Background: Robot-assisted therapy (RAT) represents an evolving approach to post-stroke upper limb rehabilitation; however, its impact on spasticity modification remains inadequately characterized. Critical knowledge gaps persist due to the diversity of robotic technologies, heterogeneous treatment protocols, and methodological limitations in outcome reporting, particularly the conflation of proximal and distal spasticity measures. This systematic review and meta-analysis evaluated the comparative effectiveness of robotic versus conventional interventions for upper limb spasticity management, examining the influence of robotic design, stroke stage, and treatment dosage on spasticity outcomes.
Methods: We searched six databases and two trial registries up to November 2024 for RCTs of RAT versus non-robotic comparators in adults with post-stroke upper limb spasticity. Secondary outcomes included upper-limb motor function, activities of daily living, and health-related quality of life. We used random (I2 ≥ 50%) or fixed effects meta-analysis to synthesize results, with subgroup analyses for stroke phase, device type, and anatomical region. The certainty of evidence was assessed using GRADE.
Results: Twenty-nine RCTs (n = 1,577) were included. Overall, RAT produced a small but statistically significant reduction in elbow flexor spasticity on the Modified Ashworth Scale (MD − 0.25; 95% CI [− 0.40 to − 0.10]). This reduction was observed primarily in trials conducted in the chronic phase. Wrist flexor spasticity was also reduced (MD = − 0.22; 95% CI −0.37 to − 0.08). Reduction was also significant in studies using multi-joint devices (SMD − 0.26; 95% CI [− 0.43 to − 0.08]). In contrast, no significant reduction was found for shoulder adductor spasticity or for hand-only devices. RAT also improved motor function with Fugl-Meyer Assessment MD 1.97; 95% CI [0.83 to 3.12] but did not consistently improve ADL and quality of life. Importantly, there was no significant difference in the odds of treatment-related adverse events (OR 1.18; 95% CI [0.53 to 2.66]).
Conclusions: This systematic review provides evidence that RAT is a safe modality that can modestly reduce elbow and wrist flexor spasticity in post-stroke patients. The benefit is specific, appearing more robust with multi-joint devices and in the chronic phase of recovery. The primary value of RAT seems to be in safely delivering high-intensity motor practice rather than acting as a first-line anti-spasticity treatment.
Registration PROSPERO CRD 42024615294.
Supplementary Information: The online version contains supplementary material available at 10.1186/s12984-026-01925-8.
Keywords: Muscle spasticity; Robotics; Stroke rehabilitation; Systematic review; Upper extremity.