A randomized, controlled trial of estradiol replacement therapy in women with hypergonadotropic amenorrhea

Premature ovarian failure is classically defined as menopause occurring before age 40 and is associated with elevated serum FSH levels. If elevated FSH levels indicate lack of ovarian feedback and depletion of primordial follicles, women with prematurely elevated FSH levels should have infertility. However, there are many reports of pregnancies in affected women occurring during estrogen therapy leading to the hypothesis that estrogen may have a salutary effect on folliculogenesis and conception. This randomized, controlled trial was designed to investigate whether estrogen replacement therapy offered a significant therapeutic benefit in hypergonadotropic amenorrhea and to evaluate the potential pathophysiologic mechanisms that would explain the reported pregnancies. Thirty seven women, aged 16 to 40, with menstrual dysfunction and documented FSH levels elevated above the 95% confidence limits of the mid-cycle gonadotropin peak of the normal menstrual cycle (> 40 IU/L 2nd IRP hMG in our RIA) on at least two occasions, entered the study. The average duration of their amenorrhea was 15.9 months (range 2-96 months). Subjects were randomized to begin estradiol replacement (micronized estradiol [Estrace TM], 2 mg orally each day) or no therapy for 6 weeks in a 12-week, cross-over design with weekly monitoring by both pelvic ultrasonography and serum hormone levels. Thirty-one women completed the entire randomized study. As expected, estradiol therapy increased mean serum estradiol levels by 98 pg/mL and was associated with a significant decrease in mean LH and FSH levels (LH: 45.4 IU/L 2nd IRP hMG vs. 37.1 IU/L, FSH: 63.4 IU/L vs. 40.6 IU/L, geometric means). However, there was no effect of estradiol replacement on mean ovarian volume, the number or size of new follicles, or the ovulation rate in all subjects or in the subset with no identified cause for their hypergonadotropic hypogonadism (n = 20). Two pregnancies occurred during the randomized trial, one on and one off estradiol. In both arms of the study, the majority of subjects developed cystic ovarian structures by ultrasound that were temporally associated with increasing serum estradiol levels, indicating functional ovarian follicles. Seventy-eight percent of all subjects grew at least one new follicle over 10 mm in diameter and 46% ovulated at least once, as determined by a serum progesterone level more than 4 ng/mL. Although ovulations were significantly more common in the 10 women subjects who had less than 3 months of amenorrhea (all of whom ovulated) than in the 27 with greater than 3 months of amenorrhea (only 7 of whom ovulated (26%), P < 0.001), there was no significant difference in eventual pregnancies (2 of the 10 women with less than 3 months of amenorrhea vs. 3 of the 27 with greater than 3 months of amenorrhea, P = 0.47). We conclude that in hypergonadotropic women with amenorrhea: 1) folliculogenesis occurs often but is less frequently followed by ovulation and rarely by pregnancy, suggesting that elevated FSH is a marker of oocyte dysfunction occurring distinct from and earlier than granulosa cell or follicular dysfunction; and 2) estrogen therapy does not improve the rate of folliculogenesis or ovulation.