Detecting an Early Response to Donepezil With Measures of Visual Attention

Acetylcholinesterase inhibitors (AChE-I) comprise a class of drugs used to treat Alzheimer's disease (AD), but controversy about their usefulness remains. Modest response rates of treated versus placebo groups, small effect sizes with respect to efficacy, drug costs, and clinical relevance of the effects are problematic. Standard efficacy measures of efficacy are not sufficiently sensitive, and trying to assess cognitive change after 4-6 months of therapy confounds the drug effect and the natural progression of the disease. Surprisingly, attention has never been included in the assessment of AChE-I drugs. The rationale for using attentional measures are that (1) Attentional deficits are recognized as a critical cognitive change in the earliest phases of AD; (2) Attentional function is directly mediated by the cholinergic system, and responds rapidly to cholinergic augmentation, particularly on tasks that tax available attentional capacity are dose dependent; and (3) Acetylcholine is depleted in AD. However, the link between attention and cholinergic depletion in AD has not been fully explored, especially with regard to response to cholinergic treatment. The study tests if attentional performance can be a more sensitive marker of response. In a longitudinal study we measure attentional, as well as cognitive and behavioral performance in de novo AD patients undergoing donepezil treatment. The investigators develop visual attentional measures and contrast them to global and domain-specific cognitive scores on three occasions (T1) baseline pre-treatment, (T2) after approximately 6 weeks, and (T3) after 6 months treatment. The T1-to-T2 arm is a double-blind placebo control period, after which members of the placebo group start open-label treatment. The assessment at 6 months allows us to determine whether the changes seen earlier at T2 can predict patients who respond, or determine which measures best predict response. We hypothesize that attention measures are more s