The efficacy and safety of aspirin in the prevention and treatment of thrombosis in essential thrombocythaemia (ET) was retrospectively analysed in a cohort of 68 ET patients. 41 patients presented with thrombosis, five patients with bleeding: two patients had a paradoxical combination of bleeding and thrombosis at presentation. At presentation, patients with bleeding had significantly higher platelet and leucocyte counts than patients with thrombosis. During long-term follow-up the incidence of thrombosis was significantly reduced in patients receiving aspirin, either as monotherapy or in combination with cytoreduction. However, treatment with aspirin (500 mg/d) was associated with an increase in (minor) bleeding complications. In patients receiving aspirin, bleeding occurred particularly at platelet counts exceeding 1000 x 10(9)/l. The overall 5- and 10-years survival probability was 93% and 84% respectively, indicating that life expectancy in ET is close to normal. Although our data need confirmation in prospective clinical trials, they suggest that aspirin, particularly in lower doses (100 mg/d), may be a safe antithrombotic agent in ET with an acceptable risk for bleeding, if applied to patients with a platelet count <1000 x 10(9)/l and/or absence of a bleeding history.