Aim: Adding polycystic ovary morphology (PCOM) as a criterion for PCOS (Rotterdam criteria) to the NIH criteria of menstrual dysfunction and hyperandrogenism doubles the number of women diagnosed with PCOS, the increase due to the combination of hyperandrogenism and PCOM. We have explored whether hyperandrogenism and PCOM function as independent diagnostic criteria.
Participants: 794 non-healthcare-seeking, euthyroid, normo-prolactinemic, women, aged 18-39 years, not recently pregnant, breast feeding or using systemic hormones.
Measurements: Modified Ferriman-Gallwey scores (mFG), sex steroids measured by LCMS/MS and anti-mullerian hormone (AMH) measured by Beckman Access 2-assay. PCOS was determined using the NIH and Rotterdam criteria for a subset who had a transvaginal ultrasound (TVU).
Results: Serum AMH was independently, positively associated with testosterone and androstenedione, adjusted for age, BMI and smoking (quantile regression β-coefficients 20.90, 95%CI 13.79-28.03; p<0.001and 5.90, 95%CI 3.76-8.03; p<0.001, respectively). For the women who had a TVU, the ovarian follicle count was positively associated with AMH (Spearman correlation coefficient 0.694, p<0.001), and serum testosterone and androstenedione (0.338 p<0.001 and 0.411 p<0.001, respectively). 10.4% of the women who had a TVU had NIH criteria PCOS and 19% had Rotterdam criteria PCOS, the difference due to 12 women with PCOM and hyperandrogenism and 2 with PCOM and menstrual dysfunction.
Conclusions: AMH is a biochemical indicator of the ovarian follicle count. Testosterone directly/indirectly stimulates AMH production during folliculogenesis, hence the positive associations between serum androgens and AMH. Therefore, serum AMH, serum testosterone and the ovarian follicle count all identify the same biological phenomenon (number of developing follicles) and are not independent. Consequently, using PCOM and hyperandrogenemia to diagnose PCOS is effectively diagnosing PCOS using two dependent indices. Our data, together with the published literature, supports consideration of a revised PCOS criteria comprising menstrual dysfunction with either PCOM/elevated AMH or hyperandrogenism, excluding diagnosis based on PCOM plus hyperandrogenism.