Can anti-müllerian hormone (AMH) predict the outcome of gonadotropin induction/intrauterine insemination (GN/IUI) cycles among infertile women?

Abstract

To evaluate the ability of AMH to predict treatment outcomes among infertile women undergoing Gn/IUI, given emerging evidence suggesting otherwise among women of normal fertility[1]. Retrospective cohort study. Sample size/Intervention: 1110 Gn/IUI cycles (from 503 women) at a large academic center were analyzed (11/2007-10/2017). Cycles were stratified by AMH serum concentration cutoffs, based on previously published literature: low (<0.7 ng/ml, n=273) and normal (≥0.7 ng/ml, n=837). Outcomes measures: Cumulative probability of achieving i) positive pregnancy test (PPr) and ii) clinical pregnancy (CPr) over a maximum of 4 completed cycles per attempt. Statistics: Pearson’s χ2, t-test, or nonparametric tests were used as appropriate. Cumulative probabilities of achieving the outcomes of interest were determined via the Kaplan-Meier failure function. Cox proportional hazard models (adjusted for age and BMI) were used to determine fecundability ratios (FR) for each AMH group. P <0.05 was considered significant. As expected, mean (SD) age, day-3 FSH, total gonadotropin dose, and duration of stimulation were different between groups: 37.5 (3.8) vs. 35.3 (3.9) years, p<0.01; 10.2 (5.1) vs. 7.0 (2.3) IU/L, p<0.01; 1836.5 (1541.1) vs. 750.8 (582.2) IU, p<0.01; and 10.2 (2.5) vs. 12.2 (4.4) days, p<0.01, for low vs. normal AMH groups, respectively. Conversely, BMI (25.3 [5.1] vs. 24.1 [5.1] kg/m2, p=0.17) and peak estradiol (508.1 [237.6] vs. 462.5 [286.4] pg/ml, p=0.15) did not differ significantly. Women with AMH >0.7 ng/ml were more likely to have either a PPr or CPr with an unadjusted cumulative probability (95%CI) of 31.4% (20.4-46.4) vs. 46.3% (37.4-56.3), p<.005, and 19.9% (11.6-32.7) vs. 29.3% (21.9-38.3), p=0.017, for low vs. normal AMH groups, respectively. Following the same trend, the estimated adjusted FRs revealed that women with AMH >0.7 ng/ml were more likely to have a PPr (adjusted FR: 1.52, 95%CI:0.99-2.35) and CPr (adjusted FR: 1.47, 95%CI:0.92-2.35). Similarly, when further stratified to low (<0.7 ng/ml), normal (0.7-8.4 ng/ml) and high (≥8.5 ng/ml) AMH groups, women with AMH ≥8.5 ng/ml were more likely to have a PPr (adjusted FR: 1.90, 95%CI:1.09-3.28) and CPr (adjusted FR: 1.76, 95%CI: 0.96-3.20) when compared to women with AMH <0.7 ng/ml. Despite the recent evidence that low AMH in a population of fertile women was not associated with reduced fecundability, our findings suggest that among infertile women undergoing Gn/IUI cycles, lower AMH might be associated with lower probability of a successful treatment outcome.