Women with polycystic ovary syndrome (PCOS: hyperandrogenism and oligomenorrhea) have been shown to have exaggerated ovarian 17-hydroxyprogesterone (17-OHP) responses after GnRH agonist stimulation, suggestive of disordered ovarian cytochrome P450c17 alpha activity. However, most hyperandrogenic women also have an exaggerated LH response to both native GnRH and GnRH agonists. To assess whether the known LH hyperresponsiveness in PCOS patients mediates their exaggerated 17-OHP response to GnRH agonist challenge or whether the 17-OHP response can be duplicated by direct stimulation of the ovary with a fixed amount of hCG, 25 PCOS women [age, 20.6 +/- 1.1 yr; body mass index (BMI), 24.9 +/- 1.3 kg/m2] and 5 controls (age, 29.4 +/- 2.3 yr; BMI, 29.2 +/- 3.0) underwent both GnRH agonist (leuprolide acetate) and hCG testing. In addition, 23 normal women (age, 26.5 +/- 1.5 yr; BMI, 27.2 +/- 1.7 kg/m2) underwent hCG testing, and 21 other normal women (age, 19.3 +/- 0.5 yr; BMI, 23.0 +/- 0.8 kg/m2) underwent leuprolide acetate challenge. Blood was sampled before and 24 h after (the previously determined time of the peak response) leuprolide acetate (500 micrograms, sc) and hCG (5000 IU, im) stimulation tests. The tests were administered during the early follicular phase in women who were ovulatory and in randomized order in the subjects receiving both stimuli. Peak serum 17-OHP levels did not differ between leuprolide acetate or hCG stimulation in PCOS patients or controls when measured at 24 h (324.9 +/- 41.9 vs. 360.4 +/- 54.0 in PCOS; 183.2 +/- 19.6 vs. 141.2 +/- 11 ng/dL in controls). Peak 17-OHP levels after hCG challenge and GnRH agonist stimulation were highly correlated (r = 0.82; P < 0.001) in the subjects receiving both stimuli. Although leuprolide acetate elicited a higher estradiol (E2) response than hCG in all subjects, serum E2 levels increased significantly after hCG treatment in both patients and controls (P < 0.001 and P < 0.0001). The small, but significant, increase in serum E2 after hCG stimulation suggests that a rigid two-cell model of ovarian steroidogenesis may be an oversimplification of in vivo physiology. Our results suggest that exaggerated 17-OHP responses to GnRH agonist stimulation in hyperandrogenic women are not mediated by the known hyperresponsiveness of LH in these patients, but are due to increased ovarian androgen sensitivity to LH stimulation.