Objective: To determine if an elevated FSH:LH ratio predicts response in infertile patients undergoing controlled ovarian hyperstimulation (COH) for IVF-ET. Design: Retrospective study. Setting: The Division of Reproductive Endocrinology at the Mount Sinai Medical Center, New York, New York. Participants: Seventy-four patients undergoing IVF-ET using similar protocols for COH with day 3 FSH, LH, and E(2) testing available for analysis. All patients were <41 years of age and had day 3 serum FSH <15 mIU/mL (conversion to SI unit, 1.00). Main Outcome Measures: Follicle-stimulating hormone:LH ratio, day 8 serum E(2), peak serum E(2), cancellation rate, pregnancy rate, and number and size of follicles. Results: An FSH:LH ratio greater than or equal to 3.6 (group I) predicted a poor response to COH (sensitivity 85.7% and specificity 95%). There were no significant differences regarding day 3 serum FSH and ampules of gonadotropins used for COH. Group I (ratio greater than or equal to 3.6) patients responded to COH with lower day 8 E(2) (97 +/- 18 versus 319 +/- 36 pg/mL; conversion factor to SI unit, 3.671), peak E(2) (422 +/- 115 versus 2,368 +/- 183 pg/mL), and fewer follicles >15 mm (1.3 +/- 0.5 versus 17.1 +/- 1.0). In group I the cycle cancellation rate (12/14) was significantly higher than the group II cycle cancellation rate (2/60) and pregnancy rate in group II (ratio <3.6) was 25%. Conclusions: The FSH:LH ratio may increase before a dramatic increase in serum FSH is observed and appears to be a useful marker of ovarian reserve.