Abstract

The aim of this study was to estimate the effect of diet on prostate and breast cancer (PC and BC) risks in smokers and nonsmokers and to explore the effect modification between smoking and dietary patterns. PC or BC incidence rates were assessed spatially according to tobacco exposure, age-adjusted standardization using lung cancer mortality as a proxy. Two case-control studies were carried out in Argentina (2008-2012). Participants were interviewed about their diet, smoking habits, and other lifestyle factors. Multilevel models were fitted including family history of cancer as the random intercept for the second level, and diet and lifestyle variables as covariates. Tobacco exposure was aggregated spatially. Family history of cancer significantly accounts for PC and BC. In smokers, high intake of fat meat increased PC and BC risks [ odds ratio (OR) 1.56, 95% confidence interval (CI) 0.81-3.05 and OR 6.01, 95% CI 1.99-8.19, respectively]. PC and BC risks were also greater in smokers with high intakes of fatty foods (OR 1.95, 95% CI 1.09-3.50 and OR 24.2, 95% CI 0.82-7.21, respectively). Moderate intake of nonstarchy vegetables and risk of PC were inversely associated in nonsmokers (OR 0.55, 95% CI 0.20-1.48). In smoker women, BC risk was associated with sweet drink consumption (OR 2.96, 95% CI 1.10-7.92) and ethanol intake (OR 5.15, 95% CI 1.88-14.16). Spatial distributions of cancer incidence rates match those of tobacco exposure. Differential effects of diet on PC and BC risks were found in smokers and nonsmokers. (C) 2014 Wolters Kluwer Health vertical bar Lippincott Williams Wilkins.