Introduction: This study was conducted to assess variability in total testosterone levels (TT) in healthy, non-obese aging men who have androgen deficiency symptoms.
Methods: There were 96 men ≥55 years of age in the study. They were selected because they had symptoms consistent with androgen deficiency. Their morning TT levels were measured 4 times over 12 months.
Results: At each time point, geometric mean (95% confidence interval) of TT levels were: 14.5 (13.4-15.7), 14.6 (13.5-15.8), 15.5 (14.4-16.8), and 15.0 (13.9-16.2) nmol/L.
Sixty percent of the men differed by ≤1 nmol/L between sample 1 and 2. The maximum intra-individual TT difference between the 4 samples (interquartile range: 2.9-6.4) was 4.1 nmol/L. If an average of 2 baseline TT levels is used, then the maximum intra-individual difference is 2.3 nmol/L (interquartile range: 1.0-3.8).
Twenty-five men had a TT level < 10 nmol/L on at least one occasion during the study, but only 1 man had <10 nmol/L at all subsequent points; 8 of 9 men with an initial TT < 10 nmol/L had subsequent levels > 10 nmol/L. Only 2 of 35 men with an initial TT > 15 nmol/L had a subsequent value <10 nmol/L.
Conclusions: For healthy, non-obese aging men with androgen deficiency symptoms, a single TT level reliably predicts repeat measures taken over a 12-month period. However, intra-individual variability is sufficient to recommend a repeat sampling after an initially low TT level; this is particularly important since diagnosing androgen deficiency is dependent on the serum TT. Averaging 2 baseline TT levels can reduce the intra-individual variability.
Commentary by Glenn R. Cunningham MD