abstract
- There are few widely accepted treatments for male subfertility. Controlled trials are necessary to evaluate different subfertility treatments because pregnancies and spontaneous improvements in semen parameters occur without treatment. Properly randomized trials can find out whether pregnancy rates are higher or lower than expected after treatment. Quantitative meta-analysis improves precision where individual trials are not powerful enough to demonstrate moderate treatment effects. Only 174 published randomized treatment trials for male subfertility were found after extensive literature review, both by computerized search and hand searches of 41 journals. In 72 of the studies, pregnancy was an outcome measure and these are analysed and discussed. The quality of most trials is poor. Meta-analysis was possible on seven themes, but the results of this exercise appear to be unduly influenced by trials of poor quality. Few conventional treatments (i.e. those not involving assisted conception techniques) result in improved fertility rates.