Hypogonadism is difficult to confirm with standard assays. A large proportion of testosterone (T) is bound to sex hormone binding globulin (SHBG) and is biologically inert. Accurate measures of free and biological-available-testosterone (Bio-T) are largely unavailable. In this study we compared formulae calculated from the total testosterone (TT) and SHBG in their ability to predict hypogonadism and/or Bio-T.
TT and SHBG were determined by ELISA and Bio-T by Ammonium Sulphate precipitation in 1072 men. An equation to calculate Bio-T from TT and SHBG was modelled on the first 715 men using multiple linear regression and validated on the remainder by correlation (r2). The derived equation was compared using the whole cohort and a subgroup with 7.5TT 12nmol/L to the Free Androgen Index (FAI) and other indices published by Nanjee(TN) and Vermeulen(TV) by area under the receiver-operator-curve (ROC).
The equation derived from our data provided the best approximation of Bio-T in the validation cohort n=357 [(r2=0.73 versus (TT r2=0.68), (FAI r2 =0.26), (TN r2=0.55), (TV r2=0.59) all, p<0.0001]. In the whole population n=1072 the derived equation was the best marker of hypogonadism as defined at our centre (area under ROC=0.94 versus TT=0.93, FAI =0.72, TN=0.91, TV=0.88). However for men with borderline hypogonadism 7.5TT 12nmol/L (n=317) our equation only had equivalent efficacy to other published indices (area under ROC 0.75 versus TT=0.63, FAI=0.74, TN=0.75, TV=0.75).
A single measure of TT is the most useful measurement of androgen status. An index derived from our data proved the best approximation of Bio-T and marker of hypogonadism. In borderline androgen deficiency published indices of free testosterone and our approximation of Bio-T have equivalent efficacy.