Purpose: This study aims to evaluate digital and conventional impressions for full-coverage restorations in terms of marginal and internal discrepancies. Study selection: The analysis included in vivo and in vitro studies reporting the marginal or internal gap of full-coverage restorations that provide both the conventional and digital impression. The PubMed, Cochrane Trials, and Scopus databases were searched. The quality of clinical trials was rated using Cochrane Collaboration's tool, and the quality of the evidence was evaluated using the Grading of Recommendations, Assessment, Development, and Evaluation tool. Heterogeneity among the studies was evaluated, and a meta-analysis with subgroup analysis was conducted whenever it was possible. Results: Thirty-three articles (8 prospective clinical trials and 26 in vitro studies) were selected to extract data after applying the predefined selection criteria. The standard mean difference (SMD) of the meta-analysis for marginal adaptation was −0.76 (95% confidence interval: −1.23 to −0.29) and −0.59 (95% confidence interval: −0.93 to −0.24) for in vitro and in vivo studies, respectively, indicating digital impressions provided significantly less marginal gap than conventional impressions in in vitro studies (P =.002). The impression technique did not significantly influence the internal adaptation. Conclusions: Differences in marginal adaptation between the digital and conventional groups are not significant for in vivo studies, but for in vitro studies, the digital impression resulted in better marginal adaptation. Based on the Grading of Recommendations, Assessment, Development, and Evaluation approach for marginal adaptation, clinical studies were classified as high confidence and in vitro studies were graded moderate because of the inconsistency. Furthermore, high-quality studies are needed to confirm our results (the International Prospective Register of Systematic Reviews; CRD42017077925). © 2019 Elsevier Inc.