A number of randomized controlled trials of acupuncture for stroke recovery were critically reviewed, beginning with an existing systematic review and meta-analysis. A number of these clinical studies suffered from methodological flaws that tended to obscure and reduce the reported effect size. These flaws included inadequate statistical analysis, failure to adequately account for differences in baseline stroke severity, and the use of an inadequate post-treatment assessment period. These three flaws, taken individually or in combination, resulted in the underreporting of acupuncture effects in 5 of the randomized controlled trials reviewed. By informally adjusting study results for these limitations, it was possible to demonstrate that acupuncture is probably much more effective in assisting stroke recovery than has generally been reported, especially when the stroke is in the moderately severe range. The negative impact of well-intended but inappropriate exclusion criteria in the meta-analysis was also illustrated. The importance of giving more attention to ancillary results, such as walking speed and mortality, was illustrated as a way to gain a deeper understanding of the true impact of a still poorly understood therapy such as acupuncture. The general conclusion of the reanalysis presented here is that there is in fact substantial evidence that acupuncture is effective as an adjunctive treatment for facilitating stroke recovery.
Topics in Stroke Rehabilitation, 2007, 14(4), 40-58