Recurrent venous thromboembolism and bleeding during anticoagulation in patients with deep-vein thrombosis treated with vitamin K antagonists or direct oral anticoagulants: insights from the RIETE registry
Direct oral anticoagulants were linked to fewer recurrent clots and major bleeds, but more bothersome non-major bleeding.
*Prospective registry cohort study; Level 2 (OCEBM).
Citation
Prandoni P, Pesavento R, Bilora F, et al. Recurrent venous thromboembolism and bleeding during anticoagulation in patients with deep-vein thrombosis treated with vitamin K antagonists or direct oral anticoagulants: insights from the RIETE registry. European Journal of Internal Medicine. 2026. doi:10.1016/j.ejim.2026.107120
Background
Prior randomized trials found direct oral anticoagulants as effective as vitamin K antagonists for venous clots, with less major bleeding. This study tested whether those results hold in routine care, including patients often excluded from trials.
Patients
24,728 adults with confirmed symptomatic lower-limb deep-vein thrombosis, with or without lung clot symptoms. Exclusions included isolated lung clots, clot sites outside the lower limb, pregnancy, age under 18 years, and anticoagulant use before diagnosis.
Intervention
Therapeutic-dose direct oral anticoagulants: rivaroxaban, apixaban, edoxaban, or dabigatran.
Control
Therapeutic-dose vitamin K antagonists.
Outcome
Recurrent venous clot, major bleeding, and clinically relevant non-major bleeding.
Follow-up Period
During active full-dose anticoagulant treatment only; median treatment was 5.1 months with direct oral agents and 6.7 months with vitamin K antagonists.
Results
| Outcome | Direct oral anticoagulants | Vitamin K antagonists | Adjusted effect |
|---|---|---|---|
| Recurrent clot | 1.81 per 100 patient-years | 2.12 per 100 patient-years | HR 0.71 (95% CI 0.58 to 0.87) |
| Major bleeding | 1.50 per 100 patient-years | 1.67 per 100 patient-years | HR 0.78 (95% CI 0.63 to 0.97) |
| Clinically relevant non-major bleeding | 4.86 per 100 patient-years | 3.08 per 100 patient-years | HR 1.35 (95% CI 1.17 to 1.55) |
HR = hazard ratio. Approximate absolute effects from crude rates: about 323 patient-years treated to prevent one recurrent clot, 588 to prevent one major bleed, and 56 to cause one non-major bleed.
Results were adjusted observational analyses during treatment. Trial-eligible patients had clearer major bleeding benefit; trial-ineligible patients did not.
Limitations
Non-randomized treatment choice may leave residual bias. Warfarin control quality was unavailable. Trial eligibility was approximate, follow-up ended at treatment stopping, and isolated lung clots were excluded.
Funding
Unrestricted educational grant from ROVI; potential industry-support bias.
Clinical Application
Use direct oral anticoagulants first for most lower-limb thrombosis patients, but individualize in high-risk patients and warn about non-major bleeding.
Discussion
Sign in to join the discussion.
In this prospective registry cohort of lower-limb DVT, DOACs versus VKAs were associated with lower recurrent VTE (HR 0.71) and major bleeding (HR 0.78) but higher CRNMB (HR 1.35); because allocation was nonrandomized, how should validity and applicability affect practice change? The study reports an adjusted hazard ratio of 0.71 for recurrent VTE with DOACs compared with VKAs during anticoagulation. Which interpretation is most appropriate?