As the number of primary total knee arthroplasties (TKA) have increased, there has been a corresponding increase in the number of revision TKAs. The purpose of the this study is to report on the follow-up results of 50 revision TKAs, and to assess their clinical outcome. From January 1990 to December 1995, 50 revision TKAs were perfomed in 43 patients, by one surgeon. Mean follow up time was 2.5 years (range, 1 to 6 years). The mean interval from the initial arthroplasty to revision was 4.6 years (range, 3 months-12.6 years). The causes of the revision were wear of polyethylene in 29 joints (58%), infection in seven (14%) and aseptic loosening in six. The revision components were fixed with cement in all cases. Additional fixations were obtained with modular stems (femoral 11 cases, tibial 13 cases), 9 tibial metal wedges, 7 femoral augmentations, and 12 allografts. The Knee Rating Scale of Hospital for Special Surgery was used for clinical assessment. The mean flexion contracture and further flexion of knee joint were 8.8degrees (range, 0-70degrees), 98.5degrees (range, 5-135degrees) preoperatively, and 3.7degrees (range, 0-15degrees), 112degrees (range, 75-135degrees) postoperatively. The Knee Rating Scale improved from 66.5 points (range, 15-83) to 87.5 points (range, 79-98). There were two cases of complications, dislocation of patella and superficial wound problem in each case. In conclusion, wear was the most common cause of failure of primary TKA. The results of revision in infected TKA were less favorable than that of wear and loosening. The prosthetic design and thickness of polyethylene were the most important factors related to wear. The tibial component of PCA and patellar component of Miller-Galante had significant wear problems.