Abstract
Background: Symptoms of behavioral variant frontotemporal dementia (bvFTD) overlap with primary psychiatric disorders (PPD) making diagnosis challenging. Serum neurofilament light (sNfL) is a candidate biomarker to distinguish bvFTD from PPD, but large-scale studies in PPD are lacking.
Objective: Determine factors that influence sNfL from a large database of PPD patients, and test its diagnostic accuracy.
Design, settings, subjects, measurements: Clinical data of people aged 40-81 were obtained from healthy subjects (n = 69), and patients with PPD (n = 848) or bvFTD (n = 82). sNfL was measured using Simoa technology on an HD-X instrument. Data were analyzed using general linear models, and Receiver Operating Characteristic (ROC) curve analyses to determine global and age-specific sNfL cutoffs to distinguish bvFTD from PPD, using the Youden Index.
Results: sNfL increased with age, while sex, BMI and diabetes status were modestly associated with sNfL. sNfL was slightly higher in PPD than healthy subjects (14.1 versus 11.7 pg/mL), when controlling for covariates. sNfL was markedly lower in PPD than bvFTD (14.1 versus 44.1 pg/mL). sNfL could differentiate PPD from bvFTD with an AUC = 0.868, but the effect was driven by the younger subjects between age 40-60 years at a cutoff of 16.0 pg/mL. No valid cutoff was detected over age 60, however, values of sNfL above 38.5 pg/mL, or below 13.9 pg/mL, provided 90% diagnostic certainty of bvFTD or PPD, respectively.
Conclusion: PPD have mildly elevated sNfL compared to healthy subjects but much lower than bvFTD. Results support the use of sNfL as a biomarker to differentiate PPD from bvFTD at age 60 or below, but accuracy decreases in older ages.
Plain Language Summary
Introduction: There is a type of early-onset dementia (behavioural variant of Frontotemporal Dementia) that is similar to Primary Psychiatric Disorders. Some of the behavior and personality aspects are similar and this means that mistakes can be made during diagnosis. When the diagnosis is incorrect, the patient will receive care that is not appropriate for them. We need a better way to tell these conditions apart.
There is a biological marker called serum neurofilament light which can be found in the blood. Serum neurofilament light has the potential to give more information that will help to tell apart the behavioural variant of Frontotemporal Dementia from Primary Psychiatric Disorders. However, this was not tested on a large-scale yet.
Methodology: We obtained clinical data of patients with Primary Psychiatric Disorders and healthy participants from Signature Biobank. Data on patients with behavioural variant of Frontotemporal Dementia were obtained from both Amsterdam Medical Center and the Comprehensive Assessment of Neurodegeneration and Dementia study. Participants were between the ages 40-81. We used statistical methods to determine factors that influence serum neurofilament light and to test for group differences in serum neurofilament light between psychiatric patients and the other participants.
Results: Age, sex, Body Mass Index, and diabetes status were associated with serum neurofilament. Serum neurofilament light was slightly higher in the psychiatric patient group (all diagnoses) compared to healthy participants. Behavioural variant of Frontotemporal Dementia patients had higher serum neurofilament light levels compared to psychiatric patients. Based on this finding a clinical cutoff of serum neurofilament light was determined to differentiate patients with the behavioural variant of Frontotemporal Dementia from those with Primary Psychiatric Disorders.
Conclusion: Serum neurofilament light values are slightly higher in psychiatric patients compared to controls, while behavioural variant of Frontotemporal Dementia patients had even higher serum neurofilament light in comparison to psychiatric patients. Results support the use of serum neurofilament light as a biomarker to accurately differentiate Primary Psychiatric Disorders from the behavioural variant of Frontotemporal Dementia to improve early diagnosis.