Comparison of Automated Anti-Müllerian Hormone Assays and Antral Follicle Count in Predicting Ovarian Response During Ovarian Stimulation

Abstract

Objective: Various parameters had been used to predict ovarian response. Among them, Anti-Müllerian Hormone (AMH) and antral follicle count (AFC) demonstrate the most favourable analytical and performance characteristics. In this pilot study, we aim to determine the cut-off levels of AMH using automated AMH assays and AFC in the prediction of poor and high responders. Study Design: Prospective study of 43 women between 21 to 45 years old scheduled for assisted reproduction. AMH levels on day 3 of menstruation were analysed using two immunoassay kits, namely the Beckman Coulter Access AMH and the Roche Elecsys AMH on the two automated analysers Beckman Coulter DxI 800 and Roche Cobas e602 respectively. AFC was also assessed on day 3 of menstruation prior to in vitro fertilization (IVF). These were compared with the number of oocytes retrieved after controlled ovarian stimulation. Results: AMH (Beckman Coulter Access AMH and Roche Elecsys AMH) highly correlated with AFC and the number of oocytes retrieved after ovarian stimulation. Beckman Coulter Access AMH was the better predictor for poor ovarian response with ROC [Formula: see text] of 0.83. For the prediction of a high response, AFC had a higher ROC [Formula: see text] of 0.95. Through ROC, the AMH cut-off level for poor ovarian response was 2.23 ng/ml with Beckman Coulter Access AMH and 2.02 ng/ml with Roche Elecsys AMH, while the AMH cut-off for a high ovarian response was 5.19 ng/ml with Beckman Coulter Access AMH and 4.60 ng/ml with Roche Elecsys AMH. For AFC, the cut-off for poor ovarian response was 18 and for high response was 34. Conclusion: AMH and AFC are reliable predictors of ovarian response. Establishment of specific levels may improve individualised controlled ovarian stimulation and optimise the oocyte yield. Larger studies are required to establish these findings.