Abstract / Summary
Aims Robotic arm-assisted unicompartmental knee arthroplasty (R-UKA) has been shown to improve alignment and accuracy of component placement compared with conventional unicompartmental knee arthroplasty (C-UKA). Despite this, there is conflicting evidence for whether this has an impact on revision rates. Methods We performed a retrospective analysis of a prospective database, comparing patients who underwent R-UKA using the Stryker Restoris MCK implant with those who underwent C-UKA using the cemented Oxford Partial Knee implant, between January 2002 and December 2024 at The Prince Charles Hospital (Brisbane, Australia), with a minimum six-month follow-up. Patient selection for UKA mirrored the recommendations by Kozinn and Scott. The primary outcome of interest was to compare the all-cause revision rate between these two groups. Secondary outcomes of interest were to compare hospital length of stay (HLOS), surgical duration, the odds of superficial surgical site infection, and the odds of periprosthetic joint infection (PJI) between the same groups. The analysis of individual outcomes were adjusted for age, sex, BMI, American Society of Anesthesiologists grade, Charlson Comorbidity Index, and primary surgeon grade. Results The hazard of all-cause revision significantly decreased by 78% in the R-UKA group as compared with the C-UKA group (p = 0.046). HLOS was also significantly shorter by 2.2 days (95% CI 1.0 to 3.3) in the R-UKA group (p < 0.001). There was no significant difference in the surgical duration (p = 0.733), odds of superficial surgical site infection (p = 0.845), nor in the odds of PJI (p = 0.415) between the two groups. Conclusion The substantial improvement in revision outcomes suggests that R-UKA is the option of choice for managing symptomatic unicompartmental knee arthritis in carefully selected patients with isolated medial compartmental osteoarthritis, especially in the hands of low-volume surgeons. Decreased HLOS should be interpreted with caution, given the non-contemporaneous nature of the two cohorts. Cite this article: Bone Jt Open 2026;7(10):1304–1311.