Background Local hypothermia induced by intravascular infusion of cold saline solution effectively reduces brain damage in stroke. We further determined the optimal temperature of local hypothermia in our study. Methods Seventy-eight adult male Sprague Dawley rats (260-300 g) were randomly divided into 3 groups: group A, ischemia/reperfusion without cold saline infusion (n=26) (control group); group B, infusion with 20 degrees C saline before reperfusion (n=26); group C: infusion with 10 degrees C saline before reperfusion (n=26). In each group, we chose 15 rats for monitoring physical indexes and the temperature of the brain (cortex and striatum) and body (anus), measurement of brain infarction volume, assessment of neurological deficits and the survival rate of reperfusion at 48 hours. Another 8 rats from each group was chosen for examining brain edema, another 3 from each group for histological observation by electron microscopy (EM) and light microscopy (LM) at 48 hours after reperfusion. Results There was no significant difference among the 3 groups for physical indexes during the examination (F-(2,F-45)= 0.577, P=0.568; F-(2,F-45)= 0.42, P=0.78 for blood pressure and blood gas analysis, respectively). The brain temperature was significantly reduced in the group C compared to the other groups (F-(2,F-45)=37.074, P=0.000; F-(2.45)=32.983, P=0.000, for cortex and striatum temperature respectively), while the difference in rectal temperature between group A and B or C after reperfusion was not significant (F-(2,F-45)= 0.17115, P=0.637). And the brain infarct volume was significantly reduced in group C (from 40%+/- 10% in group A, 26%+/- 8% in group B, to 12%+/- 6% in group C, F-(2,F-45)=43.465, P=0.000) with the neurological deficits improving in group C (X-2 = 27.626, P=0.000). The survival rate at 48 hours after 10 degrees C and 20 degrees C saline reperfusion was increased by 132.5% and 150%, respectively, as compared to the control group (X-2=10.489, P=0.005). The extent of the brain edema showed no significant difference (F-(2,F-21)=0.547, P=0.587) after cold saline infusion compared to the control group. No obvious vascular injury was found by electron or light microscopy in either infusion group. Conclusions Regional hypothermia with 10 degrees C cold saline infusion can significantly decrease the infarction volume, improve the neurological deficits, and 10 degrees C seems to be the optimal temperature in inducing a cerebral protection effect during stroke. This procedure could be adopted as a further treatment for acute stroke patients. Chin Med J 2009;122(13):1558-1563