TY - JOUR
AU - Hunter, David J.
AB - Abstract The association between anthropometric indices and the risk of breast cancer was analyzed using pooled data from seven prospective cohort studies. Together, these cohorts comprise 337,819 women and 4,385 incident invasive breast cancer cases. In multivariate analyses controlling for reproductive, dietary, and other risk factors, the pooled relative risk (RR) of breast cancer per height increment of 5 cm was 1.02 (95% confidence interval (CI): 0.96, 1.10) in premenopausal women and 1.07 (95% CI: 1.03, 1.12) in postmenopausal women. Body mass index (BMI) showed significant inverse and positive associations with breast cancer among pre- and postmenopausal women, respectively; these associations were nonlinear. Compared with premenopausal women with a BMI of less than 21 kg/m2, women with a BMI exceeding 31 kg/m2 had an RR of 0.54 (95% CI: 0.34, 0.85). In postmenopausal women, the RRs did not increase further when BMI exceeded 28 kg/m2; the RR for these women was 1.26 (95% CI: 1.09, 1.46). The authors found little evidence for interaction with other breast cancer risk factors. Their data indicate that height is an independent risk factor for postmenopausal breast cancer; in premenopausal women, this relation is less clear. The association between BMI and breast cancer varies by menopausal status. Weight control may reduce the risk among postmenopausal women. Am J Epidemiol 2000;152:514–27. body height, body weight, breast neoplasms, prospective studies BMI, body mass index, HRT, hormone replacement therapy, IGF, insulin-like growth factor, RR, relative risk The relation between body size and breast cancer risk has been the subject of numerous investigations. Many of these studies have been focused on the association between weight (typically corrected for height) and breast cancer. Attention to the relation with height has increased in recent years due to increased interest in the effects of early diet on breast cancer risk in later life (1); adult height may serve as an indicator of childhood or adolescent nutrition and energy balance (2). Studies have revealed that associations with body size may vary by menopausal status. Hunter and Willett (3) concluded that for body mass index (BMI, defined as weight (kg)/height2 (m2)), most large case-control studies, but only some cohort studies, showed a clear positive association with postmenopausal breast cancer risk and that the relative risks found in cohort studies were much closer to the null value than were those from case-control studies. On the other hand, inverse associations with BMI were found for premenopausal breast cancer in most cohort studies, whereas case-control studies showed both inverse and direct associations (3, 4). For adult height, a modest positive association with breast cancer risk was found in most, but not all, case-control studies. These positive associations were predominantly found for postmenopausal breast cancer. Cohort studies have shown positive associations with height in both pre- and postmenopausal women (3), but several used databases in which information on height, but not reproductive and other risk factors, was available, leaving open the possibility of uncontrolled confounding. Against this background, we studied the relation between height, weight, and BMI and breast cancer risk in the Pooling Project of Diet and Cancer (5). In this project, primary data from seven major prospective studies have been combined to evaluate associations between dietary factors and breast cancer risk by using a standardized approach. Since these studies also provided information on anthropometry and many covariates that can act as confounders or effect modifiers (e.g., reproductive factors, family history of cancer, and exogenous hormone use), we decided to use the pooled data for the current analysis of anthropometry and breast cancer as well. The pooling approach enables use of uniform categories of anthropometric variables and covariates and evaluation of potential effect modification by covariates. MATERIALS AND METHODS The Pooling Project has been described previously (5). Briefly, seven prospective studies (6–12) (table 1) were identified in 1992 that met the following predefined criteria: the study 1) had at least 200 incident cases of breast cancer available for analysis; 2) assessed long-term intake of foods and nutrients, including energy intake, at baseline; and 3) completed a validation study of the diet assessment method or a closely related instrument. For Pooling Project analyses (5), the Nurses' Health Study was split into two studies since it had repeated assessments of dietary intake and a longer follow-up period than did the other studies. To make the length of follow-up more comparable with the other cohorts, Nurses' Health Study (a) corresponds to the 1980–1986 follow-up period, while Nurses' Health Study (b) corresponds to the period 1986–1991. Self-administered questionnaires were used to assess anthropometric factors, reproductive factors, medical history, and family history. Each study measured dietary intake by using food frequency questionnaires. Incident breast cancers were ascertained by using follow-up questionnaires and inspection of medical records and/or tumor registry linkage. In all cohorts, follow-up was estimated to be more than 90 percent complete. TABLE 1. Characteristics of the cohort studies included in the pooled analysis of anthropometry and breast cancer, the Pooling Project of Diet and Cancer Study  Baseline cohort size  Age at baseline (years)   Years of follow-up  No. of cases*   Median  Range  Total  Premenopausal  Postmenopausal  AHS†  15,172  52  28–90  1976–1982  122  20  87  CBSS†  56,837  50  40–59  1982–1987  392  122  242  IWHS†  34,406  61  55–69  1986–1991  643  0  643  NLCS†  62,412  61  55–69  1986–1989  420  0  420  NYSC†  18,475  68  50–93  1980–1987  358  0  358  NHS(a)†  89,046  47  34–59  1980–1986  1,015  383  571  NHS(b)†  68,817  53  40–65  1986–1991  787  130  613  SMC†  61,471  52  40–76  1987–1993  648  68  274  Total  337,819        4,385  723  3,208  Study  Baseline cohort size  Age at baseline (years)   Years of follow-up  No. of cases*   Median  Range  Total  Premenopausal  Postmenopausal  AHS†  15,172  52  28–90  1976–1982  122  20  87  CBSS†  56,837  50  40–59  1982–1987  392  122  242  IWHS†  34,406  61  55–69  1986–1991  643  0  643  NLCS†  62,412  61  55–69  1986–1989  420  0  420  NYSC†  18,475  68  50–93  1980–1987  358  0  358  NHS(a)†  89,046  47  34–59  1980–1986  1,015  383  571  NHS(b)†  68,817  53  40–65  1986–1991  787  130  613  SMC†  61,471  52  40–76  1987–1993  648  68  274  Total  337,819        4,385  723  3,208  * Cases consisted of women diagnosed with invasive breast cancer, with complete data on height and weight. A total of 454 cases were considered to have uncertain menopausal status at diagnosis (see text for details). † AHS, Adventist Health Study; CBSS, Canadian National Breast Screening Study; IWHS, Iowa Women's Health Study; NLCS, Netherlands Cohort Study; NYSC, New York State Cohort; NHS(a), Nurses' Health Study (a); NHS(b), Nurses' Health Study (b); SMC, Sweden Mammography Cohort. View Large The questionnaires yielded self-reported information on height and weight at baseline. Weight at ages other than baseline age was not available for most cohorts; therefore, these variables and weight change could not be taken into account in this analysis. All height and weight data were converted into meters and kilograms, respectively. BMI (Quetelet index) was used as an indicator of relative weight. The correlation coefficients between height and BMI varied from −0.05 in the Nurses' Health Study to −0.13 in the Netherlands Cohort Study and the New York State Cohort. Participants were excluded from these analyses if they met study-specific exclusion criteria, if their reported energy intake was greater than three standard deviations from the study-specific loge-transformed mean energy intake of the baseline population, if they reported a history of cancer other than nonmelanoma skin cancer at baseline, or if they had missing data on anthropometry. As a result of these additional exclusions and expanded follow-up in some studies, the baseline cohort size and the number of cases reported in these analyses may differ from original study-specific publications (7, 10, 13–15) and from previously reported Pooling Project publications (5, 1616). The pooled analysis of anthropometry and breast cancer is based on 4,385 incident cases of invasive breast cancer (table 1). Risk factors, including anthropometric factors, may differ for premenopausal and postmenopausal breast cancer (3). However, most of the cohort studies included had information on menopausal status at baseline only. As has been described before (16), to assign changing menopause status during follow-up, an algorithm was developed based on an analysis of 42,531 Nurses' Health Study participants who were followed up biennially, were premenopausal in 1976, and remained premenopausal or had a natural menopause by 1992. On the basis of the algorithm, 723 (16.5 percent) incident cases with complete anthropometry data were premenopausal, 3,208 (73.2 percent) cases were postmenopausal, and 454 (10.3 percent) cases were of uncertain menopausal status. Because there were only 20 premenopausal cases in the Adventist Health Study, leading to very unstable effect estimates, this study was excluded from the pooled analysis among premenopausal cases. Statistical analysis Each study was analyzed by using a method consistent with its study design. Five of the cohort studies (Adventist Health Study, Iowa Women's Health Study, New York State Cohort, Nurses' Health Study (a) and (b), and the Sweden Mammography Cohort) were analyzed as nested case-control studies with a 1:10 ratio of cases diagnosed with invasive breast cancer to controls free of diagnosed cancer. Controls were matched to the cases on the basis of the year of diagnosis and the year of birth of the cases. In addition, controls did not have a history of breast cancer prior to the year of diagnosis of the cases. Controls were sampled without replacement within each year, but were eligible to be chosen again as controls or to be classified as cases in subsequent years. We used a nested case-control approach because of computational limitations. A nested case-control design also was used for the Canadian National Breast Screening Study; the investigators of that study selected two controls matched to each case on the basis of age (±2 months) and then processed previously administered questionnaires for the cases and matched controls. The Netherlands Cohort Study used a case-cohort design (17); cases were identified within the total cohort and compared with the person-time of a subcohort of 1,812 women randomly sampled at baseline. For the nested case-control studies, relative risks associated with height, weight, and BMI were estimated by conditional logistic regression with the use of SAS PROC PHREG (18). These models were stratified by year of birth and calendar year (in single years) except for the models using the Canadian National Breast Screening Study, which were stratified by year of birth and additional study-specific matching factors (8). For the Netherlands Cohort Study, EPICURE software (19) was used to estimate the relative risk; models were stratified on age at baseline. Multivariate relative risks were adjusted for menopausal status at diagnosis; age at menarche; parity; age at birth of the first child; postmenopausal hormone use; oral contraceptive use; history of benign breast disease; maternal history of breast cancer; history of breast cancer in a sister; smoking status; education; BMI or height (depending on the analysis); and intake of fat, fiber, energy, and alcohol (see table 2 for the definition of the categories used in the analysis). An indicator variable for missing responses was created for each covariate to minimize loss of observations due to missing covariate data. Two-sided 95 percent confidence intervals were calculated. The “random effects” model developed by DerSimonian and Laird (20) was used to combine loge relative risks from the multiple studies. TABLE 2. Pooled multivariate relative risks and 95 percent confidence intervals for breast cancer according to categories of height, weight, and body mass index, the Pooling Project of Diet and Cancer Anthropometric variable  Premenopausal†   Postmenopausal   Total   No. of cases  RR‡  95% CI§  No. of cases  RR‡  95% CI  No. of cases  RR¶  95% CI  Height (m)                     <1.60  149  1.0  Reference  724  1.0  Reference  970  1.0  Reference   1.60–<1.65  202  1.21  0.94, 1.55  921  1.09  0.98, 1.21  1,261  1.10  1.00, 1.20   1.65–<1.70  196  1.06  0.82, 1.36  916  1.23  1.11, 1.37  1,253  1.20  1.09, 1.32   1.70–<1.75  117  1.14  0.86, 1.52  491  1.24  1.09, 1.41  688  1.24  1.11, 1.38   ≥1.75  39  1.42  0.95, 2.12  156  1.28  0.94, 1.76  213  1.22*  0.90, 1.65   p value, test for trend  0.41  <0.001  0.001  Weight (kg)                     <60  300  1.0#  Reference  812  1.0#  Reference         60–<65  132  0.85  0.66, 1.09  608  1.16  1.03, 1.31         65–<70  100  1.09  0.83, 1.45  549  1.22  1.01, 1.47         70–<75  83  1.12  0.83, 1.52  464  1.22  1.01, 1.47         75–<80  35  0.95  0.51, 1.76  289  1.43  1.14, 1.80         ≥80  53  0.58  0.40, 0.83  486  1.25  1.02, 1.52         p value, test for trend  0.02  0.003  BMI (kg/m2)                     <21  158  1.0  Reference  363  1.0  Reference         21–<23  223  1.24  0.97, 1.57  632  1.14  0.99, 1.33         23–<25  131  1.03  0.78, 1.35  699  1.15  1.00, 1.34         25–<27  82  1.08  0.79, 1.48  564  1.26  1.09, 1.47         27–<29  47  0.97  0.66, 1.44  401  1.43  1.21, 1.67         29–<31  32  0.96  0.60, 1.52  224  1.21  1.01, 1.46         31–<33  10  0.55  0.26, 1.15  140  1.29  1.03, 1.60         ≥33  20  0.58  0.34, 1.00  185  1.27  1.03, 1.55         p value, test for trend  0.007  0.001    Anthropometric variable  Premenopausal†   Postmenopausal   Total   No. of cases  RR‡  95% CI§  No. of cases  RR‡  95% CI  No. of cases  RR¶  95% CI  Height (m)                     <1.60  149  1.0  Reference  724  1.0  Reference  970  1.0  Reference   1.60–<1.65  202  1.21  0.94, 1.55  921  1.09  0.98, 1.21  1,261  1.10  1.00, 1.20   1.65–<1.70  196  1.06  0.82, 1.36  916  1.23  1.11, 1.37  1,253  1.20  1.09, 1.32   1.70–<1.75  117  1.14  0.86, 1.52  491  1.24  1.09, 1.41  688  1.24  1.11, 1.38   ≥1.75  39  1.42  0.95, 2.12  156  1.28  0.94, 1.76  213  1.22*  0.90, 1.65   p value, test for trend  0.41  <0.001  0.001  Weight (kg)                     <60  300  1.0#  Reference  812  1.0#  Reference         60–<65  132  0.85  0.66, 1.09  608  1.16  1.03, 1.31         65–<70  100  1.09  0.83, 1.45  549  1.22  1.01, 1.47         70–<75  83  1.12  0.83, 1.52  464  1.22  1.01, 1.47         75–<80  35  0.95  0.51, 1.76  289  1.43  1.14, 1.80         ≥80  53  0.58  0.40, 0.83  486  1.25  1.02, 1.52         p value, test for trend  0.02  0.003  BMI (kg/m2)                     <21  158  1.0  Reference  363  1.0  Reference         21–<23  223  1.24  0.97, 1.57  632  1.14  0.99, 1.33         23–<25  131  1.03  0.78, 1.35  699  1.15  1.00, 1.34         25–<27  82  1.08  0.79, 1.48  564  1.26  1.09, 1.47         27–<29  47  0.97  0.66, 1.44  401  1.43  1.21, 1.67         29–<31  32  0.96  0.60, 1.52  224  1.21  1.01, 1.46         31–<33  10  0.55  0.26, 1.15  140  1.29  1.03, 1.60         ≥33  20  0.58  0.34, 1.00  185  1.27  1.03, 1.55         p value, test for trend  0.007  0.001    * Test for heterogeneity between studies, p < 0.05. † The Adventist Health Study was not included in analyses of premenopausal women. ‡ Multivariate relative risks (RR) were adjusted for age at menarche (≤11, 12, 13, 14, ≥15 years), parity (0, 1–2, ≥3), age at birth of first child (≤20, 21–25, 26–30, >30 years), postmenopausal hormone use (ever, never), oral contraceptive use (ever, never), history of benign breast disease (no, yes), maternal history of breast cancer (no, yes), history of breast cancer in a sister (no, yes, no sisters), smoking status (ever, never), education (less than high school graduation, high-school graduation, more than high-school graduation), fat intake (quintiles), fiber intake (quintiles), energy intake (continuous), and alcohol intake (0, >0–<1.5, 1.5–<5, 5–<15, 15–<30, ≥30 g/day). § CI, confidence interval. ¶ The model included all of the above terms, and menopausal status at diagnosis (premenopausal, postmenopausal, uncertain). # Further adjusted for height. View Large In these analyses, anthropometric variables were entered as categorical variables, and in addition, models were run with continuous variables. Analyses were conducted for the total group and in pre- and postmenopausal cases only. Potential effect modification by other covariates was also evaluated by examining the associations between anthropometry and breast cancer within each category of the covariate and statistical testing of the pooled linear interaction term. Pooled p values for interaction were based on the squared pooled estimate of the continuous interaction term divided by its estimated variance. We used restricted cubic splines as a graphic method of presenting dose-response curves that make no a priori assumptions about the shape of the curve. These nonparametric regression curves can also be compared with the linear model to evaluate linearity of the associations by using a likelihood ratio test, as described previously (16). For these analyses, the studies were combined into a single data set stratified by study, since there was no between-study heterogeneity in the effects of the other model covariates (21) and because height and weight were measured similarly in all studies. All p values are two-sided. RESULTS Table 1 shows the characteristics of the cohort studies and the number of incident breast cancer cases with complete anthropometry data that are available for this pooled analysis. The results of the pooled multivariate analyses using categories of anthropometric factors are presented in table 2. Height was positively associated with breast cancer risk in both pre- and postmenopausal women; the relative risks (RRs) for women 1.75 m or taller compared with women shorter than 1.60 m were 1.42 for premenopausal women, 1.28 for postmenopausal women, and 1.22 for all women. The tests for linear trend were significant in postmenopausal women and in the total group. The association of weight with breast cancer depended on menopausal status. There was an inverse association in premenopausal women, concentrated in the top weight categories. For postmenopausal women, a positive association was found, with statistically significant relative risks in all categories. The RRs for women who weighed 80 kg or more compared with those who weighed less than 60 kg were 0.58 for premenopausal and 1.25 for postmenopausal women. The tests for trend were significant in both groups. BMI showed a significant inverse association with breast cancer risk among premenopausal women (p value, test for trend = 0.007) but the inverse association was concentrated in women in the highest BMI categories. When BMI of 31 kg/m2 or more was used as the highest BMI category, the RR was 0.54 (95 percent CI: 0.34, 0.85) (data not shown). On the other hand, the association with BMI among postmenopausal women was significantly positive (p value, test for trend = 0.001), with an increased risk seen in all BMI categories. The RRs for women with a BMI of 33 kg/m2 or more compared with women with a BMI of less than 21 kg/m2 was 0.58 (95 percent CI: 0.34, 1.00) for premenopausal women and 1.27 (95 percent CI: 1.03, 1.55) for postmenopausal women. Table 3 shows study-specific and pooled relative risk estimates for multivariate models with continuous anthropometric data. The pooled RR per height increment of 5 cm was 1.02 for premenopausal women, but the estimates from the four studies for which data on premenopausal women were available varied: Only Nurses' Health Study (a) showed a positive association (p value, test for heterogeneity = 0.74). For postmenopausal women, all studies showed a positive association except the Canadian National Breast Screening Study and the New York State Cohort, in which no association was found. The pooled RR was 1.07 (p value, test for heterogeneity = 0.17) (table 3). For the total group of women, the pooled RR was 1.07 (95 percent CI: 1.02, 1.11) per height increment of 5 cm (data not shown). TABLE 3. Study-specific and pooled multivariate relative risks* of breast cancer for continuous linear effects of height, weight, and body mass index, the Pooling Project of Diet and Cancer Anthropometric variable  AHS†   CBSS†   IWHS†   NLCS†   NYSC†   NHS(a)†   NHS(b)†   SMC†   Pooled RR† 95% CI†  p value, test for heterogeneity between studies  RR 95% CI  RR 95% CI  RR 95% CI  RR 95% CI  RR 95% CI  RR 95% CI  RR 95% CI  RR 95% CI  Height (5-cm increment)                       Premenopausal  –‡  0.98  –  –  –  1.05  0.97  0.98  1.02  0.74      0.75, 1.28        0.97, 1.15  0.83, 1.12  0.78, 1.23  0.96, 1.10     Postmenopausal  1.10  0.99  1.10  1.20  1.00  1.04  1.07  1.09  1.07  0.17    0.91, 1.32  0.85, 1.16  1.03, 1.17  1.09, 1.32  0.92, 1.09  0.97, 1.12  1.00, 1.15  0.97, 1.23  1.03, 1.12    Weight§ (10-kg increment)                       Premenopausal  –‡  0.86  –  –  –  0.87  0.95  0.93  0.90  0.80      0.60, 1.22        0.79, 0.97  0.82, 1.11  0.66, 1.30  0.83, 0.97       Postmenopausal  1.18  1.11  1.11  1.03  1.04  1.01  1.03  1.15  1.06  0.36    0.97, 1.44  0.91, 1.36  1.04, 1.18  0.91, 1.16  0.95, 1.15  0.93, 1.09  0.97, 1.11  1.02, 1.30  1.03, 1.10    Body mass index (4kg/m2 increment)                       Premenopausal  –‡  0.86  –  –  –  0.86  0.95  0.90  0.89  0.82      0.60, 1.25        0.77, 0.96  0.80, 1.12  0.62, 1.32  0.81, 0.97     Postmenopausal  1.21  1.11  1.11  0.99  1.05  1.01  1.04  1.15  1.07  0.34    0.98, 1.50  0.91, 1.36  1.04, 1.19  0.86, 1.13  0.95, 1.16  0.93, 1.10  0.97, 1.12  1.01, 1.32  1.02, 1.11    Anthropometric variable  AHS†   CBSS†   IWHS†   NLCS†   NYSC†   NHS(a)†   NHS(b)†   SMC†   Pooled RR† 95% CI†  p value, test for heterogeneity between studies  RR 95% CI  RR 95% CI  RR 95% CI  RR 95% CI  RR 95% CI  RR 95% CI  RR 95% CI  RR 95% CI  Height (5-cm increment)                       Premenopausal  –‡  0.98  –  –  –  1.05  0.97  0.98  1.02  0.74      0.75, 1.28        0.97, 1.15  0.83, 1.12  0.78, 1.23  0.96, 1.10     Postmenopausal  1.10  0.99  1.10  1.20  1.00  1.04  1.07  1.09  1.07  0.17    0.91, 1.32  0.85, 1.16  1.03, 1.17  1.09, 1.32  0.92, 1.09  0.97, 1.12  1.00, 1.15  0.97, 1.23  1.03, 1.12    Weight§ (10-kg increment)                       Premenopausal  –‡  0.86  –  –  –  0.87  0.95  0.93  0.90  0.80      0.60, 1.22        0.79, 0.97  0.82, 1.11  0.66, 1.30  0.83, 0.97       Postmenopausal  1.18  1.11  1.11  1.03  1.04  1.01  1.03  1.15  1.06  0.36    0.97, 1.44  0.91, 1.36  1.04, 1.18  0.91, 1.16  0.95, 1.15  0.93, 1.09  0.97, 1.11  1.02, 1.30  1.03, 1.10    Body mass index (4kg/m2 increment)                       Premenopausal  –‡  0.86  –  –  –  0.86  0.95  0.90  0.89  0.82      0.60, 1.25        0.77, 0.96  0.80, 1.12  0.62, 1.32  0.81, 0.97     Postmenopausal  1.21  1.11  1.11  0.99  1.05  1.01  1.04  1.15  1.07  0.34    0.98, 1.50  0.91, 1.36  1.04, 1.19  0.86, 1.13  0.95, 1.16  0.93, 1.10  0.97, 1.12  1.01, 1.32  1.02, 1.11    * Relative risks were adjusted for age at menarche (≤11, 12, 13, 14, ≥15 years), parity (0, 1, 2, ≥3), age at birth of the first child (<20, 21–25, 26, 30, ≥30 years), postmenopausal hormone use (ever, never), oral contraceptive use (ever, never), history of benign breast disease (no, yes), maternal history of breast cancer (no, yes), history of breast cancer in a sister (no, yes, no sisters), smoking status (ever, never), education (less than high school graduation, high school graduation, more than high school graduation), fat intake (quintiles), fiber intake (quintiles), energy intake (continuous), and alcohol intake (0, >0–<1.5, 1.5–<5, 5–<15, 15–<30, ≥30 g/day). † AHS, Adventist Health Study; CBSS, Canadian National Breast Screening Study; IWHS, Iowa Women's Health Study; NLCS, Netherlands Cohort Study; NYSC, New York State Cohort; NHS(a), Nurses' Health Study (a); NHS(b), Nurses' Health Study (b); SMC, Sweden Mammography Cohort; RR, relative risk; CI, confidence interval. ‡ The Adventist Health Study was not included in analyses of premenopausal women. § Further adjusted for height. View Large Both continuous weight (with height in the model) and BMI showed significantly inverse associations with breast cancer risk in premenopausal women in each study, without statistical evidence of heterogeneity. The pooled RR per weight increment of 10 kg was 0.90, and for a BMI increment of 4 kg/m2, it was 0.89 (table 3). In postmenopausal women, continuous BMI was positively associated with breast cancer in all studies except the Netherlands Cohort Study and Nurses' Health Study (a), in which there was essentially no association. For weight itself, a similar pattern was seen. There was no statistical evidence of heterogeneity between the studies. The pooled RR was 1.07 per 4-kg/m2 increment in BMI and 1.06 per 10-kg increment in weight. The nonparametric regression curve (figure 1) indicated a linear association with height; the test for deviation from linearity was also not significant (p = 0.65). The results for BMI are shown in figure 2 for pre- and postmenopausal women separately. Although the test for deviation from linearity in premenopausal women was technically not significant (p = 0.06), power was limited due to small numbers in this group. Among postmenopausal women, there was significant deviation from linearity (p = 0.004). The curve among postmenopausal women suggests that the relative risk did not increase further above BMI of more than 28 kg/m2. When BMI more than 28 kg/m2 was used as the highest category in the categorical analyses, the RR was 1.26 (95 percent CI: 1.09, 1.46) (data not shown). FIGURE 1. View largeDownload slide Nonparametric regression curve for the relation between height and breast cancer, the Pooling Project of Diet and Cancer. FIGURE 1. View largeDownload slide Nonparametric regression curve for the relation between height and breast cancer, the Pooling Project of Diet and Cancer. FIGURE 2. View largeDownload slide Nonparametric regression curve for the relation between body mass index and breast cancer, the Pooling Project of Diet and Cancer. FIGURE 2. View largeDownload slide Nonparametric regression curve for the relation between body mass index and breast cancer, the Pooling Project of Diet and Cancer. The results of the analyses on interaction are given in tables 4 and 5. In these analyses, height and BMI were entered as continuous variables into the models, and relative risks for these variables were estimated in each category of the covariates. The analysis by menopausal status (table 4) showed no evidence for interaction between menopausal status and height (p value, test for interaction = 0.13), but there was significant interaction with BMI (p value, test for interaction = 0.004). Therefore, in the ensuing analyses, interactions with height were evaluated in the total group, whereas interactions with BMI were evaluated in pre- and postmenopausal women separately. Before further evaluation of interactions with other covariates, the mutual interaction between height and BMI was considered. As table 4 shows, the effect of height was not significantly different between BMI categories, nor was the effect of BMI significantly different between height categories in the pre- and postmenopausal women. When the breast cancer associations with height and BMI were evaluated according to categories of age at diagnosis, the association with height and BMI appeared to be somewhat stronger in older, postmenopausal women. In premenopausal women, the association with BMI was more inverse in older premenopausal women. However, the interactions with age at diagnosis were not significant in either pre- or postmenopausal women (table 4). TABLE 4. Pooled multivariate relative risks* (95% confidence intervals) for a 5-cm increment in height and a 4-kg/m2 increment in body mass index by menopausal status, anthropometric factors, and age at diagnosis†, the Pooling Project of Diet and Cancer Covariate and anthropometric variable  Category of covariate  p value, test for interaction  Menopausal status‡  Premenopausal   Postmenopausal            Height  1.02 (0.96, 1.10)  1.07 (1.03, 1.12)        0.13   BMI§  0.89 (0.81, 0.97)  1.07 (1.02, 1.11)        0.004  BMI (kg/m2)  <21   21–<23   23–<25   25–<29   ≥29      Height  1.10 (1.02, 1.17)  1.08 (0.98, 1.19)  1.09 (1.03, 1.15)  1.08 (1.03, 1.14)  1.02 (0.93, 1.12)  0.12  Height (m)  <1.60   1.60–<1.65   1.65–<1.70   1.70–<1.75   ≥1.75     BMI               Premenopausal  0.85 (0.62, 1.17)  0.92 (0.79, 1.07)  0.94 (0.81, 1.09)  0.87 (0.70, 1.08)  0.93 (0.64, 1.36)  0.40   Postmenopausal  1.10 (1.03, 1.18)  1.10 (1.02, 1.18)  1.06 (0.99, 1.14)  1.02 (0.92, 1.12)  1.02 (0.81, 1.29)  0.30  Age at diagnosis (years)‡ in premenopausal women  <45   45–<50   >50          Height  0.98 (0.80, 1.20)  1.06 (0.95, 1.17)  1.01 (0.85, 1.21)      0.70   BMI  1.04 (0.74, 1.46)  0.95 (0.84, 1.08)  0.73 (0.54, 0.98)      0.08  Age at diagnosis (years) in postmenopausal women  <55   55–<60   60–<65   65–<70   ≥70      Height  1.12 (1.00, 1.26)  1.06 (0.95, 1.18)  1.04 (0.94, 1.16)  1.14 (1.06, 1.23)  1.15 (1.02, 1.30)  0.24   BMI  0.97 (0.86, 1.10)  1.04 (0.94, 1.16)  1.06 (0.95, 1.17)  1.15 (1.07, 1.24)  1.09 (0.99, 1.20)  0.13  Covariate and anthropometric variable  Category of covariate  p value, test for interaction  Menopausal status‡  Premenopausal   Postmenopausal            Height  1.02 (0.96, 1.10)  1.07 (1.03, 1.12)        0.13   BMI§  0.89 (0.81, 0.97)  1.07 (1.02, 1.11)        0.004  BMI (kg/m2)  <21   21–<23   23–<25   25–<29   ≥29      Height  1.10 (1.02, 1.17)  1.08 (0.98, 1.19)  1.09 (1.03, 1.15)  1.08 (1.03, 1.14)  1.02 (0.93, 1.12)  0.12  Height (m)  <1.60   1.60–<1.65   1.65–<1.70   1.70–<1.75   ≥1.75     BMI               Premenopausal  0.85 (0.62, 1.17)  0.92 (0.79, 1.07)  0.94 (0.81, 1.09)  0.87 (0.70, 1.08)  0.93 (0.64, 1.36)  0.40   Postmenopausal  1.10 (1.03, 1.18)  1.10 (1.02, 1.18)  1.06 (0.99, 1.14)  1.02 (0.92, 1.12)  1.02 (0.81, 1.29)  0.30  Age at diagnosis (years)‡ in premenopausal women  <45   45–<50   >50          Height  0.98 (0.80, 1.20)  1.06 (0.95, 1.17)  1.01 (0.85, 1.21)      0.70   BMI  1.04 (0.74, 1.46)  0.95 (0.84, 1.08)  0.73 (0.54, 0.98)      0.08  Age at diagnosis (years) in postmenopausal women  <55   55–<60   60–<65   65–<70   ≥70      Height  1.12 (1.00, 1.26)  1.06 (0.95, 1.18)  1.04 (0.94, 1.16)  1.14 (1.06, 1.23)  1.15 (1.02, 1.30)  0.24   BMI  0.97 (0.86, 1.10)  1.04 (0.94, 1.16)  1.06 (0.95, 1.17)  1.15 (1.07, 1.24)  1.09 (0.99, 1.20)  0.13  * Relative risks were adjusted for age at menarche (≤11, 12, 13, 14, ≥15 years), parity (0, 1, 2, ≥3), age at birth of the first child (≤20, 21–25, 26–30, ≥30 years), postmenopausal hormone use (ever, never), oral contraceptive use (ever, never), history of benign breast disease (no, yes), maternal history of breast cancer (no, yes), history of breast cancer in a sister (no, yes, no sisters), smoking status (ever, never), education (less than high school graduation, high school graduation, more than high school graduation), fat intake (quintiles), fiber intake (quintiles), energy intake (continuous), and alcohol intake (0, >0–<1.5, 1.5–<5, 5–<15, 15–<30, ≥30 g/day). † The Adventist Health Study was not included in all analyses of premenopausal women. ‡ The Iowa Women's Health Study, the Netherlands Cohort Study, and the New York State Cohort were not included in this analysis. § BMI, body mass index. View Large TABLE 5. Pooled multivariate relative risks* (95% confidence intervals) for a 5-cm increment in height and a 4-kg/m2 increment in body mass index by levels of nondietary and dietary factors†, the Pooling Project of Diet and Cancer Covariate and anthropometric variable   Category of covariate  p value, test for interaction  Age at menarche (years)‡  <12   12   13   14   ≥15      Height  1.03 (0.92, 1.15)  1.12 (1.01, 1.24)  1.03 (0.97, 1.08)  1.13 (1.01, 1.26)  1.07 (0.99, 1.17)  0.70   BMI                Premenopausal  0.88 (0.72, 1.09)  1.00 (0.87, 1.16)  0.84 (0.60, 1.17)  0.89 (0.64, 1.24)  0.94 (0.66, 1.35)  0.94    Postmenopausal  1.07 (0.99, 1.16)  1.07 (1.00, 1.16)  1.05 (0.98, 1.13)  1.05 (0.94, 1.18)  1.11 (0.96, 1.29)  0.77  Parity  0   1–2   ≥3          Height  1.05 (0.97, 1.15)  1.08 (1.02, 1.15)  1.06 (1.02, 1.11)      0.89   BMI                Premenopausal  0.98 (0.58, 1.66)  0.85 (0.71, 1.02)  0.93 (0.82, 1.05)      0.76    Postmenopausal  1.08 (0.99, 1.19)  1.11 (1.02, 1.20)  1.03 (0.95, 1.08)      0.23  Age at first birth (years)  ≤20   >20–25   >25–30   >30        Height  1.07 (0.94, 1.22)  1.07 (1.03, 1.12)  1.07 (1.01, 1.13)  1.05 (0.96, 1.14)    0.54   BMI                Premenopausal  0.89 (0.64, 1.25)  0.94 (0.75, 1.17)  0.89 (0.75, 1.05)  0.87 (0.60, 1.25)    0.99    Postmenopausal  1.06 (0.96, 1.18)  1.07 (1.01, 1.13)  1.10 (1.00, 1.21)  0.98 (0.86, 1.12)    0.86  Benign breast disease‡  No   Yes            Height  1.10 (1.05, 1.15)  1.02 (0.97, 1.09)        0.06   BMI                Premenopausal  0.87 (0.78, 0.96)  0.95 (0.69, 1.33)        0.60    Postmenopausal  1.06 (1.00, 1.12)  1.07 (0.97, 1.17)        0.82  Maternal history of breast cancer‡  No   Yes            Height  1.07 (1.02, 1.12)  1.23 (1.11, 1.37)        0.005   BMI                Premenopausal  0.87 (0.79, 0.97)  0.78 (0.46, 1.34)        0.64    Postmenopausal  1.07 (1.01, 1.13)  1.08 (0.95, 1.24)        0.77  History of breast cancer in a sister‡  No   Yes            Height  1.07 (1.02, 1.12)  0.96 (0.85, 1.09)        0.08   BMI              Premenopausal  0.82 (0.73, 0.92)  1.88 (0.63, 5.61)        0.14    Postmenopausal  1.07 (1.03, 1.12)  1.08 (0.93, 1.26)        1.00  Oral contraceptive use‡  Never   Ever            Height  1.09 (1.04, 1.15)  1.03 (0.98, 1.09)        0.21   BMI                Premenopausal  0.85 (0.73, 0.98)  0.91 (0.82, 1.02)        0.44    Postmenopausal  1.08 (1.01, 1.16)  1.06 (0.95, 1.19)        0.97  Hormone replacement therapy use‡  Never   Ever        0.45   Height (postmenopausal)  1.06 (1.00, 1.13)  1.09 (1.03, 1.15)        0.24   BMI (postmenopausal)  1.09 (1.04, 1.14)  1.04 (0.92, 1.18)          Education§  Less than high school   High school   More than high school          Height  1.10 (1.04, 1.17)  1.11 (1.02, 1.20)  1.00 (0.93, 1.08)        0.28   BMI                Premenopausal  0.89 (0.62, 1.27)  0.66 (0.31, 1.41)  1.28 (0.85, 1.93)      0.17    Postmenopausal  1.12 (0.99, 1.25)  1.04 (0.93, 1.16)  1.08 (0.99, 1.17)      0.37  Dietary fiber intake  Quintile 1   Quintile 2   Quintile 3   Quintile 4   Quintile 5      Height  1.09 (1.02, 1.17)  1.01 (0.95, 1.07)  1.10 (1.03, 1.16)  1.11 (1.02, 1.21)  1.03 (0.97, 1.09)  0.84   BMI                Premenopausal  0.85 (0.60, 1.19)  0.86 (0.70, 1.04)  0.89 (0.72, 1.10)  0.95 (0.80, 1.13)  1.00 (0.78, 1.28)  0.34    Postmenopausal  1.09 (1.01, 1.18)  1.04 (0.93, 1.17)  1.01 (0.93, 1.11)  1.05 (0.95, 1.17)  1.13 (1.03, 1.23)  0.52  Total fat intake  Quintile 1   Quintile 2   Quintile 3   Quintile 4   Quintile 5      Height  1.01 (0.95, 1.08)  1.08 (1.02, 1.15)  1.11 (1.05, 1.18)  1.04 (0.98, 1.11)  1.07 (0.99, 1.15)  0.31   BMI                Premenopausal  0.86 (0.70, 1.05)  0.77 (0.62, 0.96)  0.90 (0.73, 1.11)  0.95 (0.78, 1.16)  0.96 (0.82, 1.12)  0.14    Postmenopausal  1.10 (0.98, 1.24)  1.07 (0.99, 1.16)  0.99 (0.92, 1.07)  1.10 (1.00, 1.20)  1.08 (0.97, 1.20)  0.93  Alcohol intake (g/day)  0   1–<30   ≥30          Height  1.08 (1.03, 1.13)  1.05 (1.00, 1.12)  1.06 (0.94, 1.19)      0.76   BMI                Premenopausal  0.89 (0.77, 1.02)  0.93 (0.75, 1.14)  0.92 (0.57, 1.49)      0.92    Postmenopausal  1.07 (1.02, 1.13)  1.04 (0.99, 1.10)  1.00 (0.75, 1.35)      0.78  Smoking¶  Never   Ever            Height  1.05 (1.00, 1.11)  1.07 (0.99, 1.16)        0.56   BMI                Premenopausal  0.89 (0.78, 1.01)  0.88 (0.78, 1.00)        0.94    Postmenopausal  1.05 (1.00, 1.10)  1.06 (1.00, 1.14)        0.39  Covariate and anthropometric variable   Category of covariate  p value, test for interaction  Age at menarche (years)‡  <12   12   13   14   ≥15      Height  1.03 (0.92, 1.15)  1.12 (1.01, 1.24)  1.03 (0.97, 1.08)  1.13 (1.01, 1.26)  1.07 (0.99, 1.17)  0.70   BMI                Premenopausal  0.88 (0.72, 1.09)  1.00 (0.87, 1.16)  0.84 (0.60, 1.17)  0.89 (0.64, 1.24)  0.94 (0.66, 1.35)  0.94    Postmenopausal  1.07 (0.99, 1.16)  1.07 (1.00, 1.16)  1.05 (0.98, 1.13)  1.05 (0.94, 1.18)  1.11 (0.96, 1.29)  0.77  Parity  0   1–2   ≥3          Height  1.05 (0.97, 1.15)  1.08 (1.02, 1.15)  1.06 (1.02, 1.11)      0.89   BMI                Premenopausal  0.98 (0.58, 1.66)  0.85 (0.71, 1.02)  0.93 (0.82, 1.05)      0.76    Postmenopausal  1.08 (0.99, 1.19)  1.11 (1.02, 1.20)  1.03 (0.95, 1.08)      0.23  Age at first birth (years)  ≤20   >20–25   >25–30   >30        Height  1.07 (0.94, 1.22)  1.07 (1.03, 1.12)  1.07 (1.01, 1.13)  1.05 (0.96, 1.14)    0.54   BMI                Premenopausal  0.89 (0.64, 1.25)  0.94 (0.75, 1.17)  0.89 (0.75, 1.05)  0.87 (0.60, 1.25)    0.99    Postmenopausal  1.06 (0.96, 1.18)  1.07 (1.01, 1.13)  1.10 (1.00, 1.21)  0.98 (0.86, 1.12)    0.86  Benign breast disease‡  No   Yes            Height  1.10 (1.05, 1.15)  1.02 (0.97, 1.09)        0.06   BMI                Premenopausal  0.87 (0.78, 0.96)  0.95 (0.69, 1.33)        0.60    Postmenopausal  1.06 (1.00, 1.12)  1.07 (0.97, 1.17)        0.82  Maternal history of breast cancer‡  No   Yes            Height  1.07 (1.02, 1.12)  1.23 (1.11, 1.37)        0.005   BMI                Premenopausal  0.87 (0.79, 0.97)  0.78 (0.46, 1.34)        0.64    Postmenopausal  1.07 (1.01, 1.13)  1.08 (0.95, 1.24)        0.77  History of breast cancer in a sister‡  No   Yes            Height  1.07 (1.02, 1.12)  0.96 (0.85, 1.09)        0.08   BMI              Premenopausal  0.82 (0.73, 0.92)  1.88 (0.63, 5.61)        0.14    Postmenopausal  1.07 (1.03, 1.12)  1.08 (0.93, 1.26)        1.00  Oral contraceptive use‡  Never   Ever            Height  1.09 (1.04, 1.15)  1.03 (0.98, 1.09)        0.21   BMI                Premenopausal  0.85 (0.73, 0.98)  0.91 (0.82, 1.02)        0.44    Postmenopausal  1.08 (1.01, 1.16)  1.06 (0.95, 1.19)        0.97  Hormone replacement therapy use‡  Never   Ever        0.45   Height (postmenopausal)  1.06 (1.00, 1.13)  1.09 (1.03, 1.15)        0.24   BMI (postmenopausal)  1.09 (1.04, 1.14)  1.04 (0.92, 1.18)          Education§  Less than high school   High school   More than high school          Height  1.10 (1.04, 1.17)  1.11 (1.02, 1.20)  1.00 (0.93, 1.08)        0.28   BMI                Premenopausal  0.89 (0.62, 1.27)  0.66 (0.31, 1.41)  1.28 (0.85, 1.93)      0.17    Postmenopausal  1.12 (0.99, 1.25)  1.04 (0.93, 1.16)  1.08 (0.99, 1.17)      0.37  Dietary fiber intake  Quintile 1   Quintile 2   Quintile 3   Quintile 4   Quintile 5      Height  1.09 (1.02, 1.17)  1.01 (0.95, 1.07)  1.10 (1.03, 1.16)  1.11 (1.02, 1.21)  1.03 (0.97, 1.09)  0.84   BMI                Premenopausal  0.85 (0.60, 1.19)  0.86 (0.70, 1.04)  0.89 (0.72, 1.10)  0.95 (0.80, 1.13)  1.00 (0.78, 1.28)  0.34    Postmenopausal  1.09 (1.01, 1.18)  1.04 (0.93, 1.17)  1.01 (0.93, 1.11)  1.05 (0.95, 1.17)  1.13 (1.03, 1.23)  0.52  Total fat intake  Quintile 1   Quintile 2   Quintile 3   Quintile 4   Quintile 5      Height  1.01 (0.95, 1.08)  1.08 (1.02, 1.15)  1.11 (1.05, 1.18)  1.04 (0.98, 1.11)  1.07 (0.99, 1.15)  0.31   BMI                Premenopausal  0.86 (0.70, 1.05)  0.77 (0.62, 0.96)  0.90 (0.73, 1.11)  0.95 (0.78, 1.16)  0.96 (0.82, 1.12)  0.14    Postmenopausal  1.10 (0.98, 1.24)  1.07 (0.99, 1.16)  0.99 (0.92, 1.07)  1.10 (1.00, 1.20)  1.08 (0.97, 1.20)  0.93  Alcohol intake (g/day)  0   1–<30   ≥30          Height  1.08 (1.03, 1.13)  1.05 (1.00, 1.12)  1.06 (0.94, 1.19)      0.76   BMI                Premenopausal  0.89 (0.77, 1.02)  0.93 (0.75, 1.14)  0.92 (0.57, 1.49)      0.92    Postmenopausal  1.07 (1.02, 1.13)  1.04 (0.99, 1.10)  1.00 (0.75, 1.35)      0.78  Smoking¶  Never   Ever            Height  1.05 (1.00, 1.11)  1.07 (0.99, 1.16)        0.56   BMI                Premenopausal  0.89 (0.78, 1.01)  0.88 (0.78, 1.00)        0.94    Postmenopausal  1.05 (1.00, 1.10)  1.06 (1.00, 1.14)        0.39  * Relative risks were adjusted for age at menarche (≤11, 12, 13, 14, ≥15 years), parity (0, 1, 2, ≥3), age at birth of the first child (≤20, 21–25, 26–30, >30 years), postmenopausal hormone use (ever, never), oral contraceptive use (ever, never), history of benign breast disease (no, yes), maternal history of breast cancer (no, yes), history of breast cancer in a sister (no, yes, no sisters), smoking status (ever, never), education (less than high school graduation, high school graduation, more than high school graduation), fat intake (quintiles), fiber intake (quintiles), energy intake (continuous), and alcohol intake (0, >0–<1.5, 1.5–<5, 5–<15, 15–<30, ≥30 g/day). † The Adventist Health Study was not included in analyses of premenopausal women. ‡ The New York State Cohort was not included in this analysis. § The Nurses' Health Studies (a) and (b) were not included in this analysis. ¶ The Sweden Mammography Cohort was not included in this analysis. View Large Table 5 shows results of interaction analyses for other covariates. No significant interactions with height or BMI on breast cancer risk were found except for the interaction between a maternal history of breast cancer and height. Women with a maternal history of breast cancer had a pooled RR of 1.23 per height increment of 5 cm, whereas women without such a history had a pooled RR of 1.07. This interaction was noted in every cohort except the Sweden Mammography Cohort. However, the association with height was opposite (although not significant) in those with a history of breast cancer in a sister, which was consistent across the cohorts. The association of BMI and breast cancer among premenopausal women also showed variation according to the history of breast cancer in a sister. Although the same pattern occurred in every cohort, this interaction was not significant. No interactions were seen between overall family history of breast cancer and height or BMI (data not shown). For height, a borderline significant interaction was seen with a history of benign breast disease, which was apparent in every cohort except the Sweden Mammography Cohort. Although the interaction was not significant, it is of interest, given the possible role of estrogens in the etiology of breast cancer, that the association between BMI and breast cancer was stronger and was significant only among women who never used postmenopausal hormone replacement therapy (HRT). There was also a tendency for the association between BMI and breast cancer in postmenopausal women to decrease with increasing alcohol intake. Analyses with categorical BMI (<21, 21–<25, 25–<28, and ≥28 kg/m2) in the model revealed the same pattern of interaction with age at diagnosis, HRT use, and alcohol as was observed with continuous BMI in postmenopausal women (data not shown). DISCUSSION In this pooled analysis of seven major prospective cohort studies on risk factors for breast cancer, we found a significant positive association between height and the risk of postmenopausal breast cancer. In premenopausal women, the association was less clear and was not significant, and the interaction by menopausal status was not significant. Weight and BMI showed significant inverse associations with risk of premenopausal breast cancer and significant positive associations with postmenopausal breast cancer. This interaction by menopausal status was highly significant. There was no statistically significant heterogeneity among the studies for these associations. The associations of these anthropometric factors were independent of other risk factors for breast cancer and were not significantly modified by most of these factors, except for a possible interaction between height and maternal history of breast cancer. The studies we used are a subset of prospective studies on anthropometric factors and breast cancer risk, but represent all studies with data on food and nutrient intakes that could potentially confound the anthropometric associations. The Pooling Project includes cohort studies with detailed dietary data and other relevant breast cancer risk factors. Because, in the current analysis of anthropometric indices and breast cancer, we wanted to be able to adjust for possible confounding by both nondietary and dietary risk factors, we decided to use the originally collected pooled data set. For example, alcohol intake is a risk factor for breast cancer (16), which also may be associated with BMI. Our pooling results regarding height are in agreement with results of major cohort studies not included that lack reproductive or dietary data (22–24). For BMI, the results among premenopausal women are also comparable (22, 24), but less so in postmenopausal women (24). The availability of information on many potential confounders is a major strength of our pooled analysis to rule out confounding bias, and it also provides an opportunity to examine potential effect modification. A disadvantage is that our analysis of premenopausal breast cancer was based on only four cohorts. Although this comprised over 700 premenopausal cases, the estimates were clearly less precise than were those for postmenopausal women. Another potential drawback of this analysis is the use of self-reported data on height and weight in our cohort studies. Although some attenuation of relative risk estimates might have occurred because of imperfect measurements, the potential for bias is small, considering that validation studies of self-reported measurements show high correlations (r > 0.8) with measured anthropometric data (25–27), and all data here were reported prior to the occurrence of outcomes. Furthermore, various prospective studies in which height and weight were actually measured (e.g., 22, 24, 28, 29) generally reported associations between height, weight, and breast cancer that were consistent with those from prospective studies relying on self-reported data. Selection bias due to incomplete follow-up is unlikely because all cohorts show high degrees of completeness of follow-up. Height and breast cancer A positive association between attained height and breast cancer risk has been found in many earlier studies (3). The positive association with height was less clear for premenopausal than for postmenopausal breast cancer in large (>500 cases) case-control studies (3), but recent, large case-control studies showed comparable positive associations in both menopausal strata (30–34). Two other recent, large case-control studies reported no association in pre- or postmenopausal women (35, 36), but others reported a significant positive association among young premenopausal women (37). Most cohort studies (13, 15, 23, 24, 28, 29, 38–40), but not all (41–43), reported a positive association between height and breast cancer. Several of these studies were able to compare menopausal strata and reported weaker associations among premenopausal women (13, 24, 29, 38). Our pooled analysis showed an overall positive association with premenopausal breast cancer, but the effect was less clear than that in postmenopausal cases and showed more variability among studies. In our analysis of effect modification of height and breast cancer, we found a significant interaction only with a maternal history of breast cancer, with a much stronger association among women with a positive maternal history. Although this is in agreement with earlier findings regarding overall family history of breast cancer (44, 45), it is not consistent with our opposite finding of a lower degree of risk among women with a sister with breast cancer. The directions of the effect modification were quite consistent across the cohorts, suggesting that these interactions may not be due to chance. These interactions merit further attention. Because attained height might be an indicator of childhood energy intake in situations in which there is sufficient variation in energy intake (2), this has been proposed as an explanation for the observation that the positive association between height and breast cancer was found more frequently in European populations that experienced food shortages during World War II (13, 24). Indeed, stronger associations have generally been found in studies of the affected birth cohorts in Norway, the Channel Islands, and the Netherlands, where severe food deprivation occurred at the end of the War (15, 24, 29, 40) compared with US cohort studies (3). It is also consistent with the prior observations that associations were found in US Black women (33) as opposed to White women (3636) and in US cohorts with overrepresentation of women potentially undernourished during childhood or adolescence (28), as well as with many animal studies showing that energy restriction clearly reduces experimentally induced and spontaneous mammary tumor risk in rodents (46). However, more recent studies show that height is also rather consistently related to breast cancer risk in relatively affluent populations that have not experienced energy restriction during growth periods and in which height is presumably primarily genetically determined (30, 31, 37, 47). It has also been hypothesized that height reflects mammary gland mass (or more precisely, the number of ductal stem cells that develop in the breast in utero), which could be related to breast cancer risk (48). Considering that nutritional inadequacy does not seem to be the only factor responsible for the association between height and breast cancer, Ballard-Barbash (45) has proposed that genetic and environmental factors (e.g., diet and physical activity) may affect the hormones that influence epiphyseal closure during puberty and, thus, attained height. Furthermore, Stoll (49) has suggested that better nutrition accelerates growth hormone release which, in turn, increases levels of insulin-like growth factor (IGF). The adolescent growth spurt involves stimulation by growth hormone, insulin, IGF, and sex steroids, and Stoll hypothesizes that the combination of IGF and sex steroids results in mitogenic effects on developing mammary tissue in adolescence and a concomitant increased risk of epithelial atypia and carcinogenesis. This is compatible with the observation (50) that plasma levels of IGF-1 predicted the incidence of premenopausal, but not postmenopausal, breast cancer. Dorgan et al. (51) have suggested, on the basis of a cross-sectional study, that height might influence breast cancer risk through its positive association with follicular-phase plasma estradiol. Weight and premenopausal breast cancer Our findings of an inverse association between relative weight and premenopausal breast cancer confirm those of most earlier studies in Western, high-risk countries (13, 30, 31, 35, 37, 52–61). Ziegler et al. (34) also recently reported a decreased risk among heavy, young (age <40 years) women in a case-control study among Asian-American women living in the western United States. In our pooled analysis, the strongest inverse associations were found for women with a BMI of more than 31 kg/m2. There was no significant heterogeneity between the studies in this respect. Earlier reports also showed that the inverse association is frequently limited to the higher relative weight categories but, due to the differences in relative weight distributions and the widely differing categorizations that were used, no clear cutoff value for BMI has emerged above which there is a decreased breast cancer risk. For example, significantly decreased relative risks have been reported for women with a BMI of more than 21 (13), more than 22 (37), more than 28.8 (35), and 30 kg/m2 or more (62). When we used an upper BMI category of 29 kg/m2 or more, we also found a significantly decreased risk of 0.68 (95 percent CI: 0.48, 0.97) (data not shown). However, our extended analyses indicated that the inverse association is limited mainly to women with a BMI of 31 kg/m2 or more. Thus, although there seems to be no uniformity in the literature regarding a BMI value above which breast cancer risk decreases, the overall inverse association as such is a rather consistent finding in cohort studies and also in recent case-control studies, particularly among younger premenopausal women (30, 34, 35, 37, 63). In spite of this consistency, explanations for this finding are far from satisfactory. Detection bias could be responsible for this effect, but this was concluded to be unlikely (31, 46, 64, 65). Pathak and Whittemore (58) compared countries with high, medium, and low breast cancer risk and concluded that it is unlikely that detection of breast cancer would be masked by obesity in high-risk countries and not in medium-risk populations, where positive associations with relative weight have been found. The inverse association in case-control studies also is not likely to be due to disease or treatment effects (37). It has been proposed that, in premenopausal women, obesity may protect against breast cancer by causing more frequent anovulatory menstrual cycles (66, 67). This would result in decreased estradiol and progesterone levels and lower luteal phase progesterone levels in ovulatory cycles (68). However, the level of obesity needed to induce sufficient anovulatory cycles so that breast cancer risk is decreased is unclear (69). Interestingly, the association with BMI was somewhat less strong among women who had ever used oral contraceptives and had thus experienced anovulatory cycles due to oral contraceptive use. Swanson et al. (37) reported that irregularity of menstrual cycles, as a crude indicator of anovulatory cycles, could not explain the reduced breast cancer risk among obese women. Few population-based studies have looked at the relation between BMI and menstrual cycle length or variability or bleeding patterns (70–74). In Nurses Health Study II, a U-shaped relation between BMI and menstrual irregularity was observed, with most regularity between a BMI of 18 and 22 kg/m2 (74). In the same study, both short and long/irregular cycles were associated with reduced breast cancer risk (75). Because the pooled cohorts have few women with a BMI of less than 18 kg/m2, the inverse association here is seen only in high BMI categories, which supports the hypothesis that more anovulatory cycles lead to reduced premenopausal breast cancer risk. As shown before (66), Potishman et al. (76) found that sex hormone-binding globulin levels decreased with increasing BMI in both pre- and postmenopausal women. However, with increasing BMI, total estradiol concentrations decreased in premenopausal women, but increased in postmenopausal women. They propose that lower estradiol levels in premenopausal, obese women are the result of increased estradiol clearance due to reduced serum-hormone binding capacity. Weight and postmenopausal breast cancer The significant positive association between relative weight and postmenopausal breast cancer in this pooled analysis is consistent with results of many case-control studies (3). Cohort studies have often shown weak positive or no associations among postmenopausal women (13–15, 22, 23, 29, 39, 42, 43), although significant positive associations have also been reported (24, 47, 63). Our current results, combining several of the above studies plus additional cohorts, show a modest, significantly positive association. Although no association was seen in two cohorts (Netherlands Cohort Study and Nurses' Health Study (a)), no statistically significant heterogeneity was observed in this pooled analysis. Several studies indicate that the relation between relative weight and breast cancer risk might be stronger in older postmenopausal women (35, 63). In our current analysis, we saw no significant effect modification by age at diagnosis, but there was some suggestion that women older than age 65 years at diagnosis showed stronger associations with BMI than did younger women. Since most published cohort studies have not included large numbers of older women, this interaction may become more apparent as the cohorts mature. Fat distribution has been proposed as more predictive of breast cancer risk than is body mass (59, 77). However, waist-to-hip circumference ratio as a measure of abdominal obesity or other indices of fat distribution has not been consistently associated with breast cancer risk (14, 35, 77). In postmenopausal women, ovarian estrogen production is diminished, and estrogen, which may promote tumor growth, is derived mainly from the aromatization of androstenedione that occurs primarily in adipose tissue. Perhaps high estrogen levels could occur near breast adipose tissue (34). Furthermore, excess weight is associated with decreased sex hormone-binding globulin levels (78), resulting in increased levels of biologically active estrogens (45). Higher levels of estrone and estradiol have been found in obese postmenopausal women compared with women of normal weight (79, 80). An upper threshold for the effect of BMI, such as we observed above 28 kg/m2, is also consistent with estrogen receptor-mediated effects. Our results on potential effect modification by HRT of the association between BMI and postmenopausal breast cancer are in accordance with the elevated extraovarian estrogen hypothesis. Although no significant interaction was noted, possibly due to lack of update of HRT use in most cohorts, the BMI-breast cancer association was stronger and was significant only among women who never had used HRT, i.e., among women in whom estrogen exposure is determined only by endogenous estrogen production. Recently, Huang et al. (81) reported a similar interaction between BMI and HRT, which was statistically significant. Moderately strong positive associations between plasma estrogen levels and postmenopausal breast cancer risk have been reported recently (82), especially in nonusers of HRT. Alcohol intake is positively associated with risk of breast cancer (16), possibly through elevation of estrogen levels. As with HRT as exogenous source of estrogens, the BMI-breast cancer association was diminished in alcohol drinkers and was significant only among nondrinkers in our pooled analysis, although the interaction was not significant. In contrast to an earlier report (44), we observed no indication that family history of breast cancer modifies the relation between BMI and postmenopausal breast cancer risk, as was also found recently by others (30). In conclusion, this study, based on over 4,000 incident breast cancer cases, provides further evidence of a modest positive association between adult height and breast cancer, suggesting a possible etiologic role for factors (including diet) operating in early life on breast cancer risk later in life. This study also provides further evidence for moderate inverse and positive trends in breast cancer risk associated with relative weight in pre- and postmenopausal breast cancer, respectively. Further research is warranted regarding the timing of the change in direction of the effect of excess weight in relation to a woman's age. Because weight is one of the few modifiable breast cancer risk factors, weight control provides an important opportunity for prevention of postmenopausal breast cancer. Reprint requests to Prof. Piet A. van den Brandt, Department of Epidemiology, Maastricht University, P. O. Box 616, 6200 MD Maastricht, the Netherlands (e-mail: PA.vandenBrandt@epid.unimaas.nl). Supported by research grants NIH CA55075 and CA50597 and by a Faculty Research Award (FRA-455) to Dr. David J. Hunter from the American Cancer Society. The authors thank Jolanda Nelissen for assistance in preparing the manuscript. REFERENCES 1. de Waard F, Trichopoulos D. A unifying concept of the aetiology of breast cancer. Int J Cancer  1988; 41: 666–9. Google Scholar 2. Willett W. Nutritional epidemiology. New York, NY: Oxford University Press, 1990. Google Scholar 3. Hunter DJ, Willett WC. Diet, body size, and breast cancer. Epidemiol Rev  1993; 15: 110–32. Google Scholar 4. Ursin G, Longnecker MP, Haile RW, et al. A meta-analysis of body mass index and risk of premenopausal breast cancer. Epidemiology  1995; 6: 137–41. Google Scholar 5. Hunter DJ, Spiegelman D, Adami H-O, et al. Cohort studies of fat intake and the risk of breast cancer–a pooled analysis. N Engl J Med  1996; 334: 356–61. Google Scholar 6. Kushi LH, Sellers TA, Potter JD, et al. Dietary fat and postmenopausal breast cancer. J Natl Cancer Inst  1992; 84: 1092–9. Google Scholar 7. Mills PK, Beeson WL, Philips RL, et al. Dietary habits and breast cancer incidence among Seventh-day Adventists. Cancer  1989; 64: 582–90. Google Scholar 8. Howe GR, Friedenreich CM, Jain M, et al. A cohort study of fat intake and risk of breast cancer. J Natl Cancer Inst  1991; 83: 336–40. Google Scholar 9. Willett WC, Hunter DJ, Stampfer MJ, et al. Dietary fat and fiber in relation to risk of breast cancer: an 8-year follow-up. JAMA  1992; 268: 2037–44. Google Scholar 10. Graham S, Zielezny M, Marshall J, et al. Diet in the epidemiology of postmenopausal breast cancer in the New York State cohort. Am J Epidemiol  1992; 136: 1327–37. Google Scholar 11. van den Brandt PA, van 't Veer P, Goldbohm RA, et al. A prospective cohort study on dietary fat and the risk of postmenopausal breast cancer. Cancer Res  1993; 53: 75–82. Google Scholar 12. Holmberg L, Ohlander EM, Byers T, et al. Diet and breast cancer risk: results from a population-based, case-control study in Sweden. Arch Intern Med  1994; 154: 1805–11. Google Scholar 13. London SJ, Colditz GA, Stampfer MJ, et al. Prospective study of relative weight, height, and risk of breast cancer. JAMA  1989; 262: 2853–8. Google Scholar 14. 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. Google Scholar 15. van den Brandt PA, Dirx MJM, Ronckers CM, et al. Height, weight, weight change, and postmenopausal breast cancer risk: the Netherlands Cohort Study. Cancer Causes Control  1997; 8: 39–47. Google Scholar 16. Smith-Warner SA, Spiegelman D, Yaun S-S, et al. Alcohol and breast cancer in women: a pooled analysis of cohort studies. JAMA  1998; 279: 535–40. Google Scholar 17. Prentice RL. A case-cohort design for epidemiologic cohort studies and disease prevention trials. Biometrika  1986; 73: 1–11. Google Scholar 18. SAS/STAT software: the PHREG procedure: preliminary documentation. Cary, NC: SAS Institute, Inc., 1991. Google Scholar 19. Epicure user's guide: the PEANUTS program. Seattle, WA: Hirosoft, 1993. Google Scholar 20. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials  1986; 7: 177–88. Google Scholar 21. Hunter DJ, Spiegelman D, Adami H-O, et al. Non-dietary factors as risk factors for breast cancer, and as effect modifiers of the association of fat intake and risk of breast cancer. Cancer Causes Control  1997; 8: 49–56. Google Scholar 22. Vatten LJ, Kvinnsland S. Body mass index and risk of breast cancer: a prospective study of 23,826 Norwegian women. Int J Cancer  1990; 45: 440–4. Google Scholar 23. Törnberg SA, Holm LE, Carstensen JM. Breast cancer risk in relation to serum cholesterol, serum beta-lipoprotein, height, weight, and blood pressure. Acta Oncol  1988; 27: 31–7. Google Scholar 24. 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. Google Scholar 25. Stunkard AJ, Albaum JM. The accuracy of self-reported weights. Am J Clin Nutr  1981; 34: 1593–9. Google Scholar 26. Weaver TW, Kushi LH, McGovern PG, et al. Validation study of self-reported measures of fat distribution. Int J Obes Relat Metab Disord  1996; 20: 644–50. Google Scholar 27. Rimm EB, Stampfer MJ, Colditz GA, et al. Validity of self-reported waist and hip circumferences in men and women. Epidemiology  1990; 1: 466–73. Google Scholar 28. 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: 363–7. Google Scholar 29. 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. Google Scholar 30. Trentham-Dietz A, Newcomb PA, Storer BE, et al. Body size and risk of breast cancer. Am J Epidemiol  1997; 145: 1011–19. Google Scholar 31. Brinton LA, Swanson CA. Height and weight at various ages and risk of breast cancer. Ann Epidemiol  1992; 2: 597–609. Google Scholar 32. 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. Google Scholar 33. 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. Google Scholar 34. Ziegler RG, Hoover RN, Nomura AM, et al. Relative weight, weight change, height, and breast cancer risk in Asian-American women. J Natl Cancer Inst  1996; 88: 650–60. Google Scholar 35. 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. Google Scholar 36. 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. Google Scholar 37. 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. Google Scholar 38. 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. Google Scholar 39. de Waard F, Baanders-van Halewijn EA. A prospective study in general practice on breast cancer risk in postmenopausal women. Int J Cancer  1974; 14: 153–60. Google Scholar 40. 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. Google Scholar 41. 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. Google Scholar 42. Goodman MT, Cologne JB, Moriwaki H, et al. Risk factors for primary breast cancer in Japan: 8-year follow-up of atomic bomb survivors. Prev Med  1997; 26: 144–53. Google Scholar 43. den Tonkelaar I, Seidell JC, Collette HJA, et al. A prospective study on obesity and subcutaneous fat patterning in relation to breast cancer in postmenopausal women participating in the DOM-project. Br J Cancer  1994; 69: 352–7. Google Scholar 44. Sellers TA, Kushi LH, Potter JD, et al. Effect of family history, body-fat distribution, and reproductive factors on the risk of postmenopausal breast cancer. N Engl J Med  1992; 326: 1323–9. Google Scholar 45. Ballard-Barbash R. Anthropometry and breast cancer. Body size–a moving target. Cancer  1994; 74: 1090–100. Google Scholar 46. Albanes D. Caloric intake, body weight, and cancer: a review. Nutr Cancer  1987; 9: 199–217. Google Scholar 47. Swanson CA, Brinton LA, Taylor PR, et al. Body size and breast cancer risk assessed in women participating in the Breast Cancer Detection Demonstration Project. Am J Epidemiol  1989; 130: 1133–41. Google Scholar 48. Trichopoulos D, Lipman RD. Mammary gland mass and breast cancer risk. Epidemiology  1992; 3: 523–6. Google Scholar 49. Stoll BA. Does extra height justify a higher risk of breast cancer? Ann Oncol  1992; 3: 29–30. Google Scholar 50. 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. Google Scholar 51. Dorgan JF, Reichman ME, Judd JT, et al. The relation of body size to plasma levels of estrogens and androgens in premenopausal women (Maryland, United States). Cancer Causes Control  1995; 6: 3–8. Google Scholar 52. Paffenbarger RS Jr, Kampert JB, Chang H-G. Characteristics that predict risk of breast cancer before and after the menopause. Am J Epidemiol  1980; 112: 258–68. Google Scholar 53. Willett WC, Browne ML, Bain C, et al. Relative weight and risk of breast cancer among premenopausal women. Am J Epidemiol  1985; 122: 731–40. Google Scholar 54. Helmrich SP, Shapiro S, Rosenberg L, et al. Risk factors for breast cancer. Am J Epidemiol  1983; 117: 35–45. Google Scholar 55. Kampert JB, Whittemore AS, Paffenbarger RS Jr. Combined effect of childbearing, menstrual events, and body size on age-specific breast cancer risk. Am J Epidemiol  1988; 128: 962–79. Google Scholar 56. Bouchardy C, Le MG, 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. Google Scholar 57. Vatten LJ, Kvinnsland S. Prospective study of height, body mass index and risk of breast cancer. Acta Oncol  1992; 31: 195–200. Google Scholar 58. Pathak DR, Whittemore AS. Combined effects of body size, parity, and menstrual events on breast cancer incidence in seven countries. Am J Epidemiol  1992; 135: 153–68. Google Scholar 59. Bruning PF, Bonfrer JMG, Hart AAM, et al. Body measurements, estrogen availability and the risk of human breast cancer: a case-control study. Int J Cancer  1992; 51: 14–19. Google Scholar 60. Adami HO, Rimsten Å, 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. Google Scholar 61. Törnberg SA, Carstensen JM. Relationship between Quetelet's index and cancer of breast and female genital tract in 47,000 women followed for 25 years. Br J Cancer  1994; 69: 358–61. Google Scholar 62. Chu SY, Lee NC, Wingo PA, 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: 1197–206. Google Scholar 63. Yong LC, Brown CC, Schatzkin A, et al. Prospective study of relative weight and risk of breast cancer: the Breast Cancer Detection Demonstration Project follow-up study, 1979 to 1987. Am J Epidemiol  1996; 143: 985–95. Google Scholar 64. Albanes D. Energy balance, body size and cancer. Crit Rev Oncol Hematol  1990; 10: 283–303. Google Scholar 65. Toti A, Agugiaro S, Amadori D, et al. Breast cancer risk factors in Italian women: a multicentric case-control study. Tumori  1986; 72: 241–9. Google Scholar 66. Sherman BM, Korenman SG. Measurement of serum LH, FSH, estradiol and progesterone in disorders of the human menstrual cycle: the inadequate luteal phase. J Clin Endocrinol Metab  1974; 39: 145–9. Google Scholar 67. Stoll BA. Breast cancer: the obesity connection. Br J Cancer  1994; 69: 799–801. Google Scholar 68. Key TJA, Pike MC. The role of oestrogens and progestagens in the epidemiology and prevention of breast cancer. Eur J Cancer Clin Oncol  1988; 24: 29–43. Google Scholar 69. 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. Google Scholar 70. Harlow SD, Matanoski GM. The association between weight, physical activity, and stress and variation in the length of the menstrual cycle. Am J Epidemiol  1991; 133: 38–49. Google Scholar 71. Harlow SD, Campbell BC. Host factors that influence the duration of menstrual bleeding. Epidemiology  1994; 5: 353–5. Google Scholar 72. Belsey EM, d'Arcangues C, Carlson N. Determinants of menstrual bleeding patterns among women using natural and hormonal methods of contraception. II. The influence of individual characteristics. Contraception  1988; 38: 243–57. Google Scholar 73. Cooper GS, Sandler DP, Whelan EA, et al. Association of physical and behavioral characteristics with menstrual cycle patterns in women age 29–31 years. Epidemiology  1996; 7: 624–8. Google Scholar 74. Rich-Edwards JW, Goldman MB, Willett WC, et al. Adolescent body mass index and infertility caused by ovulatory disorder. Am J Obstet Gynecol  1994; 171: 171–7. Google Scholar 75. Garland M, Hunter DJ, Colditz GA, et al. Menstrual cycle characteristics and history of ovulatory infertility in relation to breast cancer risk in a large cohort of US women. Am J Epidemiol  1998; 147: 636–43. Google Scholar 76. Potishman N, Swanson CA, Siiteri P, et al. Reversal of relation between body mass and endogenous estrogen concentrations with menopausal status. J Natl Cancer Inst  1996; 88: 756–8. Google Scholar 77. Stoll BA. Obesity and breast cancer. Int J Obes Relat Metab Disord  1996; 20: 389–92. Google Scholar 78. Moore JW, Key TJ, Bulbrook RD, et al. Sex hormone binding globulin and risk factors for breast cancer in a population of normal women who had never used exogenous sex hormones. Br J Cancer  1987; 56: 661–6. Google Scholar 79. Cauley JA, Gutai JP, Kuller LH, et al. The epidemiology of serum sex hormones in postmenopausal women. Am J Epidemiol  1989; 129: 1120–31. Google Scholar 80. Hankinson SE, Willett WC, Manson JE, et al. Alcohol, height, and adiposity in relation to estrogen and prolactin levels in postmenopausal women. J Natl Cancer Inst  1995; 87: 1297–302. Google Scholar 81. Huang Z, Hankinson SE, Colditz GA, et al. Dual effects of weight and weight gain on breast cancer risk. JAMA  1997; 278: 1407–11. Google Scholar 82. Hankinson SE, Willett WC, Manson JE, et al. Plasma sex steroid hormone levels and risk of breast cancer in postmenopausal women. J Natl Cancer Inst  1998; 90: 1292–9. Google Scholar
TI - Pooled Analysis of Prospective Cohort Studies on Height, Weight, and Breast Cancer Risk
JO - American Journal of Epidemiology
DO - 10.1093/aje/152.6.514
DA - 2000-09-15
UR - https://www.deepdyve.com/lp/oxford-university-press/pooled-analysis-of-prospective-cohort-studies-on-height-weight-and-4GNe0Lt1f8
SP - 514
EP - 527
VL - 152
IS - 6
DP - DeepDyve
ER -