[Expenses and risk of artificial knee joint: a look backward to a 20-year clinical experience]

Costs and risks of implantation of prosthetic knee joints are analyzed in this retrospective study. From 1974-1993 514 primary and 34 revision arthroplasties were done in this hospital, all by the same surgeon. 98% of the patient protocols were available and analyzed, but no systematic follow-up was attempted. 82% of the patients were female; mean age at operation was 74 years. Joint destruction was caused by osteoarthritis in 75%, aseptic osteonecrosis in 10%, rheumatoid arthritis in 9% and posttraumatic arthritis in 3.5%. 75% of the patients were obese and had a body mass index > 25 kg/m2. Non-constrained unicompartmental type prostheses were used in 66%, the non-constrained multicompartmental type in 10% and the constrained total rotation knee (Engelbrecht) in 24%. Hospital mortality rate was 0.55% due to myocardial infarction and pulmonary embolism. 3 patients died of septic prosthetic joint infections 5, 7 and 71/2 years after surgery. Perioperative morbidity, typical of the age group above 70 years, was mainly due to cardiovascular and thromboembolic events and gastrointestinal bleeding. Early infection during the first postoperative year was encountered in 3 constrained total knees, but none in nonconstrained type. The calculated operative infection rate was 2.4% for the constrained type, zero for the non-constrained type, and 0.5% for the whole series. Late prosthetic infections occurred in 8 patients up to 12 years after surgery. In comparing non-constrained unicondylar and hinged types of joint replacement, the non-constrained sledge prosthesis involves considerably lower costs in terms of duration of surgery, hospital stay, blood loss, price of the implant, infection rate and difficulties of revision arthroplasty. Lower costs and risks favour the smaller unicondylar implant for use in localized degenerative or necrotic destruction, particularly of the medial compartment of the knee. Semi-constrained total condylar systems are used for more extensive degeneration without evident instability. The indication for hinged endoprostheses is restricted to revision arthroplasty and grossly unstable knees.