OBJECTIVE To compare the use of gonadotropin-releasing hormone agonist (GnRH-a) with human chorionic gonadotropin (hCG) for triggering the final stage of follicular maturation for in vitro fertilization (IVF). METHODS In vitro fertilization outcome was determined in a randomized, prospective study. METHODS The University of Toronto IVF program at The Toronto Hospital, Toronto General Division. METHODS One hundred seventy-nine women in the IVF program were given a subcutaneous injection of leuprolide acetate (500 micrograms) or an intramuscular injection of hCG (5,000 IU) 34 to 36 hours before oocyte retrieval. Vaginal progesterone (P) suppositories (50 mg) were used two times a day for luteal phase support. A subgroup of 41 women had serum estradiol (E2) and P levels determined 2 and 7 days after embryo transfer (ET). METHODS Pregnancy rates and luteal phase E2 and P were compared. RESULTS In the GnRH-a group, there were 18 pregnancies from 84 ETs (20%). In the hCG group, there were 19 pregnancies from 95 ETs (19%). Luteal phase E2 and P levels were significantly lower in the GnRH-a group compared with the hCG group, and 18% of the former group had an apparent short luteal phase. CONCLUSIONS Gonadotropin-releasing hormone agonist appears to be an effective alternative to hCG for inducing follicular maturation in IVF. The lower luteal phase E2 concentrations may potentially be beneficial in preventing ovarian hyperstimulation and for enhancing implantation. Better luteal phase support or a different dose of GnRH-a is needed to prevent luteal phase deficiency.