Plasma p-Tau217 and GFAP predict widespread cognitive decline in Alzheimer's disease

Progression in Alzheimer's disease (AD) involves three main interrelated biological axes-tau deposition, neurodegeneration, and neuroinflammation-that jointly drive cognitive decline. Although several cerebrospinal fluid (CSF) and plasma biomarkers along these axes are well validated for diagnosis, their value for prognosis remains uncertain. We assessed how baseline markers of each axis predict cognitive trajectories in biomarker-confirmed AD.

We included 136 A + T + N + individuals (median follow-up = 24 months [IQR 12-24]; mean = 17.6 months [SD = 12.4]). Tau-deposition markers (CSF p-Tau181; plasma p-Tau181 and p-Tau217), neurodegeneration markers (CSF t-Tau; CSF and plasma neurofilament light chain, NfL) and a neuroinflammation marker (plasma glial fibrillary acidic protein, GFAP) were quantified using CLEIA, ELISA or Simoa, and stratified into tertiles. Participants were classified by age at onset, clinical phenotype, and APOE ε4 status. Cognition was assessed annually with a comprehensive neuropsychological battery. Linear mixed-effects models (MMRM) were used to test biomarker-cognition associations and interactions with clinical variables.

Elevated CSF p-Tau181 and NfL levels were associated with greater decline in memory and executive function. Among plasma biomarkers, p-Tau217 and GFAP showed the strongest associations with widespread cognitive decline, particularly in language, visuospatial, and executive domains. These associations were independent of age at onset, clinical phenotype, and APOE ε4 status.

Our findings highlight the potential prognostic value of fluid biomarkers in AD, especially CSF p-Tau181 and NfL, and plasma p-Tau217 and GFAP. These results suggest promise for improving disease monitoring, although prognostic utility at the individual level remains uncertain.