A study was undertaken to assess the clinical results of revision total knee arthroplasty in which an unresurfaced bony shell was left after removing a patellar component versus those in which a patellar component was implanted. Followup was obtained in 123 of 130 consecutive revision total knee replacements (94%) from three centers. In 21 knees a shell of patellar bone was left and 92 knees had a patellar component in place. Ten patients had a patellectomy and were excluded from consideration. The group with the bony shell had a lower postoperative knee score but the preoperative Knee Society clinical score was significantly lower as well in this group of patients. Compared with the group of patients with the patellar component in place, the group of patients with knees left with a bony shell had a significantly higher percentage of patients who had difficulty using stairs, a higher percentage of patients who were not satisfied with their surgery, and a higher percentage of patients who rated their surgery as unsuccessful in returning them to normal daily activities. When a patellar component was not able to be implanted in revision total knee arthroplasty, a lower quality result was observed.