Network Meta-Analysis of Non-Conventional Therapies for Improving Upper Limb Motor Impairment Poststroke

• Published in: Stroke (1970) 2022-12, Vol.53 (12), p.3717-3727
• Main Authors: Saikaley, Marcus, Pauli, Griffin, Sun, Hao, Serra, Julisa Rodriguez, Iruthayarajah, Jerome, Teasell, Robert
• Format: Article
• Language: English
• Subjects: Adolescent; Bayes Theorem; Humans; Motor Disorders; Network Meta-Analysis; Original Contributions; Recovery of Function; Stroke - complications; Stroke - therapy; Stroke Rehabilitation - methods; Upper Extremity
• ISSN: 0039-2499; 1524-4628; 1524-4628
• DOI: 10.1161/STROKEAHA.122.040687

Network meta-analysis is a method that can estimate relative efficacy between treatments that may not have been compared directly within the literature. The purpose of this study is to present a network meta-analysis of non-conventional interventions to improve upper extremity motor impairment after stroke. A literature search was conducted in 5 databases from their inception until April 1, 2021. Terms were used to narrow down articles related to stroke, the upper extremity, and interventional therapies. Randomized controlled trials written in English were eligible if; 50% poststroke patients; ≥18 years old; applied an intervention for the upper extremity, and/or used the Fugl-Meyer upper extremity scale as an outcome measure; the intervention had ≥3 randomized controlled trials with comparisons against a conventional care group; conventional care groups were dose matched for therapy time. A Bayesian network meta-analysis approach was taken to estimate mean difference (MD) and 95% CI. One hundred seventy-six randomized controlled trials containing 6781 participants examining 20 non-conventional interventions were identified for inclusion within the final model. Eight of the identified interventions proved significantly better than conventional care, with modified constraint induced movement therapy (MD, 6.7 [95% CI, 4.3-8.9]), high frequency repetitive transcranial magnetic stimulation (MD, 5.4 [95% CI, 1.9-8.9]), mental imagery (MD, 5.4 [95% CI, 1.8-8.9]), bilateral arm training (MD, 5.2 [95% CI, 2.2-8.1]), and intermittent theta-burst stimulation (MD, 5.1 [95% CI, 0.62-9.5]) occupying the top 5 spots according to the surface under the cumulative ranking curve. Overall, it would seem that modified constraint induced movement therapy has the greatest probability of being the most effective intervention, with high-frequency repetitive transcranial magnetic stimulation, mental imagery, and bilateral arm training all having similar probabilities of occupying the next spot in the rankings. We think this analysis can provide a guide for where future resources and clinical trials should be directed, and where a clinician may begin when considering alternative therapeutic interventions.