Diagnostic value of plasma p-tau181, NfL, and GFAP in a clinical setting cohort of prevalent neurodegenerative dementias

Abstract

Background

Increasing evidence supports the use of plasma biomarkers of neurodegeneration and neuroinflammation to screen and diagnose patients with dementia. However, confirmatory studies are required to demonstrate their usefulness in the clinical setting.

Methods

We evaluated plasma and cerebrospinal fluid (CSF) samples from consecutive patients with frontotemporal dementia (FTD) (n = 59), progressive supranuclear palsy (PSP) (n = 31), corticobasal syndrome (CBS) (n = 29), dementia with Lewy bodies (DLB) (n = 49), Alzheimer disease (AD) (n = 97), and suspected non-AD physiopathology (n = 51), as well as plasma samples from 60 healthy controls (HC). We measured neurofilament light chain (NfL), phospho-tau181 (p-tau181), and glial fibrillary acid protein (GFAP) using Simoa (all plasma biomarkers and CSF GFAP), CLEIA (CSF p-tau181), and ELISA (CSF NfL) assays. Additionally, we stratified patients according to the A/T/N classification scheme and the CSF α-synuclein real-time quaking-induced conversion assay (RT-QuIC) results.

Results

We found good correlations between CSF and plasma biomarkers for NfL (rho = 0.668, p < 0.001) and p-tau181 (rho = 0.619, p < 0.001). Plasma NfL was significantly higher in disease groups than in HC and showed a greater increase in FTD than in AD [44.9 (28.1–68.6) vs. 21.9 (17.0–27.9) pg/ml, p < 0.001]. Conversely, plasma p-tau181 and GFAP levels were significantly higher in AD than in FTD [3.2 (2.4–4.3) vs. 1.1 (0.7–1.6) pg/ml, p < 0.001; 404.7 (279.7–503.0) vs. 198.2 (143.9–316.8) pg/ml, p < 0.001]. GFAP also allowed discriminating disease groups from HC. In the distinction between FTD and AD, plasma p-tau181 showed better accuracy (AUC 0.964) than NfL (AUC 0.791) and GFAP (AUC 0.818). In DLB and CBS, CSF amyloid positive (A+) subjects had higher plasma p-tau181 and GFAP levels than A− individuals. CSF RT-QuIC showed positive α-synuclein seeding activity in 96% DLB and 15% AD patients with no differences in plasma biomarker levels in those stratified by RT-QuIC result.

Conclusions

In a single-center clinical cohort, we confirm the high diagnostic value of plasma p-tau181 for distinguishing FTD from AD and plasma NfL for discriminating degenerative dementias from HC. Plasma GFAP alone differentiates AD from FTD and neurodegenerative dementias from HC but with lower accuracy than p-tau181 and NfL. In CBS and DLB, plasma p-tau181 and GFAP levels are significantly influenced by beta-amyloid pathology.