Maximum basal FSH predicts reproductive outcome better than cycle-specific basal FSH levels: waiting for a "better" month conveys limited retrieval benefits

Background: Elevated follicle stimulating hormone (FSH) is associated with poor vaginal oocyte retrieval (VOR) outcomes and cycle cancellations but intercycle variability in basal FSH reportedly does not predict ovarian response. Methods: We conducted a retrospective cohort study of basal FSH (n = 15573 cycles) in couples (n = 9132) who initiated IVF cycle(s) with basal estradiol (E-2) < 100 pg/mL between 2002 and 2014 to reevaluate this hypothesis. The most recent (current) FSH, maximum FSH (Max FSH) and prior cycle maximum basal FSH (PMax FSH) were computed for each cycle. Metaphase II (MII) oocyte counts were modeled by age, stimulation type, prior peak E2 level, prior MII count, Max FSH, PMax FSH and current FSH. Antral follicle counts, pregnancy, clinical pregnancy and live birth rates were modeled as secondary outcomes. Results: Max FSH level distinguished completed cycles from cancelled cycles better than PMax FSH or current FSH (AUC of 0.72, 0.71 and 0.61, respectively, p < 0.001). Fewer MIIs were retrieved (5.7 +/- 3.8) in cycles with Max FSH > 13 mIU/mL (n = 1475) than those with = 13 mIU/mL (n = 11978) (11.6 +/- 7.1) (p < 0.001). Max FSH was a better predictor of MII count than PMax FSH or current FSH after controlling for age, stimulation type, prior peak E2 level and prior MII count. Additional MIIs were retrieved on average in cycles with PMax FSH > 13 mIU/mL (n = 1930) whose current FSH was = 13 mIU/ml rather than > 13 mIU/ml (p < 0.01) after controlling for age, cycle number and stimulation type. However, no improvement in pregnancy or live birth rate was detected. Conclusions: Max FSH is the best FSH-based predictor of ovarian reserve. Retrieval benefits from waiting for a "better" month appear to exist but are limited.