Differences in short-term complications between unicompartmental and total knee arthroplasty: A propensity score matched analysis

Background: Knee arthroplasty has emerged as an effective treatment for end-stage gonarthrosis. Although total knee arthroplasty remains the gold standard, unicompartmental knee arthroplasty is an appropriate alternative for select patients. We sought to use a large, heterogeneous national database to identify differences in thirty-day complication rates between unicompartmental and total knee arthroplasty as well as to identify risk factors for complications. Methods: Patients in the ACS NSQIP (American College of Surgeons National Surgical Quality Improvement Program) database who had undergone total or unicompartmental knee arthroplasty from 2005 to 2011 were identified. CPT (Current Procedural Terminology) codes were used to select cases of elective primary knee arthroplasty. Statistical models employing univariate and multivariate logistic regression identified risk factors associated with the thirty-day incidence of morbidity and mortality after total and unicompartmental knee arthroplasty. Propensity score matching addressed demographic differences between the total and unicompartmental knee arthroplasty cohorts. Results: A total of 29,333 patients were identified; 27,745 (94.6%) underwent total knee arthroplasty and 1588 (5.41%) underwent unicompartmental knee arthroplasty. Prior to matching, the total knee arthroplasty cohort was 63.7% female and had a mean BMI of 32.8 ± 7.3 kg/m², whereas the values for the unicompartmental cohort were 55.3% and 31.5 ± 6.5 kg/m² (p < 0.0001). The mean ages of these cohorts were 67.2 ± 10.1 and 64.0 ± 10.7 years, respectively (p < 0.0001). A previously developed and implemented propensity score matching algorithm was used to address the demographic differences. Following matching, the total complication rate did not differ significantly between the total and unicompartmental knee arthroplasty cohorts (5.29% compared with 4.16%, p = 0.35), whereas the rate of deep venous thrombosis (1.50% compared with 0.50%, p = 0.02) and the duration of hospital stay (3.4 compared with 2.2 days, p < 0.0001) were significantly higher in the total knee arthroplasty cohort. Conclusions: Comparison of total and unicompartmental knee arthroplasty revealed no differences in overall short-term (thirty-day) morbidity and mortality. Although this study does not address long-term subjective outcomes or implant survival, these findings should provide helpful information for surgeons counseling patients considering total and/or unicompartmental knee arthroplasty. Level of Evidence: Therapeutic Level III. See Instructions for Authors for a complete description of levels of evidence.