Angiographic studies performed within 6 hours of stroke onset have demonstrated that 75-80% of patients with an acute ischemic stroke have an angiographically visible occlusion of an extracranial and/or intracranial artery that is the cause of the ischemic stroke. The NINDS t-PA Stroke Study demonstrated that recanalization of occluded brain arteries can successfully salvage ischemic brain if intravenous tissue plasminogen activator (t-PA) is initiated within 3 hours of stroke onset. The effectiveness and safety of intravenous t-PA beyond 3 hours has yet to be shown. Large clots in the internal carotid artery, vertebral or basilar artery, or middle cerebral artery trunk are infrequently lysed by intravenous t-PA during the first hour of treatment. Intraarterial delivery of a thrombolytic agent appears to lyse clots in cerebral arteries more effectively than intravenous t-PA. However, a head-to-head comparison of full-dose intravenous t-PA and intraarterial therapy has yet to be done, and recanalization rates reported in the intraarterial studies are at longer times after the start of therapy than the rates of recanalization reported in intravenous t-PA studies. Intraarterial delivery of thrombolytic therapy takes longer than intravenous delivery, and even intraarterial delivery takes 2 hours on average to partially or completely lyse 75% of clots. Intraarterial delivery of a thrombolytic agent has yet to be proven effective and safe in improving long-term outcome of ischemic stroke patients. The approach of combining intravenous and intraarterial delivery of t-PA or other pharmacologic agents is an attractive option that deserves further study. Mechanical lysis of clots, in combination with pharmacologic therapies, is a more effective way to recanalize coronary arteries and salvage heart muscle than intravenous thrombolytic therapy alone. This approach is likely to be effective in ischemic stroke as well. What is needed is advances in catheter technology that meet the special requirements of clot lysis in cerebral vessels. Whether neuroprotective agents can extend the therapeutic window for effective clot lysis and recanalization remains to be seen. Thrombolytic therapy and other pharmacologic treatments of clot in cerebral vessels will likely remain a two-edged sword. Pharmacologic therapies that increase the likelihood of clot lysis and recanalization, such as thrombolytic agents, the platelet GIIbIIIa receptor blockers, defibrinogenating agents, and even the newer more potent thrombolytic agents, also concomitantly increase the risk of bleeding into the brain. What we will be searching for in the coming decade is the correct mechanical strategy, dose of a given pharmacologic agent, or combination of agents that maximizes recanalization and minimizes the risk of intracerebral hemorrhage.