Revision knee arthroplasty in rheumatoid arthritis.

Seventy-six revised total knee arthroplasties (TKAs) performed for mechanical, nonseptic failure of the primary arthroplasty in 62 rheumatoid arthritis patients were investigated. The revisions were performed with McIntosh, Marmor, Attenborough, Guepar, and various types of tricompartmental prostheses. Twenty-one of the 76 revisions subsequently failed. Fifteen were mechanical failures and six were deep infections. Thirteen of the 21 failures were treated with another arthroplasty, six with arthrodesis, and two with antibiotics only. After an average of 60 months, 48 of the surviving arthroplasties were examined clinically and roentgenographically. At follow-up examination, eight were clinical failures. One-half of the McIntosh and Marmor arthroplasties and one-third of the Attenborough and Guepar arthroplasties, altogether 17 cases, showed signs of potential roentgenographic failure. The majority of the revised TKAs classified as roentgenographic failures were clinically successful or acceptable. Revision of TKAs in noninfected rheumatoid patients can be performed with acceptable clinical results but with a significantly higher failure rate than after primary procedures.