Since hypothermia may be a potential treatment for perinatal cerebral hypoxic-ischemic injury, we used an established neonatal model of hypoxia-ischemia to determine the time after injury at which cooling had the best protective effect. Fourteen-day-old Wistar rats were subjected to right carotid artery ligation and hypoxia (8% O(2) for 90 min). Immediately at the end of hypoxia (defined as 0h), animals were either maintained at normal body temperature until sacrifice (normothermia) or subjected to hypothermia. In a preliminary study, the effects of a reduction in temperature and the duration of such cooling were investigated; animals were cooled (until brain temperature reached 33 degrees C or 30 degrees C) for 2, 4, or 6 h commencing immediately after hypoxia. In a second study, animals were cooled (brain temperature 30 degrees C) for 6 h commencing at either 0, 2, 4, or 6 h after the end of hypoxia. Sham-operated animals were used as controls. Twenty-four hours after hypoxia-ischemia, cerebral energy metabolism was measured by phosphorus magnetic resonance spectroscopy, and at 5 d cerebral infarction was measured by planimetry. In normothermic animals the ratio of phosphocreatine/inorganic phosphate (PCr/Pi) had fallen markedly 24 h following hypoxia-ischemia. In contrast, animals cooled between 6 and 12 h displayed high PCr/Pi ratios similar to those in control animals. Similarly, after 5 d, infarct area was significantly reduced only in animals cooled between 6 and 12 h after injury. These results indicate that cooling between 6 and 12 h after hypoxia-ischemia is more effective in reducing cerebral injury than other cooling regimes and suggest that the physiologic events during this period are critical for understanding cerebral infarction.