Failed total knee arthroplasty. Revision and arthrodesis for infection and noninfectious complications

Forty-eight selected cases of failed total knee arthroplasty were followed up for a minimum of two years. Thirty-seven failed because of sepsis; 11 had multiple operations for loosening, instability, fracture, or pain. Management of infection by parenteral antibiotic therapy alone produced one satisfactory result in seven cases. Treatment by antibiotic therapy and tissue debridement alone resulted in retention of only three of 27 prostheses. In none of these cases, including the three salvaged prostheses, was the final result satisfactory. Prosthesis removal, antibiotic therapy, and later insertion of a new prosthesis gave satisfactory results in three of three cases. In 16 of 16 formal arthrodesis procedures with strict attention to surgical technique and fixation, all infected cases fused without further drainage. Temporizing measures in the face of established infection should be expected to fail. The organisms must be cultured from joint aspirations. Appropriate patenteral antibiotics must be used for a period of at least four to six weeks. Abscesses must be evacuated. The prosthesis and all cement must be removed. Devitalized bone adjacent to the cement boundary must be excised. Arthrodesis should be obtained by bone surfaces in tight contact, even at the expense of shortening the extremity. Rigid immobilization with a four-plane external fixation using the Hoffman device is probably the single most important factor in a high arthrodesis success rate. Occasionally, a sliding bone graft can expedite fusion and avoid delayed union. In this series, the success rate of arthrodesis following infection was comparable with the fusion rate in nonseptic cases.