Hyperglycemia Increases Infarct Size in Collaterally Perfused but Not End-Arterial Vascular Territories

Abstract

Hyperglycemia exacerbates neuronal injury in the setting of reversible brain ischemia, but its effect on focal thrombotic infarction has been less extensively characterized. We investigated this problem in two rat models of focal vascular occlusion. In Model I, the right middle cerebral artery (MCA) was exposed via a subtemporal craniotomy in halothane- and nitrous oxide-anesthetized Wistar rats and was occluded photochemically by irradiation with an argon ion laser the intravenous administration of the photosensitizing dye rose bengal. Permanent MCA occlusion was combined with temporary bilateral common carotid artery ligation. In Model II, similarly anesthetized Sprague-Dawley rats were subjected to permanent photochemical occlusion of the right MCA without common carotid occlusion. In both models, rats were food deprived for 24 h and were administered varying amounts of 50% dextrose (or saline) 15 min prior to vascular occlusion to produce a spectrum of plasma glucose values, ranging from 5 to 44 μmol/ml. Brains were examined histologically 7 days following vascular occlusion, and computer-assisted planimetry was used to compute infarct volumes. In Model I, the volume of neocortical infarction ranged from 30.3 to 108.4 mm3 and exhibited a strong linear correlation with increasing preischemic plasma glucose values ( r = 0.70). In contrast, the size of the smaller striatal infarct in this model was not correlated with plasma glucose level. In Model II, there was a prominent striatal infarct, ranging in volume from 14.4 to 96.4 mm3, while neocortical infarction occurred inconstantly. As in Model I, striatal infarct volume in Model II showed no correlation with plasma glucose level. These results are consistent with the view that infarcted regions having collateral circulation [neocortex in MCA occlusion (Model I)] are vulnerable to the deleterious effects of hyperglycemia, whereas regions of nonanastomosing (end-arterial) vascular supply are not [striatum in Models I and II, and neocortex in previously reported model of photochemically induced primary microvascular thrombosis (Ginsberg et al., 1987)]. Thus, the harmful effects of elevated plasma glucose in stroke appear to be complex and may depend critically upon the degree to which collateral perfusion is available to the specific brain regions affected, as well as the extent to which local blood flow is reduced and the timing of glucose administration.