Calculated free testosterone and radioimmunoassay free testosterone as a predictor of subnormal levels of total testosterone.

Purpose: To evaluate the relationship among aFT, cFT, and total testosterone (TT) and the best method in diagnosing subnormal levels of TT.

Methods: A total of 213 men were analyzed. Fasting blood samples were drawn for the determination of the lipid profile as well as of plasmatic glucose and serum levels of albumin, TT, aFT, and sex hormone-binding globulin (SHBG). The values of cFT were determined by Vermeulen's formula.

Results: No correlation between aFT and cFT was observed (r = 0.062; P = 0.368), except after controlling for confounders (r = 0.188; P = 0.007). Only 44.8% of hypogonadal men (TT ≤ 300 ng/dL) were classified by aFT, whereas 72.4% of hypogonadal men were classified by both TT and cFT. Sensitivity, specificity, positive and negative predictive values, and positive likelihood ratio were greater in cFT when compared with aFT.

Conclusions: Our results suggest that cFT is more accurate in diagnosing subnormal levels of TT. Furthermore, we do not recommend using aFT due to its lack of accuracy. Further studies should be performed in order to evaluate the correlation between aFT and cFT with clinical signs and symptoms of androgen deficiency.