Diet and risk of colon cancer in a large prospective study of older women: an analysis stratified on family history (Iowa, United States)

Objective: The purpose was to investigate whether dietary associations with risk of colon cancer in women differ by family history of the disease. Methods: Data were analyzed from a prospective cohort study of 35,216 Iowa (United States) women aged 55 to 69 years at baseline. Through 31 December 1995, 241 colon cancers were identified through record linkage with the State Health Registry. The cohort was stratified on family history of colon cancer in first-degree relatives; nutrient intakes were divided into tertiles. Results: Analyses using Cox regression revealed that the association of most dietary components with colon cancer incidence were similar for individuals with and without a family history. However, total calcium intake was associated inversely with colon cancer among women with a negative family history (relative risk [RR] = 0.50 for upper cf lower tertile, P < 0.001), but was unrelated to incidence for women with a positive family history (RR = 1.1 for upper cf lower tertile, P = 0.69). Similarly, total vitamin E intake was associated with lower risk among women with a negative family history (RR = 0.67 for upper cf lower tertile, P = 0.04), but not among women with a positive family history (RR = 0.87 for upper cf lower tertile, P = 0.67). High intakes of fiber, fruits, and vegetables were each weakly inversely associated with risk among family-history negative women, but not among family-history positive women. Conclusions: These data, if corroborated, suggest that dietary factors typically associated with lower risk may be less effective risk-reduction interventions against colon cancer for individuals with a family history of colon cancer.