Michael F Leitzmann, Sabine Rohrmann

At present, only three risk factors for prostate cancer have been firmly established; these are all nonmodifiable: age, race, and a positive family history of prostate cancer. However, numerous modifiable factors have also been implicated in the development of prostate cancer. In the current review, we summarize the epidemiologic data for age, location, and selected behavioral factors in relation to the onset of prostate cancer. Although the available data are not entirely consistent, possible preventative behavioral factors include increased physical activity, intakes of tomatoes, cruciferous vegetables, and soy. Factors that may enhance prostate cancer risk include frequent consumption of dairy products and, possibly, meat. By comparison, alcohol probably exerts no important influence on prostate cancer development. Similarly, dietary supplements are unlikely to protect against the onset of prostate cancer in healthy men. Several factors, such as smoking and obesity, show a weak association with prostate cancer incidence but a positive relation with prostate cancer mortality. Other factors, such as fish intake, also appear to be unassociated with incident prostate cancer but show an inverse relation with fatal prostate cancer. Such heterogeneity in the relationship between behavioral factors and nonadvanced, advanced, or fatal prostate cancers helps shed light on the carcinogenetic process because it discerns the impact of exposure on early and late stages of prostate cancer development. Inconsistent associations between behavioral factors and prostate cancer risk seen in previous studies may in part be due to uncontrolled detection bias because of current widespread use of prostate-specific antigen testing for prostate cancer, and the possibility that certain behavioral factors are systematically related to the likelihood of undergoing screening examinations. In addition, several genes may modify the study results, but data concerning specific gene-environment interactions are currently sparse. Despite large improvements in our understanding of prostate cancer risk factors in the past two decades, present knowledge does not allow definitive recommendations for specific preventative behavioral interventions.