Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Serum Folate and Risk of Fatal Coronary Heart Disease Howard I. Morrison, PhD; Douglas Schaubel; Marie Desmeules, MSc; Donald T. Wigle, MD, PhD Objective.\p=m-\Toassess the relationship between serum folate level and the risk of fatal coronary heart disease (CHD) among men and women. Design.\p=m-\Retrospectivecohort study with serum folate levels measured from September 1970 to December 1972, with follow-up through 1985. Setting.\p=m-\Participantsin the Nutrition Canada Survey. Participants.\p=m-\Atotal of 5056 Canadian men and women aged 35 to 79 years with no history of self-reported CHD. Main Outcome Measure.\p=m-\Fifteen-yearCHD mortality. Results.\p=m-\Atotal of 165 CHD deaths were observed. We found a statistically significant association between serum folate level and risk of fatal CHD, with rate ratios for individuals in the lowest serum folate level category (<6.8 nmol/L [3 ng/mL]) compared with the highest category (>13.6 nmol/L [6 ng/mL]) of 1.69 (95% confidence interval, 1.10-2.61). Conclusions.\p=m-\Thesedata indicate that low serum folate levels are associated with an increased risk of fatal CHD. (JAMA. 1996;275:1893-1896) WELL-ESTABLISHED risk factors for coronary heart disease (CHD) include cigarette smoking, diabetes, hyperten¬ sion, and elevated serum cholesterol levels.13 Although there is only limited epidemiologie evidence that folate con¬ centration is associated with risk of CHD,4 folate consumption has been found to be a determinant of homocysteine-related carotid artery thickening.56 Moderate hyperhomocysteinemia has been linked to both cerebrovascular and CHD.78 Several mechanisms have been proposed,6 including the toxic effects of homocysteine on vascular endothelium, and in proliferation of arterial wall smooth muscle cells. Folate is required for the conversion of homocysteine to methionine.9 A strong inverse correlaFrom the Cancer Bureau, Health Canada, Ottawa, Ontario. Reprints: Howard Morrison, PhD, Cancer Bureau, Laboratory Centre for Disease Control Building, Tunney's Pasture, Ottawa, Ontario, Canada K1A 0L2. tion has been noted between serum folate and homocysteine levels among subjects with elevated homocysteine levels.10 The purpose of this article is to assess the relationship between serum folate level and the risk of fatal CHD among both men and women in the Nutrition Canada Survey (NCS) cohort. METHODS The NCS was conducted between Sep¬ tember 1970 and December 1972 using a 3-stage stratified cluster design. A to¬ tal of 12 795 people responded to the initial invitation to participate (46% re¬ sponse rate); 3295 unsolicited volunteers also participated.11,12 Each survey cen¬ ter included a laboratory for the collec¬ tion and initial processing of blood and urine samples. Blood samples were centrifuged to obtain serum, and a stabi¬ lizing solution was added to a portion of the serum for later determination of vi¬ tamin C level. All samples were then immediately frozen and shipped to a cen¬ tral laboratory for further biochemical analyses.11 Participants were followed up retro¬ spectively between the ages of 35 and 79 years. Thus, those younger than 35 years at interview began follow-up when they reached the age of 35 years, while those aged 80 years or older at interview were excluded. For editorial comment see 1929. The results of this article are based on the mortality experience of both solic¬ ited and unsolicited respondents. Among the 16090 NCS participants, individu¬ als who were interviewed as members of specially targeted populations (preg¬ nant women, Native Indians, and Inuit) were excluded (n=2828). A total of 3568 subjects were excluded because they did not contribute follow-up to the 35- to 79-year-old age category. Of the 8792 remaining individuals, an additional 3736 were excluded, either because of selfreported heart disease (n=987) or be¬ cause of missing data (n=2749), result¬ ing in a final study population of 5056 subjects. Work to extend the follow-up period to the end of 1993 is ongoing. We examined the following survey variables in our analysis: age; sex; re¬ spondent status (solicited or volunteer); smoking habits (never, former, current); diastolic blood pressure; history of dia¬ betes or presence of glucose in the urine (as detected with a glucose oxidase re¬ agent strip); serum total cholesterol level; serum vitamin C, serum vitamin A, and serum vitamin E levels; dietary folie acid consumption; use of folie acid supplements; use of vitamin and/or min- Downloaded from jama.ama-assn.org by guest on April 8, 2011 eral supplements; serum iron level; and folate level. Folate levels were determined microbiologically using Lactobacillus casei. Serum homocysteine levels were not measured. Age was categorized as 35 to 54,55 to 59, 60 to 64, 65 to 69, 70 to 74, and 75 to 79 years. The serum folate levels were categorized into approximate quartiles, ranging from <6.8 nmol/L (3.0 ng/mL) to >13.6 nmol/L (6.0 ng/mL). We as¬ sessed food consumption using a 24-hour food recall questionnaire. United States Department of Agriculture13 and Cana¬ dian Health Protection Branch tables14 were used to convert the information on diet to nutrient intake. Diastolic blood pressure was recorded in the survey only for subjects with a reading of 100 mm Hg or greater; only subjects who had this level or reported receiving treatment for hypertension from a physician were considered hy¬ pertensive. Fifth-phase diastolic blood pressure was measured while subjects serum were sitting. The mortality history of the cohort was obtained by linking the survey file to Statistics Canada's National Mortal¬ ity Database for the years 1970-1985 using the generalized iterative record Table 1.—Characteristics of Nutrition Canada Survey Cohort Study Subjects, Men, No. Characteristic (%) Women, No. (%) Serum cholesterol, mmol/L (mg/dL) <4.14(<160) 213(9.7) 279(9.8) 4.14-<5.17(160-<20 ) 788 (35.7) 941 (33.0) 5.17-<6.21 (200-<240) 838 (38.0) 978 (34.3) ==6.21 (==240) 368(16.7) 651(22.9) Diabetes status 2131 Nondiabetic Diabetic (96.6) 2742(96.2) 76(3.4) 107(3.8) Cigarette smoking Never Former 633(28.7) 1617(58.8) 526(23.8) 218(7.7) 1048(47.5) 1014(35.6) Current Hypertension No Yes 1839(83.3) 2222(78.0) 368(16.7) 627(22.0) Serum folate, nmol/L (ng/mL) (22.7) (21.9) 676 (30.6) <6.8 501 6.8-0.1 484 (<3.0) (3.0-<4.0) 9.1-<13.6(4.0-<6.0) ==13.6 (==6.0) 546 (24.7) Table 2.—Risk Factors for Fatal Coronary Heart Disease, Nutrition Canada Survey Cohort* 763 654 748 684 (26.8) (23.0) (26.3) (24.0) Rate Ratio (95% CI) 1.00 1.17 1.72 (0.73-1.85) (1.19-2.47) Cigarette smoking _ Never Former Current Hypertension No Yes Diabetes status Nondiabetic 1.00 2.37 (1.73-3.24) 1.00 2.26 (1.38-3.72) <4.14(<160) 4.14-<5.17(160-<200) 5.17-<6.21 (200-<240) ==6.21 (==240) 1.00 1.71 1.96 2.91 (0.61-4.78) (0.71-5.40) (1.05-8.09) 9.1-<13.6(4.0-<6.0) 6.8-0.1 (3.0-<4.0) <6.8(<3.0) 1.20 1.69 Diabetic Serum cholesterol, mmol/L (mg/dL) Serum folate, nmol/L (ng/mL) ==13.6 (==6.0) 1.00 1.11 (0.71-1.74) (0.75-1.92) (1.10-2.61) *Model included terms for age, sex, serum choles¬ terol level, smoking status, diabetes status, and diastolic blood pressure. CI indicates confidence Interval. 3.5 3.0 2.5 linkage computer system.15 Coronary heart disease (8th revision of the Inter¬ national Classification ofDiseases,16 ru¬ brics 410 to 414) was selected when listed as the underlying cause of death. We created Poisson regression mod¬ els that included terms for age, sex, se¬ rum folate level, cigarette smoking, hy¬ pertension, serum cholesterol level, and diabetes status. Volunteer status and serum levels of vitamins A, C, and E did not confound the folate-CHD associa¬ tion, and were therefore not included in the final models. Sex-specific and agespecific models were also created, as were models examining serum folate in¬ teraction by serum cholesterol levels. RESULTS Table 1 displays the baseline charac¬ teristics of study subjects. Approxi¬ mately 23% of the men and 27% of the women had serum folate levels less than 6.8 nmol/L (3.0 ng/mL). There was rela¬ tively poor correlation between serum folate and dietary folate (as determined by a 24-hour food recall, data not shown). As seen in Table 2, an increased risk of fatal CHD was associated with ciga¬ rette smoking, hypertension, elevated serum cholesterol level, diabetes, and decreased serum folate levels (lowest quartile compared with highest quartile rate ratio =1.69, 95% confidence inter¬ val [CI] =1.10-2.61). Risk increased in a step-wise fashion as the serum folate levels decreased (Figure). Sex-specific 0.5 0 4.5(2) 9.1(4) Serum Folate, 13.6(6) 18.1 (8) 22.7(10) nmol/L(ng/mL) Serum folate level and fatal coronary heart disease, Nutrition Canada Survey Cohort, 1970-1985. The rate ratio (95% confidence interval) was adjusted for age, sex, smoking, diabetes status, and serum cholesterol level. Rate ratios are plotted against median serum folate levels for each category. estimated relative risks (RRs) for folate are displayed in Tables 3 and 4 for men and women, respectively. Estimated RRs for the lowest folate category com¬ pared with the highest were greater for women (RR, 2.83; 95% CI, 1.30-6.18) than for men (RR, 1.36; 95% CI, 0.80-2.31); however, there was no statistically sig¬ nificant interaction by sex. Although there was no statistically significant interaction by age, the doseresponse relationship appears to be slightly stronger for individuals younger than 65 years compared with those 65 years and older (Table 5). To test whether the serum choles¬ terol level modifies the relationship be¬ tween serum folate level and fatal CHD risk, models were fitted with serum cho¬ lesterol levels less than 6.2 mmol/L (240 mg/dL) and 6.2 mmol/L (240 mg/dL) or higher. Coronary heart disease rate ra¬ tios according to serum folate level did not differ significantlyby cholesterol sta¬ tus (data not shown). There were too few exposed to folate supplements to assess risk (<1% of re¬ spondents). We found no important as¬ sociations between either use of vitamin and/or mineral supplements or dietary folie acid consumption and risk of fatal CHD. COMMENT Using computerized record linkage, a total of 165 CHD deaths were identified in the Canadian Mortality Database among the 5056 subjects included in the analysis. The potential for deaths hav¬ ing been missed is expected to have been small given the quality of the personal identifiers, the completeness of the Ca¬ nadian Mortality Database, and the ac¬ curacy of the record linkage process used by Statistics Canada.17 This study observed a statistically sig¬ nificant inverse relationship between se¬ rum folate level and risk of fatal CHD. A previous case-control study of early onset coronary artery disease noted a reduction in risk with increased serum folate levels.4 Folate consumption is a Downloaded from jama.ama-assn.org by guest on April 8, 2011 Coronary Heart Disease, Table 3.—Serum Folate Level and Risk of Fatal 1970 Through 1985, Men* Serum Folate, nmol/L (ng/mL) 213.6 (==6.0) 9.1-<13.6(4.0-<6.0) 6.8-0.1 (3.0-<4.0) <6.8(<3.0) Survey Cohort, Nutrition Canada Deaths, No. Person-Years RRt 30 7577 1.08 30 5511 1.36 95% CI 0.80-2.31 *RR indicates estimated relative risk, and CI, confidence interval. Test for trend, P=.02. tAdjusted for age, serum cholesterol level, smoking status, diabetes status, and diastolic blood pressure. Coronary Heart Disease, Nutrition Canada Survey Cohort, Table 4.—Serum Folate Level and Risk of Fatal 1970 Through 1985, Women* Serum Folate, nmol/L Deaths, No. (ng/mL) Person-Years 95% CI RRt (==6.0)_S)_8252_ _ 9.1-<13.6(4.0-<6.0)_13_8995_L24_0.53-2.91 6.8-<9.1 (3.0-<4.0)_9_7614_1^32_0.52-3.27 ==13.6 22 <6.8(<3.0) 8663 2.83 1.30-6.18 *RR indicates estimated relative risk, and CI, confidence interval. Test for trend, P=.04. tAdjusted for age, serum cholesterol level, smoking status, diabetes status, and diastolic blood pressure. Coronary Heart Disease Estimated Relative Risks by Age, Table 5.—Serum Folate Level and Risk of Fatal Nutrition Canada Survey Cohort* Age, Serum Folate, nmol/L (ng/mL) = Deaths, No. 13.6(26.0) 9.1-<13.6(4.0-<6.0) (3.0-<4.0) <6.8 (<3.0) 6.8-0.1 10 <65 y Age, Person-Years RRf 10 539 1.0 12156 1.49 0.58-3.80 9747 1.58 0.60-4.17 1.81 0.69-4.71 95% CI Deaths, No. 265 y 95% CI 28 Person-Years RRf 3888 1.0 4416 1.02 0.61-1.70 25 1.10 0.64-1.90 1.67 1.03-2.71 3560 *RR indicates estimated relative risk, and CI, confidence interval. tAdjusted for age, sex, serum cholesterol level, smoking status, diabetes status, and diastolic blood pressure. determinant of homocysteine-related carotid artery thickening.5 Arterial thickening has been associated with both stroke18·19 and CHD.20·21 This study is one of the first to note an association between serum folate level and fatal CHD. It has 2 significant ad¬ vantages over previous work. The NCS cohort was derived from a national nu¬ tritional survey. Previous studies of fo¬ late examined white men4 and the elder¬ ly.5 The current study noted significant associations between serum folate level and fatal CHD in both men and women. It did not rely on indirect measures of an intermediate health effect, such as ca¬ rotid artery thickening. Increased risks were not restricted to individuals with extremely low serum fo¬ late levels, but were observed for indi¬ viduals with normal levels as well, sug¬ gesting that current definitions of appropriate serum folate levels be reas¬ sessed. Although a significant exposureresponse relationship was observed, it is possible that some of the observed re¬ duction in risk associated with serum fo¬ late level reflects unmeasured dietary fac¬ tors for which folate is a good marker. However, serum levels of vitamins A, C, and E, also markers of a good diet, did not confound the association. Although neither interaction was sta¬ tistically significant, the magnitude of the protective effect of folie acid ap¬ peared to be greater for women than for men and slightly greater for those younger than 65 years compared with those aged 65 years and older. Relative risks for most CHD risk factors decline with age.1 It is not known ifthe increased RR among women as compared with men was a chance finding, or whether it reflects true differences in risk between the sexes. We found no clear relationship be¬ tween dietary folie acid consumption es¬ timated by 24-hour food recall and fatal CHD. Our crude measure of dietary fo¬ lie acid consumption did not include the contribution from vitamin supplements. Although a 24-hour food recall question¬ naire is a valid means of determining population folate levels,22 within-indi¬ vidual variation in diet severely limits its usefulness in classifying individuals. In addition, determination of dietary fo¬ late consumption was particularly prob¬ lematic in the NCS,14 and there was only a weak correlation between estimated dietary folate consumption and serum folate levels. There were too few ex¬ posed people to adequately assess folie acid supplements. No effect was noted for individuals reporting use of vitamin and/or mineral supplements. However, it is not known how many of these were actually taking supplements that con¬ tained folie acid. The failure to note a dietary effect should not be taken as evidence that dietary intervention is not to be encour¬ aged. Dietary folate was inversely as¬ sociated with carotid artery stenosis in the Framingham Study.5 Dietary folie acid supplementation has also been as¬ sociated with a reduced risk of neural tube defects.23 Low serum folate levels have been reported to enhance the ef¬ fect of risk factors for cervical dysplasia,24 and low folate intake has been as¬ sociated with an increased risk of rectal cancer.25 Folie acid fortification of staple foods has been recommended for consideration as a means to reduce the risk of cancer26 and neural tube defects.23 Folie acid for¬ tification of grain to prevent neural tube defects alone has been estimated to be cost-effective.27 Although folie acid is generally considered safe even at high levels of intake, high intake can mask the clinical signs of pernicious anemia,28 and can interfere with the effectiveness of anticonvulsants.29 Roughly one half of US adults on a given day consume less than the newly lowered recommended dietary allowance for folate,30,31 and an estimated 88%32 consume less than the levels needed to produce low, stable ho¬ mocysteine levels.6 Efforts should be made to increase folate consumption by a modification of the diet to include more vegetables and legumes. However, in spite of health promotion efforts, not all people will in¬ crease their dietary consumption of fo¬ late and the bioavailability of folate from food is significantly less than that from folie acid supplements.32 The recent de¬ cision of the Food and Drug Adminis¬ tration to require folie acid fortification of food products to prevent spina bifida may have the additional benefit of re¬ ducing CHD risk.33 References 1. Semenciw R, Morrison H, Mao Y, Johansen H, Davies JW, Wigle D. Major risk factors for cardiovascular disease mortality in adults: results from the Nutrition Canada Survey. Int J Epidemiol. 1988;17:317-324. 2. Kannel B. An overview of the risk factors for cardiovascular disease. In: Kaplan N, Stamler J, eds. Prevention of Coronary Heart Disease. Philadelphia, Pa: WB Saunders Co; 1983:1-19. 3. Martin M, Browner WS, Wentworth D, Hulley SB, Kuller LH. Serum cholesterol, blood pressure, and mortality: implications from a cohort of 361662 men. Lancet. 1986;2:923-926. 4. Pancharuniti N, Lewis CA, Sauberlich HE, et al. Plasma homocysteine, folate, and vitamin B-12 concentrations and risk for early onset coronary artery disease. Am J Clin Nutr. 1994;59:940-948. 5. Selhub J, Jacques PF, Bostom AG, et al. Association between plasma homocysteine concentra- Downloaded from jama.ama-assn.org by guest on April 8, 2011 tions and extracranial carotid-artery stenosis. N Engl J Med. 1995;332:286-291. 6. Stampfer MJ, Malinow MR. Can lowering homocysteine levels reduce cardiovascular risk? N Engl J Med. 1995;332:328-329. 7. Ueland PM, Refsum H, Brattstr\l=o"\mL. Plasma homocysteine and cardiovascular disease. In: Francis RB Jr, ed. Atherosclerotic Cardiovascular Disease, Hemostasis, and Endothelial Function. New York, NY: Marcel Dekker Inc; 1992:183-236. 8. Kang S, Wong PW, Malinow MR. Hyperhomocysteinemia as a risk factor for occlusive vascular disease. Annu Rev Nutr. 1992;12:279-298. 9. Gaby SK, Bendich A. Folic acid. In: Gaby SK, Bendich A, Singh VN, et al, eds. Vitamin Intake and Health: A Scientific Review. New York, NY: Marcel Dekker Inc; 1991. 10. Kang SS, Wong PW, Norusis M. Homocysteinemia due to folate deficiency. Metabolism. 1987;36: 458-462. 11. Department of National Health and Welfare. Nutrition Canada National Survey. Ottawa, Ontario: Information Canada; 1973. Catalog No. H58\x=req-\ 36/1973. 12. Johansen H, Semenciw R, Morrison HI, et al. Important risk factors for adult mortality: a ten year follow-up of the Nutrition Canada Survey cohort. Can Med Assoc J. 1987;136:823-828. 13. Watt BK, Merrill AL. Composition of Foods. Washington, DC: Agricultural Research Service, US Dept of Agriculture; 1963. Agricultural Handbook No. 8. 14. Department of National Health and Welfare. Nutrition Canada Food Consumption Patterns Report. Ottawa, Ontario: Information Canada; 1973. Catalog No. H58-36/1973. 15. Howe G, Lindsay J. A generalized iterative record linkage computer system for use in medical follow-up studies. Comput Biomed Res. 1981;14: 327-340. 16. National Center for Health Statistics. International Classification of Diseases, Adapted for Use in the United States. 8th rev. Washington, DC: US Dept of Health, Education, and Welfare, Public Health Service; 1967. 17. Shannon HS, Jamieson E, Walsh C, Julian JA, Fair ME, Buffet A. Comparison of individual follow\x=req-\ up and computerized linkage using the Canadian Mortality Data Base. Can J Public Health. 1989; 80:54-57. 18. Aronow WS, Ahn C, Schonenfeld M, Gutstein H. Extracranial carotid arterial disease: a prognostic factor for atherothrombotic brain infarction and cerebral transient ischemic attack. N Y State J Med. 1992;92:424-425. 19. O'Leary DH, Polak JF, Kronmal RA, et al. Distribution and correlates of sonographically detected carotid artery disease in the Cardiovascular Health Study. Stroke. 1992;23:1752-1760. 20. Salonen JT, Salonen R. Ultrasonographically assessed carotid morphology and the risk of coronary heart disease. Arterioscler Thromb. 1991;11: 1245-1249. 21. Aronow WS, Schonenfeld MR. Forty-five month follow-up of extracranial carotid arterial disease for new coronary events in elderly patients. Coron Artery Dis. 1992;3:249-251. 22. Mojonnier L, Hall Y. The National Diet Heart Study\p=m-\assessmentof dietary adherence. J Am Diet Assoc. 1968;52:288-292. 23. MRC Vitamin Study Research Group. Prevention of neural tube defects: results from the Medical Research Council Vitamin Study. Lancet. 1991;338: 131-137. 24. Butterworth CE Jr, Hatch KD, Macaluso M, et al. Folate deficiency and cervical dysplasia. JAMA. 1992;267:528-533. 25. Freudenheim JL, Graham S, Marshall JR, Haughey BP, Cholewinski S, Wilkinson G. Folate intake and carcinogenesis of the colon and rectum. Int J Epidemiol. 1991;20:368-374. 26. Block G. Micronutrients and cancer: time for action? J Natl Cancer Inst. 1993;85:846-848. 27. Romano PS, Waitzman NJ, Scheffler RM, Pi RD. Folic acid fortification of grain: an economic analysis. Am J Public Health. 1995;85:667-676. 28. Chanarin I. Adverse effects of increased dietary folate: relation to measures to reduce the incidence of neural tube defects. Clin Invest Med. 1994;17:253-255. 29. Bayless EM, Crowley JM, Preece JM, et al. Influence of folic acid on blood phenytoin levels. Lancet. 1971;1:62-64. 30. National Research Council. Recommended Dietary Allowances, 10th ed. Washington, DC: National Academy Press; 1989. 31. Subar AF, Block G, James LD. Folate intake and food sources in the US population. Am J Clin Nutr. 1989;50:508-516. 32. Sauberlich HE, Dretsch MJ, Skala JH, Johnson HL, Taylor PC. Folate requirement and metabolism in nonpregnant women. Am J Clin Nutr. 1987; 46:1016-1028. 33. Foulke J. Folic acid to fortify U.S. food products to prevent birth defects. Food and Drug Administration press release, February 29, 1996. Downloaded from jama.ama-assn.org by guest on April 8, 2011