Anti-mullerian hormone cut-off values for predicting poor ovarian response to exogenous ovarian stimulation in in-vitro fertilization

OBJECTIVES: (a) To establish the cut-off levels for anti-Mullerian hormone (AMH) in a population of Indian women that would determine poor response. (b) To determine which among the three ie.,: age, follicle stimulating hormone (FSH), or AMH, is the better determinant of ovarian reserve. STUDY DESIGN: Prospective observational study. SETTING: In vitro fertilization (IVF) unit of a tertiary hospital. MATERIALS AND METHODS: The inclusion criterion was all women who presented to the center for in-vitro fertilization/Intracytoplasmic sperm injection (IVF/ICSI). The exclusion criteria were age >45 years, major medical illnesses precluding IVF or pregnancy, FSH more than 20 IU/L, and failure to obtain consent. The interventions including baseline pelvic scan, day 2/3 FSH, luteinizing hormone (LH), estradiol estimations, and AMH measurement on any random day of cycle were done. Subjects underwent IVF according to long agonist or antagonist protocol regimen. Oocyte recovery was correlated with studied variables. The primary outcome measure was the number of oocytes aspirated (OCR). Three categories of ovarian response were defined: poor response, OCR ≤ 3; average response, OCR between 4 and 15; hyperresponse, OCR > 15. RESULTS: Of the 198 patients enrolled, poor, average, and hyperresponse were observed in 23%, 63%, and 14% respectively. Correlation coefficient for AMH with ovarian response was r = 0.591. Area under the curve (AUCs) for poor response for AMH, subject's age, and FSH were 0.768, 0.624, and 0.635, respectively. The discriminatory level of AMH for prediction of absolute poor response was 2 pmoL/l, with 98% specificity and 20% sensitivity. CONCLUSIONS: AMH fares better than age and FSH in predicting the overall ovarian response and poor response, though it cannot be the absolute predictor of non-responder status. A level of 2 pmol/l is discriminatory for poor response.