Background: Increased body mass index (BMI) is an established colorectal cancer risk factor. High waist circumference or waist-hip-ratio (WHR) may better reflect an abnormal metabolic state and be more predictive of colorectal cancer risk than BMI.

Methods: We examined BMI, waist circumference, WHR, and hip circumference in relation to colorectal cancer risk among 203,177 participants followed for 10 years. We derived standardized colorectal cancer risk estimates for each anthropometric parameter and compared predictive characteristics (Harrell's C-index). In women, we examined whether hormone replacement therapy (HRT) use modified the associations between anthropometric measures and colorectal cancer.

Results: We ascertained 2,869 colorectal cancers. In men, increased colon cancer risks were associated with BMI [HR per SD, 1.14; 95% confidence interval (CI), 1.08–1.20], waist circumference (HR per SD, 1.17; 95% CI, 1.08–1.27), and WHR (HR per SD, 1.09; 95% CI, 1.04–1.14). In women, anthropometric variables were unrelated to colon cancer. For men and women, anthropometric variables were unrelated to rectal cancer. Compared with BMI, waist circumference and WHR did not materially influence colon cancer prediction models [C-index changes: −0.0041 and 0.0046 (men); 0.0004 and 0.0005 (women)]. In current HRT users, colon cancer was inversely or suggestively inversely associated with waist circumference (HR per SD, 0.78; 95% CI, 0.63–0.97) and WHR (HR per SD, 0.88; 95% CI, 0.76–1.01), but positively related to hip circumference (HR per SD, 1.39; 95% CI, 1.13–1.71).

Conclusion: BMI, waist circumference, and WHR show comparable positive associations with colon cancer in men. Associations between anthropometric measures and colon cancer are weak or null in women, but there is some evidence for effect modification by HRT.

Impact: These findings may improve our understanding of the relation of adiposity to colorectal cancer. Cancer Epidemiol Biomarkers Prev; 22(8); 1383–94. ©2013 AACR.

Increased body mass index (BMI), as an approximation of total body adiposity, is an established colorectal cancer risk factor. Our systematic review and meta-analysis of 29 prospective studies reported relative risks for colon cancer per 5 kg/m2 BMI increment of 1.24 and 1.09 in men and women, respectively, and that associations are greater for colon versus rectal cancer (1).

Few prospective studies have evaluated relations between abdominal adiposity, approximated by waist circumference and waist-to-hip ratio (WHR), and colorectal cancer risk.

Five cohorts in men (2–6) and four in women (2, 4, 7, 8) found positive associations between high versus low waist circumference and/or WHR and colon cancer risk, with summarized relative risks of 1.81 in men and 1.50 in women (9). A small number of studies (3–7) have suggested that measures of abdominal adiposity may be more predictive for colon cancer risk than BMI.

This supposition is appealing as hypothesized biologic mechanisms underpinning the link between adiposity and colon cancer include insulin resistance and metabolic dysfunction (10), and conventionally, abdominal adiposity is considered to be more closely related with insulin resistance than total body fat (11), and may be a stronger predictor for cardiovascular disease and type II diabetes than BMI (12, 13). However, as pointed out by Huxley and colleagues (14), this comparison is not trivial—it requires comparisons of equivalent population-specific standardized increments per anthropometric measure, and formal discrimination testing. Furthermore, waist circumference is a composite of subcutaneous and visceral adipose tissues (VAT)—proportions of which vary by age and gender—and the latter being most strongly related to insulin resistance and metabolic dysfunction. Finally, there is a paucity of prospective data on the association between hip circumference and colorectal cancer risk. A higher hip circumference may be associated with reduced cardiovascular risk (15, 16) but parallel hypotheses have seldom been explored in relation to cancer risk.

For women, associations between measures of abdominal adiposity and colon cancer risk may be weaker than in men (4, 5), raising a hypothesis that additional sex-related biologic mechanisms may prevail. Data from the Women's Health Initiative randomized trial showed that hormonal replacement therapy (HRT) use is associated with reduced colon cancer risk (17), and in turn, the European Prospective Investigation into Cancer and Nutrition (EPIC) cohort reported that associations between BMI and colon cancer risk in women may be attenuated (to null) in the presence of HRT use (2).

The primary objective of the current study was to derive standardized colorectal cancer risk estimates for each anthropometric parameter and compare predictive characteristics with respect to colorectal cancer risk. The secondary objective tested the hypothesis that the associations between anthropometric measures and colorectal cancer risk in women vary according to HRT use. We extend previous studies of associations of BMI (18) and weight gain (19) in relation to colorectal cancer risk using the NIH-AARP Diet and Health Study, in which comprehensive, prospectively collected data for anthropometric exposures and HRT are available.

Study population and case identification

The NIH-AARP Diet and Health Study is a large prospective cohort study of AARP members, recruited in 1995 to 1996 from 6 U.S. states (California, Florida, Pennsylvania, New Jersey, North Carolina, and Louisiana) and two metropolitan areas (Atlanta, Georgia; and Detroit, Michigan). A self-administered baseline questionnaire that requested information on diet, physical activity, anthropometry, chronic diseases, and lifestyle factors was completed by a total of N = 566,401 individuals, ages 50 to 71 years. A subcohort of N = 334,907 participants completed a second questionnaire mailed in 1996, which included determination of waist circumference and hip circumference, which in turn, were used to derive WHR. The Special Studies Institutional Review Board (IRB) of the U.S. National Cancer Institute, Rockville, MD approved the study.

Of participants who completed the second questionnaire, we excluded those with a previous diagnosis of cancer at baseline (N = 14,565) and before the second questionnaire (N = 4,295). Furthermore, we excluded participants without data on height (N = 2,766), weight (N = 3,513), waist circumference (N = 73,055), or hip circumference (N = 11,015). We also excluded subjects with extreme values (defined as more than three interquartiles ranges outside the 75th and 25th percentiles) of waist circumference or hip circumference (N = 3,037) and those with BMI less than 18.5 kg/m2 (N = 2,342) or more than 60 kg/m2 (N = 32). As smoking is an important risk factor for colon cancer, we also excluded N = 6,618 participants with missing information on smoking status. We also excluded N = 10,383 proxy respondents and N = 109 persons with less than 18 days of follow-up. After exclusions, this subcohort comprised 203,177 participants (124,208 men and 78,969 women) and formed the basis for the present analysis. We determined that the characteristics of this subcohort were similar to those of the overall cohort (Supplementary Table S1).

Study participants were followed up through December 31, 2006 by periodic linkage of the cohort to the National Change of Address database maintained by the U.S. Postal Service, other address change update services, and from cohort members' notifications. Incident colon and rectal cancers were identified through 10 state cancer registries (including Texas and Nevada) using the International Classification of Diseases for Oncology (ICD-O-3) codes for colon cancer (C18.0–C18.9) and rectal cancer (C19 and C20).

Quality control studies report that there is only 5% loss of follow-up in this cohort (20). Vital status was ascertained by linking the cohort to the U.S. Social Security Administration Death Master File. Follow-up searches of presumed deaths in the National Death Index (NDI) Plus were used to verify deaths and to gain information on the cause of death.

Exposure ascertainment

Self-reported body weight and height were assessed in the baseline questionnaire; waist and hip circumferences were obtained from the second questionnaire. Participants were instructed to measure their waist circumference and hip circumference using a tape measure to the nearest 0.25 inch while standing. Waist circumference was to be measured 1 inch above the navel, if this was not the waistline. Hip circumference was defined as the largest circumference between the upper edge of the pelvis and the femur. In a cohort study contemporaneous to ours, comparisons between self-reported and technician-measured waist circumference have yielded correlation coefficients of 0.80 in men and 0.83 in women and for hip circumference have yielded values of 0.74 in men and 0.86 in women, and deemed sufficiently accurate for identifying relationships in epidemiologic studies (21).

Correlations between self-reported and technician-measured BMI values exceed 0.9 for both men and women (22).

Statistical analysis

We used age as the underlying time metric from the age of return of the second questionnaire to the age of diagnosis of colon or rectal cancer, move out of the registry ascertainment area, death, or December 31, 2006, whichever came first. The relationships between BMI, waist circumference, WHR, and hip circumference were evaluated using Spearman correlation coefficients in men and women separately. Cox proportional hazards regression was used to estimate the HRs and 95% confidence intervals (CIs) for colon and rectal cancer, separately, according to gender-specific quintiles of BMI, waist circumference, WHR, and hip circumference. In addition, participants were categorized into predefined World Health Organization (WHO) categories for waist circumference and WHR (23).

Three models were constructed to examine the associations between BMI, waist circumference, WHR, hip circumference, and colon and rectal cancer risk. In model 1, we adjusted for age only. In model 2—the main model—we additionally adjusted for multiple lifestyle factors (listed in footnote to Table 2). Height at baseline was included as a continuous variable in analyses of waist circumference and hip circumference, and waist circumference and hip circumference were mutually adjusted to assess whether each were independently related to colon or rectal cancer risk. In model 3, we additionally adjusted for BMI (continuous) to assess the impact of abdominal fat distribution on colon and rectal cancer independent of its relation with general adiposity.

Table 2.

HRs and 95% CIs for risk of incident colon cancer by gender across categories of anthropometric measures in the NIH-AARP Diet and Health Study

MenWomen
MeasurePYCasesaModel 1Model 2Model 3PYCasesaModel 1Model 2Model 3
BMIb (quintile) 
227,769 242 1.0 1.0  147,050 125 1.0 1.0  
228,277 255 1.06 (0.89–1.26) 1.04 (0.87–1.24)  148,971 133 1.03 (0.81–1.32) 1.04 (0.81–1.32)  
229,123 281 1.18 (0.99–1.40) 1.13 (0.95–1.35)  148,758 135 1.03 (0.81–1.32) 1.01 (0.79–1.29)  
227,671 330 1.40 (1.19–1.66) 1.31 (1.10–1.55)  147,802 143 1.10 (0.87–1.40) 1.02 (0.80–1.31)  
225,231 355 1.59 (1.35–1.87) 1.42 (1.19–1.68)  148,408 144 1.14 (0.90–1.45) 0.96 (0.74–1.23)  
Ptrendd   <0.001 <0.001    0.251 0.477  
Per SD   1.19 (1.13–1.25) 1.14 (1.08–1.20)    1.05 (0.97–1.13) 0.97 (0.90–1.05)  
WC (quintile) 
248,658 223 1.0 1.0 1.0 141,001 114 1.0 1.0 1.0 
184,690 213 1.24 (1.03–1.49) 1.18 (0.97–1.43) 1.15 (0.95–1.40) 155,739 122 0.91 (0.71–1.18) 0.90 (0.70–1.17) 0.91 (0.70–1.18) 
235,126 303 1.38 (1.16–1.63) 1.26 (1.05–1.52) 1.22 (1.00–1.47) 132,999 119 1.01 (0.78–1.31) 0.95 (0.73–1.25) 0.97 (0.74–1.28) 
227,550 314 1.46 (1.23–1.74) 1.28 (1.05–1.55) 1.21 (0.99–1.49) 163,568 176 1.19 (0.94–1,51) 1.06 (0.81–1.40) 1.10 (0.83–1.45) 
242,048 410 1.83 (1.56–2.16) 1.45 (1.16–1.82) 1.32 (1.03–1.70) 147,683 149 1.14 (0.89–1.45) 0.90 (0.63–1.27) 0.95 (0.66–1.37) 
Ptrendd   <0.001 <0.001 0.007   0.042 0.838 0.859 
Per SD   1.24 (1.18–1.30) 1.17 (1.08–1.27) 1.14 (1.04–1.26)   1.08 (1.00–1.16) 0.99 (0.88–1.12) 1.01 (0.89–1.16) 
WHRc (quintile) 
227,096 245 1.0 1.0 1.0 149,032 121 1.0 1.0 1.0 
236,193 292 1.11 (0.93–1.31) 1.08 (0.91–1.28) 1.06 (0.89–1.25) 151,823 126 0.96 (0.75–1.23) 0.92 (0.72–1.18) 0.92 (0.72–1.19) 
223,657 267 1.06 (0.89–1.26) 1.01 (0.85–1.21) 0.98 (0.82–1.16) 146,523 140 1.06 (0.83–1.36) 0.98 (0.77–1.26) 0.99 (0.77–1.27) 
205,108 264 1.14 (0.96–1.35) 1.06 (0.89–1.27) 1.00 (0.84–1.20) 147,387 140 1.03 (0.81–1.32) 0.90 (0.70–1.15) 0.91 (0.70–1.17) 
246,016 395 1.46 (1.24–1.71) 1.29 (1.10–1.52) 1.17 (0.99–1.38) 146,225 153 1.12 (0.88–1.42) 0.90 (0.70–1.15) 0.91 (0.70–1.18) 
Ptrendd   <0.001 <0.001 0.044   0.045 0.997 0.836 
Per SD   1.13 (1.08–1.19) 1.09 (1.04–1.14) 1.05 (1.00–1.11)   1.08 (1.00–1.17) 1.00 (0.92–1.08) 1.01 (0.93–1.10) 
HC (quintile) 
217,058 243 1.0 1.0 1.0 146,956 131 1.0 1.0 1.0 
255,289 249 0.84 (0.71–1.01) 0.78 (0.65–0.93) 0.77 (0.65–0.93) 150,767 121 0.90 (0.70–1.15) 0.93 (0.72–1.19) 0.94 (0.73–1.21) 
235,050 307 1.13 (0.95–1.34) 0.95 (0.79–1.14) 0.94 (0.78–1.13) 146,136 143 1.07 (0.85–1.36) 1.10 (0.85–1.43) 1.13 (0.87–1.48) 
212,115 307 1.25 (1.06–1.48) 0.96 (0.78–1.17) 0.95 (0.77–1.16) 163,789 151 1.01 (0.80–1.27) 1.01 (0.76–1.33) 1.06 (0.78–1.43) 
218,559 357 1.43 (1.21–1.68) 0.91 (0.72–1.16) 0.89 (0.70–1.14) 133,341 134 1.13 (0.89–1.44) 1.04 (0.74–1.48) 1.15 (0.77–1.71) 
Ptrendd   <0.001 0.952 0.829   0.201 0.940 0.557 
Per SD   1.18 (1.12–1.24) 1.00 (0.93–1.09) 0.99 (0.91–1.08)   1.05 (0.97–1.13) 1.01 (0.89–1.13) 1.04 (0.90–1.21) 
MenWomen
MeasurePYCasesaModel 1Model 2Model 3PYCasesaModel 1Model 2Model 3
BMIb (quintile) 
227,769 242 1.0 1.0  147,050 125 1.0 1.0  
228,277 255 1.06 (0.89–1.26) 1.04 (0.87–1.24)  148,971 133 1.03 (0.81–1.32) 1.04 (0.81–1.32)  
229,123 281 1.18 (0.99–1.40) 1.13 (0.95–1.35)  148,758 135 1.03 (0.81–1.32) 1.01 (0.79–1.29)  
227,671 330 1.40 (1.19–1.66) 1.31 (1.10–1.55)  147,802 143 1.10 (0.87–1.40) 1.02 (0.80–1.31)  
225,231 355 1.59 (1.35–1.87) 1.42 (1.19–1.68)  148,408 144 1.14 (0.90–1.45) 0.96 (0.74–1.23)  
Ptrendd   <0.001 <0.001    0.251 0.477  
Per SD   1.19 (1.13–1.25) 1.14 (1.08–1.20)    1.05 (0.97–1.13) 0.97 (0.90–1.05)  
WC (quintile) 
248,658 223 1.0 1.0 1.0 141,001 114 1.0 1.0 1.0 
184,690 213 1.24 (1.03–1.49) 1.18 (0.97–1.43) 1.15 (0.95–1.40) 155,739 122 0.91 (0.71–1.18) 0.90 (0.70–1.17) 0.91 (0.70–1.18) 
235,126 303 1.38 (1.16–1.63) 1.26 (1.05–1.52) 1.22 (1.00–1.47) 132,999 119 1.01 (0.78–1.31) 0.95 (0.73–1.25) 0.97 (0.74–1.28) 
227,550 314 1.46 (1.23–1.74) 1.28 (1.05–1.55) 1.21 (0.99–1.49) 163,568 176 1.19 (0.94–1,51) 1.06 (0.81–1.40) 1.10 (0.83–1.45) 
242,048 410 1.83 (1.56–2.16) 1.45 (1.16–1.82) 1.32 (1.03–1.70) 147,683 149 1.14 (0.89–1.45) 0.90 (0.63–1.27) 0.95 (0.66–1.37) 
Ptrendd   <0.001 <0.001 0.007   0.042 0.838 0.859 
Per SD   1.24 (1.18–1.30) 1.17 (1.08–1.27) 1.14 (1.04–1.26)   1.08 (1.00–1.16) 0.99 (0.88–1.12) 1.01 (0.89–1.16) 
WHRc (quintile) 
227,096 245 1.0 1.0 1.0 149,032 121 1.0 1.0 1.0 
236,193 292 1.11 (0.93–1.31) 1.08 (0.91–1.28) 1.06 (0.89–1.25) 151,823 126 0.96 (0.75–1.23) 0.92 (0.72–1.18) 0.92 (0.72–1.19) 
223,657 267 1.06 (0.89–1.26) 1.01 (0.85–1.21) 0.98 (0.82–1.16) 146,523 140 1.06 (0.83–1.36) 0.98 (0.77–1.26) 0.99 (0.77–1.27) 
205,108 264 1.14 (0.96–1.35) 1.06 (0.89–1.27) 1.00 (0.84–1.20) 147,387 140 1.03 (0.81–1.32) 0.90 (0.70–1.15) 0.91 (0.70–1.17) 
246,016 395 1.46 (1.24–1.71) 1.29 (1.10–1.52) 1.17 (0.99–1.38) 146,225 153 1.12 (0.88–1.42) 0.90 (0.70–1.15) 0.91 (0.70–1.18) 
Ptrendd   <0.001 <0.001 0.044   0.045 0.997 0.836 
Per SD   1.13 (1.08–1.19) 1.09 (1.04–1.14) 1.05 (1.00–1.11)   1.08 (1.00–1.17) 1.00 (0.92–1.08) 1.01 (0.93–1.10) 
HC (quintile) 
217,058 243 1.0 1.0 1.0 146,956 131 1.0 1.0 1.0 
255,289 249 0.84 (0.71–1.01) 0.78 (0.65–0.93) 0.77 (0.65–0.93) 150,767 121 0.90 (0.70–1.15) 0.93 (0.72–1.19) 0.94 (0.73–1.21) 
235,050 307 1.13 (0.95–1.34) 0.95 (0.79–1.14) 0.94 (0.78–1.13) 146,136 143 1.07 (0.85–1.36) 1.10 (0.85–1.43) 1.13 (0.87–1.48) 
212,115 307 1.25 (1.06–1.48) 0.96 (0.78–1.17) 0.95 (0.77–1.16) 163,789 151 1.01 (0.80–1.27) 1.01 (0.76–1.33) 1.06 (0.78–1.43) 
218,559 357 1.43 (1.21–1.68) 0.91 (0.72–1.16) 0.89 (0.70–1.14) 133,341 134 1.13 (0.89–1.44) 1.04 (0.74–1.48) 1.15 (0.77–1.71) 
Ptrendd   <0.001 0.952 0.829   0.201 0.940 0.557 
Per SD   1.18 (1.12–1.24) 1.00 (0.93–1.09) 0.99 (0.91–1.08)   1.05 (0.97–1.13) 1.01 (0.89–1.13) 1.04 (0.90–1.21) 

NOTE: Model 1: The model is derived from Cox regression using age as the underlying time metric; Model 2: The multivariate models were adjusted for education (less than 12 y, 12 y or high school, post-high school or some college, college graduate, or postgraduate), race/ethnicity (Caucasian or non-Caucasian), smoking status (never smoker, former smoker with 20 cig./d or less, former smoker with more than 20 cig./d, current smoker with 20 cig./d or less, or current smoker with more than 20 cig./d), marital status (married or living as married; divorced, widowed, separated, or never married), physical activity (never or rarely, 1 to 3 times/mo, 1 to 2 times/wk, 3 to 4 times/wk, 5 or more times/wk), NSAID use (yes or no), family history of colorectal cancer (yes or no), diabetes status (yes or no), dietary intakes of total energy, fiber, folate, calcium, red meat, fruits and vegetables (all in quartiles), alcohol (yes or no), HRT (women only; never, current, or past). In addition, the multivariate model for waist circumference was adjusted for hip circumference (continuous) and the multivariate model for hip circumference was adjusted for waist circumference (continuous). Multivariate models for waist circumference, hip circumference, and WHR were further adjusted for height (continuous); Model 3: The multivariate model 2 was further adjusted for BMI (continuous).

Abbreviations: HC, hip circumference; PY, person years; WC, waist circumference.

aNumber of patients with colon cancer.

bWeight/height2.

cWaist circumference/hip circumference.

dPtrend (two-sided) across categories is based on the continuous variable and was calculated using Cox regression.

Tests for linear trend were conducted using the continuous variables. For comparability across the four anthropometric measures, we calculated the HRs per SD for each anthropometric parameter. For cancer risk discrimination, we used model 2 Cox models and the resulting predictions to assess measures of discrimination for censored time-to-event data using Harrell's C-index (24). We tested for potential interactions between anthropometric parameters, gender, age, colonic sub-site, and HRT use using likelihood ratio tests.

All statistical tests were two-sided and P values were considered statistically significant at the 0.05 level. Analysis was conducted using SAS version 9.2 (SAS Institute).

Baseline characterizations

During 1,879,060 person-years of follow-up, 2,143 incident colon cancers (1,463 men; 680 women) and 726 incident rectal cancers (536 men; 190 women) were identified. Baseline characteristics of the participants by gender-specific quintiles of waist circumference are summarized in Table 1. Men and women in the highest quintile of waist circumference tended to be less educated and less physically active than those in the lowest quintile. In contrast, participants with high waist circumference were more likely to have diabetes, to have formerly smoked 20 or more cigarettes per day, to use nonsteroidal anti-inflammatory drugs (NSAID), and to have higher intakes of total energy, red meat, and alcohol than those with a low waist circumference. In women, current HRT use was inversely related to waist circumference.

Table 1.

Baseline characteristics of study participants in the NIH-AARP Diet and Health Study by waist circumference

Quintile of waist circumference (range in cm)
MenWomen
1234512345
Characteristic<89.589.5–<93.593.5–<99.099.0–<106.5≥106.5<73.673.6–<80.080.0<–86.086.0–<94.5≥94.5
Participants, n 26,934 19.994 25,583 24,840 26,857 14,933 16,512 14,145 17,464 15,915 
Age at entry, y 62.6 63.2 63.3 63.4 63.1 61.7 62.5 62.9 63.2 63.0 
Height, m 1.75 1.78 1.79 1.79 1.80 1.62 1.63 1.64 1.64 1.64 
Weight, kg 72.4 79.3 83.9 89.0 100.9 56.7 62.0 67.0 73.2 85.2 
BMI, kg/m2a 23.5 25.2 26.3 27.7 31.1 21.7 23.3 25.1 27.3 31.6 
Waist circumference, cm 85.3 91.5 95.9 101.9 112.9 68.7 76.3 82.5 89.7 103.8 
Hip circumference, cm 94.8 99.1 102.0 105.5 112.8 94.0 98.4 102.3 106.8 116.4 
WHRb 0.90 0.93 0.94 0.97 1.00 0.73 0.78 0.81 0.84 0.90 
Caucasian, % 91.0 93.9 95.2 95.8 96.0 93.3 93.3 91.8 91.6 91.9 
Family history of cancer, % 46.7 48.2 48.9 49.3 49.3 52.0 52.5 53.1 53.7 53.4 
Family history of CRC, % 15.2 14.9 15.1 15.8 15.5 19.1 19.8 19.8 19.8 19.9 
Diabetes status, % 5.3 6.7 7.4 8.9 14.6 1.7 2.1 3.4 5.8 12.8 
Education level, % 
 Less than high school 3.9 3.9 3.9 3.9 5.0 3.0 3.4 4.1 5.0 5.9 
 High school graduate 12.5 13.0 13.1 13.3 15.1 20.9 22.2 23.3 25.0 25.0 
 Some college 30.4 30.7 31.7 32.1 34.8 37.2 36.1 36.9 37.6 37.7 
 College graduate 24.2 24.8 24.5 24.5 22.1 18.8 18.6 17.4 15.5 14.8 
 Postgraduate 29.1 27.6 26.8 26.2 23.1 20.0 19.7 18.3 17.0 16.6 
Married, % 83.1 86.8 87.9 88.8 87.1 47.5 48.9 47.9 47.3 42.4 
Smoking status, % 
 Never 36.9 33.6 32.6 29.8 26.0 48.2 47.8 46.6 46.2 45.2 
 Former (<20 cig./d) 31.5 32.7 31.4 30.4 27.3 28.5 28.9 28.5 28.0 27.3 
 Former (≥20 cig./d) 19.6 24.3 27.8 32.1 38.7 8.3 10.5 11.7 13.3 16.6 
 Current (<20 cig./d) 7.3 5.7 4.7 4.2 4.0 11.7 9.6 9.3 8.8 7.1 
 Current (≥20 cig./d) 4.6 3.8 3.5 3.5 4.1 3.3 3.3 3.9 3.7 3.8 
Physical activity, % 
 Low 17.5 19.8 22.3 26.0 34.7 23.3 25.9 29.7 34.8 45.1 
 Intermediate 50.3 53.9 54.0 53.7 49.6 50.3 52.0 51.9 49.9 43.4 
 High 32.2 26.3 23.8 20.3 15.7 26.5 22.1 18.4 15.3 11.5 
Dietary intakes 
 Total energy, kcal/d 1,990 1,982 2,013 2,045 2,147 1,508 1,527 1,560 1,620 1,682 
 Alcohol drinker, % 80.1 81.9 82.3 82.0 80.3 76.8 77.4 75.9 73.2 66.4 
  Alcohol, g/dd 20.7 20.4 21.8 23.1 24.0 8.4 8.9 8.9 8.9 7.8 
 Fiber, g/d 22.1 21.1 20.9 20.6 20.8 18.7 18.1 18.2 18.1 18.2 
 Folate, mcg/d 467.0 451.5 449.7 446.0 453.0 380.6 376.0 377.9 379.7 382.5 
 Calcium, mg/d 817.9 811.0 817.9 822.7 860.5 713.4 717.7 724.5 735.9 752.4 
 Fruits and vegetables, servings/d 6.6 6.3 6.2 6.1 6.1 6.5 6.3 6.2 6.1 6.0 
 Red meat, g/d 64.5 70.3 74.8 79.5 92.7 36.9 40.3 43.5 49.4 55.6 
Regular NSAID use, %c 63.7 67.0 68.7 69.4 71.1 60.0 63.6 65.5 67.4 69.0 
HRT use, % 
 Never — — — — — 36.8 37.3 40.3 44.1 50.4 
 Current — — — — — 55.0 54.0 50.1 46.1 39.2 
 Past — — — — — 8.0 8.5 9.4 9.7 10.3 
Duration of HRT use, y — — — — — 6.9 6.8 6.8 6.6 6.4 
Quintile of waist circumference (range in cm)
MenWomen
1234512345
Characteristic<89.589.5–<93.593.5–<99.099.0–<106.5≥106.5<73.673.6–<80.080.0<–86.086.0–<94.5≥94.5
Participants, n 26,934 19.994 25,583 24,840 26,857 14,933 16,512 14,145 17,464 15,915 
Age at entry, y 62.6 63.2 63.3 63.4 63.1 61.7 62.5 62.9 63.2 63.0 
Height, m 1.75 1.78 1.79 1.79 1.80 1.62 1.63 1.64 1.64 1.64 
Weight, kg 72.4 79.3 83.9 89.0 100.9 56.7 62.0 67.0 73.2 85.2 
BMI, kg/m2a 23.5 25.2 26.3 27.7 31.1 21.7 23.3 25.1 27.3 31.6 
Waist circumference, cm 85.3 91.5 95.9 101.9 112.9 68.7 76.3 82.5 89.7 103.8 
Hip circumference, cm 94.8 99.1 102.0 105.5 112.8 94.0 98.4 102.3 106.8 116.4 
WHRb 0.90 0.93 0.94 0.97 1.00 0.73 0.78 0.81 0.84 0.90 
Caucasian, % 91.0 93.9 95.2 95.8 96.0 93.3 93.3 91.8 91.6 91.9 
Family history of cancer, % 46.7 48.2 48.9 49.3 49.3 52.0 52.5 53.1 53.7 53.4 
Family history of CRC, % 15.2 14.9 15.1 15.8 15.5 19.1 19.8 19.8 19.8 19.9 
Diabetes status, % 5.3 6.7 7.4 8.9 14.6 1.7 2.1 3.4 5.8 12.8 
Education level, % 
 Less than high school 3.9 3.9 3.9 3.9 5.0 3.0 3.4 4.1 5.0 5.9 
 High school graduate 12.5 13.0 13.1 13.3 15.1 20.9 22.2 23.3 25.0 25.0 
 Some college 30.4 30.7 31.7 32.1 34.8 37.2 36.1 36.9 37.6 37.7 
 College graduate 24.2 24.8 24.5 24.5 22.1 18.8 18.6 17.4 15.5 14.8 
 Postgraduate 29.1 27.6 26.8 26.2 23.1 20.0 19.7 18.3 17.0 16.6 
Married, % 83.1 86.8 87.9 88.8 87.1 47.5 48.9 47.9 47.3 42.4 
Smoking status, % 
 Never 36.9 33.6 32.6 29.8 26.0 48.2 47.8 46.6 46.2 45.2 
 Former (<20 cig./d) 31.5 32.7 31.4 30.4 27.3 28.5 28.9 28.5 28.0 27.3 
 Former (≥20 cig./d) 19.6 24.3 27.8 32.1 38.7 8.3 10.5 11.7 13.3 16.6 
 Current (<20 cig./d) 7.3 5.7 4.7 4.2 4.0 11.7 9.6 9.3 8.8 7.1 
 Current (≥20 cig./d) 4.6 3.8 3.5 3.5 4.1 3.3 3.3 3.9 3.7 3.8 
Physical activity, % 
 Low 17.5 19.8 22.3 26.0 34.7 23.3 25.9 29.7 34.8 45.1 
 Intermediate 50.3 53.9 54.0 53.7 49.6 50.3 52.0 51.9 49.9 43.4 
 High 32.2 26.3 23.8 20.3 15.7 26.5 22.1 18.4 15.3 11.5 
Dietary intakes 
 Total energy, kcal/d 1,990 1,982 2,013 2,045 2,147 1,508 1,527 1,560 1,620 1,682 
 Alcohol drinker, % 80.1 81.9 82.3 82.0 80.3 76.8 77.4 75.9 73.2 66.4 
  Alcohol, g/dd 20.7 20.4 21.8 23.1 24.0 8.4 8.9 8.9 8.9 7.8 
 Fiber, g/d 22.1 21.1 20.9 20.6 20.8 18.7 18.1 18.2 18.1 18.2 
 Folate, mcg/d 467.0 451.5 449.7 446.0 453.0 380.6 376.0 377.9 379.7 382.5 
 Calcium, mg/d 817.9 811.0 817.9 822.7 860.5 713.4 717.7 724.5 735.9 752.4 
 Fruits and vegetables, servings/d 6.6 6.3 6.2 6.1 6.1 6.5 6.3 6.2 6.1 6.0 
 Red meat, g/d 64.5 70.3 74.8 79.5 92.7 36.9 40.3 43.5 49.4 55.6 
Regular NSAID use, %c 63.7 67.0 68.7 69.4 71.1 60.0 63.6 65.5 67.4 69.0 
HRT use, % 
 Never — — — — — 36.8 37.3 40.3 44.1 50.4 
 Current — — — — — 55.0 54.0 50.1 46.1 39.2 
 Past — — — — — 8.0 8.5 9.4 9.7 10.3 
Duration of HRT use, y — — — — — 6.9 6.8 6.8 6.6 6.4 

NOTE: All values, except age, waist circumference, and number of participants were directly standardized to the age distribution of the cohort.

Abbreviations: CRC, colorectal cancer; —, not applicable.

aWeight/height2.

bWaist circumference/hip circumference (no units).

cDefined as 2 to 3 times use per month or more.

dAmong drinkers only; numbers do not add to 100% because of rounding.

The correlation coefficients between BMI and waist circumference, hip circumference, and WHR were 0.75, 0.65, and 0.36 in men and 0.78, 0.78, and 0.41 in women (Supplementary Table S2). The mean, SDs, and quintile-cutoff points for the 4 parameters are shown in Supplementary Table S3.

Associations with cancer risk

Age-adjusted and multivariate-adjusted associations between BMI, waist circumference, WHR, and hip circumference and colon cancer risk are presented in Table 2. In men in age-adjusted models (model 1), the HRs per SD increases in BMI, waist circumference, WHR, and hip circumference were statistically significantly positive. After additional adjustment for potential confounding variables (model 2), risk estimates were attenuated but remained statistically significant for BMI, waist circumference, and WHR, but not hip circumference. The HRs per SD increase in model 2 were 1.14 (95% CI, 1.08–1.20) for BMI; 1.17 (95% CI, 1.08–1.27) for waist circumference; and 1.09 (95% CI, 1.04–1.14) for WHR. After adjustment for BMI (model 3), the associations for waist circumference and WHR remained statistically significantly positive.

For women, borderline positive associations were observed for waist circumference and WHR in relation to colon cancer in the age-adjusted model (model 1). After multivariate adjustment, no significant associations between BMI, waist circumference, WHR, or hip circumference and colon cancer risk were evident (model 2). Tests for interaction with gender revealed significant effect modification for waist circumference (P = 0.042) but not for BMI (P = 0.170), WHR (P = 0.308), or hip circumference (P = 0.117).

We analyzed associations for BMI, waist circumference, WHR, and hip circumference with rectal cancer risk using the 3 models and found no statistically significant associations in the multivariate models (Supplementary Table S4).

We repeated analyses in model 2 for colon cancer risk, categorizing the indicator parameters by WHO cutoff-points for waist circumference and WHR (23). In men, we found significant associations for waist circumference (HR for >94 cm vs. ≤ 94 cm, 1.14; 95% CI, 1.00–1.29; HR for waist circumference >102 cm vs. ≤ 102 cm, 1.16; 95% CI, 1.01–1.33); but not for WHR (HR for >0.9 vs. ≤ 0.9, 1.12; 95% CI, 0.97–1.28). Among women, there were no associations for waist circumference (HR for >80 cm vs. ≤ 80 cm, 1.06; 95% CI, 0.88–1.29; HR for >88 cm vs. ≤ 88 cm, 1.06; 95% CI, 0.87–1.29) and WHR (HR for >0.85 cm vs. ≤0.85 cm, 0.99; 95% CI, 0.84–1.16).

We undertook stratified analyses to assess the effect of age (above vs. below the median) on associations in women. We noted a suggestive positive relation between BMI and colon cancer among younger women (those younger than age 67 years) in the age-adjusted analysis (HR per SD, 1.08; 95% CI, 0.99–1.19) but that association became null after multivariate adjustment (Supplementary Table S5). We also conducted stratified analyses by colonic sub-site (proximal vs. distal location) in women. We noted suggestive positive relations between anthropometric variables and distal colon cancer in age-adjusted analyses (HR for BMI per SD, 1.12; 95% CI, 1.00–1.28; HR for waist circumference per SD, 1.12; 95% CI, 0.98–1.27; HR for WHR per SD, 1.14; 95% CI, 1.00–1.30), but those associations were considerably attenuated and were statistically nonsignificant in the multivariate models (Supplementary Table S6).

Direct comparisons between anthropometric measures

To test the relative predictive characteristics of the four anthropometric measures, we defined the null model including all confounding variables but none of the anthropometric variables, and compared whether the addition of BMI, waist circumference, WHR, and hip circumference, separately, improved predictive characteristics model fit (Table 3). BMI, waist circumference, WHR, and hip circumference conducted equally well in colon cancer risk prediction. Single and joint additions of BMI, waist circumference, WHR, and hip circumference to a null model did not materially improve the model fit. Compared with BMI, waist circumference and WHR did not materially influence the prediction models for colon cancer (C-index changes, −0.0040 and 0.0046 for men; 0.0004 and 0.0005 for women).

Table 3.

Changes in risk discrimination for colon cancer risk for men and women after addition of anthropometric measures to the null model

Null modelBMIaWaist circumferenceWHRbHip circumferenceBMIa + waist circumferenceBMIa + WHRbBMIa + hip circumference
Men 
 −2 log L 33,666.8 33,640.9 33,627.8 33,654.8 33,642.5 33,627.4 33,637.0 33,636.6 
 AIC 33,746.8 33,722.9 33,709.8 33,736.8 33,724.5 33,711.4 33,721.0 33,720.6 
 C-index 0.7455 (0.6996, 0.7889) 0.7380 (0.6916, 0.7818) 0.7339 (0.6874, 0.7780) 0.7425 (0.6964, 0.7861) 0.7376 (0.6913, 0.7815) 0.7338 (0.6873, 0.7779) 0.7371 (0.6907, 0.7810) 0.7363 (0.6899, 0.7803) 
 C-index change Reference −0.0076 (−0.0079, −0.0071) −0.0116 (−0.0122, −0.0109) −0.0030 (−0.0032, −0.0029) −0.0079 (−0.0083, −0.0074) −0.0117 (−0.0123, −0.0110) −0.0085 (−0.0089, −0.0080) −0.0092 (−0.0097, −0.0087) 
Women 
 −2 log L 14,974.2 14,973.8 14,974.2 14,997.2 14,974.2 14,973.6 14,973.8 14,973.0 
 AIC 15,058.2 15,059.8 15,060.2 15,060.2 15,060.2 15,061.6 15,061.8 15,061.0 
 C-index 0.6833 (0.6112, 0.7512) 0.6828 (0.6107, 0.7507) 0.6832 (0.6112, 0.7512) 0.6833 (0.6112, 0.7512) 0.6834 (0.6113, 0.7513) 0.6827 (0.6106, 0.7507) 0.6827 (0.6107, 0.7507) 0.6825 (0.6104, 0.7505) 
 C-index change Reference −0.0005 (−0.0006, −0.0005) −0.0001 (−0.0001, −0.0001) 0.0000 (0.0000, 0.0000) 0.0001 (0.0001, 0.0001) −0.0006 (−0.0006, −0.0006) −0.0005 (−0.0006, −0.0005) −0.0008 (−0.0009, −0.0008) 
Null modelBMIaWaist circumferenceWHRbHip circumferenceBMIa + waist circumferenceBMIa + WHRbBMIa + hip circumference
Men 
 −2 log L 33,666.8 33,640.9 33,627.8 33,654.8 33,642.5 33,627.4 33,637.0 33,636.6 
 AIC 33,746.8 33,722.9 33,709.8 33,736.8 33,724.5 33,711.4 33,721.0 33,720.6 
 C-index 0.7455 (0.6996, 0.7889) 0.7380 (0.6916, 0.7818) 0.7339 (0.6874, 0.7780) 0.7425 (0.6964, 0.7861) 0.7376 (0.6913, 0.7815) 0.7338 (0.6873, 0.7779) 0.7371 (0.6907, 0.7810) 0.7363 (0.6899, 0.7803) 
 C-index change Reference −0.0076 (−0.0079, −0.0071) −0.0116 (−0.0122, −0.0109) −0.0030 (−0.0032, −0.0029) −0.0079 (−0.0083, −0.0074) −0.0117 (−0.0123, −0.0110) −0.0085 (−0.0089, −0.0080) −0.0092 (−0.0097, −0.0087) 
Women 
 −2 log L 14,974.2 14,973.8 14,974.2 14,997.2 14,974.2 14,973.6 14,973.8 14,973.0 
 AIC 15,058.2 15,059.8 15,060.2 15,060.2 15,060.2 15,061.6 15,061.8 15,061.0 
 C-index 0.6833 (0.6112, 0.7512) 0.6828 (0.6107, 0.7507) 0.6832 (0.6112, 0.7512) 0.6833 (0.6112, 0.7512) 0.6834 (0.6113, 0.7513) 0.6827 (0.6106, 0.7507) 0.6827 (0.6107, 0.7507) 0.6825 (0.6104, 0.7505) 
 C-index change Reference −0.0005 (−0.0006, −0.0005) −0.0001 (−0.0001, −0.0001) 0.0000 (0.0000, 0.0000) 0.0001 (0.0001, 0.0001) −0.0006 (−0.0006, −0.0006) −0.0005 (−0.0006, −0.0005) −0.0008 (−0.0009, −0.0008) 

NOTE: Null model included age at baseline (continuous), education (less than 12 y, 12 y or high school, post-high school or some college, college graduate, or postgraduate), race/ethnicity (Caucasian or non-Caucasian), smoking status (never smoker, former smoker with 20 cig./d or less, former smoker with more than 20 cig./d, current smoker with 20 cig./d or less, or current smoker with more than 20 cig./d), marital status (married or living as married, divorced, widowed, separated, or never married), physical activity (never or rarely, 1 to 3 times/mo, 1 to 2 times/wk, 3 to 4 times/wk, 5 or more times/wk), NSAID use (yes or no), family history of colorectal cancer (yes or no), diabetes status (yes or no), dietary intakes of total energy, fiber, folate, calcium, red meat, fruits and vegetables (all in quartiles), alcohol (yes or no), HRT (never, current, or past).

Abbreviations: AIC, Akaike Information Criterion; −2 log L, −2 log likelihood.

aWeight/height2.

bWaist circumference/hip circumference.

Associations stratified by HRT

To assess whether the null associations observed between anthropometric measures and risk of colon and rectal cancer in women were due to effect modification by HRT, we conducted stratified analyses in women according to current versus never HRT use at baseline (Table 4). After multivariate adjustment (model 2), current HRT use was inversely associated with colon cancer (compared with never HRT users, HR per SD, 0.63; 95% CI, 0.53–0.75). Among current users of HRT, waist circumference was inversely (HR per SD, 0.78; 95% CI, 0.63–0.97) and WHR was suggestively inversely (HR per SD, 0.88; 95% CI, 0.77–1.01) associated with colon cancer in model 2, whereas hip circumference was positively associated with colon cancer (HR per SD, 1.39; 95% CI, 1.13–1.71), but there was no association for BMI. Among never users of HRT, BMI was inversely associated with colon cancer (HR per SD, 0.89; 95% CI, 0.80–0.99). There were no associations between waist circumference, WHR, or hip circumference with colon cancer risk in never HRT users. Tests for interaction by HRT use revealed significant effect modification for hip circumference (P = 0.018) but not for waist circumference (P = 0.728), WHR (P = 0.171), or BMI (P = 0.165).

Table 4.

HRs and 95% CIs for risk of incident colon cancer across categories of anthropometric measures in women stratified by HRT use in the NIH-AARP Diet and Health Study

Never HRT useCurrent HRT use
MeasureCasesaModel 1Model 2Model 3CasesaModel 1Model 2Model 3
BMIb (quintile) 
 1 70 1.0 1.0  49 1.0 1.0  
 2 73 1.01 (0.73–1.41) 1.00 (0.73–1.40)  49 0.99 (0.67–1.47) 1.00 (0.67–1.49)  
 3 73 0.92 (0.67–1.28) 0.91 (0.65–1.27)  46 0.98 (0.65–1.47) 0.97 (0.65–1.46)  
 4 82 0.98 (0.71–1.35) 0.94 (0.67–1.30)  45 1.05 (0.70–1.57) 0.99 (0.66–1.50)  
 5 78 0.83 (0.60–1.14) 0.74 (0.53–1.04)  50 1.43 (0.97–2.12) 1.27 (0.84–1.92)  
Ptrendd  0.171 0.034   0.043 0.241  
 Per SD  0.93 (0.84–1.03) 0.89 (0.80–0.99)   1.14 (1.00–1.30) 1.09 (0.95–1.25)  
WC (quintile) 
 1 60 1.0 1.0 1.0 47 1.0 1.0 1.0 
 2 66 0.91 (0.64–1.29) 0.96 (0.67–1.37) 0.98 (0.69–1.40) 49 0.96 (0.65–1.44) 0.81 (0.53–1.21) 0.80 (0.53–1.21) 
 3 60 0.86 (0.60–1.24) 0.93 (0.64–1.35) 0.97 (0.66–1.42) 47 1.02 (0.68–1.52) 0.82 (0.53–1.26) 0.82 (0.52–1.27) 
 4 102 1.07 (0.78–1.48) 1.18 (0.81–1.70) 1.26 (0.87–1.85) 53 1.07 (0.72–1.61) 0.67 (0.42–1.07) 0.67 (0.41–1.08) 
 5 88 0.92 (0.66–1.27) 1.04 (0.66–1.64) 1.17 (0.72–1.91) 43 1.50 (1.01–2.22) 0.50 (0.27–0.91) 0.49 (0.26–0.94) 
Ptrendd  0.982 0.553 0.229  0.186 0.022 0.029 
 Per SD  1.00 (0.91–1.10) 1.05 (0.90–1.23) 1.11 (0.94–1.33)  1.09 (0.96–1.25) 0.78 (0.63–0.97) 0.77 (0.61–0.97) 
WHRc (quintile) 
 1 52 1.0 1.0 1.0 67 1.0 1.0 1.0 
 2 68 1.10 (0.77–1.58) 1.09 (0.76–1.56) 1.12 (0.78–1.61) 47 0.70 (0.48–1.02) 0.68 (0.47–0.99) 0.65 (0.45–0.95) 
 3 84 1.26 (0.89–1.79) 1.24 (0.88–1.76) 1.31 (0.92–1.86) 43 0.69 (0.47–1.01) 0.63 (0.43–0.93) 0.59 (0.40–0.87) 
 4 77 1.02 (0.72–1.45) 0.97 (0.68–1.39) 1.04 (0.73–1.49) 44 0.76 (0.52–1.12) 0.66 (0.45–0.98) 0.59 (0.40–0.88) 
 5 95 1.17 (0.83–1.63) 1.05 (0.75–1.49) 1.17 (0.82–1.68) 38 0.72 (0.49–1.08) 0.58 (0.38–0.87) 0.49 (0.32–0.76) 
Ptrendd  0.203 0.578 0.219  0.468 0.069 0.018 
 Per SD  1.07 (0.97–1.18) 1.03 (0.93–1.15) 1.07 (0.96–1.20)  0.95 (0.83–1.09) 0.88 (0.76–1.01) 0.83 (0.72–0.97) 
HC (quintile) 
 1 73 1.0 1.0 1.0 48 1.0 1.0 1.0 
 2 72 1.01 (0.73–1.39) 1.01 (0.72–1.41) 1.04 (0.74–1.46) 40 0.78 (0.52–1.19) 0.85 (0.55–1.30) 0.84 (0.54–1.29) 
 3 79 1.07 (0.78–1.47) 1.06 (0.75–1.50) 1.13 (0.79–1.61) 47 0.97 (0.65–1.45) 1.11 (0.71–1.71) 1.08 (0.69–1.68) 
 4 85 0.98 (0.71–1.33) 0.92 (0.63–1.35) 1.03 (0.69–1.54) 51 1.00 (0.67–1.48) 1.20 (0.74–1.92) 1.14 (0.69–1.88) 
 5 67 0.85 (0.61–1.18) 0.73 (0.45–1.18) 0.91 (0.53–1.56) 53 1.54 (1.05–2.28) 2.05 (1.15–3.64) 1.87 (0.98–3.59) 
Ptrendd  0.263 0.120 0.714  0.004 0.002 0.015 
 Per SD  0.95 (0.86–1.04) 0.88 (0.75–1.03) 0.97 (0.80–1.17)  1.21 (1.07–1.38) 1.39 (1.13–1.71) 1.37 (1.06–1.77) 
Never HRT useCurrent HRT use
MeasureCasesaModel 1Model 2Model 3CasesaModel 1Model 2Model 3
BMIb (quintile) 
 1 70 1.0 1.0  49 1.0 1.0  
 2 73 1.01 (0.73–1.41) 1.00 (0.73–1.40)  49 0.99 (0.67–1.47) 1.00 (0.67–1.49)  
 3 73 0.92 (0.67–1.28) 0.91 (0.65–1.27)  46 0.98 (0.65–1.47) 0.97 (0.65–1.46)  
 4 82 0.98 (0.71–1.35) 0.94 (0.67–1.30)  45 1.05 (0.70–1.57) 0.99 (0.66–1.50)  
 5 78 0.83 (0.60–1.14) 0.74 (0.53–1.04)  50 1.43 (0.97–2.12) 1.27 (0.84–1.92)  
Ptrendd  0.171 0.034   0.043 0.241  
 Per SD  0.93 (0.84–1.03) 0.89 (0.80–0.99)   1.14 (1.00–1.30) 1.09 (0.95–1.25)  
WC (quintile) 
 1 60 1.0 1.0 1.0 47 1.0 1.0 1.0 
 2 66 0.91 (0.64–1.29) 0.96 (0.67–1.37) 0.98 (0.69–1.40) 49 0.96 (0.65–1.44) 0.81 (0.53–1.21) 0.80 (0.53–1.21) 
 3 60 0.86 (0.60–1.24) 0.93 (0.64–1.35) 0.97 (0.66–1.42) 47 1.02 (0.68–1.52) 0.82 (0.53–1.26) 0.82 (0.52–1.27) 
 4 102 1.07 (0.78–1.48) 1.18 (0.81–1.70) 1.26 (0.87–1.85) 53 1.07 (0.72–1.61) 0.67 (0.42–1.07) 0.67 (0.41–1.08) 
 5 88 0.92 (0.66–1.27) 1.04 (0.66–1.64) 1.17 (0.72–1.91) 43 1.50 (1.01–2.22) 0.50 (0.27–0.91) 0.49 (0.26–0.94) 
Ptrendd  0.982 0.553 0.229  0.186 0.022 0.029 
 Per SD  1.00 (0.91–1.10) 1.05 (0.90–1.23) 1.11 (0.94–1.33)  1.09 (0.96–1.25) 0.78 (0.63–0.97) 0.77 (0.61–0.97) 
WHRc (quintile) 
 1 52 1.0 1.0 1.0 67 1.0 1.0 1.0 
 2 68 1.10 (0.77–1.58) 1.09 (0.76–1.56) 1.12 (0.78–1.61) 47 0.70 (0.48–1.02) 0.68 (0.47–0.99) 0.65 (0.45–0.95) 
 3 84 1.26 (0.89–1.79) 1.24 (0.88–1.76) 1.31 (0.92–1.86) 43 0.69 (0.47–1.01) 0.63 (0.43–0.93) 0.59 (0.40–0.87) 
 4 77 1.02 (0.72–1.45) 0.97 (0.68–1.39) 1.04 (0.73–1.49) 44 0.76 (0.52–1.12) 0.66 (0.45–0.98) 0.59 (0.40–0.88) 
 5 95 1.17 (0.83–1.63) 1.05 (0.75–1.49) 1.17 (0.82–1.68) 38 0.72 (0.49–1.08) 0.58 (0.38–0.87) 0.49 (0.32–0.76) 
Ptrendd  0.203 0.578 0.219  0.468 0.069 0.018 
 Per SD  1.07 (0.97–1.18) 1.03 (0.93–1.15) 1.07 (0.96–1.20)  0.95 (0.83–1.09) 0.88 (0.76–1.01) 0.83 (0.72–0.97) 
HC (quintile) 
 1 73 1.0 1.0 1.0 48 1.0 1.0 1.0 
 2 72 1.01 (0.73–1.39) 1.01 (0.72–1.41) 1.04 (0.74–1.46) 40 0.78 (0.52–1.19) 0.85 (0.55–1.30) 0.84 (0.54–1.29) 
 3 79 1.07 (0.78–1.47) 1.06 (0.75–1.50) 1.13 (0.79–1.61) 47 0.97 (0.65–1.45) 1.11 (0.71–1.71) 1.08 (0.69–1.68) 
 4 85 0.98 (0.71–1.33) 0.92 (0.63–1.35) 1.03 (0.69–1.54) 51 1.00 (0.67–1.48) 1.20 (0.74–1.92) 1.14 (0.69–1.88) 
 5 67 0.85 (0.61–1.18) 0.73 (0.45–1.18) 0.91 (0.53–1.56) 53 1.54 (1.05–2.28) 2.05 (1.15–3.64) 1.87 (0.98–3.59) 
Ptrendd  0.263 0.120 0.714  0.004 0.002 0.015 
 Per SD  0.95 (0.86–1.04) 0.88 (0.75–1.03) 0.97 (0.80–1.17)  1.21 (1.07–1.38) 1.39 (1.13–1.71) 1.37 (1.06–1.77) 

NOTE: Model 1: The model is derived from Cox regression using age as the underlying time metric; Model 2: The multivariate models were adjusted for education (less than 12 y, 12 y or high school, post-high school or some college, college graduate, or postgraduate), race/ethnicity (Caucasian or non-Caucasian), smoking status (never smoker, former smoker with 20 cig./d or less, former smoker with more than 20 cig./d, current smoker with 20 cig./d or less, or current smoker with more than 20 cig./d), marital status (married or living as married, divorced, widowed, separated, or never married), physical activity (never or rarely, 1 to 3 times/mo, 1 to 2 times/wk, 3 to 4 times/wk, 5 or more times/wk), NSAID use (yes or no), family history of colorectal cancer (yes or no), diabetes status (yes or no), dietary intakes of total energy, fiber, folate, calcium, red meat, fruits and vegetables (all in quartiles), alcohol (yes or no), HRT (women only; never, current, or past). In addition, the multivariate model for waist circumference was adjusted for hip circumference (continuous) and the multivariate model for hip circumference was adjusted for waist circumference (continuous). Multivariate models for waist circumference, hip circumference and WHR were further adjusted for height (continuous); Model 3: The multivariate model 2 was further adjusted for BMI (continuous).

Abbreviations: HC, hip circumference; WC, waist circumference.

aNumber of patients with colon cancer.

bWeight/height2.

cWaist circumference/hip circumference.

dPtrend (two-sided) across categories is based on the continuous variable and was calculated using Cox regression.

We noted in the stratified analyses of women with current HRT use that colon cancer risk estimates changed substantially between model 1 and model 2 for waist circumference. We explored this further, adding one potential confounder at a time to model 1 and noted that the greatest change in risk estimate was through the additions of hip circumference and BMI (Supplementary Table S7).

To further examine potential effect modification of HRT of the relations of anthropometric variables to colon cancer risk, we analyzed HRs for BMI, waist circumference, WHR, and hip circumference with risk expressed relative to quintile 1 in never HRT users (Fig. 1). After multivariate adjustment (model 2), current HRT use was associated with reduced risk of colon cancer across the lower (but not uppermost) quintiles of BMI and hip circumference and reduced risk associated with increasing waist circumference and WHR (Supplementary Table S8).

Figure 1.

Hazard of colon cancer by anthropometric parameters and HRT status referent: never HRT (no HRT) and quintile 1 of anthropometric parameters. Plots of colon cancer HRs for (A) BMI, (B) waist circumference (WC), (C) WHR, and (D) hip circumference (HC) by HRT status expressed relative to quintile 1 for never HRT users. Never HRT users represented as solid line; current HRT users represented as dotted line. For each parameter, density histograms and corresponding mid-quintile points are shown. HRs are derived as per model 2 and expressed on the log scale. The number of cases for non-HRT users and current HRT users per quintile are shown as headers.

Figure 1.

Hazard of colon cancer by anthropometric parameters and HRT status referent: never HRT (no HRT) and quintile 1 of anthropometric parameters. Plots of colon cancer HRs for (A) BMI, (B) waist circumference (WC), (C) WHR, and (D) hip circumference (HC) by HRT status expressed relative to quintile 1 for never HRT users. Never HRT users represented as solid line; current HRT users represented as dotted line. For each parameter, density histograms and corresponding mid-quintile points are shown. HRs are derived as per model 2 and expressed on the log scale. The number of cases for non-HRT users and current HRT users per quintile are shown as headers.

Close modal

We repeated all analyses for rectal cancer risk and found no associations or interactions (Supplementary Tables S9 and S10).

We excluded the first 2 years of follow-up from date of the second questionnaire, and found no material influences on our findings (Supplementary Table S11).

Main findings

In this large prospective study of men and women, our primary findings showed significant positive associations with colon cancer in men for BMI, waist circumference, and WHR, but not for hip circumference. BMI, waist circumference, and WHR were equally discriminatory for colon cancer risk. There were no associations with BMI, waist circumference, WHR, or hip circumference for colon cancer risk in women, and no associations between anthropometric variables and rectal cancer risk in men or women. In secondary analyses, there was some evidence for effect modification by HRT, but the patterns of direction of these relations were complex. Specifically, among current users of HRT, waist circumference was inversely associated with colon cancer, whereas hip circumference was positively associated with colon cancer.

Context of other studies

The observations in this study on the associations between BMI and colon cancer risk in men are consistent with our updated meta-analyses of 22 prospective studies (25). For women, associations between BMI and colon cancer risk are generally either modestly positive or null. For associations among waist circumference, WHR, and colon cancer risk, our findings for men are consistent with those reported in five other studies (2–6), which reported relative risk estimates for high versus low waist circumference or WHR ranging from 1.39 (2) to 3.3. (5). In contrast, for women, 4 (2, 4, 7, 8) of 6 prospective studies (2, 4, 5, 7, 8, 26) observed positive associations between waist circumference and colon cancer, with relative risks ranging from 1.4 to 1.6. (7, 8).

Additional adjustment for BMI generally attenuates the associations of waist circumference and WHR with colon cancer to a null association, as noted by others (4, 5, 26) or to a more modest positive association, as noted in the present study.

We noted an inverse association between abdominal adiposity and colon cancer risk in postmenopausal women with HRT use. Four previous studies have reported on the association between parameters of adiposity and colon cancer risk in postmenopausal women stratified by HRT use—two prospective studies observing positive associations between adiposity and colon cancer among postmenopausal women without HRT use, but not among HRT users (2, 27); one case–control study observing an inverse association between BMI and colorectal adenoma in HRT users (28); and by contrast, one case–control study showing a positive association between BMI and colorectal cancer risk among HRT users (29). Several reasons may explain these inconsistencies, including different criteria to define HRT use and different types of HRT between study populations.

Furthermore, we found that hip circumference was positively associated with colon cancer risk in HRT users. We are aware of another report on hip circumference and colon cancer risk stratified by HRT use (2); that study observed no association of hip circumference with colon cancer risk in any of the HRT strata. In contrast, one study found a positive association between hip circumference and colorectal cancer in postmenopausal women, but did not present results stratified by HRT use (30). One study (31) found a positive association between hip circumference and premenopausal breast cancer, which among the subgroup of women with hormone receptor–positive tumors only emerged after adjustment for BMI, suggesting an interaction between gluteofemoral fat accumulation, tumor hormone receptor status, and breast cancer risk.

Biologic mechanisms

Findings from our study may provide insight into the biologic mechanisms underpinning links between obesity and colon cancer. First, waist circumference and WHR—as approximations of VAT—are positively correlated with insulin resistance, hyperinsulinemia, and chronic inflammation, conditions hypothesized to play an important role in colon carcinogenesis (10, 32). For example, insulin directly stimulates the growth of colonic cancer cells in vitro (32, 33), whereas hyperinsulinemia indirectly increases the bioavailability of insulin-like growth factor I (IGF-I), a key promoter of tumor development (33).

Classic adipokines, such as leptin and adiponectin, are also preferentially secreted from VAT—adiponectin is inversely correlated with increasing fatness, is antiangiogenic, anti-inflammatory, insulin-sensitizing, and inhibits tumor growth in animals. Circulating levels are higher in women than men, an observation which may contribute to the gender differences in risk associations noted for colon cancer (34).

Second, as HRT use is associated with decreased colon cancer risk [this study and others (35, 36)], bioavailable estrogens may be relevant. Estrogens inhibit colonic epithelial cell growth via stimulation of estrogen receptors in colon cells (37, 38); they reduce bile acid synthesis, putative promoters of colon carcinogenesis (39); and they reduce bioavailable IGF-I levels in postmenopausal women (40).

The observations in our study that current use of HRT was associated with reduced risk of colon cancer with increasing waist circumference and WHR but elevated risk with increasing hip circumference, point to a complexity of biologic processes. In postmenopausal women, the primary production of estrogen occurs in the adipose tissue via aromatization of androstenedione to estrone, such that levels of bioavailable estrogen are greater in obese than lean women (41). The ratio of estradiol to testosterone may also be relevant (42).

VAT may be a particularly strong stimulus to aromatization through release of interleukin 6 (IL)-6 and TNF-α (43, 44). Sex-hormone–binding globulin is also reduced in obese postmenopausal women (45), serving to further drive higher levels of bioavailable estrogen. Thus, HRT use combined with an increased amount of VAT may offer “double” estrogen protection against colon cancer risk and outweigh the adverse effects of the VAT-associated insulin pathway.

Finally, among women, high hip circumference is strongly related to subcutaneous gluteofemoral fat accumulation (11, 46), which in turn, is the main source of leptin production (44). Leptin stimulates colonic cell growth in vitro (47, 48) and in human studies is associated with increased colon cancer risk independent of insulin (49). In postmenopausal women, as leptin mRNA expression in subcutaneous fat correlates positively with estrogen (50)—we speculate that hip circumference may be associated with increased colon cancer risk via an estrogen-related synthesis of leptin in the subcutaneous fat.

Strengths and limitations

Strengths of the present study include its large sample size, prospective design, colorectal sub-site classification, high follow-up rate, and detailed anthropometry data. To our knowledge, this is the first prospective study to directly compare risk estimates for colon cancer related to these anthropometric measures at 2 levels: first, by deriving standardized HRs for each parameter; and second, by comparing predictive characteristics using Harrell's C-index. This is analogous to analyses comparing BMI and abdominal adiposity measures and cardiovascular disease risk reported by the Emerging Risk Factors Collaboration (51). Finally, we addressed the hypothesis that the weaker associations observed between anthropometric measures and colorectal cancer risk in women reflects HRT use—we found that these relations are complex and warrant further research.

There are potential study limitations. First, anthropometric data were self-reported, and findings could have been affected by imprecise assessments. However, self-reported waist circumference, hip circumference, weight, and height values are sufficiently precise for use in large epidemiologic investigations (52). Second, the data informing waist circumference and hip circumference are from respondents to the second questionnaire, who represent a proportion of the overall cohort. However, characteristics of this subcohort were similar to those of the overall cohort, suggesting that selection bias was not likely responsible for our findings. Third, this study represents the largest cohort to date to evaluate associations across individual anthropometric parameters and risks of colon and rectal cancer, but larger sample sizes may be required to adequately explore interactions between highly correlated measures.

Unanswered questions and future research

Important questions remain. First, there is a need to replicate these analyses in other populations to test the generalizability of our findings. Second, previous studies indicate that estrogen therapy and combined estrogen–progesterone therapy may have distinct associations with colon cancer (17, 53, 54)—these associations need to be tested in the context of interactions with a variety of anthropometric measures. Finally, there is a need to test alternative measures of VAT—such as computed tomography (CT) scan–derived estimates—and associations with colorectal cancer risk. We hypothesize that these indicators have stronger predictive characteristics than currently used anthropometry, and more likely to identify clinically relevant high-risk individuals.

A.G. Renehan is a consultant/advisory board member of Novo Nordisk. A.R. Hollenbeck is a consultant/advisory board member of Love/Avon Army of Women Scientific Advisory Committee and Society of Psychologists in Management Board of Directors. No potential conflicts of interest were disclosed by the other authors.

The funding agencies had no role in the design, conduct, decision to report, or writing of the study.

Conception and design: A.G. Renehan, A.R. Hollenbeck, M.F. Leitzmann

Development of methodology: A.G. Renehan, M.F. Leitzmann

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A.R. Hollenbeck, A.J. Cross

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): M. Keimling, A.G. Renehan, G. Behrens, A.J. Cross, M.F. Leitzmann

Writing, review, and/or revision of the manuscript: M. Keimling, A.G. Renehan, G. Behrens, B. Fischer, A.R. Hollenbeck, A.J. Cross, M.F. Leitzmann

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): B. Fischer, A.R. Hollenbeck

Study supervision: A.G. Renehan, M.F. Leitzmann

The authors thank the participants in the NIH-AARP Diet and Health Study for their outstanding cooperation.

This research was supported by the Intramural Research Program of the NIH, National Cancer Institute.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Renehan
AG
,
Tyson
M
,
Egger
M
,
Heller
RF
,
Zwahlen
M
. 
Body-mass index and incidence of cancer: a systematic review and meta-analysis of prospective observational studies
.
Lancet
2008
;
371
:
569
78
.
2.
Pischon
T
,
Lahmann
PH
,
Boeing
H
,
Friedenreich
C
,
Norat
T
,
Tjonneland
A
, et al
Body size and risk of colon and rectal cancer in the European Prospective Investigation Into Cancer and Nutrition (EPIC)
.
J Natl Cancer Inst
2006
;
98
:
920
31
.
3.
MacInnis
RJ
,
English
DR
,
Hopper
JL
,
Haydon
AM
,
Gertig
DM
,
Giles
GG
. 
Body size and composition and colon cancer risk in men
.
Cancer Epidemiol Biomarkers Prev
2004
;
13
:
553
9
.
4.
Wang
Y
,
Jacobs
EJ
,
Patel
AV
,
Rodriguez
C
,
McCullough
ML
,
Thun
MJ
, et al
A prospective study of waist circumference and body mass index in relation to colorectal cancer incidence
.
Cancer Causes Control
2008
;
19
:
783
92
.
5.
Moore
LL
,
Bradlee
ML
,
Singer
MR
,
Splansky
GL
,
Proctor
MH
,
Ellison
RC
, et al
BMI and waist circumference as predictors of lifetime colon cancer risk in Framingham Study adults
.
Int J Obes Relat Metab Disord
2004
;
28
:
559
67
.
6.
Giovannucci
E
,
Ascherio
A
,
Rimm
EB
,
Colditz
GA
,
Stampfer
MJ
,
Willett
WC
. 
Physical activity, obesity, and risk for colon cancer and adenoma in men
.
Ann Intern Med
1995
;
122
:
327
34
.
7.
MacInnis
RJ
,
English
DR
,
Hopper
JL
,
Gertig
DM
,
Haydon
AM
,
Giles
GG
. 
Body size and composition and colon cancer risk in women
.
Int J Cancer
2006
;
118
:
1496
500
.
8.
Folsom
AR
,
Kushi
LH
,
Anderson
KE
,
Mink
PJ
,
Olson
JE
,
Hong
CP
, et al
Associations of general and abdominal obesity with multiple health outcomes in older women: the Iowa Women's Health Study
.
Arch Intern Med
2000
;
160
:
2117
28
.
9.
Ma
Y
,
Yang
Y
,
Wang
F
,
Zhang
P
,
Shi
C
,
Zou
Y
, et al
Obesity and risk of colorectal cancer: a systematic review of prospective studies
.
PLoS ONE
2013
;
8
:
e53916
.
10.
Tsugane
S
,
Inoue
M
. 
Insulin resistance and cancer: epidemiological evidence
.
Cancer Sci
2010
;
101
:
1073
9
.
11.
Krotkiewski
M
,
Bjorntorp
P
,
Sjostrom
L
,
Smith
U
. 
Impact of obesity on metabolism in men and women. Importance of regional adipose tissue distribution
.
J Clin Invest
1983
;
72
:
1150
62
.
12.
Ritchie
SA
,
Connell
JM
. 
The link between abdominal obesity, metabolic syndrome and cardiovascular disease
.
Nutr Metab Cardiovasc Dis
2007
;
17
:
319
26
.
13.
Fox
CS
,
Massaro
JM
,
Hoffmann
U
,
Pou
KM
,
Maurovich-Horvat
P
,
Liu
CY
, et al
Abdominal visceral and subcutaneous adipose tissue compartments: association with metabolic risk factors in the Framingham Heart Study
.
Circulation
2007
;
116
:
39
48
.
14.
Huxley
R
,
Mendis
S
,
Zheleznyakov
E
,
Reddy
S
,
Chan
J
. 
Body mass index, waist circumference and waist:hip ratio as predictors of cardiovascular risk—a review of the literature
.
Eur J Clin Nutr
2010
;
64
:
16
22
.
15.
Canoy
D
,
Boekholdt
SM
,
Wareham
N
,
Luben
R
,
Welch
A
,
Bingham
S
, et al
Body fat distribution and risk of coronary heart disease in men and women in the European Prospective Investigation Into Cancer and Nutrition in Norfolk cohort: a population-based prospective study
.
Circulation
2007
;
116
:
2933
43
.
16.
Yusuf
S
,
Hawken
S
,
Ounpuu
S
,
Bautista
L
,
Franzosi
MG
,
Commerford
P
, et al
Obesity and the risk of myocardial infarction in 27,000 participants from 52 countries: a case–control study
.
Lancet
2005
;
366
:
1640
9
.
17.
Chlebowski
RT
,
Wactawski-Wende
J
,
Ritenbaugh
C
,
Hubbell
FA
,
Ascensao
J
,
Rodabough
RJ
, et al
Estrogen plus progestin and colorectal cancer in postmenopausal women
.
N Engl J Med
2004
;
350
:
991
1004
.
18.
Adams
KF
,
Leitzmann
MF
,
Albanes
D
,
Kipnis
V
,
Mouw
T
,
Hollenbeck
A
, et al
Body mass and colorectal cancer risk in the NIH-AARP cohort
.
Am J Epidemiol
2007
;
166
:
36
45
.
19.
Renehan
AG
,
Flood
A
,
Adams
KF
,
Olden
M
,
Hollenbeck
AR
,
Cross
AJ
, et al
Body mass index at different adult ages, weight change, and colorectal cancer risk in the National Institutes of Health-AARP Cohort
.
Am J Epidemiol
2012
;
176
:
1130
40
.
20.
Michaud
DS
,
Midthune
D
,
Hermansen
S
. 
Comparison of cancer registry case ascertainment with SEER estimates and self-reporting in a subset of the NIH-AARP Diet and Health Study
.
J Registry Manag
2005
;
32
:
70
5
.
21.
Spencer
EA
,
Roddam
AW
,
Key
TJ
. 
Accuracy of self-reported waist and hip measurements in 4492 EPIC-Oxford participants
.
Public Health Nutr
2004
;
7
:
723
7
.
22.
McAdams
MA
,
Van Dam
RM
,
Hu
FB
. 
Comparison of self-reported and measured BMI as correlates of disease markers in US adults
.
Obesity
2007
;
15
:
188
96
.
23.
WHO
. 
Waist circumference and waist–hip ratio—report of a WHO expert consultation
.
Geneva, Switzerland
:
World Health Organization (WHO)
; 
2008
. p.
1
47
.
24.
Collaboration FS
. 
Measures to assess the prognostic ability of the stratified Cox proportional hazards model
.
Stat Med
2009
;
28
:
389
411
.
25.
Harriss
DJ
,
Atkinson
G
,
George
K
,
Cable
NT
,
Reilly
T
,
Haboubi
N
, et al
Lifestyle factors and colorectal cancer risk (1): systematic review and meta-analysis of associations with body mass index
.
Colorectal Dis
2009
;
11
:
547
63
.
26.
Martinez
ME
,
Giovannucci
E
,
Spiegelman
D
,
Hunter
DJ
,
Willett
WC
,
Colditz
GA
. 
Leisure-time physical activity, body size, and colon cancer in women. Nurses' Health Study Research Group
.
J Natl Cancer Inst
1997
;
89
:
948
55
.
27.
Lin
J
,
Zhang
SM
,
Cook
NR
,
Rexrode
KM
,
Lee
IM
,
Buring
JE
. 
Body mass index and risk of colorectal cancer in women (United States)
.
Cancer Causes Control
2004
;
15
:
581
9
.
28.
Wolf
LA
,
Terry
PD
,
Potter
JD
,
Bostick
RM
. 
Do factors related to endogenous and exogenous estrogens modify the relationship between obesity and risk of colorectal adenomas in women?
Cancer Epidemiol Biomarkers Prev
2007
;
16
:
676
83
.
29.
Slattery
ML
,
Ballard-Barbash
R
,
Edwards
S
,
Caan
BJ
,
Potter
JD
. 
Body mass index and colon cancer: an evaluation of the modifying effects of estrogen (United States)
.
Cancer Causes Control
2003
;
14
:
75
84
.
30.
Oxentenko
AS
,
Bardia
A
,
Vierkant
RA
,
Wang
AH
,
Anderson
KE
,
Campbell
PT
, et al
Body size and incident colorectal cancer: a prospective study of older women
.
Cancer Prev Res
2010
;
3
:
1608
20
.
31.
Fagherazzi
G
,
Chabbert-Buffet
N
,
Fabre
A
,
Guillas
G
,
Boutron-Ruault
MC
,
Mesrine
S
, et al
Hip circumference is associated with the risk of premenopausal ER/PR breast cancer
.
Int J Obes
2012
;
36
:
431
9
.
32.
Giovannucci
E
. 
Insulin, insulin-like growth factors and colon cancer: a review of the evidence
.
J Nutr
2001
;
131
:
3109S
20S
.
33.
Sandhu
MS
,
Dunger
DB
,
Giovannucci
EL
. 
Insulin, insulin-like growth factor-I (IGF-I), IGF binding proteins, their biologic interactions, and colorectal cancer
.
J Natl Cancer Inst
2002
;
94
:
972
80
.
34.
Rose
DP
,
Komninou
D
,
Stephenson
GD
. 
Obesity, adipocytokines, and insulin resistance in breast cancer
.
Obes Rev
2004
;
5
:
153
65
.
35.
Crandall
CJ
. 
Estrogen replacement therapy and colon cancer: a clinical review
.
J Womens Health Gend Based Med
1999
;
8
:
1155
66
.
36.
Staren
ED
,
Omer
S
. 
Hormone replacement therapy in postmenopausal women
.
Am J Surg
2004
;
188
:
136
49
.
37.
Lointier
P
,
Wildrick
DM
,
Boman
BM
. 
The effects of steroid hormones on a human colon cancer cell line in vitro
.
Anticancer Res
1992
;
12
:
1327
30
.
38.
Di Leo
A
,
Messa
C
,
Russo
F
,
Misciagna
G
,
Guerra
V
,
Taveri
R
, et al
Prognostic value of cytosolic estrogen receptors in human colorectal carcinoma and surrounding mucosa. Preliminary results
.
Dig Dis Sci
1994
;
39
:
2038
42
.
39.
Everson
GT
,
McKinley
C
,
Kern
F
 Jr
. 
Mechanisms of gallstone formation in women. Effects of exogenous estrogen (Premarin) and dietary cholesterol on hepatic lipid metabolism
.
J Clin Invest
1991
;
87
:
237
46
.
40.
Renehan
AG
,
Frystyk
J
,
Howell
A
,
O'Dwyer
ST
,
Shalet
SM
,
Flyvbjerg
A
. 
The effects of sex steroid replacement therapy on an expanded panel of IGF-related peptides
.
Growth Horm IGF Res
2007
;
17
:
210
9
.
41.
Liedtke
S
,
Schmidt
ME
,
Vrieling
A
,
Lukanova
A
,
Becker
S
,
Kaaks
R
, et al
Postmenopausal sex hormones in relation to body fat distribution
.
Obesity
2012
;
20
:
1088
95
.
42.
Lin
JH
,
Zhang
SM
,
Rexrode
KM
,
Manson
JE
,
Chan
AT
,
Wu
K
, et al
Association between sex hormones and colorectal cancer risk in men and women
.
Clin Gastroenterol Hepatol
2013
;
11
:
419
24
.
43.
Hamdy
O
,
Porramatikul
S
,
Al-Ozairi
E
. 
Metabolic obesity: the paradox between visceral and subcutaneous fat
.
Curr Diab Rev
2006
;
2
:
367
73
.
44.
Ibrahim
MM
. 
Subcutaneous and visceral adipose tissue: structural and functional differences
.
Obes Rev
2010
;
11
:
11
8
.
45.
Tchernof
A
,
Despres
JP
. 
Sex steroid hormones, sex hormone-binding globulin, and obesity in men and women
.
Horm Metab Res
2000
;
32
:
526
36
.
46.
Montague
CT
,
Prins
JB
,
Sanders
L
,
Digby
JE
,
O'Rahilly
S
. 
Depot- and sex-specific differences in human leptin mRNA expression: implications for the control of regional fat distribution
.
Diabetes
1997
;
46
:
342
7
.
47.
Hardwick
JC
,
Van Den Brink
GR
,
Offerhaus
GJ
,
Van Deventer
SJ
,
Peppelenbosch
MP
. 
Leptin is a growth factor for colonic epithelial cells
.
Gastroenterology
2001
;
121
:
79
90
.
48.
Stattin
P
,
Lukanova
A
,
Biessy
C
,
Soderberg
S
,
Palmqvist
R
,
Kaaks
R
, et al
Obesity and colon cancer: does leptin provide a link?
Int J Cancer
2004
;
109
:
149
52
.
49.
Ho
GY
,
Wang
T
,
Gunter
MJ
,
Strickler
HD
,
Cushman
M
,
Kaplan
RC
, et al
Adipokines linking obesity with colorectal cancer risk in postmenopausal women
.
Cancer Res
2012
;
72
:
3029
37
.
50.
Fajardo
ME
,
Malacara
JM
,
Martinez-Rodriguez
HG
,
Barrera-Saldana
HA
. 
Hormone and metabolic factors associated with leptin mRNA expression in pre- and postmenopausal women
.
Steroids
2004
;
69
:
425
30
.
51.
Wormser
D
,
Kaptoge
S
,
Di Angelantonio
E
,
Wood
AM
,
Pennells
L
,
Thompson
A
, et al
Separate and combined associations of body-mass index and abdominal adiposity with cardiovascular disease: collaborative analysis of 58 prospective studies
.
Lancet
2011
;
377
:
1085
95
.
52.
Rimm
EB
,
Stampfer
MJ
,
Colditz
GA
,
Chute
CG
,
Litin
LB
,
Willett
WC
. 
Validity of self-reported waist and hip circumferences in men and women
.
Epidemiology
1990
;
1
:
466
73
.
53.
Lin
KJ
,
Cheung
WY
,
Lai
JY
,
Giovannucci
EL
. 
The effect of estrogen vs. combined estrogen-progestogen therapy on the risk of colorectal cancer
.
Int J Cancer
2012
;
130
:
419
30
.
54.
Dinger
JC
,
Heinemann
LA
,
Mohner
S
,
Thai
do M
,
Assmann
A
. 
Colon cancer risk and different HRT formulations: a case–control study
.
BMC Cancer
2007
;
7
:
76
.

Supplementary data