Abstract:We previously used longitudinal datasets from cognitively unimpaired persons at risk for late-onset and autosomal dominant Alzheimer's disease (AD) to characterize the combination of cognitive test scores most sensitive to track preclinical AD and estimate sample sizes for preclinical treatment trials using the resulting Alzheimer's Prevention Initiative (API) Composite Cognitive Test Score. We now extend this approach to a larger multi-center dataset acquired September 2005-September 2012 from the 32 past and present Alzheimer's Disease Centers and provided by National Alzheimer's Coordinating Center (NACC) to demonstrate the generalizability of our initial findings, optimize the age range and provide sample size estimates for proposed preclinical APOE-ε4 homozygote (HM) trial. 2–5 years of Uniform Data Set data from 4193 cognitively unimpaired subjects (3044 APOE-ε4 non-carriers, 1043 heterozygotes (HTs), and 106 HMs were used to compute longitudinal mean-to-standard-deviation ratios (MSDRs) to 1) confirm the power of the pre-specified composite cognitive score most closely resembling that derived from the Rush cohorts (despite differences in test batteries), 2) perform an exhaustive search of the combination of 1–8 UDS test scores most sensitive to distinguishing between persons who did or did not subsequently progress to clinical AD, providing an indicator of cognitive decline associated with preclinical AD after controlling for practice/aging effects, 3) identify the optimal age ranges for tracking composite score decline in the HMs and HTs, and 4) provide sample size estimates for preclinical trials within those age ranges. 1) The pre-specified composite cognitive score resulted in MSDRs similar to the Rush cohorts. 2) the UDS-based composite score most sensitive to preclinical AD decline was similar to that identified using the Rush dataset. 3) decline in this composite was greatest in 60–75 year-old HMs and in 65–80 year-old HTs. 4) 217 60–75 year-old HMs or 1717 65–80 year-old HTs per group are needed to detect 30% treatment effect with 80% power, p=0.05 in a 60-month trial. The API Composite Cognitive Test Score for evaluation of preclinical late-onset AD treatments was confirmed to be sensitive to track preclinical AD in a large multi-center cohort and used to inform the design of API's proposed trial in APOE-ε4 HMs.

O2-03-01: Validation of the Alzheimer's Prevention Initiative Composite Cognitive Test Score