The premenstrual syndrome (PMS) has been proposed to result from excessive exposure to and/or withdrawal of brain opioid activity during the luteal phase. Because hypothalamic opioids are believed to modulate GnRH secretion, in part under the influence of ovarian steroids, we performed longitudinal studies of gonadotropin and ovarian steroid secretion across ovulatory, symptomatic cycles of 17 PMS patients and 8 normal volunteers. Pulsatile LH secretion was measured every 10 min for 8 hr at times when central opioid activity was expected to be low (early follicular phase), high (mid-luteal phase; ML), and declining (late luteal phase). In both subject groups, a cycle-phase effect was observed for LH pulse frequency (p = < 0.001) and amplitude (p = 0.002), and for the transverse mean concentrations of LH (p = 0.05), FSH (p < = 0.001), estradiol (E2) (p = < 0.001) and progesterone (P) (p = < 0.001). ML P secretion in PMS patients was pulsatile, and mean concentrations (over 30-60 min) were similar to those of normal controls. The changes in pulsatile LH secretion across the cycle were not different in the PMS patients compared to the normal women, though mean FSH in the ML phase was higher in the PMS group (p = < 0.05). The similar changes in luteal LH pulse frequency fail to provide evidence that GnRH secretion is impaired, thus challenging the view that the neuroregulation of the menstrual cycle in women with PMS is markedly altered.