Cooling the arrested brian.

Therapeutic hypothermia following an Out-of-Hospital Cardiac arrest reduces morbidity (NNT 4 for an additional good neurological outcome).

Level of Evidence 1⁺

Lead author's name and fax: Dr Stephen A Bernard, Dandenong Hospital, David street, Dundenong, VIC 3175, Australia. s.bernard@southernhealth.org.au

Three-part Clinical Question: Does the use of mild hypothermia following out-of-hospital cardiac arrest improve outcome.

The Study: Single-blinded randomised controlled trial with intention-to-treat.

The Study Patients: Patients admitted to four participation hospitals following an out-of-hospital cardiac arrest in Melbourne Australia between September 1996 and June 1999.Inclusion criteria: Initial rhythm of ventricular fibrillation, successful return of spontaneous circulation (ROSC), persistent coma after ROSC, transfer to one of the four participating hospitals. Exclusion criteria: age<18 for men, age <50 for women, cardiogenic shock (a systolic blood pressure <90mmHg despite epinephrine infusion), or possible causes of coma other than cardiac arrest.

Control group (N = 34; 34 analysed): Routine pre-hospital care. In hospital; sedated, paralysed, ventilated, inotropic support with epinephrine. Aiming for PO2 100mmHg (13.1kPa), pCO2 40mmHg (5.2kPa), MAP 90-100mmHg. Potassium and blood sugar controlled to 4 and 10 mmol/l respectively. Thrombolysis if evidence of acute infarction (and not contra-indicated), heparin infusion for acute coronary syndrome. Lignocaine bolus (1mg/kg) and infusion (2mg/min) for 24 hours. Aspirin. Body temperature measured initially by tympanic thermometer and then either bladder probe or PAC. If hypothermic on arrival then passively rewarmed to 37C. Temperature maintained at 37C. After 24hrs care followed the normal ICU protocols. Patients who regained consciousness underwent extubation and were transferred to the CCU. Treatment was withdrawn if patients were still comatosed at 72hrs. Patients with an uncertain prognosis underwent tracheostomy and were discharged from ICU.

Experimental group (N = 43; 43 analysed): Cooled at scene by the ambulance service, after ROSC, using cool pack to the head and torso–also i.v. cold fluids – novel and very effective at cooling the brain. As for control group with the addition of: Vigorous cooling in the emergency department. (ice packs to head, torso, neck, limbs). Cooled to 33C within 2 hours. Maintained at 33C for 12 hours. At 18hrs active rewarming, for 6 hours, using warm air blankets. Sedation at paralysis was continued to prevent shivering. After 24hours, patients care followed the normal ICU protocols.

1. Do the methods allow accurate testing of the hypothesis? Yes. There was a significant difference in bystander CPR rates (Normothermia > Hypothermia), which you would expect, if anything, to decrease the impact of cooling. It was not feasible to blind treating physicians but the assessing rehab consultant was blinded to treatment.

2. Do the statistical tests correctly test the results to allow differentiation of statistically significant results? Yes. Univarate analysis revealed that a patients’ age and the time from collapse to ROSC significantly affected the outcome. (For each 2yr increase in age there was a 9% decrease in the likelihood of a good outcome, and each additional 1.5 mins in delay of ROSC led to a 14% decrease.)
With multivarate analysis and adjustment for base-line differences in age and time to ROSC the odds ratio for good outcome in the hypothermia group compared with normothermia was 5.25 (95% CI; 1.47-18.76, p=0.011).

4. Did results get omitted, and why? No (there were a few protocol violations but these patients were included appropriately - intention-to-treat)

5. Did they suggest areas of further research? Yes (fully blinded study, studies to determine the optimum period of cooling)

6. Did they make any recommendations based on the results and were they appropriate? No

10. What grade of recommendation can I make when this study is considered along with other available evidence? A, but the N/S result for death is different from the HACA trial. We have to wonder why?

11. Should I change my practice because of these results? Yes – if not doing so already, patients meeting the trial inclusion criteria should be managed as the intervention group. Hypothermia was well tolerated and not associated with clinically significant adverse side effects. It is encouraging that hypothermia did not result simply in an increased number of poor neurological outcome survivors.

12. Should I audit my current practice because of these results? Yes – to determine whether all patients meeting the inclusion criteria are being managed appropriately, and their outcome.

Appraised by: Dr Chris Cairns, Department of Anaesthesia and Intensive Care, Stirling Royal Infirmary, Livilands, Stirling, FK8 2AU, UK: Monday, September 4, 2004

Cooling the arrested brain.