# Comparative effects of six rehabilitation therapies on lower limb function and gait function in stroke patients: a network meta-analysis of 33 RCTs

## TL;DR

This study compares six rehabilitation therapies for stroke patients, finding that intelligent and neuromodulation rehabilitation are most effective for lower limb and gait function, respectively.

## Contribution

A network meta-analysis ranks six rehabilitation therapies for stroke recovery, identifying optimal interventions for specific functional goals.

## Key findings

- Intelligent rehabilitation had highest probability of improving lower limb function in stroke patients.
- Neuromodulation rehabilitation showed greatest potential for enhancing gait function in stroke patients.
- Standardized rehabilitation demonstrated relatively weaker effects compared to other therapies.

## Abstract

This study aims to systematically compare the relative efficacy of six non-invasive rehabilitation interventions—Standardized Rehabilitation (SR), Aerobic Rehabilitation (AR), Resistance Training Rehabilitation (RTR), Intelligent Rehabilitation (IR), Traditional Chinese Rehabilitation (TCR), and Neuromodulation Rehabilitation (NR)—in improving lower limb function and gait function in stroke patients using a network meta-analysis approach. It also ranks the efficacy of each intervention.

We systematically searched five databases—PubMed, Embase, EBSCO, Web of Science, and Scopus—for randomized controlled trials (RCTs) published from January 2003 to November 2025. Two researchers independently screened studies, extracted data, and assessed risk of bias. Stata 18.0 software was used for statistical analysis. Standardized mean differences (SMDs) and their 95% confidence intervals (CIs) were calculated. Probabilistic ranking of intervention efficacy was performed using the cumulative ordered ranking curve area (SUCRA) value.

A total of 33 RCTs were included. Network meta-analysis revealed: (1) For lower limb function improvement (using the Fugl-Meyer lower limb score as the core indicator), the efficacy ranking was IR (SUCRA = 75.7) > AR (60.7) > RTR (53.1) > SR (10.5). Smart rehabilitation had the highest probability (46.7%) of being the optimal approach. (2) For gait function improvement (core measures: walking speed, 6-min walk test), the efficacy ranking was NR (SUCRA = 82.2) > IR (71.6) > AR (50.8) > RTR (38.8) > TCR (35.7) > SR (20.9). Neuromodulation rehabilitation had the highest probability of being the optimal solution (36.5%). Direct and indirect comparison results were largely consistent, with funnel plots showing no significant publication bias.

Based on existing evidence, intelligent rehabilitation may offer relative advantages in improving lower limb function in stroke patients, while neuromodulation rehabilitation demonstrates greater potential for enhancing gait function. Standardized rehabilitation, as a conventional baseline approach, demonstrated relatively weaker effects. The ranking results from this study provide evidence-based guidance for clinicians selecting individualized rehabilitation programs targeting different functional goals. Future high-quality research is needed to validate and refine intervention recommendations for different disease stages.