IP 1115 [IPGXXX]

NATIONAL INSTITUTE FOR HEALTH AND

CARE EXCELLENCE
INTERVENTIONAL PROCEDURES PROGRAMME
Interventional procedure overview of therapeutic
hypothermia for acute ischaemic stroke
An ischaemic stroke happens when a blood clot stops the flow of blood to the
brain. Brain tissue is then damaged because it does not get enough oxygen. In
this procedure, a cooling device is used to reduce the body’s temperature by
2ºC to 4°C (creating hypothermia) for several hours immediately after a stroke.
When the brain is cooler it needs less oxygen from the blood. The aim is to
limit the damage to brain cells caused by the stroke.

Contents
Introduction
Description of the procedure
Efficacy summary
Safety summary
The evidence assessed
Validity and generalisability of the studies
Existing assessments of this procedure
Related NICE guidance
Additional information considered by IPAC
References
Literature search strategy
Appendix

Introduction
The National Institute for Health and Care Excellence (NICE) prepared this
interventional procedure overview to help members of the interventional
IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 1 of 42
 IP 1115 [IPGXXX]

procedures advisory committee (IPAC) make recommendations about the safety

and efficacy of an interventional procedure. It is based on a rapid review of the
medical literature and specialist opinion. It should not be regarded as a definitive
assessment of the procedure.

Date prepared
This overview was prepared in August 2018.

Procedure name
 Therapeutic hypothermia for acute ischaemic stroke.

Specialist societies
 Royal College of Physicians
 British Association of Stroke Physicians (BASP)
 Faculty of Intensive Care Medicine
 Association of British Neurologists.

Description of the procedure

Indications and current treatment

Acute ischaemic stroke is characterised by the sudden loss of blood circulation to
an area of the brain and a corresponding loss of neurological function. This may
lead to symptoms such as numbness or weakness of the face, arm or leg on
1 side of the body, and often problems with speech and swallowing. Stroke
severity can be measured using scales such as the National Institutes of Health
Stroke Scale (NIHSS) and the Modified Rankin Scale (mRS). Broadly, strokes
are classified as either haemorrhagic or ischaemic. Acute ischaemic stroke refers
to stroke caused by an arterial thrombosis or embolism. It is more common than
haemorrhagic stroke.

Patients suspected to be having an acute ischaemic stroke should have rapid

assessment and early intervention with specialist care according to NICE’s
guideline on stroke and transient ischaemic attack in over 16s: diagnosis and
initial management. Recanalisation strategies, such as thrombolysis, attempt to
re-establish blood flow so that cells starved of oxygen can be rescued before
they are irreversibly damaged. Effective stroke care also includes specialised
supportive care and rehabilitation when appropriate.

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 2 of 42
 IP 1115 [IPGXXX]

What the procedure involves

Evidence suggests that reducing central body temperature decreases cerebral
metabolism, neurotransmitter production, inflammation, free radical generation
and oedema, and disrupts cellular apoptotic pathways. These factors are all
believed to aggravate neurological outcome after cerebral hypoxia or ischaemia.
The optimum timing and duration of therapeutic hypothermia has not yet been
determined. However, in active trials, cooling is attempted as close to stroke
onset as possible (usually within 6 hours). Cooling may be continued for at least
12 to 24 hours, and body temperature maintained at 33°C to 35°C. The ideal
target temperature for neuroprotection is unknown.

Before the procedure, the patient’s temperature is measured, usually with an

infrared tympanic thermometer. Further temperature monitoring is done
continuously with an internal (intravesical, rectal or oesophageal) probe
connected to the cooling device. Cooling devices can be classified into surface
(ice-cold saline, surface cooling, cooling helmets and nasal cooling) and
endovascular methods. After cooling, the body is slowly rewarmed, at a rate of
0.3°C to 0.5ºC every hour. Rewarming takes about 8 hours. During cooling,
patients need close cardiovascular monitoring in an intensive care environment,
and may also need intubation and sedation. Drugs such as buspirone or
meperidine may be used to manage shivering. The procedure may be used with
thrombolysis (intravenous alteplase), mechanical thrombectomy or other vascular
reperfusion techniques. The aim of the procedure is to limit the damage to the
brain cells.

Outcome measures

The NIHSS
This is used to measure the severity of a stroke. It scores areas such as level of
consciousness, vision, sensation, movement, speech and language. A maximum
of 42 points represents the most severe symptoms.

The levels of stroke severity on the NIHSS are categorised as:

 0: no stroke
 1–4: minor stroke
 5–15: moderate stroke
 16–20: moderate/severe stroke
 21–42: severe stroke.

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 3 of 42
 IP 1115 [IPGXXX]

mRS
This is a functional assessment scale that measures the degree of disability or
dependence of people who have suffered a stroke.

The scale runs from perfect health without symptoms to death:

 0: no symptoms
 1: no significant disability; able to carry out all usual activities, despite some
symptoms
 2: slight disability; able to look after own affairs without assistance, but unable
to carry out all previous activities
 3: moderate disability. Needs some help, but able to walk unassisted
 4: moderately severe disability; unable to attend to own bodily needs without
assistance, and unable to walk unassisted
 5: severe disability; needs constant nursing care and attention, bedridden,
incontinent
 6: dead.

Efficacy summary
Neurological outcome

In a systematic review of 252 patients (6 randomised controlled trials [RCTs]), the

risk ratio for a favourable outcome (defined as an mRS score of 0 or 1) after
therapeutic hypothermia (endovascular or surface-based cooling) was 0.85 (95%
confidence interval [CI] 0.56 to 1.29; p=0.46, I2=0%, 5 studies). The risk ratio for
a poor outcome (defined as an mRS score of 3 to 6) was 1.20 (95% CI 0.88 to
1.64; p=0.24, I2=0%; 3 studies).1

In an RCT of 120 patients who had endovascular therapeutic hypothermia or

standard care, the mRS was 0 in 33% of patients and 1 in 38% of patients at
90-day follow-up (odds ratio 0.81, 95% CI 0.36 to 1.85).The NIHSS score at
7 days was similar in the 2 groups (10.4±10.3 compared with 10.6±11.3).2

In an RCT of 33 patients who had endovascular therapeutic hypothermia or

standard care, at 6 months, a good neurological outcome (an mRS score of 0 to
3) was reported in 88% (7/8) and 40% (4/10) of surviving patients respectively
(p=0.066).3

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 4 of 42
 IP 1115 [IPGXXX]

In a non-randomised comparative study of 75 patients, 37 patients who had

therapeutic hypothermia with endovascular cooling and 2 who had surface
cooling were compared with 36 patients who had standard care (controls). At
3 month follow-up, there was a statistically significantly higher proportion of
patients with an mRS score of 0 to 1 in the hypothermia group compared with the
control group (31% [12/39] versus 8% [3/36], p=0.015) or an mRS score of 0 to 2
(49% [19/39] versus 22% [8/36], p=0.017). The proportion of patients with an
mRS score of 0 to 3 at 3-month follow-up was higher in the hypothermia group
but the difference was not statistically significant (56% [22/39] versus 39%
[14/36], p=0.129).4

In a non-randomised comparative study of 111 patients who had

hemicraniectomy after suspected middle cranial artery infarction (with or without
surface or endovascular therapeutic hypothermia), 25% (13/53) of patients who
had therapeutic hypothermia and 41% (24/58) of controls had an mRS score of 0
to 3 at 12 months (risk ratio 0.59, 95% CI 0.34 to 1.04).7

In a case series of 50 patients who had therapeutic hypothermia with surface

cooling, the mean NIHSS score was 29 at 4 weeks and the mRS score was 2.9
(range 2 to 5) at 3 month follow-up.5

In a case series of 19 patients who had an extended period of therapeutic

hypothermia (surface cooling), the mean NIHSS score was 8.3±2.7 at 12 month
follow-up and the mean mRS was 3.2±0.9.6

Mortality

In the systematic review of 252 patients, the risk ratio for mortality from all causes
after therapeutic hypothermia was 1.12 (95% CI 0.62 to 2.05; p=0.70, I2=0%,
6 studies). Mortality in the hypothermia group in the individual studies was 28%,
21%, 12%, 0%, 8%, and 13% (pretreatment mean NIHSS score was 15.2, 14.3,
8.0 [median], 12.0, 18.0 and 11.4 respectively). None of the deaths were
considered to be related to hypothermia.1

In the RCT of 120 patients, mortality was 16% with endovascular therapeutic
hypothermia and 9% with standard care (odds ratio 1.95, 95% CI 0.56 to 7.79).2

In the RCT of 33 patients, mortality was 50% (8/16) with endovascular

therapeutic hypothermia and 41% (7/17) with standard care at 6 months
(p=0.732).3

In the non-randomised comparative study of 75 patients, mortality at 1 month

was similar in the 2 groups (15% [6/39] with endovascular or surface cooling
compared with 14% [5/36] with standard care, p=0.855).4

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 5 of 42
 IP 1115 [IPGXXX]

In the non-randomised comparative study of 111 patients, survival at 12 months

was 49% (26/53) in the hypothermia group (surface or endovascular cooling) and
79% (46/58) in the control group (risk ratio 0.62, 95% CI 0.46 to 0.84).7

In a non-randomised comparative study of 47 patients who had hemicraniectomy

following suspected middle cranial artery infarction with or without hypothermia
(surface cooling), in-hospital mortality was 15% in the hypothermia group and
41% in the control group (p=0.056). Hypothermia was identified as the strongest
factor that affected survival in the multivariable analysis (odds ratio 6.21, 95% CI
1.04 to 37.05; p=0.045). Mortality at 1-year follow-up was 28% (5/18) in the
hypothermia group and 52% (13/25) in the control group (p=0.112).8

Cerebral oedema

In the non-randomised comparative study of 75 patients, 54% (21/39) of patients

who had therapeutic hypothermia had no cerebral oedema compared with 17%
(6/36) of patients in the control group (p=0.001). Absence of cerebral oedema
was an independent predictor of a good outcome (odds ratio 5.4, 95% CI 1.6 to
18.2).4

Stroke recurrence

Stroke recurrence was reported in 26% (5/19) of patients in the case series of 19
patients who had surface cooling (4 happened during the cooling period and 1 at
5 days after rewarming); 3 patients had severe clinical deterioration and died,
and 2 survived (1 of whom had a hemicraniectomy).6

Safety summary
Pneumonia

Pneumonia was reported at a statistically significantly higher rate in patients who

had therapeutic hypothermia (endovascular, intravenous or surface-based) in the
systematic review of 252 patients (risk ratio 3.30, 95% CI 1.48 to 7.34; p=0.003,
I2=0%).1

It was reported in 19% of patients who had endovascular therapeutic

hypothermia and 11% of patients in the control group in the RCT of 120 patients
(odds ratio 1.99, 95% CI 0.63 to 6.98).2

It was reported in a similar proportion of patients in each group in the RCT of 33

patients (44% [7/16] endovascular hypothermia and 47% [8/17] standard care,
p=1.00).3

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 6 of 42
 IP 1115 [IPGXXX]

It was reported at a statistically significantly lower rate in patients who had

therapeutic hypothermia in the non-randomised comparative study of 75 patients
(5% [2/39] compared with 31% [11/36], p=0.004) of patients in the control group.4

It was reported in 48% (24/50) and 42% (8/19) of patients respectively in the
case series of 50 and 19 patients who had therapeutic hypothermia with surface
cooling.5,6

Pneumonia was the most common adverse event in patients in intensive care in
the non-randomised comparative study of 47 patients, but there was no
statistically significant difference between the groups (25% [5/20] in the
hypothermia group compared with 19% [5/27] in the control group, p=0.723).8

Intracerebral haemorrhage

In the non-randomised comparative study of 75 patients, 39% (15/39) of patients

who had therapeutic hypothermia had no haemorrhagic transformation compared
with 14% (5/36) of patients in the control group (p=0.016).4

In the non-randomised comparative study of 47 patients, haemorrhagic

transformation after hemicraniectomy was reported in 35% (7/20) of patients who
had therapeutic hypothermia compared with 26% (7/27) of patients in the control
group (p=0.501).8

Symptomatic intracerebral haemorrhage was reported in 2% of patients who had

endovascular hypothermia and 4% of patients in the control group in the RCT of
120 patients (odds ratio 0.45, 95% CI 0.01 to 8.80).2

Cardiac arrhythmia or bradycardia

Bradycardia was reported in 44% (7/16) of patients who had endovascular

hypothermia and 12% (2/17) of patients in the control group (p=0.057) in the RCT
of 33 patients.3

Bradycardia was reported in 8% (3/39) of patients who had therapeutic

hypothermia and 3% (1/36) of patients in the control group in the non-randomised
comparative study of 75 patients (p value not reported).4

Cardiac arrhythmia or bradycardia was reported in 62% (31/50) of patients in the

case series of 50 patients who had surface cooling.5

All patients had bradycardia during the hypothermia procedure, without

arrhythmia, in the case series of 19 patients who had surface cooling.6

Cardiac arrhythmia was reported in 28% (15/53) of patients who had therapeutic
hypothermia with surface or endovascular cooling and 10% (2/20) of patients

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 7 of 42
 IP 1115 [IPGXXX]

who had surface cooling in the non-randomised comparative studies of 111 and
47 patients respectively.7,8

Electrolyte disorder

Electrolyte disorder was reported in 100% (16/16) of patients who had

endovascular hypothermia and 53% (9/17) of patients in the control group
(p=0.003) in the RCT of 33 patients.3

Hypoalbuminaemia

Hypoalbuminaemia was reported in 69% (11/16) of patients who had

endovascular hypothermia and 41% (7/17) of patients in the control group
(p=0.166) in the RCT of 33 patients.3

Stress hyperglycaemia

Stress hyperglycaemia was reported in 38% (6/16) of patients who had

endovascular hypothermia and 6% (1/17) of patients in the control group
(p=0.039) in the RCT of 33 patients.3

Hypotension

Severe hypotension (mean arterial pressure less than 50 mmHg) was reported in
4% (2/50) of patients in the case series of 50 patients who had therapeutic
hypothermia with surface cooling.5

Hypotension was reported in 38% (6/16) of patients who had endovascular

hypothermia and 18% (3/17) of patients in the control group (p=0.259) in the RCT
of 33 patients.3

All patients had hypotension, which needed continuous infusion of

norepinephrine, during the hypothermia procedure (surface cooling) in the case
series of 19 patients.6

Coagulation dysfunction

Severe coagulopathy was reported in 4% (2/50) of patients in the case series of

50 patients who had therapeutic hypothermia with surface cooling. This was fatal
in 1 patient.5

Coagulation dysfunction was reported in 88% (14/16) of patients who had

endovascular hypothermia and 88% (15/17) of patients in the control group
(p=1.00) in the RCT of 33 patients.3

Sepsis

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 8 of 42
 IP 1115 [IPGXXX]

Severe sepsis was reported in 11% (6/53) of patients who had therapeutic
hypothermia (surface or endovascular cooling) in the non-randomised
comparative study of 111 patients. It was the cause of death in 1 patient.7

Sepsis was reported in 1 patient who had therapeutic hypothermia (surface

cooling) in the non-randomised comparative study of 47 patients. The
hypothermia treatment was stopped.8

Gastrointestinal bleeding

Gastrointestinal bleeding was reported in 50% (8/16) of patients who had

endovascular hypothermia and 12% (2/17) of patients in the control group
(p=0.026) in the RCT of 33 patients.3

Gastric retention

Gastric retention was reported in 94% (15/16) of patients who had endovascular
hypothermia and 41% (7/17) of patients in the control group (p=0.002) in the RCT
of 33 patients.3

Deep venous thrombosis

Deep venous thrombosis was reported in 9% (5/53) of patients who had

therapeutic hypothermia (surface or endovascular cooling) in the non-randomised
comparative study of 111 patients.7

Lower extremity deep venous thrombosis was reported in 25% (4/16) of patients
who had endovascular hypothermia and 12% (2/17) of patients in the control
group (p=0.398) in the RCT of 33 patients.3

Anecdotal and theoretical adverse events

In addition to safety outcomes reported in the literature, specialist advisers are
asked about anecdotal adverse events (events which they have heard about) and
about theoretical adverse events (events which they think might possibly occur,
even if they have never happened). For this procedure, specialist advisers did not
list any additional anecdotal or theoretical adverse events.

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 9 of 42
 IP 1115 [IPGXXX]

The evidence assessed

Rapid review of literature

The medical literature was searched to identify studies and reviews relevant to
therapeutic hypothermia for acute ischaemic stroke. The following databases
were searched, covering the period from their start to 25 May 2018: MEDLINE,
PREMEDLINE, EMBASE, Cochrane Library and other databases. Trial registries
and the Internet were also searched. No language restriction was applied to the
searches (see the literature search strategy). Relevant published studies
identified during consultation or resolution that are published after this date may
also be considered for inclusion.

The following selection criteria (table 1) were applied to the abstracts identified by
the literature search. Where selection criteria could not be determined from the
abstracts the full paper was retrieved.

Table 1 Inclusion criteria for identification of relevant studies

Characteristic Criteria
Publication type Clinical studies were included. Emphasis was placed on
identifying good quality studies.
Abstracts were excluded if no clinical outcomes were reported,
or if the paper was a review, editorial, or a laboratory or animal
study.
Conference abstracts were also excluded because of the
difficulty of appraising study methodology, unless they reported
specific adverse events that were not available in the published
literature.
Patient Patients with ischaemic stroke
Intervention/test Therapeutic hypothermia
Outcome Articles were retrieved if the abstract contained information
relevant to the safety and/or efficacy.
Language Non-English-language articles were excluded unless they were
thought to add substantively to the English-language evidence
base.

List of studies included in the overview

This overview is based on about 800 patients from 1 systematic review, 3 RCTs
(1 of which was a conference abstract and included for safety data only), 3 non-
randomised comparative studies and 2 case series.1–9

Other studies that were considered to be relevant to the procedure but were not
included in the main extraction table (table 2) are listed in the appendix.
IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 10 of 42
 IP 1115 [IPGXXX]

Table 2 Summary of key efficacy and safety findings on therapeutic hypothermia

for acute ischaemic stroke

Study 1 Wan Y-H (2014)

Details
Study type Systematic review and meta-analysis
Country Study countries not reported
Recruitment period Trials that were completed and published by February 2014 were included.
Study population and n=252 (6 randomised controlled trials)
number Patients with acute ischaemic stroke
Age and sex Mean age varied from 49 to 69 years; 50% (126/252) male
Patient selection Patients older than 18 years with acute ischaemic stroke, and cerebral haemorrhage had been excluded
criteria as the cause of the symptoms based on CT or MRI. The intervention of interest was therapeutic
hypothermia including thrombolytic therapy if indicated. Information about the depth, duration, and
rewarming speed had to be available for the study to be included.
Technique Studies used endovascular, intravenous or surface-based cooling. One of the 6 studies used
hemicraniectomy as the control, the other studies used standard treatment, including thrombolytic therapy.
Follow-up Range 1 to 6 months
Conflict of Not reported
interest/source of
funding

Analysis
Follow-up issues: Complete outcome data were reported in 4 trials. It was unclear whether data were complete in the
other 2 trials.

Study design issues: Systematic review and meta-analysis complying with the Preferred Reporting Items for Systematic
Reviews and Meta-Analysis guidelines. The Cochrane risk of bias tool was used to assess the risk of bias in the included
studies and the Grading of Recommendations, Assessment, Development, and Evaluation system was used to grade the
quality of evidence for the outcomes. A sensitivity analysis was done by sequentially removing each study.

Adequate methods of random sequence generation were used in 4 trials, 1 was inadequate and 1 did not report the
methods of random sequence generation. Four trials used allocation concealment, whereas the description was unclear in
the other 2 trials. Patients and staff were blinded in only 1 trial. The assessors of outcomes were blinded in 2 trials, but the
other 4 trials did not provide information on the blinding of outcome assessment.

The evidence quality for each outcome was low or very low.

Characteristics of included studies (6 randomised controlled trials)

Author Patients Hypothermia treatment Control treatment Follow-up

De Georgia et al., n=40 (18 versus 22) Endovascular cooling: Standard medical 1 month
2004 Patients with acute mean time from stroke treatment,
circulation ischaemic onset to cooling 9.0 including
stroke ≤12 hours of hours; depth 33°C, thrombolytic
symptom onset duration 24 hours, therapy if
rewarming speed indicated
0.2°C/hour

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 11 of 42
 IP 1115 [IPGXXX]

Oesophageal
temperature monitoring
Els et al., 2006 n=25 (12 versus 13) 10 intravenous cooling, 2 Hemicraniectomy 6 months
Patients with severe external cooling
supratentorial Depth 35°C, duration
ischaemic stroke 48 hours, rewarming
speed 0.04°C/hour
Continuous oesophageal
temperature monitoring
Hemmen et al., 2010 n=58 (28 versus 30) Intravenous cooling Standard 3 months
Patients with acute Time from stroke onset treatment,
ischaemic stroke to cooling start (median): including
≤6 hours of symptom 5.9 hours; depth 33°C, thrombolytic
onset duration 24 hours, therapy.
rewarming speed
0.3°C/hour
Krieger, 2013 n=31 (17 versus 14) 7 endovascular cooling, Standard 3 months
Patients with acute 10 surface-based cooling treatment,
ischaemic stroke Mean time from stroke including
≤24 hours of onset to cooling start thrombolytic
symptom onset 12.2 hours; depth 33°C, therapy if
duration 24 hours, indicated.
rewarming speed
0.25°C/hour to
0.5°C/hour
Temperatures measured
by bladder thermistor
probe.
Piironen et al., 2014 n=36 (18 versus 18) Surface-based cooling Treatment 3 months
Patients with acute and saline infusions according to in-
ischaemic stroke Time from stroke onset house guidelines
after intravenous to cooling start (median)
thrombolysis 6 hours; depth 34.5°C to
35.5°C; duration 10.5
hours, rewarming speed
0.2°C/hour to 0.5°C/hour
Continuous bladder
temperature monitoring
Tong, 2011 n=62 (31 versus 31) Surface-based cooling Thrombolytic 3 months
Patients with acute Mean time from stroke therapy
ischaemic stroke onset to cooling start 3.9
≤6 hours of symptom hours; depth 32°C to
onset 34°C; duration 24 hours,
Continuous rectal
temperature monitoring

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 12 of 42
 IP 1115 [IPGXXX]
Key efficacy and safety findings
Efficacy Safety
Number of patients analysed: 252 Complications
 Pneumonia: RR=3.30, 95% CI 1.48 to 7.34, p=0.003, I2=0%
Favourable neurological outcome (modified Rankin Scale  Symptomatic intracranial haemorrhage: RR=1.07, 95% CI
score 0 to 1) 0.37 to 3.04, p=0.90, I2=0%
5 studies reported data on favourable neurological outcome.  Fatal intracranial haemorrhage: RR=1.40, 95% CI 0.33 to
RR=0.85 (95% CI 0.56 to 1.29, p=0.46, I2=0%) 5.99, p=0.65, I2=0%
 Deep vein thrombosis: RR=2.25, 95% CI 0.68 to 7.44,
None of the subanalyses (32°C to 33.9°C versus 34°C to p=0.18, I2=12%
35.9°C, duration ≤24 hours, rewarming speed ≤12 hours versus  Atrial fibrillation: RR=1.14, 95% CI 0.40 to 3.25, p=0.80,
>12 hours) were statistically significant. I2=0%

Poor neurological outcome (modified Rankin Scale score 3

to 6)
3 studies reported data on poor neurological outcome.
RR=1.20 (95% CI 0.88 to 1.64, p=0.24, I2=0%)

None of the subanalyses (32°C to 33.9°C versus 34°C to

35.9°C, duration ≤24 hours, rewarming speed ≤12 hours versus
>12 hours) were statistically significant.

Mortality
6 studies reported mortality from all causes.
RR=1.12 (95% CI 0.62 to 2.05, p=0.70, I2=0%)

None of the subanalyses (32°C to 33.9°C versus 34°C to

35.9°C, duration ≤24 hours, rewarming speed ≤12 hours versus
>12 hours) were statistically significant.

Mortality in the hypothermia group in the individual studies were

27.8%, 21.4%, 11.8%, 0%, 8.3%, and 12.9% (pretreatment
mean National Institutes of Health Stroke Scale score was 15.2,
14.3, 8.0 (median), 12.0, 18.0, and 11.4 respectively).
None of the deaths were considered to be related to
hypothermia.
Abbreviations used: CI, confidence interval; RR, risk ratio

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 13 of 42
 IP 1115 [IPGXXX]

Study 2 Lyden P (2016)

Details
Study type Randomised controlled trial (ICTus 2)
Country US
Recruitment period Recruitment was suspended in December 2014 and final follow-up visits were completed by May 2015
Study population and n=120 (63 hypothermia, 57 normothermia)
number Patients with acute ischaemic stroke.
Age and sex Hypothermia: mean age 66 years; 54% male
Normothermia: mean age 68 years; 61% male
Patient selection Patients with ischaemic stroke had treatment within 3 hours of symptom onset with intravenous
criteria recombinant tissue-type plasminogen activator (r-tPA); NIHSS ≥7 and ≤20 (left brain stroke) or ≤24 (right
brain stroke); age 22 to 82 years. Intra-arterial recanalisation procedures were disallowed.
Technique Patients randomised to hypothermia immediately had 2 litres of normal saline (4°C) as fast as possible;
patients in the normothermia group had the same volume of room temperature saline.
Hypothermia technique: target temperature 33°C, intravascular cooling device (Celsius Catheter,
Innercool, Carlsbad) for 24 hours, followed by controlled rewarming for 12 hours. Anti-shivering measures
(meperidine, buspirone, and skin warming) were used for all 36 hours. If shivering could not be controlled
without respiratory compromise, the target temperature was raised in 0.5°C increments until shivering
stopped (permissive hypothermia).
Time from arrival to r-tPA treatment=60 minutes; mean time from symptom onset to
cooling=287.6±65.8 minutes.
Follow-up 90 days
Conflict of Study was part of an NINDS (National Institute of Neurological Disorders and Stroke) funded SPOTRIAS
interest/source of (Specialized Program of Translational Research in Acute Stroke), a collaboration among the NINDS,
funding University of California and University of Texas.
The first author serves on the data safety monitoring board for an industry sponsored trial of therapeutic
hypothermia for cardiac arrest.

Analysis
Follow-up issues: One patient in each group was lost to follow-up; 90-day outcome data were missing for 2 patients in
the normothermia group.

Study design issues: Prospective, randomised, single-blind, multicentre study. The study was planned to have 2 stages:
a phase 2 study to assess safety and feasibility of various protocol changes and a phase 3 efficacy study. Phase 2 was
intended to have 400 patients who would have been included in the total 1600 patient phase 3 study. During the
application to renew funding for the trial, several publications established the efficacy of intra-arterial neurothrombectomy
for selected patients. The Steering Committee decided to stop the trial, examine the data and redesign the trial to include
neurothrombectomy. No statistically significant primary results were expected. A per-protocol and intention-to-treat
analysis was done. The primary endpoint was favourable outcome, defined as a 90-day mRS score of 0 or 1.

Study population issues: The baseline demographics were similar between the 2 groups.

Other issues: Of the 120 enrolled patients, 16 (13%) showed a rapid reversal of their deficit before the study procedures
were started, making them ‘early responders’ (9 hypothermia and 7 normothermia). These patients were included in the
intention-to-treat population but not the per-protocol analysis.

The authors noted that the efficacy results in this small study reflect the effect of mild (<35°C) hypothermia.

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 14 of 42
 IP 1115 [IPGXXX]
Key efficacy and safety findings
Efficacy Safety
Number of patients analysed: 120 (63 versus 57) Serious adverse events (ITT population)
 Hypothermia=41%
mRS score of 0 or 1 at 90 days (primary outcome)  Normothermia=35% (OR 1.30, 95% CI 0.58 to 2.92)
ITT population
 Hypothermia=33% Mortality
 Normothermia=38%, OR 0.81, 95% CI 0.36 to 1.85  Hypothermia=15.9%
PP population  Normothermia=8.8% (OR 1.95, 95% CI 0.56 to 7.79)
 Hypothermia=24%
 Normothermia=38%, OR 0.50, 95% CI 0.19 to 1.30 Pneumonia
 Hypothermia=19%
Using severity-adjusted outcomes based on initial NIHSS, the  Normothermia=10.5% (OR 1.99, 95% CI 0.63 to 6.98)
ITT-adjusted OR for a good outcome was 1.37 (0.60 to 3.19).
The PP-adjusted OR was 0.89 (0.34 to 2.30).
Fluid overload
 Hypothermia=4.8% (3/63)
ITT OR for mRS score of 2 to 6 in the hypothermia group was
1.03 (0.55 to 1.95). The PP OR was 1.39 (0.69 to 2.80).  Normothermia=12.3% (7/57)

Multivariable analysis (using age, baseline NIHSS, diabetes Symptomatic intracerebral haemorrhage
mellitus history, admission glucose, and time from stroke onset  Hypothermia=1.6%
to start of r-tPA)
 Normothermia=3.5% (OR 0.45, 95% CI 0.01 to 8.80)
OR for a good outcome in the hypothermia group
 ITT population=0.62 (95% CI 0.26 to 1.51)
 PP population=0.29 (95% CI 0.10 to 0.85), p<0.05 There were no adverse events related to the intravascular
cooling catheter, such as haematoma or arterial injury.
Time to reach target temperature did not seem to change
outcome. In the 33 patients with core body temperature <35°C
within 6 hours of cooling onset, the proportion with 90-day mRS
score of 0 or 1 was 24% (OR 0.52, 95% CI 0.16 to 1.52).

NIHSS score at 7 days

 Hypothermia=10.4±10.3 (n=58)
 Normothermia=10.6±11.3 (n=55)

There was no statistically significant difference between the

groups with regard to the Barthel index at 90 days and the mRS
score at 60 days.

A >7-point decrease in the NIHSS at 7 days after stroke

significantly correlated with 90-day mRS score.
Abbreviations used: CI, confidence interval; ITT, intention to treat; mRS, modified Rankin Scale; NIHSS, National Institutes of Health
Stroke Scale; OR, odds ratio; PP, per-protocol; r-tPA, recombinant tissue-type plasminogen activator.

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 15 of 42
 IP 1115 [IPGXXX]

Study 3 Su Y (2016)
Details
Study type Randomised controlled trial
Country China
Recruitment period 2010 to 2013
Study population and n=33 (16 hypothermia, 17 standard medical treatment)
number Patients with massive cerebral hemispheric infarction
Age and sex Hypothermia: mean age 60 years (range 46 to 75); 94% male
Control: mean age 69 years (range 53 to 79); 59% male
Patient selection Patients aged 18 to 80 years; acute unilateral ischaemic stroke <48 hours ago; infarction involving at least
criteria two thirds of the middle cerebral artery territory on cranial CT or MRI; a National Institute of Health Stroke
Scale (NIHSS) score ≥15 if the non-dominant hemisphere was affected or ≥20 if the dominant hemisphere
was affected; a reduced level of consciousness indicated by an NIHSS item 1a score ≥1; unable to have
decompressive craniectomy because of the premorbid use of antiplatelet or anticoagulant drugs or
because the patient declined the procedure. Patients were excluded if they met any of the following
criteria: premorbid modified Rankin Scale score >2; secondary haemorrhage involving more than a third of
the infarction territory with a space-occupying effect; Glasgow Coma Scale without a verbal response item
score <6; rapidly improving symptoms; both pupils fixed and dilated; simultaneous other brain lesion,
including tumours and contralateral or infratentorial infarctions; platelet count <75,000/mm 3, severe
coagulopathy or cardiac, liver or kidney disease; vasospastic disease, haematological disease with
increased risk of thrombosis, or paramyotonia congenita; sepsis; premorbid treatment with a monoamine
oxidase inhibitor or an allergy to pethidine; inferior vena cava fistula or a filter in its place, a mass near the
inferior vena cava, or a height of <1.5 m; pregnancy; life expectancy <6 months.
Technique The temperature of the patients in the control group was sustained between 36.5°C and 37.5°C to
maintain normothermia.
Patients randomised to the hypothermia group had endovascular hypothermia treatment using a mobile
temperature management device (CoolGard, Alsius Corporation). A Foley temperature catheter was used
to monitor the bladder temperature and was also connected to the temperature management system. The
device adjusted the saline temperature according to the patient’s temperature and the target temperature.
Hypothermia was started as soon as possible after admission. The target bladder temperature was 33°C
or 34°C. Hypothermia was maintained for a minimum of 24 hours and could be prolonged to 72 hours if
necessary. Shivering was suppressed with gloves, socks and a quilt. Patients also had oral buspirone,
intravenous (IV) pethidine, IV midazolam, and an IV muscle relaxant. The IV infusion drugs started from
the minimum doses and were adjusted according to the severity of shivering. Intracranial pressure was
invasively and continuously monitored with a transducer-tipped pressure-temperature monitoring catheter.
Follow-up 6 months
Conflict of Study was supported by a grant from the National Key Department of Neurology funded by the National
interest/source of Health and Family Planning Commission of the People’s Republic of China and the National Key
funding Department of Critical Care Medicine funded by the National Health and Family Planning Commission of
the People’s Republic of China.

Analysis
Follow-up issues: There were no losses to follow-up.

Study design issues: Prospective, single-centre randomised controlled trial. Randomisation was done by using sealed
envelopes, with odd numbers being assigned to the hypothermia group and even numbers to the control group. The
envelopes were opened by an investigator who was not involved in patient screening, treatment, data collection or
analysis. The primary outcomes were mortality and the modified Rankin Scale score at 6 months. The sample size was
calculated to be 168, which was not achieved because of slow recruitment.

Study population issues: Patients in the hypothermia group were younger than those in the control group (p=0.006) and
there was a higher proportion of males in the hypothermia group (p=0.039). An additional 2 patients were randomised to

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 16 of 42
 IP 1115 [IPGXXX]

the hypothermia group but the catheter could not be inserted because of severe bleeding at the puncture site or lumen
malfunction. These patients were excluded from the analysis.

Key efficacy and safety findings

Efficacy Safety
Number of patients analysed: 33 (16 versus 17) Complications
Good neurological outcome (modified Rankin Scale score of Of the 33 patients, only 1 patient in the control group did not
0 to 3) in surviving patients at 6 months have a complication.
 Hypothermia=87.5% (7/8)
 Control=40.0% (4/10), p=0.066 Complications Hypothermia Control p value
OR=10.5, 95% CI 0.9 to 121.4 group group
Adjusted OR=4.8, 95% CI 0.3 to 71.1 n=16 n=17
Recurrent 0 2 0.485
infarction
Mortality at 6 months
Haemorrhagic 1 2 1.00
 Hypothermia=50.0% (8/16) transformation
 Control=41.2% (7/17), p=0.732 Bradycardia 7 2 0.057
7 patients in the hypothermia group and 6 patients in the control Tachycardia 9 10 1.00
group died of herniation.
Arrhythmia 1 3 0.601
Hypotension 6 3 0.259
Intracranial pressure
10 patients in the hypothermia group had intracranial pressure Pneumonia 7 8 1.00
monitoring; 5 developed transtentorial herniation. In these Lower extremity 4 2 0.398
5 patients, the intracranial pressure was 18.7±5.5 mmHg at the deep venous
start of cooling and decreased to 12.4±6.0 mmHg when the thrombosis
target temperature was reached. During the maintenance period, Electrolyte disorder 16 9 0.003
it increased, reaching 18.8±7.4 mmHg at the start of rewarming.
The pressure increased significantly during the rewarming Coagulation 14 15 1.00
period, reaching 39.4±9.6 mmHg at the end of rewarming. All dysfunction
5 patients died from herniation. Gastrointestinal 8 2 0.026
bleeding
Gastric retention 15 7 0.002
Stress 6 1 0.039
hyperglycaemia
Hypoalbuminaemia 11 7 0.166
Acute liver injury 2 1 0.601
Acute kidney injury 2 0 0.227
Total 109 74 0.001
Abbreviations used: CI, confidence interval; OR, odds ratio

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 17 of 42
 IP 1115 [IPGXXX]

Study 4 Hong JM (2014)

Details
Study type Non-randomised comparative study
Country South Korea (2 tertiary referral stroke centres)
Recruitment period 2010 to 2012
Study population and n=75 (39 hypothermia, 36 control)
number Patients with acute ischaemic stroke in the anterior circulation, who had successful recanalisation
Age and sex Hypothermia: mean age 65 years; 59% (23/39) male
Control: mean age 68 years; 50% (18/36) male
Patient selection Inclusion criteria: ischaemic stroke involving the anterior circulation (National Institute of Health Stroke
criteria Scale [NIHSS] ≥10); acute infarction with diffusion-weighted imaging confirmation; endovascular
recanalisation (thrombolysis in cerebral ischaemia, grade ≥2b) within 6 hours after symptom onset, or
spontaneous recanalisation.
Technique Hypothermia: patients were cooled with either an endovascular cooling catheter (Alsius) placed in the
inferior vena cava (n=37) or a surface cooling device (Arctic Sun, n=2). All cooled patients were
mechanically ventilated, midazolam was used for sedation and vecuronium for neuromuscular blockade.
The target temperature was 34.5°C. Core body temperatures were plotted with an oesophageal
temperature probe. Hypothermia therapy was maintained for 48 hours, and rewarming was done under
sedation at a rate of 0.5°C per every 12 hours. Antipyretic and anti-shivering agents other than paralytics
were not used during hypothermia.
All patients had treatment according to international guidelines for stroke care. Surgical hemicraniectomy
was done when necessary.
Follow-up 90 days
Conflict of None
interest/source of
funding

Analysis
Follow-up issues: No losses to follow-up were described.

Study design issues: Prospective cohort study at 2 tertiary referral stroke centres. All patients in 1 centre had mild
hypothermia treatment and patients in the other centre had standard treatment without hypothermia. Neurological deficit
was assessed by serial NIHSS scores by certificated stroke nurses or stroke neurologists at least every 4 hours.
Neurological scales were checked daily until discharge and every 3 months thereafter. The clinical outcome was primarily
dichotomised into good (0 to 2 points) and poor (3 to 6 points) groups using a mRS score at 90 days.

Study population issues: Baseline characteristics, stroke risk factors, and initial laboratory findings were similar between
the 2 groups.

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 18 of 42
 IP 1115 [IPGXXX]
Key efficacy and safety findings
Efficacy Safety
Number of patients analysed: 75 (39 vs 36) Medical complications

Radiological and clinical outcomes Hypothermia Control p value

Hypothermia Control p value n=39 n=36
n=39 n=36 Patients with 1 or 11 (28.2) 17 (47.2) 0.089
HT more medical
complication
None 15 (38.5) 5 (13.9) 0.016
Bradycardia 3 (7.7) 1 (2.8) -
HT type 1 8 (20.5) 9 (25.0)
Elevated creatine 2 (5.1) 2 (5.6) -
HT type 2 1 (2.6) 7 (19.4) kinase
PH type 1 8 (20.5) 8 (22.2) Cardiac events (T- 1 (2.6) 0 (0) -
PH type 2 7 (17.9) 7 (19.4) wave inversion,
Cerebral oedema non-STEMI)
None 21 (53.8) 6 (16.7) 0.001 Hypokalaemia 2 (5.1) 0 (0) -
Mild 9 (23.1) 14 (38.9) Pulmonary oedema 2 (5.1) 1 (2.8) -
Midline shift 9 (23.1) 16 (44.4) Decreased blood 1 (2.6) 0 (0) -
pressure
Clinical outcomes
Pneumonia 2 (5.1) 11 (30.6) 0.004*
mRS 0 to 1 at 3 months 12 (30.8) 3 (8.3) 0.015
Urinary tract 0 (0) 2 (5.6) -
mRS 0 to 2 at 3 months 19 (48.7) 8 (22.2) 0.017 infection
mRS 0 to 3 at 3 months 22 (56.4) 14 (38.9) 0.129 Deep vein 0 (0) 0 (0) -
Mortality at 1 month 6 (15.4) 5 (13.9) 0.855 thrombosis
Hemicraniectomy 4 (10.3) 5 (13.9) 0.629 Gastrointestinal 0 (0) 1 (2.8) -
HT type 1 was defined as small petechiae along the margins of bleeding
the infarct; type 2 was defined as confluent petechiae within the * this p value is given in the body of the text, which states the
infarcted area but without a space-occupying effect. rate of pneumonia was 8% in the hypothermia group rather than
PH type 1 was defined as haematoma in <30% of the infarcted 5% that was quoted in the table.
area with some space-occupying effect; type 2 was defined as
haematoma in >30% of the infarcted area with a substantial
space-occupying effect.

After adjustment for potential confounders, therapeutic

hypothermia (OR 3.0, 95% CI 1.0 to 8.9, p=0.047) and distal
occlusion (OR 7.3, 95% CI 1.3 to 40.3, p=0.022) were
independent predictors for a good outcome. Absence of cerebral
oedema (OR 5.4, 95% CI 1.6 to 18.2, p=0.006) and no medical
complications (OR 9.3, 95% CI 2.2 to 39.9, p=0.003) were also
independent predictors for good outcome during the treatment.
Abbreviations used: CI, confidence interval; HT, haemorrhagic transformation; mRS, modified Rankin Scale; OR, odds ratio; PH,
parenchymal haematoma; STEMI, ST-segment-elevation myocardial infarction

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 19 of 42
 IP 1115 [IPGXXX]

Study 5 Schwab S (2001)

Details
Study type Case series
Country Germany (4 neurocritical care units)
Recruitment period Not reported
Study population and n=50
number Patients with acute ischaemic cerebral infarction, involving at least the complete territory of the middle
cerebral artery
Age and sex Mean 57 years; 70% (35/50) male
Patient selection Patients with acute ischaemic cerebral infarction, involving at least the complete territory of the middle
criteria cerebral artery.
Patients with history of disabling neurological disease, terminal illness, or severe cardiac failure (New York
Heart Association class III or IV) were excluded.
Technique None of the patients had thrombolytic therapy. The timing of hypothermia treatment was based on
individual decision making; some patients had treatment immediately after initial CT scan and some had
treatment only after clinical or radiological signs of midline shift became evident (mean interval between
onset of symptoms and start of hypothermia was 22±9 hours [range 4 to 75 hours]).
Hypothermia was induced by cooling blankets and alcohol and ice bags. Target temperature: 32°C to
33°C (moderate hypothermia) for 24 to 72 hours. After this period, passive rewarming over 24 hours to
normal temperature was allowed; the pace of rewarming was decided by the local physician. All patients
were sedated with midazolam or propofol; morphine or fentanyl was used for analgesia. All patients had
neuromuscular blockade before the start of moderate hypothermia, continued until rewarming at 36°C was
achieved.
Follow-up 3 months
Conflict of Not reported
interest/source of
funding

Analysis
Follow-up issues: 6% (3/50) of patients were lost to follow-up.

Study design issues: Prospective observational study, consecutive patients. Initial patient status and patient status at 4
weeks were assessed with the NIHSS. Clinical outcome was assessed with the Rankin Scale and the 100-point Barthel
Index at 3 month follow-up. The study aimed to assess the feasibility and safety of the procedure.

Study population issues: The mean NIHSS on admission was 25 (range 15 to 32). All patients had complete middle
cerebral artery stroke; 5 patients had an additional anterior or posterior artery territory infarction.

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 20 of 42
 IP 1115 [IPGXXX]
Key efficacy and safety findings
Efficacy Safety
Number of patients analysed: 50 Complications
Variable Number of
Intracranial pressure patients (%)
Mean ICP before start of hypothermia=19.8±14.2 mmHg (range 4 to 36) Platelet count <100,000/nanolitre 35 (70%)
Cardiac arrhythmia/bradycardia 31 (62%)
When steady state of hypothermia was reached, ICP reduced to Pneumonia 24 (48%)
12.4±5.3 mmHg (p<0.05) Pancreatitis 3 (6%)
Severe hypotension (mean arterial 2 (4%)
A rise in ICP was seen in all patients during rewarming, to mean values pressure <50 mmHg)
of 23.4±8.7 mmHg (range 17 to 71) Severe coagulopathy 2 (4%)
Cardiac failure 1 (2%)
A shorter (<16 hours) rewarming period was associated with a more
NB: the figures above are as reported in the table of the
pronounced rise in ICP (15±10% versus 26±15%, p=not significant). study publication, but there are some discrepancies
between the text and the table. The body of the text
Neurological outcomes states that 3 patients had severe arterial hypotension.
This was refractory under high doses of vasopressor
 NIHSS score at 4 weeks=29
agents in 1 patient and the other 2 patients finally died of
 Mean Barthel Index at 3 months=65 (range 10 to 85) cardiac failure. It also states that 3 patients had severe
 Mean Rankin Scale score at 3 months=2.9 (range 2 to 5) coagulopathy, which was fatal in 1 patient.

The pancreatitis was treated conservatively.

The prevalence of pneumonia increased with longer

duration of hypothermia.

Mortality=38% (19/50)
2 patients died of cardiac failure, 1 patient died of severe
coagulopathy, 1 patient died of herniation on day 3 while
still under moderate hypothermia, and 15 patients died
after hypothermia was stopped and were caused by an
extensive increase in ICP.
Abbreviations used: ICP, intracranial pressure; NIHSS, National Institutes of Health Stroke Scale

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 21 of 42
 IP 1115 [IPGXXX]

Study 6 Mourand I (2012)

Details
Study type Case series
Country France
Recruitment period Not reported
Study population and n=19
number Patients with severe acute ischaemic stroke
Age and sex Mean 53 years (range 27 to 65); 68% male
Patient selection Inclusion criteria: age between 18 and 65 years; large middle cerebral artery infarction involving more than
criteria two-thirds of the area; an NIHSS score of ≥20 for left-hemispheric stroke and ≥17 for right-hemispheric
stroke, with an item level of consciousness of 1 (drowsy, but arousable by minor stimulation); starting
moderate hypothermia within 24 hours of symptom onset; not eligible for intravenous thrombolysis.
Exclusion criteria: sepsis within the 72 hours before stroke onset; any previous disabling neurological
disease or terminal illness that gave a life expectancy of <6 months.
Technique Hypothermia was induced as soon as possible, but always within the first 24 hours. A cooling blanket was
used (Blanketrol II, Cincinnati Sub-Zero) and adjusted to maintain a rectal temperature of between 32°C
and 33°C. Hypothermia was stopped when the midline shift was <5 mm without major lateral-ventricle
compression. Rewarming was induced at a rate of <1.5°C/24 hours. All patients were mechanically
ventilated, and paralysed for the first 72 hours to avoid shivering. Sedation was induced by high-dose
gamma hydroxybutyrate and low-dose midazolam and fentanyl. Norepinephrine was continually infused. A
hemicraniectomy was done when necessary. An intracranial pressure probe was not used.

Mean duration of hypothermia=22.6±4.9 days.

Follow-up 12 months
Conflict of None
interest/source of
funding

Analysis
Follow-up issues: No losses to follow-up were described.

Study design issues: Single-centre, prospective observational study with consecutive patients. The aim of the study was
to assess the feasibility and safety of very prolonged moderate hypothermia for severe acute ischaemic stroke.

Study population issues: The Alberta Stroke Program Early CT Score for all patients was ≤4 at study entry and was 0
for 50% of patients. The mean NIHSS was 20.9±2.0 and Glasgow Coma Score was 11.7±1.8.

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 22 of 42
 IP 1115 [IPGXXX]
Key efficacy and safety findings
Efficacy Safety
Number of patients analysed: 19 Complications
During the hypothermia procedure, all patients had hypotension,
Neurological outcomes which needed continuous infusion of norepinephrine, and
bradycardia of between 35 and 50 pulses per minute, without
The 10 patients who survived were discharged to rehabilitation arrhythmia.
programmes. All were living at home 12 months after the stroke.
Bacterial pneumonia=42% (8/19) (treated with antibiotics; 1
patient died)
 Mean NIHSS score at 12 months=8.3±2.7
 Mean Barthel Index at 12 months=67±18 Platelet counts continuously and slowly decreased in all patients;
 Mean modified Rankin Scale score at 12 months=3.2±0.9 a preventive platelet transfusion was needed in 8 patients
between days 15 and 20).

Stroke recurrence=26% (5/19) (4 happened during the cooling In-hospital and 90-day mortality=47% (9/19) (4 deaths were
period and 1 at 5 days after rewarming; 3 patients had severe during the hypothermia treatment between days 1 and 11
clinical deterioration and died, and 2 survived [1 of whom had a [1 cardiac arrest caused by a myocardial infarct present in the
hemicraniectomy]). acute phase, 1 multiple organ failure, and 2 uncontrolled
intracranial hypotension], 3 patients died during the rewarming
phase between days 15 and 27 [1 nosocomial pneumonia, 2
Uncontrolled intracranial pressure with cerebral uncontrolled intracranial hypertension] and 2 deaths were after
herniation=31.6% (6/19) (5 patients had a hemicraniectomy rewarming [both caused by stroke recurrence]. There was a
within the first 9 days of the hypothermia period; 4 patients significant difference in the mean duration to reach target
survived after the hemicraniectomy). temperature between patients who survived [8.1±5.5 hours] and
those who died [14.0±7 hours], p=0.031).

Comparisons between survivors and deceased patients showed

that only haematocrit was statistically different by day 8 between
the 2 groups (34.8% versus 42.5%, p=0.038).

Abbreviations used: NIHSS, National Institute of Health Stroke Scale

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 23 of 42
 IP 1115 [IPGXXX]

Study 7 Schneider H (2017)

Details
Study type Non-randomised comparative study
Country Germany
Recruitment period 2001 to 2010
Study population and n=111 (53 hemicraniectomy and hypothermia versus 58 hemicraniectomy alone)
number Patients with space-occupying cerebral infarction
Age and sex Mean age 48 years; 58% (64/111) male
Patient selection Inclusion criteria for hypothermia group: age 18 to 60 years; clinical deficits suggestive of large infarction
criteria in the territory of the middle cerebral artery; decrease of consciousness (score of 1 or greater on item 1a
of the NIHSS); signs on CT or MRI of an infarct of at least 50% of the middle cerebral artery territory, with
or without additional infarction in the territory of the ipsilateral anterior or posterior cerebral artery;
hemicraniectomy within 48 hours after onset of symptoms; no fixed pupils before hemicraniectomy; no
contralateral ischaemia or other brain lesion that could affect outcome; hypothermia treatment.
Inclusion criteria for control group were comparable and included age 18 to 60 years and time from
symptom onset to hemicraniectomy 48 hours.
Technique All patients had a hemicraniectomy (median 23.5 hours after symptom onset).
In the hypothermia group, hypothermia was induced immediately after surgery and patients were
transferred to the neurocritical care unit. An intracranial pressure probe was inserted ipsilaterally in the
affected brain tissue. The target temperature was 33°C to 34°C, except in patients with known cardiac
arrhythmia or coagulopathy when the target was 35°C. The target temperature was maintained for at least
96 hours with a rewarming phase in steps of 1°/day to normothermia (36.5°C). From 2001 to 2004, cooling
mats were used for surface cooling (Blanketrol, CSZ, Cincinnati) and from 2004 to 2010, surface cooling
or endovascular temperature management were used (Arctic Sun, Medivance Inc. US; CoolGard, Alsius,
US). No specific anti-shivering medication or methods were needed because patients were deeply
sedated and showed no shivering. Core body temperature was measured with a urinary bladder catheter.
Follow-up 12 months
Conflict of None
interest/source of
funding

Analysis
Follow-up issues: No losses to follow-up were described.

Study design issues: Retrospective observational study. Patients in the hypothermia group all had treatment at a single
centre. The comparison group consisted of patients who had early hemicraniectomy from 3 randomised controlled trials.
The primary outcome measure was the functional outcome, determined by the mRS score at 12 months, dichotomised 0
to 3 versus 4 to 6.

Study population issues: In the hypothermia group, mean age was higher than in the control group (50 versus 46 years)
and there were more males (64% versus 52%). Cardiovascular risk factors were more frequent in the hypothermia group
than in the control group (diabetes: 23% versus 9%; arterial hypertension: 67% versus 37%). Baseline stroke severity was
similar in the 2 groups according to NIHSS scores, but the dominant hemisphere was more often affected in the
hypothermia group (60% versus 52%). Body temperature on admission was higher in the control group (37.2°C versus
36.3°C).

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 24 of 42
 IP 1115 [IPGXXX]
Key efficacy and safety findings
Efficacy Safety
Number of patients analysed: 111 (53 versus 58) Events reported in the hypothermia group during the intensive
care unit stay:
mRS score of 0 to 3 at 12 months  Severe sepsis=11% (6/53)
 Hypothermia=25% (13/53)  Tracheobronchitis or pneumonia=74% (39/53)
 Control=41% (24/58)  Deep venous thrombosis=9% (5/53)
RR=0.59, 95% CI 0.34 to 1.04  Cardiac arrhythmia=28% (15/53)
RR adjusted for age and NIHSS score at baseline=0.66, 95% CI
0.38 to 1.13 Mortality in hypothermia group
Adjustment for a propensity score based on age, hours to 20 patients died during the first 14 days after symptom onset; 18
surgery, and hypertension: RR 0.71, 95% CI 0.40 to 1.24 deaths were caused by intracranial herniation, 1 was caused by
severe sepsis and 1 was caused by malignant cardiac
arrhythmia.
mRS score of 0 to 4 at 12 months
 Hypothermia=40% (21/53)
 Control=72% (42/58)
RR=0.54, 95% CI 0.38 to 0.79

Survival after 12 months

 Hypothermia=49% (26/53)
 Control=79% (46/58)
RR=0.62, 95% CI 0.46 to 0.84

The results for mRS score of 0 to 4 and survival at 12 months

did not change significantly after adjusting for various factors.

Abbreviations used: CI, confidence interval; mRS, modified Rankin Scale; NIHSS, National Institutes for Health Stroke Scale; RR,
risk ratio

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 25 of 42
 IP 1115 [IPGXXX]

Study 8 Park HS (2018)

Details
Study type Non-randomised comparative study
Country Korea
Recruitment period 2012 to 2016
Study population and n=47 (20 hypothermia and hemicraniectomy, 27 hemicraniectomy alone)
number Patients with malignant middle cerebral artery infarction
Age and sex Hypothermia group: mean age 62 years; 60% (12/20) male
Control group: mean age 59 years; 63% (17/27) male
Patient selection Inclusion criteria: acute cerebral infarction involving more than two-thirds of middle cerebral artery territory
criteria and a mental status decrease to stupor or worse (a Glasgow coma scale score <11) with midline shift
>10 mm or a transtentorial herniation sign (a fixed and dilated pupil).
Technique Hypothermia was induced immediately after hemicraniectomy or after the decision of hypothermia during
postoperative care. Hypothermia was achieved by a cooling blanket (Blanketrol III, Cincinnati Sub-Zero,
USA) set to maintain the rectal temperature at 34°C. The duration of hypothermia was determined
according to the findings on CT scans, done every 2 to 3 days, as well as the condition of the patient.
Active rewarming was done at a rate of 0.2°C/hour. Patients were sedated with midazolam and
vecuronium and ventilated. Hypothermia was maintained for 4±2 days (range 1 to 7).
All patients had a decompressive hemicraniectomy. Intracranial pressure monitoring was not used.
Follow-up 1 year
Conflict of None
interest/source of
funding

Analysis
Follow-up issues: 18 of the 20 patients in the hypothermia group and 25 of the 27 patients in the control group were
followed up to 1 year.

Study design issues: Retrospective analysis of prospectively maintained stroke registry. Patients in the hypothermia
group had treatment between 2012 and 2016. Control group patients had treatment between 2010 and 2012; patients
who met the inclusion criteria for hypothermia were selected. The primary outcome of the study was defined as either
death or survival within the hospital.

Study population issues: In the hypothermia group, 35% (7/20) of patients showed a transtentorial herniation sign (a
fixed and dilated pupil) before hypothermia. There were no statistically significant differences between the groups with
regard to age, preoperative infarct volume, Glasgow Coma score, National Institutes of Health Stroke Scale score before
surgery, maximal degree of midline shift, and time interval between symptom onset and hemicraniectomy.

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 26 of 42
 IP 1115 [IPGXXX]
Key efficacy and safety findings
Efficacy Safety
Number of patients analysed: 47 (20 vs 27) Hypothermia-related complications and serious adverse
In-hospital mortality events in the intensive care unit
 Hypothermia=15.0%  Arrhythmia=10% (2/20)
 Control=40.7%, p=0.056  Sepsis=5% (1/20)
 Pneumonia=5% (1/20)
The causes of death in the hypothermia group were  Hypotension=5% (1/20)
transtentorial herniation (n=2) and cardiac failure (n=1). In the In 3 patients, hypothermia was stopped because of these side-
control group, the causes of death were transtentorial herniation effects (1 patient with sepsis, 1 with hypotension and 1 with
(n=7), pneumonia (n=2), pulmonary thromboembolism (n=1), bradycardia).
and acute kidney injury (n=1).
Of all the adverse events during the stay in the intensive care
Hypothermia was identified as the strongest factor that affected unit, pneumonia was the most common (n=5, 25%), but there
survival in the multivariable analysis (odds ratio 6.21, 95% were no statistically significant differences compared with the
confidence interval 1.04 to 37.05, p=0.045). control group (n=5, 18.5%, p=0.723).

Mortality at 1 year Haemorrhagic transformation after hemicraniectomy

 Hypothermia=27.8% (5/18)  Hypothermia=35.0% (7/20)
 Control=52.0% (13/25), p=0.112  Control=25.9% (7/27), p=0.501

Functional outcomes at 1 year follow-up were not statistically Only 1 patient in the control group had revision surgery for
significantly different between the 2 groups. removal of intracerebral haemorrhage in the basal ganglia.

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 27 of 42
 IP 1115 [IPGXXX]

Study 9 Petersson J on behalf of the EuroHYP-1 investigators (October 2018) – conference

abstract (included for safety data only)
Details
Study type Randomised controlled trial (EuroHYP-1) – conference abstract
Country 18 European countries
Recruitment period Not reported
Study population and n=98
number Patients with acute ischaemic stroke
Age and sex Not reported
Patient selection Not reported
criteria
Technique Cooling was started <6 hours after onset of symptoms and <90 mins of start of thrombolysis.
Induction with 20 ml/kg refrigerated normal saline (4°C) over 30–60 mins.
Cooling was maintained at 34°C to 35°C for 12 or 24 hours with surface or endovascular techniques.
Follow-up Not reported
Conflict of Not reported
interest/source of
funding

Analysis
Study design issues: The original target sample size was 1500. Because of slow recruitment, a new target sample size
was set at 80% power with a target of 800 patients. Intention-to-treat analysis.

Key efficacy and safety findings

Efficacy Safety
Number of patients analysed: 98
Efficacy data were not extracted Adverse events
because the source is an unpublished xxxx xxxx
conference presentation.
xxxxxxxxx xx xx
xxxxxxxxx xx xx
xxxxxxxxx xx xx
xxxxxxxxx xx xx
xxxxxxxxx xx xx

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 28 of 42
 IP 1115 [IPGXXX]

Validity and generalisability of the studies

 There are different techniques and devices used for cooling, which may have
different safety and efficacy profiles.
 The target temperature and the duration of hypothermia varied between the
studies.
 Induction time and rewarming speed varied between the studies.
 Different methods of temperature monitoring were used.
 In some studies, patients were sedated and mechanically ventilated during the
hypothermia treatment.
 Some patients had additional treatments, such as thrombolysis and
hemicraniectomy, as well as therapeutic hypothermia.

Existing assessments of this procedure

The European Stroke Organisation published guidelines for the management of
temperature in patients with acute ischaemic stroke in 2015.10 With regard to
induction of hypothermia, the recommendation states:

‘In patients with acute ischemic stroke, we do not recommend induction of

hypothermia as a means to improve functional outcome and/or survival.’

The quality of evidence was graded as very low and the strength of the
recommendation was defined as weak, according to the GRADE methodology.
The report strongly encouraged recruitment of eligible patients to ongoing RCTs.

Recent guidelines from a French expert panel stated:

‘We suggest considering targeted temperature management (TTM) at

normothermia during the early phase of severe ischaemic stroke. (Expert
opinion)

Rationale: Hyperthermia or fever is a frequent complication (>50%) in

patients at the acute phase of stroke and is correlated with poor functional
outcome. However, the efficacy of therapeutic hypothermia has not yet
been shown, according to 6 randomized trials that tested hypothermia
(33–35 °C) in stroke patients. There were few patients and methodological
biases were numerous. A single study investigated patients with severe
stroke (NIHSS >15). Two randomized studies are ongoing: EuroHYP-1
explores the value of 24 h; 34–35°C hypothermia following recent stroke,
IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 29 of 42
 IP 1115 [IPGXXX]

and the recent ICTuS 2/3 trial. In that later study, intravascular therapeutic
hypothermia was found safe and feasible in patients treated with
recombinant tissue-type plasminogen activator.’11

Related NICE guidance

Below is a list of NICE guidance related to this procedure.

Interventional procedures

 Therapeutic hypothermia following cardiac arrest. NICE interventional

procedures guidance 368 (2011). Available from
https://www.nice.org.uk/guidance/ipg386

 Therapeutic hypothermia with intracorporeal temperature monitoring for

hypoxic perinatal brain injury. NICE interventional procedures guidance 347
(2011). Available from https://www.nice.org.uk/guidance/ipg347

 Mechanical clot retrieval for treating acute ischaemic stroke. NICE

interventional procedures guidance 548 (2016). Available from
https://www.nice.org.uk/guidance/ipg548

Technology appraisals

 Alteplase for treating acute ischaemic stroke. NICE technology appraisal 264
(2012). Available from http://www.nice.org.uk/guidance/TA264

NICE guidelines

 Stroke and transient ischaemic attack in over 16s: diagnosis and initial
management. NICE clinical guideline 68 (2008). Available from
http://www.nice.org.uk/guidance/CG68 [update in development]

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 30 of 42
 IP 1115 [IPGXXX]

Additional information considered by IPAC

Specialist advisers’ opinions

Specialist advice was sought from consultants who have been nominated or
ratified by their Specialist Society or Royal College. The advice received is their
individual opinion and is not intended to represent the view of the society. The
advice provided by Specialist Advisers, in the form of the completed
questionnaires, is normally published in full on the NICE website during public
consultation, except in circumstances but not limited to, where comments are
considered voluminous, or publication would be unlawful or inappropriate.
One Specialist Adviser Questionnaire for therapeutic hypothermia for acute
ischaemic stroke were submitted and can be found on the NICE website.

Patient commentators’ opinions

NICE’s Public Involvement Programme will send questionnaires to NHS trusts for
distribution to patients who had the procedure (or their carers). When NICE has
received the completed questionnaires, these will be discussed by the
committee.

Company engagement
A structured information request was sent to 5 companies who manufacture a
potentially relevant device for use in this procedure. NICE received 1 completed
submission. This was considered by the IP team and any relevant points have
been taken into consideration when preparing this overview.

Issues for consideration by IPAC

 Ongoing trials:
 Mild Hypothermia After Endovascular Treatment in Acute Ischemic Stroke
(HELMET) (NCT02985060); RCT; Korea; estimated enrolment 40;
estimated study completion date November 2017.
 Cooling Plus Best Medical Treatment Versus Best Medical Treatment Alone
for Acute Ischaemic Stroke (EuroHYP-1) (NCT01833312); RCT; Germany;
estimated enrolment 800; estimated completion date December 2020.

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 31 of 42
 IP 1115 [IPGXXX]

 Evidence on reducing the temperature to normal in patients with fever after

ischaemic stroke has not been included.

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 32 of 42
 IP 1115 [IPGXXX]

References
1. Wan Y-H, Nie C, Wang H-L et al. (2014) Therapeutic hypothermia (different
depths, durations, and rewarming speeds) for acute ischemic stroke: a
meta-analysis. Journal of stroke and cerebrovascular diseases: the official
journal of National Stroke Association 23: 2736–47
2. Lyden P, Hemmen T, Grotta J et al. (2016) Results of the ICTuS 2 Trial
(Intravascular Cooling in the Treatment of Stroke 2). Stroke 47: 2888–95
3. Su Y, Fan L, Zhang Y et al. (2016) Improved Neurological Outcome With
Mild Hypothermia in Surviving Patients With Massive Cerebral Hemispheric
Infarction. Stroke 47: 457–63
4. Hong JM, Lee JS, Song H-J et al. (2014) Therapeutic hypothermia after
recanalization in patients with acute ischemic stroke. Stroke 45: 134–40
5. Schwab S, Georgiadis D, Berrouschot J et al. (2001) Feasibility and safety
of moderate hypothermia after massive hemispheric infarction. Stroke 32:
2033–5
6. Mourand I, Escuret E, Heroum C et al. (2012) Feasibility of hypothermia
beyond 3 weeks in severe ischemic stroke: an open pilot study using
gamma-hydroxybutyrate. Journal of the neurological sciences 316: 104–7
7. Schneider H, Kruger P, Algra A et al. (2017) No benefits of hypothermia in
patients treated with hemicraniectomy for large ischemic stroke.
International journal of stroke: official journal of the International Stroke
Society 12: 732–40
8. Park H-S, Choi J-H (2018) Safety and Efficacy of Hypothermia (34°C) after
Hemicraniectomy for Malignant MCA Infarction. Journal of Korean
Neurosurgical Society 61: 267–76
9. Petersson J on behalf of the EuroHYP-1 investigators (2018) Therapeutic
Hypothermia in Acute Ischemic Stroke - Main Results. Presented at the
World Stroke Congress, Montreal, 17 to 20 October 2018.
10. Ntaios G, Dziedzic T, Michel P et al. (2015) European Stroke Organisation
(ESO) guidelines for the management of temperature in patients with acute
ischemic stroke. International journal of stroke: official journal of the
International Stroke Society 10: 941–9
11. Cariou A, Payen J-F, Asehnoune K et al. (2017) Targeted temperature
management in the ICU: guidelines from a French expert panel. Annals of
Intensive Care 7:70

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 33 of 42
 IP 1115 [IPGXXX]

Literature search strategy

Databases Date Version/files

searched

Cochrane Database of Systematic 25/05/2018 Issue 5 of 12, May 2018

Reviews – CDSR (Cochrane)
HTA database (Cochrane) 25/05/2018 Issue 4 of 4, October 2016

Cochrane Central Database of Controlled 25/05/2018 Issue 4 of 12, April 2018

Trials – CENTRAL (Cochrane)

MEDLINE (Ovid) 25/05/2018 1946 to Present with Daily

Update
MEDLINE In-Process (Ovid) 25/05/2018 May 24, 2018
MEDLINE Epubs ahead of print (Ovid) 25/05/2018 May 24, 2018

EMBASE (Ovid) 25/05/2018 1974 to 2018 May 24

BLIC (British Library) 25/05/2018 n/a

Trial sources searched 20th February 2018

 Clinicaltrials.gov
 ISRCTN
 WHO International Clinical Trials Registry

Websites searched 20th February 2018

 National Institute for Health and Care Excellence (NICE)
 NHS England
 Food and Drug Administration (FDA) - MAUDE database
 Australian Safety and Efficacy Register of New Interventional Procedures –
Surgical (ASERNIP – S)
 Australia and New Zealand Horizon Scanning Network (ANZHSN)
 EuroScan
 General internet search

The following search strategy was used to identify papers in MEDLINE. A similar
strategy was used to identify papers in other databases.

1 Stroke/
2 Cerebral Infarction/ or Brain Ischemia/ or *Arterial Occlusive Diseases/
3 ((acute or isch?emi* or thrombotic or embolic) adj3 stroke).tw.
4 ((brain or cerebral*) adj2 (isch?emi* or infarct*)).tw.

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 34 of 42
 IP 1115 [IPGXXX]

5 (arter* adj2 occlusi*).tw.

6 ((cerebrovascular* or vascular*) adj2 accident*).tw.
7 CVA.tw.
8 or/1-7
9 ((surface or endovascular or circulat* or therap* or neuroprotect* or device* or
induc*) adj4 (cool* or chill* or hypotherm*)).tw.
10 Hypothermia, Induced/
11 (temperature adj4 (manag* or target*) adj4 (cool* or chill* or reduc* or low*)).tw.
12 or/9-11
13 8 and 12
14 (iqool or braincool).tw.
15 13 or 14
16 animals/ not humans/
17 15 not 16
18 limit 17 to english language

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 35 of 42
 IP 1115 [IPGXXX]

Appendix
The following table outlines the studies that are considered potentially relevant to
the IP overview but were not included in the main data extraction table (table 2).
It is by no means an exhaustive list of potentially relevant studies. Case reports
have been excluded, unless they report a unique safety event.
Article Number of Direction of conclusions Reasons for
patients/ non-inclusion
follow-up in table 2
Abou-Chebl A, DeGeorgia, MA, Case series Surface cooling for the treatment Larger or more
Andrefsky JC et al. (2004) n=18 of acute ischemic stroke can be recent studies
Technical refinements and performed rapidly with early are included.
drawbacks of a surface cooling neuromuscular paralysis.
technique for the treatment of Advanced age and prolonged
severe acute ischemic stroke. hypothermia may be associated
Neurocritical care 1: 131-43 with an increased risk of
complications.
Abou-Chebl A, Sung G, Barbut Case series Intranasal cooling with the Small case
D et al. (2011) Local brain n=15 RhinoChill device appears safe series, with
temperature reduction through and effectively lowers brain and mixed
intranasal cooling with the core temperatures. Further study indications.
RhinoChill device: preliminary is warranted to assess the efficacy
safety data in brain-injured of hypothermia through intranasal
patients. Stroke 42: 2164-9 cooling for brain-injured patients.
Bi M, Ma Q, Zhang S et al. RCT There was no benefit of combined Trial is included
(2011) Local mild hypothermia n=93 (31 vs local hypothermia/IV rtPA in systematic
with thrombolysis for acute 31 vs 31) treatment compared to IV rtPA review by Wan
ischemic stroke within a 6-h alone. YH et al., 2014
window. Clinical neurology and FU=90 days (study 1), cited
neurosurgery 113: 768-73 as Tong, 2011.
Chen J, Liu L, Zhang H et al. Case series Selective brain cooling by intra- Larger or more
(2016) Endovascular n=26 arterial infusion of cold saline recent studies
Hypothermia in Acute Ischemic combined with endovascular are included.
Stroke: Pilot Study of Selective recanalisation therapy in acute
Intra-Arterial Cold Saline ischemic stroke seems feasible
Infusion. Stroke 47: 1933-5 and safe.
De Georgia MA, Krieger DW, RCT Induced moderate hypothermia is Study is
Abou-Chebl A et al. (2004) n=40 (18 vs feasible using an endovascular included in
Cooling for Acute Ischemic Brain 22) cooling device in most patients systematic
Damage (COOL AID): a with acute ischemic stroke. review by Wan
feasibility trial of endovascular FU=1 month Further studies are needed to YH et al., 2014
cooling. Neurology 63: 312-7 determine if hypothermia (study 1).
improves outcome.
Den Hertog HM, van der Worp Systematic There was no statistically A more recent
HB, Tseng MC et al. (2009) review significant effect of systematic
Cooling therapy for acute stroke. n=132 pharmacological or physical review is
Cochrane Database of (3 studies on temperature-lowering therapy in included.
Systematic Reviews 2009, Issue physical reducing the risk of death or All the relevant
1. Art. No.: CD001247. DOI: cooling) dependency (odds ratio (OR) 0.9, studies
10.1002/14651858.CD001247.p 95% confidence interval (CI) 0.6 identified are
ub2. to 1.4) or death (OR 0.9, 95% CI included in table
0.5 to 1.5). Both interventions 2 or the
were associated with a non-

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 36 of 42
 IP 1115 [IPGXXX]

significant increase in the appendix of this

occurrence of infections. overview.
Els T, Oehm E, Voigt S et al. RCT The present study suggests that a Study is
(2006) Safety and therapeutical n=25 (12 vs combined therapy of mild included in
benefit of hemicraniectomy 13) hypothermia and hemicraniectomy systematic
combined with mild hypothermia in malignant brain infarction does review by Wan
in comparison with FU=6 not imply additional risks by side YH et al., 2014
hemicraniectomy alone in months effects and improves functional (study 1).
patients with malignant ischemic outcome as compared with
stroke. Cerebrovascular hemicraniectomy alone.
diseases 21: 79-85
Froehler MT, Ovbiagele B review Rapid induction of hypothermia is A more recent
(2010) Therapeutic hypothermia key and is best accomplished with systematic
for acute ischemic stroke. Expert a combination of ice-cold saline review is
review of cardiovascular therapy infusion and the use of included.
8: 593-603 endovascular cooling devices.
Shivering can be overcome with
aggressive anti-shivering
protocols including meperidine,
buspirone and surface warming. If
proven efficacious, hypothermia
would be a welcome complement
to established reperfusion
therapies for ischaemic stroke
patients.
Georgiadis D, Schwarz S, Case series Induction and maintenance of Larger or more
Kollmar R et al. (2001) n=6 hypothermia with an intravenous recent studies
Endovascular cooling for cooling device are feasible. The are included.
moderate hypothermia in safety of this approach remains to
patients with acute stroke: first be evaluated.
results of a novel approach.
Stroke 32: 2550–3
Georgiadis D, Schwarz, S, Non- In patients with severe ischaemic Larger or more
Aschoff A et al. (2002) randomised stroke, hemicraniectomy results in recent studies
Hemicraniectomy and moderate comparative lower mortality and lower are included.
hypothermia in patients with study complication rates compared with
severe ischemic stroke. Stroke n=36 moderate hypothermia. Both
33: 1584-8 treatment modalities, however,
are associated with intensive
medical treatment and a
prolonged stay in the neurological
intensive care unit.
Georgiadis D, Schwarz S, Evans Case series Static cerebral autoregulation Larger or more
DH et al. (2002) Cerebral n=14 appears intact under moderate recent studies
autoregulation under moderate hypothermia with the use of alpha- are included.
hypothermia in patients with stat for pH maintenance.
acute stroke. Stroke 33: 3026–9
Geurts M, Petersson J, Brizzi M RCT In awake patients with acute Small study,
et al. (2017) COOLIST (Cooling n=22 ischaemic stroke, surface cooling which assessed
for Ischemic Stroke Trial) A is feasible to 35°C but not to feasibility and
multicentre, open, randomized, FU=3 34.5°C and 34°C. Cooling is safety of cooling
phase II, clinical trial. Stroke 48: months associated with an increased risk to different
219–21 of pneumonia. target
temperatures.
Guluma KZ, Hemmen TM, Case series The procedure was well tolerated, Larger or more
Olsen SE et al. (2006) A trial of n=10 with minimal shivering and no recent studies
therapeutic hypothermia via rebound hyperthermia. are included.
endovascular approach in
IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 37 of 42
 IP 1115 [IPGXXX]

awake patients with acute

ischemic stroke: methodology.
Academic emergency medicine:
official journal of the Society for
Academic Emergency Medicine
13: 820-7
Guluma KZ, Oh H, Yu SW et al. Case series Endovascular hypothermia Larger or more
(2008) Effect of endovascular n=18 decreases acute post-ischemic recent studies
hypothermia on acute ischemic cerebral oedema. A larger trial is are included.
edema: morphometric analysis FU=30 days warranted to determine if it affects
of the ICTuS trial. Neurocritical final infarct volume and outcome
care 8: 42-7 in stroke.

Guluma KZ, Liu L, Hemmen TM Case series Urine output decreased 5.1 Larger or more
et al. (2010) Therapeutic n=11 ml/hour for every 1°C decrease in recent studies
hypothermia is associated with a body temperature (p=0.001), with are included.
decrease in urine output in acute no associated serious adverse
stroke patients. Resuscitation events.
81: 1642-7
Harris B, Andrews PJD, Murray Systematic Out of 46 head-cooling studies in Review focuses
GD et al. (2012) Systematic review TBI and stroke, there were no on head cooling.
review of head cooling in adults RCTs of suitable quality for formal A more recent
after traumatic brain injury and outcome analysis. Some methods systematic
stroke. Health Technology of head cooling can reduce review is
Assessment 16 (45) intracranial temperature, which is included.
an important first step in
determining effectiveness, but
there is insufficient evidence to
recommend its use outside of
research trials.
Hemmen TM, Raman R, RCT This study demonstrates the Study is
Guluma KZ et al. (2010) n=58 (28 vs feasibility and preliminary safety of included in
Intravenous thrombolysis plus 30) combining endovascular systematic
hypothermia for acute treatment hypothermia after stroke with review by Wan
of ischemic stroke (ICTuS-L): FU=90 days intravenous thrombolysis. YH et al., 2014
final results. Stroke 41: 2265-70 Pneumonia was more frequent (study 1).
after hypothermia, but further
studies are needed to determine
its effect on patient outcome and
whether it can be prevented.
Horn CM, Sun CHJ, Nogueira Case series Hypothermia can be safely Larger or more
RG et al. (2014) Endovascular n=20 performed after definitive intra- recent studies
Reperfusion and Cooling in arterial reperfusion therapy in are included.
Cerebral Acute Ischemia FU=90 days patients with large pretreatment
(ReCCLAIM I). Journal of core infarcts. A phase II study
neurointerventional surgery 6: randomising patients to
91–5 hypothermia or normothermia is
needed to properly assess the
efficacy of hypothermia as a
neuroprotectant for reperfusion
injury.
Horstmann S, Koziol JA, Non- Hypothermia reduces matrix Larger or more
Martinez-Torres F et al. (2009) randomised metalloproteinase 9 activity as recent studies
Sonographic monitoring of mass comparative well as midline shift. Transcranial are included.
effect in stroke patients treated study duplex sonography may reduce
with hypothermia. Correlation n=30 the need for repetitive CT scans in
with intracranial pressure and neurological critically ill patients.
matrix metalloproteinase 2 and 9
IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 38 of 42
 IP 1115 [IPGXXX]

expression. Journal of the

neurological sciences 276: 75-8
Hwang YH, Jeon JS, Kim YW et Case series Favourable outcome (modified Small case
al. (2017) Impact of immediate n=18 Rankin Scale <=2) at 3 months series with short
post-reperfusion cooling on was observed in 10 (56%) term follow-up.
outcome in patients with acute FU=3 patients. Symptomatic intracranial
stroke and substantial ischemic months haemorrhage, malignant brain
changes. Journal of oedema, and pneumonia were
neurointerventional surgery 9: observed in 2, 6, and 8 patients,
21-25 respectively.
Jaramillo A, Illanes S, Diaz V Review Moderate hypothermia A more recent
(2008) Is hypothermia useful in n=196 (11 ameliorates ischemic injury by systematic
malignant ischemic stroke? studies) multiple mechanisms. Treatment review with
Current status and future of acute ischemic stroke patients meta-analysis is
perspectives. Journal of the is feasible, and additional studies, included (study
neurological sciences 266: 1-8 including randomised clinical 1).
trials, are warranted.
Kammersgaard LP, Rasmussen Non- Modest hypothermia can be Larger or more
BH, Jorgensen HS et al. (2000) randomised achieved in awake patients with recent studies
Feasibility and safety of inducing comparative acute stroke by surface cooling are included.
modest hypothermia in awake study with the "forced air" method, in
patients with acute stroke n=73 (17 vs combination with pethidine to treat
through surface cooling: A case- 56) shivering. It was not associated
control study: the Copenhagen with a poor outcome.
Stroke Study. Stroke 31: 2251–6 FU=6
months
Kim JY, Yenari MA (2015) Review Animal models and the method of No meta-
Hypothermia for treatment of cooling used in the laboratory are analysis.
stroke. Brain Circulation 1:14-25 quite different from those
employed clinically. Thus, an
effort to simulate the clinical The relevant
condition more precisely might cited studies are
provide solutions for better and included.
wider application of therapeutic
hypothermia in human patients.
Second, there are few
investigations into overcoming the
complications of systemic
hypothermia such as shivering,
infection, and coagulopathies.
Though these complications are
largely ignored in the laboratory,
they are significant at the clinical
level and will need to be
addressed.
Kollmar R, Schellinger PD, Case series This pilot study suggests that Larger or more
Steigleder T et al. (2009) Ice- n=10 rapid ice-cold saline infusions in recent studies
cold saline for the induction of combination with pethidine and are included.
mild hypothermia in patients with buspirone lower the body
acute ischemic stroke: a pilot temperature significantly without
study. major side effects.
Stroke 40: 1907–9
Krieger DW, De Georgia MA., Non- Induced hypothermia appears Larger or more
Abou-Chebl A et al. (2001) randomised feasible and safe in patients with recent studies
Cooling for acute ischemic brain comparative acute ischemic stroke even after are included.
damage (cool aid): an open pilot study thrombolysis. Refinements of the
study of induced hypothermia in n=19 (10 vs cooling process, optimal target
9) temperature, duration of therapy,

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 39 of 42
 IP 1115 [IPGXXX]

acute ischemic stroke. Stroke and, most important, clinical

32: 1847-54 efficacy, need further study.
Lakhan SE, Pamplona F (2012) Systematic Hypothermia does not significantly A more recent
Application of mild therapeutic review and improve stroke severity; however, systematic
hypothermia on stroke: a meta- this finding should be taken with review is
systematic review and meta- analysis caution due to the high included.
analysis. Stroke research and n=288 (7 heterogeneity and limited number
treatment 295906 trials) of included studies. No impact on
mortality was observed.
Lyden PD, Allgren RL, Ng K et Case series Endovascular cooling with a Larger or more
al. (2005) Intravascular Cooling n=18 proactive anti-shivering regimen recent studies
in the Treatment of Stroke can be accomplished in awake are included.
(ICTuS): early clinical stroke patients. Further studies
experience. Journal of stroke are needed to establish the safety
and cerebrovascular diseases: and efficacy of this approach.
the official journal of National
Stroke Association 14: 107-14
Lyden PD, Krieger D, Yenari M Review Therapeutic hypothermia is A more recent
et al. (2006) Therapeutic promising, but further Phase 1 systematic
hypothermia for acute stroke. and Phase 2 development efforts review is
International journal of stroke: are needed to ensure that cooling included.
official journal of the of stroke patients is safe, before
International Stroke Society 1: 9- definitive efficacy trials.
19
Martin-Schild S, Hallevi H, Case series Combining caffeinol with Larger or more
Shaltoni H et al. (2009) n=20 hypothermia in patients with acute recent studies
Combined Neuroprotective stroke who have IV t-PA is are included.
Modalities Coupled with feasible. A prospective placebo-
Thrombolysis in Acute Ischemic controlled randomised study is
Stroke: A Pilot Study of Caffeinol needed to further assess safety
and Mild Hypothermia. Journal and to test the efficacy of
of Stroke and Cerebrovascular caffeinol, hypothermia, or both
Diseases 18: 86-96
Meijer RJ, Visser H, Koudstaal Case series The study suggests that mild Larger or more
PJ et al. (2001) Lowering body n=8 hypothermia in non-comatose recent studies
temperature in acute ischemic stroke patients during a period of are included.
stroke without artificial 24 hours after the ictus may be
ventilation and heavy sedation: accomplished with a cooling
a feasibility study. Journal of blanket and light sedation with
stroke and cerebrovascular midazolam in a well-equipped
diseases: the official journal of stroke unit.
National Stroke Association 10:
157-60
Milhaud D, Thouvenot E, Case series Prolonged hypothermia with Larger or more
Heroum C et al. (2005) n=12 gamma-hydroxybutyrate can be recent studies
Prolonged moderate used to treat malignant cerebral are included.
hypothermia in massive FU=6 infarction patients, with a fairly
hemispheric infarction: clinical months good clinical outcome for
experience. Journal of survivors.
neurosurgical anesthesiology
17: 49–53
Ovesen C, Brizzi M, Pott FC et RCT Therapeutic hypothermia with Trial is included
al. (2013) Feasibility of n=31 (17 vs general anaesthesia is feasible in in systematic
endovascular and surface 14) stroke patients. There were review by Wan
cooling strategies in acute increased rates of pneumonia, YH et al., 2014
stroke. Acta neurologica FU=30 days while the length of hospital stay (study 1), cited
Scandinavica 127: 399-405 remained comparable. The
endovascular cooling strategy
IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 40 of 42
 IP 1115 [IPGXXX]

provides a faster induction period as Krieger,

than surface cooling. 2013.
Piironen K, Tiainen M, RCT Mild hypothermia with a surface- Study is
Mustanoja S et al. (2014) Mild n=36 (18 vs cooling device in an acute stroke included in
hypothermia after intravenous 18) unit is safe and feasible in systematic
thrombolysis in patients with thrombolysed, spontaneously review by Wan
acute stroke: a randomized FU=3 breathing patients with stroke, YH et al., 2014
controlled trial. Stroke 45: 486- months despite the adverse events. (study 1).
91
Poli S, Purrucker J, Priglinger M Case series Although the decrease of brain Larger or more
et al. (2013) Induction of cooling n=11 temperature after cooling device recent studies
with a passive head and neck application was statistically are included.
cooling device: effects on brain significant, we doubt clinical
temperature after stroke. Stroke relevance of this rather limited
44: 708-13 effect (-0.36°C). Moreover, the
transient increases of blood
pressure and ICP warrant caution.
Poli S, Purrucker J, Priglinger M RCT In intubated stroke patients, brain Small RCT,
et al. (2014) Rapid Induction of n=20 cooling is faster during cold comparing
COOLing in Stroke Patients infusions (CI) than during 2 different
(iCOOL1): a randomised pilot nasopharyngeal cooling (NPC). cooling
study comparing cold infusions Importantly, contrary to previous methods.
with nasopharyngeal cooling. expectations, brain cooling
Critical care 18: 582 stopped soon after CI cessation.
Oesophageal but neither bladder
nor rectal temperature is suited as
surrogate for brain temperature
during CI and NPC. Several
severe adverse events demand
further studying of safety.
Poli S, Purrucker J, Priglinger M Case series The RhinoChill system cools the Larger or more
et al. (2014) Safety evaluation of n=10 brain efficiently. However, steep recent studies
nasopharyngeal cooling increases in blood pressure raise are included.
(RhinoChill) in stroke patients: FU=6 serious concerns regarding the
an observational study. months safety of its use in stroke patients.
Neurocritical care 20: 98-105
Schwab S, Schwarz S, Spranger Case series Moderate hypothermia in the Larger or more
M et al. (1998) Moderate n=25 treatment of severe cerebral recent studies
hypothermia in the treatment of ischemia is not associated with are included.
patients with severe middle FU=3 severe side effects. Moderate
cerebral artery infarction. Stroke months hypothermia can help to control
29: 2461-6 critically elevated ICP values in
severe space-occupying oedema
after MCA stroke and may
improve clinical outcome in these
patients.
Tahir RA, Pabaney AH (2016) review Although hypothermia has been A systematic
Therapeutic hypothermia and used for various purposes over review with a
ischemic stroke: A literature several decades, its efficacy in the meta-analysis is
review. Surgical neurology treatment of ischemic stroke is included (study
international 7: S381-6 debatable. Several trials have 1).
proven its safety and feasibility;
however, more robust,
randomised clinical trials with
large volumes of patients are
needed to fully establish its utility
in the clinical setting.

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 41 of 42
 IP 1115 [IPGXXX]

Wang H, Olivero W, Lanzino G RCT The helmet delivers initial rapid Small RCT with
et al. (2004) Rapid and selective n=14 (8 vs 6) and selective brain cooling and mixed
cerebral hypothermia achieved maintains a significant indications
using a cooling helmet. Journal temperature gradient between the (severe stroke
of neurosurgery 100: 272-7 core and brain temperatures or head injury).
throughout the hypothermic period
to provide sufficient regional
hypothermia yet minimise
systemic complications.

IP overview: therapeutic hypothermia for acute ischaemic stroke

© NICE 2019. All rights reserved. Subject to Notice of rights.

Page 42 of 42