We reviewed our experience with total knee arthroplasties performed between 1977 and 1992 in patients who were fifty-five years old or less to determine the appropriate management for younger patients who have severe osteoarthrosis. One hundred and fourteen knee replacements were performed in eighty-eight patients who were an average of fifty-one years old (range, twenty-two to fifty-five years old). All of the operations were performed by one of the two senior ones of us (J. N. I. or W. N. S.) with the use of cementing techniques. A posterior stabilized, posterior cruciate-substituting design was used for all but one replacement, for which a semiconstrained total condylar prosthesis was used. Six knees (four patients) were lost to follow-up. Follow-up data for the remaining 108 knees (eighty-four patients) were used to perform the survivorship analysis.One hundred and three unrevised knees (eighty patients) were available for clinical evaluation with the scoring systems of The Hospital for Special Surgery and the Knee Society at an average of eight years (range, three to eighteen years) postoperatively; thirty-six knees were followed for more than ten years. In addition, the activity levels of the patients were assessed with the activity score of Tegner and Lysholm. Radiographs were examined for evidence of loosening of the component.At the latest follow-up examination, the average knee score according to the system of The Hospital for Special Surgery had improved from 55 points preoperatively to 92 points. According to the system of the Knee Society, the average knee score was 94 points and the average functional score was 89 points. The result for all 103 knees was good or excellent according to the knee scores of The Hospital for Special Surgery and the Knee Society. Ninety-seven knees (94 per cent) had good or excellent function according to the functional score of the Knee Society. The average activity score of Tegner and Lysholm improved from 1.3 points (range, 0 to 4 points) preoperatively to 3.5 points (range, 1 to 6 points) at the latest follow-up examination. All but two patients had improvement in the activity score postoperatively, and nineteen (24 per cent) of the eighty patients had an activity score of at least 5 points, indicating regular participation in activities such as tennis, skiing, bicycling, or strenuous farm or construction work. Nine (9 per cent) of the 103 knees had non-progressive tibial radiolucent lines.Two patients had a revision because of late infection, and one patient had revision of a well fixed tibial component because of wear of the polyethylene. In addition, three patellar components were revised for loosening, and one spacer was exchanged to treat instability. With failure defined as revision of either the femoral or the tibial component, the over-all rate of survival was 94 per cent at eighteen years. When the three patellar revisions were included in the failures, the survival rate was 90 per cent at eighteen years. When the exchange of the spacer was also included in the failures, the survival rate was 87 per cent at eighteen years.We consider arthroplasty with cementing of a posterior stabilized total knee prosthesis to be effective operative treatment with durable results for osteoarthrosis in younger patients when other, less invasive measures have failed. Within the average eight-year follow-up interval of this study, polyethylene wear, osteolysis, and loosening of the conforming posterior cruciate-substituting prosthesis were not major problems for these younger, active patients, although it is possible that this observation could change with an even longer duration of follow-up.