Abstract 2176: Obesity and risk of colorectal cancer: A Mendelian randomization study

Abstract

Abstract Background: Obesity has been consistently linked to higher risk of colorectal cancer (CRC) in men, whereas the association is generally lower in women. Although these findings come from well conducted epidemiologic studies, it is difficult to exclude all other explanations for associations in observational studies, such as measurement error and other forms of bias, reverse causality and confounding. Mendelian randomization studies, using instrumental variables (IVs), can overcome some of the inherent limitations of observational studies and provide an unbiased and unconfounded estimate of the causal association between an exposure and outcome. Objective: We used genetic variants that are associated with body mass index (BMI) or waist-hip ratio (WHR) to examine the causal association between obesity and CRC. Methods: We used epidemiologic and genetic data from 10 226 CRC cases and 10 286 controls of European ancestry from the Genetics and Epidemiology of Colorectal Cancer Consortium (GECCO) or the Colon Cancer Family Registry (CCFR). The primary Mendelian randomization analysis was performed using a genetic risk score, derived from 77 established “BMI-increasing” genetic variants, as an instrumental variable for higher BMI. We additionally derived a risk score for WHR using 47 variants for WHR. We compared the IV odds ratio (OR) with the OR obtained using a traditional multivariable logistic regression model adjusted for age, sex, smoking status, family history, aspirin use, and HRT use (women). Results: For men and women combined, in traditional multivariable analysis, each 5 kg/m2 increase in BMI was associated with a 24% (95% confidence interval [CI]: 1.17-1.32) increase in risk of CRC. In the IV analysis, the OR was higher (IV-OR per 5 kg/m2=1.54, 95% CI: 1.15-2.06); however, the two estimates were not statistically significantly different (Pdifference=0.07). For men, there was a statistically significant association between BMI and CRC in traditional multivariable analysis (OR per 5kg/m2=1.33, 95% CI: 1.21-1.46), but no association in the IV analysis (IV-OR per 5 kg/m2=1.13, 95% CI: 0.70-1.83). The estimates however were not significantly different (Pdifference=0.56). In contrast, for women, the IV estimate (IV-OR per 5 kg/m2=1.94, 95% CI: 1.33-2.82) was significantly higher than the estimate from the traditional multivariable analysis (OR per 5 kg/m2=1.18, 95% CI: 1.11-1.25; Pdifference=0.01). In secondary analysis, we found a positive association between WHR and CRC in men (IV-OR per 0.1 unit increase=1.84, 95% CI: 1.02-3.31) but no statistically significant association in women (IV-OR per 0.1 unit increase=1.22, 95% CI: 0.75-1.98). Conclusions: Obesity is independently associated with risk of CRC and may confer greater risk of CRC in women than previously reported. The mechanisms remain largely unknown; however these results suggest that visceral abdominal fat may be particularly important for promoting CRC in men. Citation Format: Aaron P. Thrift, Sonja I. Berndt, Andrew T. Chan, Jenny Chang-Claude, Martha L. Slattery, Michelle Cotterchio, Graham Casey, John D. Potter, Polly A. Newcomb, Emily White, Hermann Brenner, Ulrike Peters, Peter T. Campbell. Obesity and risk of colorectal cancer: A Mendelian randomization study. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 2176. doi:10.1158/1538-7445.AM2014-2176