Can Menopause Prediction Be Improved With Multiple AMH Measurements? Results From the Prospective Doetinchem Cohort Study

Context: Anti-Mullerian hormone (AMH) levels are used worldwide as a screening tool for the duration of the female reproductive lifespan. Although AMH levels are associated with age at menopause, individual predictions of menopause with a single AMH measurement are unreliable. Objective: This study investigated whether individual AMH decline patterns can improve the prediction of menopause compared with a single measurement. Design: The study population comprised 2434 premenopausal women from the population-based Doetinchem Cohort Study. Participants were followed up every 5 years for a total of 20 years, and AMH was measured in 6699 plasma samples with the picoAMH assay. Longitudinal statistical modeling was combined with time varying Cox modeling, to integrate multiple AMH measurements per woman. Results: The mean age at menopause was 50 years, and 7.4% of the women who reached menopause during follow-up did so before age 45 years. For a 25-year-old, the AMH decline rate between ages 20 and 25 years increased the C-statistic of menopause prediction from 0.64 to 0.69. Beyond that age, the AMH decline rate did not improve predictions of menopause or early menopause. For women younger than age 30 years, for whom menopause prediction is arguably most relevant, the models underestimated the risk of early menopause. Conclusion: These results suggest that knowledge of the AMH decline rate does not improve the prediction of menopause. Based on the low discriminative ability and underestimation of the risk of early menopause, the use of AMH as a screening method for the timing of menopause cannot currently be advocated.