In case of unprovoked venous thromboembolism (VTE), the screening of thrombophilia is recommended whatever the age of the patient and the type of risk factors (RF). This prospective study was conducted in patients with unprovoked VTE to detect some predictive factors to have a higher risk of thrombophilia, focusing on age, history of venous thromboembolism, and the existence of a triggering event. From July 2000 to July 2002, in an Internal Medicine Department, unrelated patients with unprovoked VTE were included. Those unprovoked thromboembolic events were defined by the absence of association between permanent and transient RF. The primary outcome measure was the positivity of the thrombophilia screening for any type of abnormality detected (deficit of protein C, S, antithrombin, presence of a lupus anticoagulant, research of V and II mutations). Seventy-four patients were included. Eight died during the follow-up. A higher risk of thrombophilia was found in patients younger than 40 (p=0.03), or with a family but not personal history of VTE (p=0.01) or with transient RF (p=0.02). The most frequent abnormality of coagulation found in patients younger than 40 was the presence of a lupus anticoagulant. As a new strategy for the screening of thrombophilia, one could propose the following attitude: only patients with transient RF or family history of VTE could undergo a complete screening; for all the remaining patients who are younger than 40, a research of a lupus anticoagulant would be only performed. This strategy should now be balanced against the currently recommended systematic attitude in further studies.