Adrenal androgen (AA) excess, primarily in the form of dehydroepiandrosterone sulfate (DHEAS), affects over 50% of women with the polycystic ovary syndrome (PCOS). Nonetheless, it is unclear what role AA excess plays in the PCOS-associated oligo-ovulation. We have hypothesized that AAs are important in the maintenance of the ovulatory dysfunction of women with PCOS and AA excess, which can be improved by glucocorticoid suppression. To test our hypothesis we prospectively studied 36 unselected women, ages 18-40 yr, with PCOS; i.e. oligomenorrhea (cycles > 35 days in length), and clinical/biochemical evidence of hyperandrogenism (i.e. hirsutism and/or hyperandrogenemia), after the exclusion of related disorders. After informed consent, all patients underwent an acute ACTH-(1- 24) stimulation test, measuring androstenedione, dehydroepiandrosterone (DHEA) and cortisol (F), and were then treated with dexamethasone 0.5 mg/day for four cycles. Ovulatory function was assessed before and during treatment using a basal body temperature calendar and day 22-24 progesterone (P4) levels. If patients were anovulatory (P4 < 4 ng/mL), a withdrawal bleed was induced by the administration of 100 mg P4 in oil i.m. Before and during treatment the levels of total and free testosterone (T), sex hormone-binding globulin, androstenedione, DHEA, DHEAS, cortisol, luteinizing hormone (LH) and follicle-stimulating hormone (FSH) were monitored. With therapy, all patients demonstrated a significant decrease in all androgens (40-60%), a 24% increase in sex hormone-binding globulin, and no change in LH/FSH. Mean body weight increased by over 4 kg (4.4%) during treatment. Of the 138 cycles monitored, 78% remained anovulatory. Twenty-five percent, 17%, 14%, and 20% of the first, second, third, and fourth treatment cycles, were ovulatory, respectively (P = 0.381). Of the 36 patients studied, 18 (50%) did not demonstrate a single ovulatory cycle (i.e. a day 22-24 P4 level > 4 ng/mL); and of the remaining, 10 (28%) had only one, five (14%) had two, and three (8%) had three ovulatory cycles. There were no significant differences either in physical features, basal hormones, adrenal response to ACTH stimulation, or hormonal levels at the end of treatment, between those women ovulating and those not. Finally, there were no differences in ovulatory response to dexamethasone therapy between women with (n = 14) and without (n = 22) DHEAS excess (i.e. DHEAS > 2750 ng/mL). In conclusion, the data from this prospective study do nor suggest that continuous dexamethasone suppression results in consistent ovulation in any PCOS patient, regardless of basal DHEAS levels. Furthermore, this treatment is associated with significant side-effects, notably weight gain. Finally, these data suggest that, while AA may be an important risk factor for PCOS, once the syndrome is established, they play a limited role in the associated ovulatory dysfunction.