To study the role of luteal estradiol (E2), we interrupted the supply of E2 during the luteal phase of E2 and progesterone (P) replacement cycles. Thirty-one women, aged 26-37 yr, with absent or inactive ovaries received three different treatment regimens: group I (n = 11) received transdermal E2 and vaginal P according to a protocol designed to approximate levels of estrone (E1), E2, and P seen during the menstrual cycle. Groups II (n = 11) and III (n = 9) received identical treatments, except that in group II no E2, and in group III no E2 or P, was administered after day 15. Endometrial biopsies were obtained on days 20 and 24 in groups I and II, and on days 14 and 20 in group III. In group I, plasma E1 and E2 reached menstrual cycle levels, whereas in groups II and III, discontinuation of the E2 supply on day 15 resulted in a prompt decrease to castrate levels of plasma E1 and E2. In groups I and II, menopausal FSH and LH levels decreased to 26 +/- 6 and 30 +/- 7 IU/L, respectively, on day 13 (mean +/- SEM). In group I, administration of E2 and P starting on day 15 further lowered plasma gonadotropin levels. In group II, administration of P only failed to induce a similar decrease in plasma FSH and LH. No uterine bleeding occurred before day 25 in women of groups I or II, while women of group III bled within 2 days of E2 withdrawal. Endometrial biopsies were similar in groups I and II. Histological features were characteristic of early and late luteal phases on days 20 and 24, respectively. Endometrial maturation assessed by estrogen and progesterone receptors identified by immunocytochemistry showed the typical distribution seen on day 24 of the menstrual cycle with no difference between groups I and II. We conclude that in women deprived of ovarian function, administration of P only after 14 days of E2 priming prevented uterine bleeding and induced normal secretory transformations of the endometrium, but failed to suppress plasma gonadotropins.