Anti-Mullerian hormone in the diagnosis of polycystic ovary syndrome: can morphologic description be replaced?

Abstract

Study question: Can anti-Müllerian hormone (AMH) level replace the morphologic description in the diagnosis of polycystic ovary syndrome (PCOS) and what is the relationship between AMH and different diagnostic criteria of PCOS?

Summary answer: AMH may be a good substitute for polycystic ovarian morphology (PCOM) in diagnosing PCOS.

What is known already: AMH has been suggested as an alternative to antral follicle count (AFC) in diagnosing PCOS. Cut-off values for AMH studied so far show an acceptable specificity but a rather poor sensitivity, leaving up to one-third of PCOS women undiagnosed.

Study design, size, duration: We used data from a cross-sectional, case-control study on women with prior preterm birth and their controls, i.e. women with prior full-term birth. Among 262 women, 56 met the Rotterdam criteria (PCOS-R) and 44 the Androgen Excess-PCOS Society (PCOS-AES) criteria of PCOS.

Participants/materials, setting, methods: Fasting blood samples were collected, a transvaginal ultrasound investigation and a clinical examination were performed. PCOS-R and PCOS-AES were re-diagnosed by replacing PCOM with AMH. Main outcome measures were the prevalence of PCOS, PCOM, hirsutism, oligoamenorrhoea and serum levels of AMH and androgens.

Main results and the role of chance: When replacing PCOM with AMH, the specificity and sensitivity for identifying PCOS were 97.1 and 94.6% according to the PCOS-R criteria and 97.2 and 95.5% according to the PCOS-AES criteria, respectively, at an AMH cut-off value of 20 pmol/l.

Limitations, reasons for caution: The results need to be confirmed when international standards and methods for AMH measurements are established.

Wider implications of the findings: AMH may be a good substitute for PCOM in diagnosing PCOS.

Study funding/competing interest(s): This study was financed by the Cooperative of Central Norway Regional Health Authority and Norwegian University of Science and Technology. The authors have no interests to disclose.

Trial registration number: This study is registered at www.clinicaltrials.gov as NCT01355536.