The Swedish Knee Arthroplasty Register. Validity and Outcome

Abstract: The Swedish Knee Arthroplasty Register (SKAR) was initiated in 1975 to prospectively follow patients operated on with knee arthroplasty. With revisions utilized as the end-point in survival analyses, it is of importance that all revisions of included patients are reported. To validate if revision status was correctly recorded, a postal survey was made among all living patients, registered as operated during 1975-1995. 1.7% of the responding knees were found to have been revised but not reported to the register, or about one fifth of all revisions. We further investigated if the missing revisions could have been localized by comparison of the SKAR and a database run by the health authorities (PAS), and found that 84% would have been revealed by this method. The SKAR was subsequently uptated with respect to the missing revisions. The survey also included a short inquiry regarding patient satisfaction. 81% were satisfied or very satisfied, 11% uncertain and 8% unsatisfied. The satisfaction rate was relatively constant regardless of the time passed since the primary operation. In revised cases 22% of patients were unsatisfied. To evaluate the economical effect of using a total knee (TKA) instead of a unicompartmental knee (UKA), for uni- compartmental arthrosis, we compared the length of hospital stay, cost of implants and rate of revision. The TKA was more expensive and had a lower rate of revisions. However, the UKA saved more money than the cost generated by the increased number of revisions and it had a lower rate of serious complications. Investigating if the volume of surgery affected the risk of revision after UKA, we found that performing few operations was associated with an increased risk. By combining data from the national census register and the SKAR we determined the age specific incidence of knee arthroplasty. We found that the past changes in the age profile of the population only explained a small part of the observed increase in surgery while future changes could be expected to increase the demand for knee arthroplasty by at least 1%/ year. While 5 years have passed since the last major report from the SKAR regarding general outcome, the number of operations has almost doubled. With the register validated and updated with respect to revisions, we report on general demographic and epidemiological outcome with special emphasis on for the most recent 10-year period. The update of the register had led to a general increase in revision rate but otherwise the conclusions did not differ from those drawn from the register in recent years, the results after knee arthroplasty being affected by factors related to patients, material, methods and time of operation.