Premium
P4‐180: DESIGN OF THE SARTAN‐AD TRIAL
Author(s) -
Black Sandra E.,
Lanctôt Krista L.,
Zotovic Ljubica,
Oh Paul,
Masellis Mario,
Greenberg Barry D.,
Borrie Michael,
Cohen Sharon,
Fischer Corinne,
Garcia Angeles,
Goldlist Barry,
Ingram K. Jennifer,
Kumar Sanjeev,
Marotta Giovanni,
Patterson Christopher,
Pollock Bruce,
Thorpe Kevin E.,
Verhoeff Paul
Publication year - 2014
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1016/j.jalz.2014.05.1697
Subject(s) - dementia , medicine , cardiology , stroke (engine) , disease , epidemiology , cognitive decline , population , heart failure , mechanical engineering , environmental health , engineering
Such differences would serve to limit the utility of these analyses to determine differential treatment responses. We assessed the extent of this problem using pooled clinical trials data and simulations reflecting cognitive outcomes and post hoc analyses based on baseline cognition assessed by the MMSE. Methods:We used a metadatabase of 18 ADCS studies and ADNI to analyze the rate of decline on the ADAS-cog among participants. We included 2,604 subjects withmild tomoderate ADwho completed theMMSE at baseline and at least one assessment on the ADAS-cog. Baseline MMSE scores were broken into categories of 0-10, 11-14, 15-18, 19-22, 23-26, and 27-30. We used mixed effects models (random coefficient models) to estimate the rate of decline in ADAS-cog scores among the MMSE groups, adjusting for age and education. Results: Individuals in lower baseline MMSE categories tended to be older, less educated, and female. Baseline MMSE category was strongly associated with baseline ADAS-cog scores (p<0.001), independent of the effects of age (p<0.001) and education (p<0.001). However, the rates of decline on the ADAS-cog were similar among the MMSE categories in both unadjusted and adjusted models (p1⁄40.18). Conclusions: Analyses of AD clinical trials often find differences in treatment response based on level of cognitive impairment at baseline. However, such observed differences from particular trials disagree with pooled data from a substantial number of the trials. Post hoc analyses based on severity of MMSE may be unreliable for predicting future trials. If warranted by the expected actions of the drug, investigators should prospectively stratify based on cognitive severity in order to minimize erroneous conclusions.