ObjectivesTo evaluate the clinical evidence for traditional medicines (TMs) used in East Asia on measures of cognition in Alzheimer disease, determine the effect sizes at different time points for the TMs and pharmacotherapies, and assess the tolerability of the TMs. MethodsWe searched 12 databases in English, Chinese, and Japanese for eligible randomised controlled trials that compared orally administered TMs with pharmacotherapy and reported cognitive outcomes. Meta-analyses were conducted for Alzheimer's Disease Assessment Scale-cognitive subscale and/or Mini-Mental State Examination (MMSE). Mean differences and 95% confidence intervals were calculated to evaluate treatment effects. ResultsThirty randomised controlled trials met inclusion criteria. Twenty-nine compared TMs with donepezil. Single studies provided comparisons with galantamine, rivastigmine, or memantine. There were no significant differences between the TM and donepezil groups at 12 or 24weeks for Alzheimer's Disease Assessment Scale-cognitive subscale or MMSE. Improvements over baseline were significant for MMSE at 12 and 24weeks within the TM and donepezil groups and remained significant at 1year. Effect sizes were reduced in the 3 double-blind studies. At 24weeks, donepezil 10mg/d generally produced greater improvements in MMSE than 5mg/d. Tolerability reporting was incomplete and inconsistent between studies. ConclusionsThe results suggested that the clinical benefits of the TMs were not less than donepezil at comparable time points, with both groups showing improvements. However, lack of blinding in most studies and other design and measurement issues are likely to have resulted in overestimation of effect sizes in both groups. Further well-designed studies are needed.