# Vascular pathology drives the association of plasma GFAP with gray matter atrophy and cognitive decline in amyloid‐negative individuals

## TL;DR

This study shows that in people without amyloid issues, high vascular problems link plasma GFAP to brain shrinkage and worse memory, especially with high blood pressure and diabetes.

## Contribution

The study reveals vascular pathology, not amyloid, drives GFAP's link to brain atrophy and cognitive decline in amyloid-negative individuals.

## Key findings

- Plasma GFAP is associated with hippocampal atrophy and cognitive decline in individuals with high vascular burden.
- Hypertension and diabetes strongly contribute to the GFAP-neurodegeneration link.
- No significant associations were found in individuals with low vascular burden.

## Abstract

Amyloid (Aβ) pathology potentiates the association between GFAP, a biomarker for reactive astrogliosis, and neurodegeneration, while it remains unclear whether GFAP is associated with neurodegeneration without Aβ abnormalities. Preclinical studies suggest vascular pathology may activate glial cells, triggering deleterious effects. Here, we tested the hypothesis that, similar to Aβ pathology, vascular pathology may also potentiate the effects of plasma GFAP on neurodegeneration and cognitive decline in Aβ‐negative individuals.

We assessed 324 cognitively impaired (CDR < 0) Aβ‐negative (Centiloid < 24) participants from a memory clinic cohort (BICWALZS), with available CDR global, MMSE, plasma GFAP, clinical assessment of peripheral vascular risk factors [PVPs: hypertension (HTN), diabetes (DBT), and dyslipidemia (DLP)], FLAIR and T1‐based volumetrics. Participants were divided into two groups according to their WMH status: Fazekas 1 (Fz1; n = 203) and Fazekas 2‐3 (Fz2‐3; n =  121). Group differences were analyzed using ANCOVA. Associations were assessed using linear regressions, and the contribution of PVPs to the effects of biomarkers was accessed through multicollinearity analysis accounting for age, sex, and years of education.

Individuals in the Fz2‐3 group showed more hippocampal atrophy (Figure 1A, 1B; p = 0.0115), plasma NfL levels (1C; p = 0.0020), no changes in plasma GFAP levels (Figure 1D) and lower MMSE score (Figure 1E). In the Fz2‐3 group, hippocampal atrophy (Table 1; β: ‐0.179, p = 0.0440) and cognitive decline (Table 1; β: ‐0.194, p = 0.0390) was associated with plasma GFAP. No abnormalities or associations between biomarkers were found in the Fz1 group. Among the PVPs evaluated, HTN and DBT presented a stronger effect in the association of plasma GFAP in hippocampal atrophy (R2: 0.2089; p = 0.0001) and cognitive decline (R2: 0.1707; p < 0.0001), respectively (Figure 3).

Our study found that plasma GFAP is strongly linked to neurodegeneration and MMSE decline in Aβ‐negative individuals with high vascular burden. HTN and DBT, prevalent in the elderly, were the main contributors. This highlights vascular pathology as a key driver of neuroinflammation‐related neurodegeneration, underscoring the importance of managing these conditions to prevent brain atrophy.