To determine if progesterone (P) does affect gonadotropin secretion by acting directly on the pituitary, six women with hypothalamic gonadotropin deficiency were studied. They were treated with 17 beta-estradiol (E2; 2 mg/day, orally) to induce P receptors and maintain constant plasma E2 levels during two 15-day periods separated by 1 month. GnRH was administered iv at a dose of 10 microgram/pulse every 90 min during the last 5 days of E2 treatment. Either P (400 mg/day) or a placebo was administered intravaginally in a cross-over randomized design during the 5 days of pulsatile GnRH therapy. A baseline study of pulsatile LH secretion was performed, with sampling performed every 10 min for 8 h. The sampling was then repeated on day 15 of each study period at the end of pulsatile GnRH administration. Plasma levels of E2 and P were measured every day during the 5 days of either GnRH and P or GnRH and placebo treatment. In the six patients, the observed apulsatile pattern of LH during the baseline study confirmed the diagnosis of complete gonadotropin deficiency. Plasma E2 levels were not significantly different at the time of each pulse analysis (288 +/- 61 vs. 252 +/- 77 pmol/L). The plasma P level achieved with the vaginal pessaries was 22 +/- 5 nmol/L. P treatment resulted in all cases in a significant increase in the mean plasma LH level (5.2 +/- 0.9 vs. 3.6 +/- 0.7 IU/L after GnRH plus placebo; P less than 0.001). Furthermore, LH pulse amplitude was significantly increased by P compared to placebo (3.1 +/- 0.3 vs. 1.4 +/- 0.1 IU/L, respectively; P less than 0.01). Mean plasma FSH levels were significantly increased by GnRH regardless of whether P or placebo was present. In conclusion, these data indicate that a short exposure to physiological levels of P in the range of early luteal phase levels has a stimulatory effect on LH secretion by acting directly at the pituitary level.