© The Author 2014. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For Permissions, please email: firstname.lastname@example.org.OBJECTIVE: Our objective was to study the use of pain medications for persistent knee pain and their predictors after revision total knee arthroplasty (TKA).METHODS: We examined whether demographic (gender, age) and clinical characteristics [BMI, co-morbidity measured by the Deyo-Charlson index (a 5-point increase), anxiety and depression] predict the use of NSAIDs and narcotic pain medications 2 and 5 years after revision TKA. Multivariable logistic regression adjusted for these predictors as well as operative diagnosis, American Society of Anesthesiologists class and distance from the medical centre.RESULTS: A total of 1533 patients responded to the 2-year questionnaire and 881 responded to the 5-year questionnaire. NSAID use was reported by 13.4% (206/1533) of patients at 2 years and 16.7% (147/881) at 5 years. Narcotic medication use was reported by 5.4% (83/1533) of patients at 2 years and 5.9% (52/881) at 5 years. Significant predictors of the use of NSAIDs for index TKA pain at 2 and 5 years were age >60-70 years [odds ratio (OR) 0.62 (95% CI 0.39, 0.98) and 0.46 (0.25, 0.85)] compared with age ≤60 years and a higher Deyo-Charlson index [OR 0.51 (95% CI 0.28, 0.93)] per 5-point increase at 5-year after revision TKA. Significant predictors of narcotic pain medication use for index TKA pain were age >60-70 years [OR 0.41 (0.21, 0.78)] and >70-80 years [0.40 (95% CI 0.22, 0.73)] at 2 years and depression [OR 4.58 (95% CI 1.58, 13.18)] at 5 years.CONCLUSION: Younger age and depression were risk factors for the use of NSAIDs and narcotic pain medications for index TKA pain at 2- and 5-years after revision TKA.