Twenty-six RCTs (22 placebo-controlled trials and four head-to-head trials) were included in the review. The total number of included patients was not reported. Where reported, 19 trials were judged as fair quality and four trials as good quality.
Compared with placebo, a significant improvement in cognition based on ADAS-cog scores was observed for donepezil (WMD -2.67, 95% CI -3.28 to -2.06; seven treatment comparisons), galantamine (WMD -2.76, 95% CI -3.17 to -2.34; 10 treatment comparisons) and rivastigmine (WMD -3.01, 95% CI -3.80 to -2.21; two RCTs). Significant heterogeneity was found only for the pooled outcome of rivastigmine (I2=70%).
Compared with placebo, a significant improvement in functional outcome change was observed for donepezil (SMD 0.31, 95% CI 0.21 to 0.40; eight treatment comparisons), galantamine (SMD 0.27, 95% CI 0.18 to 0.36; six treatment comparisons) and rivastigmine (0.26, 95% CI 0.11 to 0.40; three RCTs). No significant heterogeneity was found for these outcomes.
A significant improvement in behaviour based on the Neuropsychiatric Inventory was found for donepezil (WMD -4.3, 95% CI -5.95 to -2.65; four RCTs) and galantamine (WMD -1.44, 95% CI -2.39 to -0.48; five treatment comparisons) compared with placebo. No significant heterogeneity was found for all these outcomes.
A significant increase in responding to the treatment was observed for donepezil (RR 1.88, 95% CI 1.50 to 2.34; six treatment comparisons) and rivastigmine (RR 1.64, 95% CI 1.29 to 2.09; two RCTs) compared with placebo, but not for galantamine (RR 1.15, 95% CI 0.96 to 1.39; five treatment comparisons). No significant heterogeneity was found for these outcomes.
Potential publication bias was found only for the outcome of global assessment of change (funnel plots were not presented). Sensitivity analyses by dose did not significantly after the results.
In head-to-head comparisons, one trial found donepezil to be more efficacious than galantamine and one study found rivastigmine to be more efficacious than donepezil. Two studies showed no significant difference in efficacy between these drugs.
Adjusted indicted comparisons did not show significant differences in cognition between drugs, but found global response to be better with donepezil and rivastigmine than with galantamine. The results favored donepezil over galantamine regarding the outcome of behaviour.
The most frequently reported adverse events were nausea, diarrhoea, dizziness and weight loss. Incidence of these adverse events between trials was consistently lowest for donepezil and highest for rivastigmine.