Age-related decline in AMH is assay dependent limiting clinical interpretation of repeat AMH measures across the reproductive lifespan

Abstract Purpose: The aim of the study was to determine whether the assays exhibit an interaction with age and exhibit heterogeneous age related declines in AMH. Apart of chronological age, AMH variation was investigated with relation to menstrual cycle day (MCD). The goal implicates two questions: Are distributions of AMH concentrations homogenous after adjustment for the specific AMH assay? Does age-assay product has an effect on AMH depletion? Methods: The study was conducted by examining results of AMH tests performed for 12,917 women with four types of AMH assays: Immunotech I generation kit (IMI, 4016 samples), Beckman Coulter II generation kit RUO (BCII RUO, 3430 samples), Beckman Coulter II generation kit with IVD certificate (BCII IVD, 830 samples), and Ansh Labs I generation kit (AnshLabs, 4641 samples). Statistical analysis included ACNOVA and least square regression technique. Results: Menstrual cycle day has no effect on AMH measurements. On the other hand, AMH values differed substantially between the four assays, with a marked discordance in the rate of age-related AMH decline for the four assays (ranging from –8.16% (95% CI: –8.79, –7.54) to –11.53% (95% CI –12.20, –10.87), with a significant interaction between age and assay. Conclusions: (1) The distribution of AMH concentration is heterogeneous after controlling the age across assays; (2) the rate of AMH decline as a function of age is different for the four manual AMH ELISA assays.