RESUMO
PURPOSE: The criteria for normal testosterone have been established by expert consensus rather than by evidence. We determined whether a cutoff point for normal could be established using biomarkers. MATERIALS AND METHODS: We performed an exploratory investigation of 1,492 hypogonadal men pooled from 7 registration trials. Serum testosterone, prostate specific antigen and hematocrit were measured at baseline and after 90 days of continuous testosterone replacement therapy. RESULTS: Baseline prostate specific antigen, percent change in prostate specific antigen and hematocrit appeared to be most strongly related to baseline serum testosterone. Subgroup analysis and visual inspection of linear spline fit of these data suggested an approximate serum testosterone cutoff for normal of 300 ng/dl for percent change in hematocrit, and 200 ng/dl for baseline prostate specific antigen and percent change in prostate specific antigen. CONCLUSIONS: This exploratory study revealed considerable variation among individuals and target tissues in individuals. Further study should be performed using standardized assays in a broader population.