Twenty-two RCTs (n=9,030) were included.
The quality assessment identified numerous methodological weaknesses in the included studies. These included a lack of correction for multiple statistical comparisons (7 studies), no intention-to-treat analysis (15 trials) and an inappropriate handling of incomplete data (8 studies).
ADAS-cog (14 trials): the mean differences between the treatment and placebo groups ranged from 1.5 to 3.9 points on the 70-point scale. Twelve of the trials showed a statistically significant difference in favour of the treatment group.
CIBIC-plus (12 trials): benefits of treatment were reported in all trials that used this outcome measure, with a statistically significant difference found between groups in favour of the treatment group. In five of these studies the difference was no longer statistically significant when the multiple statistical comparisons had been accounted for. In those studies reporting a mean difference (5 trials), the mean differences ranged from 0.26 to 0.54 points on the 7-point scale.
Other outcome measures (10 trials): the findings in this group of studies were mixed, with 5 studies showing a statistically significant benefit in favour of the treatment group. However, after correcting for multiple testing, the difference was no longer statistically significant in 2 studies. There were no differences between treatment and placebo in 4 studies, and in one study there was a negative effect with treatment.
Adverse events: there was a broad spectrum of adverse events with the cholinesterase inhibitors, the most common being nausea, vomiting, diarrhoea and weight loss.