Hypothermia for Traumatic Brain Injury in

ChildrenA Phase II Randomized Controlled Trial

John Beca, FCICM, MBChB1; Brent McSharry, FCICM, MBBS1; Simon Erickson, FCICM, MBBS2;
Michael Yung, FCICM, MBBS3; Andreas Schibler, FCICM, MD4; Anthony Slater, FCICM, MBBS5;
Barry Wilkins, FCICM, MD6; Ash Singhal, FRCSC, MD7; Gary Williams, FCICM, MBBS8;
Claire Sherring, BNurs1; Warwick Butt, FCICM, MBBS9; for the Pediatric Study Group of the
Australia and New Zealand Intensive Care Society Clinical Trials Group

Department of Pediatric Intensive Care at Starship Childrens Hospital,

Auckland, New Zealand.
2
Department of Pediatric Intensive Care, Princess Margaret Hospital,
Perth, Australia.
3
Department of Pediatric Intensive Care, Women and Childrens Hospital,
Adelaide, Australia.
4
Department of Pediatric Intensive Care, Mater Childrens Hospital, Brisbane, Australia.
5
Department of Pediatric Intensive Care, Royal Childrens Hospital, Brisbane, Australia.
6
Department of Pediatric Intensive Care, Childrens Hospital at Westmead, Sydney, Australia.
7
Department of Pediatric Neurosurgery, British Columbia Childrens Hospital, Vancouver, BC, Canada.
8
Department of Pediatric Intensive Care, Sydney Childrens Hospital, Sydney, Australia.
9
Department of Pediatric Intensive Care, Royal Childrens Hospital, Melbourne, Australia.
Supported, in part, by grants from the Victorian Transport Accident Commission, the Neurotrauma Research Programme of the West Australia
Institute for Medical Research, and the Intensive Care Foundation of Australia and New Zealand.
Dr. Erickson served as a board member for the Australia and New Zealand Intensive Care Society (ANZICS) (no financial remuneration apart
from travel to board meetings or other), consulted for Ikaria, is employed
by Princess Margaret Hospital (primary employment), provided expert
opinion for medicolegal cases (various legal entities), and received grant
support from Hospira and the Princess Margaret Hospital Foundation
(unrestricted study grants for other studies). Dr. Slater received support
for travel from the Malaysian Intensive Care Society (travel and accommodation to attend the Malaysian Intensive Care Society Annual Scientific Meeting), APLS Australia (travel and accommodation for instructing
on courses), and the ANZICS (travel and accommodation for ANZICS
core business). His institution received grant support from the Royal
Childrens Hospital Foundation (grants for hospital equipment and for
research projects and innovation grants). Dr. Wilkins is employed by
Childrens Hospital at Westmead. Dr. Singhal is employed by the University of British Columbia and the British Columbia Childrens Hospital
and received grant support from Brain Tumor Foundation of Canada, BC
Childrens Hospital Foundation, and the Rare Disease Foundation (unrelated research). Dr. Williams is employed by the Sydney Childrens Hospital Randwick, NSW, Australia, and has stock (Shares Telstra, Qantas,
1

Copyright 2015 by the Society of Critical Care Medicine and Wolters

Kluwer Health, Inc. All Rights Reserved.
DOI: 10.1097/CCM.0000000000000947

Critical Care Medicine

Commonwealth Bankpublic companies on ASX unrelated in any way to

this work). The remaining authors have disclosed that they do not have
any potential conflicts of interest.
For information regarding this article, E-mail: johnbeca@adhb.govt.nz

Objectives: To perform a pilot study to assess the feasibility of

performing a phase III trial of therapeutic hypothermia started
early and continued for at least 72 hours in children with severe
traumatic brain injury.
Design: Multicenter prospective randomized controlled phase II trial.
Setting: All eight of the PICUs in Australia and New Zealand and
one in Canada.
Patients: Children 115 years old with severe traumatic brain
injury and who could be randomized within 6 hours of injury.
Interventions: The control group had strict normothermia to a
temperature of 3637C for 72 hours. The intervention group
had therapeutic hypothermia to a temperature of 3233C for
72 hours followed by slow rewarming at a rate compatible with
maintaining intracranial pressure and cerebral perfusion pressure.
Measurements and Main Results: Of 764 children admitted to PICU
with traumatic brain injury, 92 (12%) were eligible and 55 (7.2%) were
recruited. There were five major protocol violations (9%): three related
to recruitment and consent processes and two to incorrect temperature management. Rewarming took a median of 21.5 hours (1635hr)
and was performed without compromise in the cerebral perfusion
pressure. There was no increase in any complications, including infections, bleeding, and arrhythmias. There was no difference in outcomes
12 months after injury; in the therapeutic hypothermia group, four
(17%) had a bad outcome (pediatric cerebral performance category,
46) and three (13%) died, whereas in the normothermia group, three
(12%) had a bad outcome and one (4%) died.
Conclusions: Early therapeutic hypothermia in children with severe
traumatic brain injury does not improve outcome and should not
be used outside a clinical trial. Recruitment rates were lower and
outcomes were better than expected. Conventional randomized
controlled trials in children with severe traumatic brain injury are
unlikely to be feasible. A large international trials group and alterwww.ccmjournal.org

Copyright by 2015 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.

Beca et al
native approaches to trial design will be required to further inform
practice. (Crit Care Med 2015; XX:0000)
Key Words: children; pediatric intensive care; randomized
controlled trial; therapeutic hypothermia; traumatic brain injury

raumatic brain injury (TBI) is one of the leading

causes of death and disability in childhood and is a
major global public health problem. TBI is also one of
the most common reasons for children to require mechanical
ventilation in intensive care (13). Children with severe TBI
(Glasgow Coma Score [GCS] < 9) commonly have severe and
persisting neurocognitive deficits; mean intelligent quotient
scores are on average 1826 points (12 sd) less than a control group of children (4). The majority will have significant
education difficulties and be unable to sustain permanent
employment as adults. The personal, family, social, and financial consequences of TBI are profound. In the United States,
the annual burden of TBI is estimated to be $56 billion (2).
The goal of medical therapy after severe TBI is to minimize
secondary injury. Therapeutic hypothermia (TH) has beneficial effects on all aspects of secondary brain injury in animal
models (5, 6). A recent systematic review of 18 randomized
controlled trials (RCTs) showed a reduced relative risk of death
and poor neurological outcome; however, when only highquality trials were included, there was no benefit (7).
There are several potential reasons for the failure to find a
benefit of early TH. First, TBI is a heterogeneous condition,
and TH may have different effects in different types of brain
injury. Second, TH may not have been started early enough.
Third, TH may not have been continued long enough (8).
Most RCTs have begun rewarming after a predetermined
period of time, usually 48 hours or less, regardless of intracranial pressure (ICP). Peak ICP commonly occurs 48 hours
or more after injury, and rewarming may be associated with
rebound intracranial hypertension. Several trials found hypotension and reduced cerebral perfusion pressure (CPP) during
rewarming in the TH group, including one where there was
possibly a worse outcome (911).
Our hypothesis was that TH begun early, continued for at
least 72 hours and then with the rate of rewarming guided by
ICP and CPP, will improve outcome in children with severe
TBI. The objective of this pilot study was to assess the feasibility of a phase III RCT and test the protocol, including adherence, safety, and outcomes.

Patients
Children were eligible for inclusion if they were more than 1
year old and less than 16 years old, were mechanically ventilated,
and had a GCS less than 9 and an abnormal CT scan (intracranial hemorrhage or contusion, cerebral edema, or diffuse axonal injury). Children were excluded if they were not able to be
randomized within 6 hours after injury or had any of the following: 1)a penetrating brain injury, 2) fixed dilated pupils and
a GCS=3, 3) proven cervical spinal cord injury, 4) more than
mild developmental disability prior to injury, 5) an acute epidural
hematoma evacuated and were expected to recover rapidly, 6) a
posttraumatic clinical seizure with a normal CT scan, 7) refractory shock (defined as a mean arterial pressure [MAP] less than
2 sds below mean for age despite > 80mL/kg of IV resuscitation
fluid), or 8) were suspected of having a nonaccidental injury.
Written informed consent was required from a parent or guardian before enrollment in five sites. Four sites had ethics committee approval to enroll children if there was no parent or guardian
available with delayed consent obtained when they were available.
Study Procedures
All patients were managed according to a standard protocol,
including a standard algorithm, with preferred order of therapies, for the management of intracranial hypertension (Fig.1).
This was developed and agreed to by all investigators and was
consistent with published guidelines (12). ICP was monitored
in all children, with the preferred method being an intraventricular catheter. Goals were an ICP (mm Hg) less than 20 and
a CPP (mm Hg) 1) greater than 4050 (age < 2 yr); 2) greater
than 50 (age < 11 yr); and 3) greater than 60 (age > 10 yr).

METHODS
The Hypothermia in Traumatic Brain Injury (HiTBIC) pilot
RCT (NCT00282269) was a multicenter trial of early and
prolonged TH in children with severe TBI conducted in all
eight of the PICUs in Australia and New Zealand and one
PICU in Canada. The Pediatric Study Group of the Australia and New Zealand Intensive Care Society Clinical Trials
Group designed the trial and the ethics committee at each
study site approved it.
2

www.ccmjournal.org

Figure 1. Management protocol showing standard traumatic brain injury

management and tiered management of intracranial hypertension. CSF =
cerebrospinal fluid, LD = loading dose, MD = maintenance dose.
XXX 2015 Volume XX Number XXX

Copyright by 2015 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.

Neurologic Critical Care

Temperature Management
Temperature was measured with a probe in the distal third of
the esophagus. Temperature control was achieved with a servocontrolled cooling blanket. Patients randomized to normothermia were maintained at 3637C while those randomized
to hypothermia were cooled to 3233C. Iced IV bolus fluids
(if indicated clinically), crushed ice to exposed surfaces, and a
second cooling blanket above the child could also be used during induction. Hypothermia was maintained for a minimum of
72 hours. The rate of rewarming was to be no faster than 0.5C
every 3 hours but was guided primarily by physiology rather
than time. Specifically, hypotension was managed aggressively
with IV fluids and vasoconstrictors, and intracranial hypertension was managed with the standard algorithm, but rewarming
was also slowed or stopped until these variables were controlled.
Screening and Randomization
All children admitted to a study site PICU were screened, and
basic data including reasons for nonenrollment were entered

in to a screening log. Randomization was stratified by center

and GCS (34 or 58). Randomization was in variable block
sizes and was performed by telephone to the coordinating center where sequentially numbered opaque envelopes were used.
Outcome Measures and Analysis
The prospectively defined primary outcomes were 1) pediatric
cerebral performance category (PCPC) (13) 12 months after
injury dichotomized to good (13) and poor (46) outcome; 2)
eligibility and recruitment rates; 3) protocol violations; and 4)
major adverse events. Secondary outcomes were 1) ICP and CPP
during the first 5 days and treatments required; 2) duration of
mechanical ventilation and PICU and hospital lengths of stay;
and 3) adverse events including infectious complications (ventilator-associated pneumonia, bacteremia, cerebral abscess, meningitis, and urinary tract infection), bleeding, pancreatitis, acute
respiratory distress syndrome, and arrhythmias (excluding sinus
bradycardia). The study neuropsychologist performed all PCPC
scores for all sites. Low MAP was defined as any MAP less than
40 + (1.5 age) and low CPP
as less than 50 for those older
than 10 years and less than 40
for those who were 10 years old
or younger. The pilot planned
to recruit 50 children who completed the full protocol.
Continuous data are presented as median with interquartile range. Comparisons
between groups were done with
the Wilcoxon rank-sum test or
Fisher exact test as appropriate.
Multivariable logistic regression
was used to compare the primary outcome between groups
and to investigate specific a
priori defined covariables. The
net effect of hypothermia on
longitudinal physiological data
was analyzed using generalized
estimation equation assuming
a normal distribution with an
exchangeable correlation. Data
analysis was performed using
Stata v12 (StataCorp, College
Station, TX).

RESULTS

Figure 2. Consort diagram showing overview of screened and randomized patients. ITT = intention to treat,
PP = per protocol.

Enrollment occurred from

November 2006 to May 2010.
The trial was suspended in June
2008 following publication
of a phase III trial that found
that early TH might increase
mortality (10). The suspension was to allow independent

Critical Care Medicine

www.ccmjournal.org

Copyright by 2015 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.

50
100
Head Injuries Screened

No. Fitting Criteria

10
15
20

25

150

Beca et al

E
Unit

Consent Not Attempted

Protocol Violation
Head Injuries Screened

Consent Refused
Per Protocol

Figure 3. Site performance. Numbers of children screened and randomized

and reasons for failure to randomize in eligible children for each site.

review by the Data Safety and Monitoring Committee, which

recommended continuation of the study. The duration of the
suspension was 26 months and varied between sites due to differences in time required for review by local ethics committees.
Seven hundred sixty-four children with TBI were screened
of whom 92 (12%) were eligible and 55 (7.2%) were randomized (see consort diagram; Fig.2). Of the 37 (40%) who
were eligible but not randomized, 15 parents (16%) declined
enrollment and in the remaining 22 (24%), consent was
Table 1.

not attempted. In 12 of these 22, consent was not attempted

because the study was suspended at the time. Eligibility and
randomization rates ranged from 0.9% to 19.7% and 0% to
18.3%, respectively, between sites. Sites were compared to the
coordinating center site in terms of eligibility and randomization rates and rates of consent not being attempted and consent
being refused (Fig.3). One center had a significantly lower rate
of eligibility and randomization (10% vs 1%, p=0.003 and
9% vs 0%, p=0.001, respectively). A separate site had a significantly higher rate of consent refusal (0% vs 58%; p=0.01).
There were no other significant differences between sites.
There were five major protocol violations (9%), four of which
occurred in children randomized to hypothermia. Three were
related to remote or deferred consent and led to the child being
withdrawn from the study; two children did not meet severity criteria on review after arrival at the trial site and in one the
parents withdrew consent after deferred consent had been used.
The other two major protocol violations were associated with
incorrect temperature management: in one, a child randomized
to hypothermia was rewarmed very early as the treating neurosurgical team believed that the injury was not severe and in the
other, a child randomized to normothermia received hypothermia as the initial second-tier therapy for raised ICP at the insistence of the neurosurgical team. These five protocol violations
resulted in patients not getting the prescribed temperature management. So of the 55 recruited, 50 were managed per protocol.
Baseline characteristics are shown in Table 1. Children in
the hypothermia group took a median of 0.9 hours longer to

Baseline Characteristics by Treatment Group of Patients Treated Per Protocol

Variable

Normothermia (26)

Male (%)

16 (62)

Age, yr

9.5 (5.213.8)

Age < 7 yr (%)

9 (35)

Hypothermia (24)

11 (46)
11.0 (6.914.2)

0.40
0.48

6 (25)

0.55

31.5 (2453)

0.47

Weight (kg)

30 (2050)

Injury to ED (hr)

0.8 (0.61.3)

0.7 (0.51.7)

0.68

Injury to study site (hr)

3.6 (2.44.8)

4.5(3.57.5)

0.02

Injury to randomization (hr)

5.0 (4.25.8)

5.3 (4.46.0)

0.34

GCS at scene

5 (47)

5 (37)

0.59

GCS in ED

4.5 (37)

5.5 (3.57)

0.45

Pediatric trauma score

4.5 (46)

5 (3.56)

0.63

Pediatric Index of Mortality 2 risk of death

0.048 (0.0420.058)

0.044 (0.0400.050)

0.12

3 (12)

0 (0)

0.24

0 (0)

1 (4)

0.48

Transported from another hospital (%)

13 (50)

13 (54)

1.0

Direct admission to PICU (%)

10 (39)

12 (50)

0.57

Received muscle relaxants before randomization (%)

19 (73)

22 (92)

0.14

Hypotension (%)

Hypoxia (%)

ED = emergency department, GCS = Glasgow Coma Score.

a
As defined in Methods: Outcome Measures and Analysis.
Categorical data are shown as n (frequency) and continuous data as median (interquartile range).

www.ccmjournal.org

XXX 2015 Volume XX Number XXX

Copyright by 2015 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.

Neurologic Critical Care

arrive at a study site, but there was no significant difference in

the median time to randomization, which was approximately 5
hours. There were no other baseline differences.
Outcomes
Twelve months after injury, there was no difference in the primary
outcome. Three children in the normothermia group (12%)
and four in the hypothermia group (17%) had a bad outcome
(p=0.70). This included one (4%) in the normothermia group
and three (13%) in the hypothermia group who died. Table 2
shows primary and secondary outcomes for the two groups. The
distribution of PCPC scores for both groups is shown in Figure4. The rate of poor outcomes was not significantly different
to the control arm of the Canadian TH study (22% using PCPC;
p=0.41) (10), but was lower than the CoolKids study (42%
using Glasgow Outcome Scale; p = 0.009) (14).
None of the predefined covariables was significant in a
multivariable model (age < 7; p = 0.46; GCS in the emergency
department of 3 or 4, p=0.05; temperature at admission
< 35C; p=0.38). Of note, all but one of the children with a
bad outcome had a low GCS (34) at admission. Seventy-two
percent of children with a GCS of 34 and 97% with a GCS of
58 had a good outcome.
The median time from injury to target temperature in the
hypothermia group was 9.3 hours (8.010.9hr). The median
Table 2.

time from randomization to target temperature in the hypothermia group was 4.6 hours (3.05.8hr). In the normothermia group, there were 12 episodes in eight patients (30%)
where the temperature was more than 38C. The median duration of these temperature deviations was 1.75 hours (13hr).
Figure5 shows graphs of the differences in physiology
between the two groups. During the cooling phase, the net
effect of hypothermia was a drop in the heart rate of 23.4 bpm
(95% CI, 15.930.9; p < 0.001) and the ICP was 1.8mm Hg
lower (95% CI, 0.33.4; p=0.02). Hypothermia was maintained for a median of 93.5 hours (87104hr) that included a
median rewarming time of 21.5 hours (1635hr). There was
no significant difference between groups in MAP or CPP during cooling (p=0.44 and 0.77, respectively) or in MAP, ICP, or
CPP during rewarming (p=0.68, 0.59, and 0.07, respectively).
During rewarming in the hypothermia group, there were one
or more episodes of systemic hypotension in four children
(17%), intracranial hypertension in 13 children (54%), and
low CPP in three children (13%). One of the four children
who had hypotension during rewarming had a bad outcome.
During the same time period, in the normothermia group,
these events occurred in 0 (p=0.05), 15 (60%; p=0.78), and
three (12%; p=0.99) children, respectively. There was no difference in the number of patients requiring inotropes. For
those requiring inotropes, there was a nonsignificant trend

Primary and Secondary Outcomes When Analyzed per Protocol

Outcome Variable

Normothermia (26)

Hypothermia (24)

Bad outcome (%)

3 (12)

4 (17)

0.70

Died (%)

1 (4)

3 (13)

0.34

PICU length of stay (d) in survivors

11 (814)

12 (815)

0.87

Hospital length of stay (d) in survivors

40 (2760)

48 (2877)

0.70

Mechanical ventilation (d) in survivors

10 (613)

8 (714)

0.77

6 (25)

1.0

Infection (%)
Pneumonia

7 (27)

Septic shock

Urinary tract

2 (8)

1 (4)

1.0

Wound

3 (12)

3 (13)

1.0

Other

7 (27)

10 (42)

0.25

Bleeding (%)
Intracranial

1 (4)

1.0

Gastrointestinal

Other

1 (4)

Ventricular tachycardia/ventricular fibrillation

Other

1 (4)

0.48

Acute respiratory distress syndrome (%)

2 (8)

0.23

0.48

Arrhythmia (%)

Data are shown as n (frequency).

Critical Care Medicine

www.ccmjournal.org

Copyright by 2015 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.

Beca et al

rate of major protocol violations, primarily related to remote

randomization at nonstudy sites. Since this is a pilot study, it
was not powered to detect even a large treatment effect with
hypothermia. However, it is noteworthy that the outcomes
were equivalent in each group, both in terms of dichotomized outcomes and in the distribution of PCPC scores.
This study also supports that hypothermia, used as it was in
this study, may be safe. After 72 hours of hypothermia, slow
rewarming was performed without significant compromise
in CPP. The rates of adverse events were also equivalent
betweengroups.

Good Outcome

40

60

80

100

Poor Outcome

20

Cumulative Proportion

4
3
PCPCS at 12 months
Hypothermia

Normothermia

Good Outcome

40

60

80

100

Poor Outcome

20

Cumulative Proportion

4
3
PCPCS at 12 months
Hypothermia

Normothermia

Figure 4. Pediatric cerebral performance category score (PCPCS)

for children treated with hypothermia compared with normothermia. A,
Intention to treat (55 children). B, Per protocol (50 children).

toward higher total vasoactive inotrope score (15) in the hypothermic arm (Table3), and inotropes were required for longer
(38hr [3443 hr] vs 25.5hr [737.5 hr]; p=0.02).
There was also no difference in the use of adjuvant therapies to manage intracranial hypertension (Table3). There were
no differences between groups in the prevalence of any of the
adverse events recorded (Table2). These included measures
of infection, bleeding, and arrhythmias. Pancreatitis did not
occur in any patients. The duration of mechanical ventilation
and ICU and hospital lengths of stay were not different.

DISCUSSION
This pilot RCT of early and prolonged hypothermia for severe
TBI in children was designed to investigate the f easibility and
safety of performing a phase III study. We found that the randomization rate of 7.2% was much lower than expected and
that outcomes were much better than expected, with an overall
bad outcome rate of 14% and mortality of 8%. In terms of the
recruitment and treatment protocols, we found an important
6

www.ccmjournal.org

Feasibility
This study, together with other recent RCTs in severe TBI in children and adults, has important implications about the feasibility
of interventional studies. Our study had a randomization rate
of 7.2%. Two other studies of TH in children with TBI had randomization rates on 4.2% (14) and 15.6% (10), whereas recent
adult RCTs of hypothermia (16) and decompressive craniectomy (17) had rates of 4.7% and 4.4%, respectively. In addition,
we found a mortality of 8%, similar to 12% in the control arm
of the Canadian TH study (10), 15.6% in the CoolKids study
(14), and 9.3% in a survey of PICUs in England and Wales (18).
Using dichotomized PCPC values, we found that poor outcome
occurred in 14% of children. This was not significantly different to the control arm of the Canadian TH trial (22%), but was
significantly lower than the CoolKids study (42%). Even if the
bad outcome rate is 2535% and a RCT is looking for an absolute reduction of 10%, then at least 500700 children will be
required. If randomization rates are 715%, then 3,00010,000
children might need to be screened. We suggest that this is not
feasible and that either alternative approaches to design and
consent of RCTs or alternative trial types are required.
We found significant variability between sites in randomization rates, suggesting the potential for process improvement.
These differences might be due to a combination of logistics
(e.g., lack of available research personnel), varying degrees of
commitment to the trial, variability between local ethics committees and sites in attitudes to deferred consent, and problems
with obtaining consent.
Obtaining consent from parents for a time-critical intervention in the setting of a sudden, accidental, and life-threatening
event is particularly fraught. Lack of parental competence, inadequate information, and inadequate comprehension is common
in parents giving consent in emergency situations. Most doctors also have concerns about parental competence because of
their vulnerability, emotional state, and lack of understanding
(19). Emergency care research often meets criteria for waived or
deferred consent (20), and community consultation has been
shown to be potentially supportive of this approach (21).
In our study, four of the nine sites had approval of deferred
consent. Eight children were randomized using this process,
and there were no problems with the consent process in any
of these eight participants. At one other site (where deferred
consent was approved but contentious within that PICU), consent was withdrawn by the parents of one child after arrival at
XXX 2015 Volume XX Number XXX

Copyright by 2015 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.

Neurologic Critical Care

90

CPP (mm Hg)

60
70
80
50

Heart Rate (beats per minute)

70 80 90 100 110 120

25

ICP (mm Hg)

10
15
20
5

Core Temperature (Celcius)

32
34
36
38

an ordinal score is collapsed in

to a dichotomized outcome,
important information is lost
and priority is given to one
specific transition (that from
severe to moderate disability).
This means that patients with
very good or bad prognoses
may contribute little to the
trial. It has been estimated that
0
20
40
60
80
100
0
20
40
60
80
100
Hours after randomization
Hours after randomization
this may lead to a loss of statistical power of 40% or more
(23). This may be particularly
true in severe TBI where there
is considerable heterogeneity of
lesion types and prognoses. By
using broad inclusion criteria
and covariate-adjusted ordinal
analysis (either proportional
odds regression or a sliding
0
20
40
60
80
100
0
20
40
60
80
100
dichotomy), statistical power
Hours after randomization
Hours after randomization
can be greatly enhanced and
might allow for more feasible
sample sizes (23, 24). However,
Figure 5. Physiology during hypothermia and rewarming compared with normothermia. Data points are the mean
value at each time point. The I bars are 95% CIs. CPP = cerebral perfusion pressure, ICP = intracranial pressure.
such an approach requires
robust risk modeling to be perthe site and the option of deferred consent was subsequently
formed on large high-quality datasets as has been done in adult
withdrawn at that site. We believe that there is an urgent need TBI by the International Mission for Prognosis and Analysis of
for consensus and standardization about processes for deferred
Clinical Trials in TBI investigators (25). We suggest that a similar
consent between researchers and ethics committees (22).
process is required in pediatric TBI.
An alternative approach to trial design is not to dichotoAlternative trial types, such as adaptive clinical trials (26)
mize but to use ordinal analysis of the outcome scores. When
and comparative effectiveness research, also offer potential for

Comparison of Adjuvant Therapies Used for Intracranial Hypertension and

Vasoactive Drugs Used
Table 3.
Variable

Normothermia (26) (%)

Hypothermia (24) (%)

Cerebrospinal fluid drainage

16 (62)

12 (50)

0.57

Hypertonic saline used

27 (100)

23 (96)

0.48

20% mannitol used

11 (42)

15 (63)

0.17

Phenobarbitone given

5 (19)

3 (13)

0.70

Thiopentone bolus given

2 (8)

4 (17)

0.41

Thiopentone infusion used

5 (19)

8 (33)

0.34

Decompressive craniectomy

7 (27)

5 (21)

0.75

> 3 therapies used for intracranial hypertension

5 (19)

8 (33)

0.34

Vasoactive drugs required

12 (46)

11 (46)

0.98

First 72 hr

12 (46)

11 (46)

1.0

After 72 hr

7 (28)

11 (46)

0.24

Vasoactive inotrope score (range)

22 (840)

21 (1438)

0.98

First 72 hr

36 (2350)

27 (1847)

0.42

After 72 hr

8 (716)

18 (528)

0.22

Critical Care Medicine

www.ccmjournal.org

Copyright by 2015 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.

Beca et al

TBI research. Comparative effectiveness research aims to measure the benefits and risks of interventions or systems of care
in broad populations in standard settings (27). This approach
is being used in pediatric TBI in the Approaches and Decisions
in Acute Pediatric TBI study (http://www.adapttrial.org).
The Safety of TH
We found no difference in complication rates between the two
groups. Although the study was not adequately powered for these
outcomes, this is in keeping with the findings of the other multicenter RCTs of TH in children and adults (9, 10, 14, 16, 2830). We
found that the management of rewarming is challenging; despite a
protocol emphasizing the need for strict medical titration of therapy, hypotension occurred more commonly in the hypothermia
group. However, these episodes were not associated with compromise in CPP, that is, they were all isolated episodes of hypotension
in children with normal or low ICP at that time. Using predefined
thresholds, there was no difference between groups in episodes of
intracranial hypertension or low CPP. Our protocol was therefore
successful in avoiding clinical compromise in CPP. There was a
trend to a slightly lower CPP in the hypothermia group. It is likely
that this reflects titration of therapy to the thresholds in the study
algorithm during the inherently less stable period of rewarming.
The Efficacy of TH
We found equivalent PCPC scores at 12 months after injury in
the two groups. There were no other significant differences in
hemodynamic support or therapy for intracranial hypertension
between groups. Since this pilot study supported the safety of
TH and also found that recruitment rates suggested the need for
a very large international trials group for a phase III study to be
feasible, we joined the CoolKids trial. The patients in this phase
II trial were not included in the CoolKids trial. Unfortunately,
shortly after joining, the CoolKids study was stopped for futility (14). There are now two adult and two pediatric high-quality
multicenter RCTs that show no benefit to early TH in TBI (9, 10,
14, 16). Possible reasons for the failure of early TH in TBI are that
it may not have been started early enough or continued for long
enough. The median time to target temperature was 9 hours both
in our patients and the CoolKids trial. Animal studies suggest that
that TH may need to occur within 4 hours of injury and is ineffective after around 8 hours (31). Human trials suggest that TH may
need to be continued for at least 48 hours (8). However, the most
recent adult study started TH within 22.5 hours of injury and
continued it for 48 hours (16), and it was also stopped for futility. It is also possible that TH is not efficacious in TBI because of
the heterogeneous nature of injury or that it is only beneficial in
specific subsets of patients. In the most recent adult study (16), the
subgroup with diffuse injury may have done worse and those with
surgically evacuated hematomas may have done better with TH.
Fever is associated with worse outcome in both hypoxic ischemic encephalopathy and TBI. One difference between HiTBIC
and both the Canadian and CoolKids trials was in temperature
management of the control group; our goal temperature was
3637C, whereas the goal of other two trials was 36.537.5C.
Despite these goals, two thirds of patients in the CoolKids trial
8

www.ccmjournal.org

exceeded this temperature, and approximately one third of our

patients in the control group had a temperature over 38C. A
recent large RCT of TH in adults after cardiac arrest found no
outcome benefit with a temperature targeted to 33C compared
with 36C, suggesting that any potential benefit from TH might be
predominantly explained by avoidance of fever (32). Prevention
of fever is poor in children with brain injury, and more attention
could be directed to strategies to achieve this (33).

CONCLUSIONS
This trial adds further support to the existing literature that
early TH in children with severe TBI does not improve outcome and should not be used outside a clinical trial. With our
trial protocol, 12% of children admitted to intensive care with
TBI were eligible for inclusion and 7.2% were recruited. Overall outcomes were better than expected; 8% of children died
and 14% had a bad outcome. Conventional RCTs in children
with severe TBI are unlikely to be feasible. A large international
trials group and alternative approaches to trial design will be
required to further inform practice.

ACKNOWLEDGMENTS
We acknowledge the study site coordinators (Carmel Delzoppo,
Pania Falconer, Jacqui Jauncey-Cooke, Sue Kendall, Gordon
Krahn, Georgia Letton, Flavia Li, and Janelle Young) for their
assistance with study coordination and data collection. We also
acknowledge the pediatric intensive care medical and nursing
staff and neurosurgical staff at all of the study sites for their
support of our study. We thank Peter Reed of the Starship Childrens Research Centre for statistical advice and review.

REFERENCES

1. Moreau JF, Fink EL, Hartman ME, et al: Hospitalizations of children

with neurologic disorders in the United States. Pediatr Crit Care Med
2013; 14:801810
2. Schneier AJ, Shields BJ, Hostetler SG, et al: Incidence of pediatric
traumatic brain injury and associated hospital resource utilization in
the United States. Pediatrics 2006; 118:483492
3. Stanley RM, Bonsu BK, Zhao W, et al: US estimates of hospitalized
children with severe traumatic brain injury: Implications for clinical trials. Pediatrics 2012; 129:e24e30
4. Anderson V, Godfrey C, Rosenfeld JV, et al: Predictors of cognitive
function and recovery 10 years after traumatic brain injury in young
children. Pediatrics 2012; 129:e254e261
5. Clark RS, Kochanek PM, Marion DW, et al: Mild posttraumatic hypothermia reduces mortality after severe controlled cortical impact in
rats. J Cereb Blood Flow Metab 1996; 16:253261
6. Clifton GL, Jiang JY, Lyeth BG, et al: Marked protection by moderate
hypothermia after experimental traumatic brain injury. J Cereb Blood
Flow Metab 1991; 11:114121
7. Georgiou AP, Manara AR: Role of therapeutic hypothermia in improving outcome after traumatic brain injury: A systematic review. Br J
Anaesth 2013; 110:357367
8. McIntyre LA, Fergusson DA, Hbert PC, et al: Prolonged therapeutic
hypothermia after traumatic brain injury in adults: A systematic review.
JAMA 2003; 289:29922999
9. Clifton GL, Miller ER, Choi SC, et al: Lack of effect of induction of hypothermia after acute brain injury. N Engl J Med 2001;
344:556563
10. Hutchison JS, Ward RE, Lacroix J, et al; Hypothermia Pediatric Head
Injury Trial Investigators and the Canadian Critical Care Trials Group:
XXX 2015 Volume XX Number XXX

Copyright by 2015 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.

Neurologic Critical Care

Hypothermia therapy after traumatic brain injury in children. N Engl J
Med 2008; 358:24472456
11. Marion DW, Penrod LE, Kelsey SF, et al: Treatment of traumatic brain
injury with moderate hypothermia. N Engl J Med 1997; 336:540546
12. Kochanek PM, Carney N, Adelson PD, et al; American Academy of
Pediatrics-Section on Neurological Surgery; American Association of
Neurological Surgeons/Congress of Neurological Surgeons; Child
Neurology Society; European Society of Pediatric and Neonatal
Intensive Care; Neurocritical Care Society; Pediatric Neurocritical
Care Research Group; Society of Critical Care Medicine; Paediatric
Intensive Care Society UK; Society for Neuroscience in Anesthesiology
and Critical Care; World Federation of Pediatric Intensive and Critical
Care Societies: Guidelines for the acute medical management of
severe traumatic brain injury in infants, children, and adolescents
Second edition. Pediatr Crit Care Med 2012; 13(Suppl 1):S1S82
13. Fiser DH: Assessing the outcome of pediatric intensive care. J Pediatr
1992; 121:6874
14. Adelson PD, Wisniewski SR, Beca J, et al; Paediatric Traumatic Brain
Injury Consortium: Comparison of hypothermia and normothermia
after severe traumatic brain injury in children (Cool Kids): A phase 3,
randomised controlled trial. Lancet Neurol 2013; 12:546553
15. Gaies MG, Gurney JG, Yen AH, et al: Vasoactive-inotropic score as
a predictor of morbidity and mortality in infants after cardiopulmonary
bypass. Pediatr Crit Care Med 2010; 11:234238
16. Clifton GL, Valadka A, Zygun D, et al: Very early hypothermia induction
in patients with severe brain injury (the National Acute Brain Injury Study:
Hypothermia II): A randomised trial. Lancet Neurol 2011; 10:131139
17. Cooper DJ, Rosenfeld JV, Murray L, et al; DECRA Trial Investigators;
Australian and New Zealand Intensive Care Society Clinical Trials
Group: Decompressive craniectomy in diffuse traumatic brain injury.
N Engl J Med 2011; 364:14931502
18. Tasker RC, Fleming TJ, Young AE, et al: Severe head injury in children:
Intensive care unit activity and mortality in England and Wales. Br J
Neurosurg 2011; 25:6877
19. Mason SA, Allmark PJ: Obtaining informed consent to neonatal randomised controlled trials: Interviews with parents and clinicians in the
Euricon study. Lancet 2000; 356:20452051
20. Truog RD, Robinson W, Randolph A, et al: Is informed consent always
necessary for randomized, controlled trials? N Engl J Med 1999;
340:804807
21. Morris MC, Nadkarni VM, Ward FR, et al: Exception from informed
consent for pediatric resuscitation research: Community consultation
for a trial of brain cooling after in-hospital cardiac arrest. Pediatrics
2004; 114:776781
22. Roberts I, Prieto-Merino D, Shakur H, et al: Effect of consent rituals on
mortality in emergency care research. Lancet 2011; 377:10711072

Critical Care Medicine

23. Maas AI, Steyerberg EW, Marmarou A, et al: IMPACT recommendations for improving the design and analysis of clinical trials in moderate
to severe traumatic brain injury. Neurotherapeutics 2010; 7:127134
24. Roozenbeek B, Lingsma HF, Perel P, et al; IMPACT (International
Mission on Prognosis and Clinical Trial Design in Traumatic Brain
Injury) Study Group; CRASH (Corticosteroid Randomisation After
Significant Head Injury) Trial Collaborators: The added value of ordinal analysis in clinical trials: An example in traumatic brain injury. Crit
Care 2011; 15:R127
25. Roozenbeek B, Lingsma HF, Lecky FE, et al; International Mission on
Prognosis Analysis of Clinical Trials in Traumatic Brain Injury (IMPACT)
Study Group; Corticosteroid Randomisation After Significant Head
Injury (CRASH) Trial Collaborators; Trauma Audit and Research
Network (TARN): Prediction of outcome after moderate and severe
traumatic brain injury: External validation of the International Mission
on Prognosis and Analysis of Clinical Trials (IMPACT) and Corticoid
Randomisation After Significant Head injury (CRASH) prognostic
models. Crit Care Med 2012; 40:16091617
26. Meurer WJ, Lewis RJ, Tagle D, et al: An overview of the adaptive
designs accelerating promising trials into treatments (ADAPT-IT) project. Ann Emerg Med 2012; 60:451457
27. Maas AI, Menon DK, Lingsma HF, et al: Re-orientation of clinical
research in traumatic brain injury: Report of an international workshop on comparative effectiveness research. J Neurotrauma 2012;
29:3246
28. Adelson PD, Ragheb J, Kanev P, et al: Phase II clinical trial of moderate hypothermia after severe traumatic brain injury in children.
Neurosurgery 2005; 56:740754
29. Gluckman PD, Wyatt JS, Azzopardi D, et al: Selective head cooling with mild systemic hypothermia after neonatal encephalopathy:
Multicentre randomised trial. Lancet 2005; 365:663670
30. Shankaran S, Laptook AR, Ehrenkranz RA, et al; National Institute of
Child Health and Human Development Neonatal Research Network:
Whole-body hypothermia for neonates with hypoxic-ischemic encephalopathy. N Engl J Med 2005; 353:15741584
31. Gunn AJ, Gunn TR, de Haan HH, et al: Dramatic neuronal rescue with
prolonged selective head cooling after ischemia in fetal lambs. J Clin
Invest 1997; 9:248256
32. Nielsen N, Wetterslev J, Cronberg T, et al; TTM Trial Investigators:
Targeted temperature management at 33C versus 36C after cardiac arrest. N Engl J Med 2013; 369:21972206
33. Bembea MM, Nadkarni VM, Diener-West M, et al; American Heart
Association National Registry of Cardiopulmonary Resuscitation
Investigators: Temperature patterns in the early postresuscitation
period after pediatric inhospital cardiac arrest. Pediatr Crit Care Med
2010; 11:723730

www.ccmjournal.org

Copyright by 2015 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.