Reoperations after 3200 revision TKAs: rates, etiology, and lessons learned.

The purpose of this study was to evaluate the prevalence, etiology, and evolution with time of reoperations done after index revision total knee arthroplasties. After exclusion criteria were met, 1814 index knee revisions in 1627 patients were included in the final cohort with more than 50% of the revisions done for loosening and extensor mechanism problems and more than 80% of the revisions involving the femoral, tibial, or both components. Three hundred seventy-three knees subsequently have been reoperated on one or more times. The average time from index revision total knee arthroplasty to the first reoperation was 3.5 years (range, 1 day-19 years). Of the 1814 index revision total knee arthroplasties, 373 (20%) had 593 reoperations in 336 patients. The cumulative risks of first reoperation at 5, 10, and 15 years were 16.1% (95% CI, 14.2, 17.9), 26% (95% CI, 23.4, 28.6), and 31.4% (95% CI, 30.2, 39), respectively. There was no difference in risk to first reoperation when comparing the decades in which the index revisions were done (1970-1980, 1981-1990, and 1991-2000). There was a trend toward a higher cumulative risk of first time reoperations for deep infection, loosening, and instability in the last decade, but with the numbers available this was not statistically significant. The prevalence of reoperations in this large series of index revision total knee arthroplasties done for aseptic reasons was surprisingly high. Despite substantial improvements during the past 3 decades in component design, surgical technique, and prevention of infection, patients who have a revision total knee arthroplasty are at substantial risk of having one or more subsequent problems that result in a reoperation.