Hypothermia during brain ischemia can improve neurological outcome. This study tested whether local cranial cooling during the low-flow state of cardiopulmonary resuscitation (CPR) could produce clinically significant cerebral cooling. Ice was applied to the heads and necks of subjects (hypothermia group) with out-of-hospital cardiac arrest (OOHCA) during CPR. Nasopharyngeal and tympanic temperatures were measured as surrogates for cerebral temperature. The rate of cranial cooling in the hypothermia group (-0.06 +/- 0.06 degrees C/min) was not significantly increased compared with a control group without ice (-0.04 +/- 0.07 degrees C/min), although older age was associated with more rapid cranial cooling. Of note, many subjects with OOHCA are already mildly hypothermic (mean cranial temperature= 35.0 +/- 1.2 degrees C) when they are first encountered in the field. This study suggests that brief cranial cooling is ineffective for rapidly lowering brain temperature. However, most cardiac arrest victims are spontaneously mildly hypothermic and preventing rewarming may provide some of the desired benefits of cerebral hypothermia.