To examine whether Cholinesterase Inhibitors (ChEIs) and memantine are superior to placebo in the treatment a) of Alzheimer’s disease (AD) for at least 6 months b) of Vascular Dementia (VaD). Methods: Extensive literature search in three electronic databases (PubMed, Cochrane Database, EMBASE) and four registers (Cochrane Collaboration Trial Register, Clinical Trial.gov, EU-Psi PsiTri, International Federation of Pharmaceutical Manufacturers & Associations Clinical Trial Portal) was done to retrieve all relevant records up to November 2007 for AD and up to June 2008 for VaD. A standarized data extraction form, developed especially for this review, was used. Due to variation expected, due to the wide variations found in AD and in VaD and in their response to treatment, a random effects model was used to calculate combined estimates of all treatment effects. The reporting quality was assessed by Jadad, Chalmers modified scales and CONSORT 22-items, Delphi checklists. Results: 22 trials met the inclusion criteria for AD (10 trials for donepezil, 3 for galantamine, 4 for rivastigmine and 5 for memantine). Data suggest rather modest benefit for ChEIs and memantine in cognition (Weighted Mean Difference-WMD for ADAS-cog -2.56, 95% CI -3.12, -2.01). Benefits have been observed in global function (Number-Needed-to-Treat-ΝΝΤ=10 for patients stabilized or improved in CIBIC-Plus & WMD for GDS 0.14, 95% CI 0.05-0.24). Available data suggest that memantine may also have a positive effect on behaviour (WMD for NPI -2.90, 95% CI -4.57, -1.22). Dropouts and total adverse events were higher in the treatment group for the ChEIs (Number-Needed-to-Harm-NNH 13 & 11, respectively) and not statistically significant for memantine. A subgroup analysis for severe dementia defined as MMSE?12) revealed modest benefits of donepezil and memantine in cognition, global function, activities of daily living and behavior , but as the number of trials is small the results can not be conclusive. Height trials met the inclusion criteria for VaD (2 on donepezil, 2 on galantamine, 2 on rivastigmine and 2 on memantine). Data on ADAS-cog suggest a rather small benefit of ChEIs and memantine in cognition deficits in VaD (WMD=-1.70; 95% CI -2.20, -1.20 for ChEIs & WMD=-1.85; 95% CI -2.83, -0.88 for memantine), while the NNT for global responders was not statistically significant. Results for activities on daily living were available only for the donepezil trials. Dropouts due to any cause were higher in treatment arms with a NNH 12(10-17) for ChEIs and 15(9-100) for memantine, respectively. Conclusions: The use of ChEIs and memantine were associated with modest benefits in cognition and global function in patients with AD, while dropouts and total adverse events were higher in the treatment group of ChEIs. Generalization of the results should be done with caution, as the included population in the trials does not fully represent the general AD population. However, it should be taken into account that even small benefits matter a great deal to patients and families facing a long neurodegenerative disease. ChEIs and memantine seems to provide modest benefits in cognition in VaD, with a significant dropout rate. Their use in VaD can not be generalized with available data to date