LAUSR

The meta-analysis by Corona et al [1] has comprehensively analyzed the impact of testosterone therapy (TTh) in men with sexual dysfunction, relying on the International Index of Erectile Function (IIEF) questionnaire as an effective outcome indicator. After thoroughly analyzing data from 2298 men, the authors concluded that TTh is associated with sexual function improvement (particularly erectile function) in hypogonadal men. Although testosterone (T) deficiency is known to possibly have a detrimental impact on sexual function [2], the authors should be complimented for applying a rigorous statistical analysis capable of better clarifying this complex relationship, which is still the source of an infinite critical disagreement today. Respectfully, some aspects deserve further discussion. First, the original sin of T trials is to evaluate questionnaire-based improvement of erectile function (EF) on a continuous scale ignoring the clinical meaning of this improvement. Every study included presents a generic estimate of EF improvement in men with wide-ranging levels of erectile dysfunction (ED) severity (some of them actually potent according to the Cappelleri’s criteria) [3] and comorbid conditions other than low T levels. The authors tried to overcome this inherent fault by interpreting the pooled mean improvement of EF according to the minimal clinically important differences [3], thus concluding that the observed IIEF improvement is clinically significant only in men with mild ED. However, the authors report that the improvement in overall EF component was unaffected by mean basal IIEF at meta-regression analysis. Since basal mean IIEF values varied consistently among the included studies, we argue that a stratification approach could have been more appropriate in order to avoid this apparent contradiction (ie, reporting subgroup estimates according to the baseline ED severity rather than relying on