Calculation of bioavailable and free testosterone in men: A comparison of 5 published algorithms

Background: Estimation of serum concentrations of free testosterone (FT) and bioavailable testosterone (bioT) by calculation is an inexpensive and uncomplicated method. We compared results obtained with 5 different algorithms. Methods: We used 5 different published algorithms [described by Sodergard et al. (bioT(S) and FTS), Vermeulen et al. (bioT(V) and FTV), Emadi-Konjin et al. (bioT(E)), Morris et al. (bioT(M)), and Ly et al. (FTL)] to estimate bioT and FT concentrations in samples obtained from 399 independently living men (ages 40-80 years) participating in a cross-sectional, single-center study. Results: Mean bioT was highest for bioT(S) (10.4 nmol/L) and lowest for bioT(E) (3.87 nmol/L). Mean FT was highest for FTS (0.41 nmol/L), followed by FTV (0.35 nmol/L), and FTL (0.29 nmol/L). For bioT concentrations, the Pearson correlation coefficient was highest for the association between bioT(S) and bioT(V) (r = 0.98) and lowest between bioT(M) and bioT(E) (r = 0.66). FTL was significantly associated with both FTS (r = 0.96) and FTV (r = 0.88). The Pearson correlation coefficient for the association between FTL and bioT(M) almost reached 1.0. Bland-Altman analysis showed large differences between the results of different algorithms. BioT(M), bioT(E), bioT(V), and FTL were all significantly associated with sex hormone binding globulin (SHBG) concentrations. Conclusion: Algorithms to calculate FT and bioT must be revalidated in the local setting, otherwise over- or underestimation of FT and bioT concentrations can occur. Additionally, confounding of the results by SHBG concentrations may be introduced. (c) 2006 American Association for Clinical Chemistry