Résumé
Purpose: Randomised controlled trials (RCTs) comparing robotic-assisted (RA-TKA) and conventional total knee arthroplasty (C-TKA) have demonstrated evolving trends in alignment strategies, surgical workflows and outcomes measurements. This review synthesises current evidence to clarify the evolving role of RA-TKA over the past two decades. Methods: A PRISMA systematic review was conducted searching PubMed, Cochrane Library, and Google Scholar (2000-2024). Eligible studies were Level I RCTs comparing RA-TKA to C-TKA with well-defined alignment strategies and surgical workflows. Data included demographics, robotic and prosthetic systems, learning curves, outcome measures and complications. Results: From 850 records, 9 RCTs were included (627 TKAs: 317 RA-TKA, 310 C-TKA). Patients averaged 66.9 years, BMI 28.1 kg/m(2), 27.9% male, 69.7 knees, follow-up 23.7 months. Coronal deformity was predominantly varus (<10 degrees). Robotic systems included Robodoc, Mako, and Navio using cruciate-retaining or posterior-stabilised prostheses. All earlier studies followed mechanical alignment (MA); after 2021 all RA-TKAs adopted personalised alignment strategies. Patellar management varied widely, with no RA-TKA workflow addressing the patellofemoral space. Most procedures were performed by one or two experienced surgeons; learning curves were only noted in early MA studies. RA-TKA reduced alignment outliers by 10%-24% (p < 0.05) versus C-TKA regardless of the alignment strategy, with no significant difference in functional scores (WOMAC, HSS, KSS, OKS and FJS), range of motion or complications. Recent trials increasingly assessed inflammatory markers, periarticular tissue injury, and quality-of-life outcomes. Conclusions: Over the past two decades, RA-TKA has evolved substantially, transitioning from purely bone-cutting technology to decision-making platforms enabling alignment personalisation and dynamic soft-tissue assessment, improving surgical precision regardless of alignment strategy. Clinical superiority over C-TKA remains inconclusive in relatively low-risk patients, predominantly older females with varus deformity, highlighting the need for studies with more diverse populations, prosthetic design, long-term follow-up, patient-specific outcomes, standardised workflow and learning curve reporting, and patellofemoral space optimisation.Level of Evidence Level I, systematic review of non-homogeneous RCTs.