Cross‐Sectional Analysis of Plasma GFAP Across Alzheimer's Disease Stages

John R. Hoffman, Kristen A. Russ, Alex J. Schwefel, Tatiana M. Foroud, Naazneen Khan, Yi Zhao, Jeffrey L. DageAlzheimer 2025 · DOI

TL;DR

This study investigates plasma Glial Fibrillary Acidic Protein (GFAP) as a potential biomarker for Alzheimer's Disease (AD). They found that GFAP levels are significantly associated with AD-related factors like amyloid PET, hippocampal atrophy, and cognitive status, reinforcing its promise as an early diagnostic marker. Importantly, they highlight that sex differences in GFAP levels are significant and must be considered in future AD biomarker research.

What Problem Does This Paper Address?

Alzheimer's Disease (AD) is a devastating neurodegenerative disorder, and identifying reliable biomarkers for early detection and monitoring is crucial. Glial acidic fibrillary protein (GFAP) is a protein found in astrocytes, which are activated in response to AD pathology, particularly amyloid-beta plaques. Previous research has shown GFAP levels increase in both cerebrospinal fluid (CSF) and plasma as AD progresses, suggesting its potential as an early biomarker. However, GFAP levels can be influenced by various factors beyond AD itself, such as age, sex, and other medical conditions. This variability makes it challenging to interpret GFAP measurements accurately. This paper aims to systematically explore these influencing factors to better understand how plasma GFAP levels vary across different demographic and clinical characteristics, thereby refining its utility as an AD biomarker.

Methods

The researchers utilized data from the National Alzheimer's Coordinating Center (NACC) Uniform Data Set (UDS) Version 3.0, a rich longitudinal dataset from Alzheimer's Disease Centers (ADCs). Plasma GFAP levels were measured using the Quanterix Simoa-HDx platform. To analyze the data, GFAP levels were log10-transformed and z-scored. The study employed a cross-sectional analysis on 3,456 unique subjects aged 60 and older. Statistical methods included t-tests, one-way ANOVAs, Tukey's HSD post-hoc tests, and linear regression to examine associations between GFAP levels and various factors, including diagnosis, pathology (amyloid PET, hippocampal atrophy), demographics (age, sex, race, ethnicity), and APOE genotype.

Datasets

The primary dataset used in this study is the National Alzheimer's Coordinating Center (NACC) Uniform Data Set (UDS) Version 3.0. This dataset contains longitudinal subject data from 23 Alzheimer's Disease Centers (ADCs). The GFAP measurements were performed by the National Centralized Repository for Alzheimer's Disease and Related Dementias (NCRAD). No new datasets were introduced.

Key Insights and Ideas

  • GFAP as a Promising AD Biomarker: The study reinforces that plasma GFAP is strongly associated with core AD pathologies (amyloid PET positivity) and structural changes (hippocampal atrophy), as well as cognitive status. This strengthens its case as a valuable biomarker for AD diagnosis, potentially even in its early stages.
  • The "Sex Difference" Revelation: A critical finding is the significant difference in plasma GFAP levels between males and females. This is a crucial consideration that has likely been overlooked or underappreciated in previous studies, and it emphasizes the need to stratify analyses or adjust for sex when using GFAP as a biomarker.
  • Age is a Major Player: Age is a significant confounder for GFAP levels, as expected. However, the study found that while age is associated with GFAP, the relationships between GFAP and other AD-related factors (like amyloid PET or atrophy) were not significantly altered when age was accounted for. This suggests that GFAP's association with AD pathology is somewhat independent of its age-related fluctuations.

Key Results

  • Plasma GFAP levels showed significant associations with age, sex, APOE genotype, elevated amyloid PET, hippocampal atrophy, and cognitive status (all p < 0.05).
  • The non-demographic associations (e.g., with amyloid PET, atrophy) remained significant even after adjusting for age.
  • Plasma GFAP levels were significantly different between males and females.

Limitations and Future Work

Limitations: * Cross-sectional Design: The study is cross-sectional, meaning it captures a snapshot in time. This limits the ability to establish causality or track changes in GFAP over time in relation to disease progression. * Potential for Unmeasured Confounders: While many factors were analyzed, there might be other unmeasured comorbidities or lifestyle factors that influence GFAP levels.

Future Work: * The authors suggest future efforts should focus on exploring the impact of changes in GFAP over time on changes in atrophy and cognitive decline. * Investigating how these relationships are modified by comorbid conditions is also highlighted as an important next step. * Further research is needed to understand the biological mechanisms underlying the observed sex differences in GFAP levels.

Summary updated

Abstract

Abstract Background Alzheimer's Disease (AD) is the leading cause of dementia worldwide, with Late‐Onset Alzheimer's Disease (LOAD) comprising over 90% of cases. Multiple biomarkers are being investigated for their association with pathologies associated with AD including glial acidic fibrillary protein (GFAP). GFAP levels have been observed to increase in plasma and cerebrospinal fluid of AD patients, particularly in response to amyloid‐β, before behavioral symptoms appear. This indicates GFAP's utility as an early‐AD biomarker. However, GFAP levels vary significantly between subjects due to various factors, such as demographics or medical comorbidities. To use plasma GFAP levels as an accurate biomarker in AD studies, it is essential to fully understand and identify all potential reasons for variation in its levels. Method The uniform data set (UDS), which contains longitudinal subject data from AD Centers (ADC) was obtained from the National Alzheimer's Coordinating Center (NACC). Total GFAP levels from 23 ADC's were measured by the National Centralized Repository for Alzheimer's Disease and Related Dementias (NCRAD) using Quanterix Simoa‐HDx (N4PE and N2PB). Results were log10‐transformed and z‐scored for analysis. A total of 3,456 unique subjects at a single visit, aged 60+, were included in this cross‐sectional analysis. T‐tests, one‐way ANOVAs, Tukey's‐ HSD Post‐Hoc and linear regression were used to examine the association of GFAP levels with diagnosis, pathology, demographics and comorbid conditions. Result Analysis showed that GFAP levels varied significantly ( p ‐value &lt; 0.05) with age in addition to sex, APOE genotype, elevated amyloid PET, hippocampal atrophy, and cognitive status. The non‐demographic associations were not significantly modified when adjusting for the subjects' age despite a significant association of GFAP levels with age overall. Conclusion Plasma GFAP levels differed significantly between males and females, emphasizing the need to account for sex differences in biomarker research. Plasma GFAP levels show strong associations with risk alleles, elevated amyloid PET levels, hippocampal atrophy, and cognitive status, supporting its potential as a biomarker for AD diagnosis. Future efforts will focus on exploring the impact of changes in GFAP on changes in atrophy, cognitive decline, and how these relationships are modified by comorbid conditions.
BibTeX
@inproceedings{hoffman2025crosssectional,
  author = {John R. Hoffman and Kristen A. Russ and Alex J. Schwefel and Tatiana M. Foroud and Naazneen Khan and Yi Zhao and Jeffrey L. Dage},
  title = {{Cross‐Sectional Analysis of Plasma GFAP Across Alzheimer's Disease Stages}},
  booktitle = {Alzheimer},
  year = {2025},
  doi = {10.1002/alz70856\_106901},
}
Figure