We evaluated the relation between dose and response for the neuroprotective effect of propofol in a rat model with incomplete cerebral ischaemia. For clarification of the mechanism of neuroprotection, plasma catecholamines and tumour necrosis factor-α levels were measured. Three doses (low, moderate and high-dose) of propofol were tested. These produced, respectively, a low amplitude, slowing and a burst-suppression pattern of electroencephalographic activity. Incomplete cerebral ischaemia was produced by right carotid artery occlusion combined with haemorrhagic hypotension (35 mmHg) for 30 min. Neurological outcome at 72 h post-ischaemia in the high-dose group was significantly better than that in both low-dose and moderate-dose groups. Propofol exhibited a trend in the dose-related attenuation of the increases in plasma adrenaline and noradrenaline during ischaemia. Tumour necrosis factor-α increased during and after ischaemia in all groups with no intergroup differences. The results indicate that a burst-suppression dose of propofol provides neuroprotection. The protective effect can not be completely explained by the attenuating effect on circulating catecholamines.