Prognostic value of free testosterone and free androgen index in detecting the polycystic ovary syndrome

Objectives: The mainstay of the diagnosis of polycystic ovary syndrome (PCOS) is hyperandrogenemia. Recent observations suggest that total testosterone may not be a sensitive marker of androgen excess. The aim of the present study was to compare the diagnostic value of different androgen determinations in supporting the diagnosis of PCOS. Methods: 80 untreated PCOS patients (mean age 28 years) diagnosed by the NIH-criteria and 112 healthy control women (mean age 27 years) were included in the study. Measurements of total testosterone (T) and sex hormone binding globuline (SHBG) (both by IMMULITE 2000, DPC Biermann, Bad Nauheim, Germany) as well as LH, FSH, androstendione, DHEAS and albumin levels were performed. In addition, the free androgen index (FAI=T[nmol/l]/SHBG[nmol/l] x 100), free (cfT) and bioavailable (cbT) testosterone (www.issam.ch/freetesto.htm) were calculated. Clinical signs of hyperandrogenism were evaluated by physical examination. Area under the receiver operating curve (AUC-ROC) was used to compare the sensitivity and specificity of different androgen determinations. Results: All biochemical parameters of hyperandrogenism were significantly higher in PCOS women than in controls (all p<0.0001). The highest AUC-ROC was found for FAI (0.842) followed by cfT (0.834) and cbT (0.828). Lower AUC–ROCs were found for SHBG, T and androstendione (0.783, 0.774 and 0.742, respectively). When FAI levels of 3.8 were taken as the cut-off value, sensitivity was 75% and specificity 83%. For the cfT (cut-off level 0.834 ng/dl) a sensitivity of 66% and a specificity of 90% was found. FAI and cfT correlated significantly (all p<0.05) with T, androstendione and DHEAS. In addition, FAI and cfT but not T correlated significantly with acne. However, no correlation was found between alopecia or the Ferriman-Gallwey-hirsutism scores and all biochemical parameters of hyperandrogenism. Conclusion: ROC analysis provided evidence that FAI and calculated free testosterone are useful parameters for the discrimination of PCOS women and healthy controls.