Cognitive decline across five cognitive batteries: Sample size implications for clinical trials

We evaluated the statistical power for a theoretical randomized trial of anti-amyloid treatment in preclinical Alzheimer's Disease across five cognitive composites in preclinical Alzheimer's Disease across five cognitive composites: Alzheimer's Prevention Initiative Preclinical Composite Cognitive Test (APCC); Preclinical Alzheimer's Composite with Semantic Processing (PACC5); Preclinical Alzheimer's Cognitive Composite (PACC); and global and episodic memory composites.

We utilized annual cognitive assessments from 517 decedents (78.2 ± 4.7years; 72% female) with post mortem pathologic Alzheimer's disease (AD) to represent amyloid positivity. We calculated sample sizes to detect 30% reduction in 5-year slopes of cognitive decline for equal size treatment versus placebo groups across composites.

Estimated sample sizes for APCC (n = 1633, 95% confidence interval [CI] 1400-1823), PACC (n = 1822, 95% CI 1612-2122), and episodic memory (n = 3141 95%CI 2563-3732) were larger than for PACC5 (n = 1424, 95% CI 1249-1575). Sample size estimates were similar between PACC5 and the global composite (n = 1267, 95%CI 1336-1407).

Small changes in composites, such as addition of semantic fluency in PACC5, could be considered as part of approaches to improve statistical power.

We evaluated statistical power of a theoretical 5-year randomized trial testing anti-amyloid treatments in early Alzheimer's across five cognitive composite endpoints.We leveraged annual cognitive assessment in Rush Alzheimer's Disease Center cohorts and used post mortem pathologic AD to represent amyloid positivity.Preclinical Alzheimer's Composite with Semantic Processing (PACC5) required significantly lower sample size to achieve power for a 30% reduction in cognitive slope than Alzheimer's Disease Cooperative Study-Preclinical Alzheimer's Cognitive Composite (PACC).PACC5 had better statistical power than Alzheimer's Prevention Initiative Preclinical Composite Cognitive Test (APCC) and an episodic memory composite.Small changes in cognitive composites can improve detection of cognitive decline.