The relationship between body mass index (BMI) and prostate cancer risk may be complex because obesity is associated with various hormonal factors and because the influence of BMI may differ according to whether the cancers are hereditary or sporadic. We used data from the Health Professionals Follow-Up Study, in which 2896 incident cases of prostate cancer were reported from February 1, 1986, through January 31, 2000, to determine prospectively whether BMI was associated with the risk of hereditary (men <60 years of age or with a positive family history of prostate cancer) and sporadic (men > or =60 years of age and without such a family history) prostate cancer. The risk of prostate cancer in men with a higher BMI (> or =30 kg/m2) was lower than that in men with a lower BMI (23-24.9 kg/m2) but only if they were younger (<60 years old) (relative risk = 0.52, 95% confidence interval = 0.33 to 0.83; P(trend)<.001) or had a family history of prostate cancer (relative risk = 0.74, 95% confidence interval = 0.45 to 1.19; P(trend) =.01). However, for groups with more sporadic cancers, BMI had a weak, non-statistically significant positive association with prostate cancer. We observed statistically significant interactions between BMI and age (P(interaction)<.001, two-sided Wald test) and between BMI and family history of prostate cancer (P(interaction) =.006, two-sided Wald test). Patterns for BMI and waist circumference were similar. Because obesity is associated with lower circulating concentrations of testosterone, our results suggest the hypothesis that androgens may play a more direct role for early-onset or hereditary prostate cancers than for sporadic prostate cancers.