Ten studies (n=656) were included: 3 randomised controlled trials (RCTs; n=571) and 7 open studies (n=85).
RCTs (sample size: 14 to 541).
All of the RCTs were analysed using the last-observation-carried- forward method. Two were double-blind placebo-controlled and one used a crossover design.
The largest RCT (n=541) found that, compared with placebo, rivastigmine (3 to 12 mg daily) significantly improved cognition and neuropsychiatric symptoms: the change in MMSE was 0.8 versus -0.2 with placebo (P=0.03); the decrease in Alzheimer's Disease Assessment Scale-Cognitive subscale was 2.1 versus 0.7 with placebo (P<0.001); and the change in NPI was -2.0 versus 0.0 with placebo (P=0.02). However, this RCT found no difference between treatment for Parkinsonian symptoms (change in UPDRS, P=0.83). Rivastigmine was associated with higher withdrawal rates than placebo (27.1% versus 17.9%), significantly increased nausea (29% versus 11.2%, P<0.001), vomiting (16.6% versus 1.7%, P<0.001), Parkinsonian symptoms (27.3% versus 15.6%, P=0.002) and tremor (10.2% versus 3.9%, P=0.01).
The other 2 RCTs were small crossover double-blind RCTs (n=14 and n=16, respectively). One crossover RCT (n=14) found that donepezil significantly improved cognition: the change in MMSE score at 10 weeks was 2.1 versus 0.3 with placebo (P=0.013). The other (n=16) found no significant difference between donepezil and placebo in cognition: the Dementia Rating Scale total delta score was -0.68 (95% confidence interval: -7.63, 6.82, P=0.84). Neither RCT found any statistically significant difference in Parkinsonian. Donepezil increased the mean number of adverse effects compared with placebo in both RCTs (4.2 versus 2.8 and 71.4% versus 44.4%, respectively).
Open trials (sample size from 5 to 28).
The results were mixed. Three trials showed no effect of anticholinesterase inhibitors on cognition, three showed an improvement, and the remaining trial did not use the MMSE or any standard measure.