Fifty-nine patients with rheumatic destruction of the elbow received 20 St Georg, 20 GSB III, 13 Souter-Strathclyde, and 13 Kudo endoprostheses. Among the various prosthetic categories, 43.9% of the joints had had preceding rheumatoid surgery (a previous synovectomy had been performed in 10 joints at a mean of 4.1 +/- 3.7 years and a resection interposition arthroplasty had been performed in 19 cases 4.2 +/- 1.8 years before endoprosthetic replacement). We examined 51 patients with 54 prostheses after a mean follow-up of 5.7 +/- 4.1 years using the Inglis score and analyzing all radiographs. Complications occurred in 20% of the St Georg prostheses, 25% of the GSB III prostheses, and 23% of the Souter-Strathclyde prostheses. Of the St Georg prostheses, 6 (30%) had to be exchanged, as well as 4 (20%) of the GSB III prostheses and 4 (30.7%) of the Souter-Strathclyde prostheses. Of the primarily implanted joints, the St Georg prostheses measured 77.7 +/- 7.7 on the Inglis score, GSB III 89.6 +/- 7.2, Souter-Strathclyde 88.4 +/- 6.5, and Kudo 89.7 +/- 4.4. Radiolucent lines greater than 1 mm were observed in 26% of the St Georg prostheses, 23% of the GSB III prostheses, 27% of the Souter-Strathclyde prostheses, and 9% of the Kudo prostheses. In contrast to the clinical results, the intraoperative and postoperative complications, as well as the rate of failure and radiolucent lines, showed a statistically significant relationship to previous operations of the joints, especially with the resection interposition arthroplasty. We conclude that resection interposition arthroplasty seems to be associated with complications and failures when a subsequent endoprosthesis is used.
