Hypoxic-ischemic injuries can evolve over several days, and recent studies suggest that further neuronal death may occur 6 to 72 h later. Because cerebral temperature is an important determinant of outcome during the primary injury, we investigated the effect of temperature, on outcome, during the later phases of injury. Hypoxic-ischemic injury was induced in 21-d-old rats by unilateral ligation of the right carotid artery followed by exposure to 15 min of hypoxia of 8% O2 at 34 degrees C. Cerebral temperature changes were induced by modifying environmental temperature. The rats were divided into four treatment groups: group 1 (n = 15) remained at 34 degrees C for 72 h; group 2 (n = 14) were kept at 34 degrees C for 6 h and then at 22 degrees C for the remaining 66 h; group 3 (n = 17) remained at 22 degrees C for 6 h and 34 degrees C for the next 66 h; group 4 (n = 16) remained at 22 degrees C for 72 h. Rats kept at 22 or 34 degrees C had cortical temperatures of 35.5 +/- 0.1 degrees C and 37.9 +/- 0.2 degrees C, respectively. Histologic outcome was assessed 72 h after hypoxia. The area of cortical infarction was reduced in group 4 compared with groups 1-3 (p < or = 0.05). Striatal damage was reduced in group 4 (p = 0.05). Hippocampal neuronal loss was not significantly altered. In a subsequent study the area of cortical infarction was 12.1 +/- 3 mm2 in group 1 (n = 11) compared with 3.4 +/- 1.5 mm2 group 4 treated rats (n = 10) 21 d after the injury (p < 0.01). Thus hypothermia spanning both the first 6 h and from 6 to 72 h after injury was needed to improve outcome. Conversely exposure to the thermoneutral environment exacerbated the injury. These observations suggest that prolonged moderate cerebral hypothermia can be used to suppress the cytotoxic processes that occur after hypoxic-ischemic injury.