Background: Treatment of pulmonary embolism with low-molecular-weight heparin and vitamin K antagonists, such as warfarin, is not ideal. We aimed to assess non-inferiority of idrabiotaparinux, a reversible longlasting indirect inhibitor of activated factor X, to warfarin in patients with acute symptomatic pulmonary embolism. Methods: In our randomised, double-blind, double-dummy, non-inferiority trial, we enrolled adults with objectively documented acute symptomatic pulmonary embolism attending 291 centres in 37 countries. We excluded patients who were pregnant, had active bleeding, kidney failure, or malignant hypertension, or were at high risk of death, bleeding, or adverse reactions to study drugs. We randomly allocated patients to receive 5-10 days' enoxaparin 1·0 mg/kg twice daily followed by subcutaneous idrabiotaparinux (starting dose 3·0 mg) or adjusted-dose warfarin (target international normalised ratio 2·0-3·0); regimens lasted 3 months or 6 months dependent on clinical presentation. Block randomisation was done with a central interactive computerised system, stratified by study centre and intended treatment duration. The primary efficacy outcome was recurrent venous thromboembolism at 99 days after randomisation. We estimated the odds ratio and 95 CI with a Mantel-Haenzsel χ 2 analysis (non-inferiority margin 2·0) in the intention-to-treat population. The main safety outcome was clinically relevant bleeding (major or non-major) in all patients at day 99. This study is registered with ClinicalTrials.gov, number NCT00345618. Findings: Between Aug 1, 2006, and Jan 31, 2010, we enrolled 3202 patients aged 18-96 years. 34 (2) of 1599 patients randomly allocated to receive enoxaparin-idrabiotaparinux and 43 (3) of 1603 patients randomly allocated to receive enoxaparin-warfarin had recurrent venous thromboembolism (odds ratio 0·79, 95 CI 0·50-1·25; p non-inferiority= 0·0001). 72 (5) of 1599 patients in the enoxaparin-idrabiotaparinux group and 106 (7) of 1603 patients in the enoxaparin-warfarin group had clinically relevant bleeding (0·67, 0·49-0·91; p superiority=0·0098). We noted similar differences in outcomes in those patients treated to 6 months. Interpretation: Idrabiotaparinux could provide an attractive alternative to warfarin for the long-term treatment of pulmonary embolism, and seems to be associated with reduced bleeding. Funding: Sanofi-Aventis (Paris, France).