The authors retrospectively analyzed the initial characteristics, treatment, and clinical course in 147 patients with essential thrombocythemia (ET). Median age was 60 years and the M:F ratio was 0.69. At diagnosis, 53 patients were asymptomatic; 50 patients had functional symptoms (mainly vasomotor disturbances); 27 patients had large vessel thrombosis; 27 patients had a bleeding diathesis; and seven patients had both bleeding and thrombosis. The platelet count ranged from 0.7 to 2.92 X 10(12)/l. Forty-five of the 61 tested patients (61%) had prolonged bleeding time and/or platelet hypoaggregation. Three patients had in vitro spontaneous aggregation. No significant correlations were found between hemostatic findings and in vivo bleeding or thrombosis. The incidence of bleeding, however, was higher in patients with more than 2 x 10(12)/l platelets. Of 87 karyotypes performed with banding techniques, only four were abnormal. One hundred twenty-nine patients received one or more cytoreductive agents at diagnosis or during follow-up. Sixty patients received an antiaggregating agent. First-line therapy was radiophosphorus (32P) in 22 patients; busulfan in 35 patients; and hydroxyurea in 72 patients. Hydroxyurea required continuous maintenance therapy and had to be changed to another treatment in 12 of the initial responders because of inadequate control of thrombocythemia. During follow-up, 14 treated patients experienced one or several major thrombotic events. Two untreated patients also had major thrombosis. Only one major bleeding event was seen during follow-up. Median actuarial survival was 73.5% at 7 years and only one patient progressed to acute non-lymphocytic leukemia (ANLL). These results suggest that large vessel thrombosis is the main complication of ET. It appears largely unpredictable in a given patient at diagnosis but can be largely prevented by the control of thrombocythemia. Because of the low incidence of side effects of treatment in this experience, the authors believe that cytoreductive therapy is indicated in most patients with ET, as long as a group of patients with very low risk of thrombosis is not defined in prospective studies.