Men with low testosterone concentrations are usually hypogonadal. However, because variations in the testosterone transport protein, sex hormone-binding globulin (SHBG), directly influence the total testosterone concentration, confirmation of a low testosterone with a measurement of free testosterone or "bioavailable" testosterone (BAT) is recommended. In the present study, we examined the relationship of SHBG with free testosterone (Coat-A-Count assay, Diagnostic Products) and with BAT in men (n = 29) and women (n = 28) who participated in a study of the metabolic determinants of body composition. As expected, total testosterone was strongly positively correlated with SHBG among men (r = 0.68; P <0.01). Although the BAT was independent of SHBG in men (r = 0.02), SHBG was an important predictor of free testosterone (r = 0.62; P <0.01). In contrast, in women serum concentrations of total testosterone (r = -0.26; P = 0.17), free testosterone (r = -0.30; P = 0.17), and BAT (r = -0.46; P = 0.013) all tended to be lower with increasing SHBG. Free testosterone was nearly perfectly positively correlated with total testosterone (r = 0.97) in men, among whom free testosterone represented a relatively constant percentage of the total testosterone (0.5-0.65%), and the percentage of free testosterone was unrelated to SHBG. Thus the Coat-A-Count free testosterone concentration in men, like the total testosterone concentration, is determined in part by plasma SHBG. Accordingly, androgen deficiency may be misclassified with this assay in men with low SHBG. Moreover, the previous findings of reduced free testosterone concentrations with hypertension or hyperinsulinemia or as a risk factor for developing type 2 diabetes, conditions in which SHBG is reduced, may have been methodology-related.