CSF Tau Is a Biomarker of Hippocampal Injury in Cryptogenic New-Onset Refractory Status Epilepticus

Abstract

Objective: Cryptogenic new-onset refractory status epilepticus (cNORSE) is a devastating condition characterized by the de novo onset of status epilepticus with unclear etiology. The identification of relevant early biomarkers in cNORSE is important to elucidate pathophysiology, aid clinical decision-making, and prognosticate outcomes in cNORSE.

Methods: CSF samples were obtained within 7 days of NORSE onset from an adult cNORSE cohort in a national referral center in South Korea. Nineteen patients with cNORSE were studied: 9 were male (47.4%) and the median age was 35.0 [IQR: 27.0-54.3] years. CSF from 21 patients with other neurological diseases (atypical parkinsonism, postural orthostatic hypotension syndrome, epilepsy, and cerebellar ataxia) was used as controls. Proteomic analysis was conducted using the Olink platform, and potential biomarker candidates were correlated with clinical data and MRI findings.

Results: Based on correlation analyses between proteomic data and clinical outcomes, total tau (t-tau) was selected as a potential biomarker. Patients with cNORSE had higher CSF t-tau levels than controls (p < 0.001). Early detection of high CSF t-tau was associated with the presence of hippocampal atrophy in the postacute phase of cNORSE (p = 0.044). The initial elevation of t-tau levels also correlated with a higher number of anti-seizure medications used (p = 0.031) and less improvement in Clinical Assessment Scale in Autoimmune Encephalitis (CASE) scores 1 month after NORSE onset (p = 0.066). T-tau levels were correlated with CSF pro-inflammatory cytokines/chemokines and mediators of neuronal damage.

Interpretation: Elevated CSF t-tau levels detected early after cNORSE onset may be a useful marker of initial brain injury and predict subsequent hippocampal atrophy.

Keywords: hippocampal atrophy; new onset refractory status epilepticus; tau.

FIGURE 1

Flowchart illustrating the process of selecting the cNORSE patients in the study cohort.

FIGURE 2

Selection of MAPT (t‐tau) as a biomarker. Heatmap representing adjusted correlation analysis between CSF candidate protein levels and clinical outcomes indicated that MAPT (t‐tau) and NMNAT1 were potential biomarkers (A). Violin plot of CSF t‐tau levels in patients with cNORSE and controls shows the significant elevation of early CSF t‐tau in cNORSE compared to controls (B). The dashed lines and dotted lines on the violin plot represent median values and interquartile ranges, respectively. Violin plot of CSF t‐tau levels showed a significant difference in patients with hippocampal atrophy compared those without. This did not hold true for NMNAT1 and hence MAPT was chosen as a biomarker for further analyses (C).

FIGURE 3

Clinical and radiological differences between patients with high and low t‐tau. Coronal images of patients with available interval MRIs at 2 months from NORSE onset (or next available interval scan after 2 months) (n = 17) are depicted, with yellow arrows showing the presence of hippocampal atrophy. cNORSE patients with high t‐tau had more frequent hippocampal atrophy on interval MRI. The numbers in blue represent the number of anti‐seizure medications (ASMs) used within the first month of NORSE onset, followed by the total number of ASMs used during clinical course. Heatmaps show each patient's seizure frequency at 4 weekly intervals for the first year, and missing seizure frequency data are represented in gray. cNORSE patients with high t‐tau had more frequent seizures.

FIGURE 4

Correlation analysis between CSF cytokines/chemokines and CSF t‐tau in cNORSE patients. Scatter plot depicting Spearman's rho on the x‐axis and corresponding p‐value on the y‐axis for each CSF cytokine/chemokine. The red dotted box contains cytokines/chemokines with p < 0.05 and rho > 0.5, which are enlarged on the top right panel for greater clarity.