Bleeding Risk with Apixaban vs. Rivaroxaban in Acute Venous Thromboembolism (COBRRA)

Authors, Journal, Affiliations, Type, DOI

• Castellucci LA, Chen VM, Kovacs MJ, Lazo-Langner A, Greenstreet P, Kahn S, Côté B, Schulman S, de Wit K, Douketis J, Suryanarayan D, Wan T, Yeo E, Le Templier G, Tran HA, Willcox A, Crowther HJ, Prasad R, Shivakumar S, Umana E, Ni Ainle F, Tritschler T, Barco S, Galanaud J-P, Blondon M, Baumann Kreuziger L, Solymoss S, Kearon C†, Thomas E, Ramsay T, Le Gal G, Rodger M; for the COBRRA Trial Investigators
• Journal: New England Journal of Medicine 2026;394:1051–60
• Affiliations: Ottawa Hospital Research Institute (lead); 32 sites in Canada, Australia, and Ireland
• Type: Prospective, randomized, open-label, blinded endpoint (PROBE design) superiority RCT
• DOI: https://doi.org/10.1056/NEJMoa2510703
• Funding: Canadian Institutes of Health Research; Medical Research Future Fund (Australia); Royal College of Surgeons in Ireland; International Network of Venous Thromboembolism Clinical Research Networks. No industry role in design, conduct, or analysis.

Overview

The COBRRA trial is the first head-to-head RCT directly comparing apixaban and rivaroxaban — the two most widely used NOACs for acute VTE — with respect to bleeding risk over a 3-month treatment period. Among 2,760 patients with acute symptomatic pulmonary embolism or proximal DVT, clinically relevant bleeding occurred in 3.3% on apixaban versus 7.1% on rivaroxaban (RR 0.46; P<0.001), representing a 54% relative risk reduction. VTE recurrence rates were equivalent (1.1% vs 1.0%). Prior guidelines did not recommend one agent over the other; these data provide a direct evidence base to prefer apixaban for bleeding safety.

Keywords

Venous thromboembolism, deep-vein thrombosis, pulmonary embolism, direct oral anticoagulants, apixaban, rivaroxaban, bleeding, major bleeding, ISTH criteria, PROBE design

Key Takeaways

Trial Design and Population

• Design: PROBE (prospective randomized open-label blinded endpoint); superiority trial; industry-independent; 32 centres, 3 countries (Canada, Australia, Ireland); December 2017–January 2025
• Population: Adults ≥18 years with symptomatic acute proximal lower-limb DVT or segmental/more-proximal PE; prior therapeutic anticoagulation ≤72 hours permitted
• Exclusions: Active cancer, CrCl <30 ml/min (Cockcroft-Gault), weight >120 kg (ISTH guidance at time of design), hepatic disease (Child-Pugh B/C), contraindicated interacting medications, pregnancy/breastfeeding, AF or other indication for long-term anticoagulation
• Randomization: 1:1, centralized, permuted blocks (sizes 4 and 6), stratified by renal function (CrCl <50 vs ≥50 ml/min), continued antiplatelet use, and participating center
• n = 2,760 randomized: 1,370 apixaban / 1,390 rivaroxaban; ITT n = 1,345 / 1,355 after 60 (2.2%) post-randomization exclusions

Dosing Regimens

• Apixaban: 10 mg twice daily × 7 days → 5 mg twice daily × remaining ~82 days
• Rivaroxaban: 15 mg twice daily × 21 days → 20 mg once daily × remaining ~68 days
• Key difference: rivaroxaban loading dose (15 mg BID) is 50% higher than its maintenance dose; most bleeding difference appears to cluster in the first 21 days

Baseline Characteristics

• Mean age 58.3 years; 43.5% female; 10% non-White
• 77.3% unprovoked VTE; 52.2% DVT alone; 47.8% PE ± DVT; 15.9% prior VTE
• Groups well balanced

Primary Outcome: Clinically Relevant Bleeding

• Apixaban 3.3% vs Rivaroxaban 7.1% (RR 0.46; 95% CI 0.33–0.65; P<0.001)
• Adjusted marginal RR in sensitivity analysis: 0.45 (95% CI 0.32–0.64) — fully consistent
• Benefit consistent across all prespecified subgroups

Major Bleeding

• Apixaban 0.4% vs Rivaroxaban 2.4% (RR 0.16; 95% CI 0.06–0.40) — 84% reduction
• All 5 apixaban major bleeds: haemoglobin drop or transfusion
• 32 rivaroxaban major bleeds: 28 haemoglobin drop/transfusion; 4 critical site/organ
• No fatal bleeding in either group

Clinically Relevant Nonmajor Bleeding (CRNMB)

• Apixaban 2.9% vs Rivaroxaban 4.9% (RR 0.59; 95% CI 0.40–0.86)
• Most common apixaban: vaginal bleeding (2.7%), GI (0.6%)
• Most common rivaroxaban: vaginal bleeding (3.8%), hematuria (1.3%), GI (1.0%)

Secondary Outcomes

• Recurrent symptomatic VTE: 1.1% vs 1.0% (RR 1.08; 95% CI 0.52–2.23) — no difference
• All-cause mortality: 0.1% vs 0.3% (RR 0.25; 95% CI 0.03–2.26) — NS, not powered
• Serious adverse events (unrelated to bleeding/VTE): 2.7% vs 2.2% — comparable

Medication Adherence

• Complete adherence: apixaban 65.7% vs rivaroxaban 75.1% — paradoxically lower in apixaban arm, yet VTE recurrence identical; apixaban twice-daily dosing may reduce adherence

Mechanistic Interpretation

• Bleeding difference predominantly in first 3 weeks when rivaroxaban loading dose is 50% above maintenance; rivaroxaban's higher initial anticoagulant exposure likely drives excess early bleeding
• Cannot exclude pharmacodynamic differences between factor Xa inhibitors beyond dosing

Limitations of the Document

• Open-label design: Ascertainment bias possible, though unlikely given overt nature of qualifying bleeding events and blinding of adjudication committee
• 3-month follow-up only: Whether bleeding advantage persists with extended treatment is unknown
• Weight >120 kg excluded: Limits applicability to patients with severe obesity
• Cancer excluded: Results do not apply to cancer-associated VTE
• Limited racial diversity: ~10% non-White
• Not powered for VTE recurrence: Secondary outcome only; 1.08 RR with wide CI does not rule out meaningful difference
• Adherence paradox: Lower adherence in apixaban arm could attenuate its bleeding advantage; recurrence remained equivalent despite lower adherence

Key Concepts Mentioned

• concepts/Venous-Thromboembolism-Anticoagulation — head-to-head NOAC comparison; apixaban superior for bleeding
• concepts/Cancer-Associated-VTE — cancer excluded from COBRRA; separate evidence base

Key Entities Mentioned

• Apixaban — NOAC, factor Xa inhibitor; favoured in COBRRA for bleeding safety
• Rivaroxaban — NOAC, factor Xa inhibitor; comparator

Wiki Pages Updated

• Created wiki/sources/NOAC-VTE-COBRRA-NEJM-2026.md
• Created wiki/concepts/Venous-Thromboembolism-Anticoagulation.md
• Updated wiki/wikiindex.md
• Updated wiki/sourceindex.md