A Systematic Review and Meta-analysis on the Effectiveness of Transcranial Direct Current Stimulation on Swallowing Function of Poststroke Patients.

Only 10 trials and 343 patients were included, and the authors state plainly that the results cannot be confidently interpreted. The trials used different swallowing scales and different stimulation settings (site, dose, duration), which the authors describe as substantial clinical heterogeneity. Only four of the 10 trials described how they generated the randomisation sequence, and only two said how they concealed which treatment each patient would get, so allocation may often have been known in advance. No trial protocols were available, so the review could not check whether the outcomes reported were the ones originally planned. Studies published in languages other than English or Chinese were excluded, and the authors note that most of the newly added trials came from China. The review recorded whether trials used exercises or other therapy alongside stimulation but does not report it, so it cannot separate the effect of stimulation from whatever else patients were receiving at the same time. The combined estimate is sensitive to single trials: removing one trial from the both-hemispheres analysis removed the significant effect.

The work was paid for by a government health grant, the Fujian Health and Family Planning rural and urban community appropriate technology projects fund (No. 2020TG019). No commercial sponsor or device manufacturer is reported. The authors state that no conflicts of interest have been reported by them or by anyone else in control of the content of the article.

Do not read this as proof that brain stimulation works for post-stroke swallowing difficulty. The combined result comes from 10 small trials that used different swallowing scales and different stimulation settings, and only two of them described how patients were allocated to groups, so the estimate is fragile. The paper's most quotable finding, that stimulating the unaffected hemisphere is best, is a subgroup comparison the authors themselves say needs further confirmation.