Background The optimal time window for the administration of hypothermia following cerebral ischemia has been studied for decades,with disparity outcomes.In this study,the efficacy of mild brain hypothermia beginning at different time intervals on brain endogenous antioxidant enzyme and energy metabolites was investigated in a model of global cerebral ischemia.Methods Forty-eight male Sprague-Dawley rats were divided into a sham-operated group,a normothermia (37℃-38℃) ischemic group and a mild hypothermic (31℃-32℃) ischemia groups.Rats in the last group were subdivided into four groups:240 minutes of hypothermia,30 minutes of normothermia plus 210 minutes of hypothermia,60 minutes of normothermia plus 180 minutes of hypothermia and 90 minutes of normothermia plus 150 minutes of hypothermia (n=8).Global cerebral ischemia was established using the Pulsinelli four-vessel occlusion model for 20minutes and mild hypothermia was applied after 20 minutes of ischemia.Brain.tissue was collected following 20 minutes of cerebral ischemia and 240 minutes of reperfusion,and used to measure the levels of superoxide dismutase (SOD),glutathione peroxidase (GSH-Px),reduced glutathione (GSH) and adenosine triphosphate (ATP).Results Mild hypothermia that was started within 0 to 60 minutes delayed the consumption of SOD,GSH-Px,GSH,and ATP (P ＜0.05 or P ＜0.01) in ischemic tissue,as compared to a normothermic ischemia group.In contrast,mild hypothermia beginning at 90 minutes had little effect on the levels of SOD,GSH-Px,GSH,and ATP (P＞0.05).Conclusions Postischemic mild brain hypothermia can significantly delay the consumption of endogenous antioxidant enzymes and energy metabolites,which are critical to the process of cerebral protection by mild hypothermia.These results show that mild hypothermia limits ischemic injury if started within 60 minutes,but loses its protective effects when delayed until 90 minutes following cerebral ischemia.