The effectiveness of hypothermia treatment for severe subarachnoid hemorrhage (SAH) was evaluated at the same facility under the same director. A total of 187 patients with SAH, 67 admitted before the introduction of hypothermia treatment in May 1999 (early cases) and 120 treated thereafter (late cases), were transported to the National Cardiovascular Center and treated in the acute phase between November 1997 and September 2001. Brain hypothermia treatment was performed in 19 patients of the 120 late cases, 10 males and 9 females aged 33-72 years (mean 57. 6 years), treated by direct surgery in 15 and endovascular surgery in 4. The indications for hypothermia treatment were age of 75 years or younger, SAH due to rupture of a cerebral aneurysm, Japan Coma Scale score of 100 or higher, and initiation of treatment within 24 hours after the onset. The body core temperature was sustained at 34°C for 48 hours, rewarming was performed over 48 hours, and normothermia was maintained thereafter. The outcome, evaluated according to the modified Rankin scale (m-RS) on transfer to another hospital or after 3 months, was m-RS 3 in 1 patient, m-RS 4 in 4, m-RS 5 in 3, and death in 11. Before the introduction of hypothermia treatment (early period), 16 patients showed the indications for the treatment, and their outcomes were m-RS 3 in 2, m-RS 4 in 3, m-RS 5 in 2, and death in 9. Cerebral vasospasm was important as a prognostic factor, markedly deteriorating the outcome. Hyperthermia after therapeutic hypothermia induced brain swelling and markedly affecting the outcome. Brain hypothermia treatment did not improve the outcome of severe SAH compared with the period before its introduction. The emphasis in treating severe SAH should be placed on the maintenance of normothermia to prevent brain swelling and elimination of factors that may induce cerebral vasospasm, rather than interventional hypothermia for aggressive brain protection.