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
Epidemiologic Reviews Copyright © 2001 by the Johns Hopkins University Bloomberg School of Public Health All rights reserved Vol. 23, No. 2 Printed in U.S.A. Height, Leg Length, and Cancer Risk: A Systematic Review D. Gunnell,1 M. Okasha,1 G. Davey Smith,1 S. E. Oliver,1 J. Sandhu,1 and J. M. P. Holly2 INTRODUCTION across all cancer sites. In this review, we have aimed to 1) synthesize the evidence linking height to site-specific cancers and 2) review possible mechanisms underlying the observed associations. While most studies have focused on associations between overall height and cancer risk, a few have also investigated associations of the two main components of height—leg length and trunk length—with cancer risk. Leg length is a marker for growth before puberty, since prepubertal increases in stature arise more from increases in leg length than from increases in trunk length. This is demonstrated by changes in the ratio of sitting height (mainly trunk length) to height during growth. At birth, the ratio is approximately 0.66, but by puberty it has declined to 0.52 (21). Thus, studies examining the relation of cancer risk to the components of stature—trunk length and leg length—may indicate periods during childhood growth when risk factors underlying height-cancer associations operate (22-24). Therefore, we have included such studies in this review. Associations between height and cancer risk have been reported in a number of prospective studies (1, 2). Taller individuals appear to be at increased risk. Furthermore, ecologic analyses indicate that geographic patterns of cancer incidence and mortality are associated with variations in population height (3-5). The most consistent associations have been found in relation to breast cancer (6-8), although associations have also been reported for many other cancer sites. Therefore, common mechanisms may underlie these associations, but their precise nature remains unclear. Models of cancer pathogenesis suggest that cancer arises as a result of DNA damage at a number of specific loci important in the regulation of the cell cycle or DNA repair (9, 10). While specific oncogenes, tumor suppresser genes, and cancer susceptibility genes have been identified, it has become increasingly apparent that epigenetic pathways also underlie the development of some malignancies. Several mechanisms may be common to many different cancers. For example, angiogenesis (the formation of new blood vessels) is an absolute requirement for the growth of all solid tumors (11). Similarly, apoptosis is a mechanism for eliminating damaged or dangerous cells from the body, and thus it provides a natural defense against cancer. The most potent cell survival factor controlling apoptosis is insulin-like growth factor I (IGF-I). Raised levels of IGF-I and reduced levels of its main binding protein, insulin-like growth factor (IGF)binding protein 3, may diminish this defense against a range of cancers. In support of this notion, recent prospective research has demonstrated that raised levels of IGF-I are associated with increased risks of prostate (12-14), breast (15), and colorectal (16, 17) cancers. While a number of reviews have synthesized results from studies reporting associations between height and breast (6-8), colorectal (18), and thyroid (19) cancer, and some have examined associations across a limited range of cancers (20), none have attempted to describe such associations SEARCH STRATEGY We systematically searched MEDLINE databases (US National Library of Medicine) for all relevant articles published between 1966 and 2000. The search strategy used was as follows: ["body height" or "height" or "anthropometry" or "skeletal formation" or "leg length" or "sitting height" or "body constitution" or "body composition"] AND ["neoplasms" or "cancer" or [the name of any site-specific neoplasm] ]. Terms shown in italic type were used as keywords; the remainder were mapped to Medical Subject Headings. Where the title of an article suggested that height-cancer relations might be presented, the abstract was reviewed by hand. Both English-language and non-English-language papers were reviewed. In cases where no results for height or leg length were mentioned in the abstract of a paper, the full article was examined to determine whether it should be included in the review. Relevant papers cited in the retrieved articles were also obtained. Received for publication June 2, 2000, and accepted for publication July 26, 2001. Abbreviations: Cl, confidence interval; IGF, insulin-like growth factor; IGF-I, insulin-like growth factor I. 1 Department of Social Medicine, Faculty of Medicine, University of Bristol, Bristol BS8 2PR, United Kingdom. 2 Bristol Royal Infirmary, Division of Surgery, Faculty of Medicine, University of Bristol, Bristol BS2 8HW, United Kingdom. Correspondence to Dr. David J. Gunnell, Department of Social Medicine, Canynge Hall, University of Bristol, Whiteladies Road, Bristol BS8 2PR, United Kingdom (e-mail: [email protected]). Presentation of findings Because of the large number of studies identified, we restricted our tabulations to cohort studies and nested casecontrol studies, since these types of studies are less prone to bias than case-control studies. We refer to nested casecontrol studies as cohort studies throughout this paper. Results of case-control studies are summarized; a summary 313 314 Gunnell et al. of these findings is available from the authors upon request. Few studies have examined the relation of components of stature—leg length and sitting height—to cancer risk; we tabulated the findings for all of these studies. Smoking may confound height-cancer associations, since smoking, height, and cancer incidence are all socially patterned. Therefore, we have presented height associations separately for cancers in which smoking is not thought to play an important etiologic role and those in which it is. In these latter studies, failure to adequately control for smoking may lead to biased estimates of height-cancer associations. We did not review studies examining associations of height with all cancer sites combined. All imperial units were converted to metric units. Relative risks are presented for taller groups compared with shorter (reference) groups. Where risks were presented in relation to a decrease in stature in the source publication, the inverse of the relative risk was computed. Relative risks are reported according to the number of decimal places used in the original paper; we rounded these to two decimal places if more than two decimal places were given in the data reported. Height-cancer associations for cancers in which smoking is not thought to be of etiologic importance Colorectal cancer. Table 1 summarizes the results of 16 cohort or nested case-control studies that have reported on associations between height and colorectal cancer (1, 2, 24—37). Seven of the cohort studies reviewed reported that greater stature was associated with increased cancer risk. A 20-60 percent increased risk among persons in the top height categories compared with those in the bottom categories was seen in most studies. There were no consistent differences in the associations according to gender or cancer site (colon or rectum) or according to whether mortality or incident disease was examined. In studies that controlled for socioeconomic position (2, 29), there was no clear evidence of socioeconomic confounding. In the three studies in which data were controlled for weight or body mass index, this adjustment led to a partial attenuation of the height-cancer associations (1, 27, 33). We identified 15 case-control studies reporting on heightcolorectal cancer associations. Two of these reported statistically significantly (p < 0.05) increased risks among taller individuals (38, 39). In one study, associations were seen in North American patients but not in Chinese patients (39). Three studies (40—42) found some evidence of increased risks (odds ratios > 1.2) among taller men or women, but these effects were not statistically significant at the 5 percent level. The remaining 10 case-control studies (43-52), including some with over 200 cases (46, 48-51), showed no height-cancer associations. In contrast, a meta-analysis of 13 case-control studies with a combined total of 5,287 cases of colorectal cancer reported weak but consistent heightcancer associations in males and females (18). The odds ratio in the top quintile of height compared with the bottom quintile was 1.31 (95 percent confidence interval (CI): 0.94, 1.82) among males (across height quintiles, p M d = 0.006) and 1.31 (95 percent CI: 0.96,1.79) among females {pmni = 0.19). Surprisingly, only two of the 13 case-control studies on which this meta-analysis was based (44, 50) had published their findings regarding height-cancer associations separately. Prostate cancer. Twenty-two reports on nested casecontrol or cohort studies (1,2, 13, 14, 24-26, 28, 29, 31, 53-64) described associations between height and prostate cancer (table 2). Some of these reports were based on longer follow-ups of the same cohort, and in some there may have been overlap in the populations studied (24-26, 28, 53-55, 58, 64). Where relative risks were specified, with the exception of two studies (61, 64), risks were all greater than 1.0 in the taller height categories. Most studies reported a 20-40 percent increased risk in the top height groupings compared with the bottom groups. Controlling for possible socioeconomic confounding attenuated the associations in the two cohorts reporting positive associations that examined this issue (29, 60). Height-prostate cancer associations were also apparent within the relatively socially homogenous US Physicians and Health Professionals cohorts (1, 56). As was seen with colorectal cancer, adjustment for measures of adiposity partially attenuated observed associations in two (1, 31) of the three (1, 31, 56) cohorts where its possible confounding effect was assessed. In the Kaiser Permanente cohort, race and year of birth appeared to strongly confound height-cancer associations (62). Adjustment for these factors resulted in a reduction in the relative risk in the tall group (versus the short group) from 1.45 (95 percent CI: 1.26, 1.67) to 1.15 (95 percent CI: 1.00, 1.33). In the one study where incidence and mortality associations were compared, associations were stronger for cancer mortality than for cancer incidence (59). This effect may have been due to the greater preponderance of screen-detected, localized cases among persons with incident (nonfatal) prostate cancer; many of such cancers may never become invasive. Inclusion of these cases dilutes the pool of "true" (invasive) cases and leads to measurement error in effect estimates. This effect may explain the stronger associations with advanced disease reported for one cohort (56). However, in the other two studies that compared height-cancer associations among persons with advanced disease with all cases, little difference in effect sizes was seen (61, 62). Of the 24 case-control studies examining height-prostate cancer associations, two reported associations that were statistically significant at the 5 percent level (65, 66). In one of these (65), significant associations were seen in US Whites but not in Blacks. Six papers based on five studies (40, 67-71) reported positive associations with height (odds ratios > 1.2) that did not reach conventional levels of statistical significance. One study reported a significant protective effect of height, but control patients in this study were men with prostatic hyperplasia (control selection bias) (72). The remaining 16 case-control studies, some based on over 1,000 cases, found no clear relation with height (73-88). Breast cancer. Twenty-four papers describing cohort and nested case-control studies have reported on associations between height and breast cancer (table 3) (2, 24, 25, 31, 89-108). Some of these papers were based on longer follow-ups of the same cohort, and in others there was a Epidemiol Rev Vol. 23, No. 2, 2001 Height, Leg Length, and Cancer Risk possibility of overlap in the cases studied (24, 89-98). Most studies examined incident disease, including deaths. Where reported, relative risks in relation to increasing height were greater than 1.0 in all but one of the studies (2). Interestingly, this was the only study in which deaths alone were included in the analysis (2), and this may indicate that higher mortality rates among persons from lower socioeconomic (shorter) groups attenuated any association with height. Increased risks of approximately 10-60 percent were seen in the top height categories compared with the bottom categories. Eleven studies distinguished premenopausal cancers from postmenopausal cancers. Associations were stronger in relation to postmenopausal cancer risk in seven of these studies (based on six cohorts), but the differences were often small. This is in keeping with a pooled analysis of data from seven prospective studies that reported relative risks, per 5-cm increase in height, of 1.02 (95 percent CI: 0.96, 1.10) for premenopausal cancer and 1.07 (95 percent CI: 1.03, 1.12) for postmenopausal cancer (109). In this pooled analysis, associations with height were stronger among women whose mothers had had breast cancer (p for interaction = 0.005), indicating the possible importance of genetic factors in height-cancer associations (109). Three papers reported relative risks before and after adjustment for, among other factors, weight or body mass index. In one of these, the height association became nonsignificant at the 10 percent level of statistical significance after adjustment (31); in the other two, no consistent effects of adjustment were seen (91, 99). In a pooled analysis, there was weak evidence {p for interaction = 0.12) that height associations were stronger among persons with low body mass indexes (109). In the two papers reporting results from models that controlled for socioeconomic position, there was little or no evidence of confounding (91, 98). Seventy-four published case-control studies had investigated height-breast cancer associations. Of these, 22 reported statistically significant positive associations with increasing height in at least one of the groups examined (110-131). A further 15 studies found some evidence of increased risks (odds ratio > 1.2 in the tallest group), but these increases did not reach conventional levels of statistical significance (132-146). Four studies reported a significantly reduced risk associated with increased height (147-150). The remaining 33 studies (40, 151-182) showed no association. A pooled analysis of 12 case-control studies indicated that height was associated with a greater risk of premenopausal cancer than of postmenopausal cancer (8). The relative risk in the top quintile of height versus the bottom quintile was 1.41 {p = 0.004) for premenopausal cancer and 1.11 (p = 0.20) for postmenopausal cancer. This differential effect was not consistently seen, however, across the range of studies reviewed and is in the direction opposite of that reported in the pooled analysis of cohort studies (109). Endometrial/uterine cancer. Height-endometrial cancer associations have been assessed in six cohort studies (table 4) (24, 31, 183-186). Positive associations were seen in three of these (183-185). In one study in which positive associations were reported, effects were stronger among persons with raised body mass indexes (184), and in another Epidemiol Rev Vol. 23, No. 2, 2001 315 the associations were restricted to those with postmenopausal cancer (185). Seventeen case-control studies described in 18 papers were identified. In only three of these studies were statistically significant increased risks seen among taller individuals (187-189). In the analysis by Goodman et al. (189), the increased risk was greatly attenuated after data were controlled for weight (the odds ratio in the top quartile of height declined from 1.8 to 1.3). A further five studies suggested an association with increasing height (191-195), but these findings did not reach the 5 percent level of statistical significance. In one of these studies, there was a suggestion that height-cancer associations were more marked among overweight women (192). Ten studies found no evidence of increased risk among taller women (40, 196-204). One case-control study found that greater stature was associated with reduced risk (40, 201). Other cancers not caused by smoking. Table 5 summarizes results from prospective investigations of other types of cancers not thought to be caused by smoking. Three (2, 29, 205) of the five cohort studies (2, 25, 29, 31, 205) investigating height associations with hematopoietic cancers suggested that taller individuals are at an increased risk; this trend persisted after adjustment for socioeconomic position (2, 29) and body mass index (205). Two case-control studies of adults have examined height-hematopoietic cancer associations (40, 206). One reported a positive association with Hodgkin's lymphoma (206); the other found no evidence of associations with Hodgkin's or non-Hodgkin's lymphoma and found a nonsignificant inverse association with myeloma (40). Two case-control studies of childhood leukemia were also inconclusive (207, 208). A number of other studies compared the heights of children with hematopoietic cancer with population norms (209-214), but these findings are difficult to interpret, because the population data used for comparison were often based on values for children's heights some years before the study period and thus took no account of secular increases in stature. Furthermore, the growth of children with cancer may be influenced both by cancer itself and by the effects on growth of the treatment they receive. Six cohort studies (2, 25, 26, 29, 31, 215) and seven casecontrol studies (40, 46, 216-220) have investigated associations between height and stomach cancer. Results of one of the six cohort studies suggested that greater stature was associated with decreased risk (2); the other cohort studies showed no significant association. Two (217, 220) of the seven case-control studies showed an inverse association between height and cancer risk in men and women. By contrast, in a Chinese case-control study, greater stature was associated with a threefold increased risk in females but not in males (218). The other four studies showed no clear patterns of association (40, 46, 216, 219). Associations of height with central nervous system tumors were examined in four cohort studies (25, 29, 31, 221) and one case-control study (222). One of the cohort studies indicated an increased risk with increased height (221), and another suggested a borderline increased risk in taller individuals (29). The only case-control study we iden- 1. Results of prospective studies reporting on the association between height and colorectai cancer Cohort (reference no.) No. of cases (incident cases include deaths) Estimate of relative risk Height comparison Age-adjusted risk (95% Cl*) aiian Japanese males (28) 101 colon, 63 rectum; incident cases >170cm vs. <157cm Colon 0.8 Rectum 0.8 aiian males (27) 726 colorectum; incident cases Top fertile of self-reported height compared with bottom fertile Cecum/ascending colon: OR* = 1.1 Transverse/descending colon: OR = 1.3 Sigmoid colon: OR = 1.6 Rectosigmoid junction: OR = 0.9 Rectum: OR = 0.8 Mean height in cases and noncases NES I" (males and females) (24) 62 in males and 67 in females; incident cases >178.6 cm vs. <169cm (males) >165.5 cm vs. <157 cm (females) Males (colorectai): 2.1 (95% Cl: 1.0,4.5) Females (colorectai): 1.6 (95% Cl: 0.8, 3.0) es' Health Study (females) (32) 191 colon, 49 rectum; incident cases 2168 cm vs. <160 cm (self-reported height) Colon: 1.6 (95% Cl: 1.1, 2.5) Rectum: 1.2 (95% Cl: 0.6, 2.6) sh male corporate employees (33) 51 colon, 42 rectum; incident cases <172cm vs. £178 cm Colon (mean height of cases vs. noncases): 175.1 vs. 174.5$ Rectum: 3.7 (95% Cl: 1.4, 9.9) ehall Study males (29) P™« = 0.41 NS NS 309 colon and small intestine, Mean height in cases and noncases 90 rectum; incident cases 1,0f (95% Cl: 0.6, 1.7) P,re* = 0 ° 9 P»nd = 0-58 Colon and small intestine Rectum ard and University of Pennsylvania umni (males) (25) cohort (females) (34) P^ = °™ NS* 289 colon, 108 rectum; incident cases Statist significa P ^ = 0-90 P ^ = 0-94 PM« = 0.90 Colon: 163.0 cm vs. 162.8 cm Rectum: 163.1 vs 162 8 aiian males (26) Fully adjusted risk (95% Cl) Statistical significance NS NS p < 0.01 Rectum: 3.1f (95% Cl: 1.0, 9.0) 212; incident cases £170 cm vs. <160cm (self-reported height) Colon: 1.19 1.23§(95%CI:0.83, 1.84) No. not specified; deaths only Risk per 15-cm increase in height Colon: 0.96 (95% Cl: 0.65, 1.43) Rectum: 1.01 (95% Cl: 0.52, 1.96) 0.93H (95% Cl: 0.63, 1.41) male health professionals (30) 203 colon; incident cases 2185 cm vs. £173 cm (self-reported height) Colon: 1.96 (95% Cl: 1.28, 3.01) male physicians (1) 341; incident cases £185 cm vs. <170 cm (self-reported height) Colorectai: 1.60 (95% Cl: 1.09,2.34) javik Study (males and females) (31) 338; incident cases Risk per 1 -cm increase in height No significant associations in males or females egian tuberculosis screening cohort ales and females) (35) Colon: n = 6,397 in males and 7,620 in females Rectum: n = 4,393 in males and 3,482 in females Incident cases Top quintile vs. bottom quintile p=0 1.03H (95% Cl: 0.52, 2.08) ^ = 0.002 'end = 0 1 5 p > 0.1 1.76# (95% Cl: 1.13, 2.74) 1 5 3 * * <95% C l : 1 0 4 ' 225 > No significant associations in males or femalesft Male colon: 1.37$$ (95% Cl: 1.27, 1.49) Male rectum: 1.17 (95% Cl: 1.06,1.29) Female colon: 1.35 (95% Cl: 1.26, 1.45) PL-= p>0 Height, Leg Length, and Cancer Risk I J9 8 rig Us _Q) ^_. £• (0 O Ml! in •so S2. CO — P CO t*- T- O 8 d ?5 O 1 I 8 LU I ft I 3 r 3 If I" 5§s§ OOQ< # oozo - 4 + C O 5 ^ 4t Epidemiol Rev Vol. 23, No. 2, 2001 >O2 317 tified that examined the association of height with central nervous system tumors was based on only 74 cases and indicated that short stature was associated with increased risk (222); however, it is difficult to interpret studies examining associations between height and childhood cancers, since the tumor itself may influence growth. Four cohort studies (25, 64, 223, 224) and three case-control studies (225, 226, 273) have examined associations with testicular cancer. The largest of the cohort studies (224) and one of the casecontrol studies (226) showed strong evidence of a greater risk in taller people. Two (227,228) of the three cohort studies (25, 227, 228) examining malignant melanoma risk factors reported positive associations with height. One (31) of the two (24, 31) cohort studies and all three case-control studies (40, 149, 229) of cervical cancer showed inverse associations with height. These associations remained after adjustment for an array of possible socioeconomic and reproductive confounding factors (40, 229). Other prospective studies have suggested that there are no strong height associations with gallbladder or cystic duct cancers (35). Four prospective studies (2, 230-232) have grouped together all cancers not thought to be related to smoking. Three of these studies indicated an increased risk in taller individuals that was not attenuated by controlling for possible socioeconomic confounding (2, 230, 232). Other case-control studies. A meta-analysis of 12 casecontrol studies investigating risk factors for thyroid cancer (19) reported positive associations between height and cancer risk in men and women. The pooled estimated of risk per 5cm increase in height was 1.08 (95 percent CI: 1.03, 1.13). Surprisingly, height-thyroid cancer associations were presented separately in only one (233, 234) of these 12 studies when findings were originally published. Four studies of male breast cancer have reported on associations with height (235-238). All four studies provided some evidence of increased risk with increasing height, although the evidence reached conventional levels of statistical significance in only one (235). Five case-control studies have examined heightovarian cancer associations (40, 239-242); there was probably some overlap in cases and controls in two of them (241, 242). No consistent picture was seen. One study reported a statistically significant inverse trend (p < 0.01) (40), and weak (nonsignificant) evidence of increased risk among taller women was seen in two Japanese reports (241, 242). For bone tumors, one study of adults suggested that risk increases with height (243); one (244) of the two studies of children (244, 245) reported similar effects, but in that study the control population was children with other cancers. The findings from case series that used population height norms for comparison were unreliable (246-248). One study showed a positive association between height and vaginal cancer (249). For liver cancer, one study examining this cancer reported inverse associations with height (40), but another study did not (250). Inverse associations with height were also reported in women but not in men in one case-control study of soft tissue sarcoma (251); in the two other case-control studies of this cancer (252, 253), no clear association was reported, although trends consistent with increasing risk with height were seen in the smallest study (252). 2. Results of prospective studies reporting on the association between height and prostate cancer Cohort (reference no.) No. of cases (incident cases include deaths) Estimate of relative risk Height comparison Age-adjusted risk (95% Cl») ard University alumni (53) 268; deaths only Mean height in cases and noncases 173.6 cm vs. 173.3 cmf ard and University of Pennsylvania umni (25) 243; incident cases Mean height in cases and noncases No significant association aiian Japanese (28) 96; incident cases >170cm vs. <157 cm 1.93 163.5 cm vs. 162.8 cm Statistical significance Fully adjusted risk (95% Cl) Statis signific NS* P » * = oNS aiian males (26) 306; incident cases Mean height in cases and noncases aiian Household Survey (54) 198; incident cases >173 cm vs. <162 cm (selfreported height) NES I* (24) NES I (55) 95; incident cases >178.6 cm vs. <169 cm 1.1 (95% Cl: 0.6, 2.0) 154; incident cases Mean height in cases and noncases Caucasians: 173 cm vs. 174 cm African Americans: 172 cm ehall Study (29) Not specified; deaths only Risk per 15-cm increase in height 1.59 (95% Cl: 1.01,2.44) hysicians (1) 1,047; incident cases >185cm vs. S170cm (self-reported height) 1.33H(95%CI:1.06, 1.77) ealth professionals (56) 1,369; incident cases >188cm vs. S173 cm (self-reported height) All cases: 1.34 = 0.01 Advanced disease: 1.62 = 0.002 1.8* (95% Cl: 1.0, 3.2) ft-= vs. 173 cm 1.43§ (95% Cl: 0.90, 2.22) 1.26# (95% Cl: 1.00, 1.59) All cases: 1.37** (95% Cl: 1.10, 1.70) Advanced: 1.68** (95% Cl: 1.16,2.43) Rural Health Study (57) dish construction workers (59) 71; incident cases >180cm vs. <173cm (self-reported height) 1.1 (95% Cl: 0.6, 2.0) pM = 0.8 n = 2,368—incident cases; n = 708—deaths only >180cm vs. <172cm Incidence: 1.14 (95% Cl: 1.00, 1.29) Mortality: 1.28 (95% Cl: 1.02, 1.60) p ^ = 0 04 javik Study (31) 524; incident cases Risk per 1-cm increase in height 1.01 (95% Cl: 1.00, 1.03) egian health screening cohort (60) n = 642 (entire data set); n = 378 (subjects with complete data); incident cases Top quintile vs. bottom quintile Entire data set: 1.2 (95% CI:0.9, 1.6) Subjects with complete data: 1.4 (95% Cl: 1.0,2.0) egian cardiovascular disease reening cohort (64) 220; incident cases Risk per 10-cm increase in height egian cardiovascular disease reening cohort (58) 70; incident cases >182cm vs. <170cm rew and Paisley Survey (2) 59; deaths only Risk per 10-cm increase in height p ^ = 0.04 No significant associationff p , ^ = 0.16 p , ^ = 0.05 0.99 (95% Cl: 0.82, 1.19) 1.2 (95% Cl: 0.6, 2.4) 1.30 (95% Cl: 0.88, 1.89) pM = 0.95 1.3#(95%CI:0.9, 1.9) ft--0-1 rlands cohort study (61) r Permanente Medical Care ogram cohort (62) more Longitudinal Study of Aging ) n = 681; subgroup analysis in relation to stage— n = 239 localized; n = 226 advanced; incident cases All cases: 2190 cm vs. <170cm Subgroups: risk per 5-cm increase in height All cases: 0.97 (95% Cl: 0.53, 1.77) n = 2,079 cases—1,323 localized, 314 with regional spread, 264 with distant spread, and 178 unstaged; incident cases >181 cm vs. <169cm All cases: 1.45HH (95% Cl 1.26, 1.67) 72; incident cases Mean height in cases and controls 176 cm vs. 176 cm , = 0.88 All cases: 0.96§§ (95% Cl: 0.52, 1.75) Subgroups— Localized: 0.99§§ (95% CI:0.89, 1.10) Advanced: 0.98§§ (95% CI:0.88, 1.10) All cases: 1.15## (95% Cl: 1 00, 1.33) Regional/distant cases only: 1.11## (95%CI: 0.85, 1.45) men (63) 96; incident cases >180cm vs. <175 cm 1.1 (95%CI:0.7, 1.7) ern Sweden Health and Disease hort Study (14) 149; incident cases Top quartile vs. bottom quartile 1.48 (95% Cl: 0.87, 2.50) = 0.7 , confidence interval; NS, not significant; NHANES I, First National Health and Nutrition Examination Survey, ot adjusted for age. ontrolled for ethnicity and income. ontrolled for civil service grade. so controlled for p-carotene and aspirin use. ontrolled for p-carotene, aspirin, body mass index, smoking, alcohol, and exercise. ntrolled for body mass index. ontrolled for weight, blood pressure, triglycerides, glucose, and creatinine. ontrolled for smoking, physical activity, education, and marital status. ontrolled for family history of prostate cancer and socioeconomic position. ontrolled for race. ontrolled for race and year of birth (also assessed' marital status, education, smoking, alcohol, and history of venereal disease; data were not controlled for these in the multivariable model). 3. Results of prospective studies reporting on the association between height and breast cancer Cohort (reference no.) No. of cases (incident cases include deaths) Estimate of relative risk Height comparison Age-adjusted risk (95% Cl*) Fully adjusted risk (95% Cl) 70; incident cases £170 cm vs. <155 cm reast screening cohort (102) 1,292; incident cases > 165.1 cm vs. 157.5-160 cm (self-reported height) sylvania university alumnae (25) 69; incident cases Mean height in cases vs. noncases NES l« (24) 122; incident cases >165.5 cm vs. <157 cm 2.1 (95% Cl: 1.2, 3.4) NES I (90) 131; incident cases Top quartile vs. bottom quartile 2.0 NES I (89) 70 premenopausal and 112 postmenopausal; incident cases >167 cm vs. <156 cm 1.2* (95% Cl: 1.0, 1.5) No significant differences 1.9§ (95% Cl: 1.1,3.2) Premenopausal: 1.6H (95% Cl: 0.6, 3.8) Postmenopausal: 2.0H (95% Cl: 1.0,3.8) 658 premenopausal and 420 postmenopausal; incident cases £168 cm vs <160cm; (self-reported height) Statist signific 2.4f erlands (general practice-based) 00) es' Health Study (92) Statistical significance Postmenopausal: 1.3 Premenopausal: 1.1# (95% CI:0.9, 1.3) Postmenopausal: 1.3# (95% Premenopausal11.13 Postmenopausal: 1 24 Cl: 1.0, 1.7) Premenopausal: 1.11** Postmenopausal: 1.29** Premenopausal: 1.1 es'Health Study (91) 806 pre- and 1,485 postmen- >170.2 cm vs. <157.5 cm; opausal; incident cases (self-reported height) dish health screening (103) 196 premenopausal, 986 postmenopausal; incident cases Per 5-cm increase in height Premenopausal: 1.11 (95% CI:0.98, 1.27) Postmenopausal: 1.10ft (95% Cl: 1.07, 1.13) aiian Linkage Study (94) 148 aged 30-44 years, 141 aged 45-49 years, 145 aged 50-54 years, and 135 aged 55-65 years; incident cases Top fertile vs. bottom fertile (self-reported height), by age at diagnosis Age (years) at diagnosis— 30-44: 1.02*4: (0.61, 1.71) 45-49:1.12** (0.65, 1.92) 50-54: 1.554:* (0.92, 2.61) 55-65: 1.28** (0.73, 2.24) aiian Household Survey (93) £160.1 cm vs. <154.9 cm; 86 premenopausal (aged (self-reported height) <50 years) and 292 postmenopausal (aged >50 years); incident cases = 0-58 0005 Premenopausal: 1.1 §§ (95%CI:0.6, 1.9) Postmenopausal: 1.5§§ (95% Cl: 1.1,2.1) egian tuberculosis screening hort (105) 3,305 premenopausal and 5,122 postmenopausal incident cases 789 premenopausal deaths and 1,594 postmenopausal deaths Per 15-cm increase in height Incidence— Premenopausal: 1.281111 (95% Cl: 1.13, 1.44) Postmenopausal: 1.421111 (95% Cl: 1.32, 1.53) Mortality— Premenopausal: 1.19ffll (95% Cl: 0.93, 1.52) Postmenopausal: 1.381111 (95% Cl: 1.21, 1.58) women's cohort (106) 229; incident cases Mean height in cases vs. controls (self-reported height) 160 cm vs. 160 cm p = 0.50 = P«»x, : r^rond ~ ,= 0 P«« = PM = egian health screening cohort (98) 291; incident cases £167 cm vs. <159 cm 1.43 P,— < ° 0 0 1 egian health screening cohort (97) 137 premenopausal (aged <51 years) and 99 postmenopausal (aged £51 years); incident cases £167 cm vs. <159 cm Premenopausal: 2.63 (95% Cl: 1.48, 4.68) Postmenopausal: 1.62 (95% P _ = 0.001 73 premenopausal (aged <54 years) and 95 postmenopausal (aged £54 years); incident cases £166 cm vs. <158cm Premenopausal: 1.4 (95% Cl: 0.7, 2.6) Postmenopausal: 1.9 (95% Cl: 1.1, 3.2) 79 premenopausal and 101 postmenopausal; incident cases £168 cm vs. <158cm (self-reported height) Premenopausal. 0.65 (95% Cl: 0.33, 1.30) Postmenopausal: 1.90ffi| (95% Cl: 0.96, 3.78) Project (Dutch Breast Screening gramme) (95) 38; incident cases £166 cm vs. <161 cm Postmenopausal: 1.51 (95% Cl: 0.69, 3.42) Project (Dutch Breast Screening gramme) (96) 147 premenopausal and 76 postmenopausal; incident cases >169 cm vs. <160.8 cm rlands Cohort Study (101) 626; incident cases £175 cm vs. S155 cm (self-reported height) avik Study (31) Per 1 -cm increase in 91 premenopausal (aged height <55 years) and 343 postmenopausal (aged £55 years); incident cases sey cohorts (99) east screening cohort (104) ho Bernado Study (US Whites) 8) ew and Paisley Survey (2) Permanente Medical Care gram cohort (107) < 95% Cl: 1-18' 1 - 73 ) Cl: 0.93, 2.81) Premenopausal: 1.3*** (95% Cl: 0.7, 2.5) Postmenopausal: 1.9*** (95% 2.04 (95% Cl: 1.19, 3.49) 161 cm vs. 160 cm 147; deaths only Per 10-cm increase in height 0.88 (95% Cl: 0.68, 1.16) pM = 0 Pi— = ° 0 7 P - = °-18 . < 0.001 2 . 0 6 t t t (95% Cl: 1.17, 3.63) No significant association§§§ Premenopausal: 1.04 (95% CIM.00, 1.08) Postmenopausal: 1.03 (95% Cl: 1.01, 1.05) Mean height in cases vs. noncases Pi— = Cl: 1.1, 3.3) Premenopausal: 1.28 (95% Cl: 0.78, 2.11) Postmenopausal: 0 . 9 6 t t t (95% Cl: 0.46, 1.98) 45; incident cases 214; incident cases 143## pM = 0 Pt— = P«™i * O p>0.1 p = 0.90 Height at age 17 years: 1.4 (95% Cl: 1.1, 2.0) Height at age 10 years: 1.1 (95%CI:0.8, 1.4) confidence interval; NHANES I, First National Health and Nutrition Examination Survey. t controlled for age. Calculated by comparing observed cases with expected cases based on national figures. ntrolled for weight and age at last mammogram. ntrolled for age at first birth, menopausal status, education, and alcohol consumption. ontrolled for age at first birth, age at menarche, family history of cancer, and education. ntrolled for parity, age at first birth, age at menarche, benign breast disease, family history of cancer, and smoking (also time since menopause, in postmenopausal women). ntrolled for age at baseline, age at menarche, body fatness at ages 5, 10, and 20 years, maternal body fatness, family history of cancer, alcohol (ages 18-22 years), adolescent and maternal sm ioeconomic position, and adolescent benign breast disease (also age at menopause, in postmenopausal women). ntrolled for region. ntrolled for age at first birth, 1972 socioeconomic score, and 1942 Cole (adiposity) index. ntrolled for education, ethnicity, and alcohol consumption. ot controlled for age. ntrolled for age at first birth, parity, occupation, and county of residence. ntrolled for age at first birth, family history of cancer, and weight. ntrolled for age at menarche, age at first birth, and number of liveborn children (also age at menopause, in postmenopausal women). ntrolled for age at menarche, parity, age at first birth, and alcohol consumption. ntrolled for weight, blood pressure, and levels of triglycerides, glucose, and creatinine. 4. Results of prospective studies reporting on the association between height and endometriai/uterine cancer Cohort (reference no.) breast screening cohort (183) No. of cases (incident cases include deaths) Estimate of relative risk Height comparison Age-adjusted risk (95% Cl*) 43; incident cases 2170 cm vs. <160cm NES I* (24) 30; deaths only >165.5 cm vs. <157 cm 1.0 (95% Cl: 0.2, 3.9) egian tuberculosis screening cohort 4) 2,208—incident cases; 405—deaths only Above mean vs. below mean Incident cases— Below mean BMI*: 1.2 (95% Cl: 1.1, 1 3) Above mean BMI: 1.9 (95%CI:1.7, 1.9) Deaths— Below mean BMI: 1.3 (95% Cl: 0.9, 1.8) Above mean BMI: 1.9 (95% Cl: 1.4,3.5) iian population registration (186) 214; incident cases Top fertile vs. bottom fertile (self-reported height) Project (Dutch Breast Screening ogramme) (185) 147; incident cases 2170 cm vs. <160 cm Premenopausal (n = 49): 0.8 Postmenopausal (n = 98): 2.5f avik Study (31) 98; incident cases Risk per 1 -cm increase in height No significant association Statistical significance Fully adjusted risk (95% Cl) Statisti significa 3.2t 1.3* (95% Cl: 0.7, 2.2) confidence interval; NHANES I, First National Health and Nutrition Examination Survey; BMI, body mass index; NS, not significant. adjusted for age. ntrolled for parity, age at first birth, relative weight, and socioeconomic position. ntrolled for weight, blood pressure, and levels of trigiycendes, glucose, and creatinine. p>0.1 No significant association§ p>0.1 Height, Leg Length, and Cancer Risk Height-cancer associations for cancers in which smoking is thought to be of etiologic importance Lung cancer. Ten cohort studies (1, 24-26, 28, 29, 31, 254-256) and three case-control studies (257-259) have reported on associations between height and lung cancer (table 6). Five of the six cohort studies that presented their results as relative risks showed that risks were all greater than 1, although only one of these was statistically significant at conventional levels (254). In several studies (24, 29, 256), the possible confounding effects of smoking were not controlled for. One (259) of the three case-control studies (257-259) indicated that among smokers, shorter stature was associated with increased risk in males and females, with a difference in mean heights of 1 cm. In contrast, in another of these studies there was a suggestion that greater stature in women was associated with increased risk (257). Other cancers caused by smoking. Table 7 summarizes findings from cohort studies that have examined the associations of height with other smoking-related cancers. Two (29, 215) of the four cohort studies (25, 29, 31, 215) reporting associations with esophageal cancer demonstrated reductions in risk with increasing height. Neither of these analyses controlled for smoking. Similar findings were reported in the four case-control studies that examined this issue (40, 219, 220, 260). In two of these studies, each with approximately 300 cases, strong associations were seen in models adjusted for smoking (219, 220). In one of these studies, associations were seen for adenocarcinoma but not for squamous cell carcinoma (220). The cohort studies examining pancreatic cancer (25, 29, 31, 35, 261) found no clear pattern of association with height, although one (262) of the five case-control studies (40, 190, 262, 263, 335) found evidence of greater risk among taller women. A pooled analysis of five case-control studies found some evidence of increasing risk with greater height: The relative risk in the tallest quintile was 1.35 (95 percent CI: 0.87, 2.09), and the test for linear trend was not significant at conventional levels (p = 0.12) (336). The four cohort studies that examined bladder cancer showed no clear pattern of association with respect to height (24, 25, 29, 31). Neither of the two case-control studies (40, 264) examining this issue showed strong effects, although in an Italian case-control study there was some evidence of increasing risk with height in females (40). An increased risk of renal carcinoma in taller males was reported in one (31) of two cohort studies (25, 31) examining this issue. The five case-control studies that presented data on height-renal cancer associations (40, 265-268) provided no evidence of height effects, although in one (40), a twofold increased risk was seen in the tallest men and the opposite effect was seen in women. In the two case-control studies that examined associations between stature and carcinoma of the larynx (40, 260), one reported no evidence of an influence of height on cancer risk in men (40). In the other, which examined risk in men and women combined, risk increased with decreasing stature, but effects were attenuated and nonsignificant after data were controlled for (among other factors) smoking (260). The same study reported associations Epidemiol Rev Vol. 23, No. 2, 2001 323 between reduced stature and oropharyngeal cancer risk in men and women. Again, associations were greatly attenuated but remained significant (p < 0.001) after data were controlled for smoking differences (260). In three prospective studies (2, 230, 231) (table 7) and two case-control studies (269, 270), all smoking-related cancers were grouped together. None of these studies found an association between height and smoking-related cancers. Components of height and cancer risk Cohort studies. Seven research papers based on three cohorts (22, 24, 26, 89, 128, 232, 271) have reported results from prospective investigations of the associations between components of stature and cancer (table 8). Analyses of data from the First National Health and Nutrition Examination Survey (24, 89, 128) indicated either that leg length and sitting height were equally related to risk (prostate, colorectal, bladder, and uterine cancers) or that leg length was the component of stature associated with increased risks (lung, breast, and cervical cancers) (24). Neither of the two analyses of the Honolulu Heart Program cohort showed significant associations with leg length or sitting height (26, 271), although the earlier analysis of prostate cancer risk suggested that leg length was associated more strongly with increased risk than was sitting height (26). Two published articles based on the Boyd Orr cohort assessed associations between the components of stature, measured in childhood, and later cancer risk (22, 232). In the first paper, no significant associations were reported, although leg length appeared to be more strongly associated with risk than did sitting height (232). More recently, a subgroup analysis based on persons who were aged <8 years when they were measured in childhood indicated that those with longer legs, but not those with longer trunks, were at greater risk of death from cancers not thought to be related to smoking (22). Case-control studies. Eight case-control studies have investigated associations between the components of height and cancer risk (84, 86, 128, 149, 161, 197, 272, 273) (table 9). In all but one (273) of these studies, sitting height was measured and differences were reported with respect to sitting height or the sitting heightheight ratio. High values for this ratio indicate a long trunk relative to total height, and hence relatively short legs. In three studies, the source of controls is likely to have resulted in selection biases (84, 86, 149). In one of these studies, controls were approximately 15 years younger than cases (149); in the other two (both by the same group of investigators (84, 86)), urology clinic patients were used as controls. Two of the remaining five investigations indicated that leg length was the component of stature associated with increased risk of breast cancer (128) and testicular cancer (273). In Swanson et al.'s investigation of endometrial cancer (197), there was no association with overall height but there was a significant decreased risk with increasing sitting height, which suggests that leg length was associated with increased risk. A further study based on postmortem examination of children who died from leukemia indicated some disproportion in the cases (272). Reanalysis of the original data tabulated in that paper 5. Results of prospective studies reporting on the associations between height and "other" cancers (cancers other than colorectal, prostate, breast, and etrlal/uterine cancer) not thought to be caused by smoking Type of cancer and cohort (reference no.) No. of cases (incident cases include deaths) Estimate of relative risk Height comparison Age-adjusted risk (95% Cl*) atopoietic cancers arvard and University of Pennsylvania 343; deaths only alumni (males and females) (Hodgkin's and non-Hodgkin's lymphoma and all leukemias) (25) Mean height in cases and noncases No significant association hitehall Study males (leukemia and lymphoma) (29) Risk per 15-cm increase in height Leukemia: 1.14 (95% Cl: 0.58, 2.22) Lymphoma: 1.78 (95% Cl: 0.99, 3.23) No. not specified: deaths only eykjavik Study (males and females) (leukemia and lymphoma) (31) 137; incident cases Risk per 1-cm increase in height urses' Health Study (females) (non-Hodgkin's lymphoma) (205) 199; incident cases enfrew and Paisley Survey (males and females) (all hematopoietic cancers) (2) 79; deaths only 2173 cm vs. <157 cm (self-reported height) Risk per 10-cm increase in height mach cancer arvard and University of Pennsylvania 64; incident cases alumni (males and females) (25) awaiian males (26) 229; incident cases hitehall Study males (29) No. not specified; deaths only 246; incident cases eykjavik Study (males and females) (31) orwegian tuberculosis screening cohort (males and females) (215) enfrew and Paisley Survey (males and females) (2) 6,077 males and 3,737 females; incident cases 103; deaths only tral nervous system cancers arvard and University of Pennsylvania 90; incident cases alumni (males and females) (25) 1,538 males and 1,511 orwegian tuberculosis screening females; incident cases cohort (males and females) (221) hitehall Study males (brain cancer) (29) eykjavik Study (males and females) (31) Mean height in cases and noncases Mean height in cases and noncases Risk per 15-cm increase in height Risk per 1-cm increase in height Top quintile vs. bottom quintile Statist significa 1.10t (95% Cl: 0.55, 2.17) 1.89t(95%CI: 1.04,3.45) p>0.1 1.89 (95% Cl: 1.33, 2.63) No significant associations in males or females^ p>0 ft-= 1.92H (95% Cl: 1.37, 2.78) No significant association 162.7 cm vs. 162.8 cm# NS* 0.81 (95% Cl: 0.53, 1.25) No significant associations in males or females O.93f (95% Cl: 0.61, 1.45) p>0.1 No significant associations in males or females:}: p>0 Males: 1.0** (95%CI:0.9, 1.1) Females: 1.0** (95% Cl: 0.9, 1.1) 0.68 (95% Cl: 0.50, 0.92) Mean height in cases and noncases No significant association 0.75H (95% Cl: 0.56, 1.03) Per 15-cm increase in height Risk per 1-cm increase in height Fully adjusted risk (95% Cl) 2.4§(95%CI:1.2, 4.7) Risk per 10-cm increase in height No. not specified; deaths only Risk per 15-cm increase in height 99; incident cases No significant associations in males or females Statistical significance 1.75 (95% Cl: 0.87, 3.57) No significant associations in males or females p>0.1 Males: 1.2ft (95% Cl: 1.1, 1.4) Females: 1.2ft (95% Cl: 1.1, 1.4) 1.79f (95% Cl: 0.87, 3.57) p = 0. No significant associations in males or femalesf. p>0 p = 0. cular cancer rvard and University of Pennsylvania 67; incident cases alumni (25) Mean height in cases and noncases nish military recruits (223) 224 persons aged <19 years at diagnosis and 214 persons aged >19 years at diagnosis; incident cases >190 cm vs. <170 cm rwegian cardiovascular disease screening cohort (64) 47; incident cases Risk per 10-cm increase in height 1.12 (95% Cl: 0.75, 1.67) rwegian tuberculosis screening cohort (224) 553; incident cases £183 cm vs <171 cm All testicular cancers: 1.60# (95% Cl: 1.21, 2.10) noma rvard and University of Pennsylvania alumni (males and females) (25) 104; incident cases Mean height in cases and noncases No significant association rwegian tuberculosis screening cohort (males and females) (227) 2,144 males and 2,184 females; incident cases Top quintile vs bottom quintile Males: 1.6§§ (95% Cl: 1.4, 1.8) Females: 1.6§§ (95% Cl: 1.4, 1.8) rwegian health screening cohort males and females) (228) 106 men and women; incident cases £177 cm vs. <163cm 3.11111 (95% Cl: 1.4,6.7) cal cancer ANES I* (24) 20; incident cases >165.5 cm vs. <157 cm 0.5 (95% Cl: 0.1, 2.2) 40; incident cases Per 1-cm increase in height 0.95 (95% Cl: 0.89, 1.00) rwegian tuberculosis screening ladder/cystic duct cancers cohort (males and females) (35) 350 males and 617 females; incident cases Top quintile vs. bottom quintile ncers not thought to be caused by smoking itehall Study males (230) 725; deaths only >183cm vs. <168 cm 1.29 (95% Cl: 0.96, 1.72) asgow alumni (males and females) (231) 190 males and 61 females; deaths only Per 10-cm increase in height Males: 1.06 (95% Cl: 0.84, 1.33) Females: 1.04 (95% Cl: 0.67, 1.62) ykjavik Study (31) yd Orr cohort (males and females) (232) nfrew and Paisley Survey (males and females) (2) 20 males and 39 females; deaths only 365 males and 554 females; deaths only No significant association Age <19 years at diagnosis: 1.1«(95%CI:0.4, 2.7) Age >19 years at diagnosis: 1.0tt (95%CI:0.4, 2.6) Seminoma: 1.41 (95% Cl: 1.01, 1.99) Nonseminoma: 1.21 (95% Cl: 0.75, 1.95) p<0.1 No significant associations in multivariable analysis! p>0. Males: 1.2*» (95% Cl: 0.8, 1.7) Females: 1.2** (95% Cl: 0.9, 1.5) Per 1-standard-deviation increase in childhood height Males: 1.21 (95% Cl: 0.78, 1.88) Females: 1.04 (95% Cl: 0.74, 1.47) Per 10-cm increase in height Males: 1.09 (95% Cl: 0.93, 1.27) Females: 1.07 (95% Cl: 0.93, 1.23) P™« < ° 0 1 1 - 3 6 # # < 9 5 % Cl: 1.01, 1.82) pM <0 0.92*** (95% Cl: 0.70, 1.20) 0.74*** (95% Cl: 0.45, 1.23) NS 1.30ttt (95% Cl: 0.80, 2.11) NS 0.99ttt (95% Cl: 0.69, 1.42) 1.14H(95%CI:0.97, 1.33) 1.12H(95%CI:0.98, 1.30) Table con 326 Gunnell et al. (data not shown) indicates that cases had shorter legs relative to their sitting height, although effects of treatment on growth cannot be ruled out. DISCUSSION OF MAIN FINDINGS CO 2 o cog t « 2 So co - . X aj -* "O 'c c OJ ••s= " ~ 1 2 TJ o '5 « |S| •g CO CO n CO to x" -O "w o 8"f ? i II (0 O) jS ^ o3 S o o_ t « 6 •D 0) "5 cb Iffillifl = S o .> Most prospective studies reporting associations between height and cancer have shown either that risk increases with stature or that there is no significant relation. Evidence from case-control studies is considerably weaker, but meta-analysis of their findings in relation to thyroid, breast, pancreatic, and colorectal cancer suggests a general pattern of increasing risk with greater stature (8, 18, 19, 336). Associations are similar in men and women and have been reported for a range of cancer sites—particularly for breast, prostate, colorectal, and hematopoietic cancers—but they are also seen with central nervous system tumors, malignant melanoma, endometrial cancer, and thyroid cancer. In the relatively few studies in which multivariable analyses have been carried out to control for the possible confounding effects of socioeconomic position and indicators of adiposity, the observed associations are sometimes but not generally attenuated. For breast cancer, there is some evidence that height associations are stronger in persons of lower weight (109), whereas the opposite has been reported for endometrial cancer (184). Associations are seen in relation to both cancer incidence and cancer mortality, indicating that survival bias is unlikely to explain the associations observed. The strength of association in studies with positive effects is relatively weak; there is usually, at most, a 20-60 percent increased risk in the tallest groups compared with the shortest. The magnitude of increased risk is similar for the three most common cancers not caused by smoking: colorectal cancer, prostate cancer, and breast cancer. Associations of stature with smokingrelated cancers are more difficult to interpret, since many of these studies have not controlled for the potentially confounding effects of tobacco smoking. Few studies have examined associations between the components of stature and cancer risk. Although the evidence in this area is weak, the component of height more often associated with increased risk is leg length. Leg length is a marker for growth before puberty, since prepubertal increases in stature arise more from increases in leg length than increases in trunk length. This suggests that the biologic mechanisms underlying the relation between height and cancer may have their origins in factors which influence long bone growth in childhood. The range of studies reporting height-cancer associations suggests either a common underlying mechanism or bias in the studies reviewed. Therefore, before we discuss possible mechanisms for the observed associations, alternative explanations will be considered. <D CO Chance and bias in UJ m cccc_cmcccccc _ r O O O O O O O O O O O O 0 OOOOOZOZOOOOOO Many large cohort studies and a number of case-control studies have reported on height-cancer associations, which indicates that these findings are unlikely to be due to chance alone. While bias is an important consideration in epidemiEpidemiol Rev Vol. 23, No. 2, 2001 6. Results of prospective studies reporting on the association between height and lung cancer Cohort (reference no.) No. of cases (incident cases include deaths) Estimate of relative risk Height comparison Age-adjusted risk (95% Cl*) 46; deaths only Mean height in cases and noncases iian Japanese males (28) 101; incident cases £170 cm vs. <160 cm iian survey of males and females 4) 54 in smokers and 26 in nonsmokers; incident cases Males: top fertile vs. bottom fertile (selfreported height) Females: above median vs. below median (self-reported height) iian males (26) 236; incident cases Mean height in cases and noncases 163.4 vs. 162.8t rd and University of Pennsylvania mni (males and females) (25) 335; incident cases Mean height in cases and noncases No significant association h occupational cohorts of males 5) NES I* males (24) hall Study males (29) women's cohort (256) ale physicians (1) avik Study (males and females) ) 168.0 vs. 168.9t Fully adjusted risk (95% Cl) Statistical significance Statisti significa NS* Smokers— Males: 3.7§ (95% Cl: 1.2, 11.5) Females: 1.8§ (95% Cl: 0.6, 5.2) Nonsmokers— Males: 2.9§ (95% Cl: 0.6, 13.0) Females: 3.4§ (95% Cl: 0.5, 24.1) 114; incident cases >178.6 vs. £169 cm 1.1 (95% Cl: 0.6, 2.0) No. not specified; deaths only Risk per 15-cm increase in height 0.94 (95% Cl: 0.77, 1.16) 233; incident cases <155cm vs. >165 cm (self-reported height) 0.81 (95% Cl: 0.57, 1.14) 170; incident cases £185 cm vs. 5170 cm (self-reported height) 1.2#(95%CI:0.7, 1.9) n = 472; incident cases Risk per 1 -cm increase in height No significant associations in males or females , confidence interval; NS, not significant; NHANES I, First National Health and Nutrition Examination Survey, ot adjusted for age. ontrolled for age and smoking. ontrolled for socioeconomic position and race. ontrolled for civil service grade. so adjusted for p-carotene intake and aspirin use. ntrolled for p-carotene, aspirin, body mass index, smoking, alcohol, and exercise, ontrolled for weight, blood pressure, triglycerides, glucose, and creatinine. NS 1.12H (95%CI: 0.91, 1.37) Pwnd = ° ' 4 6 p>0.1 1 1 - ** ( 9 5 % Cl: °'6' 1 8 ) No significant associations in males or femalesft p>0. 7. Results of prospective studies reporting on the associations between height and "other" cancers (cancers other than lung cancer) thought to be caused g Type of cancer and cohort (reference no.) No. of cases (incident cases include deaths) hageal cancer rvard and University of Pennsylvania 48; incident cases alumni (males and females) (25) Estimate of relative risk Height comparison Age-adjusted risk (95% Cl*) Mean height in cases and noncases No significant association hitehall Study males (29) No. not specified; deaths only Risk per 15-cm increase in height 0.43 (95% Cl: 0.22, 0.85) ykjavik Study (males and females) (31) 49; incident cases Risk per 1-cm increase in height No significant associations rwegian tuberculosis screening cohort (males and females) (215) 742 males and 274 females; incident cases Top quintile vs. bottom quintile reatic cancer rvard and University of Pennsylvania alumni (males and females) (25) 127; incident cases Mean height in cases and noncases Statistical significance Fully adjusted risk (95% Cl) 0.47f (95% Cl: 0.23, 0.93) p > 0.1 No significant associations^ No significant association "Several other anthropometric measurements, including height, were not related to subsequent pancreatic cancer" (261) 450; incident cases hitehall Study males (29) No. not specified; deaths only Risk per 15-cm increase in height 0.90 (95% Cl: 0.53, 1.54) 0.93t (95% Cl: 0.54, 1.59) ykjavik Study (males and females) (31) 65 males and 36 females; incident cases Per 1 -cm increase in height Males: 1.0 (95% Cl: 1.0, 1.1) Females: 1.1 (95%CI:1.0, 1.1) No significant association^ rwegian tuberculosis screening cohort (males and females) (35) 4,939; incident cases Top quintile vs. bottom quintile der cancer rvard and University of Pennsylvania alumni (males and females) (25) 108; incident cases Mean height in cases and noncases No significant association HANES I* males (24) 27; deaths only >178.6 cm vs. <169 cm 0.7 (95% Cl: 0.2, 2.8) hitehall Study males (29) No. not specified; deaths only Risk per 15-cm increase in height 1.02 (95% Cl: 0.54, 1.92) ykjavik Study (males and females) (31) 215; incident cases Risk per 1-cm increase in height No significant associations in males or females 77; incident cases Mean height in cases and noncases No significant association 109 males and 58 females; incident cases Increase in risk per 1-cm increase in height Males: 1.1 (95%CI:1.0, 1.1) Females ykjavik Study (males and females) (31) p>0.1 Males: 0.6§ (95% Cl: 0.5, 0.8) Females. 0.7§ (95% Cl: 0.4, 1.0) iser Permanente Medical Care Program cohort (males and females) (261) ey cancer rvard and University of Pennsylvania alumni (males and females) (25) Statist significa p>0.1 No significant association^ 1.00t (95% Cl: 0.53, 1.89) p > 0.1 NS* No significant associations in males or females^ Males: 1.0* (95%CI: 1.0, 1.1) Females p>0 NS Height, Leg Length, and Cancer Risk 8 d II —v CM CO o o ••- O o <=^ S S5 S 8 O II CO 00 ° d a a 5 E £ CO CO A <n CO .2 >. a si (D — JE 5 cr tcl c c c cH £ 5 c3 cj cj (S c3 cS Epidemiol Rev Vol. 23, No. 2, 2001 329 ologic research, cohort studies—the results of which we have given the most emphasis to—are less prone to selection and information biases that often beset other observational designs. Methodological bias in the reported studies. Five possible sources of bias in the conduct of the studies reviewed here must be considered. First, detection bias would arise if cancer were more likely to be identified in taller individuals. This might occur because taller (affluent) individuals are more likely to undergo health screening. Such an effect would lead to stronger associations for incident (screendetected) disease compared with fatal disease, under the assumption that some screen-detected cancers would not have been fatal or were detected earlier in taller individuals. The example of prostate cancer provides us with an opportunity to assess this possibility. With the advent of prostate cancer screening, one might expect height associations to be stronger for incidence than for mortality, because of the greater chance of detecting incidental cancers in individuals who undergo screening and the fact that individuals who request screening tend to be more affluent (taller). However, if anything, the opposite effect is seen for prostate cancer (59). More generally, in the papers reviewed, the associations are similar for incidence and mortality, making detection bias unlikely. Second, survival bias might arise in studies examining cancer mortality rather than incidence if stature were a marker of the chances of surviving once cancer had been diagnosed, rather than a marker of cancer risk per se. Such an effect might be expected to diminish rather than increase height-cancer associations, since higher socioeconomic position is associated with both greater stature and better prospects of cancer survival (274). Interestingly, in the breast cancer cohorts (see table 3), the only study not to show a positive association with height was the one based on deaths alone (2). Third, measurement bias would occur if there were systematic differences in the measurement or reporting of stature among those who were to go on to become cases. At the time of measurement in cohort studies, observers are unlikely to be aware of subjects' likelihood of developing cancer. While measurement error may arise when height assessment is based on self-reports rather than on measurements, such errors are likely to be nondifferential with respect to later cancer occurrence, and this is likely to attenuate associations. Furthermore, since shorter individuals tend to overreport their height, any height-cancer associations found are likely to be attenuated (275). Fourth, it has been suggested that height-cancer associations may arise because cancer risk increases with age and taller individuals are more likely to live longer (276). While this phenomenon may influence the findings of some studies, it is unlikely to explain associations seen with cancers in young people (such as premenopausal breast cancer and testicular cancer). In addition, the statistical methods used in some analyses (Cox's proportional hazards models) ensure that persons who die of cancer are compared only with persons of the same age who are alive at the time cancer is detected or cancer death occurs (277). 8. Results of prospective studies reporting on the associations between components of height (leg length, sitting height, and trunk length) and cancer Estimate of relative risk No. of cases (re.e^nc 'no.) ANES I • males and femalest 4) ANES I (89) olulu Heart Program (271) olulu Heart Program (26) Height comparison and source of controls Incident cases—all sites Top quartile vs. bottom quartile Males: 114 lung, 95 prostate, 62 colorectal, 27 bladder, and 170 at all other sites Females: 122 breast, 67 colorectal, 30 uterus, 20 cervix, and 168 at all other sites Age-adjusted risk (95% Cl*) Fully adjusted risk (95% Cl) Males: LungLeg length: 1.6 (95% Cl: 0.9, 2.7) Sitting height: 0.6 (95% Cl: 0.3, 1.1) Prostate— Leg length- 1.2 (95% Cl: 0.7, 2.2) Sitting height: 1.2 (95% Cl: 0.6, 2.3) Colon/rectum— Leg length: 1.5 (95% Cl: 0.7, 3.0) Sitting height: 1.4 (95% Cl: 0.6, 3.0) BladderLeg length: 1.1 (95% Cl: 0.4, 3 6) Sitting height: 1.2 (95% Cl: 0.3, 4.0) All other sites— Leg length: 2.0 (95% Cl: 1.2, 3.2) Sitting height: 1.2 (95% Cl: 0.7, 1.8) Leg len Neither Neither Neither Leg len Females: Breast— Leg length: 1.6 (95% Cl- 1.0, 2.6) Sitting height: 1 3 (95% Cl: 0.7, 2.2) Colon/rectum— Leg length: 1.4 (95% Cl: 0.7, 2.8) Sitting height: 1.4 (95% Cl: 0.7, 3.0) Uterus— Leg length: 1.1 (95% Cl: 0.4, 2.8) Sitting height: 0.9 (95% Cl. 0.3, 3.2) CervixLeg length: 3.8 (95% Cl: 0.8, 18.3) Sitting height: 1.0 (95% Cl: 0.3, 3.1) All other sites— Leg length: 0.8 (95% Cl: 0.5, 1.2) Sitting height: 0.8 (95% Cl: 0.5, 1.4) Incident cases of breast cancer (n= 182) Mean height and sitting height of cases vs. noncases Incident cases of prostate cancer (n= 174) Leg length >78 cm vs. <75cm Sitting height £88 cm vs. <86cm Incident cases of prostate (n = 306), colon (n = 289), rectum (n= 108), lung (n = 236), and stomach (n = 229) cancer Mean sitting height and leg length in cases vs. noncases (cm) Leg len Neither Neither Leg len Neither Height: 161.7 cm vs. 161.2 cmt (p = 0.28) Sitting height: 85.4 cm vs 85.1 cm* (p = 0.27) Leg length: 1.2 (95% Cl- 0.9, 1.8) Compon stature as with incr cancer Neither Leg len Sitting height: 0 9 (95% Cl: 0.7, 1.3) Prostate— Leg length: 76.7 vs. 76.3 Sitting height: 86.7 vs 86.5 Colon— Leg length: 76.1 vs. 76.3 Sitting height: 86.9 vs. 86.5 Rectum— Leg length: 76.5 vs. 76 3 Sitting height: 86.7 vs. 86.5 Neither Height, Leg Length, and Cancer Risk » o o O o o O 3 I °~ CO •a S q !5 i 8 2 ^ O) c5 d d O O 5 o U) I CO ^ £s 0} O) I e o a I Q] Ol 1 I S ! 8 S"! J, i? I O) o J IB .•2 1 CO CO 3 •§ 5 o>§ CD 2 -O "DOC 1 5 11 § « X LU CO <D >-• -D <D 3 c =: Ti CO 331 Lastly, case-control studies may be subject to control selection biases. In some studies, control subjects were chosen from patients with other cancers (79, 117, 164, 258) or from persons with other conditions that may themselves be associated with height (46, 251). This bias could work in different directions depending on height-disease associations in the control population. Publication bias. The final type of bias that should be considered is publication bias. The relatively high proportion of positive findings in the studies reviewed here could reflect such an effect. Many of the cohort studies reporting site-specific height-cancer associations have not reported on cancer at other sites. However, a number of studies carried out in larger cohorts—the Whitehall (29), Reykjavik (31), US Physicians (1), and Renfrew and Paisley (2) cohorts, the Norwegian tuberculosis screening cohort (35, 215, 227), and the Nurses' Health Study cohort (32, 91, 205)—have reported associations between height and cancer at a range of sites, and this increases confidence in such findings' not being a result of publication bias alone. Publication bias is more likely to be a concern with cancer sites that have been examined in only one or two articles. An example of this phenomenon is seen with the strong positive association reported in the one case-control study examining the relation between height and vaginal cancer (249). Only one (233) of 12 case-control studies (19) investigating the etiology of thyroid cancer presented height-cancer associations when results were published; however, a recent meta-analysis of these studies (19), using data obtained from the investigators, confirmed that height was associated with increased risk in most of the studies but the association had not been reported. Thus, chance and bias appear unlikely to explain reported height-cancer associations. Below, we outline possible explanations for the observed relation. The explanations can be divided into two basic categories: 1) height as a biomarker for other exposures that influence cancer risk and 2) height as a biomarker for biologic mediators of risk. 51 §11 Previously identified confounders O CO _0J £ io E I 5 §f o2 5 ±= M n §!„ I 8 S •». I i SI f I (2 t _© CD CO O> CD Q) IS O O) 5 5i S f _ « = CO O-o co o l si Epidemiol Rev Vol. 23, No. 2, 2001 0} Ol ® •£ "D -D T3 TJ -= w J2 ^ ^ ^ S"Isss OS c c H c —•.£ o o o o OU-OOOO The roles of several possible confounding factors, particularly body mass index, socioeconomic position, and smoking, have been investigated in the studies included in this review. Controlling for these factors in multivariable analyses only partially attenuates the height-cancer associations reported, or even strengthens them (1, 2, 29, 31, 33, 60, 91, 98, 99, 128). While it is possible that residual confounding remains in some of these analyses as a result of the relatively crude adjustments made, in most cases the attenuation of risk is slight (or nonexistent), indicating a residual height "effect." Socioeconomic position, like height, is a marker for a range of exposures, some of which may influence height. Interpretation of the effects of controlling for socioeconomic position in models examining height-cancer associations is therefore not straightforward. Some of the observed associations may arise as a result of failure to control for other important confounders, and in some analyses there is little assessment of possible confound- 9. Results of case-control studies reporting on the associations between components of height (leg length and sitting height) and cancer Estimate of relative risk Case-control study population (reference no.) No. of cases (incident cases include deaths) Height comparison and source of controls Compon Age-adjusted risk (95% Cl») mortem study (272) Children who died of leukemia compared with "normal" infants (275 children; not all had sitting height measured) Height and sitting height Authors reported in cases and controls statistically significant deviations in the "distribution of the group according to standing and sitting heights" (272); reanalysis of original data indicated that cases had shorter leg length relative to sitting height otherapy center patients 49) Incident cases of breast cancer (n= 150), cervical cancer (n = 60), and other cancers (n = 50) Mean stature in cases Controls were taller and and controls (other had greater sitting hospitalized patients heights than cases (n = 50) and women The relative sitting height undergoing screening of breast cancer cases was less than for tuberculosis (n = that of all other 199)) groups}: Incident cases of prostate cancer (n= 14) Cases vs. controls— Mean stature in cases and controls (other Height: 174.9 cm vs. 177.8 cm urology clinic patients Sitting height: 85.9 (n=14)) cm vs. 81.7 cm Relative sitting height§: 0.49 vs. 0.46 ogy clinic patients (84) ogy clinic patients (86) Incident cases of prostate cancer (n= 156) 1) Mean stature in cases and controls 2) Increase in odds of cancer per 1interquartile- range increase in stature Controls: other urology clinic patients (n 155) Cases vs. controls— Height: 175.5 cm vs. 175.4 cm Relative sitting height):: 0.52 vs. 0.52 st screening patients (161) Incident cases of breast cancer (n = 67) Mean stature in cases and controls (n = 59, other screening center patients) Cases vs. controls— Height: 159.8 cm vs. 159.1 cm Head to pubis: 81.5 cm vs. 80.7 cm Pubis to ground: 78.0 cm vs. 78.4 cm Statistical significance Fully adjusted ,g5"^ c ) . statistical significance s a t u n 3 as cancer Sitting p < 0.01 Breast cancer patients had lower sitting heights than controls! Leg len p < 0.001 Sitting p = 0.14 p = 0.06 p = 0.015 Relative sitting height: 0.96H (95% Cl: 0.71, 1.30) Neither Neither NS» NS NS Height, Leg Length, and Cancer Risk 333 ing structures. For example, the increased risk of esophageal cancer among shorter individuals (29, 215) may be confounded by smoking. Below, we describe other factors that may confound or explain height-cancer associations. Height as a biomarker for genotype or environmental exposures q d o II II 5 a £« O * s f2 s O1.E O i?5 — <o ^ , .!_• > in co ~ CM" "O § I •g,d cO-sdd ro-.e « ! = •-. • = ; 'o a CO o » IS ^ 5? a^ci in ^- in if £ co" § JJ 1 Eo oii- '.fe !A 'S "? i •c d < en v < S.S o o "-p c P "S >.2 d 2 r <D CM o "* oio) di ^ V ' m A •= CO loss 's o if 2 2 5 in SI la 51 ^ 8| 1 I§ {I O a) g) S « "5 w . m.° f t5 « S a . co 6 * •= S s g 8 E I Epidemiol Rev Vol. 23, No. 2, 2001 _- o o o o o o OOZIOOo • ^--H•«n5=^t * Height per se clearly does not cause cancer. In the associations reported here, it is simply acting as a biomarker for some other exposure(s). Thus, strictly speaking, the heightcancer associations could be considered confounded by these factors; however, since these factors are as yet undetermined and we have ruled out confounding by social class, smoking, and body mass index (see above), height may be considered a biomarker for this (these) unknown etiologic factor(s). To assess possible candidates for these exposures, it is important to first consider genetic and environmental influences on growth and hence final adult height and their relation to the etiology of cancer. Genetic influences on height. Twin studies highlight genetic influences on stature. Correlations between the adult heights of monozygotic twins are approximately 0.90, as compared with 0.50 for dizygotic twins (278, 279). These differences probably exaggerate the genetic contribution to height, since in-utero (environmental) influences on growth and pregnancy outcome differ for monozygotic twins and dizygotic twins (280). The particular influence of the intrauterine environment on later growth is highlighted by studies demonstrating height differences of over 1 cm among 6-year-old monozygotic twins who were discordant for birth weight (278). The relative contribution of genetic and environmental factors to adult height depends on the degree of environmental adversity experienced during growth. A twin study from Finland highlights this by showing that the genetic contribution to adult height differences increased in the first half of the 20th century, presumably as a result of environmental improvements' reducing this source of variation (279). Likewise, parent-child correlations for stature are higher in well-nourished nations than in poorly nourished nations; in the latter countries, environmental adversity may override genetic influences (281). Growth and final adult height are influenced by a number of genes. It is likely that prepubertal growth, the timing of puberty, and peak height velocity are under separate genetic control. The range of genes involved in growth is highlighted by the number of specific genes involved in growth disorders. These include the GH-1 (282), SHOX (283), FMR1 (284), and GHRHR (285) genes. While it is possible that a gene which is important in growth is closely linked to one that influences cancer risk, height-cancer associations have been reported in dizygotic twins who are discordant for adult height (273). To fully assess this issue, however, studies of cancer risk in relation to height in monozygotic twins would be required; to date, no such studies seem to have been conducted. Environmental influences on height. The three main environmental factors influencing childhood growth are 334 Gunnell et al. diet, ill health, and psychological well-being (286). Some of these factors may have long-term influences on cancer risk. It has long been recognized that animals on calorierestricted diets are smaller, live longer, and have a reduced incidence of cancer (287, 288). A recent analysis suggests that in humans, a lower calorie intake in childhood is also associated with reduced cancer risk (289). Therefore, it is possible that stature acts as a biomarker either for diets which protect against cancer (short stature) or diets which are associated with an increased risk of cancer (tallness). The importance of prepubertal diet and growth is supported by the observation that height-cancer associations appear to be strongest among birth cohorts exposed to periods of food shortage during the prepubertal growth period (97, 290). Likewise, the stronger associations with leg length suggest that nutritional status before puberty may be most important, since leg length is the component of stature responsible for the greatest amount of prepubertal growth and thus may be a more sensitive marker of nutritional influences at that time. A number of pathogens are known to be important in the etiology of cancer (291), and prolonged illness with such infections in childhood may lead to stunting (292, 293). Infection with Helicobacter pylori may underlie the increased risk of stomach cancer seen among shorter individuals in some studies (2, 217, 220), since it is important in the etiology of stomach cancer and is associated with poor childhood growth (294, 295). For most cancers, however, risk increases with stature. Greater height may reflect a lower infection load in childhood, and this may have two consequences in relation to cancer risk. First, it may leave older children susceptible to infections that, if experienced in later life rather than in early childhood, carry a greater risk of malignancy. Second, it may lead to a more generalized underdevelopment of immune function. In support of these notions, it has been shown that risk of lymphoma is increased among individuals from small families (2, 296); small family size is, in turn, associated with greater stature. Psychological stress in childhood may influence growth, leading to short stature (297, 298). However, it is unlikely that this could underlie the association between greater stature and cancer risk, unless childhood stress protects against later cancer risk. Height as a biomarker of exposures in utero, childhood growth, and the timing of puberty Associations between birth weight and the risks of breast and prostate cancer have been reported in some studies (299, 300). These associations may reflect the influence of fetal nutrition, pregnancy steroids, or maternal growth factors on long-term cancer risk (301). Since birth weight is associated with height later in life (correlations between birth weight and adult height of 0.22-0.26 have been reported (302, 303)), it is possible that the relation between height and cancer may simply reflect the long-term effects of prenatal exposures on cancer risk. While none of the studies reporting height-cancer associations have controlled for birth weight, the observation that leg length is the component of height most consistently associated with cancer risk and the fact that associations of birth weight with leg length and trunk length are similar (304) suggest that different pathways may underlie birth weight-cancer and height-cancer associations. Adult height is also influenced by the timing of puberty, and this too may affect cancer risk. Early menarche is a recognized risk factor for breast cancer (91), and recently it has been suggested that early attainment of adult height is independently associated with increased breast cancer risk (170). These risk factors may indicate either a longer exposure of sensitive tissues to sex hormones or an earlier exposure. Early menarche also signals an earlier cessation of growth, since the sex hormone surges associated with puberty result in fusion of the epiphyseal growth plates (286). Genetic and environmental influences affect the timing of puberty, and while there is some inconsistency in the evidence, most studies indicate that early puberty is associated with short adult stature (302, 305, 306). On this basis alone, one would predict that short stature would be associated with increased breast cancer risk, whereas the opposite is seen (see table 3). Therefore, the height-breast cancer associations are unlikely to arise as a result of the timing of puberty. Furthermore, controlling for age at menarche appears to have little or no effect on height-breast cancer associations (91, 101, 128). Height as a biomarker for biologic mediators of risk The explanations given above are all based on the assumption that height is simply acting as a marker for another exposure. In the two possible mechanisms discussed below, body size is more directly related to cancer risk. Of course, because they may influence total body size, all of the mechanisms reviewed above may be relevant to the pathways discussed here. The first pathway links height to the number of cells in the body. Larger bodies contain more cells than smaller bodies, and it has been suggested that the more cells one has the greater the chance that a cell will undergo malignant transformation, escape the body's cancer defense mechanisms, and progress to cancer (307). Calorie restriction in early life reduces organ cellularity (308), and it has been proposed that this mechanism might underlie height-cancer associations. Some support for this notion comes from studies demonstrating associations between breast size and cancer risk (128). The second pathway links body size to circulating levels of hormones or other substances that influence cancer risk. Two groups of hormones have been investigated: sex hormones and growth hormones. However, no consistent associations have been observed between sex hormone levels and height (309-311) or cancer risk (312, 313). More recently, IGFs have been implicated as having a role in the development of cancer, and height may act as a marker for levels of these growth factors. Evidence supporting the role of IGFs in explaining height-cancer associations is outlined below. Epidemiol Rev Vol. 23, No. 2, 2001 Height, Leg Length, and Cancer Risk IGFs, height, and cancer IGF biology. IGFs, particularly IGF-I, play a fundamental role in somatic growth. Cross-sectional studies of children (314, 315) demonstrate that IGF-I levels are associated with stature: Differences of over 20 percent have been observed between the lowest and highest height quartiles (314). IGF-I is mainly secreted from the liver in response to secretion of pituitary growth hormone, although dietary intake also influences IGF-I levels and some IGF-I synthesis occurs in peripheral tissues. The availability of IGFs to tissues depends on circulating levels of IGF-binding proteins. The principal binding protein for IGF-I is IGF-binding protein 3 (316). Epidemiologic evidence. Recent research based on prospective studies has demonstrated that raised levels of IGF-I or low levels of IGF-binding protein 3 are associated with increased risks of prostate cancer (12-14), premenopausal breast cancer (15), and colorectal cancer (16, 17) over prolonged periods of follow-up. In these studies, blood samples were collected many years before the onset of cancer, so it is likely that the raised levels predated the cancers investigated rather than arose as a result of them. It is noteworthy that height-cancer associations are found for these same cancers (tables 1-3). Associations between IGFI and lung cancer risk have also been reported (317). Because this finding was restricted to a case-control study in which samples were taken from people with diagnosed lung cancer, the direction of causality is uncertain. Studies of cancer risk in acromegaly, a condition characterized by high circulating levels of growth hormone (and hence IGF-I), provide further support for the role of the growth hormone-IGF axis in cancer risk (318-320). While findings were inconsistent, results from two of the largest prospective studies of people with acromegaly (319, 320) suggested increased risks of malignant melanoma and of gastrointestinal, breast, and hematopoietic cancers. Further evidence that IGFs may underlie height-cancer associations comes from analyses that have reported on cancer risk in relation to height and its components, leg length and trunk length. Some studies have indicated that leg length is the constituent of height most strongly related to risk of malignancy. In individuals with Laron's syndrome, an inherited condition characterized by low circulating levels of IGF-I, leg length tends to be low in relation to overall height (321, 322). Thus, it is possible that in the normal population, relative leg length provides an indication of an individual's circulating levels of IGF and his or her cancer risk. Biologic mechanisms linking IGFs to cancer risk. There are several plausible biologic mechanisms through which IGFs may influence cancer risk (323-325). First, IGFs protect damaged cells from apoptosis (cell suicide); hence, higher levels of IGF may enable damaged cells to survive and become cancerous. Second, IGFs are potent stimulators of cell turnover, and the mitotic rate of stem cells is an important determinant of cancer risk. In primate models, growth hormone and IGF-I both stimulate glandular and epithelial cell proliferation, which indicates their potential to initiate and promote neoplasia (326). Lastly, in vitro studEpidemiol Rev Vol. 23, No. 2, 2001 335 ies show that raised levels of IGF-I may amplify the effects of DNA-damaging agents (327). Observational and experimental evidence attests to the important role of nutrition, including intrauterine nutrition, in regulating IGF-I (328, 329), with high-energy diets increasing circulating levels and energy restriction decreasing circulating levels. Thus, IGFs could provide a mechanism explaining associations between nutrition and cancer in addition to a mechanism underlying the association between stature and cancer. It is possible that IGF-I levels in adulthood are "programmed" by nutrition in early life, and such programming may be responsible for associations between early life exposures and adult disease. While there is some attractive evidence in support of a role for IGFs in explaining height-cancer associations, there are some inconsistencies in the research record. In studies of children receiving growth hormone for growth disorders, for example, neither growth hormone deficiency nor growth hormone treatment appears to lead to marked disproportion in leg length relative to height (330). Furthermore, the strong associations between height and IGF-I reported for children (314, 315) are weaker in adults (14, 331) or are not found in adults (12, 332). However, it is possible that adult IGF levels reflect patterns of childhood and adolescent growth, such as growth velocity and age at peak height velocity, that are only weakly related to final adult height. CONCLUSIONS A large body of literature supports the existence of a real, albeit relatively weak, association between height and cancer risk. Taller individuals appear to be at a 20-60 percent increased risk of a range of cancers. This relation may be explained by genes or by prenatal or childhood exposures, including diet and infection. Further research is required to determine whether the influences of childhood diet on growth and circulating growth factor levels underlie the observed associations. Observations that height-cancer associations differ depending on an individual's weight require further detailed review, and mechanisms for observed interactions should be investigated. The finding in some studies that leg length is the component of height that is generating the height-cancer associations requires replication. Furthermore, the differential association of growth factor levels with leg length and trunk length should be investigated. Such research might provide insight into periods of life that are critical for the development of cancer risk. From the population perspective, one must remember that while greater stature is associated with increased cancer mortality, it has long been recognized that taller individuals generally have lower overall mortality rates than shorter people (333, 334). This is largely because associations between height and cardiorespiratory disease point in a direction opposite that of associations with cancer (2, 29). However, understanding the biologic mechanisms through which height and cancer risk are linked may lead both to possible interventions and to an appreciation of critical times in the development of cancer risk in childhood and adolescence. 336 Gunnell et al. REFERENCES 1. Hebert PR, Ajani U, Cook NR, et al. Adult height and incidence of cancer in male physicians (United States). Cancer Causes Control 1997;8:591-7. 2. Davey Smith G, Hart C, Upton M, et al. Height and risk of death among men and women: aetiological implications of associations with cardiorespiratory disease and cancer mortality. J Epidemiol Community Health 2000;54:97-103. 3. Barker DJ, Osmond C, Golding J. Height and mortality in the counties of England and Wales. Ann Hum Biol 1990;17:l-6. 4. Albanes D, Taylor PR. International differences in body height and weight and their relationship to cancer incidence. Nutr Cancer 1990;14:69-77. 5. Baanders AN, De Waard F. Breast cancer in Europe: the importance of factors operating at an early age. Eur J Cancer Prev 1992; 1:285-91. [Published erratum appears in Eur J Cancer Prev 1993;2:189]. 6. Hunter DJ, Willett WC. Diet, body build, and breast cancer. Annu Rev Nutr 1994;14:393-418. 7. Cold S, Hansen S, Overvad K, et al. A woman's build and the risk of breast cancer. Eur J Cancer 1998;34:1163-74. 8. Howe G, Hirohata T, Hislop T, et al. Dietary factors and risk of breast cancer: combined analysis of 12 case-control studies. J Natl Cancer Inst 1990;82:561-9. 9. Perera FP. Environment and cancer: who are susceptible? Science 1997 ;278:1068-73. 10. Weinberg RA. How cancer arises. Sci Am 1996;September: 32-40. 11. Folkman J, Hahnfeldt P, Hlatky L. Cancer: looking outside the genome. Nat Rev Mol Cell Biol 2000; 1:76-9. 12. Chan JM, Stampfer MJ, Giovannucci E, et al. Plasma insulin-like growth factor-I and prostate cancer risk: a prospective study. Science 1998;279:563-6. 13. Harman SM, Metter EJ, Blackman MR, et al. Serum levels of insulin-like growth factor I (IGF-I), IGF-U, IGF-binding protein-3, and prostate-specific antigen as predictors of clinical prostate cancer. J Clin Endocrinol Metab 2000;85:4258-65. 14. Stattin P, Bylund A, Rinaldi S, et al. Plasma insulin-like growth factor-I, insulin-like growth factor-binding proteins, and prostate cancer risk: a prospective study. J Natl Cancer Inst2000;92:1910-17. 15. Hankinson SE, Willett WC, Colditz GA, et al. Circulating concentrations of insulin-like growth factor-I and risk of breast cancer. Lancet 1998;351:1393-6. 16. Ma J, Pollak MN, Giovannucci E, et al. Prospective study of colorectal cancer risk in men and plasma levels of insulinlike growth factor (IGF)-l and IGF-binding protein-3. J Natl Cancer Inst 1999,91:620-5. 17. Giovannucci E, Pollak MN, Platz EA, et al. Plasma insulinlike growth factor-1 and binding protein-3 and risk of colorectal neoplasia in women. Cancer Epidemiol Biomarkers Prev 2000;9:345-9. 18. Howe GR, Aronson KJ, Benito E, et al. The relationship between dietary fat intake and risk of colorectal cancer: evidence from the combined analysis of 13 case-control studies. Cancer Causes Control 1997;8:215-28. 19. Dal Maso L, La Vecchia C, Franceschi S, et al. A pooled analysis of thyroid cancer studies. V. Anthropometric factors. Cancer Causes Control 2000; 11:137^14. 20. Le Marchand L. Anthropometry, body composition, and cancer. In: Micozzi MS, Moon TE, eds. Macronutrients. New York, NY: Marcel Dekker, Inc, 1992:321^2. 21. Gerver WJ, Bruin RD. Relationship between height, sitting height and subischial leg length in Dutch children: presentation of normal values. Acta Paediatr 1995;84:532-5. 22. Gunnell DJ, Davey Smith G, Holly JM, et al. Leg length and risk of cancer in the Boyd Orr cohort. BMJ 1998;317: 1350-1. 23. Gunnell DJ, Davey Smith G, Frankel SJ, et al. Socioeconomic and dietary influences on leg length and trunk length in childhood: a reanalysis of the Carnegie (Boyd Orr) survey of diet and health in prewar Britain (1937-39). Paediatr Perinat Epidemiol 1998;12:96-113. 24. Albanes D, Jones DY, Schatzkin A, et al. Adult stature and risk of cancer. Cancer Res 1988;48:1658-62. 25. Whittemore AS, Paffenbarger RS Jr, Anderson K, et al. Early precursors of site-specific cancers in college men and women. J Natl Cancer Inst 1985,74:43-51. 26. Chyou H, Nomura AM, Stemmermann GN. A prospective study of weight, body mass index and other anthropometric measurements in relation to site-specific cancers. Int J Cancer 1994;57:313-17. 27. Le Marchand L, Wilkens LR, Mi MP. Obesity in youth and middle age and risk of colorectal cancer in men. Cancer Causes Control 1992,3:349-54. 28. Nomura A, Heilbrun LK, Stemmermann GN. Body height and lung cancer risk. (Letter). Lancet 1983;1:1162. 29. Leon DA, Davey Smith G, Shipley M, et al. Adult height and mortality in London: early life, socioeconomic confounding, or shrinkage? J Epidemiol Community Health 1995;49:5-9. 30. Giovannucci E, Ascherio A, Rimm EB, et al. Physical activity, obesity, and risk for colon cancer and adenoma in men. Ann Intern Med 1995;122:327-34. 31. Tulinius H, Sigfusson N, Sigvaldason H, et al. Risk factors for malignant diseases: a cohort study on a population of 22,946 Icelanders. Cancer Epidemiol Biomarkers Prev 1997; 6:863-73. 32. Chute CG, Willett WC, Colditz GA, et al. A prospective study of body mass, height, and smoking on the risk of colorectal cancer in women. Cancer Causes Control 1991;2:117-24. 33. Suadicani P, Hein HO, Gyntelberg F. Height, weight, and risk of colorectal cancer. Scand J Gastroenterol 1993;28:285-8. 34. Bostick RM, Potter JD, Kushi LH, et al. Sugar, meat, and fat intake, and non-dietary risk factors for colon cancer incidence in Iowa women (United States). Cancer Causes Control 1994; 5:38-52. 35. Robsahm TE, Tretli S. Height, weight and gastrointestinal cancer: a follow-up study in Norway. Eur J Cancer Prev 1999;8:105-13. 36. Kaaks R, Toniolo P, Akhmedkhanov A, et al. Serum C-peptide, insulin-like growth factor (IGF)-I, IGF-binding proteins, and colorectal cancer risk in women. J Natl Cancer Inst 2OO0;92:1592-600. 37. Van Wayenburg CA, Van der Schouw YT, Van Noord PA, et al. Age at menopause, body mass index, and the risk of colorectal cancer mortality in the Dutch Diagnostisch Onderzoek Mammacarcinoom (DOM) cohort. Epidemiology 2000; 11:304-8. 38. Ghadirian P, Maisonneuve P, Perret C, et al. Epidemiology of sociodemographic characteristics, lifestyle, medical history, and colon cancer: a case-control study among French Canadians in Montreal. Cancer Detect Prev 1998;22:396-404. 39. Whittemore AS, Wu-Williams AH, Lee M, et al. Diet, physical activity, and colorectal cancer among Chinese in North America and China. J Natl Cancer Inst 1990;82:915-26. 40. La Vecchia C, Negri E, Parazzini F, et al. Height and cancer risk in a network of case-control studies from northern Italy. Int J Cancer 1990;45:275-9. 41. Dietz AT, Newcomb PA, Marcus PM, et al. The association of body size and large bowel cancer risk in Wisconsin (United States) women. Cancer Causes Control 1995;6:30-6. 42. Caan BJ, Coates AO, Slattery ML, et al. Body size and the risk of colon cancer in a large case-control study. Int J Obes 1998;22:178-84. 43. Manousos O, Souglakos J, Bosetti C, et al. IGF-I and IGF-II in relation to colorectal cancer. Int J Cancer 1999;83:15-17. 44. Papadimitriou C, Day N, Tzonou A, et al. Biosocial correlates of colorectal cancer in Greece. Int J Epidemiol 1984;13: 155-9. 45. Dales LG, Friedman GD, Ury HK, et al. A case-control study of relationships of diet and other traits to colorectal cancer in American blacks. Am J Epidemiol 1978;109:132-44. 46. Higginson J. Etiological factors in gastrointestinal cancer in man. J Natl Cancer Inst 1966;37:527^t5. 47. Weiss NS, Daling JR, Chow WH. Incidence of cancer of the Epidemiol Rev Vol. 23, No. 2, 2001 Height, Leg Length, and Cancer Risk 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. large bowel in women in relation to reproductive and hormonal factors. J Natl Cancer Inst 1981;67:57-60. Russo A, Franceschi S, La Vecchia C, et al. Body size and colorectal-cancer risk. Int J Cancer 1998;78:161-5. Freudenheim JL, Graham S, Marshall JR, et al. A casecontrol study of diet and rectal cancer in western New York. Am J Epidemiol 1990;131:612-24. Jain M, Cook GM, Davis FG, et al. A case-control study of diet and colo-rectal cancer. Int J Cancer 1980;26:757-68. Le Marchand L, Wilkens LR, Kolonel LN, et al. Associations of sedentary lifestyle, obesity, smoking, alcohol use, and diabetes with the risk of colorectal cancer. Cancer Res 1997; 57:4787-94. Peters RK, Garabrant DH, Yu MC, et al. A case-control study of occupational and dietary factors in colorectal cancer in young men by subsite. Cancer Res 1989;49:5459-68. Greenwald P, Damon A, Kirmss V, et al. Physical and demographic features of men before developing cancer of the prostate. J Natl Cancer Inst 1974;53:341-6. Le Marchand L, Kolonel LN, Wilkens LR, et al. Animal fat consumption and prostate cancer: a prospective study in Hawaii. Epidemiology 1994;5:276-82. Clarke G, Whittemore AS. Prostate cancer risk in relation to anthropometry and physical activity: the National Health and Nutrition Examination Survey I Epidemiological Follow-up Study. Cancer Epidemiol Biomarkers Prev 2000;9:875-81. Giovannucci E, Rimm EB, Stampfer MJ, et al. Height, body weight, and risk of prostate cancer. Cancer Epidemiol Biomarkers Prev 1997;6:557-63. Cerhan JR, Torner JC, Lynch CF, et al. Association of smoking, body mass, and physical activity with risk of prostate cancer in the Iowa 65+ Rural Health Study (United States). Cancer Causes Control 1997;8:229-38. Veierod MB, Laake P, Thelle DS. Dietary fat intake and risk of prostate cancer: a prospective study of 25,708 Norwegian men. Int J Cancer 1997;73:634-8. Andersson SO, Wolk A, Bergstrom R, et al. Body size and prostate cancer: a 20-year follow-up study among 135,006 Swedish construction workers. J Natl Cancer Inst 1997;89: 385-9. Lund Nilsen TI, Vatten LJ. Anthropometry and prostate cancer risk: a prospective study of 22,248 Norwegian men. Cancer Causes Control 1999; 10:269-75. Schuurman AG, Goldbohm A, Dorant E, et al. Anthropometry in relation to prostate cancer risk in the Netherlands Cohort Study. Am J Epidemiol 2000;151:541-9. Habel LA, Van den Eeden SK, Friedman GD. Body size, age at shaving initiation, and prostate cancer in a large, multiracial cohort. Prostate 2000;43:136-43. Putnam SD, Cerhan JR, Parker AS, et al. Lifestyle and anthropometric risk factors for prostate cancer in a cohort of Iowa men. Ann Epidemiol 2000;10:361-9. Thune I, Lund E. Physical activity and the risk of prostate and testicular cancer: a cohort study of 53,000 Norwegian men. Cancer Causes Control 1994;5:549-56. Hayes RB, Ziegler RG, Gridley G, et al. Dietary factors and risks for prostate cancer among blacks and whites in the United States. Cancer Epidemiol Biomarkers Prev 1999;8:25-34. Signorello LB, Tzonou A, Mantzoros CS, et al. Serum steroids in relation to prostate cancer risk in a case-control study (Greece). Cancer Causes Control 1997;8:632-6. Hayes RB, De Jong FH, Raatgever J, et al. Physical characteristics and factors related to sexual development and behaviour and the risk for prostatic cancer. Eur J Cancer Prev 1992; 1:239^5. Andersson SO, Baron J, Wolk A, et al. Early life risk factors for prostate cancer: a population-based case-control study in Sweden. Cancer Epidemiol Biomarkers Prev 1995;4:187-92. Andersson SO, Baron J, Bergstrom R, et al. Lifestyle factors and prostate cancer risk: a case-control study in Sweden. Cancer Epidemiol Biomarkers Prev 1996;5:509-13. Fincham SM, Hill GB, Hanson J, et al. Epidemiology of prostatic cancer: a case-control study. Prostate 1990,17:189-206. Epidemiol Rev Vol. 23, No. 2, 2001 337 71. Norrish AE, McRae CU, Holdaway IM, et al. Height-related risk factors for prostate cancer. Br J Cancer 2000;82:241-5. 72. Furuya Y, Akimoto S, Akakura K, et al. Smoking and obesity in relation to the etiology and disease progression of prostate cancer in Japan. Int J Urol 1998;5:134-7. 73. Hsing AW, Deng J, Sesterhenn IA, et al. Body size and prostate cancer: a population-based case-control study in China. Cancer Epidemiol Biomarkers Prev 2000;9:1335-41. 74. Walker AR, Walker BF, Tsotetsi NG, et al. Case-control study of prostate cancer in black patients in Soweto, South Africa. Br J Cancer 1992;65:438^tl. 75. Villeneuve PJ, Johnson KC, Kreiger N, et al. Risk factors for prostate cancer: results from the Canadian National Enhanced Cancer Surveillance System. The Canadian Cancer Registries Epidemiology Research Group. Cancer Causes Control 1999;10:355-67. 76. Whittemore AS, Kolonel LN, Wu AH, et al. Prostate cancer in relation to diet, physical activity, and body size in blacks, whites, and Asians in the United States and Canada. J Natl Cancer Inst 1995;87:652-61. 77. Kolonel LN, Yoshizawa CN, Hankin JH. Diet and prostatic cancer: a case-control study in Hawaii. Am J Epidemiol 1988;127:999-1012. 78. Koppel M, Meranze DR, Shimkin MB. Characteristics of patients with prostatic carcinoma: a control case study on 83 autopsy pairs. J Urol 1967;98:229-33. 79. Wynder EL, Mabuchi K, Whitmore WF Jr. Epidemiology of cancer of the prostate. Cancer 1971;28:344-60. 80. Talamini R, La Vecchia C, Decarli A, et al. Nutrition, social factors and prostatic cancer in a Northern Italian population. Br J Cancer 1986;53:817-21. 81. Ross RK, Shimizu H, Paganini-Hill A, et al. Case-control studies of prostate cancer in blacks and whites in southern California. J Natl Cancer Inst 1987 ;78:869-74. 82. Honda GD, Bernstein L, Ross RK, et al. Vasectomy, cigarette smoking, and age at first sexual intercourse as risk factors for prostate cancer in middle-aged men. Br J Cancer 1988; 57:326-31. 83. Mettlin C, Selenskas S, Natarajan N, et al. Beta-carotene and animal fats and their relationship to prostate cancer risk: a case-control study. Cancer 1989;64:605-12. 84. Demark-Wahnefried W, Paulson DF, Robertson CN, et al. Body dimension differences in men with or without prostate cancer. J Natl Cancer Inst 1992;84:1363^1. 85. Mbakop A, Angwafo FF, Tsingain K. Prostatic tumors (adenoma, adenocarcinoma) and obesity in Cameroon. (In French). Bull Cancer 1995;82:384-5. 86. Demark-Wahnefried W, Conaway MR, Robertson CN, et al. Anthropometric risk factors for prostate cancer. Nutr Cancer 1997;28:302-7. 87. Key TJ, Silcocks PB, Davey GK, et al. A case-control study of diet and prostate cancer. Br J Cancer 1997;76:678-87. 88. Hsieh CC, Thanos A, Mitropoulos D, et al. Risk factors for prostate cancer: a case-control study in Greece. Int J Cancer 1999;80:699-703. 89. Freni SC, Eberhardt MS, Turturro A, et al. Anthropometric measures and metabolic rate in association with risk of breast cancer (United States). Cancer Causes Control 1996;7:358-65. 90. Swanson CA, Jones DY, Schatzkin A, et al. Breast cancer risk assessed by anthropometry in the NHANES I Epidemiological Follow-up Study. Cancer Res 1988;48:5363-7. 91. Berkey CS, Frazier AL, Gardner JD, et al. Adolescence and breast carcinoma risk. Cancer 1999;85:2400-9. 92. London SJ, Colditz GA, Stampfer MJ, et al. Prospective study of relative weight, height, and risk of breast cancer. JAMA 1989;262:2853-8. 93. Galanis DJ, Kolonel LN, Lee J, et al. Anthropometric predictors of breast cancer incidence and survival in a multiethnic cohort of female residents of Hawaii, United States. Cancer Causes Control 1998;9:217-24. 94. Le Marchand L, Kolonel LN, Earle ME, et al. Body size at different periods of life and breast cancer risk. Am J Epidemiol 1988; 128:137-52. 338 Gunnel! et al. 95. Den Tonkelaar I, Seidell JC, Collette HJ. Body fat distribution in relation to breast cancer in women participating in the DOM-project. Breast Cancer Res Treat 1995;34:55-61. 96. Kaaks R, Van Noord PA, Den Tonkelaar I, et al. Breastcancer incidence in relation to height, weight and body-fat distribution in the Dutch "DOM" cohort. Int J Cancer 1998; 76:647-51. 97. Vatten LJ, Kvinnsland S. Body height and risk of breast cancer: a prospective study of 23,831 Norwegian women. Br J Cancer 1990;61:881-5. 98. Vatten LJ, Kvinnsland S. Prospective study of height, body mass index and risk of breast cancer. Acta Oncol 1992;31: 195-200. 99. De Stavola BL, Wang DY, Allen DS, et al. The association of height, weight, menstrual and reproductive events with breast cancer: results from two prospective studies on the island of Guernsey (United Kingdom). Cancer Causes Control 1993;4:331-40. 100. Waard FD, Baanders-Van Halewijn EA. A prospective study in general practice breast-cancer risk in postmenopausal women. Int J Cancer 1974; 14:153-60. 101. Van den Brandt PA, Dirx MJ, Ronckers CM, et al. Height, weight, weight change, and postmenopausal breast cancer risk: The Netherlands Cohort Study. Cancer Causes Control 1997;8:39-47. 102. Dubin N, Pasternack BS, Strax P. Epidemiology of breast cancer in a screened population. Cancer Detect Prev 1984; 7:87-102. 103. Tornberg SA, Holm LE, Cartensen JM. Breast cancer risk in relation to serum cholesterol, serum beta lipoprotein, height, weight, and blood pressure. Acta Oncol 1988;27:31-7. 104. Toniolo P, Riboli E, Shore RE, et al. Consumption of meat, animal products, protein, and fat and risk of breast cancer: a prospective cohort study in New York. Epidemiology 1994; 5:391-7. 105. Tretli S. Height and weight in relation to breast cancer morbidity and mortality: a prospective study of 570,000 women in Norway. Int J Cancer 1989;44:23-30. 106. Folsom AR, Kaye SA, Prineas RJ, et al. Increased incidence of carcinoma of the breast associated with abdominal adiposity in postmenopausal women. Am J Epidemiol 1990;131: 794-803. 107. Herrinton L, Husson G. Relationship of childhood height and risk of breast cancer. Am J Epidemiol 2000;151(suppl):S65. 108. Jernstrom H, Barrett-Connor E. Obesity, weight change, fasting insulin, proinsulin, C-peptide, and insulin-like growth factor-1 levels in women with and without breast cancer: The Rancho Bernardo Study. J Womens Health Gend Based Med 1999;8:1265-72. 109. Van den Brandt PA, Spiegelman D, Yaun SS, et al. Pooled analysis of prospective cohort studies on height, weight, and breast cancer risk. Am J Epidemiol 2000;152:514-27. 110. Brinton L, Swanson C. Height and weight at various ages and risk of breast cancer. Ann Epidemiol 1992;2:597-609. 111. Hirose K, Tajima K, Hamajima N, et al. Comparative casereferent study of risk factors among hormone related female cancers in Japan. Jpn J Cancer Res 1999,90:255-61. 112. Swanson C, Brinton L, Taylor P, et al. Body size and breast cancer risk assessed in women participating in the Breast Cancer Detection Demonstration Project. Am J Epidemiol 1989;130:1133^1. 113. Adebamowo CA, Adekunle OO. Case-controlled study of the epidemiological risk factors for breast cancer in Nigeria. Br J Surg 1999;86:665-8. 114. Hsieh CC, Trichopoulos D, Katsouyanni K, et al. Age at menarche, age at menopause, height and obesity as risk factors for breast cancer: associations and interactions in an international case-control study. Int J Cancer 1990;46:796-800. 115. Paffenbarger RS Jr, Kampert JB, Chang HG. Characteristics that predict risk of breast cancer before and after the menopause. Am J Epidemiol 1980;112:258-68. 116. Ravnihar B, MacMahon B, Lindtner J. Epidemiologic features of breast cancer in Slovenia, 1965-1967. Eur J Cancer 1971;7:295-306. 117. Staszewski J. Breast cancer and body build. Prev Med 1977; 6:410-15. 118. Valaoras V, MacMahon B, Trichopoulos D, et al. Lactation and reproductive histories of breast cancer patients in Greater Athens, 1965-67. Int J Cancer 1969;4:350-63. 119. Chie W-C, Chen C, Lee W, et al. Body size and risk of preand post-menopausal breast cancer in Taiwan. Anticancer Res 1996; 16:3129-32. 120. De Waard F, Cornelis J, Aoki K, et al. Breast cancer incidence according to weight and height in two cities of the Netherlands and in Aichi prefecture, Japan. Cancer 1977;40: 1269-75. 121. Fillmore CM, Ziegler R, Hoover R, et al. Height as a determinant of breast cancer risk in a study of Asian-American women. (Abstract). Am J Epidemiol 1997;145(suppl):S47. 122. Hall I, Newman B, Millikan R, et al. Body size and breast cancer risk in black women and white women. Am J Epidemiol 2000; 151:754-64. 123. Kalish L. Relationships of body size with breast cancer. J Clin Oncol 1984;2:287-93. 124. Kodama M, Kodama T, Miura S, et al. Nutrition and breast cancer risk in Japan. Anticancer Res 1991; 11:745-54. 125. Magnusson C, Baron J, Persson I, et al. Body size in different periods of life and breast cancer risk in post-menopausal women. Int J Cancer 1998;76:29-34. 126. McCredie M, Dite G, Giles G, et al. Breast cancer in Australian women under the age of 40. Cancer Causes Control 1998,9:189-98. 127. Ng E, Gao F, Ji C-Y, et al. Risk factors for breast carcinoma in Singaporean Chinese women: the role of central obesity. Cancer 1997;80:725-31. 128. Swanson CA, Coates RJ, Schoenberg JB, et al. Body size and breast cancer risk among women under age 45 years. Am J Epidemiol 1996;143:698-706. 129. Trentham-Dietz A, Newcomb P, Storer B, et al. Body size and risk of breast cancer. Am J Epidemiol 1997;145:1011-19. 130. Ziegler R, Hoover R, Nomura A, et al. Relative weight, weight change, height, and breast cancer risk in AsianAmerican women. J Natl Cancer Inst 1996;88:650-60. 131. Palmer JR, Rosenberg L, Harlap S, et al. Adult height and risk of breast cancer among US black women. Am J Epidemiol 1995;141:845-9. 132. De Waard F. Breast cancer incidence and nutritional status with particular reference to body weight and height. Cancer Res 1975;35:3351-6. 133. Hu Y-H, Nagata C, Shimizu H, et al. Association of body mass index, physical activity and reproductive histories with breast cancer: a case-control study in Gifu, Japan. Breast Cancer Res Treat 1997;43:65-72. 134. Wakai K, Ohno Y, Watanabe S, et al. Risk factors for breast cancer among Japanese women in Tokyo: a case-control study. J Epidemiol 1994;4:65-71. 135. Bouchardy C. Le M, Hill C. Risk factors for breast cancer according to age at diagnosis in a French case-control study. J Clin Epidemiol 1990;43:267-75. 136. Lin T, Chen K, MacMahon B. Epidemiologic characteristics of cancer of the breast in Taiwan. Cancer 1971 ;27:1497-504. 137. Paffenbarger R, Greenberg L, Chang H-G, et al. Epidemiologic characteristics of breast cancer in three menopausal stages: preliminary findings. Natl Cancer Inst Monogr 1979;53:195-202. 138. Tung H, Tsukuma H, Tanaka H, et al. Risk factors for breast cancer in Japan, with special attention to anthropometric measurements and reproductive history. Jpn J Clin Oncol 1999;29:137^6. 139. Bruning P, Bonfrer J, Hart A, et al. Body measurements, estrogen availability and the risk of human breast cancer: a case-control study. Int J Cancer 1992,51:14—19. 140. Ghadirian P, Lacroix A, Perret C, et al. Sociodemographic characteristics, smoking, medical and family history and breast cancer. Cancer Detect Prev 1998;22:485-94. 141. Hislop T, Coldman A, Elwood J, et al. Childhood and recent Epidemiol Rev Vol. 23, No. 2, 2001 Height, Leg Length, and Cancer Risk 142. 143. 144. 145. 146. 147. 148. 149. 150. 151. 152. 153. 154. 155. 156. 157. 158. 159. 160. 161. 162. 163. 164. 165. 166. 167. eating patterns and risk of breast cancer. Cancer Detect Prev 1986;9:47-58. Kolonel L, Nomura A, Lee J, et al. Anthropometric indicators of breast cancer risk in postmenopausal women in Hawaii. Nutr Cancer 1986;8:247-56. Lund E, Adami H-O, Bergstrom R, et al. Anthropometric measures and breast cancer in young women. Cancer Causes Control 1990; 1:169-72. Mannistd S, Pietinen P, Pyy M, et al. Body-size indicators and risk of breast cancer according to menopause and estrogen-receptor status. Int J Cancer 1996;68:8-13. Radimer K, Siskind V, Bain C, et al. Relation between anthropometric indicators and risk of breast cancer among Australian women. Am J Epidemiol 1993;138:77-89. Soini I. Risk factors of breast cancer in Finland. Int J Epidemiol 1977;6:365-73. Brisson J, Morrison AS, Kopans DB, et al. Height and weight, mammographic features of breast tissue, and breast cancer risk. Am J Epidemiol 1984; 119:371-81. Del Giudice ME, Fantus IG, Ezzat S, et al. Insulin and related factors in premenopausal breast cancer risk. Breast Cancer Res Treat 1998;47:111-20. Brinkley D, Carpenter RG, Haybittle JL. An anthropometric study of women with cancer. Br J Prev Soc Med 1971; 25:65-75. Parazzini F, La Vecchia C, Negri E, et al. Anthropometric variables and risk of breast cancer. Int J Cancer 1990;45: 397^02. Reuter K, Baker S, Krolikowski F. Risk factors for breast cancer in women undergoing mammography. Am J Roentgenol 1992;158:273-8. Kocic B, Jankovic S, Petrovic B, et al. The significance of physical activity and anthropometric parameters for the development of breast cancer. Vojnosanit Pregl 1996;53:301—4. Mirra AP, Cole P, MacMahon B. Breast cancer in an area of high parity: Sao Paulo, Brazil. Cancer Res 1971;31:77-83. Choi N, Howe G, Miller A, et al. An epidemiologic study of breast cancer. Am J Epidemiol 1978;1O7:51O-21. Li C, Stanford J, Daling J. Anthropometric variables in relation to risk of breast cancer in middle-aged women. Int J Epidemiol 2000;29:208-13. Burch JD, Howe GR, Miller AB. Breast cancer in relation to weight in women aged 65 years and over. Can Med Assoc J 1981;124:1326, 1328. Chie WC, Li CY, Huang CS, et al. Body size as a factor in different ages and breast cancer risk in Taiwan. Anticancer Res 1998;18:565-70. Franceschi S, Favero A, La Vecchia C, et al. Body size indices and breast cancer risk before and after menopause. Int J Cancer 1996;67:181-6. Kyogoku S, Hirohata T, Takeshita S, et al. Anthropometric indicators of breast cancer risk in Japanese women in Fukuoka. Jpn J Cancer Res 1990;81:731-7. Levshin V. Importance of anthropometric parameters in breast cancer. Vopr Onkol 198O;26:69-73. Mondina R, Borsellino G, Poma S, et al. Breast carcinoma and skeletal formation. Eur J Cancer 1992;28A: 1068-70. Taioli E, Barone J, Wynder E. A case-control study on breast cancer and body mass. Eur J Cancer 1995;31 A:723-8. Tavani A, Braga C, La Vecchia C, et al. Height and breast cancer risk. Eur J Cancer 1998;34:543-7. Wynder E, MacCornack F, Stellman S. The epidemiology of breast cancer in 785 United States Caucasian women. Cancer 1978;41:2341-54. Zhang Y, Rosenberg L, Colton T, et al. Adult height and risk of breast cancer among white women in a case-control study. Am J Epidemiol 1996;143:1123-8. Adami HO, Rimsten A, Stenkvist B, et al. Influence of height, weight and obesity on risk of breast cancer in an unselected Swedish population. Br J Cancer 1977;36:787-92. Chu S, Lee N, Wingo P, et al. The relationship between body mass and breast cancer among women enrolled in the Cancer and Steroid Hormone Study. J Clin Epidemiol 1991 ;44: Epidemiol Rev Vol. 23, No. 2, 2001 339 1197-206. 168. Ewertz M. Influence of non-contraceptive exogenous and endogenous sex hormones on breast cancer risk in Denmark. Int J Cancer 1988;42:832-8. 169. Ingram D, Nottage E, Ng S, et al. Obesity and breast disease: the role of female sex hormones. Cancer 1989;64:1049-53. 170. Li CI, Malone KE, White E, et al. Age when maximum height is reached as a risk factor for breast cancer among young U.S. women. Epidemiology 1997,8:559-65. 171. Lubin F, Ruder A, Wax Y, et al. Overweight and changes in weight throughout adult life in breast cancer etiology: a casecontrol study. Am J Epidemiol 1985; 122:579-88. 172. McCredie M, Paul C, Skegg D, et al. Reproductive factors and breast cancer in New Zealand. Int J Cancer 1998;76: 182-8. 173. McNee R, Mason B, Neave L, et al. Influence of height, weight and obesity on breast cancer incidence and recurrence in Auckland, New Zealand. Breast Cancer Res Treat 1987; 9:145-50. 174. Peacock S, White E, Daling J, et al. Relation between obesity and breast cancer in young women. Am J Epidemiol 1999; 149:339-46. 175. Schapira D, Kumar N, Lyman G, et al. Abdominal obesity and breast cancer risk. Ann Intern Med 1990; 112:182-6. 176. Ursin G, Paganini-Hill A, Siemiatycki J, et al. Early adult body weight, body mass index and premenopausal bilateral breast cancer: data from a case-control study. Breast Cancer Res Treat 1994;33:75-82. 177. Le M, Flamant R. Nutrition and carcinoma of the breast: presentation of a French investigation. (In French). Bull Cancer 1978;65:65-9. 178. Atalah E, Urteaga C, Rebolledo A, et al. Breast cancer risk factors in women in Santiago, Chile. (In Spanish). Rev Med Chil 2000; 128:137-43. 179. Varela G, Carbajal A, Nunez C, et al. Influence of energy intake and body mass index in the incidence of cancer of the breast: case-control study in a sample from 3 Spanish hospital populations. (In Spanish). Nutr Hosp 1996;11:54-8. 180. Lam PB, Vacek PM, Geller BM, et al. The association of increased weight, body mass index, and tissue density with the risk of breast carcinoma in Vermont. Cancer 2000;89: 369-75. 181. Thomas E, Cade J, Vail A. Risk factor analysis of data from assessment clinics in the UK Breast Screening Programme: a case-control study in Portsmouth and Southampton. J Epidemiol Community Health 1996;50:144-8. 182. Petridou E, Papadiamantis Y, Markopoulos C, et al. Leptin and insulin growth factor I in relation to breast cancer (Greece). Cancer Causes Control 2000; 11:383-8. 183. Baanders-Van Halewijn EA, Poortman J. A case control study of endometrial cancer within a cohort. Maturitas 1985; 7:69-76. 184. Tretli S, Magnus K. Height and weight in relation to uterine corpus cancer morbidity and mortality: a follow-up study of 570,000 women in Norway. Int J Cancer 1990;46:165-72. 185. De Waard F, De Ridder CM, Baanders-Van Halewijn EA, et al. Endometrial cancer in a cohort screened for breast cancer. Eur J Cancer Prev 1996;5:99-104. 186. Le Marchand L, Wilkens LR, Mi MP. Early-age body size, adult weight gain and endometrial cancer risk. Int J Cancer 1991;48:807-ll. 187. Kalandidi A, Tzonou A, Lipworth L, et al. A case-control study of endometrial cancer in relation to reproductive, somatometric, and life-style variables. Oncology 1996;53:354—9. 188. Koumantaki Y, Tzonou A, Koumantakis E, et al. A casecontrol study of cancer of endometrium in Athens. Int J Cancer 1989;43:795-9. 189. Goodman MT, Hankin JH, Wilkens LR, et al. Diet, body size, physical activity, and the risk of endometrial cancer. Cancer Res 1997;57:5077-85. 190. Ghadirian P, Simard A, Baillargeon J, et al. Nutritional factors and pancreatic cancer in the francophone community in Montreal, Canada. Int J Cancer 1991;47:l-6. 340 Gunnell et al. 191. Brinton LA, Berman ML, Mortel R, et al. Reproductive, menstrual, and medical risk factors for endometrial cancer: results from a case-control study. Am J Obstet Gynecol 1992;167:1317-25. 192. Wynder EL, Escher GC, Mantel N. An epidemiological investigation of cancer of the endometrium. Cancer 1966; 19:489-520. 193. Elwood JM, Cole P, Rothman KJ, et al. Epidemiology of endometrial cancer. J Natl Cancer Inst 1977;59:1055-60. 194. Shu XO, Brinton LA, Zheng W, et al. A population-based case-control study of endometrial cancer in Shanghai, China. Int J Cancer 1991;49:38^3. 195. Ewertz M, Machado SG, Boice JD, et al. Endometrial cancer following treatment for breast cancer: a case-control study in Denmark. Br J Cancer 1984;50:687-92. 196. Weiderpass E, Persson I, Adami HO, et al. Body size in different periods of life, diabetes mellitus, hypertension, and risk of postmenopausal endometrial cancer (Sweden). Cancer Causes Control 2000; 11:185-92. 197. Swanson CA, Potischman N, Wilbanks GD, et al. Relation of endometrial cancer risk to past and contemporary body size and body fat distribution. Cancer Epidemiol Biomarkers Prev 1993,2:321-7. 198. Jensen H. Endometrial carcinoma: a retrospective, epidemiological study. Dan Med Bull 1985;32:219-28. 199. Blitzer PH, Blitzer EC, Rimm AA. Association between teenage obesity and cancer in 56,111 women: all cancers and endometrial carcinoma. Prev Med 1976;5:2O-31. 200. Ewertz M, Schou G, Boice JD Jr. The joint effect of risk factors on endometrial cancer. Eur J Cancer Clin Oncol 1988; 24:189-94. 201. La Vecchia C, Parazzini F, Negri E, et al. Anthropometric indicators of endometrial cancer risk. Eur J Cancer 1991; 27:487-90. 202. Villani C, Pucci G, Pietrangeli D, et al. Role of diet in endometrial cancer patients. Eur J Gynaecol Oncol 1986;7:139-43. 203. Schapira DV, Kumar NB, Lyman GH, et al. Upper-body fat distribution and endometrial cancer risk. JAMA 1991 ;266: 1808-11. 204. Olson SH, Trevisan M, Marshall JR, et al. Body mass index, weight gain, and risk of endometrial cancer. Nutr Cancer 1995;23:141-9. 205. Zhang S, Hunter DJ, Rosner BA, et al. Dietary fat and protein in relation to risk of non-Hodgkin's lymphoma among women. J Natl Cancer Inst 1999; 19:1751-8. 206. Hancock BW, Mosely R, Coup AJ. Height and Hodgkin's disease. (Letter). Lancet 1976;2:1364. 207. Mehes K, Signer E, Pluss HJ, et al. Increased prevalence of minor anomalies in childhood malignancy. Eur J Pediatr 1985; 144:243-54. 208. Bessho F. Height at diagnosis in acute lymphocytic leukaemia. Arch Dis Child 1986;61:296-8. 209. Broomhall J, May R, Lilleyman JS, et al. Height and lymphoblastic leukaemia. Arch Dis Child 1983;58:3OO-l. 210. McWhirter WR, McWhirter KM, Taylor D. Height and lymphoblastic leukemia. (Letter). Arch Dis Child 1983;58:839-40. 211. Robison LL, Nesbit ME Jr, Sather HN, et al. Height of children successfully treated for acute lymphoblastic leukemia: a report from the Late Effects Study Committee of Childrens Cancer Study Group. Med Pediatr Oncol 1985; 13:14-21. 212. Westphal M, Morgan SK, Grush OC, et al. Nutrition and growth in children with acute lymphoblastic leukemia (ALL). Clin Res 1979;27:816A. 213. Oakhill A, Mann JR. Poor prognosis of acute lymphoblastic leukaemia in Asian children living in the United Kingdom. Br Med J (Clin Res Ed) 1983;286:839-41. 214. Berry DH, Elders MJ, Crist W, et al. Growth in children with acute lymphocytic leukemia: a Pediatric Oncology Group study. Med Pediatr Oncol 1983; 11:39-45. 215. Tretli S, Robsahm TE. Height, weight and cancer of the oesophagus and stomach: a follow-up study in Norway. Eur J Cancer Prev 1999;8:115-22. 216. Risch HA, Jain M, Choi NW, et al. Dietary factors and the incidence of cancer of the stomach. Am J Epidemiol 1985; 122:947-59. 217. Hansson LE, Baron J, Nyren O, et al. Early-life risk indicators of gastric cancer: a population-based case-control study in Sweden. Int J Cancer 1994,57:32-7. 218. Ji BT, Chow WH, Yang G, et al. Body mass index and the risk of cancers of the gastric cardia and distal stomach in Shanghai, China. Cancer Epidemiol Biomarkers Prev 1997; 6:481-5. 219. Lagergren J, Bergstrom R, Nyren O. Association between body mass and adenocarcinoma of the esophagus and gastric cardia. Ann Intern Med 1999; 130:883-90. 220. Chow WH, Blot WJ, Vaughan TL, et al. Body mass index and risk of adenocarcinomas of the esophagus and gastric cardia. J Natl Cancer Inst 1998;90:150-5. 221. Helseth A, Tretli S. Pre-morbid height and weight as risk factors for development of central nervous system neoplasms. Neuroepidemiology 1989;8:277-82. 222. Howe GR, Burch JD, Chiarelli AM, et al. An exploratory case-control study of brain tumors in children. Cancer Res 1989;49:4349-52. 223. Davies TW, Prener A, Engholm G. Body size and cancer of the testis. Acta Oncol 1990;29:287-90. 224. Akre O, Ekbom A, Sparen P, et al. Body size and testicular cancer. J Natl Cancer Inst 2000,92:1093-6. 225. UK Testicular Cancer Study Group. Social, behavioural and medical factors in the aetiology of testicular cancer: results from the UK study. Br J Cancer 1994;70:513-20. 226. Gallagher RP, Huchcroft S, Phillips N, et al. Physical activity, medical history, and risk of testicular cancer (Alberta and British Columbia, Canada). Cancer Causes Control 1995;6: 398-406. 227. Thune I, Olsen A, Albrektsen G, et al. Cutaneous malignant melanoma: association with height, weight and body-surface area. A prospective study in Norway. Int J Cancer 1993; 55:551-61. 228. Veierod MB, Thelle DS, Laake P. Diet and risk of cutaneous malignant melanoma: a prospective study of 50,757 Norwegian men and women. Int J Cancer 1997;71:600-4. 229. Yoo KY, Kang D, Koo H-W, et al. Risk factors associated with uterine cervical cancer in Korea: a case-control study with special reference to sexual behavior. J Epidemiol 1997; 7:117-23. 230. Davey Smith G, Shipley M, Leon DA. Height and mortality from cancer among men: prospective observational study. BMJ 1998;317:1351-2. 231. Okasha M, McCarron P, McEwen J, et al. Height and cancer mortality: results from the Glasgow university student cohort. Public Health 2000;114:451-5. 232. Gunnell DJ, Davey Smith G, Frankel S, et al. Childhood leg length and adult mortality: follow up of the Carnegie (Boyd Orr) Survey of Diet and Health in Pre-War Britain. J Epidemiol Community Health 1998;52:142-52. 233. Goodman MT, Kolonel LN, Wilkens LR. The association of body size, reproductive factors and thyroid cancer. Br J Cancer 1992;66:1180-4. 234. Kolonel LN, Hankin JH, Wilkens LR, et al. An epidemiologic study of thyroid cancer in Hawaii. Cancer Causes Control 1990; 1:223-34. 235. D'Avanzo B, La Vecchia C. Risk factors for male breast cancer. Br J Cancer 1995;71:1359-62. 236. Casagrande JT, Hanisch R, Pike MC, et al. A case-control study of male breast cancer. Cancer Res 1988;48:1326-30. 237. Thomas DB, Jimenez LM, McTiernan A, et al. Breast cancer in men: risk factors with hormonal implications. Am J Epidemiol 1992;135:734-48. 238. Mabuchi K, Brass DS, Kessler U. Risk factors for male breast cancer. J Natl Cancer Inst 1985;74:371-5. 239. Shu XO, Gao YT, Yuan JM, et al. Dietary factors and epithelial ovarian cancer. Br J Cancer 1989;59:92-6. 240. Tzonou A, Day NE, Trichopoulos D, et al. The epidemiology of ovarian cancer in Greece: a case-control study. Eur J Epidemiol Rev Vol. 23, No. 2, 2001 Height, Leg Length, and Cancer Risk Cancer Clin Oncol 1984;20:1045-52. 241. Mori M, Nishida T, Sugiyama T, et al. Anthropometric and other risk factors for ovarian cancer in a case-control study. Jpn J Cancer Res 1998;89:246-53. 242. Mori M, Nishimura H, Nishida T, et al. A case-control study of ovarian cancer to identify its risk factors. (In Japanese). Nippon Sanka Fujinka Gakkai Zasshi 1996;48:875-82. 243. Gelberg KH, Fitzgerald EF, Hwang S, et al. Growth and development and other risk factors for osteosarcoma in children and young adults. Int J Epidemiol 1997,26:272-8. 244. Fraumeni JF Jr. Stature and malignant tumors of bone in childhood and adolescence. Cancer 1967;20:967-73. 245. Operskalski EA, Preston-Martin S, Henderson BE, et al. A case-control study of osteosarcoma in young persons. Am J Epidemiol 1987;126:118-26. 246. Brostrom LA, Adamson U, Filipsson R, et al. Longitudinal growth and dental developments in osteosarcoma patients. Acta Orthop Scand 1979;51:755-9. 247. Scranton PE Jr, DeCicco FA, Totten RS, et al. Prognostic factors in osteosarcoma: a review of 20 years' experience at the University of Pittsburgh Health Center Hospitals. Cancer 1975:36:2179-91. 248. Pendergrass TW, Foulkes MA, Robison LL, et al. Stature and Ewing's sarcoma in childhood. Am J Pediatr Hematol Oncol 1984;6:33-9. 249. Sharp GB, Cole P. Identification of risk factors for diethylstilbestrol-associated clear cell adenocarcinoma of the vagina: similarities to endometrial cancer. Am J Epidemiol 1991:134:1316-24. 250. Chen WJ, Chung YC. Energy expenditure in patients with hepatocellular carcinoma. Cancer 1994;73:59O-5. 251. Tavani A, Soler M, La Vecchia C, et al. Body weight and risk of soft-tissue sarcoma. Br J Cancer 1999:81:890-2. 252. Serraino D, Franceschi S, Talamini R, et al. Non-occupational risk factors for adult soft-tissue sarcoma in northern Italy. Cancer Causes Control 1991;2:157-64. 253. Zahm SH, Blair A, Holmes FF, et al. A case-control study of soft-tissue sarcoma. Am J Epidemiol 1989; 130:665-74. 254. Lee J, Kolonel LN. Body height and lung cancer risk. (Letter). Lancet 1983;1:877. 255. Cochrane AL, Moore F. Body height and lung cancer risk. (Letter). Lancet 1983;1:1162. 256. Drinkard CR, Sellers TA, Potter JD, et al. Association of body mass index and body fat distribution with risk of lung cancer in older women. Am J Epidemiol 1995:142:600-7. 257. Goodman MT, Wilkens LR. Relation of body size and the risk of lung cancer. Nutr Cancer 1993;20:179-86. 258. Kabat GC, Wynder EL. Body mass index and lung cancer risk. Am J Epidemiol 1992;135:769-74. 259. Wynder EL, Goodman MT. Body height and lung cancer risk. (Letter). Lancet 1983:1:1162-3. 260. D'Avanzo B, La Vecchia C, Talamini R, et al. Anthropometric measures and risk of cancers of the upper digestive and respiratory tract. Nutr Cancer 1996;26:219-27. 261. Friedman GD, Van den Eeden SK. Risk factors for pancreatic cancer: an exploratory study. Int J Epidemiol 1993:22:30-7. 262. Bueno de Mesquita HB, Maisonneuve P, Moerman CJ, et al. Anthropometric and reproductive variables and exocrine carcinoma of the pancreas: a population-based case-control study in the Netherlands. Int J Cancer 1992;52:24-9. 263. Ji BT, Hatch MC, Chow WH, et al. Anthropometric and reproductive factors and the risk of pancreatic cancer: a case-control study in Shanghai, China. Int J Cancer 1996:66:432-7. 264. Nakata S, Sato J, Ohtake N, et al. Epidemiological study of risk factors for bladder cancer. (In Japanese). Hinyokika Kiyo 1995:41:969-77. 265. Lindblad P, Wolk A, Bergstrom R, et al. The role of obesity and weight fluctuations in the etiology of renal cell cancer: a population-based case-control study. Cancer Epidemiol Biomarkers Prev 1994;3:631-9. 266. Mellemgaard A, Lindblad P, Schlehofer B, et al. International renal-cell cancer study. III. Role of weight, height, physical activity, and use of amphetamines. Int J Cancer 1995;60: Epidemiol Rev Vol. 23, No. 2, 2001 341 350-4. 267. Mellemgaard A, Engholm G, McLaughlin JK, et al. Risk factors for renal-cell carcinoma in Denmark. III. Role of weight, physical activity and reproductive factors. Int J Cancer 1994; 56:66-71. 268. Chow WH, McLaughlin JK, Mandel JS, et al. Obesity and risk of renal cell cancer. Cancer Epidemiol Biomarkers Prev 1996:5:17-21. 269. Pui CH, Dodge RK, George SL, et al. Height at diagnosis of malignancies. Arch Dis Child 1987:62:495-9. 270. La Vecchia C, Decarli A, Negri E, et al. Height and the prevalence of chronic disease. Rev Epidemiol Sante Publique 1992:40:6-14. 271. Severson RK, Grove JS, Nomura AM, et al. Body mass and prostatic cancer: a prospective study. BMJ 1988;297:713-15. 272. Stowens D, Cork MG, Mallardi A. La patologia della leucemia acuta infantile. Minerva Pediatr 1961; 13:865-90. 273. Swerdlow AJ, De Stavola BL, Swanwick MA, et al. Risk factors of testicular cancer: a case-control study in twins. Br J Cancer 1999;80:1098-102. 274. Rosso S, Faggiano F, Zanetti R, et al. Social class and cancer survival in Turin, Italy. J Epidemiol Community Health 1997;51:30-4. 275. Roberts RJ. Can self-reported data accurately describe the prevalence of overweight? Public Health 1995;109:275-84. 276. McCarthy M. Height and risk of cancer: survival selection bias should have been ruled out. (Letter). BMJ 1999:318:1074-5. 277. Gunnell DJ, Davey Smith G, Holly JM, et al. Height and risk of cancer—authors' reply. (Letter). BMJ 1999:318:1075. 278. Wilson RS. Twin growth: initial deficit, recovery, and trends in concordance from birth to nine years. Ann Hum Biol 1979; 6:205-20. 279. Silventoinen K, Kaprio J, Lahelma E, et al. Relative effect of genetic and environmental factors on body height: differences across birth cohorts among Finnish men and women. Am J Public Health 2000:90:627-30. 280. Phillips DI. Twin studies in medical research: can they tell us whether diseases are genetically determined? Lancet 1993; 341:1008-9. 281. Mueller WH. Parent-child correlations for stature and weight among school aged children: a review of 24 studies. Hum Biol 1976;48:379-97. 282. Hasegawa Y, Fujii K, Yamada M, et al. Identification of novel human GH-1 gene polymorphisms that are associated with growth hormone secretion and height. J Clin Endocrinol Metab 2000:85:1290-5. 283. Binder G, Schwarze CP, Ranke MB. Identification of short stature caused by SHOX defects and therapeutic effect of recombinant human growth hormone. J Clin Endocrinol Metab 2000:85:245-9. 284. Parvari R, Mumm S, Galil A, et al. Deletion of 8.5 Mb, including the FMR1 gene, in a male with the fragile X syndrome phenotype and overgrowth. Am J Med Genet 1999;83: 302-7. 285. Baumann G. Mutations in the growth hormone releasing hormone receptor: a new form of dwarfism in humans. Growth Horm IGF Res 1999;9(suppl B):24-9. 286. Tanner JM. Foetus into man: physical growth from conception to maturity. Ware, United Kingdom: Castlemead Publications, 1989. 287. McCay CM, Maynard LA, Sperling G, et al. Retarded growth, life span, ultimate body size and age changes in the albino rat after feeding diets restricted in calories. J Nutr 1939;18:1-13. 288. Hart RW, Turturro A. Dietary restrictions and cancer. Environ Health Perspect 1997; 105:989-92. 289. Frankel S, Gunnell DJ, Peters TJ, et al. Childhood energy intake and adult cancer—the Boyd Orr cohort study. BMJ 1998:316:499-504. 290. Moller H. Clues to the aetiology of testicular germ cell tumours from descriptive epidemiology. Eur Urol 1993,23:8-13. 291. Hall AJ, Peckham CS. Infections in childhood and pregnancy as a cause of adult disease—methods and examples. Br Med 342 Gunnell et al. Bull 1997;53:10-23. 292. Hewitt D, Westropp CK, Acheson RM. Oxford Child Health Survey. Effect of childish ailments on skeletal development. Br J Prev Soc Med 1955;9:179-86. 293. Martorell R, Habicht JP. Growth in early childhood in developing countries. In: Falkner F, ed. Human growth: a comprehensive treatise. New York, NY: Plenum Press, 1986:241-62. 294. Perri F, Pastore M, Leandro G, et al. Helicobacter pylori infection and growth delay in older children. Arch Dis Child 1997;77:46-9. 295. Forman D, Newell DG, Fullerton F, et al. Association between infection with Helicobacter pylori and risk of gastric cancer: evidence from a prospective investigation. BMJ 1991;302:1302-5. 296. Bonelli L, Vitale V, Bistolfi F, et al. Hodgkin's disease in adults: association with social factors and age at tonsillectomy. A case-control study. Int J Cancer 1990;45:423-7. 297. Widdowson EM. Mental contentment and physical growth. Lancet 1951;1:1316—18. 298. Montgomery SM, Bartley MJ, Wilkinson RG. Family conflict and slow growth. Arch Dis Child 1997;77:326-30. 299. Michels KB, Trichopoulos D, Robins JM, et al. Birthweight as a risk factor for breast cancer. Lancet 1996;348:1542-6. 300. Tibblin G, Eriksson M, Cnattingius S, et al. High birthweight as a predictor of prostate cancer risk. Epidemiology 19956423t 301. Trichopoulos D, Lipworth L. Is cancer causation simpler than we thought, but more intractable? Epidemiology 1995;6: 347-9. 302. Miller FJ, Billewicz WZ, Thomson AM. Growth from birth to adult life of 442 Newcastle upon Tyne children. Br J Prev Soc Med 1972;26:224-30. 303. Leon DA, Koupilova I, Lithell HO, et al. Failure to realise growth potential in utero and adult obesity in relation to blood pressure in 50-year-old Swedish men. Br Med J 1996; 312:401-6. 304. Gunnell D, Davey Smith G, McConnachie A, et al. Separating in-utero and postnatal influences on later disease. (Letter). Lancet 1999;354:1526-7. 305. Okasha M, McCarron P, Davey Smith G, et al. Age at menarche: secular trends and its association with adult anthropometric measures. Ann Hum Biol 2001 ;28:68-78. 306. Shangold MM, Kelly M, Berkeley AS, et al. Relationship between menarcheal age and adult height. South Med J 1989; 82:443-5. 307. Albanes D, Winick M. Are cell number and cell proliferation risk factors for cancer? J Natl Cancer Inst 1988;80:772-5. 308. Winick M, Noble A. Cellular response in rats during malnutrition at various ages. J Nutr 1966;66:300-6. 309. Hankinson SE, Willett WC, Manson JE, et al. Alcohol, height, and adiposity in relation of estrogen and prolactin levels in postmenopausal women. J Natl Cancer Inst 1995; 87:1297-302. 310. Nagata C, Kabuto M, Takatsuka N, et al. Associations of alcohol, height, and reproductive factors with serum hormone concentrations in postmenopausal Japanese women. Breast Cancer Res Treat 1997;44:235-41. 311. Madigan MP, Troisi R, Potischman N, et al. Serum hormone levels in relation to reproductive and lifestyle factors in postmenopausal women (United States). Cancer Causes Control 1997,9:199-207. 312. Hsing AW, Comstock GW. Serological precursors of cancer—serum hormones and risk of subsequent prostatecancer. Cancer Epidemiol Biomarkers Prev 1993;2:27-32. 313. Gann PH, Hennekens CH, Ma J, et al. Prospective study of sex hormone levels and risk of prostate cancer. J Natl Cancer Inst 1996;88:1118-26. 314. Fall CH, Pandit AN, Law CM, et al. Size at birth and plasma insulin-like growth factor-1 concentrations. Arch Dis Child 1995;73:287-93. 315. Juul A, Dalgaard P, Blum WF. Serum levels of insulin-like growth factor (IGF)-binding protein-3 (IGFBP-3) in healthy infants, children, and adolescents: the relation to IGF-I, IGFII, IFGBP-1, IGFBP-2, age, sex, body mass index and pubertal maturation. J Clin Endocrinol Metab 1995;80:2534-^t2. 316. Le Roith D. Insulin-like growth factors. N Engl J Med 1997; 336:633^10. 317. Yu H, Spitz MR, Mistry J, et al. Plasma levels of insulin-like growth factor-I and lung cancer risk: a case-control analysis. J Natl Cancer Inst 1999;91:151-6. 318. Nabarro JD. Acromegaly. Clin Endocrinol 1987;26:481-512. 319. Ron E, Gridley G, Hrubec Z, et al. Acromegaly and gastrointestinal cancer. Cancer 1991 ;68:1673-7. 320. Orme SM, McNally RJ, Cartwright RA, et al. Mortality and cancer incidence in acromegaly: a retrospective cohort study. J Clin Endocrinol Metab 1998;83:2730-4. 321. Rosenbloom AL, Aguirre JG, Rosenfeld RG, et al. The little women of Loja—growth hormone-receptor deficiency in an inbred population of southern Ecuador. N Engl J Med 1990; 323:1367-74. 322. Laron Z, Lilos P, Klinger B. Growth curves for Laron syndrome. Arch Dis Child 1993;68:768-70. 323. Holly JM, Gunnell DJ, Davey Smith G. Growth hormone, IGF-I and cancer. Less intervention to avoid cancer? More intervention to prevent cancer? J Endocrinol 1999; 162:321-30. 324. Holly J. Insulin-like growth factor-I and new opportunities for cancer prevention. Lancet 1998;351:1373—5. 325. Davey Smith G, Gunnell D, Holly J. Cancer and insulin-like growth factor-I: a potential mechanism linking the environment with cancer risk. (Editorial). BMJ 2000;321:847-8. 326. Ng ST, Zhou J, Adesanya OO, et al. Growth hormone treatment induces mammary gland hyperplasia in aging primates. Nat Med 1997;3:1141-4. 327. Cianfarani S, Tedeschi B, Germani D, et al. In vitro effects of growth hormone (GH) and insulin-like growth factor I and II (IGF-I and -II) on chromosome fragility and p53 protein expression in human lymphocytes. Eur J Clin Invest 1998; 28:41-7. 328. Ketelslegers JM, Maiter D, Maes M, et al. Nutritional regulation of insulin-like growth factor-I. Metabolism 1995;44(suppl 4):50-7. 329. Langford K, Blum W, Nicolaides K, et al. The pathophysiology of the insulin-like growth factor axis in fetal growth failure: a basis for programming by undernutrition. Eur J Clin Invest 1994;24:851-6. 330. Tanner JM, Whitehouse RH, Hughes CR, et al. Effect of human growth hormone treatment for 1 to 7 years on growth of 100 children, with growth hormone deficiency, low birthweight, inherited smaliness, Turner's syndrome, and other complaints. Arch Dis Child 1971;46:745-82. 331. Signorello LB, Kuper H, Lagiou P, et al. Lifestyle factors and insulin-like growth factor 1 levels among elderly men. Eur J Cancer Prev 2000;9:173-8. 332. Goodman-Gruen D, Barrett-Connor E. Epidemiology of insulin-like growth factor-I in elderly men and women: The Rancho Bernardo Study. Am J Epidemiol 1997;145:970-6. 333. Waaler HT. Height, weight and mortality: the Norwegian experience. Acta Med Scand 1984;679(suppl):l-56. 334. Floud R, Wachter K, Gregory A. Height, health and history: nutritional status in the United Kingdom, 1750-1980. Cambridge, United Kingdom: Cambridge University Press, 1990. 335. Kalapothaki V, Tzonou A, Hsieh C, et al. Tobacco, ethanol, coffee, pancreatitis, diabetes mellitus and cholelithiasis as risk factors for pancreatic carcinoma. Cancer Causes Control 1993;4:375-82. 336. Howe GR, Ghadirian P, Bueno de Mesquita HB, et al. A collaborative case-control study of nutrient intake and pancreatic cancer within the SEARCH programme. Int J Cancer 1992;51:365-72. Epidemiol Rev Vol. 23, No. 2, 2001