Variability in SHBG assays and the effect thereof on calculated estimates of free testosterone

BackgroundSerum free testosterone is commonly used as a parameter to evaluate testosterone exposure and is mostly calculated using mathematical approximations. As the principal testosterone-binding protein, SHBG concentration is always included in such calculations. However, variability in SHBG measurements may affect reported SHBG levels and consequently free testosterone calculations. In this study, we re-evaluate the effects of SHBG assay choice and interlaboratory variability on calculated free testosterone (cFT).MethodsSerum samples from 113 men and 106 women were collected. SHBG levels were measured using three different SHBG immunoassays (Roche, Abbott and Siemens). Testosterone levels were measured using LC-MS/MS. Afterwards, cFT was calculated using the Vermeulen formula and measured directly. SHBG concentrations, and derived cFT concentrations, from different assays were compared. To simulate interlaboratory SHBG variability, measured levels were modified by 15% after which cFT was recalculated using the Vermeulen, Ly, Sartorius and Södergard formulae. The proportions of diagnoses of hypogonadism or hyperandrogenism were compared.ResultsAssessed SHBG assays showed very good conformity. The largest difference was 7%, between the Abbott and Siemens assay. The difference in cFT levels was at most 3% between the Abbott and Siemens assay. Interlaboratory variability affected the proportion of diagnoses depending on the used formula.ConclusionsOur results do not show large differences between SHBG assays and only minor effects on cFT levels. Therefore, SHBG assay choice is not expected to greatly influence clinical decision making. In contrast, interlaboratory variation in SHBG measurements and choice of formula might considerably affect cFT results and their interpretation.