Endocrine Abstracts

Anti-Müllerian hormone (AMH) levels strongly correlate with the number of antral follicles in the ovary. In women with polycystic ovary syndrome (PCOS), this is reflected by significantly increased serum AMH levels. AMH levels are therefore suggested as a proxy for polycystic ovary morphology (PCOM) in PCOS diagnosis. Different assays are available to measure serum AMH levels. However, lack of a golden standard and the use of different antibodies to detect AMH, have led to inter-assay variability. Little is known about inter-assay correlation in women with increased AMH levels, especially in PCOS. Hence, our aim was to investigate the correlation of AMH values between different AMH assays and with total follicle count (TFC) in a large cohort of PCOS patients. Serum AMH levels were measured in 1660 PCOS patients, diagnosed using the Rotterdam criteria. AMH levels were measured by three different AMH assays: (1) Gen II ELISA [Beckman Coulter]; (2) picoAMH assay [AnshLabs]; and (3) Automated Elecsys assay [Roche]. Patients were divided in subgroups based on the reported AMH cutoff values for PCOM: low AMH (<2.80 ng/ml), mid AMH (2.80−7.04 ng/ml) and high AMH level (>7.04 ng/ml). Passing Bablok regression was used for the comparison between assay methods. Spearman’s correlation rank was used to assess the correlation between AMH levels and TFC. The inter-assay correlations over the total range of AMH levels were: Gen II vs Elecsys: 0.81; picoAMH vs Gen II: 0.81; picoAMH vs Elecsys: 0.94. Stratification in three AMH subgroups revealed an AMH level dependent inter-assay correlation. A strong inter-assay correlation was present in both low and high AMH subgroups, ranging from 0.62−0.86. The correlation in the mid AMH level subgroup was only moderate, with correlation coefficients ranging between 0.28−0.56. A positive correlation was present over the total range of AMH levels and TFC, with correlation values ranging from 0.57−0.62. However, subgroup analysis showed that independently of assay method used, the correlation decreased in all three AMH subgroups and became moderate at best with coefficients ranging between 0.11−0.45. In conclusion, in our cohort of PCOS patients both inter-assay correlation and correlation between AMH level and follicle count depend on the range of serum AMH level. While a high AMH level may reflect the presence of PCOM, our results suggest that it does not accurately reflect the total number of follicles in PCOS. This once more emphasizes the need of a standardization of AMH measurement.