Testosterone supplementation for erectile dysfunction: Results of a meta-analysis

Purpose: To our knowledge a causal relationship between altered levels of a ndrogens and erectile dysfunction has not yet been established. We reviewed the literature to assess the usefulness of androgen replacement for erecti le dysfunction. Materials and Methods: Meta-analysis was chosen as the method of evaluating the literature. Study inclusion criteria were testosterone given as the on ly therapy for erectile dysfunction and a clearly stated definition of resp onse for evaluating treatment success or failure. Results: We evaluated 73 articles obtained by a MEDLINE search of 1966 to 1 998 and included 16 in our study. The overall response rate was 57%. In the 9 series with response rate by etiology patients with primary versus secon dary testicular failure had a response rate of 64% versus 44% (p (0.001). I ntramuscular and oral methods of delivery were equivalent with a response r ate of 51.3% and 53,2%, respectively. However, the response to transdermal therapy was significantly different from that of intramuscular and oral tre atment (80.9% versus 51.3% and 53,2%, respectively, p (0.001). The mean con fidence level response for testosterone treatment was 16.7% in the placebo and 65.4% in the treated group (p (0.0001). Conclusions: Our meta-analysis of the usefulness of androgen replacement th erapy for erectile dysfunction indicates that the response rate for a prima ry etiology was improved over that for a secondary etiology, transdermal te stosterone therapy was more effective than intramuscular or oral treatment, and intramuscular and oral treatments were equivalent. In addition, there was a statistically significant difference in favor of testosterone over pl acebo, implying a role for supplementation in select groups.