Studies in humans suggest that regions that show maximal increases in brain oxygen extraction fraction (OEF) in the hours following an ischemic episode are those most vulnerable for infarction and are often, although not always, associated with the final site of infarction. To clarify this issue, we followed the hemodynamic and metabolic characteristics of regions with an initially maximally increased OEF and compared them with the ultimately infarcted region in an experimental stroke model. Positron emission tomography (PET) was used to obtain functional images of the brain prior to and following reversible unilateral middle cerebral artery occlusion (MCAO) in 11 anesthetized baboons. To model early reperfusion, the clips were removed 6 h after occlusion. Successive measurements of regional CBF (rCBF), regional CMRO2 (rCMRO2), regional cerebral blood volume, and regional OEF (rOEF) were performed during the acute (up to 2 days) and chronic (> 15 days) stage. Late magnetic resonance imaging (MRI) scans (co-registered with PET) were obtained to identify infarction. Reversible MCAO produced an MRI-measurable infarction in 6 of 11 baboons; the others had no evidence of ischemic damage. Histological analysis confirmed the results of the MRI investigation but failed to show any evidence of cortical ischemic damage. The lesion was restricted to the head of the caudate nucleus, internal capsule, and putamen. The infarct volume obtained was 0.58 +/- 0.31 cm3. The infarcts were situated in the deep MCA territory, while the area of initially maximally increased OEF was within the cortical mantle. The mean absolute rCBF value in the infarct region of interest (ROI) was not significantly lower than in the highest-OEF ROI until 1-2 days post-MCAO. Cerebral metabolism in the deep MCA territory was always significantly lower than that of the cortical mantle; decreases in CMRO2 in the former region were evident as early as 1 h post-MCAO. In the cortical mantle, the rOEF was initially significantly higher than in the infarct-to-be zone. Subsequently, the OEF declined in both regions. The differences in the time course of changes in CMRO2 and OEF between these two regions, with the eventually infarcted area showing earlier metabolic degradation and in turn decline in OEF, presumably underlie their different final outcomes. In conclusion, following MCAO, the region that shows an early maximal increase in the OEF is both topographically and physiologically distinct from the region with final consolidated infarction if reperfusion is allowed at 6 h. This high OEF, although indicative of a threatened condition, is not an indicator of inescapable consolidated infarction and is thus a situation in which therapy could be envisaged. Whether or not it is at risk of infarction and thus constitutes one target for therapy remains to be seen.