The benzodiazepine, midazolam, is commonly used for sedation and anesthesia in the operating room and the intensive care unit where there is a risk of cerebral ischemia. We therefore examined its ability to reduce damage subsequent to cerebral ischemia. Male Wistar rats were randomly assigned to a high-dose midazolam group or a matched vehicle group and a lower dose of midazolam group or a matched vehicle group. The rats underwent 90 minutes of middle cerebral artery occlusion. In the midazolam groups, the first dose of midazolam (10 or 25 mg/kg) was given by a 10-minute intravenous infusion before ischemia; a second dose of (1/2) the initial dose (5 or 12.5 mg/kg) was given 1 hour after the onset of ischemia. In the vehicle groups, a similar volume of vehicle was given at the same time intervals. Infarct size, NeuN immunopositive cells in the ischemic penumbral and core regions, and neurologic outcome were determined 7 days after ischemia. Compared with vehicle-treated rats, the higher-dose midazolam (25 mg/kg)-treated rats had a smaller infarct size (93.9+/-63.5 mm vs. 152.0+/-53.7 mm, P<0.05), more NeuN immunopositive cells in the ischemic core region (206.7+/-211.3/mm vs. 40.0+/-66.3/mm, P<0.01), and better neurologic outcome (P<0.05). Midazolam at a lower dose (10 mg/kg) had no significant effects. Although midazolam generated dose-dependent hemolysis, this hemolysis was transient. Midazolam (25 mg/kg) caused a loss of the righting reflex in rats that lasted until 19.9+/-1.3 min after the injection, the anesthetic dose of midazolam in rats is approximately 100x greater than the anesthetic dose for humans. An anesthetic dose of midazolam in rats reduced neuronal damage and improved neurologic outcome 7 days after focal cerebral ischemia, however, it also caused transient hemolysis.