A pilot study of the utility of cerebrospinal fluid neurofilament light chain in differentiating neurodegenerative from psychiatric disorders: A 'C-reactive protein' for psychiatrists and neurologists?

Abstract

Neurofilament light has shown promise as a biomarker for diagnosis, staging and prognosis in a wide range of neurological and neurodegenerative disorders. This study explored the utility of cerebrospinal fluid neurofilament light in distinguishing primary psychiatric disorders from neurodegenerative and neurological disorders, a common diagnostic dilemma for psychiatrists and neurologists. This cross-sectional retrospective pilot study assessed cerebrospinal fluid neurofilament light on patients referred to a tertiary neuropsychiatry service from 2009 to 2017 for diagnostic assessment of neuropsychiatric and neurocognitive symptoms, where a neurodegenerative disorder was a differential diagnosis, who received lumbar punctures as part of a comprehensive workup. The most recent gold-standard clinical consensus diagnosis was categorised into psychiatric disorder or neurodegenerative or neurological disorder. Data from healthy controls were available for comparison. Data extraction and diagnostic categorisation was blinded to neurofilament light results. A total of 129 participants were included: 77 neurodegenerative or neurological disorder (mean age 57 years, including Alzheimer's dementia, frontotemporal dementia), 31 psychiatric disorder (mean age 51 years, including schizophrenia, major depressive disorder) and 21 healthy controls (mean age 66 years). Neurofilament light was significantly higher in neurodegenerative or neurological disorder (M = 3560 pg/mL, 95% confidence intervals = [2918, 4601]) compared to psychiatric disorder (M = 949 pg/mL, 95% confidence intervals = [830, 1108]) and controls (M = 1036 pg/mL, 95% confidence intervals = [908, 1165]). Neurofilament light distinguished neurodegenerative or neurological disorder from psychiatric disorder with an area under the curve of 0.94 (95% confidence intervals = [0.89, 0.98]); a cut-off of 1332 pg/mL was associated with 87% sensitivity and 90% specificity. Cerebrospinal fluid neurofilament light shows promise as a diagnostic test to assist with the often challenging diagnostic dilemma of distinguishing psychiatric disorders from neurodegenerative and neurological disorders. Further studies are warranted to replicate and expand on these findings, including on plasma neurofilament light.