Serial soluble neurofilament heavy chain in plasma as a marker of brain injury after cardiac arrest.
ABSTRACT INTRODUCTION: Induced hypothermia has been shown to improve outcome after cardiac arrest, but early prognostication is hampered by the need for sedation. Here we tested whether a biomarker for neurodegeneration, the neurofilament heavy chain (NfH), may improve diagnostic accuracy in the first days after cardiac arrest. METHODS: This prospective study included 90 consecutive patients treated with hypothermia after cardiac arrest. Plasma levels of phosphorylated NfH (SMI35) were quantified using standard ELISA over a period of 72 h after cardiac arrest. The primary outcome was the dichotomized Cerebral Performance Categories scale (CPC). A best CPC 1-2 during 6 months follow-up was considered a good outcome, a best CPC of 3-4 a poor outcome. Receiver operator characteristics and area under the curve were calculated. RESULTS: The median age of the patients was 65 years, and 63 (70%) were male. A cardiac aetiology was identified in 62 cases (69%). 77 patients (86%) had out-of-hospital cardiac arrest. The outcome was good in 48 and poor in 42 patients. Plasma NfH levels were significantly higher 2 and 36 hours after cardiac arrest in patients with poor outcome (median 0.28 ng/mL and 0.5 ng/mL, respectively) compared to those with good outcome (0 ng/mL, p = 0.016, p < 0.005, respectively). The respective AUC were 0.72 and 0.71. CONCLUSIONS: Plasma NfH levels correlate to neurological prognosis following cardiac arrest. In this study, 15 patients had neurological co-morbidities and there was a considerable overlap of data. As such, neurofilament should not be used for routine neuroprognostication until more data are available.
- [show abstract] [hide abstract]
ABSTRACT: When a patient resuscitated from cardiac arrest remains unconscious the clinician would like to have a reliable early method for predicting the outcome. The objective of our study was to predict cerebral outcome after cardiac arrest by clinical neurological examination. The data were drawn from an international multicentre controlled clinical trial of thiopentone. Twelve hospitals in nine countries took part. 262 comatose cardiac arrest survivors were followed up for one year. These patients were given advanced life support (American Heart Association guidelines) followed by intensive care to a standardised protocol. Glasgow and Glasgow-Pittsburgh coma scores and their constituent signs were recorded at fixed times. Outcome was taken to be the best cerebral performance at any time during follow-up, and for that purpose we used cerebral performance categories (CPC 1-5) of the Glasgow outcome categories. A poor outcome (CPC 3-5) could be predicted immediately after reperfusion (at entry into the study) with an accuracy ranging from 52% to 84% for various signs and scores. On the third day it was possible to identify severely disabled or permanently comatose survivors without false predictions using both coma scores and several of their constituent variables. The best predictor was absence of motor response to pain. This modelling exercise now needs to be repeated on a new series of patients but the results do suggest that, after 3 days, stringent ethical criteria can be met and used in decision-making about termination of care in comatose cardiac arrest survivors.The Lancet 05/1994; 343(8905):1055-9. · 39.06 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: To systematically review outcomes in comatose survivors after cardiac arrest and cardiopulmonary resuscitation (CPR). The authors analyzed studies (1966 to 2006) that explored predictors of death or unconsciousness after 1 month or unconsciousness or severe disability after 6 months. The authors identified four class I studies, three class II studies, and five class III studies on clinical findings and circumstances. The indicators of poor outcome after CPR are absent pupillary light response or corneal reflexes, and extensor or no motor response to pain after 3 days of observation (level A), and myoclonus status epilepticus (level B). Prognosis cannot be based on circumstances of CPR (level B) or elevated body temperature (level C). The authors identified one class I, one class II, and nine class III studies on electrophysiology. Bilateral absent cortical responses on somatosensory evoked potential studies recorded 3 days after CPR predicted poor outcome (level B). Burst suppression or generalized epileptiform discharges on EEG predicted poor outcomes but with insufficient prognostic accuracy (level C). The authors identified one class I, 11 class III, and three class IV studies on biochemical markers. Serum neuron-specific enolase higher than 33 microg/L predicted poor outcome (level B). Ten class IV studies on brain monitoring and neuroimaging did not provide data to support or refute usefulness in prognostication (level U). Pupillary light response, corneal reflexes, motor responses to pain, myoclonus status epilepticus, serum neuron-specific enolase, and somatosensory evoked potential studies can reliably assist in accurately predicting poor outcome in comatose patients after cardiopulmonary resuscitation for cardiac arrest.Neurology 08/2006; 67(2):203-10. · 8.25 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: Neurofilaments (Nf) are the major constitutents of the axoskeleton and body fluid Nf levels are an important tool for estimating axonal degeneration in vivo. This paper presents a new sandwich ELISA allowing quantification of the NfH(SMI35) phosphoform from CSF, brain tissue and cell culture homogenates. The sensitivity of the NfH(SMI35) ELISA is 0.2 ng/ml with a recovery of 119% and a mean within- and between-batch precision of 10.6% and 23%, respectively. CSF NfH(SMI35) was stable at 4 degrees C, is not influenced by freeze-thaw cycles, and proteolysis present at room temperature could be prevented by adding protease inhibitors. Aggregate formation was observed for HPLC-purified bovine NfH and could be resolved by sonication. The upper reference value for CSF NfH(SMI35) levels (0.73 ng/ml) was defined as the 95% cumulative frequency from 416 CSF samples. Based on this cutoff, a significantly higher proportion of patients with amyotrophic lateral sclerosis, space-occupying lesions, disc prolapse and subarachnoid haemorrhage had pathologically elevated NfH(SMI35) levels compared to patients with cluster headache or demyelinating disease.A new nomenclature is proposed to facilitate the comparison between ELISA, immunoblotting and immunocytochemistry.Journal of Immunological Methods 08/2003; 278(1-2):179-90. · 2.23 Impact Factor
Serial soluble neurofilament heavy chain in
plasma as a marker of brain injury after
Malin Rundgren1*, Hans Friberg1, Tobias Cronberg2, Bertil Romner3and Axel Petzold4,5
Introduction: Induced hypothermia has been shown to improve outcome after cardiac arrest, but early
prognostication is hampered by the need for sedation. Here we tested whether a biomarker for
neurodegeneration, the neurofilament heavy chain (NfH), may improve diagnostic accuracy in the first days after
Methods: This prospective study included 90 consecutive patients treated with hypothermia after cardiac arrest.
Plasma levels of phosphorylated NfH (SMI35) were quantified using standard ELISA over a period of 72 h after
cardiac arrest. The primary outcome was the dichotomized Cerebral Performance Categories scale (CPC). A best
CPC 1-2 during 6 months follow-up was considered a good outcome, a best CPC of 3-4 a poor outcome. Receiver
operator characteristics and area under the curve were calculated.
Results: The median age of the patients was 65 years, and 63 (70%) were male. A cardiac aetiology was identified
in 62 cases (69%). 77 patients (86%) had out-of-hospital cardiac arrest. The outcome was good in 48 and poor in
42 patients. Plasma NfH levels were significantly higher 2 and 36 hours after cardiac arrest in patients with poor
outcome (median 0.28 ng/mL and 0.5 ng/mL, respectively) compared to those with good outcome (0 ng/mL, p =
0.016, p < 0.005, respectively). The respective AUC were 0.72 and 0.71.
Conclusions: Plasma NfH levels correlate to neurological prognosis following cardiac arrest. In this study, 15
patients had neurological co-morbidities and there was a considerable overlap of data. As such, neurofilament
should not be used for routine neuroprognostication until more data are available.
Post cardiac arrest intensive care is complex and the
dynamic multi-organ failure is often referred to as the
post-resuscitation syndrome . The major cause of
death in patients with return of spontaneous circulation
(ROSC) is ischemic brain damage , which evolves over
several days . In many patients who remain comatose
after cardiac arrest, a reliable assessment of neurological
prognosis therefore needs to be postponed by several
days [4-6]. The use of induced hypothermia to improve
outcomes has further complicated matters because it
mandates sedation and intermittent use of muscle relax-
ants during the intervention. Moreover, hypothermia
delays the metabolism of drugs  and makes a clinical
neurological examination less reliable [8,9]. Therefore,
we need to reassess and improve our prognostic instru-
ments and explore novel and complementary methods of
a clinical neurological examination .
Several biomarkers have been evaluated to assess brain
damage after cardiac arrest with inconsistent results -
for example, neuronspecific enolase (NSE), S-100B, and
glial fibrillary acidic protein (GFAP) [5,11-13]. Among
these, NSE and S-100B are the most extensively studied
and NSE is the only one that has been integrated into
clinical guidelines . Although highly correlated to the
extent of ischemic brain injury , weaknesses with
NSE include its sensitivity to false positives due to sam-
ple hemolysis and the lack of a standard [14,15]. The
advantage of a high sensitivity of blood S-100B levels for
parenchymal brain damage is hampered by a lower
* Correspondence: email@example.com
1Department of Intensive-and Perioperative Care, Skåne University Hospital;
Department of Clinical Sciences Lund University, Lund, Sweden
Full list of author information is available at the end of the article
Rundgren et al. Critical Care 2012, 16:R45
© 2012 Rundgren et al.; licensee BioMed Central Ltd. This is an open access article distributed under the terms of the Creative
Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and
reproduction in any medium, provided the original work is properly cited.
specificity due to presence in extra-nervous system tis-
sues such as bone marrow and fatty tissue [16,17].
The neurofilament heavy chain (NfH) is a specific pro-
tein biomarker of neurons and axons . This 190 to
210 kDa protein of various degrees of phosphorylation
is an obligate heteropolymer with the Nf light (NfL) and
medium (NfM) chains. Together NfH, NfL and NfM
belong to the class IV group of intermediate filaments,
which support the axonal cytoskeleton . Accumula-
tion or increased levels of NfH and NfL were observed
in a number of diseases including stroke and cardiac
To investigate the time-course of NfH release and its
prognostic value, we studied serial plasma samples in a
cohort of hypothermia-treated cardiac arrest patients.
Materials and methods
This prospective study was performed in the general and
cardio-thoracic ICUs of Lund University Hospital, Swe-
den, from August 2003 to March 2007. The study was
approved by the Regional Ethical Review Board at Lund
University (411/2004, 223/2008), and informed consent
was sought from next of kin or, retrospectively, from
Epidemiological data and cardiac arrest data were col-
lected prospectively. All cardiac arrest patients, regard-
less of location of arrest or initial rhythm, with ROSC
and sustained unconsciousness (Glasgow coma scale
(GCS) ≤7), were considered for induced hypothermia.
Exclusion criteria for hypothermia treatment were term-
inal disease, intracerebral hemorrhage, aortic dissection,
or major trauma. Hypothermia was initiated as soon as
possible after ROSC in the emergency room or catheter-
ization laboratory using 30 ml/kg cold saline and subse-
quent treatment in the ICU was performed as described
A cardiologist initially evaluated all patients. Urgent
angiography, percutaneous cardiac intervention and, if
necessary, circulatory support using intraaortic balloon
pump counter pulsations was undertaken when indi-
cated. Patients received hypothermia for 24 hours at 33
+/-1°C and rewarming was controlled at 0.5°C/hour.
Patients were sedated using propofol 2 to 4 mg/kg/hour
and fentanyl 1 to 3 μg/kg/hour .
In patients remaining comatose, full intensive care was
provided for at least three days after normothermia, at
which time a clinical neurological evaluation was per-
formed. In addition, somatosensory evoked potentials
(SSEP), amplitude-integrated electroencephalogram
(aEEG) and diffusion-weighted magnetic resonance ima-
ging (DW-MRI) were added as a basis for a decision on
level of care . The patients were evaluated at ICU
and hospital discharge by an intensivist, and by a neu-
rologist six months later, using the five-graded Cerebral
Performance Categories (CPC) scale: CPC 1 = good cer-
ebral performance, CPC 2 = moderate cerebral disabil-
ity, independent, CPC 3 = severe cerebral disability,
conscious but dependent, CPC 4 = coma, CPC 5 =
death . To assess for clinically relevant neurological
injury, the best CPC score during six months follow up
was regarded as the primary outcome. A best CPC score
of 1 to 2 at any time was considered a good outcome
and a best CPC of 3 to 4 a poor outcome. Our second-
ary outcome was survival at six months.
As the release profile of soluble NfH after cardiac arrest
was not known, we chose to collect several samples over
the first three days to assess the release profile in a simi-
lar fashion as previously done for neuron-specific enolase
and S-100B , thus in part using a cohort on which we
have previously published. Plasma-samples for soluble
neurofilament analysis were collected at admission and at
2, 6, 12, 24, 36, 48, and 72 hours after cardiac arrest. The
plasma samples were centrifuged and frozen (-70°C)
immediately after collection. After the end of the study,
samples were thawed once, centrifuged at 4,000 rpm for
five minutes, aliquoted, and refrozen (-70°C) for later
analysis. During aliquoting, the samples were kept on ice.
We compared the levels of plasma NfH with the pre-
viously analysed levels of plasma NSE at the 48 hour
time-point. This time-point for NSE was chosen because
it had the best sensitivity/specificity for a poor prognosis
in our previous Receiver Operating Characteristic (ROC)
curve analysis . Levels of plasma NfH were also com-
pared with the results of SSEP-recordings approximately
72 hours after normothermia.
All samples were coded. Plasma NfH levels were mea-
sured in duplicates with the analyst being blinded to all
other information using a standard in-house ELISA .
Adhering to a previously proposed nomenclature the
capture antibody (SMI 35 for variously phosphorylated
NfH) is shown in the superscript as NfHSMI35. To mini-
mize the analytical error all samples were batch analysed
. Batch analysis improved the analytical error (coeffi-
cient of variation) to 5.4% in the present study. The
reported sensitivity of the ELISA is 0.2 ng/mL . Fol-
lowing batch analysis the detection limit in this study
was 0.001 ng/mL. Non-measurable NfHSMI35levels were
reported as 0 ng/mL.
The coded biomarker data and clinical data were elec-
tronically transferred to a third centre (Department of
Neurosurgery, Copenhagen University, Denmark). Only
after merging all code data the joint dataset was released
to the investigators for analysis of the predefined
hypotheses. Statistical analysis was performed using the
Rundgren et al. Critical Care 2012, 16:R45
Page 2 of 7
SPSS software software version 15.0 (SPSS Inc., Chicago,
IL, USA). Continuous data are presented as median and
inter quartile range (IQR). Categorical data are given as
counts and percentages. The material was dichotomized
using the primary endpoints good (best CPC 1 to 2 dur-
ing six months follow up), and poor neurological out-
come (best CPC 3 to 4 during six months follow up).
The data were not normally distributed, consequently
non-parametrical statistics were used. The Mann Whit-
ney U test was used to assess differences between the
good and poor outcome groups at the different time-
points. Bonferroni-corrections were used to correct for
multiple comparisons. Correlations between continuous
variables were assessed using Spearman correlation. Chi
squared test was used to compare frequencies. ROC
analysis was performed. P < 0.05 was considered
Ninety-two patients were included and two were
excluded due to all samples missing (n = 1) and failed
analysis (n = 1), resulting in 90 remaining patients. For
patient characteristics, see Table 1. The median time
from cardiac arrest to return of spontaneous circulation
was 20 minutes (IQR 14 to 30 minutes). Therapeutic
hypothermia was induced at a median of 66 minutes
(IQR 54 to 94 minutes) after cardiac arrest, and goal
temperature was reached at a median of 210 minutes
(IQR 135 to 295 minutes). Forty-eight patients (53%)
had a good outcome, defined as best CPC 1 to 2 during
follow up, and 42 (47%) had a poor outcome. At six
months follow up, 46 patients were still alive in the
good outcome group and one more in the poor out-
come group resulting in a six-month survival of 52% (47
of 90). For time and cause of death and withdrawal of
intensive care, see Table 2. The patients were heteroge-
neous with regard to their past medical history with sig-
nificant co-morbidity. Fifteen of the included patients
had a previously known neurological disease; stroke or
intracranial bleeding for more than one year ago (n =
8), epilepsy (n = 2), posttraumatic para-paresis (n = 1),
diabetes with polyneuropathy (n = 1), Guillian-Barré
syndrome (n = 1), diffuse memory problems (n = 1),
and placement of a ventriculoperitoneal shunt for post-
traumatic hydrocephalus (n = 1).
A total of 591 samples were collected from the 90
patients (Table 3). The main reasons for missing sam-
ples were due to intra- or inter-hospital transfer (early
samples), patients dying or regaining consciousness and
physically leaving the ICU (late samples).
After Bonferroni corrections, there were significant
differences between the good and the poor outcome
groups at two hours (P = 0.016) and 36 hours (P <
0.005) after cardiac arrest. Plasma NfHSMI35levels were
also detectable following cardiac arrest in patients with
a good clinical outcome. Due to overlapping NfHSMI35
values, no clinically meaningful cut-off levels separating
the good and the poor outcome groups could be identi-
fied (Figure 1). The ROC-analysis showed the highest
area under curve (AUC) for two hours (AUC 0.72, 95%
confidence interval (CI) 0.54 to 0.90), 36 hours (AUC
0.71, 95% CI 0.52 to 0.90) and 72 hours (AUC 0.68, 95%
CI 0.50 to 0.88) respectively (Figure 2). The AUC for
the other measurement times were 0.51 to 0.60 also
with wide CIs. Regarding six months survival, a group
difference remained after Bonferroni corrections at 36
hours (P = 0.008) but was also significant at 72 hours (P
= 0.04) after cardiac arrest. The plasma NfHSMI35con-
centration at 36 hours was selected for comparison with
NSE at 48 hours to limit the number of comparisons,
Table 1 Patient characteristics
Good outcome, n = 48Poor outcome, n = 42
Age (median, IQR)
Out-of-hospital cardiac arrest
Initial rhythm VT/VF (n = 85)
n = 90, if not otherwise stated
Good outcome is defined as a best cerebral performance categories scale 1 to 2 during six months follow up; P value from Mann Whitney U test (age), Fisher’s
exact test for remaining comparisons.
IQR, inter quartile range; VT/VF, pulse-less ventricular tachycardia or ventricular fibrillation.
Table 2 Time and cause of death
Number of patients833 2
Time of death (d)2 (2-32)7 (3-17)29 and 46
Time of withdrawal (d)NA6 (3-10)NA
Assessment of cause of death in dying patients, including time of death, the
number of patients where intensive care was withdrawn, and time of
withdrawal. In patients where intensive care was withdrawn due to
neurological reasons the cause of death was assessed as due to neurological
reasons. Other includes premorbid disease and other interfering disease
(septicaemia) unrelated to the cardiac arrest. Time is shown in days (d) as
median (range). NA = not applicable; MOF = Multi-organ failure
Rundgren et al. Critical Care 2012, 16:R45
Page 3 of 7
but no significant correlation was found (P = 0.068, data
not shown). Plasma NfHSMI35levels at 36 hours were
not significantly different between patients lacking SSEP
bilaterally and patients with at least unilateral identifi-
able N20 peaks (data not shown).
In approximately one third of the patients the cardiac
arrest was of non-cardiac cause (asphyxia/respiratory
failure (n = 9), pulmonary embolism (n = 5), hanging
and drowning (n = 3), hyperkalemia (n = 1), alcohol/epi-
lepsy/acidosis (n = 3), and unknown (n = 5)). A post hoc
analysis of the NfH values using only the patients with
cardiac arrest of a cardiac cause (n = 62) was made. The
results after Bonferroni-corrections (x8) were significant
for two hours (P = 0.008), 24 hours (P = 0.048), and 36
hours (P < 0.005). As in the primary analysis no cut-off
values were identifiable.
This study evaluated serial samples of NfHSMI35in
plasma from patients resuscitated after cardiac arrest
and treated with induced hypothermia. The main find-
ings were that NfHSMI35was measurable in plasma from
patients with poor as well as a good outcome, with sig-
nificantly higher acute levels (2 and 36 hours) in the
poor outcome group. The respective AUC (0.72 and
0.71) were promising but the considerable overlap of
data between the groups precluded calculation of a defi-
nite cut-off value to be recommended for use in indivi-
An ideal biomarker should be easily collected and is,
in practice, often drawn from blood. Regarding biomar-
kers assessing neurological damage, the cerebrospinal
(CSF) compartment may allow for more specific detec-
tion of central nervous system-related damage and is
therefore often used [22,29]. Patients exposed to modern
cardiac care, including most cardiac arrest patients, are
heavily anti-coagulated and exposed to drugs affecting
platelet function as well as low molecular weight
heparin. This contraindicates lumbar puncture for col-
lection of CSF due to the risk of epidural hematoma. In
Table 3 Plasma NfHSMI35and neurological outcome
Good outcome, n = 48
Poor Outcome, n = 42
medianIQR Rangen IQRRange
h, hours after cardiac arrest; IQR, interquartile range; NfH, neurofilament heavy chain; P, Mann-Whitney U-tests without Bonferroni corrections.
Figure 1 Box-plot comparing good and poor outcome groups
of patients at the different sampling times. Boxes show the inter
quartile range with median marked in the box, whiskers denote
values within 1.5 times the interquartile range. Outliers are omitted.
CA, cardiac arrest; NfH, neurofilament heavy chain.
Figure 2 ROC-curve showing 2, 36 and 72 hours after the
cardiac arrest. NfH, neurofilament heavy chain; ROC, receiver
Rundgren et al. Critical Care 2012, 16:R45
Page 4 of 7
the ICU setting, there is instead a considerable advan-
tage in blood sampling, allowing for serial measure-
ments, reducing the risk of making an erroneous
decision based on one single sample. Hence, although
the CSF may be a more relevant compartment, assess-
ment of biochemical neuro-markers in cardiac arrest
patients is better performed in blood.
NfH is a large protein with a molecular weight of 190
to 210 kD depending on the degree of phosphorylation
. Following neuro-axonal injury NfH is released into
the extracellular fluid from where it can be measured
using ELISA . From the extracellular fluid NfH dif-
fuses into the CSF compartment. From the CSF, NfH
reaches the blood via the blood CSF barrier at the lum-
bar level or it may diffuse directly through the cortical
arachnoid villi to the blood stream . The physiologi-
cal wash out pattern of NfHSMI35from the CSF is not
known in cardiac arrest. For reasons already discussed,
it was not possible to evaluate the assessment of
NfHSMI35kinetic from the CSF to the blood in this
Our time-course analysis showed an initial increase in
plasma levels of NfH, with significant differences
between the good and the poor outcome groups as early
as two hours after cardiac arrest (Figure 1). This may
reflect early neuronal damage . From experimental
data it was shown that blood NfH levels reach their
peak between 30 minutes and two hours after ‘cardiac
arrest’ . This fits well with the in vivo human data
from this study. Alternatively, NfHSMI35could be
released from peripheral axons [18,19] damaged during
resuscitation. The levels of plasma NfHSMI35in our
study were similar to those seen within six hours after a
mild stroke . However, despite a uniform protein
standard, one needs to be cautious when comparing
these studies, because one was performed from serum
and the other from plasma. The long-term temporal
dynamics of blood NfH following cardiac arrest are less
well known. A secondary increase of NfHSMI35occurred
in the poor outcome group at later sampling times
(Figure 1). We also found significant differences between
NfHSMI35levels in surviving and dying patients at 36
and 72 hours after cardiac arrest. This delayed release of
NfH in this exploratory study fits with the delayed apop-
totic cell death pattern described in experimental mod-
els of CA  and with the release pattern of the more
established biomarker NSE, which peaks at 48 to 72
hours after the cardiac arrest . Part of the lack of
significance may relate to the fact that approximately
one third of the patients had cardiac arrests of non-car-
diac cause, as suggested by the post hoc subgroup analy-
sis. It is also possible that complications in the post-
arrest period add to development of secondary brain
damage contributing to the later rise in blood levels
Still, we were not able to identify any cut-off levels or
predictive values to be recommended for individual
patients due to a considerable data overlap. One reason
may be that sampling was stopped at 72 hours after car-
diac arrest and that the optimal time for NfH sampling
thereby may have been missed. An increase in NfHSMI35
approximately one week after subarachnoidal hemor-
rhage has been observed in ventricular CSF  and a
recent study showed that NfHSMI35levels increased in
serum three weeks after stroke . From a practical
point of view, a later maximum level of NfHSMI35(> 1
week) is of less value for the acute prognostication and
the decision on level of care. As the sensitivity of estab-
lished methods for neurological prognostication, such as
SSEP, is limited, biomarkers, such as serum NfHSMI35
may provide complimentary information. We did not,
however, find NfHSMI35levels to be related to a lost
SSEP N20 potential or to the serum levels of NSE. In a
previous study we found NSE to correlate well with
neurological outcome , SSEP, diffusion-weighted
MRI and also neuropathology in a fraction of the pre-
sent cohort . It is possible that the heterogeneous
patient population and methodological limitations
further discussed below, might have affected the results
of the analyses. A future, larger prospective study with a
prolonged sampling may allow to address these issues.
Such a study should also include an appropriate control
group without neurological co-morbidities to allow for
better calculation of a cut-off level.
Limitations of the study
In this study, the neurofilament analyses were per-
formed on stored samples and the results were not
available during the treatment and evaluation of
patients, and have thus not influenced clinical decision-
making. We encountered large variability of the plasma
levels of NfH, which can be noted in the range of values
in Table 3. Occasional values in one patient were high
without an apparent clinical correlate and occasional
patients with definite brain damage did not have mea-
surable plasma levels of NfH. This variability needs to
be taken into account when planning future studies.
Long-term storing and handling of samples may have
affected NfHSMI35levels and contributed to the
As presented in the results section, some (n = 15) of
the patients had neurological comorbidities. Importantly,
in all of these co-morbidities, an increase of NfHSMI35
levels has been described. In an experimental setting
these limitations can be overcome, revealing a more
meaningful temporal time course of blood NfHSMI35
levels , but this is not realistic for clinical practice.
The relevance of pre-existing neurological co-morbidity
is further underlined by a post-hoc analysis showing
Rundgren et al. Critical Care 2012, 16:R45
Page 5 of 7
significant difference in plasma NfHSMI35levels 36 hours
after cardiac arrest between the patients with previous
neurological disease compared with the ones without (P
= 0.018, n = 76). We speculate that an already injured
brain may be more vulnerable to ischemic damage fol-
lowing cardiac arrest but the present study was under-
powered to investigate this post-hoc hypothesis. The
presence of anti-neurofilament auto-antibodies may
have masked the relevant binding epitopes for the
ELISA causing false-negative results  but this was
not investigated. Could injury to the peripheral nervous
system have caused an increase of plasma NfHSMI35
levels? This is a possibility and has been discussed else-
where . Particularly in patients with significant vas-
cular co-morbidity and diabetes there is an increased
likelihood for presence of coexisting peripheral neuropa-
thy. Taken together these pre-analytical and analytical
problems need to be taken into account for the inter-
pretation of plasma NFHSMI35levels.
Plasma NfHSMI35levels were quantifiable in comatose
patients after cardiac arrest. The plasma levels were sig-
nificantly higher levels in the poor outcome group as
compared with the good outcome group in the early
phase after cardiac arrest when sedation limits the clini-
cal assessment. Due to a considerable overlap between
the groups no clinically relevant cut-off level could be
proposed. It is too early to recommend the routine use
of plasma NfHSMI35in this context for clinical decision-
• NfHSMI35is quantifiable in plasma after cardiac
• Levels are significantly higher in the poor outcome
group at 2 and 36 hours after cardiac arrest.
• This exploratory study needs validation in a larger
and more homogenous cohort before plasma NfH can
be recommended for routine neuroprognostication.
aEEG: amplitude integrated electroencephalogram; AUC: area under curve;
CI: confidence interval; CPC: cerebral performance categories scale; CSF:
cerebrospinal fluid; DW-MRI: diffusion-weighted magnetic resonance
imaging; ELISA: enzyme-linked immunosorbent assay; GCS: Glasgow coma
scale; GFAP: glial fibrillary acidic protein; IQR: inter quartile range; NfH:
neurofilament heavy chain; NfL: neurofilament light chain; NfM:
neurofilament medium chain; NSE: neuronspecific enolase; ROC: receiver
operating characteristic; ROSC: return of spontaneous circulation; SSEP:
somatosensory evoked potential.
This study was supported by ALF (Academic Learning and Research grants),
Lund University Medical Faculty to Hans Friberg and Tobias Cronberg.
Regional Research Support, Region Skåne, Skåne University Hospital, to Malin
Rundgren, Hans Friberg, and Tobias Cronberg, and A-L & S-E Lundgrens
1Department of Intensive-and Perioperative Care, Skåne University Hospital;
Department of Clinical Sciences Lund University, Lund, Sweden.
2Department of Neurology, Skåne University Hospital; Department of Clinical
Sciences, Lund University, Lund, Sweden.3Department of Neurosurgery,
Copenhagen University, Copenhagen, Denmark.4Department of
Neuroimmunology, UCL Institute of Neurology, Queen Square, London, UK.
5Deptartment of Neurology, VU medisch centrum Amsterdam, Amsterdam,
MR and HF performed the collection of patient data, informed consents and
plasma samples. TC was primarily responsible for the neurological
assessment, BR combined the clinical and biochemical data sets to keep the
analyst blinded to clinical outcome, AP performed the NfH analyses. HF and
AP had the ideas for this study. All authors contributed in the writing
process. All authors read and approved the final manuscript.
The authors declare that they have no competing interests.
Received: 17 October 2011 Revised: 19 January 2012
Accepted: 12 March 2012 Published: 12 March 2012
1.Nolan J, Neumar R, Adrie C, Aibiki M, Berg R, Böttinger B, Callaway C, Clark R,
Geocadin R, Jauch E, Kern K, Laurent I, Longstreeth W, Merchant R, Morley P,
Morrison L, Nadkarni V, Peberdy M, Rivers E, Rodriguez-Nunez A, Sellke F,
Spaulding C, Sunde K, Hoek T: Post-cardiac arrest syndrome: epidemiology,
pathophysiology, treatment, and prognostication: a Scientific Statement
from the International Liaison Committee on Resuscitation; the American
Heart Association Emergency Cardiovascular Care Committee; the
Council on Cardiovascular Surgery and Anesthesia; the Council on
Cardiopulmonary, Perioperative, and Critical Care; the Council on Clinical
Cardiology; the Council on Stroke (Part II). Int Emerg Nursing 2010, 18:8-28.
2.Laver S, Farrow C, Turner D, Nolan J: Mode of death after admission to an
intensive care unit following cardiac arrest. Intensive Care Med 2004,
3. Pulsinelli WA, Brierley JB, Plum F: Temporal profile of neuronal damage in
a model of transient forebrain ischemia. Ann Neurol 1982, 11:491-498.
4.Edgren E, Hedstrand U, Kelsey S, Sutton-Tyrrell K, Safar P: Assessment of
neurological prognosis in comatose survivors of cardiac arrest. BRCT I
Study Group. Lancet 1994, 343:1055-1059.
5.Zandbergen E, Hijdra A, Koelman J, Hart A, Vos P, Verbeek M, de Haan R,
PROPAC Study Group: Prediction of poor outcome within the first 3 days
of postanoxic coma. Neurology 2006, 66:62-68.
6.Wijdicks EF, Hijdra A, Young G, Bassetti CL, Wiebe S, Quality Standards
Subcommittee of the American Academy of Neurology: Practice
parameter: prediction of outcome in comatose survivors after
cardiopulmonary resuscitation (an evidence-based review): report of the
Quality Standards Subcommittee of the American Academy of
Neurology. Neurology 2006, 67:203-210.
7. Tortorici M, Kochanek P, Poloyac S: Effects of hypothermia on drug
disposition, metabolism, and response: A focus of hypothermia
-mediated alterations on the cytochrome P450 enzyme system. Crit Care
Med 2007, 35:2196-2204.
8.Rossetti AO, Oddo M, Logroscino G, Kaplan PW: Prognostication after cardiac
arrest and hypothermia: a prospective study. Ann Neurol 2010, 67:301-307.
9.Thenayan EA, Savard M, Sharpe M, Norton L, Young B: Predictors of poor
neurologic outcome after induced mild hypothermia following cardiac
arrest. Neurology 2008, 71:1535-1537.
10.Cronberg T, Rundgren M, Westhall E, Englund E, Siemund R, Rosén I,
Widner H, Friberg H: Neuron-specific enolase correlates with other
prognostic markers after cardiac arrest. Neurology 2011, 77:623-630.
11.Kaneko T, Kasaoka S, Miyauchi T, Fujita M, Oda Y, Tsuruta R, Maekawa T:
Serum glial fibrillary acidic protein as a predictive biomarker of
neurological outcome after cardiac arrest. Resuscitation 2009, 80:790-794.
Rundgren et al. Critical Care 2012, 16:R45
Page 6 of 7
12. Rundgren M, Karlsson T, Nielsen N, Cronberg T, Johnsson P, Friberg H:
Neuron specific enolase and S-100B as predictors of outcome after
cardiac arrest and induced hypothermia. Resuscitation 2009, 80:784-789.
Fugate JE, Wijdicks EF, Mandrekar J, Claassen D, Manno E, White R, Bell M,
Rabinstein A: Predictors of neurologic outcome in hypothermia after
cardiac arrest. Ann Neurol 2010, 68:907-914.
Stern P, Bartos V, Uhrova J, Bezdickova D, Vanickova Z, Tichy V, Pelinkova K,
Prusa R, Zima T: Performance characteristics of seven neuron-specific
enolase assays. Tumour Biol 2007, 28:84-92.
Johnsson PBS, Luhrs C, Malmkvist G, Alling C, Solem JO, Stahl E: Neuron-
specific enolase increases in plasma during and immediately after
extracorporeal circulation. Ann Thorac Surg 2000, 69:750-754.
Anderson R, Hansson L, Nilsson O, Liska J, Settergren G, Vaage J: Increase
in serum S100A1-B and S100BB during cardiac surgery arises from
extracerebral sources. Ann Thorac Surg 2001, 71:1512-1517.
Undén J, Bellner J, Eneroth M, Alling C, Ingebrigtsen T, Romner B: Raised
serum S100B levels after acute bone fractures without cerebral injury. J
Trauma 2005, 58:59-61.
Petzold A: Neurofilament phosphoforma: Surrogate markters for axomal
injury, degeneration and loss. J Neurol Sci 2005, 233:183-198.
Jensen HA, Loukogeorgakis S, Yannopoulos F, Rimipiläinen E, Petzold A,
Tuominen H, Lepola P, Macallister RJ, Deanfield JE, Mäkelä T, Alestalo K,
Kiviluoma K, Anttila V, Tsang V, Juvonen T: Remote ischemic
preconditioning protects the brain against injury after hypothermic
circulatory arrest. Circulation 2011, 123:714-721.
Kuhle J, Leppert D, Petzold A, Regeniter A, Schindler C, Mehlin M,
Anthony DC, Kappos L, Lindberg RL: Neurofilament heavy chain in CSF
correlates with relapses and disability in multiple sclerosis. Neurology
Petzold A, Michel P, Stock M, Schluep M: Glial and axonal body fluid
biomarkers are related to infarct volume, severity, and outcome. J Stroke
Cerebrovasc Dis 2008, 17:196-203.
Rosen H, Karlsson JE, Rosengren L: CSF levels of neurofilament is a
valuable predictor of long-term outcome after cardiac arrest. J Neurol Sci
Sellner J, Patel A, Dassan P, Brown M, Petzold A: Hyperacute detection of
neurofilament heavy chain in serum following stroke: a transient sign.
Neurochem Res 2011, 36(12):2287-2291.
Sellner J, Petzold A, Sadikovic A, Esposito L, Weber M, Heider P, Eckstein H-
H, Memmer B, Poppert H: The value of the serum neurofilament protein
heavy chain as a biomarker for peri-operative brain injury after carotid
endarterectomy. Neurochem Res 2009, 34:1969-1974.
Rundgren M, Rosen I, Friberg H: Amplitude-integrated EEG (aEEG) predicts
outcome after cardiac arrest and induced hypothermia. Intensive Care
Med 2006, 32:836-842.
Jennett B, Bond M: Assessment of outcome after severe brain damage.
Lancet 1975, 1:480-484.
Petzold A, Keir G, Green AJ, Giovannoni G, Thompson EJ: A specific ELISA
for measuring neurofilament heavy chain phosphoforms. J Immunol
Methods 2003, 278:179-190.
Petzold A, Verwey NA, van Uffelen K, Blankenstein MA, Teunissen C: Batch
prepared protein standards for cerebrospinal fluid (CSF) biomarkers for
neurodegeneration. J Neurosci Methods 2010, 193:296-299.
Eikelenboom MJ, Uitdehaag BMJ, Petzold A: Blood and CSF Biomarker
Dynamics in Multiple Sclerosis: Implications for Data Interpretation. Mult
Scler Int 2011, 2011:823176.
Petzold A, Tisdall MM, Girbes AR, Martinian L, Thom M, Kitchen N, Smith M:
In vivo monitoring of neuronal loss in traumatic brain injury: a
microdialysis study. Brain 2011, 134:464-483.
Petzold A, Keir G, Kay A, Kerr M, Thompson EJ: Axonal damage and
outcome in subarachnoid haemorrhage. J Neurol Neurosurg Psychiatry
Singh P, Yan J, Hull R, Read S, O’Sullivan J, Henderson RD, Rose S, Greer JM,
McCombe PA: Levels of phosphorylated axonal neurofilament subunit H
(pNfH) are increased in acute ischemic stroke. J Neurol Sci 2011,
Cite this article as: Rundgren et al.: Serial soluble neurofilament heavy
chain in plasma as a marker of brain injury after cardiac arrest. Critical
Care 2012 16:R45.
Submit your next manuscript to BioMed Central
and take full advantage of:
• Convenient online submission
• Thorough peer review
• No space constraints or color figure charges
• Immediate publication on acceptance
• Inclusion in PubMed, CAS, Scopus and Google Scholar
• Research which is freely available for redistribution
Submit your manuscript at
Rundgren et al. Critical Care 2012, 16:R45
Page 7 of 7