Comparative analyses of Alzheimer's disease blood biomarkers and cognitive domains.

INTRODUCTION: Whether Alzheimer's disease (AD) blood biomarker-cognition associations differ across cognitive domains, analytic context, and biomarker modeling strategy in population-based cohorts is unclear. METHODS: In 1170 older adults from the Health and Retirement Study Harmonized Cognitive Assessment Protocol (HRS-HCAP), we examined cross-sectional (2016) and prospective (2016 to 2022) associations of blood phosphorylated tau at threonine 181 (p-tau181), glial fibrillary acidic protein (GFAP), neurofilament light (NfL), and amyloid beta 42/40 ratio (Aβ42/40) with memory, executive function, language, visuospatial ability, and global cognition using individual-biomarker, principal component analysis-derived composite, and multi-biomarker panel models. RESULTS: Cross-sectionally, NfL and GFAP showed the broadest associations with cognitive performance. In simultaneous models of baseline-adjusted follow-up cognition, p-tau181 was associated with lower memory (β = -0.080, q < 0.001) and global cognition (β = -0.066, q < 0.001), while GFAP was associated with lower executive function (β = -0.047, q = 0.007), memory (β = -0.047, q = 0.032), and global cognition (β = -0.066, q = 0.002). NfL and Aβ42/40 were not independently associated after mutual adjustment. P-tau181 also showed relative domain selectivity between memory and executive function. Relative model support differed by cognitive domain and analytic context. DISCUSSION: AD blood-based biomarker-cognition associations were domain-differentiated and context-dependent: Concurrent associations did not reliably indicate associations with cognition 6 years later, and only p-tau181 showed formal evidence of relative memory-versus-executive selectivity. Exploratory pairwise comparisons highlighted p-tau181 + GFAP for follow-up memory and global cognition, warranting independent evaluation.