The n e w e ng l a n d j o u r na l of m e dic i n e

Original Article

Hypothermia for Neuroprotection

in Convulsive Status Epilepticus
Stephane Legriel, M.D., Virginie Lemiale, M.D., Maleka Schenck, M.D.,
Jonathan Chelly, M.D., Virginie Laurent, M.D., Fabrice Daviaud, M.D.,
Mohamed Srairi, M.D., Aicha Hamdi, M.D., Guillaume Geri, M.D., Ph.D.,
Thomas Rossignol, M.D., Julia HillyGinoux, M.D., Julie BoisramHelms, M.D.,
Benjamin Louart, M.D., Isabelle Malissin, M.D., Nicolas Mongardon, M.D., Ph.D.,
Benjamin Planquette, M.D., Ph.D., Marina Thirion, M.D., Sybille Merceron, M.D.,
Emmanuel Canet, M.D., Ph.D., Fernando Pico, M.D., Ph.D.,
YvesRoger TranDinh, M.D., Ph.D., JeanPierre Bedos, M.D., Ph.D.,
Elie Azoulay, M.D., Ph.D., Matthieu RescheRigon, M.D., Ph.D.,
and Alain Cariou, M.D., Ph.D., for the HYBERNATUS Study Group*

A BS T R AC T

BACKGROUND
Convulsive status epilepticus often results in permanent neurologic impairment. The authors affiliations are listed in the
We evaluated the effect of induced hypothermia on neurologic outcomes in patients Appendix. Address reprint requests to
Dr. Legriel at the MedicalSurgical Inten-
with convulsive status epilepticus. sive Care Unit, Centre Hospitalier de Ver-
saillesSite Andr Mignot, 177 rue de
METHODS Versailles, 78150 Le Chesnay CEDEX,
In a multicenter trial, we randomly assigned 270 critically ill patients with convul- France, or at slegriel@ch-versailles.fr.
sive status epilepticus who were receiving mechanical ventilation to hypothermia * A complete list of the members of the
(32 to 34C for 24 hours) in addition to standard care or to standard care alone; HYBERNATUS (Hypothermia for Brain
Enhancement Recovery by Neuroprotec-
268 patients were included in the analysis. The primary outcome was a good func- tive and Anticonvulsivant Action after
tional outcome at 90 days, defined as a Glasgow Outcome Scale (GOS) score of 5 Convulsive Status Epilepticus) study
(range, 1 to 5, with 1 representing death and 5 representing no or minimal neuro- group is provided in the Supplementary
Appendix, available at NEJM.org.
logic deficit). The main secondary outcomes were mortality at 90 days, progression
to electroencephalographically (EEG) confirmed status epilepticus, refractory status N Engl J Med 2016;375:2457-67.
DOI: 10.1056/NEJMoa1608193
epilepticus on day 1, super-refractory status epilepticus (resistant to general Copyright 2016 Massachusetts Medical Society.
anesthesia), and functional sequelae on day 90.
RESULTS
A GOS score of 5 occurred in 67 of 138 patients (49%) in the hypothermia group
and in 56 of 130 (43%) in the control group (adjusted common odds ratio, 1.22;
95% confidence interval [CI], 0.75 to 1.99; P=0.43). The rate of progression to
EEG-confirmed status epilepticus on the first day was lower in the hypothermia
group than in the control group (11% vs. 22%; odds ratio, 0.40; 95% CI, 0.20 to
0.79; P=0.009), but there were no significant differences between groups in the
other secondary outcomes. Adverse events were more frequent in the hypothermia
group than in the control group.
CONCLUSIONS
In this trial, induced hypothermia added to standard care was not associated with
significantly better 90-day outcomes than standard care alone in patients with con-
vulsive status epilepticus. (Funded by the French Ministry of Health; HYBERNATUS
ClinicalTrials.gov number, NCT01359332.)

n engl j med 375;25nejm.org December 22, 2016 2457

The New England Journal of Medicine
Downloaded from nejm.org on July 22, 2017. For personal use only. No other uses without permission.
Copyright 2016 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l
C
onvulsive status epilepticus is
among the most challenging life-threat-
ening neurologic emergencies.1 Antiepi-
leptic drug treatment is initiated rapidly and
adjusted in response to clinical and electroen-
cephalographic (EEG) findings.2 The in-hospital
mortality associated with this condition is ap-
proximately 20% and, if status epilepticus re-
mains refractory to treatment, can be as high as
40%.3-7 Functional impairments after status epi-
lepticus may be due to the effect of seizures on
cortical neurons or to cerebral damage due to the
underlying process that caused the seizures. In a
prospective, multicenter study, 50% of survivors
of convulsive status epilepticus who were admit-
ted to an intensive care unit (ICU) had residual
functional impairments 90 days after the event.
Factors independently associated with poorer out-
comes were older age, longer duration of motor
seizures, focal neurologic signs at the onset, and
refractory seizures,8 all of which support the
need for neuroprotective strategies for patients
with status epilepticus.2,9
Therapeutic hypothermia has antiepileptic and
neuroprotective properties in studies of seizures
in animals. In humans, therapeutic hypothermia
has been used as adjuvant treatment for super-
refractory seizures that are resistant to general
anesthetic agents.10 Therapeutic hypothermia has
also been tested as a neuroprotective treatment
in diseases that underlie status epilepticus, such
as ischemic or hemorrhagic stroke,11-15 subarach-
noid hemorrhage, and traumatic brain injury,16-20
but the results have generally been negative. We
conducted a randomized, controlled trial to de-
termine whether therapeutic hypothermia (32 to
34C for 24 hours) added to standard care would
yield better neurologic outcomes than standard
care alone in patients with convulsive status
epilepticus.
Me thods
Trial Design and Oversight
The HYBERNATUS (Hypothermia for Brain En-
hancement Recovery by Neuroprotective and Anti-
convulsivant Action after Convulsive Status
Epilepticus) trial was a multicenter, open-label,
parallel-group, randomized, controlled trial that
was conducted in 11 French ICUs from March
2011 through January 2015. The trial protocol
2458 n engl j med 375;25nejm.org December 22, 2016
The New England Journal of Medicine
Downloaded from nejm.org on July 22, 2017. For personal use only. No other uses without permission.
Copyright 2016 Massachusetts Medical Society. All rights reserved.
of m e dic i n e
has been published previously21 and is available
with the full text of this article at NEJM.org. Ap-
proval was obtained from an independent ethics
committee (Comit de Protection des Personnes
dIle de France IV, Saint Louis) and from the
French health authorities (Agence Franaise de
Scurit Sanitaire des Produits de Sant). A steer-
ing committee provided guidance throughout the
trial. An independent data and safety monitoring
board (see the Supplementary Appendix, available
at NEJM.org) reviewed the data and performed a
scheduled, blinded interim analysis. The trial was
investigator-initiated, and there was no industry
support or involvement. The authors vouch for
the conduct of the trial, adherence to the proto-
col, and the accuracy and completeness of the
data and analyses.
The trial complied with the provisions of the
Declaration of Helsinki, Good Clinical Practice
guidelines, and French regulatory requirements.
The patients were unable to provide informed
consent at inclusion. In accordance with French
law, the ethics committee waived the requirement
for obtaining informed consent from patients be-
cause of the emergency setting of the research.
Before inclusion, investigators sought consent for
participation from surrogates. If the surrogates
chose not to give consent, the patients were not
included. Written consent was then requested
from all included patients as soon as they re-
gained competence.
Inclusion and Exclusion Criteria
All patients older than 18 years of age who were
admitted with any seizure activity to the partici-
pating ICUs were screened for eligibility by the
ICU physicians, and a log of all screened patients
was retained. Eligible patients had convulsive
status epilepticus, defined as 5 minutes or more of
continuous clinical seizure activity or more than
two seizures without a return to baseline in the
interval; had been admitted less than 8 hours
after the onset of seizures; and were receiving
mechanical ventilation. Exclusion criteria were
full recovery from seizure (defined as a return to
the baseline state of consciousness), a need for
emergency surgery that would preclude thera-
peutic hypothermia, postanoxic status epilepticus,
imminent death, and do-not-resuscitate orders;
an exclusion for bacterial meningitis was added
in January 2013, after the results of a study indi-
 Hypothermia in Convulsive Status Epilepticus

cated that moderate hypothermia might be harm- 48 hours after randomization. Further details of
ful in persons with severe bacterial meningitis.22 treatments are provided in the Supplementary
Appendix.
Randomization
Consecutive eligible patients were assigned in a
1:1 ratio to one of the two treatment groups with
the use of a computer-generated scheme with
permuted blocks and stratification according to
center, age (65 or >65 years), and seizure dura-
tion (60 or >60 minutes). Concealment of trial-
group assignments was ensured by the use of a
secure, noninteractive Web-based system that was
accessible from each trial center and that was
managed by the biostatistical unit of the Saint
Louis University Hospital (Paris), which had no
other role in patient recruitment.
Trial Intervention
In the hypothermia group, the objective was to
lower the core body temperature to 32 to 34C as
rapidly as possible after randomization and to
maintain this target temperature for 24 hours.
Hypothermia was induced with ice-cold intrave-
nous fluids at 4C and was maintained with ice
packs at the groin and neck and a cold-air tunnel
around the patients body (for details, see the
Methods section in the Supplementary Appen-
dix). Sedation was implemented with propofol
and neuromuscular blockade. In the control group,
if the physician determined that sedation was
needed, a propofol regimen that was the same as
that in the therapeutic hypothermia group was
used. In both groups, continuous EEG monitor-
ing was started within 2 hours after random-
ization and was maintained for 48 hours after
randomization or until body temperature was
normalized in the hypothermia group.23 A 30-
minute standard EEG segment extracted from
the continuous EEG recording was interpreted at a
central site by a neurophysiologist within 2 hours
after the initiation of continuous EEG. In both
groups, if the results showed ongoing seizures,
repeated propofol boluses were administered,
followed by a maintenance intravenous infusion
to obtain and maintain a burst-suppression EEG
pattern for 24 hours. Core body temperature
was monitored continuously with the use of an
esophageal probe and was recorded every 4 hours
during the first 48 hours after randomization.
Patients were screened for propofol-related in-
fusion syndrome every 8 hours during the first
Outcomes and Assessments
The primary outcome was the absence of func-
tional impairment 90 days after status epilepti-
cus, as defined by a score of 5 on the Glasgow
Outcome Scale (GOS), indicating survival with
good function that allowed a return to former
occupational or academic activities, with or with-
out minor physical or mental deficits (the GOS
ranges from 1 to 5, with 1 representing death
and 5 representing no or minimal neurologic
deficit). The score was determined during a struc-
tured interview that was conducted at 907 days
by an independent assessor.
Secondary outcomes were the rates of death
in the ICU, in the hospital, and by day 90; pro-
gression to EEG-confirmed status epilepticus,
defined as coma with or without subtle convul-
sive movements but with generalized or lateral-
ized ictal discharges on the EEG between 6 and
12 hours after randomization; refractory status
epilepticus on day 1, defined as continuous or
intermittent clinical seizures, EEG-confirmed sei-
zures, or both despite two types of antiepileptic
drugs within 24 hours after the onset of status
epilepticus; super-refractory status epilepticus,
defined as ongoing or recurrent status epilepti-
cus between 24 and 48 hours after the initiation
of anesthetic treatment; total seizure duration;
lengths of stay in the ICU and in the hospital;
and, in a subset of patients, impairments on day
90 (as determined by new seizures, recurrent
status epilepticus after discharge, the number of
antiepileptic drugs being used, and the score
on the MiniMental State Examination [MMSE;
range, 0 to 30, with higher scores representing
better performance]).
Physicians who assessed outcomes, trial ad-
ministrators, and statisticians were unaware of the
trial-group assignments. At 907 days, 60 of the
270 patients agreed to an additional evaluation
by a neurologist who was unaware of the trial-
group assignments; the evaluation included an
interview and determination of the GOS and
MMSE scores. Before this visit, the patients un-
derwent magnetic resonance imaging of the brain
and conventional EEG at their local hospitals and
brought the scans and EEG results to the visit.

n engl j med 375;25nejm.org December 22, 2016 2459

The New England Journal of Medicine
Downloaded from nejm.org on July 22, 2017. For personal use only. No other uses without permission.
Copyright 2016 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l
The neurologist noted from patient interviews
and hospital charts any episodes of recurrent
convulsive status epilepticus since randomiza-
tion; the frequency of any recurrent seizures, as
well as their type and associated symptoms; and
the nature and dose of antiepileptic drugs and
other treatments.
Statistical Analysis
All analyses were conducted according to a pre-
viously published statistical analysis plan.21 We
calculated that 135 patients per group would
need to be enrolled for the trial to have 90%
power to detect an absolute difference of 20 per-
centage points in the percentage of patients with
a GOS score of 5 on day 90 (40% in the standard-
care group4,8 vs. 60% in the hypothermia group),
at an alpha level of 0.05.
A single scheduled interim analysis was per-
formed to assess efficacy after enrollment of
50% of the planned cohort, with the use of a
two-sided, symmetric OBrienFleming design
and a two-sided alpha level of 0.005. This analy-
sis was reviewed by the independent data and
safety monitoring board. It yielded a P value of
0.18, a result that led to continuation of the trial
and enrollment of the full sample.
Continuous variables are summarized as me-
dians and interquartile ranges and categorical
variables as counts and percentages. The primary
outcome was compared between groups with the
use of a logistic model stratified according to
the randomization stratification variables with
the use of a Wald test. Other binary outcomes
(mortality in the ICU, in-hospital mortality, the
percentage of patients with refractory status epi-
lepticus, total seizure duration, and recurrence
of status epilepticus after discharge) were simi-
larly analyzed. Survival within each group was esti-
mated by the KaplanMeier method with censoring
on day 90, and survival rates were compared
with the use of the log-rank test. Median dura-
tions of hospital and ICU stays were estimated
in both groups with the use of a KaplanMeier
estimator and tested with the use of a log-rank
test, with discharge alive as the event of interest
and death as the censoring event. Total seizure
duration, number of antiepileptic drugs, and
MMSE score were compared with the use of the
Wilcoxon rank-sum test. Seizure and status epi-
lepticus recurrences were compared with the use
of the Fishers exact test. We performed 13 tests
for secondary outcomes; thus, fewer than 1 test
2460 n engl j med 375;25nejm.org December 22, 2016
The New England Journal of Medicine
Downloaded from nejm.org on July 22, 2017. For personal use only. No other uses without permission.
Copyright 2016 Massachusetts Medical Society. All rights reserved.
of m e dic i n e
would be expected to have a significant result
(P<0.05) on the basis of chance alone.
We conducted exploratory analyses of the pri-
mary outcome in subgroups, building logistic-
regression models to compare odds ratios, with
their 95% confidence intervals, for achievement
of a GOS score of 5. We first searched for quan-
titative interaction between the intervention and
randomization stratification variables using the
Gail and Simon heterogeneity test.24 In case of
opposite direction in odds ratios, qualitative in-
teraction was further tested with the use of the
Silvapulle test.25
All analyses were performed according to the
intention-to-treat principle; missing values were
handled with the use of the last-observation-
carried-forward approach. Sensitivity analyses of
the primary outcome were performed, including
the use of multiple imputation for missing data
by means of chained equations (20 data sets im-
puted) and a per-protocol approach that included
all patients who received the intervention and
were enrolled in the trial only once.
All reported P values are two-sided, and P values
of less than 0.05 were considered to indicate statis-
tical significance. Analyses were performed with
R software, version 3.3.1 (www.R-project.org/).
R e sult s
Trial Patients
From March 2011 through January 2015, a total
of 270 patients were enrolled, 2 of whom with-
drew consent; of the 268 patients who were in-
cluded in the analysis, 138 were assigned to the
hypothermia group and 130 to the control group
(Fig. S1 in the Supplementary Appendix). One
patient in each group had a second episode of
status epilepticus, and each was enrolled twice.
Demographic and clinical characteristics at
baseline were similar in the two treatment groups
(Table1). The median age was 57 years (inter-
quartile range, 45 to 69), and 173 patients (65%)
were male. A total of 130 patients (49%) had a
history of epilepsy. Most episodes of status epi-
lepticus (173 of 268; 65%) began outside the
hospital. The median time from seizure onset
to the initiation of an antiepileptic drug was 40
minutes (interquartile range, 10 to 75), and a
median of 2 drugs (interquartile range, 1 to 3)
were administered before the ongoing seizures
were controlled. The median time to the control
of electrical seizure activity was 80 minutes
 Hypothermia in Convulsive Status Epilepticus

Table 1. Baseline Characteristics.*

Characteristic Hypothermia (N=138) Control (N=130)

Demographic characteristics
Age yr
Median 57 57
Interquartile range 4467 4271
Age >65 yr no. (%) 40 (29) 43 (33)
Male sex no. (%) 95 (69) 78 (60)
Medical history
History of epilepsy no./total no. (%) 71/138 (51) 59/129 (46)
Use of long-term antiepileptic therapy no./total no. (%) 55/71 (77) 46/59 (78)
Characteristics of the status epilepticus at randomization
Out-of-hospital onset no. (%) 90 (65) 83 (64)
Bystander-witnessed onset no. (%) 120 (87) 113 (87)
Seizure type no. (%)
Primary or secondary generalized seizures 124 (90) 110 (85)
Partial motor seizures 14 (10) 20 (15)
Tonicclonic motor manifestations no./total no. (%) 109/135 (81) 105/130 (81)
Continuous seizures no. (%) 63 (46) 72 (55)
Focal neurologic signs at the scene no. (%) 47 (34) 40 (31)
Time from seizure onset to initiation of antiepileptic drug min
Median 40 40
Interquartile range 785 1072
Refractory status epilepticus no. (%) 32 (23) 35 (27)
Glasgow Coma Scale score
Median 4 4
Interquartile range 37 37
Core temperature C
Median 37.0 37.0
Interquartile range 36.437.7 36.137.7
First-line antiepileptic drug no. (%) 138 (100) 130 (100)
Benzodiazepine alone 115 (83) 108 (83)
Phenobarbital, fosphenytoin, phenytoin, valproate sodium, or levetiracetam 6 (4) 11 (8)
Midazolam, propofol, or thiopental 16 (12) 9 (7)
Second-line antiepileptic drug no. (%) 84 (61) 71 (55)
Etiologic category of status epilepticus no. (%)
Acute symptomatic 98 (71) 85 (65)
Remote symptomatic 6 (4) 8 (6)
Progressive symptomatic 4 (3) 5 (4)
Unknown 30 (22) 32 (25)

* Patients in the hypothermia group were assigned to induced hypothermia (32 to 34C) plus standard care, and patients in the control group
were assigned to standard care alone. Between-group differences were not assessed for statistical significance.
Focal neurologic signs were defined as symptoms or signs consistent with damage to, or dysfunction of, a specific site in the central nervous
system. Signs were described as unifocal or multifocal and as transient or persistent.
Refractory status epilepticus at randomization was defined as continuous or intermittent seizures despite treatment with an intravenous benzo-
diazepine (clonazepam, diazepam, or midazolam) and intravenous fosphenytoin, phenytoin, phenobarbital, or valproate sodium.
Scores on the Glasgow Coma Scale range from 3 to 15, with lower scores indicating a reduced level of consciousness. Data were missing
for three patients in the hypothermia group and four patients in the control group.
Data were missing for one patient in the control group.
Etiologic categories of status epilepticus are reported as defined by the 2015 report of the International League against Epilepsy Task Force on Classi
fication of Status Epilepticus. The acute symptomatic category corresponds to status epilepticus occurring in close temporal relationship with an
acute central nervous system disorder (e.g., metabolic, toxic, structural, infectious, or owing to inflammation). The remote symptomatic category
corresponds to status epilepticus related to a past central nervous system injury that is stable over time. The progressive symptomatic category corre-
sponds to status epilepticus related to a progressive and degenerative central nervous system injury.

n engl j med 375;25nejm.org December 22, 2016 2461

The New England Journal of Medicine
Downloaded from nejm.org on July 22, 2017. For personal use only. No other uses without permission.
Copyright 2016 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

(interquartile range, 40 to 210). At the time of (Table2). Sensitivity analyses, including the
randomization, status epilepticus was refractory, as overall distribution of GOS scores and imputa-
previously defined, in 67 patients (25%). The me- tion for missing data, showed similar results
dian time to first EEG interpretation was 6.6 hours (Table S4 in the Supplementary Appendix).
(interquartile range, 5.4 to 8.3). Table S1 in the
Supplementary Appendix provides details on the Secondary Outcomes
management of status epilepticus in the trial. Mortality in the ICU, in-hospital mortality, and
90-day mortality did not differ significantly be-
Temperature Management tween the two groups (Table2, and Fig. S4 in
The esophageal mean core temperature at ran- the Supplementary Appendix). A total of 44 pa-
domization was 37.0C (interquartile range, 36.4 tients (15 in the hypothermia group and 29 in
to 37.7) in the hypothermia group and 37.0C the control group) had progression to EEG-con-
(interquartile range, 36.1 to 37.7) in the control firmed status epilepticus, designated as one of
group. Cooling was initiated a median of 5.8 hours four main secondary outcomes (odds ratio, 0.40;
(interquartile range, 3.5 to 8.0) after seizure 95% CI, 0.20 to 0.79; P=0.009). No other sec-
onset. The target temperature (32 to 34C) was ondary efficacy outcomes differed significantly
reached in 135 of 138 patients (98%), within a between the two groups (Table2). Details of
median of 5.2 hours (interquartile range, 3.5 to 7.1). functional sequelae at day 90 were assessed by a
Figure1 shows core body temperatures during neurologist during an additional face-to-face
the first 48 hours after randomization. evaluation in 60 of the 270 included patients;
hospital charts allowed the collection of data
Primary Outcome regarding seizure recurrence or status epilepticus
In the intention-to-treat analysis, a GOS score of recurrence within 90 days in up to 19 additional
5 occurred in 67 of 138 patients (49%) in the patients (Table2).
hypothermia group and in 56 of 130 patients
(43%) in the control group (odds ratio, 1.22; 95% Subgroup Analysis According to Age
confidence interval [CI], 0.75 to 1.99; P=0.43) The trial was not powered to find significant ef-
fects according to age. The intervention showed
a significant quantitative interaction with age
39
subgroup (P=0.02) that is, a heterogeneity in
38 Control

the intervention effect according to age. The odds

ratio of a GOS score of 5 in patients who were

37




Core Temperature (C)

36
treated with hypothermia as compared with con-
trols was 1.75 (95% CI, 0.98 to 3.16) among pa-

35

tients 65 years of age or younger, whereas it was

34

0.49 (95% CI, 0.19 to 1.25) among patients older

33 Hypothermia

32 than 65 years of age (Fig.2).

31
Adverse Events
30
One or more adverse events of any grade of se-
0
0 4 8 12 16 20 24 28 32 36 40 44 48 verity occurred in 117 of 138 patients (85%) in the
Hours since Randomization hypothermia group and in 100 of 130 patients
(77%) in the control group. Pneumonia attrib-
Figure 1. Body Temperature during the Intervention Period. uted by the investigators to aspiration occurred
Shown are the mean esophageal core temperatures in the group assigned in 51% of the patients in the hypothermia group
to induced hypothermia (32 to 34C) plus standard care and in the group and in 45% of the patients in the control group.
assigned to standard care alone (control). Data were obtained from con
Of the 261 patients who received propofol, 1 pa-
tinuous recordings but are indicated here at 4-hour intervals. The area
between the dashed lines is the 32 to 34C range. At randomization, the tient (in the hypothermia group) had propofol-
median temperature in both groups was 37C (dotted line). I bars indicate related infusion syndrome. Table S2 in the Sup-
95% confidence intervals. plementary Appendix lists adverse events. There
was no significant difference between the groups

2462 n engl j med 375;25nejm.org December 22, 2016

The New England Journal of Medicine

Downloaded from nejm.org on July 22, 2017. For personal use only. No other uses without permission.
Copyright 2016 Massachusetts Medical Society. All rights reserved.
 Hypothermia in Convulsive Status Epilepticus

Table 2. Primary and Secondary Outcomes.*

Hypothermia Control Odds Ratio P

Outcome (N=138) (N=130) (95% CI) Value
Primary outcome: GOS score of 5 at day 90 67 (49) 56 (43) 1.22 (0.751.99) 0.43
no. (%)
Secondary outcomes
Total seizure duration min
Median 75 90 0.26
Interquartile range 37180 45255
Progression to EEG-confirmed status epilepticus 15 (11) 29 (22) 0.40 (0.200.79) 0.009
no. (%)
Refractory status epilepticus from day 1 to day 3
no. (%)
Refractory status epilepticus on day 1 43 (31) 50 (38) 0.68 (0.401.15) 0.15
Super-refractory status epilepticus 23 (17) 30 (23) 0.64 (0.341.19) 0.16
Length of ICU stay days
Median 8 7 0.44
Interquartile range 514 316
Length of hospital stay days
Median 21 19 0.89
Interquartile range 1038 1040
Death in ICU no. (%) 13 (9) 15 (12) 0.83 (0.381.82) 0.64
Death in hospital, including in ICU no. (%) 17 (12) 20 (15) 0.81 (0.401.64) 0.55
Death between randomization and 90 days after 18 (13) 20 (15) 0.86 (0.431.72) 0.67
discharge no. (%)
Functional impairments within 90 days
No. of antiepileptic drugs on day 90**
Median 1 1 0.63
Interquartile range 01 01
Seizure recurrence within 90 days 5/45 (11) 1/34 (3) 0.22
no./total no. (%)
Status epilepticus recurrence within 90 days 4/44 (9) 1/34 (3) 0.38
no./total no. (%)
MMSE score on day 90
Median 26 25 0.27
Interquartile range 2429 2128

* ICU denotes intensive care unit.

Odds ratios were stratified according to age (>65 or 65 years) and seizure duration at inclusion (>60 or 60 minutes).
Scores on the Glasgow Outcome Scale (GOS) range from 1 to 5, with 1 representing death and 5 representing no or
minimal neurologic deficit.
EEG-confirmed status epilepticus was diagnosed when the patient was found in a coma with or without subtle con-
vulsive movements but with generalized or lateralized ictal discharges on the EEG between 6 and 12 hours after ran-
domization.
Refractory status epilepticus on day 1 was defined as continuous or intermittent clinical seizures, EEG-confirmed sei-
zures, or both despite two lines of antiepileptic drugs within 24 hours after the onset of status epilepticus.
Super-refractory status epilepticus was defined as ongoing or recurrent status epilepticus between 24 and 48 hours
after the initiation of anesthetic treatment.
** Data were available for 33 patients in the hypothermia group and 27 patients in the control group.
Scores on the MiniMental State Examination (MMSE) range from 0 to 30, with higher scores indicating better per-
formance. Data were available for 33 patients in the hypothermia group and 27 patients in the control group.

n engl j med 375;25nejm.org December 22, 2016 2463

The New England Journal of Medicine
Downloaded from nejm.org on July 22, 2017. For personal use only. No other uses without permission.
Copyright 2016 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

Control Hypothermia Odds Ratio for Primary Outcome P Value for

Subgroup Group Group (95% CI) Interaction
no. with primary outcome/total no.
All patients 56/130 67/138 1.22 (0.751.99)
Seizure duration at 0.62
randomization
60 min 46/99 52/103 1.17 (0.682.05)
>60 min 10/31 15/35 1.58 (0.584.41)
Age at randomization 0.02
65 yr 39/87 57/97 1.75 (0.983.16)
>65 yr 17/43 10/41 0.49 (0.191.25)

0.20 0.33 0.50 0.67 1.00 1.50 2.00 3.00 5.00

Control Group Hypothermia

Better Group Better

Figure 2. Prespecified Subgroup Analyses.

The primary outcome was the absence of functional impairment at 90 days, as defined by a score of 5 on the Glasgow Outcome Scale
(GOS; range, 1 to 5, with 1 representing death and 5 representing no or minimal neurologic deficit). The sizes of the data markers reflect
the relative size of each subgroup. P values were calculated with the use of the Gail and Simon test.

in the number of deaths in the hospital or during seizure cessation or for the achievement of a
the subsequent 90 days (Table2). burst-suppression EEG pattern in patients with
refractory status epilepticus.33
The finding that 43% of the patients in the
Discussion
control group had a GOS score of 5 was consis-
The results of this trial do not support a benefi- tent with our working hypothesis that outcomes
cial effect of therapeutic hypothermia as com- after status epilepticus are frequently unsatisfac-
pared with standard care alone in patients with tory, and this finding is consistent with results
convulsive status epilepticus who were receiving of previous studies.4,8 With respect to the inten-
mechanical ventilation. The percentage of pa- sity of treatment applied during the trial, we
tients with a GOS score of 5, signifying no or attempted to attain a burst-suppression EEG pat-
minimal neurologic deficit, was similar in the tern and usually succeeded. Among all patients,
two groups, and the overall distribution of GOS the median time from seizure onset to the ini-
scores was similar with and without the addi- tiation of an antiepileptic drug was 40 minutes
tion of hypothermia to antiepileptic drugs and (interquartile range, 10 to 75), and electrical
sedation. seizure activity was controlled in a median time
Therapeutic hypothermia has been investigat- of 80 minutes (interquartile range, 40 to 210).
ed as a potential neuroprotective intervention in Normothermia was maintained in the control
various forms of brain injury and has been used group to avoid fever that could have been a risk
in comatose patients after cardiac arrest.26-28 Its factor for further secondary neuronal damage.
use is controversial in traumatic brain injury,16 but The number of reported adverse events was
it is incorporated into the treatment program higher in the hypothermia group than in the
for refractory intracranial hypertension, ischemic control group, mainly because of a higher rate of
stroke, and intracerebral hemorrhage.12 It may aspiration pneumonia and minor adverse events.
be harmful in comatose patients who have bac- Aspiration has been reported in patients receiv-
terial meningitis.22 In small-animal models, ther- ing hypothermia therapy for other conditions,34
apeutic hypothermia has been shown to have but pneumonia was not associated with an in-
antiepileptic effects.29-32 In a systematic review, creased incidence of septic shock in our trial.
there was limited evidence for the benefit of Moreover, there were fewer episodes of hypovo-
therapeutic hypothermia as adjuvant therapy for lemic shock in the hypothermia group than in the

2464 n engl j med 375;25nejm.org December 22, 2016

The New England Journal of Medicine

Downloaded from nejm.org on July 22, 2017. For personal use only. No other uses without permission.
Copyright 2016 Massachusetts Medical Society. All rights reserved.
 Hypothermia in Convulsive Status Epilepticus

control group (Table S2 in the Supplementary
Appendix). This could have resulted from the
hemodynamic effect of the volume of cold intra-
venous fluid received at the initiation of hypo-
thermia.35 No significant differences in 90-day
mortality and functional impairment were seen
between the two groups. Under a protocol that
limited the dose and duration of propofol ther
apy, a single case of propofol-related infusion
syndrome occurred in the hypothermia group.36
The length of stay in the ICU was similar in the
two groups. Thus, our trial showed no overall
benefits from hypothermia in the treatment of
convulsive status epilepticus under the circum-
stances encompassed by the trial protocol.
The intervention showed a quantitative inter-
action with prespecified age subgroups in an
exploratory analysis; however, the trial was not
powered to find significant effects according to
age. The ostensible lack of benefits in older pa-
tients may have been related to the poorer out-
come of status epilepticus that was reported pre-
viously in this population.37 Antiepileptic drugs
are challenging to use in older patients owing to
differences in the pharmacokinetic and pharmaco-
dynamic properties of the drugs in these patients,
as well as interactions with medications used on
a long-term basis.37 Further studies of thera-
peutic hypothermia in young patients may be
warranted.
Refractory status epilepticus on day 1, despite
two types of antiepileptic drugs, and progression
to EEG-confirmed status epilepticus were less
common in the hypothermia group than in the
control group, which resulted in a shorter total
seizure duration in the hypothermia group, but
only progression to EEG-confirmed status epi-
lepticus reached statistical significance.38 In pa-
tients with refractory status epilepticus, ongoing
or recurrent status epilepticus between 24 and
48 hours after the initiation of anesthetic treat-
ment (super-refractory status epilepticus) was
also less common with hypothermia but did not
reach statistical significance.39 These findings are
consistent with experimental and clinical data
but apparently had no effect on ensuring a good
clinical outcome in our trial.33
Our trial has several limitations. First, the
applicability of our findings to all patients with
convulsive status epilepticus is uncertain. Patients
must be receiving mechanical ventilation before
therapeutic hypothermia can be initiated, and
we therefore excluded spontaneously breathing
patients, who account for about one third of pa-
tients with convulsive status epilepticus.40 Second,
we used external cooling techniques to sustain
hypothermia, and we cannot exclude the possi-
bility that the results would have been different
if hypothermia had been implemented earlier
with the use of intravascular devices. Although
it is possible that earlier attainment of hypother-
mia would have altered the trial results, we en-
deavored to induce hypothermia as quickly and
safely as possible. Third, a GOS score of 4 may
be viewed as a good outcome, but it nevertheless
indicates clinically meaningful functional impair-
ments. We therefore defined a good outcome as
a GOS score of 5. Finally, the choice of propofol
as a sedative and antiepileptic anesthetic drug
could conceivably have been harmful and may
have negated the effect of hypothermia. How-
ever, in the absence of strong evidence support-
ing the choice of a particular initial anesthetic
agent, propofol was used to ensure patient awak-
ening as early as possible after the end of the
intervention.
In conclusion, this trial involving critically ill
patients with convulsive status epilepticus who
were receiving mechanical ventilation showed no
significant benefit with respect to good func-
tional outcome from the addition of therapeutic
hypothermia to standard antiepileptic therapy.
Supported by a research grant from the French Ministry of
Health (Programme Hospitalier de Recherche Clinique, PHRC
AOM 09-P081249).
Disclosure forms provided by the authors are available with
the full text of this article at NEJM.org.
We thank Antoinette Wolfe, M.D. (Issy-les-Moulineaux, France),
for helping to prepare an earlier version of the manuscript.

Appendix
The authors affiliations are as follows: the MedicalSurgical Intensive Care Unit (S.L., V. Laurent, J.H.-G., B.P., S.M., J.-P.B.) and the
Neurology and Stroke Department (F.P.), Centre Hospitalier de VersaillesSite Andr Mignot, Versailles, INSERM Unit 970 (Team 4),
Paris Cardiovascular Research Center (S.L., A.C.), the Medical Intensive Care Unit (V. Lemiale, E.C., E.A.) and Department of Biostatis-
tics and Medical Information (M.R.-R.), Saint Louis University Hospital, Assistance PubliqueHpitaux de Paris (AP-HP), the Medical
Intensive Care Unit, Cochin University Hospital, Hopitaux UniversitairesParis Centre, AP-HP (F.D., G.G., N.M., A.C.), the Medical
Intensive Care Unit (I.M.) and Neurophysiology Department (Y.-R.T.-D.), Lariboisire University Hospital, AP-HP, INSERM Unit 1153

n engl j med 375;25nejm.org December 22, 2016 2465

The New England Journal of Medicine
Downloaded from nejm.org on July 22, 2017. For personal use only. No other uses without permission.
Copyright 2016 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

(ECSTRA Team), Universit Paris Diderot, Sorbonne Paris Cit (M.R.-R.), and Paris Descartes University, Sorbonne Paris CitMedical
School (A.C.), Paris, the Medical Intensive Care Unit, Hpital de Hautepierre, Hpitaux Universitaires de Strasbourg (M. Schenck), and
the Medical Intensive Care Unit, Nouvel Hpital Civil, Hpitaux Universitaires de Strasbourg, and Fdration de Mdecine Translation-
nelle de Strasbourg, Facult de Mdecine, Universit de Strasbourg (J.B.-H.), Strasbourg, the MedicalSurgical Intensive Care Unit,
Centre Hospitalier de Melun, Melun (J.C.), the Anesthesiology and Critical Care Department, Toulouse University Hospital, University
Toulouse 3 Paul Sabatier, Toulouse (M. Srairi), the MedicalSurgical Intensive Care Unit, Centre Hospitalier de Montreuil, Montreuil
(A.H.), the MedicalSurgical Intensive Care Unit, Centre Hospitalier du Mans, Le Mans (T.R.), the Intensive Care Units, Division of
Anesthesia, Intensive Care, Pain, and Emergency Medicine, University Hospital of Nmes, Nmes (B.L.), and the MedicalSurgical In-
tensive Care Unit, Centre Hospitalier Victor Dupouy, Argenteuil (M.T.) all in France.

References
1. Lowenstein DH, Alldredge BK. Status
epilepticus. N Engl J Med 1998;338:970-6.
2. Brophy GM, Bell R, Claassen J, et al.
Guidelines for the evaluation and man-
agement of status epilepticus. Neurocrit
Care 2012;17:3-23.
3. Claassen J, Lokin JK, Fitzsimmons BF,
Mendelsohn FA, Mayer SA. Predictors of
functional disability and mortality after
status epilepticus. Neurology 2002; 58:
139-42.
4. Gao Q, Ou-Yang TP, Sun XL, et al. Pre-
diction of functional outcome in patients
with convulsive status epilepticus: the
END-IT score. Crit Care 2016;20:46.
5. Rossetti AO, Lowenstein DH. Manage-
ment of refractory status epilepticus in
adults: still more questions than answers.
Lancet Neurol 2011;10:922-30.
6. Legriel S, Mourvillier B, Bele N, et al.
Outcomes in 140 critically ill patients with
status epilepticus. Intensive Care Med
2008;34:476-80.
7. Sutter R, Kaplan PW, Regg S. Inde-
pendent external validation of the Status
Epilepticus Severity Score. Crit Care Med
2013;41(12):e475-9.
8. Legriel S, Azoulay E, Resche-Rigon M,
et al. Functional outcome after convulsive
status epilepticus. Crit Care Med 2010;38:
2295-303.
9. Outin H, Blanc T, Vinatier I. Emer-
gency and intensive care unit manage-
ment of status epilepticus in adult pa-
tients and children (new-born excluded):
Socit de Ranimation de Langue Fran-
aise experts recommendations. Rev Neu-
rol (Paris) 2009;165:297-305. (In French.)
10. Legriel S, Resche-Rigon M, Cariou A.
Dual anticonvulsant and neuroprotective
effects of therapeutic hypothermia after
status epilepticus. Clin Neurol Neurosurg
2015;131:87-8.
11. Blanco M, Campos F, Rodrguez-Yez
M, et al. Neuroprotection or increased
brain damage mediated by temperature in
stroke is time dependent. PLoS One 2012;
7(2):e30700.
12. Kollmar R, Juettler E, Huttner HB,
et al. Cooling in Intracerebral Hemor-
rhage (CINCH) trial: protocol of a ran-
domized German-Austrian clinical trial.
Int J Stroke 2012;7:168-72.
13. Kollmar R, Staykov D, Drfler A,
Schellinger PD, Schwab S, Bardutzky J.
Hypothermia reduces perihemorrhagic
edema after intracerebral hemorrhage.
Stroke 2010;41:1684-9.
14. Steiner T, Bsel J. Options to restrict
hematoma expansion after spontaneous
intracerebral hemorrhage. Stroke 2010;
41:402-9.
15. Rincon F, Lyden P, Mayer SA. Relation-
ship between temperature, hematoma
growth, and functional outcome after in-
tracerebral hemorrhage. Neurocrit Care
2013;18:45-53.
16. Andrews PJD, Sinclair HL, Rodriguez
A, et al. Hypothermia for intracranial
hypertension after traumatic brain injury.
N Engl J Med 2015;373:2403-12.
17. Polderman KH. Induced hypothermia
and fever control for prevention and treat-
ment of neurological injuries. Lancet 2008;
371:1955-69.
18. Flynn LM, Rhodes J, Andrews PJ.
Therapeutic hypothermia reduces intra-
cranial pressure and partial brain oxygen
tension in patients with severe traumatic
brain injury: preliminary data from the
Eurotherm3235 Trial. Ther Hypothermia
Temp Manag 2015;5:143-51.
19. Kochanek PM, Safar PJ. Therapeutic
hypothermia for severe traumatic brain
injury. JAMA 2003;289:3007-9.
20. Sandestig A, Romner B, Grnde PO.
Therapeutic hypothermia in children and
adults with severe traumatic brain injury.
Ther Hypothermia Temp Manag 2014;4:
10-20.
21. Legriel S, Pico F, Tran-Dinh YR, et al.
Neuroprotective effect of therapeutic hypo-
thermia versus standard care alone after
convulsive status epilepticus: protocol of
the multicentre randomised controlled
trial HYBERNATUS. Ann Intensive Care
2016;6:54. arrest: characteristics, risk factors and
22. Mourvillier B, Tubach F, van de Beek
D, et al. Induced hypothermia in severe
bacterial meningitis: a randomized clini-
cal trial. JAMA 2013;310:2174-83.
23. Claassen J, Taccone FS, Horn P, Holt-
kamp M, Stocchetti N, Oddo M. Recom-
mendations on the use of EEG monitor-
ing in critically ill patients: consensus
statement from the Neurointensive Care
Section of the ESICM. Intensive Care Med
2013;39:1337-51.
24. Gail M, Simon R. Testing for qualita-
tive interactions between treatment effects
and patient subsets. Biometrics 1985;41:
361-72.
25. Silvapulle MJ. Tests against qualita-
tive interaction: exact critical values and
robust tests. Biometrics 2001;57:1157-65.
26. Bernard SA, Gray TW, Buist MD, et al.
Treatment of comatose survivors of out-
of-hospital cardiac arrest with induced hy-
pothermia. N Engl J Med 2002;346:557-63.
27. Nielsen N, Wetterslev J, Cronberg T,
et al. Targeted temperature management
at 33C versus 36C after cardiac arrest.
N Engl J Med 2013;369:2197-206.
28. The Hypothermia after Cardiac Arrest
Study Group. Mild therapeutic hypother-
mia to improve the neurologic outcome
after cardiac arrest. N Engl J Med 2002;
346:549-56.
29. Liu Z, Gatt A, Mikati M, Holmes GL.
Effect of temperature on kainic acid-induced
seizures. Brain Res 1993;631:51-8.
30. Lundgren J, Smith ML, Blennow G,
Siesj BK. Hyperthermia aggravates and
hypothermia ameliorates epileptic brain
damage. Exp Brain Res 1994;99:43-55.
31. Maeda T, Hashizume K, Tanaka T. Ef-
fect of hypothermia on kainic acid-induced
limbic seizures: an electroencephalograph-
ic and 14C-deoxyglucose autoradiographic
study. Brain Res 1999;818:228-35.
32. Schmitt FC, Buchheim K, Meierkord H,
Holtkamp M. Anticonvulsant properties
of hypothermia in experimental status
epilepticus. Neurobiol Dis 2006;23:689-96.
33. Zeiler FA, Zeiler KJ, Teitelbaum J, Gill-
man LM, West M. Therapeutic hypother-
mia for refractory status epilepticus. Can
J Neurol Sci 2015;42:221-9.
34. Perbet S, Mongardon N, Dumas F,
et al. Early-onset pneumonia after cardiac
influence on prognosis. Am J Respir Crit
Care Med 2011;184:1048-54.
35. Polderman KH, Rijnsburger ER, Peer-
deman SM, Girbes AR. Induction of hypo-
thermia in patients with various types of
neurologic injury with use of large vol-
umes of ice-cold intravenous fluid. Crit
Care Med 2005;33:2744-51.
36. Rossetti AO, Milligan TA, Vullimoz S,
Michaelides C, Bertschi M, Lee JW. A ran-
domized trial for the treatment of refrac-

2466 n engl j med 375;25nejm.org December 22, 2016

The New England Journal of Medicine

Downloaded from nejm.org on July 22, 2017. For personal use only. No other uses without permission.
Copyright 2016 Massachusetts Medical Society. All rights reserved.
 Hypothermia in Convulsive Status Epilepticus

tory status epilepticus. Neurocrit Care
2011;14:4-10.
37. Legriel S, Brophy GM. Managing sta-
tus epilepticus in the older adult. J Clin
Med 2016;5:5.
38. Treiman DM, Meyers PD, Walton NY,
et al. A comparison of four treatments for
generalized convulsive status epilepticus.
N Engl J Med 1998;339:792-8.
39. Shorvon S, Ferlisi M. The treatment
of super-refractory status epilepticus:
a critical review of available therapies and
a clinical treatment protocol. Brain 2011;
134:2802-18.
40. Rossetti AO, Bleck TP. Whats new in
status epilepticus? Intensive Care Med
2014;40:1359-62.
Copyright 2016 Massachusetts Medical Society.

n engl j med 375;25nejm.org December 22, 2016 2467

The New England Journal of Medicine
Downloaded from nejm.org on July 22, 2017. For personal use only. No other uses without permission.
Copyright 2016 Massachusetts Medical Society. All rights reserved.