The diagnostic value of plasma free testosterone in non-tumorous and tumorous hyperandrogenism.

The clinical significance of total and free testosterone (T) estimates for the diagnostic approach to hirsute patients was assessed. Plasma T was measured by radioimmunoassay and its non-protein-bound fraction was determined by equilibrium dialysis, thus facilitating the calculation of apparent free T (AFT). The cases of 162 subjects were investigated; the subjects included 75 women in whom glandular androgen release had been defined by selective catheterization. A positive linear correlation was observed between both parameters over a wide range of concentrations (T, 153 to 10,700 pg/ml; AFT, 0.8 to 342 pg/ml; P less than 0.001). Significant differences of mean T and AFT levels were found between healthy control subjects (n = 8) and subjects with non-tumorous hyperandrogenism (n = 60; P less than 0.005). Individual values overlapped considerably; elevated T (greater than 640 pg/ml; 48%) or AFT (greater than 7.2 pg/ml; 52%) were present in only half the hirsute women. However, the upper 95% confidence limits of normal for both indices were exceeded in all patients with androgen-secreting ovarian tumors (n = 7). It is concluded that the indirect estimation of AFT in addition to T is time-consuming, costly, without practical value in selecting the proper treatment, and therefore not mandatory in the routine evaluation of androgenized women.