Implications of Androgen Assay Accuracy in the Phenotyping of Women With Polycystic Ovary Syndrome

Context/Objective: Hyperandrogenism is a common feature of women with polycystic ovary syndrome (PCOS) and is considered a cardinal element for the diagnosis and phenotyping of this condition. Unfortunately, routinely available methods for measuring serum androgens suffer from major limitations. No data are available on the impact of androgen assay quality on the assignment of PCOS women to the different clinical phenotypes of PCOS, when defined according to the Rotterdam criteria for diagnosis. Patients: Two hundred four consecutive Caucasian women with PCOS, diagnosed by the Rotterdam criteria participated in the study. Design: Assessment of total T (TT), free T (FT), and androstenedione (A) by both a chemiluminescent assay, routinely available in our hospital, and gold standard methodology, ie, liquid chromatography-mass spectrometry and equilibrium dialysis, was performed. The results were compared and the associations of these data with clinical and metabolic features of PCOS women were analyzed. Results: By using gold standard assays, TT was high in 36.3% of women, whereas A only marginally contributed to identifying hyperandrogenemic patients. However, gold standard FT measurement was elevated in 70.6% of the PCOS patients, identifying them as hyperandrogenemic. Routine TT and A assays, and the derived calculated FT, were strikingly inaccurate, with substantial overestimation. These assays erroneously classified 60 patients (29.4%), 32 as false hyperandrogenemic, and 28 as false normoandrogenemic, with incorrect assignment of many patients to the clinical phenotypes of PCOS and inappropriate estimation of their metabolic risk. In particular, women misclassified as normoandrogenic had amoresevere metabolic profile than true normoandrogenic subjects. Conclusions: FT alone, as measured by equilibrium dialysis or calculated by using the Vermeulen formula, provided that TT is assayed by gold standard methodology, can be used to identify hyperandrogenemic PCOS subjects. The use of routine androgen assays may misclassify the phenotype of many PCOS women, with errors in the estimation of individual metabolic risk.