Objective: Bioavailable testosterone (BioT) is measured directly by bioassay or estimated by calculating the free androgen index (FAI) and reflects hyperandrogenaemia more accurately than measurement of total testosterone in plasma. A comparison of the biological variation of FAI and measured BioT was undertaken to determine the relative suitability of these estimates in PCOS.
Method: The biological variation of BioT and FAI was assessed in 12 PCOS patients (median age 28yrs, range 18-31) and 11 controls (median age 30yrs, range 19-33) by measuring total testosterone, BioT and sex hormone binding globulin (SHBG) at four day intervals on 10 consecutive occasions. The FAI was obtained as the quotient 100 T/SHBG. The study was approved by the Local Research Ethics committee.
Results: The PCOS group had higher mean BioT and FAI than the control group (mean plus/minus SD, 1.32 plus/minus 0.42 vs 0.88 plus/minus 0.45 nmol per L, p=0.001, and 16.5 plus/minus 7.5 vs 7.6 plus/minus 5.3, p=0.001, respectively). In both the PCOS and control groups the within subject (CVi) coefficient of variation was significantly lower for the measured BioT than for the calculated FAI (PCOS, 16.4% vs 20.8%, p= 0.041; controls, 8.4% vs 17.6%, p=0.01). In the PCOS and control groups the between subject coefficient of variation (CVg) for BioT was 26.7% and 49.7% and for FAI was 39.8% and 67.6% respectively. The critical difference between two consecutive BioT samples in an individual patient with PCOS was 45%, indicating that at any level of BioT, a subsequent sample must rise or fall by greater than 45% to be considered significantly different from the first.
Conclusion: Both BioT and FAI detect hyperandrogenaemia in PCOS. Both estimates have a high between subject variation but the lower within subject variation in BioT compared to FAI suggests that it is the more reliable test for serial measurement.
24 - 26 Mar 2003
British Endocrine Societies