized peanut oil. Lipids 19, 11–16. Tucker KL, Mahnken B, Wilson PW, Jacques P & Selhub J (1996) Folic acid fortification of the food supply. Potential benefits and risks for the elderly population. Journal of the American Medical Association 276, 1879–1885. Turpeinen AM, Sauer R, Freese R, Pajari A-M & Mutanen M (1997) Platelets are similarly activated by both oleic acid and linoleic acid rich diets in healthy humans. Prostaglandins, Leukotrienes and Essential Fatty Acids 57, 228. Turpeinen AM, Wubert J, Aro A, Lorenz R & Mutanen M (1998) Similar effects of diets rich in stearic acid or trans fatty acids on platelet function and endothelial prostacyclin production in humans. Arteriosclerosis, Thrombosis and Vascular Biology 18, 316–322. Ubbink JB, Vermaak WJ, van der Merwe A & Becker PJ (1993) Vitamin B-12, vitamin B-6, and folate nutritional status in men with hyperhomocysteinemia. American Journal of Clinical Nutrition 57, 47–53. Ueda S, Elliott HL, Morton JJ & Connell JMC (1995) Enhanced pressor response to angiotensin I in normotensive men with a deletion genotype (DD) for angiotensin-converting enzyme. Hypertension 25, 1266–1269. Uhleman ER, Ten Pas JH, Lucky AW, Schulman JD, Mudd SH & Shulamn NR (1976) Platelet survival and morphology in homocysteinuria due to cystathionine synthase deficiency. New England Journal of Medicine 295, 1283–1286. Van den Berg M, Boers GH, Franken DG, Blom HJ, Van Kamp GJ, Jakobs C, Rauwerda JA, Kluft C & Stehouwert CD (1995) Hyper-homocysteinaemia and endothelial dysfunction in young patients with peripheral arterial occlusive disease. European Journal of Clinical Investigation 25, 176–181. van de Vijver LPL, Kardinaal AFM, Grobbee DE, Princen HMG & van Poppel G (1997) Lipoprotein oxidation, antioxidants and cardiovascular risk. Epidemiologic evidence. Prostaglandins, Leukotrienes and Essential Fatty Acids 57, 479–487. Vanhanen HT, Blomqvist S, Ehnholm C, Hyvonen M, Jauhiainen M, Torstila I & Miettinen TA (1993) Serum cholesterol, cholesterol precursors, and plant sterols in hypercholesterolemic subjects with different apoE phenotypes during dietary sitostanolester treatment. Journal of Lipid Research 34, 1535–1544. Verhoef P, Stampfer MJ, Buring JE, Gaziano JM, Allen RH, Stabler SP, Reynolds RD, Kok FJ, Hennekens CH & Willett WC (1996) Homocysteine metabolism and risk of myocardial infarction: relation with vitamins B6, B12, and folate. American Journal of Epidemiology 143, 845–859. Vollset SE, Heuch I & Bjelke E (1985) Fish consumption and mortality from coronary heart disease. New England Journal of Medicine 313, 820–821. von Eckardstein A, Malinow MR, Upson B, Heinrich J, Schulte H, Schonfeld R, Kohler E & Assmann G (1994) Effects of age, lipoproteins and hemostatic parameters on the role of homocyst(e)ine as a cardiovascular risk factor in men. Arteriosclerosis and Thrombosis 14, 460–464. Wahlqvist ML, Krivokuca-Bogetic Z, Lo CS, Hage B, Smith R & Lukito W (1992) Differential serum responses of tocopherols and tocotrienols during vitamin supplementation in hypercholesterolemic individuals without change in coronary risk factors. Nutrition Research 12, Suppl. 1, S181–S201. Wang J, Dudman NP & Wilcken DE (1993) Effects of homocysteine and related compounds on prostacyclin production by cultured human vascular endothelial cells. Thrombosis and Haemostasis 70, 1047–1052. Warshafsky S, Kamer RS & Sivak SL (1993) Effect of garlic on total serum cholesterol. A meta-analysis. Annals of Internal Medicine 119, 599–605. Weber C, Erl W, Pietsch A, Danesch U & Weber PC (1995) Docosahexaenoic acid selectively attenuates induction of vascular cell adhesion molecule-1 and subsequent monocytic cell adhesion to human endothelial cell stimulated by tumor necrosis factor-a. Arteriosclerosis, Thrombosis and Vascular Biology 15, 622–628. Weidtmann A, Scheithe R, Hrboticky N, Pietsch A, Lorenz R & Siess W (1995) Mildly oxidized LDL induces platelet aggregation through activation of phospholipase A2. Arteriosclerosis, Thrombosis and Vascular Biology 15, 1131–1138. Weinberger MH (1990) Clinical studies of the role of dietary sodium in blood pressure. In Hypertension: Pathophysiology, Diagnosis, and Management, pp. 1999–2010 [JH Laragh and BM Brenner, editors]. New York, NY: Raven Press. Welborn TG, Breckenridge A, Rubinstein AH, Dollery CT & Fraser TR (1966) Serum-insulin in essential hypertension and peripheral vascular disease. Lancet 1, 1336–1337. Westerband A, Mills JL, Marek JM, Heimark RL, Hunter GC & Williams SK (1997) Immunocytochemical determination of cell type and proliferation rate in human vein graft stenoses. Journal of Vascular Surgery 25, 64–73. Weylandt KH, Kang JX & Leaf A (1996) Polyunsaturated fatty acids exert antiarrhythmic actions as free acids rather than in phospholipids. Lipids 31, 977–982. Widhalm K, Stargel WW, Burns TS & Tschanz C (1994) Evaluation of clinical and biochemical parameters in children after consumption of microparticulated protein fat substitute (Simplesse). Journal of the American College of Nutrition 13, 392–396. Williams RR, Hasstedt ST, Hunt SC, Wu LL, Hopkins PN, Berry TD, Stults BM, Barlow GK & Kuida H (1991) Genetic traits related to hypertension and electrolyte metabolism. Hypertension 17, Suppl. 1, I-69–I-73. Wince LC, Hugman LE & Brenner GM (1984) Alteration of inotropic response in rat atria by dietary fat. Federation Proceedings 43, 2677. World Hypertension League (1991) Alcohol and hypertension. Journal of Hypertension 5, 227–232. Xu Q, Dietrich H, Steiner HJ, Gown AM, Mikuz G, Kaufmann SHE & Wick G (1992) Induction of atherosclerosis in normocholesterolemic rabbits by immunization with heat shock protein 65. Arteriosclerosis and Thrombosis 12, 789–799. Yaqoob P & Calder PC (1993) The effects of fatty acids on lymphocyte functions. International Journal of Biochemistry 25, 1705–1714. Zhao B (1996) Role of lipoproteins in platelet activation. Blood Coagulation and Fibrinolysis 2, 270–273. Zilversmit DB (1979) Atherogenesis: a postprandial phenomenon. Circulation 60, 473–485. Zock PL, de Vries JH, de Fouw NJ & Katan MB (1995) Positional distribution of fatty acids in dietary triglycerides: effects on fasting blood lipoprotein concentrations in humans. American Journal of Clinical Nutrition 61, 48–55. Zock PL, de Vries JHM & Katan MB (1994) Impact of myristic acid versus palmitic acid on serum lipid and lipoprotein levels in healthy women and men. Arteriosclerosis and Thrombosis 14, 567–575. # Nutrition Society 1998 British Journal of Nutrition (1998), 80, Suppl. 1, S147–S171 S147 Functional food science and gastrointestinal physiology and function S. Salminen1 , C. Bouley2 , M.-C. Boutron-Ruault3 , J. H. Cummings4¬ , A. Franck5 , G. R. Gibson6 , E. Isolauri7 , M.-C. Moreau8 , M. Roberfroid9 and I. Rowland10 1 Department of Biochemistry and Food Chemistry, University of Turku, SF-20500 Turku, Finland 2 Groupe Danone, 15, Av. Galilée, F-92350 Le Plessis-Robinson, France 3 U290 INSERM, Hôpital St Lazare, 107, rue du Faubourg Saint-Denis, F-75010 Paris, France 4 Dunn Clinical Nutrition Centre, Hills Road, Cambridge CB2 2DH, UK 5 Raffinerie Tirlemontoise – ORAFTI, Aandorenstraat 1, B-3300 Tienen, Belgium 6 Institute of Food Research, Reading Laboratory, Earley Gate, Reading RG6 6BZ, UK 7 University of Tampere Medical School, PO Box 607, SF-33101 Tampere, Finland 8 INRA – Unité d’Ecologie et de Physiologie du Système Digestif, Bâtiment 440 R-2, Domaine de Vilvert, F-78352 Jouy-en-Josas Cedex, France 9 UCL, Ecole de Pharmacie, Tour Van Helmont, Avenue E. Mounier, 73, B-1200 Brussels, Belgium 10 University of Ulster, Coleraine BT52 1SA, UK Contents 1. Introduction 2. Intestinal microflora: physiology and functions 2.1. The normal flora 2.2. Fermentation and short-chain fatty acids 2.2.1. Physiology and health 2.2.2. Acetate 2.2.3. Propionate 2.2.4. Butyrate 2.3. Interactions between the intestinal microflora and epithelial cells 2.4. The concept of healthy microflora 3. The gastrointestinal immune system 3.1. Gut-associated lymphoid tissue (GALT) 3.2. The structure of GALT and cell distribution 3.3. Immunophysiological regulation 3.4. Regulation of antigen transfer 3.5. Interactions between the intestinal microflora and the GALT 4. Mucosal cell proliferation and differentiation 4.1. Cell proliferation 4.2. Differentiation 4.3. Apoptosis 4.4. Mucosal enzymes 5. Gastrointestinal function and disease 5.1. Gastrointestinal infections 5.2. Normal bowel habit 5.3. Constipation 5.4. Irritable bowel syndrome (IBS) 5.5. Inflammatory bowel disease 5.5.1. Crohn’s disease 5.5.2. Ulcerative colitis 5.6. Food allergy S148 S148 S148 S149 S149 S149 S149 S149 S151 S151 S152 S152 S152 S152 S153 S153 S154 S154 S154 S154 S154 S154 S154 S155 S155 S155 S155 S155 S156 S156 5.7. Colorectal cancer 6. Methodology 6.1. Human intestinal microflora 6.2. Functional analysis of the gut microflora 6.2.1. Bacterial enzymes 6.2.2. Bacterial metabolites in faeces 6.2.3. Assessment of cytotoxicity, genotoxicity and mutagenicity of faeces 6.2.4. Susceptibility of functional markers to dietary change 6.3. Digestibility and bioavailability of foods 6.4. Large-bowel function 6.5. Gut-associated lymphoid tissue 6.6. Epithelial cell proliferation and colon carcinogenesis 6.6.1. Biological markers for colorectal carcinogenesis 6.6.2. Cell proliferation 6.6.3. Differentiation 6.6.4. Apoptosis 6.6.5. Products used in experimental carcinogenesis 6.6.6. Types of lesion 6.6.7. Transgenic mouse models for colon cancer studies 6.6.8. Limits of experimental models 7. Human studies on the effects of food and food components 7.1. Prebiotics 7.2. Probiotics 7.2.1. Alleviation of lactose intolerance symptoms 7.2.2. Immune enhancement S156 S156 S156 S157 S158 S158 S158 S158 S158 S158 S159 S160 S160 S160 S161 S161 S161 S161 S161 S162 S162 S162 S163 S163 S163 Abbreviations: GALT, gut-associated lymphoid tissue; IBS, irritable bowel syndrome; Ig, immunoglobulin; ILSI, International Life Sciences Institute; IQ, 2amino-3-methyl-7H-imidazo[4,5-f ]quinoline; MTT, 3-[4,5-dimethyl-thiazol-2-yl]-2,5-diphenyltetrazolium bromide; 7-OHIQ, 7-hydroxy-2-amino-3,6dihydro-3-methyl-7H-imidazo[4,5-f ]quinoline-7-one; rRNA, ribosomal RNA; SCFA, short-chain fatty acids. *Corresponding author: Dr J. H. Cummings, fax þ44 (0)1223 413763, email [email protected] S148 S. Salminen et al. 7.2.3. Acute gastroenteritis 7.2.4. Faecal mutagenicity and enzymes 7.3. Diet and colon cancer 7.3.1. Dietary protective factors 8. Safety issues 8.1. Prebiotics 8.2. Probiotics 9. Critical evaluation of present knowledge 9.1. Intestinal microflora 9.2. Mucosal function 9.3. Gastrointestinal physiology 9.4. Methodology S163 S163 S163 S163 S164 S164 S164 S164 S164 S164 S165 S165 9.5. Human studies on health benefits 9.5.1. Prebiotics 9.5.2. Probiotics 9.5.3. Diet and colon cancer 9.6. Safety 10. Recommendations for future research priorities 10.1. Intestinal microflora 10.2. Short-chain fatty acids and intestinal microflora 10.3. Diet and cancer 10.4. Immune system 10.5. Gut mucosa S165 S165 S165 S165 S165 S165 S166 S166 S166 S166 S166 Abstract The gut is an obvious target for the development of functional foods, acting as it does as the interface between diet and the metabolic events which sustain life. The key processes in digestive physiology which can be regulated by modifying diet are satiety, the rate and extent of macronutrient breakdown and absorption from the small bowel, sterol metabolism, the colonic microflora, fermentation, mucosal function and bowel habit, and the gut immune system. The intestinal microflora is the main focus of many current functional foods. Probiotics are foods which contain live bacteria which are beneficial to health whilst prebiotics, such as certain nondigestible oligosaccharides which selectively stimulate the growth of bifidobacteria in the colon, are already on the market. Their claimed benefits are to alleviate lactose maldigestion, increase resistance to invasion by pathogenic species of bacteria in the gut, stimulate the immune system and possibly protect against cancer. There are very few reports of well-designed human intervention studies with prebiotics as yet. Certain probiotic species have been shown to shorten the duration of rotavirus diarrhoea in children but much more work is needed on the mechanism of immunomodulation and of competitive exclusion and microflora modification. The development of functional foods for the gut is in its infancy and will be successful only if more fundamental research is done on digestive physiology, the gut microflora, immune system and mucosal function. Gastrointestinal function: Microflora: Immune system 1. Introduction One of the most promising areas for the development of functional foods lies in modification of the activity of the gastrointestinal tract by use of probiotics, prebiotics and synbiotics. To understand the potential value of these functional foods and to be able to develop new approaches it is necessary to study the normal human intestinal flora, fermentation, the gut immune system, mucosal function and the principal gut-related diseases. 2. Intestinal microflora: physiology and functions 2.1. The normal flora (Gibson & Macfarlane, 1995) Bacterial numbers and composition vary considerably along the human gastrointestinal tract. The total bacterial count in gastric contents is usually below 103 /g, with numbers being kept low due to the acid lumen pH. In the small intestine, numbers range from approximately 104 /ml contents to about 106 –107 /ml at the ileocaecal region. The main factors limiting growth in the small bowel are the rapid transit of contents and secretion of bile and pancreatic juice. The human large intestine is an intensely populated microbial ecosystem. Several hundred species of bacteria are usually present, with typical numbers of about 1011 –1012 /g. The majority of these bacteria are strict anaerobes. Table 1 lists bacteria commonly isolated from the human colon. Bacterial counts of individual species range over several orders of magnitude, and the nutrition and metabolic products of different bacterial groups vary considerably. Most bacteria growing in the colon are non-sporing anaerobes and include members of the genera Bacteroides, Bifidobacterium and Eubacterium among many others. Clostridia are also represented, although they are outnumbered by the non-sporing anaerobes, as are facultative anaerobes such as streptococci and enterobacteria. Quantitatively, the most important genera of intestinal bacteria in animals and man are the bacteroides and bifidobacteria, which can account for 30 % and 25 % of the total anaerobic counts respectively. The Gram-negative Bacteroides group (e.g. Bacteroides ovatus, Bacteroides fragilis, Bacteroides thetaiotaomicron) are thought to be numerically predominant. The genus contains both proteolytic and saccharolytic species. Amongst the Gram-positive, non-sporing rods, several genera are numerically significant. Obligate anaerobes include eubacteria and bifidobacteria, such as Bifidobacterium bifidum and Bifidobacterium infantis, which are prominent in the faeces of breast-fed infants. The genus Lactobacillus contains many species that occur in the gut of most warm-blooded animals. Although numerically important in the alimentary tract, their ecological significance has not been conclusively determined. Gastrointestinal physiology and function S149 Table 1. Bacteria, their substrates and products in the human large intestine (From Macfarlane et al. 1995) Concentration (log10 /g dry wt faeces) Bacteria Description Mean Range Substrate Saccharolytic Saccharolytic, some amino acid fermenting species Saccharolytic Saccharolytic and amino acid fermenting species Saccharolytic Saccharolytic As for the clostridia Amino acid fermenters Chemolithotrophic Various Saccharolytic, lactate fermenting Saccharolytic Carbohydrate and amino acid fermenting Amino acid fermentation, carbohydrate also assimilated As for streptococci Bacteroides Eubacteria G¹ rods Gþ rods 11·3 10·7 9·2–13·5 5·0–13·3 Bifidobacteria Clostridia Gþ rods Gþ rods 10·2 9·8 4·9–13·4 3·3–13·1 Lactobacilli Ruminococci Peptostreptococci Peptococci Methanobrevibacter Desulfovibrios Propionibacteria Actinomyces Streptococci Fusobacteria Gþ Gþ Gþ Gþ Gþ G¹ Gþ Gþ Gþ G¹ 9·6 10·2 10·1 10·0 8·8 8·4 9·4 9·2 8·9 8·4 3·6–12·5 4·6–12·8 3·8–12·6 5·1–12·9 7·0–10·5 5·2–10·9 4·3–12·0 5·7–11·1 3·9–12·9 5·1–11·0 Escherichia G¹ rods 8·6 3·9–12·3 rods cocci cocci cocci cocco bacilli rods rods rods cocci rods Fermentation products A, P, S A, B, L A, L, f, e A, P, B, L, e L A A, L A, B, L CH4 A A, P A, L, S L, A B, A, L Mixed acids Gþ, Gram-positive; G¹, Gram-negative; A, acetate; P, propionate; B, butyrate; L, lactate; S, succinate; f, formate; e, ethanol. Several types of spore-forming rods and cocci are also inhabitants of the gut. The genus Clostridium is probably the most common: C. perfringens, C. bifermentans and C. tetani are regularly isolated, albeit in relatively low numbers, and are of significance in human and veterinary medicine. Facultative and obligately anaerobic Gram-positive cocci are also numerically important. The strict anaerobes include Peptostreptococcus, Ruminoccus, Megasphaera elsdenii and Sarcina ventriculi. The facultatively anaerobic streptococci are well represented by many species from Lancefield group D, including S. faecalis, S. bovis and S. equinus, and some from group K, such as S. salivarius, which is usually associated with the mouth. Gram-negative anaerobic cocci include Veillonella and Acidaminococcus. Although not numerous, the Gram-negative facultative anaerobic rods include a number of important pathogens. For example, members of the Enterobacteriaceae, particularly Escherichia coli, are usually thought of as characteristic intestinal bacteria. The large-gut microflora is acquired at birth. Initially, facultatively anaerobic strains dominate. Thereafter, differences exist in the species composition that develops and this is largely controlled by the type of diet. The faecal flora of breast-fed infants is dominated by bifidobacteria. In contrast, formula-fed infants have a more complex microbiota with bifidobacteria, bacteroides, clostridia and streptococci all being prevalent. After weaning, a pattern that resembles the adult flora becomes established (Ducluzeau, 1993). The principal role of the intestinal microflora is to salvage energy from carbohydrates not digested in the upper gut, through fermentation. The major substrates for fermentation are dietary carbohydrates that have escaped digestion in the upper gastrointestinal tract. These include starch that enters the colon (resistant starch), as well as NSP, e.g. cellulose, hemicelluloses, pectins and gums. Other carbohydrate sources available for fermentation are non-digestible oligosaccharides, various sugars and sugar alcohols (Cummings et al. 1997). In addition, proteins and amino acids can be effective as growth substrates for colonic bacteria. These include elastin, collagen and albumin, as well as bacterial protein released following cell lysis. Pancreatic enzymes represent a source of N. Bacterial secretions, lysis products, sloughed epithelial cells and mucins may also make a contribution as fermentation substrates. Total substrate availability in the human adult colon is 20–60 g carbohydrate and 5–20 g protein/d (Cummings & Englyst, 1987; Cummings et al. 1989). Significant regional differences occur in bacterial activity in the colon. The right (proximal) colon is characterized by a high substrate availability (due to dietary input), low pH (from acids produced in fermentation) and rapid transit. The left, or distal, colon has a lower concentration of available substrate, the pH is approximately neutral and bacteria grow more slowly. The proximal region tends to be a more saccharolytic environment than the distal gut, the latter having higher bacterial proteolysis. In addition to its role in fermentation the large-intestinal microflora contributes towards health in a number of other ways. The development of the intestinal microflora provides the basis for a barrier that prevents pathogenic bacteria from invading the gastrointestinal tract. The composition of the intestinal microflora together with the gut immune system allows resident bacteria to exert a protective function. In addition gut bacteria are involved in vitamin synthesis (especially vitamins B and K) and in the metabolism of xenobiotics. Thus, modification of the flora by dietary means offers one of the most effective opportunities for development of functional foods. 2.2. Fermentation and short-chain fatty acids (Binder et al. 1994; Cummings, 1995; Cummings et al. 1995) Through fermentation, bacterial growth is stimulated (biomass), and short-chain fatty acids (SCFA) and the gases H2 , CO2 and CH4 are produced. S150 S. Salminen et al. SCFA are the major end-products of bacterial fermentative reactions in the colon and are the principal anions in the hindgut of man and all other mammals. The SCFA are acetate, propionate and butyrate but other significant endproducts of carbohydrate fermentation include lactate, ethanol, succinate, formate, valerate and caproate (Table 1). Branched-chain fatty acids such as isobutyrate, 2-methylbutyrate and isovalerate may be formed from the fermentation of amino acids that originate in proteolysis. The other end-products from bacterial metabolism of proteins include NH3 , phenols, indoles and amines, some of which have toxic properties (Macfarlane & Macfarlane, 1995). The amount of SCFA, which is usually in excess of 100 mmol/kg contents, and the molar ratios of the three principal acids produced by fermentation, vary substantially, depending on the substrate. This has been studied extensively in vitro using single-chamber chemostat models of the gut inoculated with intestinal micro-organisms. Yields vary from 40–60 % (g SCFA/100 g substrate utilized), with molar ratios of acetate from 60–80, propionate 14–22 and butyrate 8–23 (Cummings, 1995). Whilst acetate is produced in all fermentation systems in vitro, it is the major product of pectin breakdown. Similarly, the highest molar ratios of propionate are seen characteristically with arabinogalactan and guar gum as substrate. Amounts of butyrate vary perhaps more than any other according to substrate but the polysaccharide that is associated with the highest relative amounts is starch. In animal studies, wheat bran seems to give rise to high concentrations of SCFA in the gut, despite the fact that it is relatively poorly fermented, especially in human subjects (Cheng et al. 1987; McIntyre et al. 1991). Studies in human subjects to determine amounts of SCFA in the gut are difficult, but evidence suggests that caecal concentrations of SCFA are approximately double those in the recto-sigmoid area (Cummings et al. 1987). The amount of SCFA produced in human subjects is very difficult to determine. Studies of arterio–venous differences across the gut indicate that 300–500 mmol are produced each day, whilst in individual cases this may reach 1–2 mol. Few dynamic studies have been carried out in man because of problems accessing the portal vein and differential metabolism of SCFA by individual tissues. The situation is complicated by endogenous production of acetate by the liver. Future stable-isotope studies may give more information in this area. SCFA production in the large intestine can be observed qualitatively by measuring levels in blood. However, only acetate appears in significant amounts in peripheral blood, although this responds in both time and amount to substrate fermentation in the large intestine (Pomare et al. 1985; Lifschitz et al. 1995). 2.2.1. Physiology and health. All SCFA are rapidly absorbed from the hindgut and stimulate salt and water absorption. They are then metabolized principally by the gut epithelium, liver and muscle, with virtually none appearing in urine and only small amounts in faeces. One of the most important properties of SCFA is their trophic effect on the intestinal epithelium. All three major SCFA are trophic when infused into the large intestine, although butyrate seems to be the most effective and propionate the least. What is perhaps more interesting is that infusion of SCFA into the hindgut leads to trophic effects in the small intestine (Sakata, 1987; Frankel et al. 1994) although the mechanisms for this are not fully determined. These trophic properties of SCFA have important implications, particularly for patients receiving enteral or parenteral nutrition, and in maintaining the mucosal defence barrier against invading organisms. 2.2.2. Acetate. Acetate is the principal SCFA in the gut. It is taken up by the epithelium, appears in portal blood and eventually passes through the liver to peripheral tissues where it is metabolized by muscle. In animal studies, the liver secretes free acetate when levels in portal blood fall below a critical level. Uptake and utilization of acetate by many tissues has been shown and is the principal route whereby the body obtains energy from carbohydrates not digested and absorbed in the small intestine. Current evidence suggests that the energy value of fermented carbohydrate is 6·3–8·4 kJ/g (1·5–2 kcal/g) (Livesey, 1990; Roberfroid et al. 1993). 2.2.3. Propionate. In ruminant species, propionate is a major glucose precursor but this is not an important role in hindgut fermenting species such as man. Propionate is largely cleared by the liver and has not been shown consistently to have significant effects on carbohydrate metabolism in human subjects. In vitro, propionate inhibits uptake of acetate into the cholesterol synthesis pathway, and in both rats and pigs propionate supplementation of the diet reduces cholesterol levels in blood. In human feeding studies of propionate only one out of three currently reported shows any change in blood cholesterol levels (Venter et al. 1989; Todesco et al. 1991; Stephen, 1994). 2.2.4. Butyrate. Butyrate is the most interesting of the SCFA, since in addition to its trophic effect on the mucosa it is an important energy source for the colonic epithelium and regulates cell growth and differentiation. Butyrate is almost entirely cleared by the colonic epithelium and is the principal energy source for the epithelial cells (Bugaut & Bentejac, 1993; Cummings, 1995). A defect in butyrate metabolism has been identified in ulcerative colitis patients and may be induced by S compounds generated in the largebowel lumen (Roediger et al. 1993; Pitcher & Cummings, 1995). The effect of butyrate on cell growth and differentiation is of great importance and has been the subject of a number of studies (Boffa et al. 1992; McIntyre et al. 1993). Butyrate brings about a concentration-dependent slowing of the rate of transformed cell growth and promotes expression of differentiation markers in vitro, thus leading to reversion of cells from a neoplastic to a non-neoplastic phenotype (Kim et al. 1980, 1994; Whitehead et al. 1986; Gibson et al. 1992). In vitro studies with colonocytes suggest an interaction between long-chain fatty acids, which result in decreased viability and differentiation of the cells, and butyrate, which has the opposite effect (Awad et al. 1991). In carcinogen-induced animal models of largebowel cancer, however, butyrate, either from fermentable carbohydrate sources such as resistant starch or purified NSP such as pectin, leads to increased cell turnover and in some studies increased tumour formation (Sakamoto et al. 1996; Young et al. 1996). The proliferative effects of Gastrointestinal physiology and function butyrate are probably not of pathological significance. Butyrate increases the proliferative index at the bottom of the crypt and thereby has a trophic effect on the mucosa. It does not, however, increase the proliferative index of the surface of the crypt (type II abnormality), which is closely connected with risk of colorectal cancer. Moreover, the relevance of these models to human carcinogenesis is doubtful for a number of reasons (see section 6.6). The expression of several genes is affected by butyrate and butyrate response factors have been identified in the upstream element of certain genes (Kim et al. 1994). 2.3. Interactions between the intestinal microflora and epithelial cells Although attachment to the epithelium is thought to be an important factor whereby bacteria colonize the gut, the mechanisms that allow certain species to maintain themselves in specific locations in the intestinal tract are largely unknown. An interesting new observation is that the intestinal microflora can influence expression of epithelial glycoconjugates, which may serve as receptors for S151 attachment of (pathogenic) micro-organisms. Recent papers by Bry et al. (1996) and Umesaki et al. (1995, 1997) report that host epithelial cells in the small intestine express fucosylated glycoconjugates in response to the presence of specific, strictly anaerobic bacteria (B. thetaiotaomicron and a segmented filamentous bacterium SIF13). Attachment of some pathogenic micro-organisms is decreased by the mutually beneficial crosstalk between the indigenous microflora and the host (Umesaki, 1989). The observation that one species can induce epithelial surface structures which influence attachment of other bacteria has significance for the use of CaCo-2 cell lines or gnotobiotic animals as model systems to study adherence or infectious diseases as well as for strategies to prevent and to treat gastrointestinal diseases. Thus, adhesion of bacteria to mucosal cell lines is important, but their mucus-adhering and degrading properties also need to be addressed. 2.4. The concept of healthy microflora It is a long-held belief, originating probably with Metchnikoff at the turn of the century, that some gut bacteria are Fig. 1. Generalized scheme of predominant groups of colonic bacteria, indicating how the genera may exhibit potentially harmful and beneficial functions. Gþ, Gram positive; cfu, colony-forming units. S152 S. Salminen et al. beneficial to health, whilst others may be harmful (Fig. 1). That bacteria in the gut can be harmful through production of toxins causing diarrhoea, mucosal invasion and activation of carcinogens is self-evident. Such bacteria are thought to include the Clostridia, sulfate reducers and amino acid-fermenting species. Potentially health-giving bacteria are thought to include principally the bifidobacteria and lactobacilli. These two genera do not include any significant pathogenic species and their dominance in the faeces of breast-fed babies is thought to provide protection against infection. In adults they may be the principal species responsible for barrier function and for stimulating healthy immune function. In addition, bacteria act in symbiosis with the host through fermentation. However, the colonic microflora is a complex interactive community of organisms and its functions are a consequence of the combined activities of the microbial components. Thus manipulation of the human intestinal flora offers the potential to improve health through a variety of mechanisms. 3. The gastrointestinal immune system 3.1. Gut-associated lymphoid tissue (GALT) (Brandtzaeg et al. 1989) The first, and in normal individuals only, contact that ingested bacteria, including probiotics, have with the immune system is with the GALT. The human intestine represents the largest mass of lymphoid tissue in the body, containing over 106 lymphocytes/g tissue. In addition, about 60 % of the total immunoglobulin (Ig; several grams) produced daily is secreted into the gastrointestinal tract. GALT is part of the mucosal immune system (i.e. gastrointestinal tract, respiratory tract, oral cavity, urogenital tract and mammary glands) and has unique cell types and mechanisms of immunity. The special nature of intestinal immunity has evolved under constant exposure to environmental antigens, whilst requiring an effective response to an invading pathogen despite the presence of dietary antigens. The difference between immune responses to dietary proteins and antigens of colonizing bacteria may play a role in the prevention of hypersensitivity reactions to food proteins. 3.2. The structure of GALT and cell distribution Intestinal immune cells are organized in different compartments: aggregated in follicles and the Peyer’s patches; distributed within the mucosa as diffuse lymphocyte populations; and in the epithelium (reviewed by McKay & Perdue, 1993). The GALT T-lymphocytes are not homogeneous. These are classified as CD4þ helper/inducer cells and CD8þ suppressor/cytotoxic cells, generating different cytokine profiles with distinct yet unproven functions (reviewed by Brandtzaeg et al. 1989; Brandtzaeg, 1995). The majority of the intra-epithelial T-cells have a suppressor/cytotoxic phenotype, contrasting with the lamina propria cells, which show mainly a helper/inducer phenotype. The lamina propria is also endowed with lymphocytes belonging to the B-cell lineage. These are mainly memory cells and plasmocytes, where 70–90 % of them are IgA-producing cells. The epithelial layer of the small-intestinal mucosa is arranged in folds, consisting of villi and crypts, which increase the absorptive surface area. The epithelium consists of a single layer of absorptive columnar epithelial cells, goblet cells and intra-epithelial lymphocytes. The intraepithelial lymphocytes are a heterogeneous population of cells. In the mouse, the primary intra-epithelial lymphocytes are CD3þ, CD8þ, T-cells with a g/d-T-cell receptor (T-cell receptor 1) and in man CD8 T-cells expressing an a/b-cell receptor (T-cell receptor 2). The proportion of T-cell receptor 1 cells in the epithelium is greater than in peripheral blood. The g/d-T-cell receptor cells are thought to mature in the epithelium rather than in the thymus, thus their development might be more susceptible to environmental exposures. Intra-epithelial lymphocytes are known to mediate both non-major-histocompatibility-complexrestricted and major-histocompatibility-complex-restricted cytotoxicity, and regulate neighbouring immune and epithelial cells by secreting cytokines. The epithelium is surrounded by the lamina propria, which comprises lymphoid organs such as reticular tissue and which contains plasma cells, T-helper cells, granulocytes and mast cells. The lamina propria is surrounded by smooth-muscle tissue. Along the small intestine are Peyer’s patches, which are organized lymphoid follicle aggregates. The Peyer’s patches are more accessible to micro-organisms than other epithelial surfaces of the gut, because they have reduced numbers of the mucus-secreting goblet cells. In addition, the epithelial layer of the Peyer’s patches contains specialized transport cells called M-cells, which lack microvilli and are able to phagocytose both soluble antigens and micro-organisms. 3.3. Immunophysiological regulation (Brandtzaeg, 1995) Different components of the mucosal immune system act to focus a specific response against offending antigens. The first line in this defence, immune exclusion involving IgA antibodies, is non-inflammatory (Brandtzaeg, 1995). The best-characterized component of the mucosal immune defence is the secretory IgA system (Brandtzaeg, 1995). IgA antibody production is abundant at mucosal surfaces. IgG-, IgM- and IgE-secreting cells function also, but at a significantly lower frequency in GALT. In contrast to IgA in serum, secretory IgA is present in dimeric or polymeric form in the gut. The predominance of IgA in the mucosal immune system results from IgA-selective T-cell regulation in GALT, particularly in the Peyer’s patches, where specific immune responses are generated (Biewenga et al. 1993). After being synthesized by IgA precursor cells, polymeric IgA is transported to the mucosal surface by epithelial transcytosis mediated by the polymeric immunoglobulin receptor, the secretory component. Secretory IgA is resistant to intraluminal proteolysis, and does not activate complement or inflammatory responses, which makes IgA ideal for protecting the mucosal surfaces. Hence, the main function of secretory antibodies is, in cooperation with nonimmunological defence mechanisms (Sanderson & Walker, 1993), to mediate immune exclusion of foreign antigens by Gastrointestinal physiology and function preventing epithelial adherence and penetration of invasive pathogenic micro-organisms, neutralizing toxins and viral multiplication. Although GALT is mainly involved in specific immune protection of the gut, there is evidence for a ‘common mucosal immune system’: an immune response initiated in GALT can affect immune responses at other mucosal surfaces (Brandtzaeg, 1995). The lymphocytes activated within Peyer’s patches disseminate via mesenteric lymph nodes, thoracic duct and the bloodstream back to the lamina propria, and traffic between other secretory tissues, including the respiratory tract and the lachrymal, salivary and mammary glands. There are differences between the upper and lower parts of the human GALT in isotype distribution of immunoglobulin-producing cells. Two IgA subclasses are available (IgA1 and IgA2 ). IgA1 immunocytes predominate in the small-intestinal mucosa, while IgA2 are most frequent in normal colonic mucosa. IgA2 in the colon is resistant to most bacterial proteases that cleave IgA1 (Brandtzaeg, 1995). Immune elimination is directed towards removal of foreign antigens that have penetrated the mucosa. This second line of defence involves antibodies such as IgG and a large number of mediators such as inflammatory cytokines, which are considered to be responsible for the pathophysiology associated with local inflammation (Brandtzaeg, 1995). Immune regulation pertains to the state of specific hyporesponsiveness induced by prior oral administration of antigens, inducing oral tolerance. Consequently, hyporesponsiveness to ubiquitous antigens such as dietary antigens is a hallmark of the intestinal immune system (Weiner et al. 1994). This has been taken to be a combined effect of immune exclusion and suppression of the systemic immune response, but it is still a matter of debate. 3.4. Regulation of antigen transfer (Isolauri et al. 1993a, b) Antigens are proteins foreign to the host. Factors that influence antigenicity include molecular complexity, solubility and concentration. Most antigens are macromolecules in the molecular mass range 10 000–70 000 Da. Apart from the barrier function, the intestinal mucosa is efficient in assimilating antigens. For this purpose, there are specialized antigen transport mechanisms in the villous epithelium and particularly in the Peyer’s patches (Heyman et al. 1982). The manner in which an antigen is transported across the mucosa determines the subsequent immune response (Heyman et al. 1982; Heyman & Desjeux, 1992; Isolauri et al. 1993a, b; Sanderson & Walker, 1993). Most antigens are excluded by a well-functioning mucosal barrier but an immunologically important fraction of antigen does bypass it (Heyman et al. 1982; Isolauri et al. 1993a, b). Antigens are absorbed across the epithelial layer by transcytosis, and the main degradative pathway entails lysosomal processing of the antigen. A minor pathway allows the transport of unprocessed antigens. The Peyer’s patches are covered by a unique epithelium and antigen transport across this is characterized by rapid uptake and S153 reduced degradation of antigens. In health, paracellular leakage of macromolecules is prevented because intact intercellular tight junctions maintain the barrier to macromolecules. Consequently, in healthy subjects antigen transfer is well controlled and aberrant antigen absorption does not occur. There is evidence that during the process of absorption across the intestinal mucosa, dietary antigens are altered into a tolerogenic form (Weiner et al. 1994). By interfering with this process intestinal inflammation is an important risk factor for the development of hypersensitive disorders (Fargeas et al. 1995). 3.5. Interactions between the intestinal microflora and the GALT (Moreau & Coste, 1993) After birth, the intestine is rapidly colonized by bacteria, which probably act as a source of antigens and non-specific immunomodulators. The dual role of the digestive flora on the immune system should be emphasized. Bacteria can be considered as antigens able to elicit specific systemic and local immune responses. Furthermore, they exert a considerable influence on the number and distribution of the GALT cell populations and play an important role in the regulation of immune responses. These data have emerged mainly from animal studies using germ-free and gnotobiotic animal models (see section 6.5). As direct evidence from human subjects is scarce, we can only extrapolate from experimental results obtained in mice. Such studies are important to determine the exact role played by different bacteria present in the digestive flora, with the aim of improving the bacterial equilibrium and allowing the best immune modulation by functional foods. The cellular and molecular events by which the digestive flora modulates the immune system are still poorly understood. The digestive flora is the major antigenic stimulus responsible for the migratory pathway and maturation of precursor lymphoid cells present in the Peyer’s patches. Consequently, it acts on the development and maturation of the IgA plasmocytes. In germ-free mice, IgA-plasmocyte number is decreased tenfold as compared with controls. It has been shown that the sequential establishment of the digestive flora from birth to weaning is responsible for the progressive increase in IgA plasmocyte numbers in the lamina propria of the small intestine in the growing normal mouse. In addition, Gram-negative bacteria such as Escherichia coli and Bacteroides play an important role in this immunologically non-specific effect. The digestive flora also modulates the specific immune responses at local and systemic levels. It allows the persistence of the systemic unresponsiveness to an antigen, induced by a previous feeding with the same antigen (oral tolerance) (Moreau & Gaboriau-Routhiau, 1996) and shortens the abrogation of oral tolerance mediated by cholera toxin or E. coli toxin (Gaboriau-Routhiau & Moreau, 1996), which seems to be a property of Gramnegative bacteria (M. C. Moreau and V. Gaboriau, unpublished results). In another study the presence of the gut flora modulated the intestinal antibody IgA response to rotavirus. Recently, the development of an experimental model of adult germ-free mice infected with a heterologous strain of S154 S. Salminen et al. rotavirus allowed investigation of the immunomodulating properties of a strain of Bifidobacterium on the enhancement of the intestinal anti-rotavirus IgA antibody response at cellular and faecal levels (Moreau et al. 1998). At the systemic level, in gnotobiotic mice harbouring a human strain of Bifidobacterium in the intestine or two bacterial strains from yoghurt, Lactobacillus bulgaricus and Streptococcus thermophilus, increases of the specific antibody response in serum and in the phagocytic activity of peritoneal phagocytes were observed respectively (Moreau et al. 1994). 4. Mucosal cell proliferation and differentiation (Wright & Alison, 1984) The intestinal mucosa of rodents and other mammals is renewed every 2–3 d (Wright & Alison, 1984). Maintenance of the architecture of the colonic mucosa, in particular of mucosal crypts, is a consequence of the balance amongst a number of factors. Proliferation of stem cells occurs near the base of the crypt. As enterocytes migrate up the crypt they differentiate, mature and become functional in terms of absorption and mucin secretion. 4.1. Cell proliferation This has been well described and a wide variety of methods are available for its estimation in vitro and in vivo in both animals and man (Goodlad & Wright, 1982; Goodlad, 1989). 4.2. Differentiation Light and electron microscopic examination of human colonic tissue has revealed that stem cells differentiate into a number of cell types, including mucus-secreting cells, columnar cells (thought to have an absorptive and a secretory function) and intestinal endocrine cells. Histochemical studies have shown alterations in secreted glycoproteins between differentiated and undifferentiated regions of the small and large intestine, indicating that a modification of carbohydrate structures accompanies goblet cell differentiation in rat and man (Boland et al. 1992). Furthermore, the mucin of normal colonic mucosa differs markedly from that in cancerous tissue and ‘transitional tissue’ in the early stages of neoplastic development (Boland et al. 1992). A characteristic of tumours is the presence of poorly differentiated cells: consequently, a dietary treatment that encourages differentiation is potentially beneficial. 4.3. Apoptosis Apoptosis (genetically programmed, autonomous cell death) associated with the removal of damaged cells is considered to be a protective event. 4.4. Mucosal enzymes (Szarka et al. 1995) Phase I, cytochrome P450 enzymes and phase II drugmetabolizing enzymes such as glutathione S-transferase (EC 2.5.1.18) and UDP-glucuronosyl transferase (EC 2.4.1.17) are widely distributed in the intestinal mucosa. These enzymes are involved in the biotransformation of mutagens, procarcinogens, steroids and other compounds of exogenous and endogenous origin. Modulation by dietary compounds may result in protection against toxic and carcinogenic damage to tissues (Wattenberg, 1983). In terms of deactivation, the enzyme glutathione S-transferase is of particular importance. It is present in many tissues in a variety of forms (p, m and a) and plays a critical role in protecting tissues from xenobiotics and carcinogens. Glutathione S-transferase activity has been found to be lower in individuals at high risk from colon cancer when compared with controls (Szarka et al. 1995). 5. Gastrointestinal function and disease 5.1. Gastrointestinal infections (Gracey, 1993; Savarino & Bourgeois, 1993) Acute infections of the gut are usually self-limiting, characterized by diarrhoea and often vomiting. The principal pathogens are viruses and bacteria such as Escherichia coli, Campylobacter spp, Vibrio cholerae, Staphylococcus aureus, Bacillus cereus, Clostridium perfringens, Salmonella spp, Shigella spp, Yersinia spp and a number of protozoa, especially Giardia lamblia, Entamoeba histolytica and Cryptosporidium parvum. Bacteria causing infection are usually classified according to whether they secrete an enterotoxin (toxigenic) or invade the bowel wall (invasive). Toxigenic diarrhoeas include cholera, and both enteropathogenic and enterotoxigenic E. coli, whilst the classic invasive organisms are Shigella (dysentery), Salmonella (typhoid) and enteroinvasive E. coli. Rotaviruses are most commonly found in diarrhoea of children and invade the small-intestinal epithelium. Acute diarrhoea is responsible for 3–4 million deaths annually worldwide, many of which are children, in which it accounts for 20–30 % of all mortality. Rotavirus is the most common cause of acute childhood diarrhoea. It is primarily seen in infants and young children, with a peak incidence between 6 months and 2 years of age. Rotaviruses invade the highly differentiated absorptive columnar cells of the small-intestinal epithelium, where they replicate. This results in partial disruption of the intestinal mucosa with loss of microvilli and decreased villus : crypt ratio. Rotavirus infection is associated with increased intestinal permeability. Jalonen et al. (1991) found increased lactulose : mannitol urinary recovery ratios in patients with acute diarrhoea compared with nondiarrhoeal patients. Concomitantly, the levels of immune complexes containing dietary b-lactoglobulin in sera were significantly higher in patients with rotavirus diarrhoea than in non-diarrhoeal patients. Enhanced macromolecular absorption in rotavirus gastroenteritis has been shown in several studies (Heyman et al. 1987; Isolauri et al. 1993a, b). A local immunoinflammatory reaction impairs the intestine’s barrier function. Impaired barrier function and defective handling of intraluminal antigens in the epithelial cells may be an important pathogenic mechanism in acute and chronic gastrointestinal disorders. It may abrogate tolerance to ubiquitous antigens, Gastrointestinal physiology and function including bacteria residing in the intestine (Duchmann et al. 1995). Chronic infection of the gut is much rarer and seen only in persons who have anatomical abnormalities of the gut such as blind loops, strictures or fistulas. Chronic infection with Tropheryma whippelli causes Whipple’s disease and intestinal bacteria are responsible for tropical sprue. Tuberculosis affects the gut, especially the ileo-caecal region, and chronic carrier states occur with amoebas. The main indigenous bacteria of the large intestine resist invasion by pathogenic species and this is part of the human host defence against diarrhoeal illness. This barrier function provided by the gut flora may be impaired during antibiotic use, where diarrhoea is common. Antibiotic-associated diarrhoea is usually due to invasion with toxin-producing species such as Clostridium difficile or Clostridium septicum. 5.2. Normal bowel habit (Cummings, 1993, 1994) Bowel habit is defined by the amount of stool passed, frequency of defecation and consistency of stool. It varies very widely throughout the world with daily stool weights in the range 100–400 g/d and stool frequency of three times per day to three times per week. In European countries and North America, daily stool weight is of the order of 100–150 g/d (Cummings et al. 1992). Bowel habit is controlled principally by two factors, first diet, and second gut motor activity (transit time). The foods that affect bowel habit are those which reach the large intestine, i.e. are nondigestible. The dietary components falling into this category are lactose (in lactase-deficient individuals), sugar alcohols, non-digestible oligosaccharides, resistant starch and NSP. Dietary fat and protein have little effect on bowel habit unless they are rendered non-absorbable by some technique (e.g. sucrose polyester). The mechanism by which non-digestible foods affect bowel habit depends on their fermentability. Foods that are not fermented appear in faeces and cause bulking depending on their inherent mass or water-holding capacity (i.e. bran and other intact cell-wall material). Most foods that reach the large intestine are fermented, yielding SCFA, which are absorbed and do not contribute to faecal bulk, and H2 and CH4 , which can expand faecal bulk but not mass. Fermentation also stimulates bacterial growth to produce biomass, which is the principal mechanism of increasing stool mass. A final mechanism that needs to be borne in mind is the interrelationship between intestinal bulk and motor activity. As bulk in the large intestine increases, so motor activity is stimulated and, in general, the greater the bulk the more rapid the transit. Motor activity expressed as transit time may also modulate stool output independently of dietary bulk (Cummings, 1993, 1994). 5.3. Constipation Constipation is a disorder of motor activity of the large bowel traditionally defined in terms of bowel frequency. The main symptom in constipation is straining at defecation, and discomfort, distension and incomplete rectal emptying are all considered part of the condition. Total gut transit time S155 is generally prolonged in constipated subjects. There are many causes of constipation, with diet one of the common reasons, particularly low-NSP diets, gluten-free diets, ‘low-residue’ diets and enteral feeds. Treatment of simple constipation is usually in the first instance by dietary means. The principle is to increase fermentable carbohydrates in the diet, especially NSP from whole-grain cereals. Thus, diet has a major role to play in controlling bowel habit. 5.4. Irritable bowel syndrome (IBS) (Thompson & Heaton, 1980; Thompson & Gomborone, 1993) IBS is one of the commonest disorders seen in the hospital gastroenterology clinic, but it is poorly understood. IBS (or irritable colon, mucus colitis, spastic colon) is a disorder of motor activity of the whole bowel, although colonic symptoms usually predominate. It occurs very widely throughout the world and is commoner in women. IBS has two main presenting features, abdominal pain and altered bowel habit (Thompson & Heaton, 1980). The cause of IBS is unknown but it occurs in many patients following dysentery or antibiotic use. In addition, patients often volunteer that specific foods upset them (food intolerance) and stress is clearly contributory. Wheat bran and other bulk laxatives are frequently given, but results have been very variable. They may aggravate symptoms through gas production, although in patients who are predominantly constipated they are of benefit. Because of a postulated disturbance in the colonic microflora in IBS a number of groups are currently trying the use of probiotics to ‘normalize’ the flora. 5.5. Inflammatory bowel disease (Podolsky, 1991; MacDonald, 1993; Tytgat et al. 1995) Two major disorders, Crohn’s disease and ulcerative colitis, are conventionally grouped together under the heading inflammatory bowel disease because both are characterized by chronic inflammation in the gut. However, it is best to consider them as separate conditions because they have characteristically different pathology, clinical courses, complications and management. The aetiology of neither is known. 5.5.1. Crohn’s disease. Crohn’s disease may affect any part of the gut from mouth to anus. Characteristically it occurs in the ileocaecal region and colon, and the inflammation is patchy or discontinuous. It frequently recurs after surgical resection of the affected areas of gut at or near the point of anastomosis of the bowel. The involved intestine is thickened, with ulceration of the mucosa, stricturing and fistula formation. Mouth ulcers and perianal abscesses are characteristic. Histologically there is transmural inflammation, with mononuclear cells, lymphoid aggregates and granulomata. Crohn’s disease occurs worldwide although it is uncommon in Central and South America, Africa and Asia. It is seen less frequently than ulcerative colitis, although rates have increased fivefold since 1950. It is predominantly a disease of the young with peak occurrence between the ages of 20 and 30 years with a second peak between 70 and 80 years. The cause is unknown but genetic factors are S156 S. Salminen et al. important with 10 % of patients having a close relative with the disease. Diet has been implicated, especially sugar, but a multicentre trial in which a diet excluding sugars and rich in NSP was given did not result in any conclusive benefit in patient management (Ritchie et al. 1987). More likely aetiological factors include bacteria and other micro-organisms. Diet has an important role to play in the management of Crohn’s disease. For patients who do not respond to conventional treatment in Crohn’s disease, various enteral and parenteral regimens providing bowel rest have been used. The rationale for such treatments is that the absence of food antigens in the bowel lumen reduces inflammatory immune reactions, motor and digestive activity, and gives the mucosa a chance to heal itself. Most patients relapse soon after introduction of normal food following bowel rest regimens. It has been suggested that dietary modification to exclude foods likely to cause symptoms (predominantly cereals, dairy products and meat) can lead to extended periods of remission. 5.5.2. Ulcerative colitis. Ulcerative colitis is a chronic inflammatory condition of the mucosa of the large bowel that causes bloody diarrhoea. It is one of the diseases of modern civilization, being first described with certainty in 1909 and predominantly affecting industrialized populations. It has an overall prevalence of 40–120 cases per 100 000 of the population in Western countries and is uncommon, although beginning to emerge, in Africa and India. It affects the sexes equally and usually presents between the ages of 20 and 40 years. Its cause is unknown but there is probably a genetic component, with a 15-fold increased risk amongst close relatives of patients. Animal models point to an involvement of the immune system, but also to the necessary presence of bacteria in the colon to produce colitis. The observation that healthy colonocytes use butyrate for their metabolism and that this is defective in ulcerative colitis has raised the possibility of dietary factors in its aetiology. However, no convincing evidence for a dietary factor has emerged, although the inhibition of butyrate oxidation by S-compounds leaves the possibility of diet combining with the colonic microflora as a possible initiating factor of the inflammation (Pitcher & Cummings, 1996). 5.6. Food allergy (Isolauri, 1995) Food allergy is defined as an immunologically mediated adverse reaction against dietary antigens. Food allergy can affect several organ systems, the symptoms commonly arising from the gut, skin and respiratory tract. Despite the wide spectrum of clinical manifestations, there are at least two prerequisites for the development of food allergy: dietary antigens must penetrate the intestine’s mucosal barrier, and the absorbed antigens must cause harmful immune responses. The immaturity of the immune system and the gastrointestinal barrier may explain the peak prevalence of food allergies in infancy. In food allergy, intestinal inflammation and disturbances in intestinal permeability and antigen transfer occur when an allergen comes into contact with the intestinal mucosa. During dietary elimination of the antigen, the barrier and transfer functions of the mucosa are normal. It has, therefore, been concluded that impairment of the intestine’s function is secondary to an abnormal intestinal immune response to the offending antigens. 5.7. Colorectal cancer (Faivre et al. 1985) Colorectal cancer is unequally spread throughout the world (Burkitt, 1971). It is amongst the three most common malignancies in most industrialized countries including Western Europe, and its survival rate has improved little during the past decades, being of the order of 40 % at 5 years. The most common locations in high-risk countries are the left colon and the rectum, whereas right-colon cancers are proportionally more common in low-risk areas such as Japan. About 5 % of colorectal cancers are truly genetic diseases (hereditary nonpolyposis colorectal cancer and familial polyposis coli), transmitted as autosomal dominant, but the majority of colorectal cancers are sporadic, and are mostly influenced by environmental factors, in particular diet, with a potential interaction between a genetic background and diet (Boutron et al. 1996). From 70 to 80 % of left colon and rectal cancers in Western countries follow the so-called adenoma–carcinoma pathway, with possibly less in the case of right-colon cancers (Bedenne et al. 1992). This is of major importance as it provides the opportunity of studying precancerous lesions both in aetiological studies such as case–control and cohort studies, and for intervention studies, where studying adenoma recurrence or growth is easier and brings results more rapidly (Boutron & Faivre, 1993). 6. Methodology 6.1. Human intestinal microflora (Collins & Gibson, 1998) The identification of factors controlling or influencing the composition of the human intestinal microflora, including prebiotics and probiotics, may be compromised by the precision of current methodologies for determining bacterial composition which are based almost entirely on phenotypic approaches. Whilst these have met with some success, when done properly, they are time-consuming, laborious and lack the resolving power necessary to analyse the complex microbiota at the species or subspecies level. Traditional gut microbiological methodologies are usually based on morphological and biochemical properties of the organisms (Table 2). Whilst such an approach is costeffective and allows the processing of replicate samples, the procedures used may be unreliable and may lack resolution. For example, the metabolic plasticity of organisms is problematic and the test used may not be reproducible. Phenotypic characterization does not allow a high degree of fidelity and is most useful for genus level identification. In some cases this situation is eased if the test organism exhibits a specific metabolic trait. For example, bifidobacteria may be detected, on a qualitative basis, by the production of fructose-6-phosphate phosphoketolase activity. An additional problem is that traditional cultivationbased methods may result in underestimation of microbial diversity, due to the presence of organisms that cannot be Gastrointestinal physiology and function S157 Table 2. Methods for study of the human gut microbiota Method Advantages Disadvantages Morphological and biochemical characteristics Straightforward to carry out. Can run in parallel a large number of replicates. Relatively inexpensive Specific biomarkers, e.g. certain cell-wall antigens, cellular fatty acids, plasmid profiles, serotyping, resistance to antibiotics Ribotyping (RNA polymorphisms) Cultural procedures may not be required 16 S ribosomal RNA typing Reliable. Very high discriminatory power High fidelity. Reliable. Cumulative database allows placement of unknown species. Applicable to culturable and non-culturable forms. Allows probe development cultivated, which therefore elude isolation. This has led to the development of alternative strategies for assessing microflora changes. The detection of biomarkers that may be attributed to certain components of the flora offers some potential. However, to be wholly effective there would be a need for all the major bacterial components of interest to be separable by individual biomarkers, e.g. changes in cellular fatty acids. This may be feasible but would be difficult to prove in a reliable manner. An attractive solution to the problem of determining microflora changes accurately, lies in the application of modern high-resolution molecular genetic techniques. Recent advances in the field of molecular biology are revolutionizing the characterization and identification of micro-organisms (Pace, 1996). For example, molecular sequence analysis, particularly of ribosomal RNA (rRNA), provides a powerful tool for determining the genetic interrelationships of micro-organisms, and allows systematic monitoring of the gut flora response to dietary intervention. By utilizing diagnostic sequences within the rRNA, it is possible to design gene probes that facilitate precise identification. The use of polymerase chain reaction technologies may also allow access to non-culturable micro-organisms. Over the next few years, 16 S rRNA sequence analysis is expected to rapidly advance our knowledge of the true genetic diversity of the gut microbiota, including organisms that evade traditional identification, due either to a lack of taxonomic resolution and/or non-culturability. Molecular approaches have already been used to determine changes in the composition of the microbial gut flora (Langendijk et al. 1995; Kok et al. 1996; McCartney et al. 1996; Wilson & Blitchington, 1996). Clearly, however, before such techniques are routinely used in gut microbiological applications, the fidelity and efficacy of such methods need to be rigorously evaluated. A comparative phylogenetic framework of gut micro-organisms, based on genetic material such as 16 S rRNA, would allow highly discriminatory and dependable diagnostic probes to be developed. Preferably, the probes should be validated using different procedures, such as in situ and dot blot hybridizations. Involves operator subjectively to recognize different colonial and cellular morphologies. Lack of discriminatory power and subject to metabolic plasticity of the organisms. Applicable only to culturable bacteria Cannot assign the position of hitherto unknown species. Relies on all test organisms having unique biomarker. Stability of the biomarker may be questionable Applicable only to culturable forms. Cannot assign the taxonomic position of any unknown species Costly for both reagents and large-scale equipment, e.g. automated sequencers. Recommended for partial use only Another approach to analyse the genetic diversity of complex microbial populations is denaturing gradient gel electrophoresis or temperature gradient gel electrophoresis. The technique is based on the separation of polymerase chain reaction-amplified fragments of genes coding for 16 S rRNA, all of the same length (Muyzer et al. 1993). This results in unique separation patterns for different microbial populations, and will contribute to the description of changes or differences in microflora composition of uncharacterized microbial populations. The potential benefits of such technologies in gut microbiology, especially dietary modulation for improved health, are large, particularly when used in conjunction with traditional phenotypic procedures. The use of molecular genetic approaches for qualitative and quantitative monitoring of the human intestinal microbiota will constitute an essential step forward for determining the validity of the functional food concept, when directed towards the role of the gut flora. 6.2. Functional analysis of the gut microflora (Rowland, 1995) The complexity of the gut microflora, coupled with timeconsuming procedures necessary to identify and enumerate the anaerobic components, makes their characterization by conventional methodology difficult and expensive. In particular, such methods are not suited to studies involving large numbers of subjects or treatment regimens. Less comprehensive studies (e.g. identification of major groups) are open to the criticism that any induced changes occurring in genera or species other than those being enumerated will be missed. Furthermore, although bacteriological investigations are useful in describing the basic ecology of the gut, they are of less value in studies of metabolism, nutrition and cancer. An alternative approach is to use biochemical assays that measure the functional activity of the flora as a whole and thus permit deductions to be made regarding the role of the flora in the metabolism of dietary components. In addition, by selecting microbial enzyme activities or metabolic endpoints resulting in compounds with potentially toxic or beneficial effects, probable health consequences for the host can be assessed. S158 S. Salminen et al. 6.2.1. Bacterial enzymes. The bacterial enzymes commonly assayed include b-glucuronidase (EC 3.2.1.31), bglucosidase (EC 3.2.1.21), azoreductase, nitroreductase, nitrate reductase (EC 1.7.99.4), the conversion of precarcinogen 2-amino-3-methyl-7H-imidazo[4,5-f ]quinoline (IQ) to 7-hydroxy-2-amino-3,6-dihydro-3-methyl-7Himidazo[4,5-f ]quinoline-7-one (7-OHIQ). The substrates of these enzymes and the functional and health implications of their products have been extensively reviewed (Rowland, 1995). For example, bacterial b-glucuronidase in the colon is able to release carcinogens from hepatically derived glucuronic acid conjugates and is a critical factor in the enterohepatic circulation of drugs and other foreign compounds. b-Glucosidase hydrolyses plant glycosides to release the aglycones, many of which are mutagenic, although some exert anti-carcinogenic activity (Rowland, 1995). Azo- and nitroreductases reduce their substrates to amines, which are usually more toxic than the parent compound, and nitrate reductase generates the highly reactive and toxic anion nitrite. Bacterial conversion to 7-OHIQ is one pathway of activation of the pre-carcinogen IQ (formed during cooking of meat and fish) to a genotoxic derivative. 6.2.2. Bacterial metabolites in faeces. Faecal metabolites that are indicators of bacterial activity relevant to colonic health include NH3 , a toxic product of bacterial breakdown of protein and urea (Clinton, 1992), and phenols and cresols, which are derived from amino acid catabolism by gut bacteria (Macfarlane & Macfarlane, 1995). The production of NH3 is closely related to bacterial activity and is associated with certain toxic events in the gastrointestinal tract. NH3 is considered to be a potential tumour promoter in the colon, and has been postulated to enhance neoplastic transformation in the gut. Other gut bacterial products with possible adverse effects on the colonic mucosa include N-nitroso compounds, which are potentially carcinogenic substances formed by bacterial catalysis of the reaction of nitrite and nitrogenous compounds in the colon (Rowland, 1995), diacylglycerol, a putative tumour promoter derived from lipid breakdown, and secondary bile acids, deoxycholic and lithocholic acids, also putative tumour promoters. 6.2.3. Assessment of cytotoxicity, genotoxicity and mutagenicity of faeces. An alternative approach to assaying enzymes or metabolites in faeces is to assess toxicological activity of fractions using short-term tests for toxicity, genotoxicity and mutagenicity. This provides a direct estimate of the potential of the faecal sample to damage the colonic mucosa and has been used to provide insights into possible processes involved in colon cancer. Usually, the aqueous phase of human faeces (faecal water) is used (Rafter et al. 1987). Cellular toxicity can be assessed using rapid colorimetric assays in multiwell plates. For example the 3-[4,5-dimethylthiazol-2-yl]-2,5diphenyltetrazolium bromide (MTT) assay (Mosmann, 1983) involves the pH-dependent conversion of water-soluble MTT to water-insoluble formazan and provides a reliable estimate of viable cell number. Recently, using the Comet assay, it has been shown that the genotoxicity of faecal water varies markedly between individuals, with at least some of the DNA damage occurring via an oxidative mechanism (Venturi et al. 1997). 6.2.4. Susceptibility of functional markers to dietary change. In animal models, major changes in activities of bacterial enzymes and in levels of bacterial metabolic products have been seen after a wide range of dietary changes. These include the type and level of dietary fat and protein, supplementation with dietary fibre and resistant starch, and addition of oligosaccharides (Rowland, 1991). In many cases changes were seen in the absence of alteration in composition of the gut flora. In contrast, modification by diet of bacterial metabolism in human subjects has proved more difficult. However, changes in enzyme activities and concentrations of NH3 , phenol and cresol have been detected in volunteers consuming lactulose (Terada et al. 1992) and lactobacilli (Goldin & Gorbach, 1984). A low level of NH3 production in the gut is associated with low-protein, high-fibre diets, which appear to be protective against cancer of the colon. NH3 levels have been shown to be elevated in rats consuming a diet containing high-risk factors for colon cancer (Hambly et al. 1997). The cytotoxicity of faecal water strongly correlates with bile acid concentration in faeces and is increased in individuals on high-fat diets (Rafter et al. 1987) and decreased in subjects on high-resistant-starch diets (van Munster et al. 1994). 6.3. Digestibility and bioavailability of foods The methodology for studying non-digestibility of foods is an important area for understanding the effects of food and food components on intestinal microflora and physiology. A compilation of current methodology is given in Table 3. It is important to identify the nature of molecules, their chemical bonds and molecular size to understand digestibility. In vitro digestion studies and markers for absorption and excretion are of value. Animal models offer a means of simulating different digestion extremes and human volunteer studies enhance the understanding of in vivo digestibility of foods and bioavailability of nutrients. 6.4. Large-bowel function Study of large-bowel function is extremely difficult, mainly because of its inaccessibility. However, the large bowel has unique aspects of metabolism; the principal events in the lumen are anaerobic and end-products such as H2 and SCFA are not produced by other biochemical reactions in the body. These products are absorbed and appear in blood and breath, and can be used to study intraluminal events. Many investigators have used faeces and their composition as a guide to intracolonic events. Such studies have often been criticized because of the lack of representativeness of faeces, but for some aspects of colonic metabolism, such as the gut microflora of the lumen, they are probably acceptable. The study of the large intestine has spawned a large number of in vitro models, particularly of fermentation. A short summary of methods is given in Table 4. Gastrointestinal physiology and function S159 Table 3. Methods to study the digestibility and bioavailability of foods Method Advantages 1. Chemistry Disadvantages Identifies nature of molecules, chemical bonds, molecular size, etc. (e.g. most b-glucans are not digested) Studies with pancreatic and other enzymes 2. In vitro model 3. Blood appearance 4. Breath 5. Ileostomy model 6. Intestinal intubation: jejunum, ileum, colon, perfusion/aspiration, single or multilumen tubes 7. Faecal analysis 8. In vitro fermentation 9. Animal models: germ-free/conventional, fistulated Requires dedicated carbohydrate chemistry laboratory Glucose tolerance, chylomicrons. Applicable to a large number of samples H2 or CH4 , 13 CO2 . Simple non-invasive for human studies Probably the gold standard for study of digestion in stomach and small bowel Valuable for dynamic studies of flow rate, site of digestion in small bowel, nutrient concentrations and physical form, and identification of intermediate products of digestion Relatively straightforward and gold standard for overall digestion and fermentation Various batch and continuous culture methods available. Good for modelling physiology Gives greater access to tissues. Germ-free, human associated and conventional flora Care needed to mimic conditions in gut and hence rate and extent of digestion Requires human subjects and can be affected by factors other than those determining digestion Difficult to quantify. Prolonged studies (16–24 h) needed for fermentation. Large subject variability Requires access to patient population. May underestimate losses to caecum because of microbial colonization of ileum and fermentation in bag Very invasive human studies. Need X-ray control. Presence of tube may alter normal physiology. Difficult to make quantitative Requires accurate faecal collections with balance markers and good methods validated for faeces Needs well-founded microbiology laboratory Animal facilities needed. Extrapolation to man Table 4. Methods for studying large-bowel function 6.5. Gut-associated lymphoid tissue Focus of study Method Bowel habit Motor function Diary record Motility recording Recto-anal manometry and pelvic floor neurophysiology Transit time: whole gut – non-absorbable markers Partial transit time: small bowel – isotope-labelled meal and gamma scanning – lactose breath H2 large bowel – X-ray following marker ingestion Microflora, and enzymic activities Fat, N, biomass, carbohydrate, DM, bile acids, sterols, pH Faecal water: osmolarity, pH electrolytes bile acids genotoxicity/cytotoxicity Occult blood Transport physiology Batch cultures, single chemostats, multi-chamber chemostats Short-chain fatty acids, branched-chain fatty acids, bile acids H2 , CH4 , 13 CO2 General histology and immuno-staining DNA adducts Mutational analysis (e.g. Kras) Proliferation markers Genotyping Apoptosis Composition and structure Thickness Breakdown Barium studies Plain film of abdomen Proctoscopy Sigmoidoscopy (including flexible) Colonoscopy Vasculature pH Motility Faecal analysis In vitro models Blood (including portal blood) Breath Mucosal biopsies Exfoliated cells Mucus X-ray Endoscopy Angiography Radiotelemetry The presence of the digestive flora has a considerable influence on the immune system of the host (see section 3.5) and the principal methodology to study this is the use of germ-free animals, also termed axenic animals. By comparing germ-free and conventional animals, it is possible to highlight the role of the digestive flora in immune function. Moreover, the role played by bacteria isolated from the digestive flora, or used as probiotics, can be analysed by inoculating the gut of germ-free animals with these bacteria. These are called gnotobiotic animals. Recently, germ-free mice associated with human flora have been developed allowing in vivo studies of functional properties of probiotics and prebiotics used in human nutrition. The advantages of gnotobiotic animal studies are to determine which kind of immune response a given bacteria established in the gut is able to exert, e.g. non-specific and/or specific immune response; inductive or suppressive immune response. Many methodologies are available. They are summarized in Table 5 and marked with (H) when they are applicable to human studies. The disadvantages of gnotobiotic models are the expensive animal facilities they need and the limits of the animal species studied. Moreover, it is not certain that a modulating bacterial effect observed in gnotobiotic conditions will be expressed in conventional conditions. Thus, other studies using conventional animals are needed. Current methods utilize in vitro cultures of systemic or intestinal lymphoid cells (De Simone et al. 1993), cellular assays with colonic cell lines (Schiffrin et al. 1995) and in vivo assays with conventional animal models (Perdigon et al. 1996). Two types of specific immune response can be assessed at the intestinal level. These include the suppression of humoral and cellular immune responses to chronically administered antigens at the systemic level (immune regulation) and the induction of a protective IgA antibody response at the mucosal level S160 S. Salminen et al. Table 5. Selected methods to study intestinal immune function Immune response Methodology Advantages Proliferative assays In vitro: measurement of the proliferation of systemic or intestinal lymphoid cells after stimulation with mitogens or cell components Easy for systemic assays: from blood samples (H) Cytokine production In vitro: after proliferative assays Phagocytic activity In vitro: peritoneal cells, circulating cells Development of new methodologies (H) Biomarkers (H) Easy for systemic assays: from blood samples (H) Modulation of molecular expression in intestinal cell lines In vitro: HT-29, CaCo-2 cell line cultures Facs analysis, histochemical methods Measured by ELISA or ELISPOT at several levels: (1) In serum: soluble IgA antibodies (2) In blood: circulating IgAproducing cells (3) In faeces IgA antibody response (4) In saliva (5) Whole gut lavage fluid Oral tolerance to dietary antigens Ex vivo: proliferative assays of lymphoid blood cells with specific antigen Inflammatory cytokine production by lymphoid blood cells In vivo: intestinal permeability In vitro: Ussing chamber Cell lines originated from human intestine Specialized line cells Disadvantages Not easy for intestinal assays in human subjects: need biopsy Technical difficulties Need correlation with in vivo assays to give a biological significance Need correlation with in vivo assays to give a biological significance Correlation with specific immune response poorly understood Technical difficulties for intestinal phagocytes Adenocarcinoma lines Absence of correlation with intestinal cellular environment Biomarkers Easy in human subjects (H) Reflects intestinal response (H) Easy, allows kinetic studies (H) Easy (H) Invasive, but avoids intestinal biopsy (H) Direct measure of the unresponsiveness state (H) Biomarkers: TNF-a (H) Increase in food hypersensitivities (H) Direct correlation with antigen transfer Does not reflect the intestinal response Does not reflect transepithelial transport into gut lumen Individual and daily variations, proteolytic activity, reflects only colonic response Individual and daily variations Needs hospitalization Cause or consequence of oral tolerance breakdown? Not specific to immunological changes Invasive and difficult technology H, applicable to human studies; IgA, immunoglobulin A; TNF-a, tumour necrosis factor-a; Facs, fluorescence-activated cell-sorting; ELISPOT, enzyme-linked immunospot assay. (immune exclusion). The definition of biomarkers is still incomplete. Several methodologies have been developed for IgA response measurements. Sampling needs blood collection or intestinal biopsies or stool and saliva collection. Problems in faeces and saliva handling and daily variations in expression of IgA activities have not been solved. In the case of oral tolerance, the biomarkers available in human subjects are defined as the absence of proliferation of cultured blood cells with antigen or down-regulation of inflammatory cytokines involved with food allergies (Benlounes et al. 1996; Sütas et al. 1996a, b). Other methodologies must be developed according to further knowledge about oral tolerance mechanisms (Weiner et al. 1994). 6.6. Epithelial cell proliferation and colon carcinogenesis 6.6.1. Biological markers for colorectal carcinogenesis. A relationship between colorectal carcinogenesis and abnormal cell proliferation has been demonstrated (Lipkin, 1988) in studies of the mucosa in patient groups at high risk of cancer, e.g. ulcerative colitis and familial adenomatous polyposis, and in animals exposed to carcinogens that target the colon. Two changes in colonic cell proliferation have been described (Risio, 1992): an increase in the total number of proliferating cells, or hyperproliferation, which is not a specific marker of cancer risk, and a progressive shift of proliferating cells to the crypt surface (stage II abnormality) which is more specific for tumour risk. 6.6.2. Cell proliferation. Several techniques have been developed to measure cell proliferation in colonic mucosa and assess the influence of diet. Measurement of crypt-cell production rate in microdissected crypts is considered to provide the best assessment of proliferation with the fewest artifacts. However, since it requires in vivo treatment with vincristine, it is not suitable for human studies. Change in the rate of cell proliferation in the normal mucosa may be less reliable as a biomarker for diet-related cancer risk (Wasan & Goodlad, 1996) and is just one of the processes contributing to colonic mucosal crypt architecture. Other events are differentiation, exfoliation and apoptosis. It is likely that the best predictor of cancer risk is an overall assessment of these events. Markers of early epithelial events have been used in animal models. These include DNA damage, microadenomas and aberrant crypt foci in the mucosa. The microgel electrophoresis (Comet) assay has been used for assessing Gastrointestinal physiology and function DNA damage in the colonic mucosa. Induction of aberrant crypt foci has been particularly widely used, as it is easy to observe macroscopically. However, its reliability as a marker of colorectal tumour risk is a matter of debate. In man, aberrant crypts and microadenomas, similar to those described in animals, have been described (Roncucci, 1992) but their correlation with other well-known markers of risk has not been established. Other markers of cell proliferation include PCNA and Mib 1; both are proteins appearing at specific stages in the cell cycle. 6.6.3. Differentiation. Methods exist to measure the state of differentiation of the mucosal epithelial cells by histochemical staining of mucins using binding to specific lectins. The methods can be applied to tissue sections from human and animal biopsies after fixation. Such approaches have been used to study colonic epithelium in rats and human subjects at various stages of neoplasia. They have not achieved widespread use for investigating dietary modification of neoplastic processes. 6.6.4. Apoptosis. Identification of oligonucleotide fragments by in situ end labelling using immunoperoxidase techniques forms the basis for various methods that can be applied to sections of colonic tissue from human subjects and animals and to cell suspensions (Ansari et al. 1993). Apoptosis in cell cultures from transformed colon tissues has been determined by measuring cell loss from monolayers (Hague et al. 1993). 6.6.5. Products used in experimental carcinogenesis (Martin et al. 1981). Spontaneous colorectal cancer is exceptional in animal models, but tumours can be created easily in rats, mice or hamsters using chemicals such as N-methylnitrosourea and N-methyl-N 9-nitro-N-nitrosoguanidine (Table 6). These products are direct carcinogens, which explains their local efficacy and their specificity when administered intra-rectally. Others such as 1,2-dimethylhydrazine or azoxymethane must be first metabolized by the liver, then at the level of the target cell, in order to be carcinogenic. 1,2-Dimethylhydrazine and derived azo and azoxy alkanes represent the carcinogens most commonly used to induce intestinal carcinomas in rats, mice and hamsters, by the oral or subcutaneous route. In rats, precancerous lesions such as adenomas are exceptional, cancers arising most commonly without any precursor adenoma, which is the opposite to what is observed in man. In mice, it is easier to induce adenomas than carcinomas (Maskens, 1976). S161 Secondary bile acids are cocarcinogens. They have been shown to promote colorectal carcinogenesis in animal models by increasing tumour formation rate induced by carcinogens, or increasing colonic cell hyperproliferation through the production of diacylglycerol and stimulation of protein kinase. 6.6.6. Types of lesion (Weisburger, 1973). The most studied lesion is adenocarcinoma induced either directly or, more often, indirectly. When azoxymethane is used by the subcutaneous or intramuscular route, intestinal carcinomas are formed in 100 % of cases. Small doses induce tumours in the proximal colon and the caecum, whereas larger doses produce tumours mainly in the distal colon. Such tumours arise on flat mucosa and form plaques, thus mimicking human infiltrating tumours, which arise without any detectable adenomatous tissue and are named de novo cancers. However, this type of tumour is rare in human carcinogenesis, particularly in Western countries where tumours are mainly of the fungating type and arise in a pre-existing adenoma. Aberrant foci are interesting to study as they can be easily observed macroscopically. After a single injection of azoxymethane, aberrant foci have been described after early slaughter of the animals. These lesions are considered by some authors to be an early marker of tumour risk. They are more common in the rectum (90 % of animals) than the caecum (10 % of animals), but tend to migrate with time, with a decrease in rectal lesions and an increase in caecal lesions after 4 weeks. Changes in colonic cell proliferation, which will be described in man elsewhere, are largely used to test the protective effect of products, e.g. Ca, against secondary bile acid-induced proliferation. 6.6.7. Transgenic mouse models for colon cancer studies. A number of inbred mouse models carrying germ-line mutations at the Apc gene (the murine homologue of APC, which is mutated in patients with familial adenomatous polyposis) have been developed. These animals exhibit spontaneous tumours throughout the intestinal tract, usually in the first few months of life. They are of use in studies of the interaction of diet and colon cancer and provide a model that dispenses with the need for chemical induction of carcinogenesis. The susceptibility of tumour incidence in these mice to dietary modulation is under investigation in a number of laboratories. Table 7 lists four such mouse models in current use. Table 6. Current methods for studying colon carcinogenesis in animals Products Route MNNG (N -methyl-N 9-nitro-N nitrosoguanidine) and MNU (N -methylnitrosourea) DMH (1,2-dimethylhydrazine) Intra-rectal instillations Azo and azoxy alkane derivatives Subcutaneous or intramuscular Secondary bile acids Intra-rectal or oral Oral or subcutaneous Lesions Colon adenocarcinomas, squamous cell anal carcinomas, spleen and liver haemangiomas Multiple colorectal adenocarcinomas and a few adenomas in rats In mice adenomas, and a few carcinomas; adenocarcinomas in plaques (de novo) Alone: colonic cell hyperproliferation After DMH: increased number of carcinomas S162 S. Salminen et al. Table 7. Transgenic mouse models currently in use in colon cancer studies Model Mutation Min (multiple intestinal neoplasia), usually used in the heterozygous cross with C57B16/J mice Apc1638N Apc 1638T Apc D716 Tumour yield per animal Germline nonsense mutation at codon 850 of Apc About 100 adenomas with the majority in the small intestine Apc disrupted at codon 1638, probably leading to a null allele for Apc Apc disrupted at codon 1638 leading to a truncated polypeptide Apc disrupted at codon 716, leading to a truncated protein About five intestinal tumours No intestinal tumours at 5 months 200–500 intestinal adenomas 6.6.8. Limits of experimental models (Maskens & Dujardin Loits, 1981). Experimental carcinogenesis in animals creates mainly infiltrating de novo carcinomas, and is an indirect process in most cases, i.e. multiple metabolic transformations are needed, the last occurring in the colon itself and involving bacterial enzymes such as bglucuronidase, azoreductase or nitroreductase. In man, in particular in high-risk countries such as Western Europe, North America or Australia, most colorectal tumours are polypoid, fungating, and arise on a pre-existing adenoma. It has been estimated that, in the distal colon and rectum, where most tumours arise, over 80 % of cancers arise through the adenoma–carcinoma pathway. Indirect evidence, such as the type of mutation of the p53 protein, points to a major role played by secondary bile acids, i.e. endogenous carcinogenesis, whereas carcinogenesis in animals is called exogenous carcinogenesis. The role of the latter may be more important in low-risk countries such as Japan, but is likely to be of little importance in high-risk countries. Therefore, it is difficult to extrapolate from animal models to man, in particular regarding the importance of bacterial enzymes, apart from the major role played by the 7 a-dehydroxylase, which converts primary into secondary bile acids. 7. Human studies on the effects of food and food components Among the components likely to be used in functional foods, prebiotics and probiotics are already used as food ingredients. Probiotics have been variously defined (Fuller, 1991, 1992; Havenaar & Huis in’t Veld, 1992) largely on the basis of their initial use in animal feeds. For the purposes of human nutrition we suggest that a probiotic is best defined as ‘a live microbial food ingredient that is beneficial to health’. A prebiotic is a ‘nondigestible food ingredient that beneficially affects the host by selectively stimulating the growth and/or activity of one or a limited number of bacteria in the colon, that have the potential to improve host health’ (Gibson & Roberfroid, 1995). Another approach is the use of synbiotics. A synbiotic has been defined as ‘a mixture of probiotics and prebiotics that beneficially affects the host by improving the survival and implantation of live microbial dietary supplements in the gastrointestinal tract, by selectively stimulating the growth and/or activating the metabolism of one or a limited number of health-promoting bacteria, and thus improving host welfare’ (Gibson & Roberfroid, 1995). 7.1. Prebiotics (Gibson & Roberfroid, 1995) The key criterion for a food ingredient to be classified as a prebiotic is selective stimulation in human subjects of the growth of potentially beneficial bacteria in the gut. A prebiotic may repress pathogen growth or virulence and induce systemic effects that can be beneficial to health. The prebiotics identified today and which have served to introduce the concept (Gibson & Roberfroid, 1995) are nondigestible oligosaccharides that are fermented in the colon. They are obtained either by hot-water extraction from plants, eventually followed by enzymic hydrolysis of the extracted molecules, or synthesis from a disaccharide using -osyl transferases. The prebiotics that are available in Europe at present or that are being developed, belong to two groups, characterized by the major -osyl monomer they are composed of; namely, the fructosyl prebiotics (fructooligosaccharides) and the galactosyl prebiotics (galactooligosaccharides). Daily intake of fructosyl prebiotics for a few weeks leads to a selective stimulation of the growth of bifidobacteria. Many studies using different dose regimens and different methods of microbial analysis have shown selective stimulation of bifidobacteria (Roberfroid et al. 1997). Ito et al. (1993) have reported the same effect for galactosyl prebiotics. In addition, the experiments reported by Gibson et al. (1995) demonstrate that the growth of bifidobacteria is accompanied by a reduction in the number of other populations, e.g. bacteroides, clostridia and fusobacteria, thus leading to a major modification in the composition of the colonic microflora. Lactitol (4-O-b-d galactopyranosyl-d-glucitol) has been shown to have prebiotic properties (Ballongue et al. 1997). The nutritional properties of prebiotics have been summarized in the report of an International Life Sciences Institute (ILSI) Europe workshop entitled Colonic Microflora: Nutrition and Health (Roberfroid et al. 1995) and in the proceedings of the 1st International Conference on East–West Perspectives in Functional Foods organized by ILSI (Roberfroid, 1996). In a recent monograph, Cummings (1997) also reviewed the nutritional properties of prebiotics. The health-promoting consequences of prebiotic fermentation include increased faecal biomass and consequently stool weight and/or stool frequency (Roberfroid et al. 1993; Gibson et al. 1995). Colonic fermentation of prebiotics produces SCFA mainly acetate, propionate and butyrate (Roland et al. 1995). Through fermentation and absorption of SCFA prebiotics have part of their energy salvaged. Their Gastrointestinal physiology and function digestible energy is of the order of 4–9 kJ/g (1–2 kcal/g) (Roberfroid et al. 1993; Molis et al. 1996). This property has led to their use as bulking ingredients, sugar substitutes and, for inulin, as a fat replacer (Coussement, 1996). 7.2. Probiotics (Salminen et al. 1996a,b,c) The bacterial genera most often used as probiotics are lactobacilli and bifidobacteria. They can be given with fermented foods such as yoghurt, fermented vegetables or meats and they may briefly establish in the gut. A number of health-related effects are documented. Such established effects are listed in Table 8 and each effect has been shown in at least two human clinical studies by different research groups. Colonic fermentation has been shown to be altered following probiotic intake either as fermented milks or freeze-dried cultures. Japanese studies have shown that oral administration of certain lactic-acid bacteria increases the numbers of endogenous lactobacilli in faeces. Similarly, an increase in faecal bifidobacteria has been observed, and the numbers of clostridia have decreased (Hosoda et al. 1994; Benno et al. 1996). Recent studies by Alander et al. (1996) have indicated that the potential to colonize human colonic mucosa differs from that of faecal samples. Thus, more human studies are clearly needed to understand changes in both colonic microflora and colonic mucosal microflora during probiotic intake. 7.2.1. Alleviation of lactose intolerance symptoms. The majority of the world’s population have low levels of bgalactosidase in their small-bowel mucosa. Most lactasedeficient people are symptom-free if they consume only limited amounts of milk, except those subjects who are severely lactose intolerant. A beneficial effect of probiotics on lactose digestion has been demonstrated (Sanders, 1993). In particular, in studies comparing yoghurt and milk consumption it was shown that yoghurt consumption enhances lactose digestion in lactase-deficient subjects (Kolars et al. 1984; Marteau et al. 1990) and slows orocaecal transit. 7.2.2. Immune enhancement. An enhancement of the circulating IgA antibody secreting cell response was observed in infants supplemented with a strain of Lactobacillus casei, and was correlated with shortened duration of diarrhoea in the study group when compared with a placebo group (Kaila et al. 1992). Other studies reported an enhancement in the non-specific immune phagocytic activity of granulocyte populations in the blood of human volunteers after consumption of Lactobacillus acidophilus and Bifidobacterium bifidum (Schriffin et al. 1995; Marteau Table 8. Health-related effects of currently available probiotics (for detailed references, see Marteau & Rambaud, 1993; Lee & Salminen, 1995; Salminen et al. 1996a, b, c) Alleviation of symptoms of lactose intolerance Immune enhancement Shortening the duration of rotavirus diarrhoea Decreasing faecal mutagenicity Decreasing faecal bacterial enzyme activity Prevention of recurrence of superficial bladder cancer S163 et al. 1997a, b). Since phagocytic activity is involved with natural immunity and phagocytes are implicated in antibody immune responses as antigen-presenting cells, it is possible that stimulation of intestinal IgA antibody responses induced by lactic-acid bacteria may be explained partly by an effect on phagocyte cell functions. Ingestion of yoghurt has been reported to stimulate cytokine production, including interferon-g in human blood mononuclear cells (Solis-Pereyra & Lemonnier, 1996). 7.2.3. Acute gastroenteritis. Several studies, by different groups and in different conditions, have shown that some probiotic lactobacilli significantly shorten the duration of rotavirus diarrhoea in developed (Guarino et al. 1977) and developing countries (Raza et al. 1995; Pant et al. 1996). For one strain, a similar clinical effect on the duration of rotavirus diarrhoea without an increase in immune response has been reported (Shornikova et al. 1997). Bifidobacterium bifidum (now reclassified as Bifidobacterium animalis) has been reported to prevent rotavirus diarrhoea (Saavedra et al. 1994). 7.2.4. Faecal mutagenicity and enzymes. Lactic acid bacteria influence the mutagenicity of intestinal contents and the levels of faecal microbial enzymes, such as bglucuronidase, b-glucosidase, nitroreductase and urease (EC 3.5.1.5). Lactobacillus acidophilus has been shown to decrease faecal and urinary mutagenicity in healthy volunteers consuming fried minced beef. The same strain decreased faecal E. coli levels in colon-cancer patients and reduced faecal b-glucuronidase levels. Similar results have been reported for many probiotic lactobacilli strains (Goldin & Gorbach, 1984; Goldin et al. 1992; Ling et al. 1994; Morotomi, 1996). The prophylactic effects of oral administration of a lactobacillus strain on the recurrence of superficial bladder cancer have been reported in two Japanese studies (Aso & Akazan, 1992; Aso et al. 1995). 7.3. Diet and colon cancer The principal risk factors for sporadic colorectal cancer are high intakes of animal fat and meat, particularly red meat, excess energy intake, high intakes of refined cereals and of alcohol, particularly beer for colorectal carcinogenesis, and lack of physical exercise. A number of dietary factors have been proposed as potentially protective (Potter et al. 1993). 7.3.1. Dietary protective factors. A high intake of vegetables and fruit is associated with a decreased risk of colorectal cancer and adenomas throughout the world (Trock et al. 1990). Although vegetables such as the cruciferae may have a specific protective affect via antioxidant substances such as indoles, all fresh vegetables seem to be protective, in particular against cancer of the distal bowel. Components of fruit and vegetables that have been implicated as protective against colorectal cancer include NSP (dietary fibre), micronutrients (vitamins C, D and E, b-carotene, Ca and Se), glucosinolates, phenols, lignans, flavonoids and isoflavonoids. One of the most important protective components of vegetables and unrefined cereals is dietary fibre. The mechanism may be through a diluting effect of carcinogens, a reduction in transit time, fermentation and the production S164 S. Salminen et al. of SCFA and alterations in N and bile acid metabolism. Although results of individual case–control studies have been disappointing, a meta-analysis of thirteen of these studies (Howe et al. 1992) found a significant inverse association between fibre intake and risk of colorectal cancer with a relative risk close to 0·5 for consumers of over 27 g/d. When extrapolating these data to the American population, the authors estimated it to be possible to reduce the risk of colorectal cancer by 31 % i.e. 50 000 cases per year by increasing the daily consumption by 13 g. However, the multiplicity of methods for assessing dietary fibre intake in these studies makes any quantitative recommendations difficult. Ca and vitamin D are other potentially protective factors that could easily be used as preventive agents. A recent American cohort study summarized the relations between Ca, vitamin D and dairy products (Bostick et al. 1993). Of thirteen studies (nine case–control and four intervention studies), eight suggested an inverse relationship, which was statistically significant in five. Most studies that observed a protective effect of Ca came from Nordic or Anglo-Saxon communities, whereas no study demonstrated a protective effect of Ca or dairy products in Latin communities. It can, therefore, be suggested that the way Ca is consumed, and perhaps the general level of Ca consumption, is of importance. It has been suggested that fermented dairy products could be protective. In a French study (Boutron et al. 1996), a reduced risk of large adenomas was associated with a moderate intake of yoghurt. Other dietary protective factors which have been investigated in intervention studies of recurrence of adenomas are the so-called antioxidant vitamins. The results of these studies have so far been disappointing (Greenberg et al. 1994; MacLennan et al. 1995). 8. Safety issues (Donohue & Salminen, 1996; Roberfroid et al. 1997) Safety issues with new prebiotics and probiotics, including genetically modified bacteria, should be assessed according to European Union Novel Food Regulations on a case-bycase basis. Safety aspects of currently used prebiotics and probiotics are discussed in relation to their physiological properties. 8.1. Prebiotics (Van Loo et al. 1995) Although the European Union regulation on novel foods and novel food ingredients leaves room for interpretation, it is very likely that most prebiotics placed into the European market after this regulation comes into force will fall within its scope. They will, therefore, be subject to safety and nutritional evaluation on a case-by-case basis. Fructosyl-type ingredients, for which prebiotic properties are claimed, are already in the market. Fructosyl-type ingredients are natural components of a variety of fruits, vegetables and cereals, and consequently are consumed regularly (a few g/d) as part of the current diet (Van Loo et al. 1995). They are classified as natural food ingredients and cleared as novel food ingredients. Similarly galactosyl prebiotics have been cleared as novel food ingredients by the Health Department of The Netherlands (Staatscourant, 1996). 8.2. Probiotics (Adams & Marteau, 1995; Donohue & Salminen, 1996) The use of lactic acid-producing bacteria in foods has a long history and most strains are considered commensal micro-organisms with little or no pathogenic potential. Their ubiquitous presence in intestinal epithelium and the human gastrointestinal tract, and their traditional use in fermented foods and dairy products attest to their safety. Members of the genus Lactobacillus are most commonly given safe or generally recognized as safe (GRAS) status, whereas members of the genera Streptococcus and Enterococcus contain many opportunistic pathogens. Case reports from the literature of lactic acid-producing bacteria causing clinical infection in human subjects have recently been analysed in reviews by Gasser (1994) and by Aguirre & Collins (1993). Both reviews conclude that, considering their widespread consumption, lactic-acid bacteria appear to have a very low pathogenic potential. Two recent Finnish studies confirm that the number of infections associated with lactic-acid bacteria is extremely small and no case could be linked to commercial probiotics (Saxelin et al. 1996a,b). 9. Critical evaluation of present knowledge Each of the foregoing sections has examined in depth one aspect of the science of gastrointestinal physiology and function in relation to probiotics and prebiotics. The following conclusions are derived from the discussions within each section. 9.1. Intestinal microflora The intestinal microflora has been studied using traditional methods. A current problem is the presence of nonculturable species, which require new methodologies to be developed for their detection and measurement. The establishment of the normal human intestinal microflora, its components and metabolic activities, requires further study. Similarly, knowledge is needed on the composition and activities of the flora in different ethnic groups, at different ages and in different countries. Prebiotics and a few probiotics have already been shown to have the potential to modify significantly the composition of the intestinal microflora. In particular, the stimulatory effect of fructosyl prebiotics on the growth of bifidobacteria is well established. Moreover, prebiotics and probiotics are interesting tools with which to study the physiological consequences of changes in metabolic activities following such modification. 9.2. Mucosal function There is evidence for a strong interaction between the intestinal microflora, gut mucosa and GALT and the functions and dysfunctions of the gastrointestinal tract. It has been shown that probiotics may change the gut mucosal Gastrointestinal physiology and function barrier by stabilizing the intestinal mucosa, normalizing intestinal permeability and improving gut immunology. However, there are differences in the function and activity of different probiotics. Another consequence of intake of prebiotics and probiotics is the prevention of overgrowth of pathogenic bacteria and viruses. Taken together, these modifications have been shown to influence the gut barrier system. Future human studies should consider intestinal immunity and its modulation by resident probiotic bacteria or prebiotic components in more detail. Local release of cytokines induced by inflammatory reactions may amplify adverse reactions to food components within the intestinal tract and other parts of the body. 9.3. Gastrointestinal physiology Gut bacteria play a role in bowel functions like faecal mass, stool frequency, regulation of colonic pH, production of SCFA and salvage of energy from non-digestible food components. Any modification of the microflora is likely to influence these functions, but most of the results of human studies with pro- and prebiotics are preliminary and yet to be confirmed. In particular, the interaction of probiotics and prebiotics with the endocrine activity of the gut requires investigation. 9.4. Methodology There is clearly a need for new methodology to measure and characterize the composition of faecal microflora, in particular in large human nutrition studies. New biomarkers specifically involved with immune responses need to be defined and validated for nutritional studies. Germ-free animals (gnotobiotic mice) offer a model to clarify the specific effect of a bacterial strain or a food component on a given immune response. Development of early markers of carcinogenesis is urgently required for human intervention studies. 9.5. Human studies on health benefits 9.5.1. Prebiotics (Table 9). There are currently few reports of well-designed and well-documented human Table 9. Established and postulated functional effects of currently available prebiotics in human subjects Established functional effects Non-digestibility and low energy value (< 9 kJ/g) Stool bulking effect Modulation of the gut flora, promoting bifidobacteria and repressing clostridia IBS, irritable bowel syndrome. Postulated areas for future research Prevention of intestinal disorders (IBS, ulcerative colitis) and infections, including diarrhoea Modulation of immune response Prevention of colon carcinogenesis Reduction in serum levels of triacylglycerols and cholesterol Improved bioavailability of minerals (Ca, Mg) S165 Table 10. Postulated clinical effects of probiotics Areas for future research Research on mechanisms Regulation of intestinal motility. Modulation of intestinal and systemic immune responses. Reduction and protection of radiotherapy-associated intestinal dysfunction. Prevention of intestinal cancers Intestinal microflora effects, immunomodulation effects, competitive exclusion, cholesterol lowering intervention studies with prebiotics. Future studies may indicate differences between the effects of prebiotic components. Thus, it is important that each compound is tested separately or in products designed for a particular function. 9.5.2. Probiotics (Table 10). Alleviation of lactose maldigestion symptoms is well characterized for some probiotics. The ability of certain probiotic strains to shorten the duration of rotavirus diarrhoea has been established in several studies. Immunomodulation has been demonstrated for some probiotic strains and one strain has been documented to reduce the recurrence of superficial bladder cancer in human subjects. Novel methodologies are required to provide further data on the mechanisms of competitive exclusion and microflora modification. For other areas of interest there are still few welldesigned and well-conducted human intervention studies with probiotics. There are indications of significant differences between probiotic bacterial genera and species, and between strains of the same species. Thus, it is important that each strain is tested on its own or in products designed for a particular function. Further research is needed on the immune response, intestinal cancer and the mechanisms that underlie probiotic effects. 9.5.3. Diet and colon cancer. The major challenge at present is to demonstrate within intervention studies which specific dietary factors are able to decrease the risk of colon cancer. Recurrence of large adenomas, the step prior to cancer, is a promising tool for use in human studies and could be applied to prebiotics, probiotics and synbiotics using functional food products as test products. Colon cancer-related variables need to be carefully characterized in the diet and among the microflora. 9.6. Safety The human consumption of prebiotics and probiotics, at least lactic-acid bacteria, appears to be safe. 10. Recommendations for future research priorities Evidence has accumulated to support an important role for gastrointestinal function and the intestinal microflora in maintaining health and preventing diseases. Disturbances of the intestinal microflora may lead to other disturbances and dysfunctions of the gut. Thus, understanding the normal microflora with regard to its metabolic activity and S166 S. Salminen et al. influence on the immune and endocrine systems remains a key area for future research. 10.1. Intestinal microflora (1) Develop and validate robust methods that are applicable to large-scale human studies of the intestinal microflora. (2) Characterize the normal microflora and its activities in healthy persons of all ages. (3) Identify changes in microflora composition and activity associated with major dysfunctions of the gut. (4) Identify dietary factors that lead to changes in the microflora and the mechanisms that bring about improvement in health. 10.2. Short-chain fatty acids and intestinal microflora (1) The mechanism by which mixed populations of anaerobic gut bacteria produce different amounts and patterns of SCFA needs to be further investigated. (2) The role of butyrate in cell differentiation and growth requires further study, particularly of the butyrate response elements in genes and identification of the genes involved. (3) The role of acetate in metabolism and its regulation, particularly in fasting or starving subjects, requires further study. (4) In vivo production rates of SCFA and their relation to H2 metabolism and microbial growth need to be determined. 10.3. Diet and cancer (1) Determine the role of the intestinal microflora with respect to composition and activities in carcinogenesis, in particular of the colo-rectum. (2) Modulation of these aspects by probiotics and prebiotics requires further study. (3) Novel biomarkers of colorectal carcinogenesis need to be developed and validated. (4) The influence of diet and the intestinal microflora on DNA damage and repair in normal mucosa requires further study. 10.4. Immune system (1) Improve understanding of GALT function and regulation, and interaction between GALT and the digestive epithelium. (2) Understand the role of the intestinal microflora and its modification by dietary factors in regulating immune function in health and disease. (3) Develop and validate biomarkers for immune function and their long-term effects. 10.5. Gut mucosa (1) Develop novel methodologies to study the function and changes in the intestinal mucosa in human subjects. (2) Characterize the microflora associated with a healthy mucosa and mucus metabolism. (3) Determine the effects of changes in phase I and phase II mucosal enzymes on xenobiotic metabolism and health. References Adams MR & Marteau P (1995) On the safety of lactic acid bacteria from food. International Journal of Food Microbiology 27, 262–264. Aguirre M & Collins MD (1993) Lactic acid bacteria and human clinical infection. Journal of Applied Bacteriology 75, 95–107. Alander M, Korpela R, Saxelin M, Vilpponon-Salmila T, MattilaSandholm T & von Wright A (1996) Recovery of Lactobacillus rhamnosus GG from human colonic biopsies. Letters in Applied Microbiology 24, 361–364. Ansari B, Coates PJ, Greenstein B & Hall PA (1993) In situ end labelling (ISEL) detects DNA strand breaks in apoptosis and in other physiological and pathological states. Journal of Pathology 170, 1–8. Aso Y & Akazan H (1992) Prophylactic effect of Lactobacillus casei preparation on the recurrence of superficial bladder cancer. Urology International 49, 125–129. Aso Y, Akazan H, Kotake T, Tsukamoto T, Imai K & Naito S (1995) Preventive effect of a Lactobacillus casei preparation on the recurrence of superficial bladder cancer in a double blind trial. European Journal of Urology 27, 104–109. Awad A, Horvath PJ & Anderson MS (1991) Influence of butyrate on lipid metabolism, survival, and differentiation of colon cancer cells. Nutrition and Cancer 16, 125–133. Ballongue J, Schumann C & Quignon P (1997) Effects of lactulose and lactitol on colonic microflora and enzymatic activity. Scandinavian Journal of Gastroenterology 32, Suppl. 222, 41–44. Bedenne L, Faivre J, Boutron MC, Piard F, Cauvin JM & Hillon P (1992) Adenoma–carcinoma sequence or ‘de novo’ carcinogenesis? A study of adenomatous remnants in a population-based series of large bowel cancers. Cancer 69, 883–888. Benlounes N, Dupont C, Candahl C, Blaton MA, Desjeux JF & Heyman M (1996) The threshold for immune cell reactivity to milk antigens decreases in cow’s milk allergy with intestinal symptoms. Journal of Allergy and Clinical Immunology 98, 781–789. Benno Y, He F, Hosoda M, Hashimoto H, Kojima T, Yamazaki K, Iino H, Mykkänen H & Salminen S (1996) Effects of Lactobacillus GG yogurt on human intestinal microecology in Japanese subjects. Nutrition Today 31, 9–11. Biewenga J, van Rees EP & Sminia T (1993) Induction and regulation of IgA responses in the microenvironment of the gut. Short analytical review. Clinical Immunology and Immunopathology 67, 1–7. Binder HJ, Cummings JH & Soergel KH (editors) (1994) Falk Symposium 73. Short Chain Fatty Acids. Dordrecht, The Netherlands: Kluwer Academic. Boffa LC, Lupton JR, Mariani MR, Ceppi M, Newmark HL, Scalmati A & Lipkin M (1992) Modulation of colonic epithelial cell proliferation, histone acetylation, and luminal short chain fatty acids by variation of dietary fiber (wheat bran) in rats. Cancer Research 52, 5906–5912. Boland CR, Martin MA & Goldstein IJ (1992) Lectin reactivities as intermediate biomarkers in premalignant colorectal epithelium. Journal of Cellular Biochemistry 16G, 103–109. Bostick RM, Potter JD, McKenzie DR, Sellers TA, Kushi LH, Steinmetz KA & Folsom AR (1993) Reduced risk of colon cancer with high intake of vitamin E: the Iowa women’s health study. American Journal of Epidemiology 137, 1302–1317. Boutron MC & Faivre J (1993) Diet and the adenoma–carcinoma sequence. European Journal of Cancer Prevention 2, 95–98. Gastrointestinal physiology and function Boutron MC, Faivre J, Marteau P, Couillaut C, Senesse P & Quipourt V (1996) Calcium, phosphorus, vitamin D, dairy products and colorectal carcinogenesis: a French case–control study. British Journal of Cancer 74, 145–151. Brandtzaeg P (1995) Molecular and cellular aspects of the secretory immunoglobulin system. Acta Pathologica, Microbiologica et Immunologica Scandinavica 103, 1–19. Brandtzaeg P, Halstensen TS, Kett K, Krajci P, Kvale D, Rognum TO, Scott H & Sollid LM (1989) Immunobiology and immunopathology of human gut mucosa: humoral immunity and intraepithelial lymphocytes. Gastroenterology 97, 1562–1584. Bry L, Falk PG, Midtvedt T & Gordon J (1996) A model for host– microbial interactions in an open mammalian ecosystem. Science 273, 1380–1383. Bugaut M & Bentejac M (1993) Biological effects of shortchain fatty acids in nonruminant mammals. Annual Review of Nutrition 13, 217–241. Burkitt DP (1971) Epidemiology of cancer of the colon and rectum. Cancer 28, 3–13. Cheng BQ, Trimble RP, Illman RJ, Stone BA & Topping DL (1987) Comparative effects of dietary wheat bran and its morphological components (aleurone and pericarp seed coat) on volatile fatty acid concentrations in the rat. British Journal of Nutrition 57, 69–76. Clinton SK (1992) Dietary protein and carcinogenesis. In Nutrition, Toxicity and Cancer, pp. 455–479 [IR Rowland, editor]. Boca Raton, FL: CRC Press. Collins MD & Gibson GR (1998) Nutritional modulation of microbial ecology. American Journal of Clinical Nutrition (In the Press). Coussement P (1996) Pre- and synbiotics with inulin and oligofructose. Food Technology in Europe, December 1995/January 1996 issue, 102–104. Cummings JH (1993) The effect of dietary fiber on faecal weight and composition. In CRC Handbook of Dietary Fiber in Human Nutrition, pp. 263–349 [GA Spiller, editor]. Boca Raton, FL: CRC Press. Cummings JH (1994) Non-starch polysaccharides (dietary fibre) including bulk laxatives in constipation. In Constipation, pp. 307–314 [MA Kamm and JE Lennard-Jones, editors]. Petersfield: Wrightson Biomedical Publishing Ltd. Cummings JH (1995) Short chain fatty acids. In Human Colonic Bacteria: Role in Nutrition, Physiology and Pathology, pp. 101–130 [GR Gibson and GT Macfarlane, editors]. Boca Raton, FL: CRC Press. Cummings JH (1997) The Large Intestine in Nutrition and Disease. Danone Chair Monograph, pp. 1–155. Brussels: Institut Danone. Cummings JH, Bingham SA, Heaton KW & Eastwood MA (1992) Faecal weight, colon cancer risk and dietary intake of non-starch polysaccharides (dietary fiber). Gastroenterology 103, 1783–1789. Cummings JH & Englyst HN (1987) Fermentation in the large intestine and the available substrates. American Journal of Clinical Nutrition 45, 1243–1255. Cummings JH, Gibson GR & Macfarlane GT (1989) Qualitative estimates of fermentation in the hind gut of man. Acta Veterinaria Scandanavica Suppl. 86, 76–82. Cummings JH, Pomare EW, Branch WJ, Naylor CPE & Macfarlane GT (1987) Short chain fatty acids in human large intestine, portal, hepatic and venous blood. Gut 28, 1221–1227. Cummings JH, Roberfroid MB, Andersson H, Barth C, FerroLuzzi A, Ghoos Y, Gibney M, Hermonsen K, James WPT, Korver O, Lairon D, Pascal G & Voragen AGS (1997) A new look at dietary carbohydrate: chemistry, physiology and health. European Journal of Clinical Nutrition 51, 417–423. S167 Cummings JH, Rombeau JL & Sakata T (editors) (1995) Physiological and Clinical Aspects of Short Chain Fatty Acids. Cambridge: Cambridge University Press. De Simone C, Vesely R, Bianchi Salvadori B & Jirillo E (1993) The role of probiotics in modulation of the immune system in man and animals. International Journal of Immunotherapy 9, 23–28. Donohue DC & Salminen S (1996) Safety of probiotic bacteria. Asia Pacific Journal of Clinical Nutrition 5, 25–28. Duchmann R, Kaiser I, Hermann E, Mayet W, Ewe K & Meyer zum Büschenfelde KH (1995) Tolerance exists towards resident intestinal flora but is broken in active inflammatory bowel disease (IBD). Clinical and Experimental Immunology 102, 448–455. Ducluzeau R (1993) Installation, équilibre et rôle de la flore microbienne du nouveau-né (Development, balance and role of the microbial flora of the newborn). Annales Pédiatriques (Paris) 40, 13–22. Faivre J, Boutron MC, Hillon P, Bedenne L & Klepping C (1985) Epidemiology of colorectal cancer. In Diet and Human Carcinogenesis, pp. 123–135 [JV Joosens, MJ Hill and J Geboers, editors]. Amsterdam: Elsevier Scientific Publishers. Fargeas MJ, Theodorou V, More J, Wal JM, Fioramonti J & Bueno L (1995) Boosted systemic immune and local responsiveness after intestinal inflammation in orally sensitized guinea pigs. Gastroenterology 109, 53–62. Frankel WL, Zhang W, Singh A, Klurfeld DM, Don S, Sakata T, Modlin I & Rombeau JL (1994) Mediation of the trophic effects of short-chain fatty acids on the rat jejunum and colon. Gastroenterology 106, 375–380. Fuller R (1991) Probiotics in human medicine. Gut 32, 439–442. Fuller R (editor) (1992) Probiotics. The Scientific Basis. London: Chapman and Hall. Gaboriau-Routhiau V & Moreau MC (1996) Gut flora allows recovery of oral tolerance to ovalbumin in mice after transient breakdown mediated by cholera toxin or Escherichia coli heat-labile enterotoxin. Pediatrics Research 39, 625–629. Gasser F (1994) Safety of lactic acid bacteria and their occurrence in human clinical infections. Bulletin Institut Pasteur 92, 45–67. Gibson GR, Beatty ER, Wang X & Cummings JH (1995) Selective stimulation of bifidobacteria in the human colon by oligofructose and inulin. Gastroenterology 108, 975–982. Gibson GR & Macfarlane GT (editors) (1995) Human Colonic Bacteria: Role in Nutrition, Physiology and Pathology. Boca Raton, FL: CRC Press. Gibson GR & Roberfroid MB (1995) Dietary modulation of the human colonic microbiota: introducing the concept of prebiotics. Journal of Nutrition 125, 1401–1412. Gibson PR, Moeller I, Kagelari O, Folino M & Young GP (1992) Contrasting effects of butyrate on the expression of phenotypic markers of differentiation in neoplastic and non-neoplastic colonic epithelial cells in vitro. Journal of Gastroenterology and Hepatology 7, 165–172. Goldin BR & Gorbach SL (1984) The effect of milk and lactobacillus feeding on human intestinal bacterial enzyme activity. American Journal of Clinical Nutrition 39, 756–761. Goldin BR, Gorbach SL, Saxelin M, Barakat S, Gualteri L & Salminen S (1992) Survival of Lactobacillus species (strain GG) in the human gastrointestinal tract. Digestive Diseases and Sciences 47, 573–575. Goodlad RA (1989) Gastrointestinal cell proliferation. Digestive Diseases 7, 169–177. Goodlad RA & Wright NA (1982) Quantitative studies on epithelial replacement in the gut. Techniques in the Life Sciences P212, 1–23. Gracey M (1993) Transmission and epidemiology. Bailliere’s Clinical Gastroenterology 7, 195–214. S168 S. Salminen et al. Greenberg ER, Baron JA, Tosteson TD, Freeman DH Jr, Beck GJ, Bond JH, Colacchio TA, Coller JA, Frankl HD & Haile RW (1994) A clinical trial of antioxidant vitamins to prevent colorectal adenoma. New England Journal of Medicine 331, 141–147. Guarino A, Canani RB, Spagnuolo MI, Albano F & Benedetto LD (1997) Oral bacterial therapy reduces the duration of symptoms and of viral excretion in children with mild diarrhoea. Journal of Pediatric Gastroenterology and Nutrition (In the Press). Hague A, Manning AM, Hanlon KA, Huschtscha LI, Hart D & Paraskeva C (1993) Sodium butyrate induces apoptosis in human colonic tumour cell lines in a p53-independent pathway: implications for the possible role of dietary fibre in the prevention of large-bowel cancer. International Journal of Cancer 55, 498–505. Hambly RJ, Rumney CJ, Fletcher JME, Rijken P & Rowland IR (1997) Effects of high and low risk diets on gut-microfloraassociated biomarkers of colon cancer in human-flora-associated rats. Nutrition and Cancer 27, 250–255. Havenaar R & Huis in’t Veld JHJ (1992) Probiotics; a general view. In The Lactic Acid Bacteria. Vol. 1. The Lactic Acid Bacteria in Health and Disease, pp. 151–170 [BJB Wood, editor]. Barking: Elsevier Applied Science. Heyman M & Desjeux JF (1992) Significance of intestinal food protein transport. Journal of Pediatric Gastroenterology and Nutrition 15, 48–57. Heyman M, Ducroc R, Desjeux JF & Morgat JL (1982) Horseradish peroxidase transport across adult rabbit jejunum in vitro. American Journal of Physiology 242, G558–G564. Heyman M, Gorthier G, Petit A, Meslin J-C, Moreau C & Desjeux J-F (1987) Intestinal absorption of macromolecules during viral enteritis: an experimental study on rotavirusinfected conventional and germ-free mice. Pediatrics Research 22, 72–78. Hosoda M, Ile F, Hiramaru M, Hashimoto H & Benno Y (1994) Effects of Lactobacillus GG intake on faecal microflora and defacation in healthy volunteers. Bifidus 8, 21–28. Howe G, Benito E, Casteleto R, Cornée J, Estève J, Gallagher RP, Iscovich JM, Deng-ao J, Kaaks R, Kune GA, Kune S, L’Abbé KA, Lee HP, Miller AB, Peters RK, Potter JD, Riboli E, Slattery ML, Trichopoulos D, Tuyns A, Tzonou A, Whittemore AS, Wu-Williams AH & Zeng S (1992) Dietary intake of fiber and decreased risk of cancers of the colon and rectum: evidence from the combined analysis of 13 case–control studies. Journal of the National Cancer Institute 84, 1887–1896. Isolauri E (1995) The treatment of cow’s milk allergy. European Journal of Clinical Nutrition 49, 549–555. Isolauri E, Kaila M, Arvola T, Majamaa H, Rantala I, Virtanen E & Arvilommi H (1993a) Diet during rotavirus enteritis affects jejunal permeability to macromolecules in suckling rats. Pediatrics Research 33, 548–553. Isolauri E, Majamaa H, Arvola T, Rantala I, Virtanen F & Arvilommi H (1993b) Lactobacillus casei strain GG reverses increased intestinal permeability induced by cow milk in suckling rats. Gastroenterology 105, 1643–1650. Ito M, Kimura M, Deguchi Y, Miyamori-Watabe A, Yajima T & Kan T (1993) Effects of transgalactosylated disaccharides on the human intestinal microflora and their metabolism. Journal of Nutritional Science and Vitaminology 39, 279–288. Jalonen T, Isolauri E, Heyman M, Crain-Denoyelle A-M, Sillanaukee P & Koivula T (1991) Increased b-lactoglobulin absorption during rotavirus enteritis in infants: relationship to sugar permeability. Pediatrics Research 30, 290–293. Kaila M, Isolauri E, Soppi E, Virtanen V, Laine S & Arvilommi H (1992) Enhancement of the circulating antibody secreting cell response in human diarrhea by a human Lactobacillus strain. Pediatrics Research 32, 141–144. Kim YS, Gum JR, Ho SB & Deng G (1994) Colonocyte differentiation and proliferation: overview and the butyrateinduced transcriptional regulation of oncodevelopmental placental-like alkaline phosphatase gene in colon cancer cells. In Falk Symposium 73. Short Chain Fatty Acids, pp. 119–134 [HJ Binder, JH Cummings and KH Soergel, editors]. Lancaster: Kluwer Academic Publishers. Kim YS, Tsao D, Siddiqui B, Whitehead JS, Arnstein P, Bennett J & Hicks J (1980) Effects of sodium butyrate and dimethylsulfoxide on biochemical properties of human colon cancer cells. Cancer 45, 1185–1192. Kok RG, De Waal A, Schut F, Welling GW, Weenk G & Hellingwerf KJ (1996) Specific detection and analysis of a probiotic Bifidobacterium strain in infant faeces. Applied Environmental Microbiology 62, 3668–3672. Kolars JC, Levitt MD, Aouiji M & Savaiano DA (1984) Yogurt – an autodigesting source of lactose. New England Journal of Medicine 310, 1–3. Langendijk PS, Schut F, Jansen GJ, Raangs GC, Kamphuis GR, Wilkinson MHF & Welling GW (1995) Quantitative fluorescence in situ hybridization of Bifidobacterium spp. with genus-specific 16S rRNA-targeted probes and its application in faecal samples. Applied Environmental Microbiology 61, 3069–3075. Lee YK & Salminen S (1995) The coming of age of probiotics. Trends in Food Science Technology 6, 241–245. Lifschitz CH, Carrazza FR, Feste AS & Klein PD (1995) In vivo study of colonic fermentation of carbohydrate in infants. Journal of Paediatric Gastroenterology and Nutrition 20, 59–64. Ling WH, Korpela R, Mykkänen H & Salminen S (1994) Lactobacillus strain GG and fiber supplementation decreases colonic hydrolytic and reductive enzyme activities in healthy female subjects. Journal of Nutrition 121, 18–23. Lipkin M (1988) Biomarkers of increased susceptibility to gastrointestinal cancer: new application to studies of cancer prevention in human subjects. Cancer Research 43, 1899–1904. Livesey G (1990) Energy values of unavailable carbohydrates and diets: an inquiry and analysis. American Journal of Clinical Nutrition 51, 617–637. McCartney AL, Wenzhi W & Tannock GT (1996) Molecular analysis of the composition of the bifidobacterial and lactobacillus microflora of humans. Applied Environmental Microbiology 62, 4608–4613. MacDonald TT (1993) Aetiology of Crohn’s disease. Archives of Disease in Childhood 68, 623–625. Macfarlane GT, Gibson GR, Drasar BS & Cummings JH (1995) Metabolic significance of the gut microflora. In Gastrointestinal and Oesophageal Pathology, 2nd ed., pp. 249–274 [R Whitehead, editor]. Edinburgh: Churchill Livingstone. Macfarlane GT & Macfarlane S (1995) Proteolysis and amino acid fermentation. In Human Colonic Bacteria: Role in Nutrition, Physiology, and Pathology, pp. 75–100 [GR Gibson and GT Macfarlane, editors]. London: CRC Press. McIntyre A, Gibson PR & Young GP (1993) Butyrate production from dietary fibre and protection against large bowel cancer in a rat model. Gut 34, 386–391. McIntyre A, Young GP, Taranto T, Gibson PR & Ward PB (1991) Different fibres have different regional effects on luminal contents of rat colon. Gastroenterology 101, 1274–1281. McKay DM & Perdue MH (1993) Intestinal epithelial function: the case for immunophysiological regulation. Cells and mediators (First of two parts). Digestive Diseases and Sciences 38, 1377–1387. MacLennan R, Macrae F, Bain C, Battistutta D, Chapuis P, Gratten H, Lambert J, Newland RC, Ngu M & Russell A (1995) Randomized trial of intake of fat, fiber, and beta carotene to prevent colorectal adenomas. Journal of the National Cancer Institute 87, 1760–1766. Gastrointestinal physiology and function Marteau P, Flourié B, Pochart P, Chastang C, Desjeux JF & Rambaud JC (1990) Role of the microbial lactase (EC 3.2.1.23) activity from yoghurt on the intestinal absorption of lactose: an in vivo study in lactase-deficient humans. British Journal of Nutrition 64, 71–79. Marteau P, Minekus M, Havenaar R & Huis in’t Veld JHJ (1997a) Survival of lactic acid bacteria in a dynamic model of the stomach and small intestine: validation and the effects of bile. Journal of Dairy Science 80, 1031–1037. Marteau P & Rambaud JC (1993) Potential for using lactic acid bacteria for therapy and immunomodulation in man. FEMS Microbiology Review 12, 207–220. Marteau P, Vaerman JP, Dehennin JP, Bord S, Brassart D, Pochart P, Desjeux JF & Rambaud JC (1997b) Effects of intrajejunal perfusion and chronic ingestion of Lactobacillus johnsonii strain La1 on serum concentrations and jejunal secretions of immunoglobulins and serum proteins in healthy humans. Gastroenterology and Clinical Biology 21, 293–298. Martin MS, Justrabo E, Jeannin JF, Leclerc A & Martin F (1981) Effect of dietary chenodeoxycholic acid on intestinal carcinogenesis induced by 1,2 dimethylhydrazine in mice and hamsters. British Journal of Cancer 43, 884–886. Maskens AP (1976) Histogenesis and growth pattern of 1,2dimethylhydrazine-induced rat colon adenocarcinoma. Cancer Research 36, 1585–1592. Maskens AP & Dujardin Loits RM (1981) Experimental adenomas and carcinomas of the large intestine behave as distinct entities. Cancer 47, 81–89. Molis C, Flourié B, Ouarne F, Gailing MF, Lartigue S, Guibert A, Bornet F & Galmiche JP (1996) Digestion, excretion and energy value of fructo oligosaccharides in healthy humans. American Journal of Clinical Nutrition 64, 324–328. Moreau MC, Bisetti N & Dubuquoy C (1998) Immunomodulating properties of a strain of Bifidobacterium used as probiotic on the faecal and cellular intestinal IgA anti-rotavirus responses in mice. In Functional Foods Proceedings. Royal Society of Chemistry. Moreau MC & Coste M (1993) Immune responses to dietary protein antigens. World Review of Nutrition and Dietetics 74, 22–57. Moreau MC & Gaboriau-Routhiau V (1996) The absence of gut flora, the doses of antigen ingested and aging affect the longterm peripheral tolerance induced by ovalbumin feeding in mice. Research Immunology 147, 49–59. Moreau MC, Nuyts V & Raibaud P (1994) Effet de l’ingestion d’un lait fermenté avec Lactobacillus bulgaricus et Streptococcus thermophilus sur la stimulation de l’immunité de l’hôte: étude chez la souris axénique (Effect of the ingestion of a fermented milk on the stimulation of immunity in germ-free mice). Cahiers Decarboxylation Nutrition et Diététique 29, 341–347. Morotomi M (1996) Properties of Lactobacillus casei Shirota strain as probiotics. Asia Pacific Journal of Clinical Nutrition 5, 29–30. Mosmann T (1983) Rapid colorimetric assay for cellular growth and survival: application to proliferative and cytotoxic assays. Journal of Immunology Methods 65, 55–63. Muyzer G, de Waal EC & Uitterlinden AG (1993) Profiling complex microbial populations by denaturing gradient gel electrophoresis analysis of polymerase chain reaction-amplified genes coding for 16S rRNA. Applied Environmental Microbiology 59, 695–700. Pace N (1996) New perspective on the natural microbial world: molecular microbial ecology. ASM News 62, 463–470. Pant AR, Graham SM, Allen SJ, Harikul S, Sabohaeron A, Cuevas L & Hart CA (1996) Lactobacillus GG and acute diarrhoea in young children in the tropics. Journal of Tropical Pediatrics 42, 162–165. S169 Perdigon G, Alvarez S, Rachid M, Agüero G & Gobatto N (1996) Probiotic bacteria for humans: clinical systems for evaluation of effectiveness. Journal of Dairy Science 78, 1597–1606. Pitcher MCL & Cummings JH (1995) Colonic fermentation, sulphur metabolism and ulcerative colitis. In Falk Symposium 85. Inflammatory Bowel Disease, pp. 529–543 [GNJ Tytgat, JFWM Bartelsman and SHJ van Deventer, editors]. London: Kluwer Academic Publishers. Pitcher MCL & Cummings JH (1996) Hydrogen sulphide: a bacterial toxin in ulcerative colitis? Gut 39, 1–4. Podolsky DK (1991) Inflammatory bowel disease. (First of two parts.) New England Journal of Medicine 325, 928–937. Pomare EW, Branch WJ & Cummings JH (1985) Carbohydrate fermentation in the human colon and its relation to acetate concentration in venous blood. Journal of Clinical Investigation 75, 1448–1454. Potter JD, Slattery ML, Bostock RM & Gapstur SM (1993) Colon cancer: a review of the epidemiology. Epidemiology Review 15, 499–545. Rafter JJ, Child P, Anderson AM, Alder R, Eng V & Bruce WR (1987) Cellular toxicity of faecal water depends on diet. American Journal of Clinical Nutrition 45, 559–563. Raza S, Graham SM, Allen SJ, Sultana S, Cuevas L & Hart CA (1995) Lactobacillus GG promotes recovery from acute nonbloody diarrhoea in Pakistan. Pediatric Infection and Disease 14, 107–111. Risio M (1992) Cell proliferation in colorectal tumour progression: an immunohistochemical approach to intermediate biomarkers. Journal of Cellular Biochemistry 16G, 79–87. Ritchie JK, Wadsworth J, Lennard-Jones JE & Rogers E (1987) Controlled multicentre therapeutic trial of an unrefined carbohydrate, fibre-rich diet in Crohn’s disease. British Medical Journal 295, 517–520. Roberfroid MB (1996) Functional effects of food components and the gastrointestinal system: chicory fructooligosaccharides. Nutrition Reviews 54, S38–S42. Roberfroid MB, Bornet F, Bouley C & Cummings JH (1995) Colonic microflora: nutrition and health. Nutrition Reviews 53, 127–130. Roberfroid M, Gibson GR & Delzenne N (1993) The biochemistry of oligofructose, a non-digestible fiber: an approach to calculate its caloric value. Nutrition Reviews 51, 137–146. Roberfroid MB, Van Loo JAE & Gibson GR (1997) The bifidogenic nature of chicory inulin and its hydrolysis products. Journal of Nutrition (In the Press). Roediger WEW, Duncan A, Kapaniris O & Millard S (1993) Reducing sulfur compounds of the colon impair colonocyte nutrition: implications for ulcerative colitis. Gastroenterology 104, 802–809. Roland N, Nugon-Baudon L, Andrieux C & Szylit O (1995) Comparative study of the fermentative characteristics of inulin and different types of fibre in rats inoculated with a human whole faecal flora. British Journal of Nutrition 74, 239–249. Roncucci L (1992) Early events in human colorectal carcinogenesis. Aberrant crypts and microadenoma. Italian Journal of Gastroenterology 24, 498–501. Rowland I (1991) Nutrition and gut flora metabolism. In Nutrition, Toxicity and Cancer, pp. 113–135 [IR Rowland, editor]. Boca Raton, FL: CRC Press. Rowland IR (1995) Toxicology of the colon: role of the intestinal microflora. In Human Colonic Bacteria: Role in Nutrition, Physiology and Pathology, pp. 155–174 [GR Gibson and GT Macfarlane, editors]. London: CRC Press. Saavedra JM, Bauman NA, Oung I, Perman JA & Yolken RH (1994) Feeding of Bifidobacterium bifidum and Streptococcus thermophilus to infants in hospital for prevention S170 S. Salminen et al. of diarrhoea and shedding of rotavirus. Lancet 344, 1046– 1049. Sakamoto J, Shigeyuki N, Sugawara K, Iwane S & Munakata A (1996) Comparison of resistant starch with cellulose diet on 1,2-dimethylhydrazine-induced colonic carcinogenesis in rats. Gastroenterology 110, 116–120. Sakata T (1987) Stimulatory effect of short chain fatty acids on epithelial cell proliferation in the rat intestine: a possible explanation for trophic effects of fermentable fibre, gut microbes and luminal trophic factors. British Journal of Nutrition 58, 95–103. Salminen S, Isolauri E & Salminen E (1996a) Clinical uses of probiotics for stabilizing the gut mucosal barrier: successful strains and future challenges. Anton. Leeuwenhoek 70, 347–358. Salminen S, Isolauri E & Salminen E (1996b) Probiotics and stabilisation of the gut mucosal barrier. Asia Pacific Journal of Clinical Nutrition 5, 53–56. Salminen S, Laine M, von Wright A, Vuopio-Varkila J, Korhonen T & Mattila-Sandholm T (1996c) Development of selection criteria for probiotic strains to assess their potential in functional foods: a Nordic and European approach. Bioscience Microflora 15, 61–67. Sanders ME (1993) Summary of the conclusions from a consensus panel of experts on health attributes on lactic cultures: significance to fluid milk products containing cultures. Journal of Dairy Science 76, 1819–1828. Sanderson IR & Walker WA (1993) Uptake and transport of macromolecules by the intestine: possible role in clinical disorders (an update). Gastroenterology 104, 622–639. Savarino SJ & Bourgeois AL (1993) Diarrhoeal disease: current concepts and future challenges. Epidemiology of diarrhoeal diseases in developed countries. Transcript of the Royal Society of Tropical Medicine and Hygiene 87, 7–11. Saxelin M, Chuang NH, Chassy B, Rautelin H, Mäkelä PH, Salminen S & Gorbach SL (1996a) Lactobacilli and bacteremia in Southern Finland, 1989–1992. Clinical Infection and Disease 22, 564–566. Saxelin M, Rautelin H, Salminen S & Mäkelä P (1996b) The safety of commercial products with viable Lactobacillus strains. Infection and Disease in Clinical Practice 5, 331–335. Schiffrin E, Rochat F, Link-Amster H, Aeschlimann J & DonnetHugues A (1995) Immunomodulation of blood cells following the ingestion of lactic acid bacteria. Journal of Dairy Science 78, 491–497. Shornikova AV, Casas I, Isolauri E, Mykkänen H & Vesikari T (1997) Lactobacillus reuteri as a therapeutic agent in acute diarrhea in young children. Journal of Pediatric Gastroenterology and Nutrition 24, 399–404. Solis-Pereyra B & Lemonnier D (1996) Induction of human cytokines by bacteria used in dairy foods. Nutrition Research 13, 1127–1140. Staatscourant (1996) Toelating nieuw voedingsmiddel. Galactooligosaccharides (Acceptance of New Nutrient. Galactooligosaccharides) no. 91, p. 7. Stephen A (1994) Propionate – sources and effect on lipid metabolism. In Falk Symposium 73. Short Chain Fatty Acids, pp. 260–271 [HJ Binder, JH Cummings and KH Soergel, editors]. Dordrecht, The Netherlands: Kluwer Academic Publishers. Sütas Y, Hurme M & Isolauri E (1996a) Downregulation of antiCD3 antibody-induced IL-4 production by bovine caseins hydrolysed with Lactobacillus GG-derived enzymes. Scandinavian Journal of Immunology 43, 687–689. Sütas Y, Soppi E, Korhonen H, Syväoja EL, Saxelin M, Rokka T & Isolauri E (1996b) Suppression of lymphocyte proliferation in vitro by bovine caseins hydrolysed with Lactobacillus GGderived enzymes. Journal of Allergy and Clinical Immunology 98, 216–224. Szarka CE, Pfeiffer GR & Hum ST (1995) Glutathione-Stransferase activity and glutathione S-transferase m expression in subjects with risk for colorectal cancer. Cancer Research 55, 2789–2793. Terada A, Hara H, Kataoka M & Mitsuoka T (1992) Effect of lactulose on the composition and metabolic activity of the human faecal flora. Microbial Ecology in Health and Disease 5, 43–50. Thompson WG & Gomborone JE (1993) Irritable bowel syndrome: pathogenesis and management. Lancet 341, 1569–1572. Thompson WG & Heaton KW (1980) Functional bowel disorders in apparently healthy people. American Gastroenterology Association 79, 283–288. Todesco T, Rao VA, Bosello O & Jenkins DJA (1991) Propionate lowers blood glucose and alters lipid metabolism in healthy subjects. American Journal of Clinical Nutrition 54, 860–865. Trock B, Lanza E & Greenwald P (1990) Dietary fiber, vegetables, and colon cancer: critical review and meta-analyses of the epidemiologic evidence. Journal of the National Cancer Institute 82, 650–661. Tytgat GNJ, Bartelsman JFWM & van Deventer SJH (editors) (1995) Falk Symposium 85. Inflammatory Bowel Diseases. London: Kluwer Academic Publishers. Umesaki Y (1989) Intestinal glycolipids and their role in intestinal colonization of mice. Bifidobacteria Microflora 8, 13–22. Umesaki Y, Okada Y, Imaoka A, Setoyama H & Matsumoto S (1997) Interactions between epithelial cells and bacteria, normal and pathogenic. Science 276, 964–965. Umesaki Y, Okada Y, Matsumoto S, Imaoka A & Setoyama H (1995) Segmented filamentous bacteria are indigenous intestinal bacteria that activate intraepithelial lymphocytes and induce MHC class II molecules and fucosyl asialoGM1 glycolipids on the small intestinal epithelial cells in ex-germ-free mouse. Microbial Immunology 39, 555–562. Van Loo J, Coussement P, De Leenheer L, Hoebregs H & Smits G (1995) On the presence of inulin and oligofructose as natural ingredients in the Western diet. Critical Reviews in Food Science and Nutrition 35, 525–552. van Munster IP, Tangerman A & Nagengast FM (1994) Effect of resistant starch on colonic fermentation, bile acid metabolism, and mucosal proliferation. Digestive Diseases and Sciences 39, 834–842. Venter CS, Vorster HH & Cummings JH (1989) Effects of dietary propionate on carbohydrate and lipid metabolism in man. American Journal of Gastroenterology 85, 549–553. Venturi M, Hambly RJ, Glinghammar B, Rafter JJ & Rowland IR (1997) Genotoxic activity in human faecal water and the role of bile acids: a study using the Comet assay. Carcinogenesis (In the Press). Wasan HS & Goodlad RA (1996) Fibre-supplemented foods may damage your health. Lancet 348, 319–320. Wattenberg LW (1983) Inhibition of neoplasia by minor dietary constituents. Cancer Research 43, Suppl., 2448s–2453s. Weiner HL, Friedman A, Miller A, Khoury SJ, Al-Sabbagh A, Santos L, Sayegh M, Nussenblatt RB, Trentham DE & Hafler DA (1994) Oral tolerance: immunologic mechanisms and treatment of animal and human organ-specific autoimmune diseases by oral administration of autoantigens. Annual Review of Immunology 12, 809–837. Weisburger J (1973) Chemical carcinogens and their mode of action in colonic neoplasia. Disease of the Colon and Rectum 289, 1099–1103. Gastrointestinal physiology and function Whitehead RH, Young GP & Bathal PS (1986) Effects of short chain fatty acids on a new human colon carcinoma cell line (LIM 1215). Gut 27, 1457–1463. Wilson KH & Blitchington RB (1996) Human colonic biota studied by ribosomal DNA sequence analysis. Applied Environmental Microbiology 62, 2273–2278. S171 Wright NA & Alison M (1984) The Biology of Epithelial Cell Populations. Oxford: Oxford University Press. Young GP, McIntyre A, Albert V, Folino M, Muir JG & Gibson PR (1996) Wheat bran suppresses potato-starch potentiated colorectal tumorigenesis at the aberrant crypt stage in a rat model. Gastroenterology 110, 508–514. q Nutrition Society 1998 British Journal of Nutrition (1998), 80, Suppl. 1, S173–S193 S173 Functional food science and behaviour and psychological functions F. Bellisle1¬ , J. E. Blundell2 , L. Dye2 , M. Fantino3 , E. Fern4 , R. J. Fletcher5 , J. Lambert6 , M. Roberfroid7 , S. Specter1 , J. Westenhöfer8 and M. S. Westerterp-Plantenga9 1 INSERM – Unité 341 et Service de Nutrition, Hôtel-Dieu, 1, Place du Parvis Notre-Dame, F-75181, Paris Cedex 04, France 2 PsychoBiology Group, Department of Psychology, University of Leeds, Leeds LS2 9JT, UK 3 Laboratoire de Physiologie, Faculté de Médecine, Université de Bourgogne, 7, Boulevard Jeanne d’Arc, F-21033 Dijon Cedex, France 4 Nestec Ltd, Nestlé, 55, Avenue Nestlé, CH-1800 Vevey, Switzerland 5 Kellogg Company of Great Britain Ltd, The Kellogg Building, Talbot Road, Manchester M16 0PU, UK 6 Mars Confectionery, Division of Mars UK Ltd, Dundee Road, Slough SL1 4JX, UK 7 UCL, Ecole de Pharmacie, Tour Van Helmont, Avenue E. Mounier, 73, B-1200 Brussels, Belgium 8 Department of Nutrition and Home Economics, Advanced Technical College, Lohbrügger Kirchstr. 65, D-21033 Hamburg, Germany 9 Department of Sciences, Open University of the Netherlands/Maastricht University, Postbus 2960, NL-6401 DL Heerlen, The Netherlands Contents 1. 2. 3. 4. Overview General methodological considerations 2.1. Ensuring the reliability of food intake data Functions of macronutrients in relation to appetite 3.1. Overview 3.1.1. Ethnic, educational and social determinants 3.1.2. Circadian rhythms 3.2. Effects of macronutrient intake on energy intake, hunger, satiety and metabolism 3.2.1. Fat intake in relation to energy balance 3.2.2. Fat and carbohydrate intake in relation to satiety 3.2.3. Relative levels of fat, carbohydrate and protein intake in relation to satiety and metabolism 3.2.4. Alcohol intake and satiety 3.3. Functional properties of macronutrient substitutes 3.3.1. The preload paradigm 3.3.2. The issue of compensation 3.3.3. Highlight: sucrose polyester 3.4. Modulating food intake through palatability enhancers: the case of monosodium glutamate Foods and cognitive performance 4.1. Overview 4.2. Performance tasks 4.3. Effects of nutritional manipulations on performance 4.3.1. Food v. no food first thing in the morning 4.3.2. Effects at midday S174 S174 S175 S175 S175 S175 S175 S175 S175 S176 S176 S176 S177 S177 S177 S177 S178 S178 S178 S178 S179 S179 S179 4.3.3. Effects following evening meals Influence of macronutrient ratios and individual nutrients 4.4.1. The carbohydrate–tryptophan connection 4.4.2. Dietary fat 4.4.3. Glucose 4.4.4. Dietary choline 4.4.5. Vitamins and minerals 4.4.6. Alcohol 4.5. The case of caffeine 5. Functional effects of foods on activation or sedation 5.1. Overview 5.2. Methods for assessing activation or sedation 5.2.1. Tryptophan 5.2.2. Tryptophan, tyrosine, and ‘jet lag’ 5.3. Carbohydrate v. protein: effects on performance 5.4. Caffeine-driven activation S180 4.4. 6. Influence of food and individual nutrients on affective state 6.1. Overview 6.2. Diet-related dysphoria 6.3. Central neurochemical activity and mood state 6.4. Other links between nutrition, mood and brain serotonin levels 7. Nutrition and endorphins 8. Safety issues 8.1. Overview 8.2. Tryptophan 8.3. Caffeine 8.4. Macronutrient substitutes and flavour enhancers Abbreviations: DHA, docosahexaenoic acid; MSG, monosodium glutamate; SPE, sucrose polyester. *Corresponding author: Dr F. Bellisle, fax +33 1 40 46 82 48, email [email protected] S180 S180 S181 S181 S182 S182 S182 S182 S182 S182 S182 S182 S183 S183 S183 S183 S183 S183 S184 S184 S184 S185 S185 S185 S185 S185 S174 F. Bellisle et al. 9. Hyperactivity 9.1. Overview 9.2. Hyperactivity and sugar 9.3. Hyperactivity and food additives 10. Critical assessment of the science base: identification of criteria 11. Considerations for future research S185 S185 S185 S186 S186 S187 11.1. 11.2. 11.3. 11.4. 11.5. 11.6. 11.7. Appetite and satiety Macronutrient replacers Palatability enhancers Cognitive performance Pain perception Sensitive subgroups Time-dependent effects S187 S187 S187 S187 S188 S188 S188 Abstract The impact of ingesting various foods on psychological and behavioural functions is a topic of both interest and concern to the general public. In this article, the scientific literature concerning demonstrated cause-and-effect relationships is reviewed, beginning with methodological considerations specific to the quantification of particular behaviours and psychological events. The essential function of food is to satisfy hunger and the need for essential nutrients. The contribu