Early Revision Rates and Clinical Outcomes of Two Cementless Cruciate-Retaining Total Knee Arthroplasty Designs at Two-Year Follow-Up: A Propensity Score-Adjusted Comparative Analysis

Background and Objectives: Cementless total knee arthroplasty (TKA) has gained popularity but comparative outcome data across implant systems remain limited. This study compared survivorship and clinical outcomes of two cementless cruciate-retaining TKA implants from different manufacturers. Materials and Methods: We performed a retrospective, single-institution review of 220 patients who had a cementless cruciate-retaining TKA implant between 2019 and 2024; 110 patients received Implant A and 110 patients received Implant B. Baseline variables included demographics, comorbidities, and preoperative Knee Injury and Osteoarthritis Outcome Score for Joint Replacement (KOOS, JR). Clinical outcomes included 90-day readmission, all-time revision, reoperation, manipulation under anesthesia, and postoperative KOOS, JR scores. Propensity score-based overlap weighting was used to create a balanced pseudo-population for adjusted analyses. Adjusted outcomes included weighted logistic regression for two-year revision and a weighted Cox proportional hazards model for time to revision. Results: Crude revision rates were low and not significantly different (A: 0.9 versus B: 2.7%, p = 0.622). Overlap weighting achieved balance across all measured baseline covariates. Unadjusted rates of 90-day readmission, reoperation, and manipulation under anesthesia were similar between groups. In adjusted analyses, no statistically significant difference was observed for two-year revision in Implant B vs. A cohorts (OR 2.17, 95% CI 0.10 to 157, p = 0.621) or time to revision (HR 2.22, 95% CI 0.18 to 27.09, p = 0.533). Conclusions: Both cementless cruciate-retaining TKA systems demonstrated low early revision rates in this cohort. Propensity score-adjusted analyses reduced measured baseline imbalances and did not identify a significant difference in revision risk between implants; however, interpretation is limited by sparse revision events, potential residual confounding, and the retrospective observational design.