Over a ten-year period, thirty-three consecutive revision total elbow arthroplasties were performed at our institution. These were assessed at a minimum of three years after the revision, with an average length of follow-up of sixty-one months. Eighteen (55 per cent) of the elbows had a good result and fifteen (45 per cent) had a poor result. The poor results were due to infection in three elbows, loosening of the prosthesis in six, inadequate motion in two, continued pain in two, and prosthetic failure in two. Additional surgical revision with another implant was done in the fifteen elbows that initially had a poor result. Eventually the result was good in twenty-four elbows. The three elbows that became infected after surgical revision had a resection arthroplasty and all were rated as having a fair result. The data from this study indicated that reimplantation is a viable option for the revision of a failed total elbow arthroplasty, although more than one revision may be required. They also suggested that young patients who have post-traumatic arthritis should not undergo a total joint replacement, and that revision procedures should be performed in settings that can offer several surgical options and by surgeons who have had experience with these options. Alternatives to reimplantation as a revision procedure should be considered in selected patients.