We investigated whether obesity was a marker for a neuroendocrinologically distinct form of the polycystic ovary syndrome (PCO). Further, since women with PCO have significantly higher basal and/or glucose-stimulated plasma insulin levels, we also examined the effects of chronic hyperinsulinemia on gonadotropin and gonadal steroid secretion. Ten obese women (nine with acanthosis nigricans) and five nonobese women (one with acanthosis nigricans) with PCO as well as seven obese and six nonobese women of comparable age and weight in the midfollicular phase of their cycles were studied. Pulsatile gonadotropin release was determined for 6-24 h as well as gonadotroph sensitivity to GnRH (10 micrograms, iv). The obese PCO women had significantly increased basal and glucose-stimulated plasma insulin levels compared to the other groups, the nonobese PCO and the obese normal women had similar insulin levels, and the nonobese normal women had the lowest insulin levels. All four groups had similar plasma estradiol levels. Both the obese and the nonobese PCO women had similar and significantly higher mean plasma LH levels, LH pulse amplitude, and integrated LH responses to GnRH compared to values in both normal groups (P less than 0.01 to P less than 0.001); the obese PCO women did not differ from the nonobese PCO women. The mean LH pulse frequencies per 6 h were similar in the four groups. FSH secretion did not differ significantly in the four groups. The levels of the putative gonadal feedback steroids, plasma total and non-sex hormone-binding globulin-bound testosterone, non-sex hormone-binding globulin-bound estradiol, and estrone, were similar in both PCO groups and were significantly higher than those in both normal groups (all P less than 0.001). The only independent effect of obesity was on plasma androstenedione levels and the androstenedione to estrone ratio, both of which were significantly higher in PCO women than normal women (P less than 0.01 to P less than 0.001), but significantly lower in the obese (PCO and normal) women than in nonobese (PCO and normal) women (P less than 0.05). These findings suggest that 1) the impact, if any, of obesity in PCO is not reflected in discernible changes in gonadotropin release or in the gonadal steroid feedback environment; and 2) insulin does not have a major role in the perpetuation of PCO, since obese and nonobese PCO women had similar reproductive hormone levels despite significantly different degrees of hyperinsulinemia.