The excellent study of Burn and colleagues compared the outcome of matched unicompartmental (33,000) and total (250,000) knee replacements based on routinely collected data from four U.S. and one U.K. health care databases. Although the findings from each database were different, the conclusions from the meta-analysis of all five databases are what would be expected: Following UKR, there was evidence of better clinical outcomes with a 19% (CI 10% to 27%) decrease in use of opioids, which was a proxy for persistent pain, and a lower complication rate with a 38% (CI 5% to 64%) decrease in venous thromboembolism. However, the revision rate of UKR at 5 years was 64% (CI 40% to 94%) — higher than that of TKR.
The authors compared their results to the recently published randomized study (TOPKAT) comparing UKR and TKR, and found their results were broadly similar, except that in TOPKAT the revision rates of UKR and TKR were the same.
An important weakness of the health care database study and other similar studies, acknowledged in the paper, is that the precise indications for the procedure are not known, so these cannot be matched for and may have a profound influence on the outcome. For example, many surgeons tend to use UKR for patients with early arthritis, who they feel will not do well with TKR. These patients often do not do well following UKR and have a high revision rate. Had the study been able to compare UKR and TKR done for bone-on-bone medial compartment arthritis, the primary indication for UKR, which is satisfied in about half the cases having knee replacement, they might have found the revision rate of UKR and TKR were similar as in TOPKAT.
David W. Murray, MD, FRCS
Professor of orthopedic surgery,
University of Oxford
Consultant orthopedic surgeon
Nuffield Orthopaedic Centre
Disclosures: Murray reports he receives royalties related to partial knee replacement.