Chocolate Candy and Incident Invasive Cancer Risk in the Women’s Health Initiative: An Observational Prospective Analysis

Laboratory and animal studies suggest an inverse association between chocolate consumption and the risk of cancer. Epidemiological studies have yielded inconsistent evidence. To assess the association of chocolate candy consumption with incident, invasive total, breast, colorectal, and lung cancers in a large cohort of postmenopausal American women. Prospective cohort study with a mean 14.8-year follow-up. Chocolate candy intake was assessed by food frequency questionnaire. Invasive cancer events were assessed by physician adjudication. The Women’s Health Initiative Study enrolled 161,808 postmenopausal women at 40 clinical centers nationwide between 1993 and 1998. Of these women, 114,281 with plausible food frequency or biometric data and no missing data on chocolate candy exposure were selected for analysis. Cancer risk in quartiles of chocolate candy consumption with the first quartile as referent. Multivariable Cox regression was used to calculate hazard ratios and 95% confidence intervals. There were 16,164 documented incident invasive cancers, representing an incidence rate of 17.0 per 100 participants and 12.3 per 1000 person years during follow-up among participants without any preexisting cancers or missing outcome data. There were no statistically significant associations for total invasive cancer (P-linear = .47, P-curvature = .14), or invasive breast cancer (P-linear = .77, P-curvature = .26). For colorectal cancer P-linear was .02, P-curvature was .03, and compared with women eating a 1 oz (28.4 g) chocolate candy serving