Objective: The identification of hyperandrogenism represents the cornerstone for the assessment of polycystic ovary syndrome (PCOs). However, its definition has always been troubling, mostly because of the poor accuracy shown by routine androgens assays. As suggested by literature, the application of more precise steroid measurement methods (such as liquid chromatography tandem mass spectrometry, LC-MS/MS) could improve the diagnostic workup. The aim of our study is to evaluate the impact of the assessment of testosterone (T) and androstenedione (A) by LC/MS-MS using a multisteroid kit in the diagnosis of PCOs and in the assignment to the different clinical phenotypes.
Design: We enrolled 98 consecutive patients (24.2±6.2 years) referred for suspected PCOs (menstrual irregularities, hirsutism, alopecia). 10 patients were excluded because other diagnosis were made. T and A were measured both by routine assay (ECLIA and CLIA, respectively) and LC-MS/MS multisteroid kit in the 88 subjects included. Clinical and biochemical parameters associated with metabolic risk (blood pressure, BMI, waist, glucidic and lipidic metabolism) were recorded. 34 pre-menopausal Caucasian eumenorrhoeic, without clinical hyperandrogenism volunteers served as a control group to derive LC-MS/MS T and A reference ranges.
Results: According to the Rotterdam consensus, based on T measurement by ECLIA PCOS was confirmed in 65/88 subjects: 87.7% were oligoamenorrhoeic, 84.6% had clinical hyperandrogenism, 63% had polycystic ovaries and 54% high AMH levels. Measuring T by LC-MS/MS, PCOs was diagnosed in 67/88 subjects. High T (HT) was found in 43% and 56.7%, while high A (HA) was found in 47.6% and 59.7% PCOs patients respectively by routine assays and LC-MS/MS. Based on LC-MS/MS 19 (28.4%) patients were normoandrogenic, 8 (11.9%) had HT, 10 (14%) HA and 30 (44.8%) HT+HA. Routine assays misclassified 15 patients as normoandrogenic. Hyperandrogenic PCOs patients by LC-MS/MS showed a higher total cholesterol levels than normoandrogenic ones (161.8±32.9 vs 144.8±28.7 respectively, P=0.05), glucidic metabolism and clinical parameters were comparable. No difference was observed between the two groups when hyperandrogenism was identified by routine assays.
Conclusions: LC-MS/MS is more sensitive than routine assays in identifying biochemical hyperandrogenism in PCOs patients. These patients seems to have a slightly worse metabolic profile. Regardless of the measurement method used, A is the most frequently elevated androgen in our PCOs patients cohort.
18 - 21 May 2019
European Society of Endocrinology