Cooling the Acutely Brain Injured Patient

 

The use of hypothermia to 33C within 6 hours of a severe brain injury which is maintained for 48hrs does not improve outcome compared with 'normothermia' of 37C

Level of Evidence 1+

 

Citation/s:
Lack of Effect of Induction of Hypothermia after Acute Brain Injury. The New England Journal of Medicine 2001 344; 8: 556-563


Lead author's name and fax: Guy L Clifton, Vivian L Smith Center for Neurologic Research, University of Texas-Houston Health Science Center, 6431 Fannin, Suite 7.148, Houston, TX 77030; guy.i.clifton@uth.tmc.edu

 

Three-part Clinical Question: Does the induction of hypothermia following acute brain injury improve outcome when compared to normothermia?


Search Terms: Acute brain injury, induced hypothermia, outcome, temperature regulation

 

The Study:
Single-blinded concealed randomized controlled trial without intention-to-treat.


The Study Patients: Patients were enrolled in 11 centers in the USA. Criteria for inclusion were; age 16-65yrs, a non-penetrating head injury, and a GCS of 3-8 (coma), after resuscitation. Criteria for exclusion were; GCS 3 with unreactive pupils, a life-threatening injury to an organ other than the brain, a systolic BP of <90mmHg after resuscitation, oxygen saturation of less than 94% after resuscitation, bleeding, pregnancy, or known pre-existing medical conditions (e.g. severe heart disease), or if the examiners were unable to initiate cooling within six hours of the brain injury. ICP was measured in all patients. All patients received 5-10mg/hr of morphine for at least 72hrs. Patients who were hypothermic on admission were not actively warmed. CPP was maintained >70mmHg with fluid or vasopressors. Raised ICP was treated sequentially with paralysis, ventricular drainage, hyperventilation (limited by pCO2>30mmHg), mannitol (limited by serum osmo <315mOsm/kg), barbiturate coma with thiopentone. All patients were loaded and maintained on phenytoin.


Control group (N = 193; 178 analyzed): Constant temperature measurement via indwelling bladder catheter. Body temperature maintained at 37C. Vecuronium administered as needed for respiratory management. Enteral or parenteral feeding commenced within 48hrs of the injury.


Experimental group (N = 199; 190 analyzed): Constant temperature measurement via indwelling bladder catheter. Cooling began immediately after randomization; the goal was to achieve the target bladder temperature of 33C within eight hours of the injury. Cooling procedures included; ice application, gastric lavage, ventilation with room temperature air. After the target temperature was achieved temperature control pads were used to maintain the bladder temperature between 32.5 and 34C for 48hrs. At the end of this period the patients were rewarmed slowly (<0,5C per 2 hours). All patients in this group paralyzed with vecuronium for 48hrs to prevent shivering. Enteral or parenteral feeding commenced 72 hrs after the injury.

The Evidence:

Outcome

Time to Outcome

CER

EER

RRR

ARR

NNT

Death

6 months

0.249

0.266

7%

0.017

59

95% Confidence Intervals:

42% to 28%

0.104 to 0.070

NNT = 14 to INF; NNH = 10 to INF

Poor Outcome (severe disability, vegetative state, or death)

6 Months

0.528

0.543

3%

0.015

67

95% Confidence Intervals:

22% to 16%

0.114 to 0.084

NNT = 12 to INF; NNH = 9 to INF

 

Comments:


A well-conducted multi-center study. Suggests that the routine use of hypothermia in brain-injured patients is of no benefit, however there are several additional points. Outcomes in the control group of this study were better than in a previous study by the same group. The only difference in management is that in this study control patients who were hypothermic on admission were not actively warmed. This suggests that either spontaneous hypothermia is a marker of more severe brain injury or that very early cooling is necessary to gain a neuro-protective effect. Brain-injured patients who are hypothermic on admission should NOT be rewarmed.


The results of this study are somewhat surprising when you consider the theoretical pathophysiological benefits of cooling in brain injury. A possible reason for this is that the majority of brain-injured patients will develop pyrexia within the first 48hr following their insult. Pyrexia is an independent marker of poor outcome. In view of this perhaps the control group in a study of this kind should not be maintained at 37C as this in not 'normal' and will require cooling to achieve it in most.

Appraised by: Chris Cairns, Intensive Care Unit, Western General Hospital, Crewe Road, Edinburgh; Thursday, January 03, 2002
Email: Chris.Cairns@btinternet.com
Kill or Update By: Jan 2004

 

Printer-friendly format