Critically appraised paper: Non-invasive brain stimulation does not enhance the effect of robotic-assisted upper limb training on arm motor recovery after stroke [commentary]

摘要

Summary of: Reis SB, Bernardo WM, Oshiro CA, Krebs HI, Conforto AB. Ef- fects of robotic therapy associated with noninvasive brain stimulation on upper-limb rehabilitation after stroke: systematic review and meta-analysis of randomized clinical trials. Neurorehabil Neural Repair. 2021;35:256–266. Objective: To review the evidence as to whether adding non-invasive brain stimulation enhances the effects of robotic-assisted upper limb training on upperlimbmotorrecoveryinindividualswithstroke.Datasources:MEDLINE, EMBASE, CENTRAL, LILACS, CINAHL, DORIS, and PEDro were searched up to July 2019. This search was supplemented by searching online archives of theses and trial registries. Study selection: Randomised controlled trials (parallelorcrossover)involvingpeoplewithupperlimbparesisduetostroke, in which non-invasive brain stimulation before, during or after robotic- assisted upper limb rehabilitation was compared with sham non-invasive brain stimulation or robotic-assisted upper limb rehabilitation without non- invasive brain stimulation.Outcomemeasures were upperlimbperformance in either impairment-level and/or activity-level domains. Data extraction: Two reviewers extracteddata and discrepancies were resolved byconsensus. For crossover designs, only the first-phase intervention data were extracted. Risk of bias for individual studies was assessed according to specified criteria by two reviewers and quality of the body of evidence was rated according to GRADEpro. Data synthesis: Of 1,176 articles identified by the search, eight unique trials with a total of 324 participants (161 active,163 control) met the selectioncriteriaandwereincludedinthereview.Thequalityofevidencewas high for both impairment-level and activity-level outcome measures. Based on the quantitative pooling of the available data, there was no effect of non- invasive brain stimulation on upper limb performance on the Fugl-Meyer Assessment (seven studies, MD 0.15, 95% CI 23.10 to 3.40) or on upper limb activity limitation (five studies, SMD 0.03, 95% CI 20.28 to 0.33). Planned subgroupanalysesdemonstratedsimilarresultsforbothsubacuteandchronic stroke, robotic device characteristics (end-effector and exoskeleton), upper limb joints involved in training, and unimanual and bimanual training. There was noevidence thatnon-invasive brain stimulationparadigms (increasedor decreasedcorticalexcitability),timingofstimulation(before,afterandduring robotic-assisted therapy), or number of sessions influenced the results. Conclusion: At present, there is high-quality evidence to suggest that the ef- fectsofrobotic-assistedupperlimbtrainingonupperlimbmotorimpairment ormotoractivityforindividualswithstrokearenotenhancedbyexistingnon- invasive brain stimulation approaches.