With the possible exception of aspirin, there are still no drug treatments which have been found to be so effective in randomised controlled trials, and meta-analyses of such trials, that they should be used routinely in acute ischaemic stroke. The evidence on thrombolysis is tantalising but with only about 3500 randomised patients in all the trials put together it is too early to say how, and for whom, this promising treatment should be used, particularly bearing in mind the risk of intracranial haemorrhage; the evidence so far is enough to persuade us to enter our patients into further large trials of thrombolysis but it is not enough to justify this treatment in our routine clinical practice.