Summary
A medRxiv preprint reports that plasma p-tau217 showed the broadest association with later Alzheimer’s-related brain atrophy among cognitively unimpaired participants. The observational analysis included 818 people and 2,293 longitudinal brain MRI scans.
A study of 818 cognitively unimpaired participants has linked several blood biomarkers with later changes in brain-atrophy patterns associated with Alzheimer’s disease. Among the markers examined, plasma phosphorylated tau 217, or p-tau217, had the most extensive association with longitudinal Alzheimer’s-related atrophy, including changes in the medial temporal lobe.
The findings come from a preprint posted on medRxiv on September 17, 2026. The analysis used data from the Baltimore Longitudinal Study of Aging and examined associations rather than testing an assigned treatment or intervention.
Study design and brain measures
The researchers measured five plasma biomarkers while participants were cognitively unimpaired: the amyloid-beta 42/40 ratio, glial fibrillary acidic protein (GFAP), neurofilament light chain (NfL), p-tau181 and p-tau217. These blood measurements were taken at an index, or baseline, visit.
During follow-up, 104 participants developed mild cognitive impairment or dementia. Of these, 74 cases were attributed to Alzheimer’s disease, 24 to non-Alzheimer’s causes and six had an unknown cause. The researchers also analysed 2,293 brain MRI scans collected over time.
The MRI analysis used several multidimensional measures. SPARE-BA represented brain-age-related atrophy, while SPARE-AD represented an Alzheimer’s-like atrophy pattern. Five additional R-indices captured major dimensions of structural loss, including subcortical, diffuse cortical and perisylvian patterns. Statistical models were used to examine whether baseline biomarker levels were associated with later clinical conversion and whether they were related to changes in these MRI-derived patterns.
What the biomarkers were associated with
P-tau181, p-tau217 and the amyloid-beta 42/40 ratio measured with the Lumipulse assay were associated with incident mild cognitive impairment or dementia due to Alzheimer’s disease. The study reported no corresponding association with non-Alzheimer’s causes for these markers. GFAP was associated with incident impairment or dementia attributed to both Alzheimer’s and non-Alzheimer’s causes, suggesting a less disease-specific pattern in this cohort.
All four of those biomarkers were associated with longitudinal change in the SPARE-AD measure. The amyloid-beta 42/40 ratio, p-tau181 and p-tau217 were also associated with progressive parietotemporal atrophy, a pattern involving regions toward the upper and side portions of the brain that are commonly examined in Alzheimer’s research.
P-tau217 was the only biomarker in the analysis associated with longitudinal medial temporal lobe atrophy. This region includes structures important for memory, and its involvement is a central feature of the structural changes studied in Alzheimer’s disease. The authors therefore identified p-tau217 as the marker with the most extensive relationship to the Alzheimer’s-related MRI patterns examined.
The analysis did not find associations between the plasma biomarkers and SPARE-BA, the brain-age measure, or the R-indices representing subcortical, diffuse cortical or perisylvian atrophy. This distinction is important because the study was examining whether the biomarkers tracked Alzheimer’s-like change more specifically than general ageing-related brain loss.
The results are preliminary because the report is a medRxiv preprint and the study is observational. The associations show how baseline blood measurements related to later clinical outcomes and MRI changes in this cohort; they do not establish that any biomarker causes brain atrophy. The supplied abstract also does not report effect sizes, clinical thresholds or diagnostic accuracy measures, so the findings describe a biological relationship rather than a validated standalone blood test for individual clinical use.