P3-435: plasma neurofilament light as a measure of neurodegeneration for nondemented individuals

Hao Hu,Jin-Tai Yu
DOI: https://doi.org/10.1016/j.jalz.2019.06.3469
2019-01-01
Alzheimer s & Dementia
Abstract:Neurofilament light (NFL) is a putative marker of large-caliber axonal degeneration which is an important pathological change in neurodegeneration diseases including Alzheimer disease (AD). We investigated the pathological changes of plasma NFL in a nondemented cohort classified by biomarker-based framework and compared the predictive roles of plasma NFL with established neurodegenerative biomarkers. Data used in this study were obtained from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database. Firstly, 526 nondemented participants were included in a biomarker-based framework (A-T-N-, A+T-N-, A+T+N-, A+T+N+). Intergroup differences were calculated by Kruskal-Wallis test, chi square test and one-way analysis of variance. Associations between diagnosis, CSF biomarkers, MRI measures, cognition and plasma NFL were tested by multiple linear regression models for baseline data and linear mixed-effects models (LME) for longitudinal data. Secondly we estimated the temporal ordering of plasma NFL abnormalities compared with other biomarkers by LME. Finally, we compared the predictive roles of plasma NFL with established neurodegenerative biomarkers using LME and cox proportional hazard regression models in a cohort classified by clinical phenotype and amyloid deposition (A). Plasma NFL concentrations were higher in more advanced ATN stages (A+T+N-, A+T+N+) of a nondemented cohort and had associations with baseline and longitudinal data of cognition and hippocampal volume in advanced ATN stages (A+T+N-, A+T+N+). Plasma NFL concentrations were correlated with CSF tTau and pTau181 but not with CSF Aβ42. The temporal ordering of plasma NFL abnormalities was similar to neurodegenerative biomarkers rather than to Aβ. Plasma NFL had predictive roles on neurodegeneration for nondemented individuals in the similar range as some established neurodegenerative biomarkers (Hippocampus volume, CSF tTau, FDG-PET).
What problem does this paper attempt to address?