To determine the hormonal effects of MK-906, an orally active 5 alpha-reductase inhibitor, on serum androgens and androgen conjugates, 12 healthy men were given 10, 20, 50, and 100 mg MK-906 2 weeks apart in randomized order in a 4-period crossover design. Serum testosterone (T), dihydrotestosterone (DHT), androstanediol glucuronide, and androsterone glucuronide were measured before and 24 hours after each dose. The effect of MK-906 on glucuronyl transferase activity, the enzyme responsible for androstanediol glucuronide and androsterone glucuronide formation, was assessed in vitro using rat prostate tissue. Serum T levels were unchanged after all doses. Serum DHT, androstanediol glucuronide, and androsterone glucuronide were suppressed by 70, 40, and 56%, respectively, after the 10-mg dose, and by 82, 52, and 66% after the 100-mg dose (P less than 0.02 for the comparison between the 10 and 100-mg doses for all three steroids), respectively. Baseline serum T and DHT levels were strongly correlated (R = 0.89, P = 0.0002), as were androstanediol glucuronide and androsterone glucuronide levels (R = 0.78, P = 0.003), but there was no correlation between DHT levels and the levels of either conjugated steroid. MK-906 had no effect on glucuronyl transferase activity in vitro. It was concluded that single doses of MK-906 suppress both conjugated and unconjugated 5 alpha-reduced androgens. While all three steroids appeared to be good markers of systemic 5 alpha-reductase inhibition, further research will be needed to determine which steroid best reflects tissue DHT levels in patients receiving these inhibitors.