Revision of unicompartmental knee arthroplasty: outcome in 1,135 cases from the Swedish Knee Arthroplasty study.

From 1975 through 1995, 45,025 knee arthroplasties were recorded in the prospective Swedish Knee Arthroplasty study. By the end of 1995, 1,135 of 14,772 primary unicompartmental knee arthroplasties (UKA) for localized, mainly medial arthrosis had been revised. The Marmor/Richards and St. Georg sledge/Endo-Link prostheses were used in 65%. Mean age at revision was 72 (71) years. 232 revisions were performed as an exchange UKA (partial in 97) and 750 as a total knee arthroplasty (TKA). 153 were revised by other modes. In medial UKA, the indication for revision was component loosening in 45% and joint degeneration in 25% and in lateral UKA, the corresponding figures were 31% and 35%, respectively. In 94 cases, unicompartmental components were added to the initially untreated compartment, in 14 with partial exchange of a component. The CRRR was estimated using survival statistics. After only 5 years, the risk of having a second revision was more than three times higher for failed UKAs revised to a new UKA (cumulative rerevision rate (CRRR 26%) than for those revised to a TKA (CRRR 7%). This difference remained, even if those revised before 1985, when modern operating technique was introduced, were excluded (CRRR 31% and 5%, respectively). UKA is a safe primary procedure, when performed with well-designed components and modern surgical technique. It gives documented good patient satisfaction, range of motion, pain relief and relatively few serious complications. However, once failed, the knee should be revised to a TKA. This applies to most modes of failure. Not even joint degeneration of the unoperated compartment can be safely treated by adding contralateral components; CRRR after this procedure was 17%, while it was 7% when converted to a TKA.