P2–409: A pilot study investigating the effects of rosiglitazone on performance in neuropsychological tests in patients with mild–to–moderate Alzheimer's disease

Abstract

Rosiglitazone (a PPARγ agonist) has shown beneficial effects on memory and executive function tasks in Alzheimer's disease (AD)1. To examine the use of standard clinical measures (e.g ADAS–cog) and experimental measures of simple and choice reaction time (examining attention, visuospatial abilities, executive function) to detect cognitive decline and drug effects in a subset of mild to moderate AD patients involved in a 24 week phase 2b trial of rosiglitazone2. Simple reaction time (SRT) (mean time to respond to a change in colour of a centrally presented target), and choice reaction time (CRT) (the mean time to correctly respond to a target changing colour in either the left or right visual field) data were collected at baseline and 24 weeks in addition to ADAS–cog in a subset of 56 patients from the main trial. The subset was divided approximately equally between 2mg, 4mg, 8mg rosiglitazone and placebo treatments throughout the study. Across the study as whole (511 patients in ITT population), and in the subset, no differentiation was seen between placebo and rosiglitazone groups on the ADAS–cog at 24 weeks2. Similarly, in the subset studied, SRT remained stable and did not differentiate between groups. However, placebo treated patients showed increased reaction times (slower to respond) on the CRT task at 24 weeks, relative to rosiglitazone treated patients. Their deterioration in this measure of attention and executive functions over time was greater than that of all 3 rosiglitazone groups, with an exploratory analysis reaching statistical significance indicating the greatest effects with the 2 mg and 4 mg rosiglitazone. This effect was particularly marked for targets in the left visual field. Recent evidence suggests a link between reduced attention and preferential bias to the right side of space3,4. The left–right asymmetry in performance seen in our placebo group at 24 weeks but not at baseline, offers further evidence of reduced attention and executive functions with disease progression. Simple computerised tasks, e.g. choice reaction time task, may be useful to measure decline in placebo group and beneficial drug effects in small studies.