Endocrine Abstracts

Plasma total testosterone (TT) does not provide reliable information about the biological active testosterone, since a large proportion of this hormone is bound to either specific (sex hormone binding globulin, SHBG) or non-specific binding proteins (albumin). Free testosterone (fT) represents not more than 0.7–2.2% of TT while bioavailable testosterone (BA-T) in females is about 15–48% of the TT. The aim was to investigate the analytical and clinical correlations among measured TT and calculated free testosterone index (fTi), BA-T and fT levels.

Methods: Sera from 61 women (age: 26±10 years) including 12 healthy controls and 49 patients who had either high (n=26, HRP) or low risk (n=23, LRP) for true hyperandrogenism on the basis of clinical and ultrasound findings were analysed. Samples were examined for TT and SHBG using electrochemiluminescence immunoassays (Roche) and for albumin using colorimetric method, then the fT, fTi and BA-T concentrations were calculated.

Results: All sera had normal albumin levels (mean±S.E.M., 46±3 g/l). There were significant (P<0.01) but weak correlations between TT and fTi (r=0.68), between TT and BA-T (r=0,79) and between TT and fT (r=0.79), while correlations between fTi and BA-T (r=0.94) and between fTi and fT (r=0.89) were stronger. All T fractions were the highest in HRP (P<0.001) compared to the control and LRP groups. For detecting hormonal hyperandrogenism, the sensitivities and specificities of testosterone fractions were as follows: fTi, 54 and 85%; BA-T, 54 and 88%; TT, 50 and 76%; fT, 46 and 74%, respectively.

Conclusions: In patients with normal albumin levels there are only marginal differences among sensitivities of fTi, BA-T, TT and fT. Our data confirm that when TT is out of the reference range the SHBG and albumin concentration should be measured to calculate the fTi or BA-T because of their highest specificity.