THERAPY

Criteria abstracted from The Users' Guides to the Medical Literature series in JAMA

N Engl J Med 2008 358; 2447-2456..[abstract]

Review posted January 15, 2009

I. What is being studied?:

1. The study objective:

   To determine the effect of hypothermia therapy on neurologic outcome and mortality among children with severe traumatic brain injury.

2. The study design:

   An international multicenter randomized interventional trial stratified according to age and treatment center.

   Inclusion/Exclusion Criteria (clarifications in parentheses from related pre-trial(3)):

3. The patients included:

   • Age between 1 -17 years old

   • Traumatic Brain Injury

   • Glasgow Coma Scale of 8 or less (at the scene or in the Emergency Room)(3)

   • Head CT that showed acute brain injury (contusion, intracranial hemorrhage, diffuse axonal injury or cerebral edema)(3)

   • Need for mechanical ventilation

4. The patients excluded:

   • Those screened more than 8hr after the injury

   • Refractory shock (despite intravenous colloid and red cell transfusions exceeding 80ml/kg)(3)

   • Suspected brain death (defined as fixed and dilated pupils, Glasgow Coma Scale score = 3, and no evidence of brain function on neurological examination)(3)

   • Nonaccidental injury

   • Prolonged cardiac arrest (cardiac arrest and remaining pulseless after arrival in the emergency department despite advanced cardiac life support which included at least one dose of epinephrine)(3)

   • Preexisting neurodevelopmental disability

   • Brain injury due to a gunshot wound

   • High cervical (C1 – C5) spinal cord injury

   • Acute isolated epidural hematoma (expected to recover after surgical evacuation)(3)

   • Pregnancy

5. The interventions compared:

   • Those who received hypothermia therapy (32.5 ± 0.5°C for 24hr initiated within 8hr after brain injury) to those who were kept normothermic (37 ± 0.5°C).

6. The outcomes evaluated:

   • Primary outcome: The proportion of patients who had an unfavorable outcome (severe disability, persistent vegetative state, or death) as assessed by the six-point Pediatric Cerebral Performance Category scale at 6 months after the injury.

   • Secondary outcomes: The Pediatric Cerebral Performance Category scale was also assessed at 1 week, 1, 3, and 12 months after the injury (relative to pre-morbid level of functioning estimated by parents or guardians). Other measures of functioning (intelligence, memory, information processing speed, executive functions) were assessed if possible at 3 and 12 months after the injury. A variety of physiologic parameters, ICU/hospital length of stay, adverse events, and co-interventions used were also recorded.

II. Are the results of the study valid?

Primary questions:

1. Was the assignment of patients to treatments randomized?

   Yes. Patients were randomly assigned to treatment or no treatment groups within four different blocks stratified by age (greater or less than 7 years old) and treatment center.

2. Were all patients who entered the trial properly accounted for and attributed at its conclusion?

   Was follow-up complete?

   No. “20 of the 225 patients were lost to follow up at 6 months – 6 of 108 patients (6%) in the hypothermia group and 14 of 117 (12%) in the normothermia group.� Assuming that all of the patients lost from the hypothermia group did poorly and that all of the patients from the control/normothermia group did well, hypothermia therapy was associated with an unfavorable outcome (P = 0.001); with the opposite scenario there was no increased risk of an unfavorable outcome with hypothermia therapy (P=0.82). The conclusions of the trial (see below) maintain that hypothermia therapy did not improve primary outcome and there was a trend toward increased mortality in the hypothermia group. Therefore, the loss to follow up can be considered minimal, as adjusting for these patients does not change the findings or conclusions from the study.

   Were patients analyzed in the groups to which they were randomized?

   Yes. 102 of 108 (94%) assigned to hypothermia therapy received it. 114 of 117 (97%) assigned to normothermia received it. No patient assigned to normothermia was treated with hypothermia (no crossover). Statistical analysis of the primary outcome was conducted according to the intention to treat (ITT) principle and then separately according to the treatment received. Of note, no major differences were found between these two analyses. All secondary outcomes were analyzed according to the ITT principle with additional analyses of variables “related to the process of care… [using] the nonparametric Wilcoxon rank-sum test�.

Secondary questions:

3. Were patients, health workers, and study personnel "blind" to treatment?

   Patients: no (not possible)

   Health workers: no (not possible)

   Study personnel: blinded to treatment assignment at the time of the outcome assessment

4. Were the groups similar at the start of the trial?

   Yes. At the start of the trial the two groups were highly similar in their baseline characteristics including demographics, GCS, Pediatric Trauma Score, cause of injury, clinical status upon presentation, head CT findings, and concomitant injuries.

5. Aside from the experimental intervention, were the groups treated equally?

   Yes and no. The patients were all treated according to the same strict guidelines for general management of brain injury as determined in the related pre-trial. 3 However, there were important differences in particular interventions received by each group:

   • There were significantly more patients treated with normothermia (vs. hypothermia) who also received hypertonic saline within 24hr for raised ICP (54 (46%) vs. 34 (31%); p = 0.02). According to the study treatment guidelines (3) , 3% saline was given for ICP>20mmHg (or >25 in refractory cases) despite draining CSF and giving mannitol (as long as serum was not “severely hyperosmolarâ€�). The reason more patients in the normothermic group required more aggressive osmotherapy is unclear.

   • There were significantly more patients treated with hypothermia (vs. normothermia) who also received vasoactive medications for hypotension during rewarming (92 (85%) vs. 66 (56%), p < 0.001). Blood pressure (hypotension) was managed according to study treatment guidelines (3) to maintain CPP (>60 mmHg if > 10 years old, or >50 mmHg if <10 years old) in the face of ICP < 20 mmHg. The cause of the hypotension was not discussed and was not formally compared to other clinical parameters (dysrhythmia, concurrent illnesses, bleeding, etc).

   These differences in co-interventions are potential weaknesses in this study, especially as these particular co-interventions (osmolar therapy and vasoactive medications) may be significantly effective at altering the primary outcomes in the study and as blinding is limited. We therefore rely on the rigorous randomization to eliminate sources of bias from these co-interventions.

III. What were the results?

1. How large was the treatment effect?

   The sample size of the study (n = 225) was sufficient to “detect a reduction of 20 percentage points in the absolute risk of an unfavorable outcome, from 50% in the control (normothermia) group to 30% in the hypothermia group, with a two-sided alpha level of 0.05 and a statistical power of 80%, [assuming] a 10% rate of loss to follow up�.

   Primary outcome:

   The proportion of patients who had an unfavorable outcome (severe disability, persistent vegetative state, or death) as assessed by the six-point Pediatric Cerebral Performance Category scale at 6 months after the injury.

   There was no significant effect from treatment:

   • There was no significant difference in the primary outcome when patients were treated with hypothermia (31%) versus normothermia (22%) (RR 1.41 (95% CI 0.89 to 2.22; P=0.14)).

   • Adjusting for the 20 patients lost to follow up did not change the primary outcome.

   • Adjusting for the effects of 8 predetermined factors possibly associated with outcome in children with traumatic brain injury (age, Glasgow Coma Scale score on admission, temperature on admission, hypotension or hypoxia on admission, ICP, number of therapies used to control ICP, use of hypertonic saline for ICP control, CT scan variables: extradural hematoma, cerebral edema, midline shift) does not significantly change the primary outcome. However, there was a trend towards an unfavorable outcome being more likely with hypothermia therapy than with the other factors measured (“the adjusted odds ratio for an unfavorable outcome with hypothermia therapy was 2.33 (95% CI, 0.92 to 5.93; P=0.08â€�).

     In fact, there was evidence that hypothermia therapy increased morbidity and mortality in certain subgroups:

   • Subgroup analysis of the primary outcome revealed a trend toward higher risk for an unfavorable outcome with hypothermia therapy in patients 7 years of age or older (RR 1.71, 95% CI 0.96 to 3.06, P=0.06) and a significantly higher risk for an unfavorable outcome in patients with ICP < 20mmHg (RR 2.12, 95% CI 1.07 to 4.19, P=0.03).

   • Assessment of mortality (as measured by chi-square tests and Cox proportional hazard models) revealed a RR of death with hypothermia therapy of 1.4 (95% CI, 0.9 to 2.27; P=0.06) and a hazard ratio for death with hypothermia therapy of 1.84 (95% CI, 0.95 to 3.58; P=0.07). When adjusted for factors associated with outcome in children with traumatic brain injury (as above), the hazard ratio for death is larger, 2.36 (95% CI, 1.04 to 5.37; P=0.04). This analysis suggests there is a trend towards increased mortality with hypothermia treatment.

   Secondary outcomes:

   The Pediatric Cerebral Performance Category scale assessed at 1 week, 1, 3, and 12 months after the injury (relative to pre-morbid level of functioning estimated by parents or guardians). Other measures of functioning (intelligence, memory, information processing speed, executive functions) assessed if possible at 3 and 12 months after the injury. Duration of care, adverse events, co-interventions used, and a variety of physiologic parameters (during cooling and re-warming periods) including ICP and hemodynamic measures were recorded and compared among treatment groups.

   ICP was lower during cooling (at 16hr and 24hr) and higher during re-warming (at 48hr and 72hr) (all time points with P<0.05). In the hypothermia treatment group, heart rate, and, during re-warming, mean blood pressures (and cerebral perfusion pressures) were notably lower with more episodes of hypotension. Long-term visual memory was worse in the normthermia group at 12 months. However, loss of patients (due to death, follow up, being too young to participate in testing, or inability to participate due to functional impairment) confounds this result. Also observed in the hypothermia group in the first 24hr were higher serum glucose and lower platelets; and higher prothrombin time and lactate between 25 and 72hrs (all comparisons with P<0.05). With all other secondary outcomes, there was no significant difference between treatment groups.

2. How precise was the estimate of the treatment effect?

   With regard to the primary outcome (i.e., those who had an unfavorable outcome including death), no treatment effect was found. The calculated ARR for their primary outcome was -0.09% (this was not reported). The relative risk (RR) of an unfavorable outcome with hypothermia therapy was 0.31/0.22 = 1.41 (95% CI 0.89 to 2.22; P=0.14). Due to the 95% CI crossing zero, the RR has potential to be either positive or negative. Nonetheless, the 2.2 times greater risk of an unfavorable outcome in the treatment group is important to note.

   The trend towards increased mortality was only significant when adjusted for factors associated with outcome in children with traumatic brain injury. In the subgroup analysis, a higher risk for an unfavorable outcome in patients 7 years of age or older when treated with hypothermia was noted. However, this is not greatly different from the primary analysis and therefore not as significant. The higher risk for unfavorable outcome in patients with ICP < 20mmHg (RR 2.12, 95% CI 1.07 to 4.19, P=0.03) on the other hand is important and should be considered (see below).

   The other secondary endpoints (HR, MAP/hypotension, CPP, serum glucose, lactate, PTT, and platelet count) are significantly different between treatment groups and should be considered.

IV. Will the results help me in caring for my patients?

1. Can the results be applied to my patient care?

   Yes and no. The inclusion and exclusion criteria applied in this study accurately reflect patients admitted to the Intensive Care Unit for severe traumatic brain injury. However, the exclusion criteria of 8hr might include patients too far progressed in their illness. Would hypothermia treatment benefit patients if instituted earlier? It would have been interesting to provide an a priori subgroup analysis of these patients. (A post hoc analysis apparently showed no benefit, although these data are not included.)

   The subgroup analysis by age showed a higher risk for an unfavorable outcome with hypothermia therapy in patients 7 years and older. This analysis was generated prior to initiation of the study. However, the differences in effect from treatment in the subgroup are not greatly different from those in the primary group (and have not been studied elsewhere) and therefore are less informative. The authors also report greater risk for an unfavorable outcome with hypothermia in patients with ICP < 20mmHg. While significantly different from the primary group, this analysis was generated post hoc and is therefore less reliable.

2. Were all clinically important outcomes considered?

   Yes and no. Death and severe neurologic disability were the primary outcomes considered. These are the most clinically important outcomes to consider in patients with severe traumatic brain injury. Important secondary outcomes and substitute endpoints were also included: hemodynamic parameters, ICP, duration of care, and adverse events. Importantly, seizure activity (including nonconvulsive status epilepticus) was not included as a secondary outcome. Seizures often occur after traumatic brain injury and represent a significant short- and long-term co-morbidity (4,5). Moreover, differentiating seizure activity from shivering (due to hypothermia), and detecting nonconvulsive status epilepticus often requires EEG. Diagnosis and treatment of seizures in these patients is important. The effect that hypothermia therapy has on seizure activity should have been considered.

3. Are the likely treatment benefits worth the potential harms and costs?

   Yes and no. It is clear that hypothermia therapy as provided in this study does not improve the neurologic outcome of patients, and the potential harm (death) should dissuade practitioners from using it in this patient population. These conclusions mirror those from numerous adult studies (see table below). However, these results conflict with several other adult trials (see table below), a meta-analysis of adult randomized control trials of hypothermia therapy (6), hypothermia trials in neonates with hypoxia-ischemia brain injury (7,8), and adults with brain injury after ventricular fibrillation-induced cardiac arrest (9), all of which support the limited use of therapeutic hypothermia in specific clinical situations. Whether the reported differences in outcomes reflect different physiological states in neonatal, adult, and pediatric brain injury (10), mechanistic differences between brain trauma and cerebral hypoxia-ischemia, or simply differences in study design remains to be determined. There are numerous studies which demonstrate that hypothermia can control ICP (see table below) as was seen in this trial. It may be that the potential benefit from decrease ICP was lost during re-warming when blood pressures fell. So, while using therapeutic hypothermia to improve neurologic outcome is not supported by this trial, it may be reasonable to consider using hypothermia as an adjuvant or “salvage� therapy in cases of intractable intracranial hypertension. That said, further controlled studies will be needed to determine whether different hypothermia therapy protocols (e.g. given earlier, given longer, different re-warming protocols, etc) will improve survival, neurologic outcome, or other co-morbid states such as seizures and intracranial hypertension.

   
   

   		Good Outcome ( Glasgow Outcome Score 4 or 5)			Effect on ICP
   Author (year)	N	Hypothermia	Normothermia	Significant?
   Shiozaki (1993) (11)	33	38%	6%	N
   Clifton (1993) (12)	46	52%	36%	N	None
   Marion (1997) (13)	82	56%	33%	Y	Reduced
   Jiang (2000) (14)	87	47%	27%	Y	Reduced
   Shiozaki (2001) (15)	91	23%	30%	N	None
   Clifton (2001) (16)	368	43%	43%	N	Reduced
   Zhi (2003) (17)	396	62%	38%	Y	Reduced

   
   

References:

1. http://www.cche.net/usersguides/therapy.asp
2. Hutchison JS, Ward RE, Lacroix J, Hébert PC, Barnes MA, Bohn DJ, Dirks PB, Doucette S, Fergusson D, Gottesman R, Joffe AR, Kirpalani HM, Meyer PG, Morris KP, Moher D, Singh RN, Skippen PW, Hypothermia Pediatric Head Injury Trial Investigators and the Canadian Critical Care Trials Group. (2008). Hypothermia Therapy after Traumatic Brain Injury in Children. N Engl J Med 5;358(23):2447-56.
3. Hutchison J, Ward R, Lacroix J, Hébert P, Skippen P, Barnes M, Meyer P, Morris K, Kirpalani H, Singh R, Dirks P, Bohn D, Moher D; HyP-HIT Investigators; Canadian Critical Care Trials Group. (2006). Hypothermia pediatric head injury trial: the value of a pretrial clinical evaluation phase. Dev Neurosci. 28(4-5):291-301.
4. Vespa PM, Nuwer MR, Nenov V, Ronne-Engstrom E, Hovda DA, Bergsneider M, Kelly DF, Martin NA, Becker DP. (1999). Increased incidence and impact of nonconvulsive and convulsive seizures after traumatic brain injury as detected by continuous electroencephalographic monitoring. J Neurosurg. 91(5):750-60.
5. Annegers JF, Coan SP. (2000). The risks of epilepsy after traumatic brain injury. Seizure. 9(7):453-7.
6. McIntyre L, Fergusson D, Hébert P, Moher D, Hutchinson J. (2003). Prolonged therapeutic hypothermia after traumatic brain injury in adults. A systematic review. JAMA. 289:2992-9.
7. Shankaran S, Pappas A, Laptook AR, McDonald SA, Ehrenkranz RA, Tyson JE, Walsh M, Goldberg RN, Higgins RD, Das A; NICHD Neonatal Research Network. (2008) Outcomes of safety and effectiveness in a multicenter randomized, controlled trial of whole-body hypothermia for neonatal hypoxic-ischemic encephalopathy. Pediatrics. 122(4):e791-8.
8. Jacobs S, Hunt R, Tarnow-Mordi W, Inder T, Davis P. (2007). Cooling for newborns with hypoxic ischaemic encephalopathy. Cochrane Database Syst Rev. (4):CD003311.
9. Hypothermia after Cardiac Arrest Study Group. (2002). Mild therapeutic hypothermia to improve the neurologic outcome after cardiac arrest. N Engl J Med. 346(8):549-56.
10. Giza CC, Mink RB, Madikians A. (2007). Pediatric traumatic brain injury: not just little adults. Curr Opin Crit Care. 13(2):143-52.
11. Shiozaki T, Sugimoto H, Taneda M, Yoshida H, Iwai A, Yoshioka T, Sugimoto T. (1993). Effect of mild hypothermia on uncontrollable intracranial hypertension after severe head injury. J Neurosurg. 79(30):363-8.
12. Clifton GL, Allen S, Barrodale P, Plenger P, Berry J, Koch S, Fletcher J, Hayes RL, Choi SC. (1993). A phase II study of moderate hypothermia in severe brain injury. J Neurotrauma. 10(3):263-71
13. Marion DW, Penrod LE, Kelsey SF, Obrist WD, Kochanek PM, Palmer AM, Wisniewski SR, DeKosky ST. (1997). Treatment of traumatic brain injury with moderate hypothermia. N Engl J Med. 336(8):540-6.
14. Jiang J, Yu M, Zhu C. (2000). Effect of long-term mild hypothermia therapy in patients with severe traumatic brain injury: 1-year follow-up review of 87 cases. J Neurosurg. 93(4):546-9.
15. Shiozaki T, Hayakata T, Taneda M, Nakajima Y, Hashiguchi N, Fujimi S, Nakamori Y, Tanaka H, Shimazu T, Sugimoto H. (2001). A multicenter prospective randomized controlled trial of the efficacy of mild hypothermia for severely head injured patients with low intracranial pressure. Mild Hypothermia Study Group in Japan . J Neurosurg. 94(1):50-4.
16. Clifton GL, Miller ER, Choi SC , Levin HS, McCauley S, Smith KR Jr, Muizelaar JP, Wagner FC Jr, Marion DW, Luerssen TG, Chesnut RM, Schwartz M. (2001). Lack of effect of induction of hypothermia after acute brain injury. N Engl J Med. 344(8):556-63.
17. Zhi D, Zhang S, Lin X. (2003). Study on therapeutic mechanism and clinical effect of mild hypothermia in patients with severe head injury. Surg Neurol. 59(5):381-5.

Last Updated: January 15, 2009

Fill out our Feedback Form | Editorial Notes | About Us