This multicentre open-label randomized parallel-group trial aims to evaluate the association intermittent pneumatic compression + elastick stockings versus elasting stockings alone on symptomatic or asymptomatic venous thromboembolism incidence, evaluated systematically at day 6 (+/-2days), in patients hospitalized in intensive care units and with high bleeding risk.

Background: Venous thromboembolism is a leading cause of morbi-mortality for patients hospitalized in intensive medical care units. Nevertheless, few studies evaluating prophylactic methods exist. Data from these studies demonstrate however that unfractionated heparin and low molecular weight heparin are effective to reduce the incident rate of asymptomatic deep venous thrombosis (DVT) of the lower limbs by 50 % compared to the absence of prophylaxis. Rate of asymptomatic DVT remains about 15 %. When the bleeding risk is high, drug prophylaxis with anticoagulants is contra-indicated, and mechanical devices are recommended: elastic stockings (ES) alone or associated with intermittent pneumatic compression (IPC). However, mechanical devices have not been systematically evaluated in intensive medical care units.

Objective:

To compare the association IPC + ES to ES alone in patients with high bleeding risk and hospitalized in intensive medical care units, on symptomatic or asymptomatic venous thromboembolism incidence, evaluated systematically at day 6.

Design: Multicentre open-label randomized parallel-group trial with blinded evaluation of outcomes, and an inclusion period of 24 month. Brest CIC (Centre d'Investigations Cliniques, research center) coordinates this multicentre trial.

Outcomes:

The primary endpoint is a combined criterion (blindly evaluated) between day 1 and day 6:

1. Non fatal symptomatic venous thromboembolic event (objectively confirmed) between day 1 and day 6,
2. death due to a pulmonary embolism between day 1 and day 6, and 3) asymptomatic DVT (distal or proximal) detected by ultrasonography systematically done at day 6.

Patients number:

CIREA 1 : Assuming a DVT frequency of 15 % in the control group (ES alone) , we calculated that 356 patients will be required for the study to have 80% power to detect a 60% reduction in the relative risk with a two-sided alpha level of 5%. Because of about 20 % deaths in the first days, we decide to increase the number at 392 subjects.

Statistical analysis: Efficacy analysis is performed on an 'intention-to-treat' basis. The frequencies of the combined primary outcome at day 6 are compared between groups using an exact one-sided Fischer test. Adjustment for stratification variables used Cochran-Mantel-Haenszel test. Relative risk and absolute risk reduction are computed with their 95% CIs. A logistic regression model is used to take into account potential imbalance in baseline characteristics.

Inclusion Criteria:

• Age > 18 Years,
• Admission in intensive medical care unit
• High risk for hemorrhage
• Written informed consent given by the patient or relative.

High risk for hemorrhage is defined by:

• symptomatic bleeding or organic lesions likely to bleed,
• hemophilic diseases,
• haemostatic abnormalities: platelet count < 50 000/mm 3, APTT > 2 N, prothrombin time < 40%,
• recent intra-cerebral hemorrhage,
• severe anemia (Hemoglobin < 7 g/dl) due to a bleeding or unexplained.

Exclusion Criteria:

• Age < 18 years,
• Patient refusal,
• No high risk for hemorrhage
• Admission in intensive care unit ≥ 36 hours
• Admission in intensive care unit likely for < 72 hours
• A "do not resuscitate" order
• IPC contra-indication: Recent DVT (< 3 month), severe lower limbs arteriopathy (III and IV), any arterial graft of the legs, a wound in the lower limb related either to a vascular disease (ulcer) or a trauma.
• Patient with mechanical prosthetic heart valve.