Ovarian response to recombinant human follicle-stimulating hormone in luteinizing hormone-depleted women: Examination of the two cell, two gonadotropin theory

Objective: To evaluate the relative contribution of FSH to ovarian estrogen production. Design: Nonrandomized, prospective study. Setting: University of Toronto teaching hospital reproductive biology unit. Patients: Five women who had been treated with depot GnRH agonist with hormonal add-back for 4 to 48 months and who were confirmed to be gonadotropin depleted by both bioassay and RIA. Interventions: Subjects received 75 IU SC recombinant human FSH daily for 7 days followed by 150 IU daily for 7 days and 225 IU daily for the third week. Main Outcome Measure: Serum steroid determination and vaginal sonography for follicle size and endometrial thickness were performed serially and follicular fluid hormone levels were measured in two subjects. Results: Bioactive LH and FSH activity were less than the detection limit of the assay (0.1 mIU/mL; conversion factor to SI units, 1.00 for LH and FSH) before recombinant FSH treatment in all five women. In al subjects, at least one preovulatory follicle developed by the end of two to three weeks. Endometrial thickness increased to between 7 and 9 mm in four women. Mean serum E(2) in the five subjects increased from 17 pg/mL (range: 5 to 33 pg/mL; conversion factor to SI unit, 3.671) at baseline to 230 pg/mL (range: 37 to 489 pg/mL) at the end of the study. Follicular fluid E(2) concentrations ranged from 44,296 to 69,367 pg/mL in the four follicles aspirated. Conclusion: Our results indicate that LH is not necessary for ovarian E(2) production. We speculate that the granulosa cells, in the absence of detectable LH bioactivity, can use circulating adrenal androgens or constitutive or FSH-stimulated thecal androgens, to produce intrafollicular E(2).