Considerable experimental data exist indicating a detrimental effect of elevated temperature during and after acute brain insults. Numerous clinical studies have found a consistent relationship between elevated temperature and poor outcome. Similarly, reducing temperature below normal seems to improve outcome in experimental models and in patients after cardiac arrest. A causal relationship between fever and poor outcome in patients with neurologic injury has not been established, however. No prospective controlled trial has been performed to determine if control of fever improves clinical outcome in nontraumatic neurologic ICU patients. In addition, two large prospective human trials of induced hypothermia failed to show a benefit in severe TBI and in aneurysm surgery after SAH. Decisions about clinical management must be made in the absence of definitive trials. In such a situation, the decision as to whether to use a particular intervention must balance potential benefit versus potential risk. Current technology allows control of hyperthermia in a fairly safe, efficient, and easy-to-use manner. With this minimal risk, it seems reasonable, in the absence of infection, to treat fever aggressively early after acute central nervous system insults. In infected patients, lowering temperature potentially may compromise innate defense mechanisms, however, making aggressive fever control a less preferable option. Currently, outside the setting of cardiac arrest, there does not seem to be compelling data to support the use of therapeutic hypothermia. The negative trials in TBI and SAH surgery should temper use in those settings. Ongoing trials are expected to help determine if therapeutic hypothermia has a role in the management of acute ischemic stroke.