Blood samples were collected from 52 incident cases of histologically confirmed prostate cancer, an equal number of cases of benign prostatic hyperplasia (BPH) and an equal number of apparently healthy control subjects. The three groups were matched for age and town of residence in the greater Athens area. Steroid hormones, sex hormone-binding globulin, and insulin-like growth factor 1 (IGF-1) were measured in duplicate by radioimmunoassay in a specialized US centre. Statistical analyses were performed using multiple logistical regression. The results for IGF-1 in relation to prostate cancer and BPH were adjusted for demographic and anthropometric factors, as well as for the other measured hormones. There was no relation between IGF-1 and BPH, but increased values of this hormone were associated with increased risk of prostate cancer; an increment of 60 ng ml(-1) corresponded to an odds ratio of 1.91 with a 95% confidence interval of 1.00-3.73. There was also some evidence for an interaction between high levels of testosterone and IGF-1 in relation to prostate cancer. This finding suggests that, in addition to testosterone, IGF-1 may increase the risk of prostate cancer in humans.