OBJECTIVE: Luteal phase abnormalities are known to complicate ovulation induction with gonadotropins. This study was performed to test the effect of a modified human chorionic gonadotropin (hCG) regimen on the luteal phase during gonadotropin treatment. DESIGN: Fifteen women from a private practice setting volunteered to be studied during each of two nonconception, gonadotropin-stimulated cycles. After ovarian stimulation with human menopausal gonadotropins (hMG), hCG was administered either as a single dose of 10,000 IU (single dose) or in two divided doses of 5,000 IU given 1 week apart (split dose). MAIN OUTCOME MEASURES: Early, midluteal, and late luteal estradiol (E2) and progesterone (P) levels and luteal phase lengths were measured, and their median values and intraquartile ranges (IQR) compared using nonparametric analysis. RESULTS: Early and midluteal E2 and P levels were similar regardless of which hCG regimen was administered. The median late luteal E2 level was 1,146.0 pg/mL (the IQR ranged from 633 to 1,650, IQR = 1,017) with the split-dose regimen and 240.0 pg/mL (the IQR ranged from 150 to 460, IQR = 310) with the single-dose regimen. The median late luteal P level was 108.0 ng/mL (the IQR ranged from 58.5 to 129, IQR = 70.5) with the split-dose regimen and 4.2 ng/mL (the IQR ranged from 1.9 to 11.7, IQR = 9.8) with the single-dose regimen. Median luteal phase lengths were 16 days (the IQR ranged from 15 to 17, IQR = 2) for the split-dose regimen and 11 days (the IQR ranged from 10 to 12, IQR = 2) for the single-dose regimen. CONCLUSION: In hMG-stimulated cycles, a second dose of hCG given during the midluteal phase significantly increases late luteal E2 and P levels and consistently lengthens the luteal phase.