Premium
P1‐173: Comparison of Different Cerebrospinal Fluid Biomarkers for Differential Diagnosis in Neurodegenerative Disease: A Critical Analysis
Author(s) -
Jansen Kari-Maria,
Lange Peter,
Schmidt Oliver,
Zerr Inga
Publication year - 2016
Publication title -
alzheimer's and dementia
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 6.713
H-Index - 118
eISSN - 1552-5279
pISSN - 1552-5260
DOI - 10.1016/j.jalz.2016.06.921
Subject(s) - cerebrospinal fluid , neurology , neurochemistry , disease , medicine , differential diagnosis , context (archaeology) , dementia , lumbar puncture , pathology , psychiatry , biology , paleontology
Background: There are emerging cerebrospinal fluid (CSF) biomarker candidates detecting brain synaptic dysfunction and neuronal damage that may help improve Alzheimer’s disease (AD) detection/diagnosis. CSF concentrations of neurogranin and neurofilament light (NFL) polypeptide are candidate biomarkers of synaptic degeneration and large-caliber myelinated axonal loss, respectively. In addition to the core biomarker Ab1-42 we analyzed the combined added diagnostic value of CSF neurogranin and NFL for classifying MCI, prodromal and AD dementia (ADD) against frontotemporal dementia (FTD) and healthy controls (HC). Methods:Multicenter study of CSF neurogranin, NFL, and Ab1-42 concentrations were analyzed in ADD (n1⁄441), amnestic MCI (aMCI) (n1⁄411), non-amnestic MCI (n1⁄414), prodromal AD (n1⁄412), FTD (n1⁄49), and HC (n1⁄425). Both neurogranin (in-house assay, LLOQ1⁄4125 pg/mL) and NFL concentrations were measured using ELISA (UmanDiagnostics; Ume a, Sweden; LLOD1⁄450 ng/ L). Groupwise effects were analysed using non-parametric Kruskal-Wallis tests. Relative risks (RRs) versus the HC reference category were computed using multinomial logistic regression after adjustment for age and sex as nuisance covariates. Classification accuracy was evaluated as ROC AUC in the model trained with a biomarker of interest versus the combination of the three biomarkers. Results: Median groupwise values were ADD>aMCI>prodromal AD>non-amnestic MCI>HC>FTD (neurogranin), ADD>non-amnestic MCI>aMCI>prodromal AD>FTD>HC (NFL), and HC>non-amnestic MCI>FTD> aMCI>ADD>prodromal AD (Ab1-42). An increase by 1 pg/mL in Ab1-42 was associated with a RR of -0.726% (p1⁄40.023551) for MCI, -0.556% (p1⁄40.029895) for non-amnestic MCI, -1.466% (p1⁄40.000025) for ADD, -0.725% (p1⁄40.020818) for FTD, -1.302% (p1⁄40.000397) for prodromal AD, whereas an increase by 1 pg/mL in NFL was associated with a RR of 0.354% (p1⁄40.005232) for non-amnestic MCI, 0.414% (p1⁄40.001742) for ADD, 0.279% (p1⁄40.032262) for FTD, and an increase by 1 pg/ mL in neurogranin was associated with a RR of 1.996% (p1⁄40.000098) for aMCI, 0.893% (p1⁄40.034684) for non-amnestic MCI, 1.752% (p1⁄40.000192) for ADD, 1.797% (p1⁄40.000201) for prodromal AD. ROC AUC values (“neurogranin only”/“Ab1-42 only”/“NFL only”/“All three biomarkers”) were: 0.805/0.83/ 0.819/0.912 (ADD), 0.883/0.847/0.873/0.89 (aMCI), 0.762/0.735/ 0.665/0.821 (non-amnestic MCI), 0.745/0.718/0.733/0.793 (prodromal AD), 0.814/0.75/0.74/0.886 (FTD), 0.858/0.926/0.919/ 0.979 (HC). Conclusions: We demonstrated that altered CSF Ab1-42, neurogranin and NFL concentrations were independently associated with RR. All three biomarkers combined showed a statistically significant improved diagnostic performance compared to the application of the single biomarkers.