ObjectivesTo test the hypothesis that selective brain cooling could be performed in an infant model of cardiac arrest and resuscitation without changing core temperature and to study its acute effects on regional organ blood flow, cerebral metabolism, and systemic hemodynamics.DesignProspective, randomized, controlled study.SettingResearch laboratory at a university medical center.SubjectsFourteen healthy infant piglets, weighing 3.5 to 6.0 kg.InterventionsPiglets were anesthetized and mechanically ventilated, and had vascular catheters placed. Parietal cortex (superficial brain), caudate nucleus (deep brain), esophageal, and rectal temperatures were monitored. All animals underwent 6 mins of cardiac arrest induced by ventricular fibrillation, 6 mins of external cardiopulmonary resuscitation (CPR), defibrillation, and 2 hrs of reperfusion. Normal core temperature (rectal) was regulated in all animals. In seven control animals (group 1), brain temperature was not manipulated. In seven experimental animals (group 2), selective brain cooling was begun during CPR, using a cooling cap filled with minus 30 degrees C solution. Selective brain cooling was continued for 45 mins of reperfusion, after which passive rewarming was allowed. Regional blood flow (microspheres) and arterial and sagittal sinus blood gases were measured prearrest, during CPR, and at 10 mins, 45 mins, and 2 hrs of reperfusion.Measurements and Main ResultsRectal temperature did not change over time in either group. In group 1, brain temperature remained constant except for a decrease of 0.6 degrees C at 10 mins of reperfusion. In group 2, superficial and deep brain temperatures were lowered to 32.8 plus minus 0.7 (SEM) degree C and 34.9 plus minus 0.4 degrees C, respectively, by 15 mins of reperfusion. Superficial and deep brain temperatures were further lowered to 27.8 plus minus 0.8 degrees C and 31.1 plus minus 0.3 degrees C, respectively, at 45 mins of reperfusion. Both temperatures returned to baseline by 120 mins. Cerebral blood flow was not different between groups at any time point, although there was a trend for higher flow in group 2 at 10 mins of reperfusion (314% of baseline) compared with group 1 (230% of baseline). Cerebral oxygen uptake was lower in group 2 than in group 1 (69% vs. 44% of baseline, p equals .02) at 45 mins of reperfusion. During CPR, aortic diastolic pressure was lower in group 2 than in group 1 (27 plus minus 1 vs. 23 plus minus 1 mm Hg, p equals .007). Myocardial blood flow during CPR was also lower in group 2 (80 plus minus 7 vs. 43 plus minus 7 mL/min/100 g, p equals .002). Kidney and intestinal blood flows were reduced during CPR in both groups; however, group 2 animals also had lower intestinal flow vs. group 1 at 45 and 120 mins of reperfusion.ConclusionsSelective brain cooling by surface cooling can be achieved rapidly in an infant animal model of cardiac arrest and resuscitation without changing core temperature. Brain temperatures known to improve neurologic outcome can be achieved by this technique with minimal adverse effects. Because of its ease of application, selective brain cooling may prove to be an effective, inexpensive method of cerebral resuscitation during pediatric CPR.(Crit Care Med 1996; 24:1009-1017)