Intravenous (i.v.) recombinant tissue plasminogen activator (rt-PA) should be available on a 24/7 basis in hospitals admitting patients with stroke. We aimed at evaluating the influence of the number of patients previously treated with i.v. rt-PA by neurologists on patients' outcome. For each patient consecutively treated with i.v. rt-PA for cerebral ischaemia at the Lille University Hospital, we determined the number of previous treatments with rt-PA administered by the neurologist. We performed logistic regression analyses to determine the influence of the experience on the outcome evaluated by the modified Rankin scale (mRS) after 3 months, 0-1 meaning independence, and 0-2 meaning absence of handicap. We compared outcomes of patients treated by the 25% less experienced neurologists with those of trials. Forty-four neurologists treated 800 patients. The experience of the treating neurologist was independently associated with independence (adjusted odds ratio [(adj)OR] 1.062 for 10 patients more; 95% confidence interval [CI] 1.008-1.120), and absence of handicap ((adj)OR 1.076 for 10 patients more; 95%CI 1.016-1.140) at 3 months, but not with symptomatic intracerebral haemorrhage and death. The proportions of patients from the 1st quartile who were independent and without handicap at 3 months were 37.9 and 51.1%. Patients treated by less experienced neurologists, have similar outcomes than expected from trials, suggesting they benefit from thrombolysis. However, the outcome of patients treated by more experienced neurologists was slightly better. Less experienced neurologists should not be excluded from rt-PA programmes, but their practices should be evaluated and educational programmes organised.