RESUMO
PURPOSE: To demonstrate the differences between outcomes and complications after reverse total shoulder arthroplasty (rTSA), according to the types of lateralized prosthesis designs: glenoid-based lateralization (LG) and humerus-based lateralization (LH). METHODS: PubMed, Embase, and the Cochrane Library databases were systematically searched for studies published before 1 April 2021, using the PRISMA guidelines. A network meta-analysis was applied for indirect comparison, incorporating studies including medialized or conventional Grammont prosthesis and each lateralized prosthesis (LG vs control and LH vs control). Out of 1,989 screened studies, 11 studies were included to compare functional scores, range of motion (ROM), radiologic outcomes, and revision rates. In addition, six articles from the included studies, which had repaired subscapularis after rTSA were analyzed to exclude the potential influence of subscapularis repair on the outcomes. The data were pooled using a random-effects model. The pooled estimates of the mean differences (MDs) and 95% confidence intervals (CIs) were calculated for continuous data, while dichotomous data were analyzed using the pooled relative risk (RR) and their 95% CIs. RESULTS: The ROM, complications, and functional scores were similar between the two groups. In subgroup analysis of 6 studies involving concomitant subscapularis repair, the LH group showed higher American Shoulder and Elbow Surgeons (ASES) scores and Constant scores than the control group. Regarding the ROM, LH group showed better forward elevation than the LG group (LH vs LG: MD 10.07, 95% CI -9.05-29.19). CONCLUSION: Overall, the outcomes and occurrence of complications were not significantly different between the two lateralized prosthesis groups. However, when the subscapularis was repaired, LH prosthesis seems to be more suitable to obtain a better ASES score and ROM. LEVEL OF EVIDENCE: Level III, network meta-analysis.