Revisiting the serum level of anti-Mullerian hormone in patients with functional hypothalamic anovulation

STUDY QUESTION: Are serum levels of anti-Mullerian hormone (AMH) normal in patients with functional hypothalamic anovulation (FHA)? SUMMARY ANSWER: Our study confirms that in the general FHA population, serum AMH levels are not decreased, but if patients with polycystic ovarian morphology (PCOM) are excluded, levels become significantly lower, as in other situations of gonadotropic insufficiency. WHAT IS KNOWN ALREADY: In most situations of low LH (physiological, pharmacological or pathological), serum AMH levels are low. However, paradoxically, many publications have reported normal or even increased serum AMH levels in FHA patients. STUDY DESIGN, SIZE, DURATION: Retrospective observational study conducted in an academic centre. The data concerning the study population was collected between 2006 and 2015 from a database including clinical, biological and ultrasound information. PARTICIPANTS/MATERIALS, SETTING, METHODS: A total of 45 FHA patients were compared to 37 controls matched based on age and body mass index (BMI). Serum LH, FSH, androstenedione, total testosterone, prolactin and AMH levels were measured by immunoassay. We defined PCOM with strict criteria: a follicle number per ovary (FNPO) >= 12 or >= 19 per ovary, depending on the date on which the assessment was carried out and the ultrasound device. An AMH level >= 35 pmol/l could be a substitute for an excess FNPO. Controls meeting these criteria were not included in this study. MAIN RESULTS AND THE ROLE OF CHANCE: There was no significant difference in the ranges of AMH levels between FHA and controls. Using strict criteria to define PCOM status, 46.7% of FHA patients had PCOM. After excluding these patients, the levels of AMH were significantly lower (P < 0.002) in FHA patients compared to controls. Within the FHA group, patients with PCOM had significantly higher ranks of AMH levels and BMI than those without PCOM. However, within the PCOM subgroup, the ranks of LH, FSH and A levels were still lower than in controls (P < 0.0001, <0.002 and <0.05, respectively). The positive correlation between AMH and LH was significant in the controls but not in the FHA group. However, in the FHA PCOM , there was a strong positive correlation between BM1 and LH. LIMITATIONS, REASONS FOR CAUTION: This is a retrospective study; our controls did not represent the general population as they were recruited in an ART centre; we used a modified classification for PCOM using follicle count and/or AMH level with in-house thresholds to define the follicle excess; the AMH assay used is no longer commercially available. WIDER IMPLICATIONS OF THE FINDINGS: Besides biasing the results of AMH assay in FHA patients, the presence of PCOM in FHA patients despite low gonadotropin and androgen levels raises the issue of epigenetically acquired amplification of androgen and/or FSH sensitivity within granulosa cells from polycystic ovaries. In terms of clinical practice, it seems important not to diagnose a low ovarian reserve in FHA patients too quickly on the basis of a decreased AMH level alone. On the contrary, a high AMH level in the context of a menstrual disorder and PCOM should not lead to a misdiagnosis of polycystic ovary syndrome (PCOS) if the basal LH is low.