Early Versus Late Resumption of Anticoagulation in Patients With Both High Thrombosis Risk and Major HEmoRrhage

STUDY DESIGN: This is a comparative, randomised, open study assessing after 1 month and 3 months the effect of early (H48 to 72) versus late (H120 to 144) resumption of anticoagulation in patients presenting with serious bleeding while on anticoagulants (excluding intracerebral bleeding) and with a thromboembolic risk evaluated as high (except mitral prostheses). The accumulated frequency of major hæmorrhage, thromboembolic events and deaths should be 26% in case of early resumption and 15% in case of late resumption, i.e. a relative risk reduction of 43%. Based on this hypothesis, to obtain 80% power with two-sided α being 5%, each group should include 208 patients, for a total of 416 patients. EVALUATION CRITERIA: The main criteria in this study will be the accumulated one-month incidence of hæmorrhage recurrence, thromboembolic complications and deaths. It is a combined criterion associating: * Fatal hæmorrhage proven by autopsy or sudden deaths in a clinical context strongly suggestive of hæmorrhage * Fatal thromboembolic events proven by autopsy or imagery or sudden deaths in a clinical context strongly suggestive of thrombosis * Any clinically significant hæmorrhage leading to temporary (\> 24 hours) or permanent withdrawal of anticoagulant treatment * Any symptomatic thromboembolic event in any territory, proven by imagery or surgery These events will be validated by a committee for the validation of critical events blind to the date of anticoagulant treatment resumption. The secondary evaluation criteria will be symptomatic hæmorrhages, fatal or not, symptomatic thromboembolic incidents, fatal or not, and mortality at 1 month and 3 mont