Testosterone Binding and Free Plasma Androgen Concentrations under Physiological Conditions: Characterization by Flow Dialysis Technique*

Testosterone binding to plasma has been analyzed under physiological conditions. Flow dialysis technique (FDT) has permitted precise studies of whole 37 C plasma with negligible disturbance of the equilibrium state. The information obtained by FDT was compared with that obtained by a charcoal adsorption assay (CA), which primarily assesses testosterone binding totestosterone-estradiol-binding globulin (TEBG). The results of the two methods correlated closely(r = 0.98). Consequently, the percentage of testosterone free in plasma under in vivo conditions could be readily calculated from data generated by the much easier CA technique. Plasma free androgenconcentrations were then computed for a large variety of sera. For example, the mean (±SDplasma free testosterone concentrations of normal men, hirsute women, normal women, and pregnant women were found to be 128 ± 57, 16.1 ± 11.7, 5.4 ± 2.7, and 4.6 ± 1.7 pg⁄ml, respectively. Theselevels are in close agreement with previous estimates. The plasma free testosterone level of women on estrogen-containing oral contraceptives was normal (6.6 ±0.4 pg⁄ml) in the face of frequently elevated plasma total androgen levels. The FDT data are shown by mathematical modeling and competition studies to fit a general modelfor the binding of the TEBG ligands in plasma (plasma 17β-hydroxysteroids) to TEBG and albumin. This model has permitted the calculation of the association constants of these testosterone-bindingspecies under physiological conditions. This in turn has made it possible to define the relative role of TEKG and albumin as determinants of the percentage of testosterone free in plasmaand to compute TEBG capacity directly from the total plasma 17β3-hydroxyste-roid level andthe percentage of free 17β8-hydroxysteroid obtained by CA. TEBG plays the central role in plasma testosterone binding because of its high association constant (8.4 X 108M-1) and the widevariation in itsplasma level among individuals. Changes in serum albumin concentration within thenormal range exert little influence on the percentage of free testosterone at any level of TEBG. Hypoalbuminemiadoes result in a significant increase in the percentage of free testosterone at any low TEBG titer, however. Means of correcting CA data for this effect are presented.