Approximately 80 to 90% of cerebral ischaemic events that occur within 24 hours of symptom onset are due to atherothrombotic or thromboembolic occlusions. This forms the rationale for the use of thrombolytic agents in patients with acute ischaemic stroke. Early studies determined that recanalisation occurred in approximately 21 to 72% of patients with occluded cerebral arteries after intra-arterial or intravenous administration of streptokinase, urokinase, alteplase (recombinant tissue-type plasminogen activator; rt-PA) or duteplase (a 2-chain rt-PA). Initial reports suggested that frequencies of haemorrhagic transformation and parenchymatous haematoma in the carotid territory were similar whether patients with middle cerebral artery stroke received thrombolysis via intra-arterial or intravenous administration. The Multicentre Acute Stroke Trial-Europe (MAST-E), the Australia Streptokinase (ASK), and the Multicentre Acute Stroke Trial-Italy (MAST-I) trials, which evaluated intravenous streptokinase 1.5 x 10(6) IU in patients with acute ischaemic stroke, were terminated prematurely because of excessive early mortality and symptomatic intracranial haemorrhage in streptokinase recipients compared with those treated with placebo. However, those studies had not been preceded by dose-ranging trials. Intravenous administration of alteplase 0.9 mg/kg within 3 hours [National Institute of Neurological Disorders and Stroke (NINDS) trial], or 1.1 mg/kg within 6 hours [European Cooperative Acute Stroke Study (ECASS)], of symptom onset in patients with acute ischaemic stroke resulted in an absolute 11 to 13% treatment-associated improvement in clinical measurement scales; such as the modified Rankin scale and Barthel index, compared with placebo recipients. In the ECASS trial, those results were limited to a 'target population' restricted to those who satisfied all entry criteria. In both trials, the frequency of symptomatic haemorrhage was greater in patients treated with alteplase than with placebo and reinforced the importance of careful patient selection. Strict patient selection remains central to the success of this approach.