Utility of Plasma Neurofilament Light in the 1Florida Alzheimer's Disease Research Center (ADRC)

Abstract

Background: Plasma NfL (pNfL) levels are elevated in many neurological disorders. However, the utility of pNfL in a clinical setting has not been established.

Objective: In a cohort of diverse older participants, we examined: 1) the association of pNfL to age, sex, Hispanic ethnicity, diagnosis, and structural and amyloid imaging biomarkers; and 2) its association to baseline and longitudinal cognitive and functional performance.

Methods: 309 subjects were classified at baseline as cognitively normal (CN) or with cognitive impairment. Most subjects had structural MRI and amyloid PET scans. The most frequent etiological diagnosis was Alzheimer's disease (AD), but other neurological and neuropsychiatric disorders were also represented. We assessed the relationship of pNfL to cognitive and functional status, primary etiology, imaging biomarkers, and to cognitive and functional decline.

Results: pNfL increased with age, degree of hippocampal atrophy, and amyloid load, and was higher in females among CN subjects, but was not associated with Hispanic ethnicity. Compared to CN subjects, pNfL was elevated among those with AD or FTLD, but not those with neuropsychiatric or other disorders. Hippocampal atrophy, amyloid positivity and higher pNfL levels each added unique variance in predicting greater functional impairment on the CDR-SB at baseline. Higher baseline pNfL levels also predicted greater cognitive and functional decline after accounting for hippocampal atrophy and memory scores at baseline.

Conclusion: pNfL may have a complementary and supportive role to brain imaging and cognitive testing in a memory disorder evaluation, although its diagnostic sensitivity and specificity as a stand-alone measure is modest. In the absence of expensive neuroimaging tests, pNfL could be used for differentiating neurodegenerative disease from neuropsychiatric disorders.

Keywords: Alzheimer’s disease; amyloid; diagnosis; hippocampal atrophy; magnetic resonance imaging; plasma neurofilament light; positron emission tomography.

Fig. 1

Plasma neurofilament light by etiological diagnosis. Boxplots show the distribution of plasma neurofilament light by diagnostic category. FTLD, frontotemporal lobar degeneration.

Fig. 2

ROC Curves for Prediction of Decline on the CDR. Area under the receiver operating characteristic (AUROC) curve for prediction of a decline of 2 or more points on the CDR sum of boxes in cognitively impaired subjects over 2 follow-up visits (an average of 2.3 years), using baseline measures of Hopkins Verbal Learning Test-Revised (HVLT-R) total recall, plasma neurofilament Light (pNfL) and the combination of both HVLT-R and pNfL. The logistic regression models were adjusted for age and APOE ɛ4 genotype.