Comparing the inter-cycle variation of serum anti-Mullerian hormone and antral follicle count measurements in predicting ovarian response before IVF

Abstract

STUDY QUESTION: Does serum anti-Mullerian hormone (AMH) have less intercycle variation than antral follicle count (AFC) when used as an ovarian response predictor? SUMMARY ANSWER: Inter-cycle variations were significantly less with serum AMH than AFC measurement, although both have no significant difference in their predictive performance on poor ovarian response. WHAT IS KNOWN ALREADY: Both AMH and AFC measurements are recognized as the best markers for predicting ovarian response before IVF treatment. There was one report in the European population indicating that AMH (measured using the older ELISA) had less inter-cycle variation than AFC. However, there is no report on this in the Asian population nor using the newer generation ELISA. STUDY DESIGN, SIZE, DURATION: This observational study was a secondary analysis using AMH and AFC data from control subjects of another prospective clinical trial on the effect of DHEA pre-treatment. Only control subjects (n = 45) in the placebo arm were included in this secondary analysis. AMH and AFC measurements over four consecutive menstrual cycles before they underwent standard IVF treatment using the GnRH antagonist protocol were analysed. PARTICIPANTS/MATERIALS, SETTING, METHODS: The subjects underwent IVF treatment in a university-affiliated hospital in Hong Kong. AFC was measured in the early follicular phase (between days 2–5) by transvaginal ultrasonography in four consecutive cycles, and serum was obtained on the same occasions for AMH measurement using the Gen II AMH ELISA (Beckman-Coulter). The intra-class correlation coefficients (ICC) for AFC and AMH across the four study cycles, as well as their predictive performance on poor ovarian response, were compared. MAIN RESULTS AND THE ROLE OF CHANCE: Subjects were aged between 30 and 40 years. No significant difference was observed in AFC (p = 0.114) and AMH (p = 0.375) across the four study cycles (Friedman’s test). The single-measures ICC of AFC and AMH were 0.597 (95% CI 0.459–0.726) and 0.850 (95% CI 0.778–0.906) respectively. The average-measures ICC of AFC were 0.856 (95% CI 0.773–0.914) and 0.958 (95% CI 0.933–0.975), respectively. Hence, both single-measures and average-measures ICC were significantly higher with AMH than with AFC. The areas under the ROC curve of the four AFC measurements in predicting poor ovarian response (defined as three or less oocytes retrieved) in the subsequent IVF cycle ranged from 0.693 (95% CI 0.536–0.844) to 0.818 (95% CI 0.672–0.918) with no significant difference (p > 0.05) between the four cycles, whereas those of the four AMH measurement ranged from 0.785 (95% CI 0.636–0.895) to 0.821 (95% CI 0.676–0.920) with no significant difference (p > 0.05) between the four cycles. LIMITATIONS, REASONS FOR CAUTION: This was a secondary analysis on data obtained from another clinical trial; no data on intra-cycle variation of AFC and AMH was available. WIDER IMPLICATIONS OF THE FINDINGS: Our study in the Chinese population using the newer generation AMH ELISA concurred with the findings from a previous European study. Although both AFC and AMH are good predictors of ovarian response, AMH is less subjected to inter-cycle variation than AFC, hence allowing pre-IVF assessment at more flexible timing.