Work

Feasibility, safety, and acceptability of robotic-assisted rehabilitation for patients with knee osteoarthritis

Year: 2026

Type: article

Abstract: Total knee replacement (TKR) is an effective treatment for end-stage knee osteoarthritis (KOA), yet optimal preoperative rehabilitation strategies remain undefined. Robot-assisted rehabilitation offers potential advantages in delivering controlled, intensive, and reproducible exercise protocols for patients with knee osteoarthritis. This early-phase feasibility study evaluated the safety, usability, and patient acceptability of the Knee-LOCOBOT, a novel robotic rehabilitation device designed for patients with KOA. No inference regarding efficacy can be drawn given the absence of a control group, randomization, or blinding. A single-arm feasibility study was conducted at Tan Tock Seng Hospital, Singapore (ClinicalTrials.gov: NCT07539948; NHG DSRB Reference: 2024–3529). Ten patients with KOA were recruited, with nine completing the intervention prior to TKR. Participants underwent four sessions of robot-assisted exercises incorporating cycling, leg press, and isometric quadriceps strengthening modes. Outcome measures included the Oxford Knee Score (OKS), EQ-5D-5 L, Knee Society Score (KSS), 6-Minute Walk Test (6MWT), Timed Up and Go (TUG), 30-second Chair Rise, quadriceps strength, and patient satisfaction survey. Recruitment, retention, and adherence were the primary feasibility outcomes. From 15 patients screened, 10 were enrolled (consent rate 71%, 10 of 14 eligible), of whom nine completed all four sessions, giving a retention rate of 90% (9/10) and 100% session completion and adherence among completers; the single withdrawal was for personal reasons unrelated to the device. A statistically significant within-group change in 6MWT distance was observed (median 291 m to 360 m; median within-participant change + 15 m, mean change + 31 m; p = 0.012, n = 9). In the absence of a control group this change cannot be attributed to the intervention. Right quadriceps strength showed a non-significant numerical increase ( p = 0.05, unadjusted, one of multiple comparisons). Exercised vs. non-exercised limb strength difference was small and non-significant (Cohen’s d = 0.11, p = 0.75). No adverse events were recorded. Patient acceptability was high (88.9% agreement). This small, single-arm, uncontrolled feasibility evaluation in nine selected low-pain KOA patients suggests that four supervised sessions can be completed without recorded adverse events, with high patient acceptability. No conclusions regarding clinical effectiveness or suitability for rehabilitation pathway integration can be drawn. An adequately powered randomised controlled trial is required before any efficacy claims can be made. Trial registration ClinicalTrials.gov, NCT07539948, date of first registration 27/03/2026.

Source: Scientific Reports

DOI: 10.1038/s41598-026-73345-6