Size of the treatment effect on cognition of cholinesterase inhibition in Alzheimer's disease

Background: Six cholinesterase inhibitors (ChEIs) have been tested in people with Alzheimer's disease, using methods currently required for regulatory approval. The clinical importance of their treatment effects is controversial. Objective: To determine whether cholinesterase inhibition produces treatment effects in Alzheimer's disease that are large enough to be clinically detectable. Methods: Overview analysis of published trials of ChEIs in which the Alzheimer's Disease Assessment Scale-Cognitive Subscale (ADAS-Cog) and a global clinical measure were primary outcomes. Two quantitative summary measures of the treatment effect (Cohen's d and the standardised response mean (SRM)) were calculated and presented as funnel plots. Observed cases analyses and intention to treat (ITT) with the last observation carried forward (LOCF) analyses were compared. Results: The median Cohen's d effect sizes (ES) using ITT samples with LOCF for the ADAS-Cog were: low dose of a ChEI (n = 8 studies) median ES = 0.15, range = 0.03-0.22; medium dose (n = 13) median ES = 0.23, range = 0.12-0.29; high dose (n = 9) median ES = 0.28, range = 0.01-0.31. In general, the ES were larger when calculated as SRMs (for example, high dose ChEI studies, median SRM = 0.47; range = 0.30-0.63) and highest in the observed cases analyses (for example, high dose median SRM = 0.56, range = 0.35-0.78). Global clinical scales produced similar estimates of ES (for example, high dose ChEI, ITT/LOCF median Cohen's d = 0.29, range = 0.20-0.47). Conclusions: ChEIs produce small-moderate effect sizes in clinical trials which are reproducible and demonstrate a dose response. Better descriptions of the patterns of treatment response are needed to guide individual patient decisions about the effectiveness of treatment, but group effects are evident and appear large enough to be clinically detectable.