Abstract
PURPOSE: Various assistive techniques, such as conventional cutting instruments (CON), computer-assisted navigation systems (CAS), patient-specific instruments (PSI) and robot-assisted systems (RAS), have been developed and applied in primary total knee arthroplasty (TKA). In this study, we aimed to assess the relative accuracy and efficacy of several assistive techniques for TKA through a network meta-analysis (NMA) based on multiple published randomized controlled trials (RCTs). METHODS: The PubMed, EMBASE and Cochrane databases were searched for RCTs to conduct this NMA from inception to 1 January 2024. We combined direct and indirect comparisons using a Bayesian NMA framework to assess and compare the effects of different assistive techniques on radiological and clinical outcomes. An NMA was conducted, and the study protocol was published online at PROSPERO (CRD42023402882). RESULTS: One hundred and twelve RCTs involving 14,968 TKAs with four different assistive techniques (CON, CAS, PSI and RAS) were evaluated. Inconsistency and heterogeneity were acceptable for most outcomes. Based on the surface under the cumulative ranking curve, RAS could be the best technique for accurate mechanical axis alignment and component position, followed by CAS, PSI and CON. We observed no difference in clinical outcome scores. Additionally, CAS was the best intervention for visual analogue scale scores, and PSI had the shortest operative time. No significant differences were observed in postoperative complications, range of motion or total blood loss. CONCLUSION: RAS was most likely to achieve an accurate alignment, followed by CAS, PSI and CON. No differences were observed in clinical outcome scores and postoperative complications among the four assistive techniques. LEVEL OF EVIDENCE: Level I (systematic review of Level-I randomized controlled studies).