This preprint examined whether plasma biomarkers measured in cognitively unimpaired individuals are preferentially associated with Alzheimer disease (AD)-related rather than age-related brain atrophy patterns. Clarifying these associations could inform the prognostic and diagnostic utility of blood biomarkers for preclinical AD.
The study used data from the Baltimore Longitudinal Study of Aging (BLSA). Plasma biomarkers were measured while participants were cognitively unimpaired (N=818). The measured analytes included Aβ42/Aβ40 (assayed using both Quanterix and Lumipulse platforms), glial fibrillary acidic protein (GFAP), neurofilament light chain (NfL), phospho-tau181 (p-tau181), and phospho-tau217 (p-tau217).
Longitudinal structural brain imaging comprised 2,293 MRI scans. Multidimensional atrophy pattern scores were derived: a brain age index (SPARE-BA), an AD-like atrophy index (SPARE-AD), and five dominant atrophy dimensions summarized as R-indices. These pattern scores were used to investigate spatial and multidimensional trajectories of brain atrophy.
During follow-up, 104 participants developed mild cognitive impairment (MCI) or dementia. Etiology classification recorded 74 conversions due to AD, 24 due to non-AD causes, and 6 with unknown etiology. Plasma biomarkers measured at the index visit (while participants were cognitively unimpaired) were tested for association with subsequent clinical conversion.
Associations between baseline plasma biomarkers and incident MCI/dementia due to AD versus non-AD etiologies were assessed. Longitudinal associations between plasma biomarkers and change in pattern scores were examined using linear mixed effects models on the repeated MRI measures. The abstract reports which biomarkers were associated with incident AD and with longitudinal changes in each imaging pattern; specific effect sizes, p-values, or confidence intervals were not provided in the abstract.
Among cognitively unimpaired participants, baseline plasma p-tau181, p-tau217, and Aβ42/Aβ40 measured by the Lumipulse assay were associated with incident MCI/dementia due to AD but were not associated with incident non-AD etiologies. In contrast, GFAP was associated with incident MCI/dementia due to both AD and non-AD causes. The abstract does not report whether NfL was associated with incident clinical outcomes.
All four biomarkers highlighted in results were associated with longitudinal change in the SPARE-AD score, indicating relationships between plasma measures and progression of AD-like brain atrophy patterns over time. The study did not observe associations between plasma biomarkers and the SPARE-BA index that captures brain aging, suggesting specificity for AD-like atrophy rather than general brain aging in this cohort.
When regional patterns were examined, both Aβ42/Aβ40 (Quanterix and Lumipulse), p-tau181, and p-tau217 were associated with longitudinal atrophy in parieto-temporal regions. p-tau217 was unique among the measured biomarkers in being associated with longitudinal atrophy of the medial temporal lobe, a region central to early AD pathology.
Plasma biomarkers were not associated with SPARE-BA (the brain age pattern). Nor were associations observed with the R-indices that captured subcortical, diffuse cortical, or perisylvian atrophy patterns. These null findings suggest the measured plasma biomarkers may preferentially reflect AD-like cortical patterns rather than other regional or generalized atrophy dimensions in this sample.
In this cognitively unimpaired BLSA sample, p-tau217 showed the most extensive associations with later clinical AD and with longitudinal AD-related brain atrophy, including both parieto-temporal and medial temporal changes. p-tau181 and Aβ42/Aβ40 also tracked parieto-temporal atrophy and were predictive of incident AD when measured in the preclinical state. GFAP demonstrated broader associations with incident cognitive impairment from both AD and non-AD causes, suggesting it may reflect more general neurodegenerative or neuroinflammatory processes.
These results support the view that certain plasma biomarkers, particularly p-tau217, relate to the specific neuroanatomical progression of AD pathology and could have prognostic value when measured in cognitively unimpaired individuals.
The abstract does not provide detailed effect sizes, statistical metrics, or sensitivity analyses; those details are not reported here. The work is a preprint and has not been peer reviewed, and the authors note that findings should not be used to guide clinical practice. BLSA data are available on request from the BLSA data repository, subject to data transfer agreements and review by the Data Sharing Proposal Review Committee.
The Intramural Research Program of the National Institutes of Health funded the work. Ethical approval was provided by the NIH IRB, and participants gave written informed consent at each visit. Competing interests reported by authors are documented in the source.
Measured in cognitively unimpaired adults, plasma p-tau217 most consistently identified individuals at risk of subsequent AD-related clinical decline and tracked longitudinal AD-like regional brain atrophy, with additional but more limited correspondence for p-tau181 and Aβ42/Aβ40. GFAP associated with incident impairment across etiologies. Further peer-reviewed reporting is needed for full methodological and statistical details.