OBJECTIVE: To examine racial differences in surgical complications, mortality, and revision rates after total knee arthroplasty. METHODS: We studied patients undergoing primary total knee arthroplasty using 2001-2007 Pennsylvania Health Care Cost Containment Council data. We conducted bivariate analyses to assess the risk of complications such as myocardial infarction, venous thromboembolism, wound infections, and failure of prosthesis, as well as 30-day and 1-year overall mortality after elective total knee arthroplasty, between racial groups. We estimated Kaplan-Meier 1- and 5-year surgical revision rates, and fit multivariable Cox proportional hazards models to compare surgical revision by race, incorporating 5 years of followup. We adjusted for patient age, sex, length of hospital stay, surgical risk of death, type of health insurance, hospital surgical volume, and hospital teaching status. RESULTS: In unadjusted analyses, there were no significant differences by racial group for either overall 30-day or in-hospital complication rates, or 30-day and 1-year mortality rates. Adjusted Cox models incorporating 5 years of followup showed an increased risk of revisions for African American patients (hazard ratio [HR] 1.39, 95% confidence interval [95% CI] 1.08-1.80), younger patients (HR 2.30, 95% CI 1.96-2.69), and lower risk for female patients (HR 0.81, 95% CI 0.71-0.92). CONCLUSION: In this sample of patients who underwent knee arthroplasty, we found no significant racial differences in major complication rates or mortality. However, African American patients, younger patients, and male patients all had significantly higher rates of revision based on 5 years of followup.