Focal Muscle Vibration for Upper Limb Spasticity After Stroke: A Systematic Review and Meta-Analysis.

Abstract

To estimate the effect of focal muscle vibration (FMV) on post-stroke upper limb spasticity and explore whether frequency, amplitude, or target joint moderates effects. PubMed, EMBASE, Web of Science, CINAHL, PEDro, and Cochrane Central Register of Controlled Trials were searched from inception through March 2026. Studies of adults with post-stroke upper limb spasticity were eligible if they applied FMV to the upper limb and reported the Modified Ashworth or Tardieu Scale. Controlled parallel, randomized crossover, and comparative repeated-measures studies were eligible, including peer-reviewed reports and preprints with sufficient outcome data. Two reviewers independently extracted data and assessed risk of bias. Hedges' g was calculated from pre-to-post change scores and synthesized with a three-level random-effects model (restricted maximum likelihood). Certainty was assessed with GRADE. The review was registered in PROSPERO (CRD420251181549). Eleven reports representing 11 independent participant samples contributed 32 effect estimates from 285 participants assigned to or receiving an FMV condition. The three-level pooled effect was g=-0.526 (95% CI, -0.742 to -0.310; p<.001). All three frequency-band estimates favored FMV, with 95% CIs below zero, and no differences were detected among frequency bands (Wald; p=.948); continuous frequency was not associated with effect size (β=0.00117 per Hz; 95% CI, -0.00117 to 0.00352; p=.326). All 12 finger estimates were negative. Leave-one-out analyses retained confidence intervals below zero, and Egger's regression on 10 independent report-level aggregates suggested possible small-study effects (bias coefficient=-6.10; p=.002). Certainty in the primary outcome was moderate. FMV probably reduces post-stroke upper limb spasticity. The antispasticity effect was observed across the 30-300 Hz range studied, with no frequency-related modification detected, and finger estimates showed the most uniform effect direction. These findings support flexible frequency selection within studied protocols. Larger trials are needed to refine amplitude, dose, and muscle targeting.