Objective: The measurement of serum testosterone in women is challenging due to lack of trueness, precision, and sensitivity of various available testosterone assays. Accurate assessment of testosterone in women is crucial especially in conditions associated with alleged over-or under-production of testosterone, such as in polycystic ovary syndrome (PCOS) or primary ovarian insufficiency (POI). The aim of this study was to measure and compare androgen concentrations in women with PCOS, POI, and female controls and to evaluate the performance of extraction RIA and liquid chromatographytandem mass spectrometry (LC-MS/MS) in these women. Design: Cross-sectional study. Methods: Carefully phenotyped women with POI (nZ208) or PCOS (nZ200) and 45 healthy, regularly cyclic female controls were included. Method comparison analyses were performed for total testosterone, androstenedione (AD), and DHEA, as measured by LC-MS/MS and extraction RIA. Results: All androgen levels were significantly elevated in women with PCOS compared with POI patients (P!0.05) and controls (P!0.05). Women with POI presented with similar androgen concentrations as controls, except for AD. Compared with measurements by extraction RIA, testosterone, DHEA, and AD concentrations measured by LC-MS/MS were systematically lower. However, using extraction RIA and LC-MS/MS, testosterone, DHEA, and AD measurements were shown to have good agreement as assessed by Bland-Altman analysis and intraclass correlation coefficient: 0.95 (95% confidence interval 0. 94-0.91), 0.83 (0.79-0.86), and 0.96 (0.95-0.97) respectively. Conclusions: LC-MS/MS, compared with a labor-intensive extraction RIA, shows good precision, sensitivity, and high accuracy for measuring female testosterone, DHEA, and AD concentrations under various clinical conditions. LC-MS/MS, therefore, represents a convenient and reliable assay for both clinical and research purposes, where androgen measurement in women is required.
European Journal of Endocrinology 165 925-933