The recent advent of "warm heart" surgery has resulted in reexamination of the neuroprotective effects of hypothermia in the setting of cardiopulmonary bypass (CPB). Hypothermia has been shown to confer significant protection in the setting of transient, but not permanent, ischemia. The mechanism of this neuroprotection is unclear at this time. Reduction in cerebral metabolic rate is believed to be less important compared with the effect of hypothermia on the release of excitatory neurotransmitters, catecholamines, or other mediators of cellular injury. It is for this reason that mild hypothermia (33-35 degrees C) is believed to confer significant neuroprotection. Two large randomized trials of warm versus cold heart surgery have been reported. Neither study found a difference in terms of neuropsychologic dysfunction. However, one study identified a threefold increase in strokes in the "warm" patients. The reasons for this difference are not clear; however, there were various differences in technique and patient population that may have been important. There are other reports of large series of patients undergoing normothermic bypass, with no increase in stroke rate over what is reported elsewhere in the literature. To date, the evidence would suggest that neuropsychologic function is not affected by CPB temperature, suggesting that the transient ischemia is not an important mechanism in this injury. Stroke after CPB is usually the result of permanent ischemia, and hypothermia's effect in this setting is minimal. It would seem unlikely that hypothermia offers anything more than modest benefits in the clinical situation where there is no circulatory arrest. Copyright (C) 1996 by W.B. Saunders Company.