J Orthop. 2026 Apr;74
412-418
Aims: Robot-assisted rehabilitation (RAR) has emerged as a novel strategy to enhance recovery after total knee arthroplasty (TKA). This meta-analysis evaluates the impact of RAR on postoperative pain, muscle strength, range of motion (ROM), and functional recovery in comparison with conventional physical therapy (PT).
Methods: A systematic search through October 2025 across PubMed, EMBASE, CINAHL, Web of Science, Cochrane CENTRAL, ProQuest Dissertations & Theses, Google Scholar, and ClinicalTrials.gov identified randomized controlled trials (RCTs) comparing RAR versus conventional PT after primary TKA. Outcomes were pooled using random-effects models, and risk of bias was assessed with the Cochrane tool.
Results: Six RCTs (309 patients; 155 intervention, 154 control) met inclusion, with overall moderate quality and low-to-moderate risk of bias. Pooled analysis showed no significant differences in ROM for knee extension at 1 week (MD: 0.38, p = 0.54) or 2 weeks (MD: 0.41, p = 0.47), or knee flexion at 1 week (MD: 1.55, p = 0.45). Knee extension strength was also comparable (MD: 0.01, p = 0.88), though knee flexion strength was significantly improved with RAR (MD: 0.09, p < 0.0001). No significant difference was observed in postoperative pain (MD: -10.77, p = 0.13), walking speed (MD: 0.05, p = 0.36), or length of stay (MD: -1.57, p = 0.23).
Conclusion: RAR after TKA produced outcomes similar to PT across ROM, pain, strength, walking speed, and length of stay, with only knee flexion strength showing significant improvement. Restricting to RCTs and TKA-only populations yielded more conservative results than prior reviews, likely reducing bias. However, additional large-scale, long-term trials are needed to define durability, cost-effectiveness, and optimal candidates.
Keywords: Exoskeleton; Functional recovery; Muscle strength; Pain; Robot-assisted rehabilitation; Systematic review; Total knee arthroplasty