From individual randomized studies it is unclear whether the patella should be replaced during total knee replacement. We did a meta-analysis to provide quantitative data to compare patellar resurfacing with nonresurfacing during total knee arthroplasty. Only randomized, controlled trials reported between January 1966 and August 2003 comparing patellar replacement with patella retention were included for a total of 12 studies. Two reviewers assessed trial quality and extracted data from papers. The outcomes identified were reoperations for patellar problems, anterior knee pain, knee scores, stair climbing, and patient satisfaction. The resurfaced patella performed better, and we found an increased relative risk (defined by the ratio of the risk of the event in the resurfaced group on the risk of the event in the nonresurfaced group) for reoperation, for significant anterior knee pain, and for significant pain during stair climbing when the patella was left unresurfaced. No differences were observed between the two groups for International Knee Society function score, Hospital for Special Surgery score, and for patient satisfaction. Despite these general findings, forming a definitive conclusion is difficult because many confounding factors, such as component design, surgeon experience, and technical aspects of the surgery, might influence the result in a patient.