# Plasma NfL and GFAP in the preclinical stages of neurodegenerative diseases: insights from the UK Biobank

## TL;DR

Blood markers NfL and GFAP can predict early signs of neurodegenerative diseases like Parkinson's and Alzheimer's before diagnosis.

## Contribution

The study reveals distinct pre-diagnostic profiles of plasma NfL and GFAP for various neurodegenerative diseases using UK Biobank data.

## Key findings

- Plasma NfL is linked to higher risk of ALS, APD, AD, and PD before diagnosis.
- Plasma GFAP is specifically associated with increased risk of AD.
- NfL and GFAP show different patterns in predicting neurodegenerative diseases.

## Abstract

Plasma neurofilament light chain (NfL) and glial fibrillary acidic protein (GFAP) are promising blood-based biomarkers of neuroaxonal injury and astrocytic activation, relevant in neurodegeneration. We investigated their pre-diagnostic profiles across common neurodegenerative diseases, including Parkinson’s disease (PD), atypical parkinsonian disorders (APD), Alzheimer’s disease (AD), and amyotrophic lateral sclerosis (ALS).

Forty-eight thousand five hundred and twenty-four UK Biobank participants with baseline plasma proteomic data for NfL and GFAP were included. Incident diagnoses of PD, APD, AD, and ALS were identified through ICD-10—coded health records, after careful inclusion/exclusion procedures. Baseline plasma NfL and GFAP concentrations were standardized as z-scores adjusted for relevant covariates. The hazard ratio (HR) for incident neurodegenerative diseases was estimated using Cox regression models adjusted for demographic, clinical, socioeconomic and genetic factors.

The final sample included 1196 cases (505 PD, 26 APD, 476 AD, 189 ALS) and 44,107 control subjects. In Cox regression models, NfL was associated with higher risk of incident ALS (HR 1.69; 95% CI, 1.58–1.80, p < 0.001), APD (HR 1.51; 95% CI, 1.22–1.87, p < 0.001), AD (HR 1.31; 95% CI, 1.22–1.41, p < 0.001), and PD (HR 1.14; 95% CI, 1.05–1.23, p < 0.001). GFAP was independently associated with incident AD only (HR 1.72; 95% CI, 1.63–1.82, p < 0.001).

Plasma NfL and GFAP showed distinct pre-diagnostic profiles. NfL was associated with incident diagnosis of several neurodegenerative diseases, while GFAP was specific to AD, reinforcing its role in dementia. These results may help optimize the identification of target populations at risk of neurodegenerative diseases for future neuroprotective treatments.

The online version contains supplementary material available at 10.1007/s00415-025-13498-y.