Several different strategies were used to investigate the relationship between hyperandrogenism and hyperinsulinemia associated with polycystic ovary syndrome. Ketoconazole was given orally (400 mg/day) for 9 months to evaluate the effect of reduction in ovarian and adrenal androgens on insulin response (oral glucose tolerance test) in 35 women with polycystic ovary syndrome. Androgenic steroids (testosterone, androstenedione, dehydroepiandrosterone sulphate, and free testosterone index) decreased (p < 0.01), but basal insulinemia, maximum peak insulin, and insulin/glucose ratio showed no significant changes. One month after treatment was stopped, free testosterone index, and serum concentrations of androstenedione and testosterone, increased (p < 0.05), but no alterations were noted in insulin parameters. Body mass index was stable throughout the ten-month study period. Our findings suggest that endogenous androgens, no matter whether they are of ovarian or adrenal origin, do not play a major role in the modulation of hyperinsulinemia in patients with polycystic ovary syndrome.