We previously found mild hypothermia (34-36 degrees C), induced before cardiac arrest, to improve neurologic outcome. In this study we used a reproducible dog model to evaluate mild hypothermia by head cooling during arrest, continued with systemic cooling (34 degrees C) during recirculation and for 1 h after arrest. In four groups of dogs, ventricular fibrillation (no flow) of 12.5 min at 37.5 degrees C was reversed with cardiopulmonary bypass and defibrillation in less than or equal to 5 min, and followed by controlled ventilation to 20 h and intensive care to 96 h. In Study A we resuscitated with normotension and normal hematocrit; Control Group A-I (n = 12) was maintained normothermic, while Treatment Group A-II (n = 10) was treated with hypothermia. In Study B we resuscitated with hypertension and hemodilution. Control Group B-I (n = 12) was maintained normothermic (6 of 12 were not hemodiluted), while Treatment Group B-II (n = 10) was treated with hypothermia. Best overall performance categories (OPCs) achieved between 24 and 96 h postarrest were in Group A-I: OPC 1 (normal) in 0 of 12 dogs, OPC 2 (moderate disability) in 2, OPC 3 (severe disability) in 7, and OPC 4 (coma) in 3 dogs. In Group A-II, OPC 1 was achieved in 5 of 10 dogs (p less than 0.01), OPC 2 in 4 (p less than 0.001), OPC 3 in 1, and OPC 4 in 0 dogs. In Group B-I, OPC 1 was achieved in 0 of 12 dogs, OPC 2 in 6, OPC 3 in 5, and OPC 4 in 1 dog. In Group B-II, OPC 1 was achieved in 6 of 10 dogs (p less than 0.01), OPC 2 in 4 (p less than 0.05), and OPC 3 or 4 in 0 dogs. Mean neurologic deficit and brain histopathologic damage scores showed similar significant group differences. Morphologic myocardial damage scores were the same in all four groups. We conclude that mild brain cooling during and after insult improves neurologic outcome after cardiac arrest.