Diet appears to be a major determinant in the incidence of prostate cancer. In a case-control study conducted in Athens, Greece, we found that dairy products, butter and seed oils were positively associated with risk of prostate cancer, whereas cooked and raw tomatoes were inversely associated. We utilized the data from this study to calculate the population attributable fractions under alternative assumptions of feasible dietary changes. For each subject, a dietary score was calculated and categorized into approximately quintiles, representing increasing levels of prostate cancer risk as a function of the intake of the five discriminatory food groups or items. Population attributable fractions in terms of this dietary score were calculated taking into account multivariate adjustment. We observed that, if all individuals were shifted to the baseline category, the incidence of prostate cancer in this study population would be reduced by 41% (95% confidence interval 23-59%). However, if all individuals were shifted to the adjacent lower risk quintile, the expected incidence reduction would be a more modest 19%. The incidence of prostate cancer in Greece could be reduced by about two-fifths if the population increased the consumption of tomatoes and reduced the intake of dairy products, and substituted olive oil for other added lipids.
(C) 2000 Lippincott Williams & Wilkins, Inc.