BACKGROUND Open tension-free methods of groin hernia repair have been widely adopted despite little rigorous evaluation. METHODS Information was assimilated from all randomized or quasi-randomized trials comparing open mesh with open non-mesh methods to assess benefits and safety. Electronic databases were searched and members of the EU Hernia Trialists Collaboration consulted to identify trials. Prespecified data items were extracted from reports, and quantitative or, if not possible, qualitative meta-analysis was performed. RESULTS Fifteen eligible trials, which included 4005 participants, were identified. There were similar numbers of complications in each group, with few data to address short-term pain and length of stay in hospital. Return to usual activities was quicker in the mesh group for seven of ten trials (P not significant). There were fewer reported recurrences in the mesh groups: overall 21 (1.4 per cent) of 1513 versus 72 (4.4 per cent) of 1634 (odds ratio 0.39 (95 per cent confidence interval 0.25-0. 59); P < 0.001). CONCLUSION Although the rigorous search maximized trial identification, formal meta-analysis was limited by the variation in trial reporting. Within the data available, mesh repair was associated with fewer recurrences.