Abstract
Alzheimers Dement. 2025 Oct;21(10):e70802. doi: 10.1002/alz.70802.
ABSTRACT
INTRODUCTION: Identifying at-risk individuals and selecting sensitive cognitive outcome measures are critical for designing efficient clinical trials targeting early Alzheimer's disease (AD) stages.
METHODS: We compared amyloid beta (Aβ)-positron emission tomography (PET), plasma tau phosphorylated at threonine 217 (p-tau217), Aβ42/40, and p-tau217/Aβ42-related decline across neuropsychological and functional measures in 225 individuals (176 cognitively unimpaired (CU), 49 with mild cognitive impairment (MCI). Johnson-Neyman analysis identified the biomarkers, which were used as trial inclusion criteria to estimate sample sizes needed to detect a 30% slowing across cognitive outcomes.
RESULTS: In the CU, using combined plasma Aβ42/40 and p-tau217 cut-offs for eligibility yielded the lowest sample size estimates for a comprehensive multidomain cognitive composite score, whereas sample sizes were higher for all other inclusion criteria based on single biomarkers. In MCI, estimates were substantially lower and less variable across most inclusion criteria and outcome measures.
DISCUSSION: These findings highlight the need for careful consideration of outcome measures, baseline diagnosis, and inclusion criteria, given their substantial effect on sample size estimation in trials.
HIGHLIGHTS: Aβ SUVR, plasma p-tau217, and the p-tau217/Aβ42 ratio predicted decline across multiple cognitive domains. Using cohort-specific biomarker cutoffs, sample size estimates were similar for p-tau217 combined with Aβ42/40 or Aβ SUVR. A multidomain composite best detected AD-related decline. Outcome measures and eligibility criteria strongly impact sample size estimates, especially in CU.
PMID:41117398 | PMC:PMC12538628 | DOI:10.1002/alz.70802
UK DRI Authors
