Objectives To assess the 5‐year survival rates and incidences of complications of cemented and screw‐retained implant reconstructions. Methods An electronic Medline search complemented by manual searching was conducted to identify randomized controlled clinical trials (RCTs), and prospective and retrospective studies giving information on cemented and screw‐retained single‐unit and multiple‐unit implant reconstructions with a mean follow‐up time of at least 1 year. Assessment of the identified studies and data ion were performed independently by three reviewers. Failure rates were analyzed using Poisson regression models to obtain summary estimates and 95% confidence intervals of failure rates and 5‐year survival proportions. Results Fifty‐nine clinical studies were selected from an initial yield of 4511 titles and the data were extracted. For cemented single crowns the estimated 5‐year reconstruction survival was 96.5% (95% confidence interval (CI): 94.8–97.7%), for screw‐retained single crowns it was 89.3% (95% CI: 64.9–97.1%) (P = 0.091 for difference). The 5‐year survival for cemented partial fixed dental prostheses (FDPs) was 96.9% (95% CI: 90.8–99%), similar to the one for screw‐retained partial FDPs with 98% (95% CI: 96.2–99%) (P = 0.47). For cemented full‐arch FDPs the 5‐year survival was 100% (95% CI: 88.9–100%), which was somewhat higher than that for screw‐retained FDPs with 95.8% (95% CI: 91.9–97.9%) (P = 0.54). The estimated 5‐year cumulative incidence of technical complications at cemented single crowns was 11.9% and 24.4% at screw‐retained crowns. At the partial and full‐arch FDPs, in contrast, a trend to less complication at the screw‐retained was found than at the cemented ones (partial FDPs cemented 24.5%, screw‐retained 22.1%; full‐arch FDPs cemented 62.9%, screw‐retained 54.1%). Biological complications like marginal bone loss >2 mm occurred more frequently at cemented crowns (5‐year incidence: 2.8%) than at screw‐retained ones (5‐year incidence: 0%). Conclusion Both types of reconstructions influenced the clinical outcomes in different ways, none of the fixation methods was clearly advantageous over the other. Cemented reconstructions exhibited more serious biological complications (implant loss, bone loss >2 mm), screw‐retained reconstructions exhibited more technical problems. Screw‐retained reconstructions are more easily retrievable than cemented reconstructions and, therefore, technical and eventually biological complications can be treated more easily. For this reason and for their apparently higher biological compatibility, these reconstructions seem to be preferable.