A disease progression meta-analysis model for cognitive deterioration in patients with Alzheimer's disease

Abstract

Previously, efforts have been made by various authors to evaluate disease progression in Alzheimer's disease (AD), utilizing patient data from individual clinical studies or pooled patient data across studies. In this analysis, we conducted a systematic review of public data sources from 1990 to 2008 of all available studies of acetylcholinesterase (AChE) inhibitors, as well as clinical studies evaluating the rate of deterioration in AD patients (e.g. Vitamin E study). The unique aspect of this analysis is that we have developed a model to describe the longitudinal response in ADAS-cog (change from baseline) in patients with mild-to-moderate AD and estimate disease progression in both placebo-treated and AChE inhibitor-treated patients. A total of 52 trials, with data from 19,972 patients and 84,441 individual observations, were identified. From these, we obtained 576 mean change in ADAS-cog scores from baseline values, spanning 12 weeks to 3 years of treatment. The model described the rate of disease progression, the placebo effect observed, and the symptomatic effect of AChE inhibitors. Baseline ADAS-cog, MMSE, age and publication year were tested as covariates. Disease progression in patients with mild-to-moderate AD across all available and relevant literature sources was estimated as 5.5 points per year (95% CI: 4.9-6.1) with baseline ADAS-cog of 25. An Emax-over time model best described the symptomatic drug effect for AChE inhibitors. Baseline ADAS-cog was a significant covariate on disease progression in that greater baseline impairment was associated with faster cognitive decline. The model was not able to describe any effect of baseline age, likely due to the narrow distribution of mean age (literature-level analysis). There was no significant impact of publication year in the model. Baseline ADAS-cog was a significant covariate on disease progression, and describes, or at least explains, the different rates of deterioration observed in mild versus moderate stages of AD. There was no significant impact of publication year in the model, suggesting a slower rate of disease progression has not been observed in more recent trials. Differences from various previous analyses may be due to differences in underlying disease severity in the datasets used.