We have tested whether small intraischemic variations in brain temperature influence the outcome of transient ischemia. To measure brain temperature, a thermocouple probe was placed stereotaxically into the left dorsolateral striatum of rats prior to 20 min of four-vessel occlusion. Rectal temperature was maintained at 36-37 degrees C by a heating lamp, and striatal temperature prior to ischemia was 36 degrees C in all animals. Six animal subgroups were investigated, including rats whose intraischemic striatal brain temperature was not regulated, or was maintained at 33, 34, 36, or 39 degrees C. Postischemic brain temperature was regulated at 36 degrees C, except for one group in which brain temperature was lowered from 36 degrees C to 33 degrees C during the first hour of recirculation. Energy metabolites were measured at the end of the ischemic insult, and histopathological evaluation was carried out at 3 days after ischemia. Intraischemic variations in brain temperature had no significant influence on energy metabolite levels measured at the conclusion of ischemia: Severe depletion of brain ATP, phosphocreatine, glucose, and glycogen and elevation of lactate were observed to a similar degree in all experimental groups. The histopathological consequences of ischemia, however, were markedly influenced by variations in intraischemic brain temperature. In the hippocampus, CA1 neurons were consistently damaged at 36 degrees C, but not at 34 degrees C. Within the dorsolateral striatum, ischemic cell change was present in 100% of the hemispheres at 36 degrees C, but in only 50% at 34 degrees C. Ischemic neurons within the central zone of striatum were not observed in any rats at 34 degrees C, but in all rats at 36 degrees C. In rats whose striatal temperature was not controlled, brain temperature fell from 36 to 30-31 degrees C during the ischemic insult. In this group, no ischemic cell change was seen within striatal areas and was only inconsistently documented within the CA1 hippocampal region. These results demonstrate that (a) rectal temperature unreliably reflects brain temperature during ischemia; (b) despite severe depletion of brain energy metabolites during ischemia at all temperatures, small increments of intraischemic brain temperature markedly accentuate histopathological changes following 3-day survival; and (c) brain temperature must be controlled above 33 degrees C in order to ensure a consistent histopathological outcome. Lowering of the brain temperature by only a few degrees during ischemia confers a marked protective effect.