To investigate the effect of 3D printed patient-specific instrumentation (PSI) on unicompartmental knee arthroplasty (UKA).

Inclusion criteria for literature study included: cohort studies and randomised controlled trials (RCTs), patients receiving UKA for the first time, surgical procedures involving both PSI-assisted UKA and conventional UKA. Exclusion criteria: studies that did not include the required outcome indicators, and studies with high dropout rates at follow-up, etc.. Seven databases including Web of Science, The Cochrane Library, Excerpta Medica Database (Embase), PubMed, China National Knowledge Infrastructure (CNKI), China Online Journals (Wanfang) and China Science and Technology Journal Database (VIP) were retrieved based on above criteria. Newcastle-Ottawa scale (NOS) was applied to evaluate the quality of cohort studies. The improved Jadad scale was adopted to evaluate the quality of RCTs, and meta-analysis was performed using Revman 5.3.

A total of nine works of literature were collected, including five cohort studies and four RCTs, all of which were of high quality. Results of the meta-analysis showed that PSI could reduce the coronal [mean difference (MD)=-1.09, 95% confidence interval (CI) (-1.50, -0.67), P <0.001]and sagittal [MD=-0.73, 95%CI(-1.43, -0.02), P=0.04]deviation of the tibial prosthesis compared with conventional instrumentation, but might increase the position deviation of the coronal plane of the femoral prosthesis [MD=0.89, 95%CI (0.17, 1.60), P=0.01]. The differences were statistically significant. In terms of hip-knee-ankle angle [MD=-0.13, 95%CI (-0.72, 0.46), P=0.66], operation time [MD=2.07, 95%CI (-0.14, 4.29), P=0.07], Hospital for Special Surgery (HSS) score [MD=-0.70, 95%CI (-1.71, 0.31), P=0.17] and the number of outliers in prosthesis positions [MD=0.64, 95%CI (0.38, 1.08), P=0.10], PSI-assisted UKA did not deliver better results than conventional UKA, and the differences were not statistically significant.

Prosthesis localization can benefit from the PSI-assisted UKA, but the overall surgical effect may not show remarkable difference from that of conventional UKA.