Fifteen young women with a diagnosis of secondary hypothalamic amenorrhea of at least 2 years' duration were given either 50 mg naltrexone daily or placebo, following a randomized double-blind crossover scheme. Seven patients did not menstruate with either therapy. In the other eight, the following results were recorded (mean +/- SD and range): a cycle length of 28.7 +/- 7.6 (12-45) days for naltrexone compared with 30.8 +/- 5.9 (16-43) days for placebo, a follicular phase length of 20.8 +/- 5.4 (14-34) days for naltrexone and 23.2 +/- 4.3 (19-32) days for naltrexone and 8.3 +/- 1.6 (5-10) days for placebo. The number of ovulatory cycles was 18 of 24 (75%) with naltrexone and eight of 24 (33%) with placebo (P less than .05). Most luteal phases were short. In five normally menstruating women, we gave either naltrexone or placebo in the luteal phase using a crossover blinded scheme. Steroidogenesis in the normal luteal phase was not impaired by naltrexone therapy. In functional hypothalamic amenorrheic patients with normal weight, menstruation might be restored by either placebo or naltrexone, but naltrexone provides a clinical and therapeutic advantage by increasing the ovulation rate.