Patient-specific instrumentation does not improve accuracy in total knee arthroplasty

Patient-specific instrumentation (PSI) has been introduced as a tool to increase the accuracy of total knee arthroplasty (TKA) compared with conventional instrumentation (CLI). However, previous studies have shown inconsistent results. The authors conducted a meta-analysis to compare the performance of PSI to CLI in TKA. PubMed, EMBASE, and Cochrane Central Register of Controlled Trials electronic databases were systematically searched to identify eligible trials published between 2000 and March 2014. Two reviewers independently assessed methodological quality according to the Cochrane Handbook. Subgroup analyses were performed based on the different study designs (randomized, controlled trial [RCT] vs non-randomized, controlled trial [non-RCT]), preoperative magnetic resonance imaging vs computed tomography, and systems of PSI to explore the source of heterogeneity. Fourteen studies (7 RCTs and 7 non-RCTs) involving 1906 patients were included. There were no statistical differences with respect to the outliers of mechanical axis, coronal femoral component, sagittal femoral component, femoral component rotation, operative time, blood loss, and length of hospital stay between PSI and CLI groups. The number of outliers in coronal tibial components (odds ratio, 2.29; 95% confidence interval, 1.20 to 4.35; P=.01) and sagittal tibial components (odds ratio, 1.67; 95% confidence interval, 1.16 to 2.42; P<.01) was significantly lower in the CLI group than in the PSI group. Based on the numbers available, the use of PSI compared with CLI was not likely to improve the accuracy of component alignment and treatment effects of TKA. Further high-quality RCTs are warranted to confirm the authors' results.