Reported associations between body measurements and the risk of prostate cancer are weak and inconsistent, possibly because some measures used do not differentiate between adipose and nonadipose tissue, body components that would theoretically have different associations with prostate cancer. Some studies have addressed this problem by estimating lean body mass from subjects' age, height, and weight. In a prospective cohort study of men 27-75 years of age at recruitment in 1990-1994, body measurements were taken by trained interviewers. Nonadipose and adipose mass were calculated from bioelectric impedance analysis. Incident prostate cancers were ascertained by use of the population cancer registry. Altogether 16,336 men contributed 113,535 person-years and 477 cancers, of which 79 were "aggressive," to the analysis. We found no overall association between prostate cancer and any anthropometric measurement. Analysis stratified by cancer aggressiveness revealed modest associations between measures of adiposity and the risk of aggressive disease. On the basis of the WHO cut points and compared with men in the normal range of body mass index, the risk ratio for obese men was 2.2 (95% confidence interval, 1.2-4.1). For each 10-kg increase in fat mass, the risk ratio was 1.4 (95% confidence interval, 1.0-1.8). Energy imbalance may play a role in the development of aggressive prostate cancer.