Journal of Neurology (2023) 270:2128–2138

https://doi.org/10.1007/s00415-022-11547-4

ORIGINAL COMMUNICATION

Neurofilament light (NfL) as biomarker in serum and CSF in status

epilepticus
Nils G. Margraf1 · Justina Dargvainiene2 · Emily Theel1 · Frank Leypoldt1,2 · Wolfgang Lieb3 · Andre Franke4 ·
Klaus Berger5 · Jens Kuhle6,7 · Gregor Kuhlenbaeumer1

Received: 2 October 2022 / Revised: 22 December 2022 / Accepted: 23 December 2022 / Published online: 9 January 2023
© The Author(s) 2023

Abstract
Objective We explored the potential of neurofilament light chain (NfL) in serum and cerebrospinal fluid as a biomarker for
neurodestruction in status epilepticus.
Methods In a retrospective analysis, we measured NfL in serum and cerebrospinal fluid samples of patients with status
epilepticus using a highly sensitive single-molecule array technique (Simoa). Status epilepticus was diagnosed according to
ILAE criteria. Additionally, we employed an alternative classification with more emphasis on the course of status epilepticus.
We used data from three large control groups to compare NfL in status epilepticus versus neurologically healthy controls.
Results We included 28 patients (mean age: 69.4 years, SD: 15 years) with a median status duration of 44 h (IQR: 80 h).
Twenty-one patients (75%) suffered from convulsive status epilepticus and seven (25%) from non-convulsive status epi-
lepticus. Six patients died (21%). Cerebrospinal fluid and serum NfL concentrations showed a high correlation (r = 0.73,
p < 0.001, Pearson). The main determinant of NfL concentration was the status duration. NfL concentrations did not differ
between convulsive status epilepticus and convulsive status epilepticus classified according to the ILAE or to the alterna-
tive classification without and with adjusting for status duration and time between status onset and sampling. We found no
association of NfL concentration with death, treatment refractoriness, or prognostic scores.
Conclusion The results suggest that neurodestruction in status epilepticus measured by NfL is mainly determined by status
duration, not status type nor therapy refractoriness. Therefore, our results suggest that regarding neurodestruction convulsive
and non-convulsive status epilepticus are both neurological emergencies of comparable urgency.

Keywords Status epilepticus (SE) · Neurofilament light (NfL) · Biomarker · Cerebrospinal fluid (CSF) · Neurodestruction

Nils G. Margraf and Justina Dargvainiene have contributed equally

and share first authorship.

5
* Nils G. Margraf Institute of Epidemiology and Social Medicine, University
n.margraf@neurologie.uni-kiel.de of Münster, Münster, Germany
6
1 Multiple Sclerosis Centre, Neurology, Departments of Head,
Department of Neurology, University Hospital Schleswig-
Spine and Neuromedicine, Biomedicine and Clinical
Holstein (UKSH), Christian-Albrechts-University (CAU),
Research, University Hospital Basel and University of Basel,
Arnold‑Heller‑Str. 3, 24105 Kiel, Germany
Basel, Switzerland
2
Institute of Clinical Chemistry, University Hospital 7
Research Center for Clinical Neuroimmunology
Schleswig-Holstein (UKSH), Christian-Albrechts-University
and Neuroscience (RC2NB), University Hospital
(CAU), Kiel, Germany
and University of Basel, Basel, Switzerland
3
Institute of Epidemiology and Biobank PopGen,
University Hospital Schleswig-Holstein (UKSH),
Christian-Albrechts-University (CAU), Kiel, Germany
4
Institute of Clinical Molecular Biology,
Christian-Albrechts-University of Kiel and University
Hospital Schleswig-Holstein, Kiel, Germany

13
Vol:.(1234567890)
Journal of Neurology (2023) 270:2128–2138
Introduction
Status epilepticus (SE) is one of the most important and
common neurological emergencies. The incidence in
adults varies in different geographic regions, studies, and
age groups between 1.3 and 73.7/100,000/year adults with
subgroups of 80 years and older reaching an incidence
over 150/100,000 [18]. Immediate and consistent action
is required to terminate the ongoing epileptic activity
because mortality and disability in SE are high. Mortality
rates range from 16 to 25% in adults and increase with
patient age [3]. For refractory status epilepticus (RSE), the
reported fatality rates range from 16 to 39% in adults [19].
Novy, Logroscino and Rossetti reported that only 63% of
the survivors of SE and 21% of the survivors with RSE
recovered completely [25]. Experiments in primates show
that prolonged generalized convulsive seizures cause neu-
ronal damage [23]. However, if convulsive features of the
generalized seizures are suppressed by muscle relaxation
with curare, neuronal damage still occurs but is somewhat
milder, suggesting that generalized non-convulsive sei-
zures are also neurotoxic [24]. In animal models of strictly
focal SE, bilateral brain pathology and remote neuronal
damage are rare [5]. In summary, animal models of epi-
lepsy suggest (I) that severe neuronal damage is found in
brain regions with excessive neuronal discharges leading
to glutamatergic excitotoxicity, and (II) neuronal damage
is aggravated by physiological compromise associated
with convulsions [22]. However, several recent studies of
the treatment of SE show that in contrast to former studies
neurodestruction is comparable between convulsive and
non-convulsive SE (NCSE) [15, 27, 29]. Current studies
note that the mortality is even higher in NCSE than in con-
vulsive SE [15, 18]. This lead to fundamental discussions
about the treatment of convulsive SE and NCSE [29].
Neuron-specific enolase (NSE) is to date the most
promising neurodestruction biomarker in SE, but differ-
ent studies showed equivocal results concerning the cor-
relation with status duration and outcome [7, 8, 36, 37].
However, none of the biomarkers has yet been introduced
in clinical practice [9]. The restricted accessibility of cer-
ebrospinal fluid (CSF), especially for multiple follow-up
examinations limits the use of CSF biomarkers. A recently
developed highly sensitive single-molecule array (Simoa)
immunoassay allows reproducible and precise measure-
ments of brain-specific proteins, such as neurofilaments
[28]. Neuronal death leads to the release of neurofilaments
in CSF and blood. Neurofilament light chain (NfL) has
been investigated successfully as a biomarker of neurode-
struction in various neurological diseases [42, 43].
Here, we measured NfL in CSF and serum of patients
with SE using Simoa technology to explore three
2129
hypotheses: First, we asked whether the NfL concen-
trations in SE are increased in comparison to controls.
Second, we investigated if the NfL concentrations differ
between patients with convulsive SE and patients with
NCSE diagnosed by the ILAE classification and an alter-
native classification placing more attention on the course
of the SE. Third, we explored if NfL predicts progno-
sis and is associated with established predictive scales
(STESS, EMSE-EAC), therapy refractoriness, and death.
Materials and methods
Setting and study design
This study was conducted in an epilepsy center certi-
fied by the German Society for Epileptology (DGfE) at
the Department of Neurology of the University Medical
Center Schleswig–Holstein (UKSH) in northern Germany.
The UKSH, campus Kiel serves a catchment area of about
500,000 people. The study was approved by the Ethics Com-
mittee of the Medical Faculty of the Christian-Albrechts-
University, Kiel, Germany (D 480/20). The study was con-
ducted following the World Medical Association Declaration
of Helsinki. Anonymized data will be shared by request with
any qualified investigator.
Patients with SE
We identified patients with a diagnosis of SE between the
1st of February 2015 and the 15th of May 2020 through the
hospital information system of the UKSH. We included all
patients aged above 18 years with SE and severely impaired
vigilance with available CSF and serum samples in our
biobank. We identified a SE according to the International
League Against Epilepsy (ILAE) classification described
in the “Report of the ILAE Task Force on Classification
of SE” [40] and excluded absence status. In subsequent
sections, we refer to “convulsive SE” versus “NCSE” and
add “according to ILAE” if necessary. We allowed a purely
clinical diagnosis for convulsive SE but NCSE required in
addition to severely impaired vigilance one of the following
EEG phenomena: spike-and-wave complexes, generalized or
lateralized periodic discharges (GPDs/LEDs) with additional
characteristics according to Trinka and Leitinger [41]. The
patients with NCSE included in this study met the Salzburg
criteria for NCSE [16].
In addition, we classified patients according to criteria
developed by Leitinger and Trinka which take the evolu-
tion of semiology into account [18]. In short, patients with
a convulsive semiology throughout the SE are categorized
as group A, while patients with convulsive semiology at
the beginning and later switching to a non-convulsive
13
2130
semiology or purely non-convulsive semiology are in cat-
egories B and C which we summarized as one category.
We will refer to these categories “convulsive only” and
“NCSE at any point during evolution” and add “alterna-
tive classification” if necessary [18]. The SE duration
was inferred from the complete information available in
the hospital information system including the nursing
documentation.
We excluded patients with neurological diseases
known to lead to significant increases in NfL concentra-
tion, namely: brain tumor [10], ischemic or hemorrhagic
strokes, septicemia or posthypoxic encephalopathy within
the preceding 12 months, cerebral amyloid angiopathy,
head trauma, cerebral autoimmune diseases [14], infec-
tious meningoencephalitis [21], severe peripheral neurop-
athy [e.g., critical illness polyneuropathy before lumbar
puncture (LP)] [4]s, and severe neurodegenerative illness
[20, 26, 35], e.g., dementia with Lewy bodies (DLB), cor-
ticobasal degeneration (CBD), progressive supranuclear
palsy (PSP) and amyotrophic lateral sclerosis (ALS).
Controls
We obtained control measurements of three well-estab-
lished population-based cohorts. First, we obtained and
measured 71 control samples from the Popgen control
cohort, a longitudinal study of health in northern Germany
[13]. To expand the age range and enlarge the number of
control NfL measurements, we obtained 813 additional
control values from the population-based arm (cohort 3)
of the BiDirect Study [38] and 295 controls values from
the MEMO Study, an elderly community sample within the
population-based KORA project [32]. We excluded sam-
ples with a neurodegenerative disease but not samples with
a history of other age-related comorbidities like cardiovas-
cular diseases to avoid the generation of “supercontrols”.
The Popgen samples were measured concurrently with the
patient samples on the same instrument while BiDirect and
MEMO samples were measured on an identical but not the
same instrument using the same chemistry.
Clinical data
We extracted the following clinical data from the medical
records: age, sex, date and reason of hospital admission,
pre-existing diseases, medication during the hospital stay,
duration of SE, time from status onset to sampling, semiol-
ogy and course of SE, the patient’s outcome assessed in
prognostic scores and findings in EEG, cranial MRT and
CCT investigations.
13
Journal of Neurology (2023) 270:2128–2138
NfL measurements
We obtained all serum/CSF pairs at the same time point,
namely the time of LP. Polypropylene tubes were used
for CSF and serum. We used tubes with gel separators
for blood collection. The serum was separated by spin-
ning at 2000×g for 10 min after coagulation. All samples
were frozen at − 80 °C in polypropylene cryovials within
maximal 48 h at 4 °C. Samples were thawed only once
immediately before the measurements.
Analysis was performed on a fully automated HD-X
platform (Quanterix) based on Simoa assay technique
using a commercially available multiplex kit (Neurology
4-plex: NfL, Tau, UCHL-1, and GFAP). For each run, 8
calibrators were measured in duplicates and two controls
(low and high, included in the kit) were analyzed in sigli-
cates at the beginning and end of the run for continuous
process control. All samples were measured in duplicates.
The coefficients of variance between the replicates of 10%
for both serum and CSF were tolerated, by values exceed-
ing 10% the measurements were repeated. For analysis,
mean concentration of two replicates was estimated. The
laboratory personnel was blinded to the clinical history
and other laboratory parameters of the study participants.
We analyzed only the NfL measurements of the 4-plex
Quanterix kit because NfL is already a fairly well-estab-
lished biomarker [42, 43] with a large number of available
control cohorts and shows favorable properties concerning
short transfer time to CSF and blood, the long half-life,
and well-developed precise assay [28, 34]. Two patients
had missing CSF NfL values because samples had been
used up completely. The Shapiro–Wilk test for normal-
ity showed that the variables time between status onset
and lumbar puncture (“status onset time to LP”) and NfL
concentration in CSF as well as serum required logarith-
mic transformation to conform with the assumption of a
normal distribution.
Prognostic parameters
To estimate the severity of the SE, we used the follow-
ing outcome scores: SE Severity Score (STESS), modified
STESS (mSTESS), and Epidemiology-Based Mortality
Score in SE (EMSE) [17, 30, 31]. The STESS rates the
categories consciousness, worst seizure type, age, and his-
tory of previous seizures with points: 0–3 points predict-
ing a favorable and 4–6 points predicting an unfavorable
outcome. There are different versions of the EMSE: We
used the EMSE-EAC rating the patient’s etiology, age, and
each comorbidity, which leads to an open upwards point
scale with 27 points or more predicting high mortality.
Journal of Neurology (2023) 270:2128–2138
Statistics
We used R/RStudio (version 1.5.46) for all analyses—
commands in this paragraph shown in parentheses—and
assessed cross-tables using the Fisher exact test (fisher.
test). We used the Shapiro–Wilk (shapiro.test) test to assess
consistency with the assumption of a normal distribution
and calculated mean (mean) and standard deviation (SD)
for normally distributed variables. We compared normally
distributed variables between two groups with t tests (t.test)
or the linear model (lm) command equivalent to a t test.
For variables not following a normal distribution, we calcu-
lated median (median) and interquartile ranges (IQR) and
compared them between groups with the Mann–Whitney U
test (wilcox.test). Some variables were log-transformed to
achieve consistency with a normal distribution for statistics
requiring normally distributed variables, e.g., linear regres-
sion (lm) and correlation analysis (cor) using the Pearson
method. We calculated receiver operating characteristics
using the package pROC (version 1.18.0) and extracted the
areas under the curve (AUC). Boxplots display the median as
a horizontal line, the hinges indicate the interquartile range
(IQR), and the whiskers extend to the smallest/largest value
at most 1.5 * IQR of the hinges.
Results
Clinical data of the patients with SE
We included 28 patients (24 women, 4 men) with SE and
a mean age of 69.4 years (SD =  ± 15)). The median sta-
tus duration was 44 h (IQR 80 h) and 22 patients (79%)
survived SE. Patient characteristics including basic infor-
mation regarding NfL levels are listed in Table 1. Twenty-
one patients (75%) suffered from convulsive SE and seven
(25%) from NCSE according to the ILAE classification. All
patients showed impaired consciousness. Five patients had
epilepsy before the event [unclassified epilepsy n = 2 (7%),
structural epilepsy n = 3 (11%)]. Twelve of the remaining
patients had a potential structural cause of SE (43%) and
the etiology remained unknown for 11 patients (39%) and
6 of them fit the criteria of NORSE (new-onset refractory
status epilepticus) [11]. Fifteen patients (54%) had normal
standard parameters in the CSF investigation. Ten patients
showed mild to moderate brain–blood-barrier dysfunction,
two showed an inflammatory CSF, and one of them was
due to an isolated elevated cell count. In 21 patients, the
SE was refractory to treatment with a benzodiazepine and
a classic antiepileptic drug (RSE, 75%) [2]. Burst suppres-
sion was used in 4 patients, either by propofol alone (n = 2)
or by a combination of propofol and midazolam (n = 2). Six
patients (21%) died in the hospital. The reasons for death
2131
were respiratory failure, peritonitis, cardiac failure, pneumo-
nia, cardiogenic shock, and aspiration pneumonia.
NfL in status epilepticus versus control samples
We compared the serum NfL values in SE versus a large
set of controls (n = 1186, Fig. 1). The control samples cov-
ered an age range from 20 to 83 years. Figure 1 shows that
the NfL concentration of the Popgen, BiDirect, and MEMO
control samples are all in the same range taking age into
account. Therefore, we decided to pool all three control sam-
ples for analysis. Since NfL did not differ between sexes
(p = 0.47, t test), but increases with age, we fitted a linear
model with a quadratic term for serum NfL versus age and
determined the 95% prediction intervals and extrapolated to
the age of the oldest patient with SE (93 years) (Fig. 1). Only
four out of 28 patients with SE showed NfL concentrations
within the 95% NfL prediction interval for controls (Fig. 1).
In most patients with SE the NfL concentrations grossly
exceeded those of controls up to an approximately 100-fold
increase for the highest NfL value in a patient (2142 pg/ml).
Determinants of NfL concentrations in status
epilepticus
Table 2 shows the basic statistic measures presented in
Table 1 stratified by the two different classifications of SE
differentiating convulsive SE from NCSE and by survival.
CSF and serum NfL concentrations showed a high corre-
lation (r = 0.73, p =  < 0.001, Supplementary Fig. 1). Both,
NfL in CSF and serum were correlated with “status dura-
tion” (CSF: r = 0.58, p = 0.002; serum: r = 0.59, p = 0.001)
and “status onset time to LP" (CSF: r = 0.64, p < 0.0004;
serum: r = 0.55, p = 0.003, Fig. 2). Almost identical slopes
of the regression lines for CSF and serum versus "status
duration" and "status onset time to LP" indicate an identical
relative increment of NfL concentrations in CSF and serum
despite differing absolute values, arguing against a large
influence of passage time between CSF and serum. "Status
duration” and "status onset time to LP" were correlated with
each other (r = 0.66, p = 0.0001, Supplementary Fig. 2), and
“status onset time to LP” was always longer or equal to “sta-
tus duration”. To account for the influence of these covari-
ates (hereinafter referred to only as "covariates") we adjusted
for them using linear regression analyses. Without and with
adjustment for covariates, NfL concentrations did not differ
significantly between patients with convulsive SE and NCSE
according to the ILAE classification and the “convulsive
only” versus “NCSE at any point during evolution” catego-
ries of the alternative classification (Fig. 3). Unadjusted NfL
showed a trend towards higher concentrations in treatment
RSE (p(CSF) = 0.05, p(serum) = 0.08). The status duration
13
2132 Journal of Neurology (2023) 270:2128–2138

Table 1  Basic characteristics of Characteristics SE patients Controls Significance

the SE patients and controls
Sample size n = 28 n = 1180
Sex, n (%)
Male 4 (14%) 600 (51%) p < 0.0011
Female 24 (86%) 580 (49%)
Age, years
Mean value (± SD) 69 (± 16) 58 (± 12) p < 0.0011
Serum NfL, pg/ml
Mean value 244 12 p < 0.0011
Median 99 10
IQR 169 7
Status duration
Median in h 44
≤ 24 h, n (%) 13 (46%) –
> 24 h, n (%) 15 (54%) –
Status onset time to LP
Median in h 72
Semiology, n (%)
Convulsive SE (ILAE) 21 (75%) –
NCSE (ILAE) 7 (25%)
Convulsive only (alternative classification) 15 (54%) –
NCSE at any point during evolution (alternative 13 (46%)
classification)
Refractoriness, n (%)
Responsive 7 (25%) –
Refractory 19 (68%) –
Super-refractory 2 (7%)
Etiology, n (%)
Structural 15 (54%) –
Unknown 13 (46%) –
STESS (0–6)
Median 4.5 –
IQR 2 –
mSTESS (0–8)
Median 6 –
IQR 2 –
EMSE-EAC (cut-off 27)
Median 40 –
IQR 31 –
Died, n (%) 6 (21%) –

EMSE-EAC epidemiology-based mortality score in status epilepticus using the combination of etiology, age
and comorbidity, LP lumbar puncture, mSTESS modified STESS, NCSE non-convulsive status epilepticus,
NfL neurofilament light, SE status epilepticus, STESS status epilepticus severity score

was longer in treatment RSE than in interruptible SE (mean
86.0 vs. 2.4 h).
Association of NfL with death and prognostic scales
Next, we assessed the correlation between NfL, death
(Fig. 4a) and the prognostic scales EMSE_EAC (Fig. 4b) and
STESS/mSTESS (Supplementary Fig. 3). NfL concentration
were not associated with death without adjustment for the
covariates (p(CSF) = 0.78, p(Serum) = 0.21) nor with adjust-
ment (p(CSF) = 0.98, p(Serum) = 0.43, Fig. 4a). Next, we
estimated the relation between the STESS and EMSE_EAC
and death versus survival using Receiver-Operating-Char-
acteristics (ROC) analysis. We obtained Areas-Under-the-
Curve of 0.70 for STESS, 0.62 for mSTESS and 0.71 for
EMSE-EAC (data not shown). These values point in the

13
Journal of Neurology (2023) 270:2128–2138
Fig. 1  A NfL in patients with serum NfL < 60 pg/ml and controls, B
in all patients and controls. Sample: Status epilepticus = SE patients
from this study, Popgen = Popgen control sample measured on the
same instrument concomitantly with the epilepsy samples, BiDirect,
and MEMO: additional control samples measured on an identical
but not the same instrument using the same Simoa kit. Blue inter-
expected direction and are roughly comparable to other
studies [12, 33]. We assessed the log-transformed EMSE
as an interval-scaled normally distributed variable and
STESS (7 levels) as well as mSTESS (9 levels) as ordinal
categorical variables. EMSE scores were not correlated
with NfL (Fig. 4b) neither without covariate adjustment
(p(CSF) = 0.89, p(Serum) = 0.16) nor with adjustment
(p(CSF) = 0.86, p(Serum) = 0.23).
Due to the low number of individuals in some categories,
a meaningful statistical analysis of the STESS and mSTESS
data was not possible. However, Supplementary Fig. 3 shows
that a trend towards higher NfL values in higher STESS or
mSTESS categories is not apparent.
Discussion
Unlike most discussions, we start with the evident weak-
nesses of our study, because these are important not only
for the interpretation of the data and results but also for the
design of follow-up studies. Apart from the cross-sectional
design and the relatively small patient sample, the major
weaknesses are the large differences in the time to blood/
CSF sampling and the high phenotypic variability of SE
concerning the etiology, semiology, and duration. While
the time to sampling can be standardized, the large pheno-
typic variability cannot. If standardization of time to sam-
pling leads to NfL as a clinically useful prognostic marker
remains to be shown. However, using a standardized sam-
pling time, some status will be sampled after the end of the
status, some during the status, and some status will continue
a long time after sampling, continuing to release markers of
2133
rupted line: NfL per age regression line for the joint control sample,
gray band: standard error for the mean of the NfL per age regression
line. Green interrupted lines: 95% prediction interval for the NfL per
age regression, meaning that the linear model predicts that 95% of
all population control samples will show NfL concentrations within
these limits
neurodestruction. Despite its weak points, our study allows
several meaningful conclusions. First, we show that Simoa
technology is sensitive and reliable enough to measure NfL
in serum as a proxy for CSF in SE. This is a finding in line
with the other paper on that topic by Giovannini and col-
leagues [6]. In both studies, even though Giovannini used a
different method to measure NfL, the correlation between
CSF and serum NfL values in SE patients is very similar.
Second, we conclude from the parallel increase in con-
centration in serum and CSF that the passage time from CSF
to blood must be too short to be meaningful even if the time
from seizure to sample ascertainment is as short as ~ 2–6 h.
We observe a log/log-linear increase of NfL in CSF and
serum with status duration as well as time to sampling. We
cannot tease apart with certainty whether status duration or
time to sampling is the decisive parameter but we think that
status duration has higher importance, at least if the sample
was taken later than ~ 24 h (15 samples) because studies of
“single hit” events like traumatic brain injury have shown
that the NfL levels are rather stable after one day [1]. This
observation from our study is relevant for any further use
of NfL in clinical settings. Early transfer of CSF into serum
means that measurement in serum is meaningful. In addi-
tion, the measurement of NfL in serum is easy to perform
and repeat. Medically, our most important conclusion is that
brain damage as measured by NfL might not differ by a very
large extent between convulsive SE and non-convulsive SE,
at least in patients with severely impaired consciousness. In
our study, this is true for SE classified according to the ILAE
classification but also for the alternative classification taking
the development of the semiology over time into account.
This finding suggests that concerning brain damage only,
13
2134 Journal of Neurology (2023) 270:2128–2138

Table 2  Basic characteristics of the SE patients stratified by epilepsy classification and death
ILAE classification convulsive SE NCSE p value (test)

Number 21 7 0.001* (Fisher)

Age (yrs ± SD) 70 ± 18 67 ± 13 0.563 (t test)
Sex (number female) 17 7 0.545 (Fisher)
status duration (hrs (IQR)) 39 (1–84) 60 (16–74) 0.523 (MWU)
status to LP (hrs (IQR)) 85 (30–247) 72 (55–267) 0.750 (MWU)
NFL (csf) (pg/ml (IQR)) 4149 (2060–9755) 3518 (1184–28,898) 0.836 (MWU)
NFL (serum) (pg/ml (IQR) 95 (48–180) 184 (54–384) 0.641 (MWU)
Died (Number) 5 1 1.00 (Fisher)
STESS (0–6 (IQR)) 5.00 (4.00–6.00) 6.00 (4.50–7.00) 0.634 (MWU)
mSTESS (0–8 (IQR)) 7.00 (6.00–8.00) 8.00 (6.50–8.00) 0.382 (MWU)
EMSE-EAC (0–64 (IQR)) 35.00 (25.00–47.00) 40.00 (34.50–62.50) 0.457 (MWU)
Alternative classification convulsive only NCSE at any point during evolution p value

Number 15 13 0.790 (Fisher)

Age (yrs ± SD) 71 ± 19 67 ± 13 0.463 (t test)
Sex (number female) 12 12 0.600 (Fisher)
status duration (hrs (IQR)) 6 (1–55) 60 (24–168) 0.011* (MWU)
status to LP (hrs (IQR)) 50 (29–115) 168 (60–390) 0.088 (MWU)
NFL (csf) (pg/ml (IQR)) 2319 (1853–6167) 9148 (2146–28,639) 0.085 (MWU)
NFL (serum) (pg/ml (IQR) 67 (48–128) 184 (63–558) 0.155 (MWU)
Died (Number) 3 3 1.00 (Fisher)
STESS (0–6 (IQR)) 6.00 (4.50 – 6.00) 5.00 (4.00–6.00) 0.865 (MWU)
mSTESS (0–8 (IQR)) 7.00 (6.00–8.00) 7.00 (6.00–8.00) 0.635 (MWU)
EMSE-EAC (0–64 (IQR)) 43.00 (32.00–51.00) 35.00 (25.00–57.00) 0.519 (MWU)
Survival Died Survived p value

Number 6 22 < 0.001* (Fisher)

Age (yrs ± SD) 75 ± 11 68 ± 17 0.267 (t test)
Sex (number female) 4 20 0.192 (Fisher)
status duration (hrs (IQR)) 50 (14–78) 37 (3–83) 0.674 (MWU)
status to LP (hrs (IQR)) 50 (41–113) 93 (50–300) 0.251 (MWU)
NFL (csf) (pg/ml (IQR)) 3077 (1729–4324) 6120 (1887–12,883) 0.790 (MWU)
NFL (serum) (pg/ml (IQR) 44 (35–94) 125 (59–230) 0.214 (MWU)
Died (Number) na na na
STESS (0–6 (IQR)) 6.00 (6.00–6.00) 5.00 (4.00–6.00) 0.117 (MWU)
mSTESS (0–8 (IQR)) 7.00 (7.00–7.75) 7.00 (6.00–8.00) 0.372 (MWU
EMSE-EAC (0–64 (IQR)) 61.00 (39.75–81.50) 37.50 (26.00–46.50) 0.123 (MWU)

EMSE-EAC epidemiology-based mortality score in status epilepticus using the combination of etiology, age and comorbidity, LP lumbar
puncture, mSTESS modified STESS, na not available, NCSE non-convulsive status epilepticus, NfL neurofilament light, SE status epilepticus,
STESS status epilepticus severity score

both convulsive, as well as non-convulsive SE, might turn
out to be medical emergencies of comparable urgency if our
results are corroborated by future studies. This conclusion
is in agreement with animal studies as outlined in the intro-
duction but has not yet been shown in humans [5, 22–24]. It
is currently a matter of debate if both convulsive- and non-
convulsive SE should be treated aggressively considering
the adverse effects of sedatives and anticonvulsants [44].
Especially in the elderly, non-convulsive SE is commonly
treated less aggressively than convulsive SE [29, 39]. If
future studies confirm that non-convulsive SE too leads to
significant neurodestruction, intensified treatment might be
indicated. Additional findings are that NfL concentrations in
most SE patients are higher than the age-adjusted 95th per-
centile of neurologically healthy controls which is in agree-
ment with the only other study of NfL in SE [6]. Interest-
ingly, the range of NfL concentrations overlapped between
SE and controls also in this study [6]. The comparison

13
Journal of Neurology (2023) 270:2128–2138 2135

Fig. 2  NfL versus A status duration and B status onset time to LP. r Pearson correlation coefficient, p p value for the correlation

Fig. 3  NfL in convulsive SE

and NCSE according to: A
ILAE classification, B alterna-
tive classification. p p value,
LM linear model (equivalent
to t test), LMadj linear model
adjusted for status duration and
status onset time to LP. Serum
NfL concentrations were mul-
tiplied by the factor 60 to allow
plotting them side by side with
the CSF NfL concentrations

of NfL between interruptible and treatment RSE pointed
towards a possible difference in NfL concentrations just fail-
ing to reach statistical significance. However, we think that
the main determinant of NfL concentrations is the status
duration and not treatment refractoriness because (1) status
duration was—as expected—much longer in treatment RSE
and (2) status duration was highly linearly correlated on a
log–log scale with NfL concentration across the whole range
of status durations (Fig. 2). We found no association of NfL
levels with scales predicting mortality nor with treatment
refractoriness. We cannot exclude that this is caused either
by the extreme heterogeneity of SE or the small sample size
or our clinical sampling scheme hindering the detection of
small differences. This important difference to the study by
Giovannini and colleagues might in part be explainable by
the fact that we adjusted for the status duration and time to
CSF/serum sampling in all analyses while Giovannini and
colleagues did not [6]. Larger studies taking the lessons from
this study and the study by Giovannini et al. in the study
design into account are required to draw firm conclusions.
The integration of patients with differential diagnoses of SE
is required to explore the diagnostic value of NfL in SE.

13
2136
Fig. 4  NfL in A patients who
survived or died, B versus
EMSE-EAC score. p p value,
LM linear model (equivalent
to t test), LMadj linear model
adjusted for status duration and
status onset time to LP. Serum
NfL concentrations were mul-
tiplied by the factor 60 to allow
plotting them side by side with
the CSF NfL concentrations
Conclusion
In summary, we show that NfL might be a clinically useful
biomarker of SE, that neurodestruction, as measured by NfL,
does not differ largely between convulsive SE and NCSE,
and point out several caveats and suggestions for subsequent
studies.
Supplementary Information The online version contains supplemen-
tary material available at https://d​ oi.o​ rg/1​ 0.1​ 007/s​ 00415-0​ 22-1​ 1547-4.
Author contributions All authors contributed to the study conception
and design. Material preparation, data collection and analysis were per-
formed by NGM, JD, ET, FL, WL, AF, KB, JK and GK. The first draft
of the manuscript was written by NGM, GK and JD and all authors
commented on previous versions of the manuscript. All authors read
and approved the final manuscript.
Funding Open Access funding enabled and organized by Projekt
DEAL. The authors did not receive support from any organization for
the submitted work.
Data availability Anonymized data will be shared by request from any
qualified investigator.
Declarations
Conflicts of interest NGM received travel grants from Eisai Pharma
and Angelini Pharma. Lecture fees were given by Jazz Pharma and
Angelini Pharma. Research support was granted by Jazz Pharma,
Angelini Pharma, Eisai Pharma, UCB Pharma, LivaNova and Desi-
tin Pharma. He also receives funding from the Ministry of Health
(Schleswig–Holstein) and the German Research Council (DFG).
Non-financial interests: none. JD has no relevant financial or non-
financial interests to disclose. ET has no relevant financial or non-
financial interests to disclose. FL reports having received speaker’s
honoraria from Grifols, Biogen, Novartis, Roche, Alexion, serving on
advisory boards for Roche and Biogen and receiving research funding
13
Journal of Neurology (2023) 270:2128–2138
from the German Ministry of Education and Research, the German
Society for Laboratory Medicine (DGKL) and the German Research
Council (DFG). He works for an academic institution offering com-
mercial autoantibody testing. Non-financial interests: none. WL has
no relevant financial or non-financial interests to disclose. AF has no
relevant financial or non-financial interests to disclose. KB receives for
the MEMO Study funding by the German Research Society (Deutsche
Forschungsgemeinschaft DFG, grant: BE1996/1-1). The measurement
of NfL was done through funds from the Institute of Epidemiology and
Social Medicine, University of Muenster. The BiDirect Study is sup-
ported by grants of the German Ministry of Research and Education
(BMBF) to the University of Muenster (01ER0816 and 01ER1506).
Non-financial interests: none. JK received speaker fees, research sup-
port, travel support, and/or served on advisory boards by Swiss MS
Society, Swiss National Research Foundation (320030_189140/1),
University of Basel, Progressive MS Alliance, Bayer, Biogen, Bristol
Myers Squibb, Celgene, Merck, Novartis, Octave Bioscience, Roche,
Sanofi. Non-financial interests: none. GK receives non-project-related
financial support from the BMBF project CONNECT-GENERATE:
Genetics of autoimmune encephalitis. Non-financial interests: none.
The authors declare that they have no conflict of interest concerning
this work.
Ethical standards The study was approved by the Ethics Committee
of the Medical Faculty of the Christian-Albrechts-University, Kiel,
Germany (D 480/20). The study was conducted following the World
Medical Association Declaration of Helsinki. The use of retrospective
data was reviewed and confirmed by the Ethics Committee. The data
protection guidelines were also observed. The identity of the subjects
is protected.
Open Access This article is licensed under a Creative Commons Attri-
bution 4.0 International License, which permits use, sharing, adapta-
tion, distribution and reproduction in any medium or format, as long
as you give appropriate credit to the original author(s) and the source,
provide a link to the Creative Commons licence, and indicate if changes
were made. The images or other third party material in this article are
included in the article's Creative Commons licence, unless indicated
otherwise in a credit line to the material. If material is not included in
Journal of Neurology (2023) 270:2128–2138
the article's Creative Commons licence and your intended use is not
permitted by statutory regulation or exceeds the permitted use, you will
need to obtain permission directly from the copyright holder. To view a
copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/.
References
1. Al Nimer F, Thelin E, Nystrom H, Dring AM, Svenningsson A,
Piehl F, Nelson DW, Bellander BM (2015) Comparative assess-
ment of the prognostic value of biomarkers in traumatic brain
injury reveals an independent role for serum levels of neurofila-
ment light. PLoS ONE 10:e0132177
2. Bleck TP (2005) Refractory status epilepticus. Curr Opin Crit
Care 11:117–120
3. Boggs JG (2004) Mortality associated with status epilepticus.
Epilepsy Curr 4:25–27
4. Celikbilek A, Tanik N, Sabah S, Borekci E, Akyol L, Ak H, Adam
M, Suher M, Yilmaz N (2014) Elevated neurofilament light chain
(NFL) mRNA levels in prediabetic peripheral neuropathy. Mol
Biol Rep 41:4017–4022
5. de Curtis M, Rossetti AO, Verde DV, van Vliet EA, Ekdahl CT
(2021) Brain pathology in focal status epilepticus: evidence from
experimental models. Neurosci Biobehav Rev 131:834–846
6. Giovannini G, Bedin R, Ferraro D, Vaudano AE, Mandrioli J,
Meletti S (2022) Serum neurofilament light as biomarker of
seizure-related neuronal injury in status epilepticus. Epilepsia
63:e23–e29
7. Hanin A, Demeret S, Denis JA, Nguyen-Michel VH, Rohaut B,
Marois C, Imbert-Bismut F, Bonnefont-Rousselot D, Levy P,
Navarro V, Lambrecq V (2022) Serum neuron-specific enolase: a
new tool for seizure risk monitoring after status epilepticus. Eur
J Neurol 29:883–889
8. Hanin A, Denis JA, Frazzini V, Cousyn L, Imbert-Bismut F,
Rucheton B, Bonnefont-Rousselot D, Marois C, Lambrecq V,
Demeret S, Navarro V (2022) Neuron specific enolase, S100-
beta protein and progranulin as diagnostic biomarkers of status
epilepticus. J Neurol 269:3752–3760
9. Hanin A, Lambrecq V, Denis JA, Imbert-Bismut F, Rucheton B,
Lamari F, Bonnefont-Rousselot D, Demeret S, Navarro V (2020)
Cerebrospinal fluid and blood biomarkers of status epilepticus.
Epilepsia 61:6–18
10. Hepner A, Porter J, Hare F, Nasir SS, Zetterberg H, Blennow
K, Martin MG (2019) Serum neurofilament light, glial fibrillary
acidic protein and tau are possible serum biomarkers for activity
of brain metastases and gliomas. World J Oncol 10:169–175
11. Hirsch LJ, Gaspard N, van Baalen A, Nabbout R, Demeret S,
Loddenkemper T, Navarro V, Specchio N, Lagae L, Rossetti AO,
Hocker S, Gofton TE, Abend NS, Gilmore EJ, Hahn C, Khosra-
vani H, Rosenow F, Trinka E (2018) Proposed consensus defini-
tions for new-onset refractory status epilepticus (NORSE), febrile
infection-related epilepsy syndrome (FIRES), and related condi-
tions. Epilepsia 59:739–744
12. Kang BS, Kim DW, Kim KK, Moon HJ, Kim YS, Kim HK, Lee
SY, Koo YS, Shin JW, Moon J, Sunwoo JS, Byun JI, Cho YW,
Jung KY, Chu K, Lee SK (2016) Prediction of mortality and func-
tional outcome from status epilepticus and independent external
validation of STESS and EMSE scores. Crit Care 20:25
13. Krawczak M, Nikolaus S, von Eberstein H, Croucher PJ, El
Mokhtari NE, Schreiber S (2006) PopGen: population-based
recruitment of patients and controls for the analysis of complex
genotype-phenotype relationships. Community Genet 9:55–61
2137
14. Kuhle J, Kropshofer H, Haering DA, Kundu U, Meinert R, Barro
C, Dahlke F, Tomic D, Leppert D, Kappos L (2019) Blood neu-
rofilament light chain as a biomarker of MS disease activity and
treatment response. Neurology 92:e1007–e1015
15. Lattanzi S, Giovannini G, Brigo F, Orlandi N, Trinka E, Meletti
S (2021) Status epilepticus with prominent motor symptoms
clusters into distinct electroclinical phenotypes. Eur J Neurol
28:2694–2699
16. Leitinger M, Beniczky S, Rohracher A, Gardella E, Kalss G,
Qerama E, Hofler J, Hess Lindberg-Larsen A, Kuchukhidze G,
Dobesberger J, Langthaler PB, Trinka E (2015) Salzburg con-
sensus criteria for non-convulsive status epilepticus–approach to
clinical application. Epilepsy Behav 49:158–163
17. Leitinger M, Holler Y, Kalss G, Rohracher A, Novak HF, Hofler J,
Dobesberger J, Kuchukhidze G, Trinka E (2015) Epidemiology-
based mortality score in status epilepticus (EMSE). Neurocrit
Care 22:273–282
18. Leitinger M, Trinka E, Giovannini G, Zimmermann G, Florea
C, Rohracher A, Kalss G, Neuray C, Kreidenhuber R, Hofler J,
Kuchukhidze G, Granbichler C, Dobesberger J, Novak HF, Pilz
G, Meletti S, Siebert U (2019) Epidemiology of status epilep-
ticus in adults: a population-based study on incidence, causes,
and outcomes. Epilepsia 60:53–62
19. Leitinger M, Trinka E, Zimmermann G, Granbichler CA, Kobu-
lashvili T, Siebert U (2020) Epidemiology of status epilepticus
in adults: apples, pears, and oranges—a critical review. Epilepsy
Behav 103:106720
20. Mattsson N, Andreasson U, Zetterberg H, Blennow K, Alzhei-
mer’s Disease Neuroimaging I (2017) Association of plasma
neurofilament light with neurodegeneration in patients with
Alzheimer disease. JAMA Neurol 74:557–566
21. Mattsson N, Bremell D, Anckarsater R, Blennow K, Anckar-
sater H, Zetterberg H, Hagberg L (2010) Neuroinflammation
in Lyme neuroborreliosis affects amyloid metabolism. BMC
Neurol 10:51
22. Meldrum BS (1993) Excitotoxicity and selective neuronal loss in
epilepsy. Brain Pathol 3:405–412
23. Meldrum BS, Brierley JB (1973) Prolonged epileptic seizures in
primates. Ischemic cell change and its relation to ictal physiologi-
cal events. Arch Neurol 28:10–17
24. Meldrum BS, Vigouroux RA, Brierley JB (1973) Systemic fac-
tors and epileptic brain damage. Prolonged seizures in paralyzed,
artificially ventilated baboons. Arch Neurol 29:82–87
25. Novy J, Logroscino G, Rossetti AO (2010) Refractory status epi-
lepticus: a prospective observational study. Epilepsia 51:251–256
26. Olsson B, Portelius E, Cullen NC, Sandelius A, Zetterberg H,
Andreasson U, Hoglund K, Irwin DJ, Grossman M, Weintraub
D, Chen-Plotkin A, Wolk D, McCluskey L, Elman L, Shaw LM,
Toledo JB, McBride J, Hernandez-Con P, Lee VM, Trojanowski
JQ, Blennow K (2019) Association of cerebrospinal fluid neu-
rofilament light protein levels with cognition in patients with
dementia, motor neuron disease, and movement disorders. JAMA
Neurol 76:318–325
27. Pollak L, Gandelman-Marton R, Margolin N, Boxer M, Blatt I
(2014) Clinical and electroencephalographic findings in acutely
ill adults with non-convulsive vs convulsive status epilepticus.
Acta Neurol Scand 129:405–411
28. Rissin DM, Kan CW, Campbell TG, Howes SC, Fournier DR,
Song L, Piech T, Patel PP, Chang L, Rivnak AJ, Ferrell EP, Ran-
dall JD, Provuncher GK, Walt DR, Duffy DC (2010) Single-mol-
ecule enzyme-linked immunosorbent assay detects serum proteins
at subfemtomolar concentrations. Nat Biotechnol 28:595–599
29. Rossetti AO, Hirsch LJ, Drislane FW (2019) Nonconvulsive
seizures and nonconvulsive status epilepticus in the neuro ICU
should or should not be treated aggressively: a debate. Clin Neu-
rophysiol Pract 4:170–177
13
2138
30. Rossetti AO, Logroscino G, Bromfield EB (2006) A clinical
score for prognosis of status epilepticus in adults. Neurology
66:1736–1738
31. Rossetti AO, Logroscino G, Milligan TA, Michaelides C, Ruff-
ieux C, Bromfield EB (2008) Status Epilepticus Severity Score
(STESS): a tool to orient early treatment strategy. J Neurol
255:1561–1566
32. Rubsamen N, Maceski A, Leppert D, Benkert P, Kuhle J, Wiendl
H, Peters A, Karch A, Berger K (2021) Serum neurofilament light
and tau as prognostic markers for all-cause mortality in the elderly
general population-an analysis from the MEMO study. BMC Med
19:38
33. Sairanen JJ, Kantanen AM, Hyppola HT, Kalviainen RK
(2020) Outcome of status epilepticus and the predictive value
of the EMSE and STESS scores: a prospective study. Seizure
75:115–120
34. Shahim P, Politis A, van der Merwe A, Moore B, Ekanayake V,
Lippa SM, Chou YY, Pham DL, Butman JA, Diaz-Arrastia R, Zet-
terberg H, Blennow K, Gill JM, Brody DL, Chan L (2020) Time
course and diagnostic utility of NfL, tau, GFAP, and UCH-L1 in
subacute and chronic TBI. Neurology 95:e623–e636
35. Steinacker P, Anderl-Straub S, Diehl-Schmid J, Semler E, Uttner
I, von Arnim CAF, Barthel H, Danek A, Fassbender K, Fliess-
bach K, Foerstl H, Grimmer T, Huppertz HJ, Jahn H, Kassubek
J, Kornhuber J, Landwehrmeyer B, Lauer M, Maler JM, Mayer
B, Oeckl P, Prudlo J, Schneider A, Volk AE, Wiltfang J, Schro-
eter ML, Ludolph AC, Otto M (2018) Serum neurofilament light
chain in behavioral variant frontotemporal dementia. Neurology
91:e1390–e1401
13
Journal of Neurology (2023) 270:2128–2138
36. Suzuki Y, Toribe Y, Goto M, Kato T, Futagi Y (1999) Serum
and CSF neuron-specific enolase in patients with West syndrome.
Neurology 53:1761–1764
37. Tanabe T, Suzuki S, Hara K, Shimakawa S, Wakamiya E, Tamai
H (2001) Cerebrospinal fluid and serum neuron-specific enolase
levels after febrile seizures. Epilepsia 42:504–507
38. Teismann H, Wersching H, Nagel M, Arolt V, Heindel W, Baune
BT, Wellmann J, Hense HW, Berger K (2014) Establishing the
bidirectional relationship between depression and subclinical arte-
riosclerosis–rationale, design, and characteristics of the BiDirect
Study. BMC Psychiatry 14:174
39. Treiman DM, Walker MC (2006) Treatment of seizure emergen-
cies: convulsive and non-convulsive status epilepticus. Epilepsy
Res 68(Suppl 1):S77-82
40. Trinka E, Cock H, Hesdorffer D, Rossetti AO, Scheffer IE, Shin-
nar S, Shorvon S, Lowenstein DH (2015) A definition and clas-
sification of status epilepticus—report of the ILAE task force on
classification of status epilepticus. Epilepsia 56:1515–1523
41. Trinka E, Leitinger M (2015) Which EEG patterns in coma are
nonconvulsive status epilepticus? Epilepsy Behav 49:203–222
42. Varhaug KN, Torkildsen O, Myhr KM, Vedeler CA (2019) Neu-
rofilament light chain as a biomarker in multiple sclerosis. Front
Neurol 10:338
43. Verde F, Otto M, Silani V (2021) Neurofilament light chain as
biomarker for amyotrophic lateral sclerosis and frontotemporal
dementia. Front Neurosci 15:679199
44. Walker MC (2007) Treatment of nonconvulsive status epilepticus.
Int Rev Neurobiol 81:287–297