ROBOTIC-ASSISTED TOTAL KNEE ARTHROPLASTY USING A SUBVASTUS APPROACH REDUCES BLOOD LOSS AND EARLY REOPERATIONS: A PROPENSITY-MATCHED COMPARATIVE STUDY OF 1,023 CONSECUTIVE TOTAL KNEE ARTHROPLASTIES

Introduction The subvastus approach in total knee arthroplasty (TKA) is a muscle-sparing technique that preserves the extensor mechanism and may facilitate early postoperative recovery. However, limited exposure may complicate implant positioning. Image-based robotic assistance has been introduced to improve surgical precision and reproducibility, potentially reducing soft-tissue trauma and perioperative blood loss. Nevertheless, clinical data evaluating early postoperative outcomes of robotic-assisted TKA performed through a subvastus approach remain limited. The aim of this study was to compare perioperative blood loss and early postoperative outcomes between robotic-assisted and conventional TKA performed through a subvastus approach. Methods All consecutive patients undergoing primary TKA with a Triathlon implant between January 2018 and January 2025 in a high-volume academic centre were retrospectively reviewed. Patients were divided into two groups according to the surgical technique: robotic-assisted TKA using the MAKO system and conventional mechanically instrumented TKA. Exclusion criteria included non-subvastus approach, unicompartmental or patellofemoral arthroplasty, previous infection of the knee, valgus alignment >183°, or follow-up <90 days. The primary outcome was estimated perioperative blood loss calculated using the hemoglobin-based Nadler and Gross method. Secondary outcomes included operative time, transfusion rate, and complications requiring reoperation within 90 days. Propensity score matching (1:3) based on BMI, ASA score, and sex was performed to reduce selection bias, followed by multivariable regression analyses. Results A total of 806 robotic-assisted and 217 conventional TKAs met the inclusion criteria. After propensity score matching, 548 robotic TKAs and 186 conventional TKAs were analyzed. Robotic-assisted TKA was associated with significantly lower blood loss compared with conventional TKA (mean difference 221 mL; p < 0.001). In multivariable analysis, robotic assistance remained independently associated with reduced blood loss (β −212 mL; 95% CI−273 to −151; p < 0.001). Transfusion rates were lower in the robotic group (2.5% vs 7.9%; adjusted OR 0.30; p = 0.0017). Early reoperations within 90 days were also less frequent with robotic assistance (2.1% vs 5.3%; OR 0.39; p = 0.043). Operative time was significantly shorter in the robotic group. Conclusion Robotic-assisted TKA performed through a subvastus approach was associated with reduced perioperative blood loss, lower transfusion rates, shorter operative time, and fewer early reoperations compared with conventional TKA. These findings suggest that robotic assistance may improve early perioperative safety while maintaining surgical efficiency in high-volume centres.