Abstract
OBJECTIVE: To assess the accuracy of antimullerian hormone (AMH) in predicting cumulative live birth rate (CLBR) within 1 year after treatment initiation in GnRH antagonist treatment cycles for in vitro fertilization (IVF). DESIGN: Observational (retrospective) substudy as part of an ongoing prospective cohort study. SETTING: University medical center. PATIENT(S): A total of 487 patients scheduled for IVF/intracytoplasmic sperm injection (ICSI). INTERVENTION(S): Patients starting their first IVF/ICSI cycle with 150 or 225 IU recombinant FSH and GnRH antagonist cotreatment were included. Serum samples collected before the first IVF treatment were used to determine AMH. Treatment data after treatment initiation for a maximum of 1 year were recorded. MAIN OUTCOME MEASURE(S): Prediction of CLBR with the use of AMH. RESULT(S): The model for predicting CLBR within 1 year included age at first treatment, AMH, type of infertility, and previous assisted reproductive technology treatment leading to live birth. The accuracy in discriminating between women who did or did not achieve a live birth was only 59%. AMH had intermediate added value in the prediction of CLBR as demonstrated by the net reclassification improvement (total 29.8). A nomogram based on age and AMH was developed by which a subgroup of patients could be identified with the poorest pregnancy prospects. CONCLUSION(S): The predictive accuracy of AMH for 1-year CLBR in GnRH antagonist treatment cycles was limited and did not yield much additional value on top of age. Withholding treatment based on predictors such as age and AMH, or a combination, remains problematic. CLINICAL TRIAL REGISTRATION NUMBER: www.clinicaltrials.gov, NCT02309073.
