Studies show that treatment of men with 5α-reductase inhibitors such as finasteride is effective for the primary prevention of prostate cancer. Although it is known that finasteride treatment suppresses serum levels of dihydrotestosterone (DHT) and its distal metabolite, 5α-androstane-3α,17β-diol glucuronide (3α-diol G), and increases serum testosterone (T) levels, little is known about its effect on other precursors and metabolites of DHT, as well as on the relationship of these androgens to prostate specific antigen (PSA), a marker of prostatic intraepithelial neoplasia. The present study provides new data on the effect of finasteride on precursors and metabolites of DHT. Fifty-three men, ages 57-79 years, with elevated PSA levels (>4ng/ml), were randomized to treatment with finasteride (5mg/day) or observation (controls) for 12 months. Blood samples were obtained at baseline, 1, 3, 6 and 12 months for measurement of PSA, androstenedione (A), T, DHT, 3α-diol G, androsterone glucuronide (ADT G) and DHT sulfate (DHT S) in serum by validated, highly specific radioimmunoassays. Statistical analysis was carried out using mixed model ANOVA and t-tests. In the control group, PSA and androgen levels were unchanged throughout the 12 months of treatment. In the finasteride group, PSA, DHT, DHT S, 3α-diol G and ADT G decreased from baseline to 1 month by 23.2%, 78.7%, 71.0%, 75.7% and 43.0%, respectively. The change in PSA decreased further to 46.1% and 55.1% at 3 and 12 months of treatment, respectively, whereas the decrease in androgens observed at 1 month did not change by more than 6.9% for DHT, DHT S and 3α-diol G in the subsequent months of sampling. However, the decline in ADT G was only 22.2% at month 3, and remained essentially at this level after that time. In contrast, T and A increased significantly from baseline, and the increase in A of approximately 34.5% was about 1.9 times the increase in T (approximately 18.3%). The present data suggest that either 3α-diol G or DHT S may serve as a potential diagnostic marker of intraprostatic 5α-reductase activity during treatment of patients with 5α-reductase inhibitors.