OBJECTIVE:To compare the effects of total knee arthroplasty performed with and without patellar resurfacing. METHODS:All randomized clinical trials relevant to comparison of total knee arthroplasty performed with and without patellar resurfacing between 1966 and 2008 were collected. The quality of the trials was assessed. The outcomes data including the prevalence of reoperations,postoperative anterior knee pain,patient satisfaction,and the improvement in various knee scores were extracted and reviewed systematically. RESULTS:1 731 knees were included in the 15 published trials,including 842 knees in the resurfaced group and 889 in the nonresurfaced group. During the four years following surgery,there was no significant difference between the resurfaced and the nonresurfaced groups in re-operation rate (relative risk=0.56,95% CI 0.17-1.85); whereas the re-operation rate of the nonresurfaced group was significantly greater than resurfaced group 4 years after surgery (relative risk=0.46,95% CI 0.25-0.84). On the prevalence of postoperative anterior knee pain,there was no significant difference between two groups (relative risk=0.65,95% CI 0.35-1.21). No significant difference was found in patient satisfaction (relative risk=1.00,95% CI 0.95-1.05). There was no significant difference between the resurfaced and the nonresurfaced groups in the KSS scores (weighing mean difference:-0.41; 95% CI:-3.08 to 2.26). CONCLUSION:The long-term rate of reoperations in total knee arthroplasty performed with patellar resurfacing was lower than that without patellar resurfacing. There were no significant differences in the terms of prevalence of postoperative anterior knee pain,patient satisfaction,and the improvement in various knee scores between the resurfaced and the nonresurfaced groups. The results of this systematic review should be explained prudently because of some limitations of included trials. For more reliable conclusions,more trials with high quality on methodology are needed.