One of the problems in the clinical application of differential hypothermia (DH) for malignant brain tumors, is the method by which the tumor is kept normothermic under generalized hypothermia. In this study, a unilateral carotid artery was exposed and perfused with blood warmed at 40-42℃ through a femoral-carotid shunt under generalized hypothermia with a rectal temperature of 22-32℃ for 2-10 hours in 26 dogs. There were two groups of 13 dogs each. In Group I, the internal carotid arteries were perfused, and in Group II, the common carotid arteries were. The peak brain temperature of the perfused cerebral hemisphere rose above 35℃ in 8 cases of Group I (35.0-41.6℃) and all cases of Group II (36.4-39.9℃). The flow rate to keep the perfused brain normothermic was about 40-60 ml/min. in Group I, and
50-80 ml/min. in Group II. In Group II, the maximum perfusion pressure was below the physiological level of 200 mmHg, but in Group I, the pressure often rose above 200 mmHg (230-350mmHg). The maximum temperature difference between the body and the perfused cerebral hemisphere ranged from 9 to 13℃ in 8 cases in Group I and 12 in Group II, indicating that the common carotid artery is the preferred artery to be perfused. Eight dogs expired during the DH procedure, and 11 dogs within 24 hours after rewarming. Intracranial hemorrhage was found in seven dogs, all of which belonged to Group I. Six dogs revealed intracerebral hemorrhage localized around a thermistor probe inserted in the brain, which occurred in the perfused cerebral hemisphere in 3 cases and bilaterally in 3 cases. In all of these dogs, the maximum perfusion pressure exceeded 230 mmHg. Subdural hematoma of the perfused side was observed in one dog. No brain edema was observed macroscopically or microscopically in any case. The ratio of the summed amplitude of theta waves to that of alpha waves in EEG was measured in the seven dogs who did not demonstrate any physiological and neurological abnomalities immediately after rewarming. The ratio of the perfused side was smaller during than before perfusion in six dogs, and was smaller than that of the non-perfused side in all cases. These results showed that alpha activities were increased and slow activities were decreased in normothermically perfused brain, and indicated that the perfused brain was more physiological than the non-perfused one. Of a total of 26 dogs used in this experiment, 19 died due to massive hemorrhage (about 150-250 g) in 10 dogs, intracranial hemorrhage in 3, cardiac arrest during hypothermia in 3 and uncertain causes in 3. All seven of the dogs which survived belonged to Group II, and their maximum perfusion pressure was below 200 mmHg. This result suggests that the maximum perfusion pressure is one of the most important factors affecting mortality during the carotid perfusion. In dogs which survived neither neurological deficit nor abnomalities in the histological studies and in the follow-up EEG were observed.