Retrospective study of ovarian reaction and fertility in women with discordant antimullerian hormone and follicle-stimulating hormone/luteinizing hormone ratios during in vitro fertilization.

Abstract

The objective of this study was to examine the predictive ability of follicle-stimulating hormone (FSH)/luteinizing hormone (LH) ratio and antimullerian hormone (AMH) levels in the same cohort when both are inconsistent and to identify which has an excellent ability to predict live birth and ovarian response to in vitro fertilization (IVF). A retrospective cohort study was performed within 6096 IVF cycles executed between January 2016 and August 2019 at the Center for Assisted Reproduction, The Third Affiliated Hospital of Zhengzhou University, and Jiaozuo Maternity and Child Health Care Hospital. Initially, IVF cycles were classified according to basal FSH/LH ratio and AMH values, and the primary outcome was a comparison of live birth rate per cohort. Secondary outcomes included characteristics of the study individuals such as body mass index (BMI), age, antral follicle count, ovarian sensitivity index, cycle cancellation rate, and cycle outcome data. Women with FSH/LHhigh ratio and AMHnormal levels had a meaningly higher live birth rate compared with those with FSH/LHnormal ratio and AMHlow levels (46.59% vs. 21.21%, p < 0.001). In addition, women with FSH/LHnormal ratio and AMHlow levels were found to have a higher cancellation rate in their IVF cycles (80.98%). In women with FSH/LHnormal ratio, further multivariate analysis revealed that AMH level, age, number of retrieved oocytes, and FSH dosage were relevant risk factors for live birth. The relative risk of live birth was 0.11 (95% [CI] 0.06-0.20, p < 0.001) in patients with AMHlow compared with patients with AMHnormal . It also suggested that the probability of AMHlow level may be higher as the women's age (≥35 years, odd ratio [OR] 1.94, 95% [CI] 1.44-2.61; p < 0.001) and increasing BMI (≥28 kg/m2 , OR 2.38, 95% [CI] 1.33-4.27; p=0.004). Receiver operating characteristic curve analysis indicated that AMH had higher sensitivity and specificity to predict live birth compared with FSH/LH (AUC 0.627 vs. 0.539). AMH levels can be an excellent predictor of the discrepancy between FSH/LH ratio and AMH levels regarding living birth rates in women undergoing IVF.