The purpose of this study was to determine whether the use of computer-assisted surgery can improve the clinical results in total knee arthroplasty (TKA) compared with conventional methods of TKA.A literature search of PubMed (1966 to August 2011), CENTRAL (Cochrane Controlled Trials Register; issue 3, 2011), and EMBASE (1984 to August 2011) was conducted. Randomized, controlled trials detecting the clinical outcomes of TKA with or without the use of computer-assisted surgery were identified. A meta-analysis of these clinical trials was then performed. Twenty-one articles were included in the meta-analysis. The results confirmed that operative time was significantly increased with the use of computer-assisted TKA (mean standard difference, 14.68; 95% confidence interval [CI], 11.74 to 17.62; P<.00001], whereas no significant difference existed between the 2 groups regarding the total operative blood loss (mean standard difference, -54.38; 95% CI, -119.76 to 11.00; P=.10). As for other clinical outcomes, including the Knee Society Score (mean standard difference, 4.47; 95% CI, -1.05 to 9.99; P=.36) and range of motion (mean standard difference, 1.38; 95% CI, -1.43 to 4.18; P=.34), the use of computer-assisted TKA did not help to improve function recovery postoperatively.