Convincing Evidence to Rethink Indications for Unicompartmental Knee Arthroplasty

Early failures of unicompartmental knee arthroplasty (UKA) were observed after introduction of the technique in the 1970s, leading Kozinn and Scott1 to consolidate their recommendations for successful UKA in 1989. These have emerged as our most routinely accepted traditional indications for patient selection and include an age of <60 years, a weight of <180 lb (81.6 kg), avoidance of strenuous labor, minimal baseline pain, a preoperative arc of motion of ≥90° with <5° of flexion contracture, and angular deformity of <15°; contraindications include patellofemoral or contralateral condylar degeneration, inflammatory arthropathy, chondrocalcinosis, and anterior cruciate ligament (ACL) insufficiency1,2. Current evidence on contemporary UKA designs has challenged and expanded on traditional criteria for patient selection such that today UKA is estimated to be performed in 8% to 12% of all knee arthroplasties, with 90% of these being performed in the medial compartment3.

Ekhtiari et al. add to our growing body of evidence in support of the use of UKA and are to be commended on their recent large and unfunded report “Unicompartmental Knee Arthroplasty: Survivorship and Risk Factors for Revision. A Population-Based Cohort Study with Minimum 10-Year Follow-up.” The authors demonstrate outcomes rivaling survivorship seen with total knee arthroplasty (TKA) and identify risk factors that should be considered when selecting a patient for UKA today. By using administrative records of a single-payer health-care system (Ontario, Canada), they explored patient factors including age, sex, diabetes, income quintile, and rurality. They also explored available surgical data—i.e., whether or not the UKA was performed with cement.

A notable strength of this study is that it is very well powered, with a conservative estimate that 1,537 patients were needed to detect a 50% prevalence of revision with a 2% margin of error and 95% confidence interval. Between 2002 and 2006, the authors followed 4,385 patients with a minimum 10-year follow-up—remarkably with <0.1% of data missing. Although not quite as well powered as large national registry studies, this study has a similar impact with such robust data collection and so large a sample size.

The vast majority of patients evaluated (98%) had primary osteoarthritis (OA) of the knee, with only 16 UKAs performed for osteonecrosis and none for inflammatory arthritis. Seven hundred and seventy-nine patients underwent all-cause revision, with the cumulative risk of all-cause revision being 16.5% at 10 years accounting for patients censored at the time of final follow-up. The predominant reason for UKA revision was “mechanical loosening,” accounting for 650 (83.4%) of the 779 failures. The second most common cause of revision was periprosthetic joint infection (PJI), accounting for 129 of the 779 failures. Perhaps the greatest study weakness, which the authors readily acknowledge, is that “mechanical loosening” was likely an all-encompassing designation as they were unable to distinguish between the use of mobile and fixed-bearing implants or to know whether OA developed in other compartments. It has been established that UKA may commonly fail due to progression of OA or mobile-bearing dislocation as well as to implant loosening and infection2. In the current study, these possible (or even more likely) known causes for implant revision were not captured, a limitation inherent to the databases analyzed.

The risk of revision within 10 years following UKA was higher in male patients (hazard ratio [HR]: 1.38, p < 0.001) and diabetic patients (HR: 1.49, p < 0.001). Surprisingly, stratifying by whether or not the patient had diabetes revealed no difference in the risk of revision due to PJI. However, an age older than 50 years (HR: 0.70, p = 0.008 for patients 50 to 59 years old and HR: 0.36, p < 0.001 for patients 70 to 79 years old), the use of cemented implants (HR: 0.69, p = 0.001), and more recent surgery (HR: 0.91, p < 0.001) were all protective against revision surgery following primary UKA. The median time to revision was 4.05 years, with Kaplan-Meier survivorship estimates of 97.2% at 1 year, 90.5% at 5 years, 83.5% at 10 years, and 81.9% at 15 years. The authors freely recognize limitations inherent to their administrative database study including informative censoring, which may limit any survival analysis. Additional potentially important patient factors (such as body mass index [BMI], smoking status, presurgical motion, limb alignment, or which compartment was replaced) and surgeon factors (such as surgical volume) were not captured in their data set.

The authors’ findings regarding age are perhaps the most exciting challenge to traditional thinking for patients being considered for UKA. They clearly demonstrate a progressively lower HR for revision following UKA with progressively increasing age at the time of the index procedure. This finding is consistent with contemporary reports including a recent systematic review of 20 studies evaluating patients who underwent UKA when they were older than 70 that showed favorable clinical outcomes and 10-year survivorship of 87.5% to 98%4. By reinforcing recent evidence and adding new evidence for our appraisal, the large cohort study by Ekhtiari et al. should convincingly change surgeons’ criteria for selecting UKA for older patients who may otherwise be suitable candidates but, using more traditionally accepted criteria, would instead undergo TKA. Evidence regarding other factors—such as male sex, having diabetes, and using cementless fixation—is more mixed in the literature, and these factors should be evaluated further.

As we continue to expand on traditional indications for UKA, this procedure will undoubtedly remain an important alternative in the orthopaedic armamentarium for predominantly unicompartmental OA of the knee. We should be grateful to Ekhtiari et al. for sharing their findings and enabling us to better characterize the role of UKA for the management of painful knee unicompartmental degeneration.