A canine model of cortical vein occlusion was used to evaluate whether data obtained from monitoring venous stump pressure could help predict cerebral infarction after venous obstruction. Following bilateral parasagittal craniotomy, the cortical vein in each hemisphere was temporarily occluded and the increase in pressure was directly measured. Permanent venous obstruction was subsequently produced, and parenchymal brain damage 24 hours later was classified as: Stage 0, no parenchymal damage; Stage I, mild edema; Stage II, moderate parenchymal edema and/or ischemic changes in neurons; and Stage III, moderate-to-severe hemorrhage. The histological stages correlated closely with the rise in venous pressure: mean pressure increases (+/- standard deviation) were 5.5 +/- 2.9 mm Hg in hemispheres graded as Stage 0 (12 hemispheres), 7.7 +/- 3.2 mm Hg in those graded as Stage I (five), 11.2 +/- 4.1 mm Hg in those classed as Stage II (five), and 16.4 +/- 5 in those categorized as Stage III (seven). There were significant differences between Stages 0 and II (p < 0.01) and between Stages 0 and III (p < 0.001). Disruption of the blood-brain barrier as indicated by extravasation of Evans blue dye correlated well with the pressure increment. These results may indicate the threshold for injury after cortical venous occlusion. Venous stump pressure measurements obtained during a test occlusion may be a useful adjunct in predicting brain damage and may be helpful for intraoperative vessel selection for venous resection.