Pupose:To evaluate the results of intra-arterial urokinase thrombolysis in cases of acute ischemic stroke and to define the factors affecting prognosis. MATERIALS AND METHODS:Forty-eight patients with angiographically proven occlusion of the intracranial arteries were treated with local intra-arterial infusion of urokinase within six hours of the onset of symptoms. Neurologic status was evaluated on admis-sion and on discharge using the NIH(National Institute of Health) stroke scale score (SSS). When the SSS decreased by at least four points, this was considered indicative of an improved clinical outcome. RESULTS: Complete recanalization was achieved in 17/48 patients (35%), including 8 of 13 (62%) with occlusion of the vertebrobasilar artery (VBA), 9 of 20 (45%) with occlu-sion of the middle cerebral artery (MCA), and none of 15 with occlusion of the internal carotid artery (ICA). Neurologic status improved in 12 (60%) of patients with MCA oc-clusion, in five (38%) of those with VBA occlusion and in three (20%) of those with ICA occlusion (P<0.05). Patients in whom occluded MCA was completely recanalized showed greater clinical improvement than those with partial or no recanalization (P<0.05). The overall mortality rate was 21%, 43% (9/21) in patients in whom CT revealed signs of early infarct, but only 4% (1/27) in those without this sign (P<0.05). The mortality rate of patients with parenchymal hematoma (4/5) was higher than that of those with hemorrhagic infarct (3/9) or without hemorrhage (3/34) (P<0.05). CONCLUSION: In patients in whom occluded MCA was completely recanalized, the clinical outcome was better, while patients with VBA occlusion did not benefit from re-canalization. The presence on CT scans of signs of early infarct and of parenchymal hematoma after thrombolysis correlated with a high mortality rate.