Therapeutic hypothermia after out-of-hospital cardiac arrest in children

In children who suffer out of hospital cardiac arrest, targeted hypothermia at 33.0℃ confers no benefit when compared to targeted normothermia at 36.8℃. Level of evidence: 2B (RCT with wide CIs)

Appraised by: Andrew Claxton

Citation: Moler FW, Silverstein FS, Holubkov R and the THAPCA Trial Investigators. Therapeutic hypothermia after out-of-hospital cardiac arrest in children. N Eng J Med 2015; 372: 1898–1908.

Lead author: Frank W Moler; [e-mail].

Three-part clinical question

Patient: Children older than 48 h and younger than 18 years old who had an out-of-hospital cardiac arrest with at least 2 min of chest compressions and requiring mechanical ventilation after return of spontaneous circulation (ROSC).

Intervention: Targeted temperature management at 33.0℃ vs 36.8℃ within 6 h of ROSC for 48 h. Rewarmed to 36.8℃ over 16 h.

Outcome: Primary: Survival at 12 months with a Vineland Adaptive Behaviour Scales, second edition (VABS-II) score of 70 or higher. Secondary: Survival at 12 months and change in neurobehavioural function based on pre-cardiac arrest and 12 month VABS-II.

The study

Multi-centre, prospective, single-blinded (blinded centralised outcome assessors, statisticians and authors), randomised controlled trial (RCT) with intention-to-treat.

The study patients

All children admitted to 38 paediatric intensive care units (PICUs) across Canada and the United States of America between September 2009 and December 2012.

Eligibility: Children aged between 48 h and 18 years of age who have had an out-of-hospital cardiac arrest.

Inclusion criteria: At least 2 min of cardiopulmonary resuscitation (CPR) following cardiac arrest and requiring mechanical ventilation following ROSC.

Exclusion criteria: Not randomised within 6 h of cardiac arrest. Motor score of 5 or 6 on Glasgow Coma Scale (GCS). Decision made to withhold aggressive treatment. Cardiac arrest associated with major trauma.

Study method

Children eligible for inclusion were randomly assigned to one of the two targeted temperature management groups in a 1:1 ratio with use of permuted blocks stratified according to clinical centre and age at entry (<2 years, 2 to <12 years, and >12 years). Those randomised to the therapeutic hypothermia group were pharmacologically paralysed and core body temperature was rapidly reduced using a Blanketrol III temperature management unit (Cincinnati Sub-Zero) applied anteriorly and posteriorly. Within 24 h of randomisation a parent or guardian of the child was asked to complete a standardised caregiver questionnaire, from which the baseline VABS-II score could be calculated.

The hypothermic period of 48 h commenced at randomisation and sedation for this period was mandated in both groups. After 48 h, the children in the therapeutic hypothermia group were gradually warmed to 36.8℃ over at least 16 h. This temperature was then actively controlled at 36.8℃ for the remainder of the 120-h intervention period. Patients in the control group were actively maintained at 36.8℃ (36.8–37.5) for the entire 120-h study period. Core temperature was measured using two of either oesophageal, rectal or bladder probes. Clinical teams were free to determine all other aspects of care.

Survivors were followed up at 12 months by telephone from a centralised location and VABS-II data were collected, where an outcome of 70 or greater (on a scale of 20 to 160) was defined as a good outcome.

Outcome	Time to outcome	CER	EER	RRR	ARR	NNT
Alive with VABS II ≥70	1 year	0.123	0.196	59%	7.3%	ns
	95% Confidence intervals:			128% to 9%	15.7% to 1.1%
Alive	1 year	0.287	0.377	31%	9.0%	ns
	95% Confidence intervals:			69% to 6%	19.8% to 1.8%

Key: CER: control event rate, EER: experimental event rate, RRR: relative risk reduction, ARR: absolute risk reduction, NNT: number needed to treat.

The evidence:

EBM questions:

Do the study methods accurately allow for testing of the hypothesis? Yes. This was a multi-centre trial across two countries. Patient numbers required for analysis were calculated prior to trial commencement. Randomisation was via a telephone number based system in permuted blocks and stratified according to age on admission and site. The study was single-blinded as treating clinicians would need to know the target temperature of each patient. However, centralised assessors, statisticians and the authors were blinded.

Do the statistical tests correctly test the results to allow differentiation of statistically significant results? Yes.

Are the conclusions valid in light of the results? Yes. However, the primary outcome modified intention-to-treat analysis was of 88% of children randomised; 12% were lost due to poor baseline function or incomplete data. As the primary outcome rate was 12% and 20%, this may have significantly altered the outcome.

Did any results get omitted and why? Yes. Of the 295 patients that underwent randomisation, 4 patients were withdrawn, 3 from the hypothermic group who did not receive therapeutic hypothermia treatment, and 1 from the normothermic group who was accidentally given therapeutic hypothermic treatment. Overall, 260 patients in the modified intention-to-treat population were included in the primary analysis.

Did the authors suggest areas for further research? No. However, the authors did state that due to the observed confidence intervals for treatment differences, a potential clinical benefit could not be ruled out despite the lack of a significant difference between groups. The authors suggested that a larger trial may be able to detect or reject a smaller intervention benefit.

Did the authors make any recommendations based on the results and were they appropriate? No. The authors concluded that in children who survive out-of-hospital cardiac arrests, therapeutic hypothermia confers no significant benefit when compared to therapeutic normothermia with regard to 12-month survival and neurobehavioural outcome. The authors did, however, postulate that controlled normothermia may be of benefit in these patients as fever often develops after hypoxic-ischaemic brain injury.

Is this study relevant to my clinical practice? Yes. The children included were a heterogenous group who were treated in westernised PICUs and with demographics similar to children in the UK.

What level of evidence does this study represent? 2B. RCT with wide confidence intervals and a loss of 12% of patients in the modified intention-to-treat population included in the primary analysis.

What grade of recommendation can I make based on this result alone? B. This trial showed that there is no significant advantage to therapeutic hypothermia at 33℃ compared to therapeutic normothermia at 36.8℃ in comatose children surviving out-of-hospital cardiac arrest.

What grade of recommendation can I make when this study is considered alongside other available evidence? A. The conclusion that therapeutic hypothermia is not beneficial when compared to therapeutic normothermia is consistent with results from a recent large international multi-centre trial in adults. ¹ The similarity with this trial when compared to earlier trials^2,3 is that by having a normothermic arm, pyrexia is avoided.

Should I change my practice based upon these results? Possibly. This studied showed no improvement in survival or neurobehavioural outcome at 12 months when comparing therapeutic hypothermia with therapeutic normothermia. However, the authors make note of a small possible clinical benefit in which their study was underpowered to show. When compared to previous studies in adults, there is compelling evidence to support avoidance of hyperthermia in these patients.

Should I audit my own practice based upon these results? Yes.