This study was designed to investigate whether androgens directly, independent of their aromatization to estrogens, disrupt gonadotropin secretion in hyperandrogenic women with the polycystic ovary syndrome (PCO). Pulsatile gonadotropin release and gonadotroph sensitivity to GnRH were determined on consecutive study days basally and during a primed continuous infusion of testosterone (T; n= 4; 100 micrograms/h; twice the mean production rate of T in PCO) or dihydrotestosterone (DHT; n= 5; 50 micrograms/h). To determine if the gonadotropin secretory changes during T infusion were secondary to spontaneous variation, four patients had two consecutive basal studies, and all patients received DHT on the third study day. T infusion that increased mean plasma T levels from 76+/-12 (+/-SE) to 315+/-28 ng/dl produced no significant changes in the amount or pattern of LH release or in LH sensitivity to GnRH. Mean plasma FSH levels decreased slightly but significantly during T infusion (basal, 242+/-29 vs. T 226+/-30 ng/ml LER-907; P less than 0.05 by two-tailed paired t test), but the pulsatile pattern of FSH release and FSH sensitivity to GnRH did not change. DHT infusion increased plasma DHT levels from 17+/-3 to 244+/-31 ng/dl, but did not alter the mean levels, pulsatile patterns, or sensitivity to GnRH of LH or FSH. These data suggest that androgens do not directly alter gonadotropin release in PCO. Thus, regulation of the hypothalamic-pituitary axis in women with PCO is different from that in men despite chronic exposure to hyperandrogenemia.