@ERSpublicationsBeside scores, doctor-patient relationship still key to assess individual benefit of extended anticoagulation http://ow.ly/uWaL303tb77Venous thromboembolism (VTE) is a common, multifactorial and capricious disease. Its incidence is estimated at between 1 and 1.84 per 1000 person-years, half of cases consisting of isolated distal or proximal deep vein thrombosis (DVT), a quarter of isolated pulmonary embolism (PE) and the rest being PE associated with DVT [1, 2]. The direct severity depends on the mortality risk due to haemodynamic instability in massive PE, and risk of mid-or long-term thromboembolism recurrence and associated sequelae [3]. Population-based studies have revealed that the recurrence risk is significant in the first days, intermediate during the next 3-6 months, then falls to approximately 3-5% per year, though varies widely depending on the thromboembolism context and patient characteristics [1, 3]. We can thus distinguish three therapeutic stages: initial treatment (acute phase), maintenance treatment (consolidation phase) and, at least for some patients, long-term treatment (secondary prevention phase) [4][5][6].The emergence of direct oral anticoagulants (DOACs) has made a profound difference in each stage of the process. Initially prescribed orally (rivaroxaban and apixaban) or following a few days of administering parenteral anticoagulant (dabigatran and edoxaban), these agents enable avoiding overlapped treatment phases, which carries a particular bleeding risk with vitamin K antagonists (VKAs) [7]. They also offer greater, wider treatment windows than VKAs, thus generally proving noninferior to LMWH/VKA combinations in terms of efficacy, yet superior in terms of safety. This positive effect was confirmed in a meta-analysis reviewing five studies and involving nearly 25 000 patients, reporting a 40% reduction in relative risk (RR) of major bleeding (RR 0.60, 95% CI 0.41-0.88) and >60% reduction in risk of fatal bleeding (RR 0.36, [8]. Nevertheless, managing bleeding used to be a serious concern until consensus approaches were established [9,10] and antidotes were discovered [11,12]. The latter are now recommended as first-line in the latest international guidelines for acute VTE [5]. They are also indicated in secondary prevention, though this is more complex due to the very fine line between thrombosis and bleeding risks in this context [13]. Two approaches are therefore necessary to define which patients are likely to benefit from long-term DOACs as secondary prevention: identifying first those at high risk of recurrence, then those at low bleeding risk.The study by KLOK et al. [14], published in this issue of the European Respiratory Journal, addresses this using data from two randomised studies comparing dabigatran with warfarin in the management of proximal acute VTE. In those patients randomly assigned to the dabigatran arm, six factors were shown to be independent predictors of severe bleeding: active cancer, being male with uncontrolled hypertension, anaemia, histor...