Forty-five total knee replacements in forty-one patients who had continued to have progressive osteoarthritis after a proximal tibial osteotomy were evaluated prospectively. There were 51 per cent excellent, 29 per cent good, 4 per cent fair, and 16 per cent poor results after a minimum follow-up of two years. The resection of tibial bone that was produced by the osteotomy could not be related causally to the fair or poor clinical results that were found after arthroplasty. Radiographic study, however, showed that 80 per cent of the knees had patella infera before the arthroplasty, which may contribute to altered biomechanics of the patellofemoral joint of the arthroplasty. The procedure for total knee replacement is made more difficult by the previous osteotomy, and a custom-made prosthesis may be required. The clinical results that were obtained in this series are similar to those for the revision total knee arthroplasties that have been done at this institution, but they were not as satisfactory as those that were obtained after primary total knee replacement. Our results should be considered when a surgeon compares the advantages and disadvantages of proximal tibial osteotomy with those of total knee arthroplasty for an older patient who has unicompartmental osteoarthritis of the knee.