Background: Numerous studies have evocated the clinical usefulness of serum AMH levels as a predictor of ovarian response and pregnancy in assisted reproductive technology cycles. Nevertheless, the analysis of the literature shows a great dispersion in serum AMH concentrations obtained with different methods from almost comparable populations.
Methods: We compared two commercial immunoassays (AMH Beckman Coulter ELISA and AMH DSL ELISA) and we evaluated the AMH levels in serum as a prognosis value for ovarian response and pregnancy in assisted reproductive technology cycles.
Results: We found a close linear relationship between the two methods but AMH levels were almost 4.6-fold lower with the DSL kit than with the Beckman Coulter kit. We found a significant and positive correlation between the number of mature ovocytes inseminated and AMH levels obtained with the two methods. Whatever the ELISA used, we found no significant difference between AMH level of pregnancy and non pregnancy groups. Indeed, using the Beckman Coulter method, all pregnant patients had serum AMH levels over 1.4 microg/L. Conversely, no cut-off value can be found for the DSL kit.
Conclusion: Our results show clearly for the first time that AMH results are method dependent even if the correlation obtained between the two methods remained excellent. The Beckman Coulter AMH ELISA should produce clinical agreement when used for prognosis purposes on patients undergoing assisted reproduction.