Perioperative cerebral ischemic insults are common in some surgical procedures. The notion that induced hypothermia can be employed to improve outcome in surgical patients has persisted for six decades. Its principal application has been in the context of cardiothoracic and neurosurgery. Mild (32-35 degrees C) and moderate (26-31 degrees C) hypothermia have been utilized for numerous procedures involving the heart, but intensive research has found little or no benefit to outcome. This may, in part, be attributable to confounding effects associated with rewarming and lack of understanding of the mechanisms of injury. Evidence of efficacy of mild hypothermia is absent for cerebral aneurysm clipping and carotid endarterectomy. Deep hypothermia (18-25 degrees C) during circulatory arrest has been practiced in the repair of congenital heart disease, adult thoracic aortas, and giant intracranial aneurysms. There is little doubt of the protective efficacy of deep hypothermia, but continued efforts to refine its application may serve to enhance its utility. Recent evidence that mild hypothermia is efficacious in out-of-hospital cardiac arrest has implications for patients incurring anoxic or global ischemic brain insults during anesthesia and surgery, or perioperatively. Advances in preclinical models of ischemic/anoxic injury and cardiopulmonary bypass that allow definition of optimal cooling strategies and study of cellular and subcellular events during perioperative ischemia can add to our understanding of mechanisms of hypothermia efficacy and provide a rationale basis for its implementation in humans.