Background: Bioavailable testosterone (BT) is measured [assayed BT (aBT)] or calculated (cBT) in the diagnosis of hypogonadism in men. The cBT depends, however, on the values of the association constants of total testosterone (TT) for sex hormone-binding globulin (SHBG; K(s)) and albumin (K(a)), and its use therefore remains controversial.
Methods: In 503 selected, untreated healthy men, 20-74 years old, we measured TT, dihydrotestosterone (DHT), and androstenediol (5-diol) by GC-MS, SHBG by RIA, and BT after ammonium sulfate precipitation or by calculation according to the law of mass action.
Results: A slight decrease in TT, significant decreases in BT and 5-diol, no variation in DHT, and an increase in SHBG were observed with age. In young males (< or = 39 years), the lower normal limits were between 2.30 and 2.72 nmol/L for aBT and 8.50 nmol/L for TT. For K(s) = 1 x 10(9) L/mol and K(a) = 3.6 x 10(4) L/mol, the lower cBT limit was found to be 2-fold higher than for aBT. With optimized K(s) = 1.9 x 10(9) L/mol and K(a) = 2.45 x 10(4) L/mol, cBT values close to aBT were obtained. When 5-diol was included in the model as a competitive SHBG inhibitor, the correlation between cBT and aBT was better and the cBT:aBT ratios vs 5-diol were less biased.
Conclusion: Lower normal serum aBT concentration in normal men appears to be between 2.30 and 2.72 nmol/L. Much higher serum cBT concentrations are associated with use of different association constants that may be inappropriate. When using the optimized binding constants, taking age-related 5-diol values into consideration slightly improves prediction of cBT.