Evaluation of assays available to measure free testosterone

This study compared the results of various testosterone assays in a cross-sectional study of 50 male subjects age 28 to 90 years. The purpose of the study was to determine the relationship of the various testosterone assays to one another. In addition, we determined week-to-week variability in testosterone and bioavailable testosterone in 16 subjects. The following assays were undertaken: total testosterone (T), free testosterone by equilibrium dialysis (FTD), bioavailable testosterone (BT), free testosterone by ultracentrifugation (FTU), and direct estimation of serum free T by an analog ligand radioimmunoassay (FTA). In addition, using total T and sex hormone-binding globulin (SHBG), we calculated the free androgen index (FAI = T/SHBG) and the free testosterone index (FTI) using the method of Vermeulen. In a second study, we measured the week-to-week variation in T and BT in a group of older males. Lastly, we demonstrated excellent stability of serum stored at -70degreesC for up to 7 years for T and BT. Correlations of the various assays to increasing age were significant for all assays except total T (r = -.126). The best correlation was found with BT (r = -.744, P <.001). All measures were statistically significantly correlated with FT. The best correlations were FTI (r = .807, P <.007) and BT (r = .670, P <.001). If T was used to classify hypogonadism in comparison to BT, it resulted in misclassification in 42% of cases. In addition, we demonstrated a marked week-to-week variability in T and BT in older men with the T and BT being in the eugonadal range some weeks and hypogonadal on other occasions. This occurred in 8 of 16 men for T and 10 of 16 men for BT. Based on these data, we suggest that the FTI or BT are the most practical methods to determine hypogonadism. There is a need for increased awareness that marked week-to-week variability within a single individual can occur for measurements of both T and BT. Copyright 2002, Elsevier Science (USA). All rights reserved.