New England Journal of Medicine September 08, 2011

Rivaroxaban versus Warfarin in Nonvalvular Atrial Fibrillation

Manesh R. Patel, Kenneth W. Mahaffey, Jyotsna Garg, et al.

Bottom Line

In patients with nonvalvular atrial fibrillation at increased risk for stroke, rivaroxaban was noninferior to warfarin for the prevention of stroke and systemic embolism, with comparable overall bleeding rates but significantly fewer intracranial and fatal hemorrhages.

Key Findings

1. In the per-protocol, as-treated primary analysis, the primary end point (stroke or systemic embolism) occurred at 1.7% per year with rivaroxaban vs. 2.2% per year with warfarin (HR 0.79; 95% CI, 0.66 to 0.96; P<0.001 for noninferiority) [3.4].

2. In the intention-to-treat analysis, the primary end point occurred at 2.1% per year with rivaroxaban vs. 2.4% per year with warfarin (HR 0.88; 95% CI, 0.74 to 1.03; P<0.001 for noninferiority; P=0.12 for superiority).

3. The principal safety end point (major and non-major clinically relevant bleeding) occurred at 14.9% per year with rivaroxaban vs. 14.5% per year with warfarin (HR 1.03; 95% CI, 0.96 to 1.11; P=0.44).

4. Rivaroxaban was associated with significant reductions in critical bleeding subtypes, including intracranial hemorrhage (0.5% vs. 0.7% per year, P=0.02) and fatal bleeding (0.2% vs. 0.5% per year, P=0.003).

5. Rivaroxaban was associated with higher rates of gastrointestinal bleeding and drops in hemoglobin compared to warfarin.

Study Design

Design

RCT

Double-Blind

Sample

14,264

Patients

Duration

707 days

Median

Setting

45 countries

Population Patients with nonvalvular atrial fibrillation at moderate to high risk for stroke (CHADS2 score of 2 or greater).

Intervention Rivaroxaban 20 mg once daily (15 mg daily for patients with creatinine clearance 30-49 ml/min).

Comparator Dose-adjusted warfarin targeting an INR of 2.0 to 3.0.

Outcome Composite of stroke (ischemic or hemorrhagic) and non-CNS systemic embolism.

Study Limitations

• The mean time in therapeutic range (TTR) for the warfarin arm was approximately 55%, which is lower than in some other contemporary NOAC trials, potentially exaggerating the comparative benefit of rivaroxaban [3.4].

• Although noninferiority was robust, rivaroxaban did not achieve statistical superiority over warfarin in the intention-to-treat analysis (P=0.12).

• Patients discontinuing rivaroxaban at the end of the trial experienced a transient excess of strokes, highlighting the vulnerability of transitioning between a short-acting DOAC and a delayed-onset vitamin K antagonist without robust bridging.

• The enrolled population had a relatively high baseline risk of stroke (mean CHADS2 score of 3.5), limiting direct generalizability to lower-risk AF populations.

Clinical Significance

ROCKET-AF established rivaroxaban as a highly effective, once-daily alternative to warfarin for stroke prevention in nonvalvular AF. By eliminating the need for routine INR monitoring and dramatically reducing the risk of devastating intracranial and fatal bleeds, the trial supported a major paradigm shift toward direct oral anticoagulants (DOACs) as first-line therapy.

Historical Context

For decades, dose-adjusted vitamin K antagonists like warfarin were the standard of care for stroke prevention in atrial fibrillation, despite carrying significant burdens such as narrow therapeutic windows, dietary restrictions, and frequent monitoring. ROCKET-AF was part of a landmark trio of mega-trials (alongside RE-LY for dabigatran and ARISTOTLE for apixaban) that evaluated target-specific DOACs. Rivaroxaban's approval marked the arrival of the first once-daily Factor Xa inhibitor, greatly simplifying anticoagulation management.

Guided Discussion

High-yield insights from every perspective

Med Student

Medical Student

How does the mechanism of action of rivaroxaban differ from that of warfarin, and why does this pharmacological difference theoretically contribute to the observed reduction in intracranial hemorrhage in the ROCKET-AF trial?

Key Response

Rivaroxaban directly and selectively inhibits Factor Xa, blocking the convergence point of the intrinsic and extrinsic coagulation pathways. Warfarin inhibits VKORC1, depleting functional Factors II, VII, IX, and X. Warfarin causes a profound, sometimes erratic suppression of the extrinsic pathway (Factor VII) which, combined with its dietary/genetic variability, often leads to supratherapeutic spikes and higher intracranial hemorrhage risk. Factor Xa inhibitors provide a more targeted, predictable pharmacokinetic profile that may spare some baseline tissue factor-mediated hemostasis in the brain.

Resident

A 72-year-old patient with nonvalvular AF and a history of a major gastrointestinal bleed requires anticoagulation. Given the ROCKET-AF findings, how do you weigh the risks and benefits when choosing between rivaroxaban and warfarin for this specific patient?

Key Response

ROCKET-AF demonstrated that while rivaroxaban significantly reduced intracranial and fatal hemorrhages compared to warfarin, it paradoxically increased the risk of major gastrointestinal bleeding. For a patient with a prior GI bleed, residents must recognize this specific adverse effect profile and balance the catastrophic, often fatal nature of ICH against the treatable nature of GI bleeding. This often prompts consideration of alternative DOACs or shared decision-making regarding the exact GI pathology.

Fellow

The ROCKET-AF trial enrolled a significantly higher-risk population (mean CHADS2 score of 3.5) compared to the RE-LY and ARISTOTLE trials, but the mean Time in Therapeutic Range (TTR) for the warfarin arm was relatively low at 55%. How do these two factors complicate cross-trial comparisons and the interpretation of rivaroxaban's noninferiority?

Key Response

A lower TTR in the control arm can artificially inflate the relative efficacy and safety of the experimental drug, leading critics to argue warfarin was 'poorly managed'. However, the higher baseline CHADS2 score of the ROCKET-AF cohort makes achieving a high TTR inherently more difficult due to extensive comorbidities and polypharmacy. Fellows must understand that while rivaroxaban's efficacy is proven in a uniquely high-risk cohort, the low TTR warrants caution when applying these results to patients with historically well-controlled, stable warfarin regimens.

Attending

In ROCKET-AF, there was a noted increase in stroke and systemic embolism in the rivaroxaban group during the transition period at the end of the study when patients were switched back to open-label warfarin. How did this 'end-of-study' phenomenon permanently alter real-world clinical practice for managing DOAC transitions?

Key Response

The spike in thrombotic events upon discontinuing rivaroxaban highlighted a major pharmacokinetic vulnerability: a 'coverage gap' or prothrombotic rebound effect before warfarin reached a therapeutic INR. This finding profoundly changed clinical protocols, teaching attendings that transitioning from a short-acting DOAC to a Vitamin K antagonist requires careful, protocolized overlap and bridging strategies rather than abrupt switching.

Scholarly Review

Critical appraisal through the lens of expert reviewers and guideline development

PhD

The primary efficacy analysis for noninferiority in ROCKET-AF was conducted in the per-protocol (on-treatment) population, while superiority was tested in the intention-to-treat (ITT) population, yielding a non-significant p-value of 0.12. What is the statistical rationale for preferring per-protocol over ITT for noninferiority testing, and what inherent biases does this introduce?

Key Response

In noninferiority trials, an ITT analysis can bias results toward the null (making treatments look falsely similar) due to nonadherence, dropouts, or crossover. Therefore, a per-protocol analysis is preferred to ensure the active drug is truly noninferior when actually taken. However, PP analysis breaks randomization and is highly susceptible to selection bias and confounding. Testing superiority in ITT preserves randomization but failed to reach significance, meaning rivaroxaban cannot be definitively declared superior to warfarin in the broader real-world application of this trial.

Journal Editor

ROCKET-AF utilized a complex double-blind, double-dummy design with a central algorithm generating 'sham' INR values for the rivaroxaban group to maintain blinding. As a peer reviewer, what specific threats to validity might you flag regarding this blinding method and its impact on subjective endpoint adjudication?

Key Response

Managing sham INRs via an algorithm can sometimes inadvertently unblind investigators if patterns emerge (e.g., highly stable 'sham' INRs versus erratic real ones, or lack of bleeding despite high sham INRs). If local investigators become unblinded, it introduces significant ascertainment bias, particularly for softer, subjective secondary endpoints like 'clinically relevant nonmajor bleeding.' An editor would scrutinize the fidelity of this blinding and demand robust objective criteria for the clinical event committee's adjudication process.

Guideline Committee

Based on ROCKET-AF and subsequent DOAC trials, current ACC/AHA/HRS guidelines give DOACs a Class I recommendation over warfarin for nonvalvular AF. How does the specific renal clearance profile and bleeding data from ROCKET-AF dictate current guideline recommendations for dose adjustment, particularly regarding the use of rivaroxaban in patients with severe chronic kidney disease?

Key Response

Current guidelines recommend DOACs over warfarin (Class I, LOE A) for eligible patients. However, because rivaroxaban is partially renally cleared, ROCKET-AF excluded patients with CrCl < 30 mL/min and required a dose reduction (15 mg) for CrCl 30-49 mL/min. Consequently, guidelines mandate strict renal dose adjustments. Unlike apixaban, which has specific guideline support for use in ESRD/dialysis based on subsequent pharmacokinetic studies, rivaroxaban lacks robust safety data in ESRD, leading guidelines to generally prefer warfarin or apixaban in that specific subpopulation.

Clinical Landscape

Noteworthy Related Trials

2009

RE-LY Trial

n = 18,113 · NEJM

Tested

Dabigatran 110mg or 150mg twice daily

Population

Patients with nonvalvular atrial fibrillation and at least one risk factor for stroke

Comparator

Warfarin

Endpoint

Stroke or systemic embolism

Key result: Dabigatran 150 mg was superior to warfarin in reducing stroke or systemic embolism, while the 110 mg dose was non-inferior with significantly less major bleeding.

2011

ARISTOTLE Trial

n = 18,201 · NEJM

Tested

Apixaban 5mg twice daily

Population

Patients with nonvalvular atrial fibrillation and at least one additional risk factor for stroke

Comparator

Warfarin

Endpoint

Stroke or systemic embolism

Key result: Apixaban was superior to warfarin in preventing stroke or systemic embolism, caused less bleeding, and resulted in lower mortality.

2013

ENGAGE AF-TIMI 48 Trial

n = 21,105 · NEJM

Tested

Edoxaban 30mg or 60mg once daily

Population

Patients with nonvalvular atrial fibrillation and moderate-to-high risk of stroke

Comparator

Warfarin

Endpoint

Stroke or systemic embolism

Key result: Edoxaban was noninferior to warfarin for the prevention of stroke or systemic embolism and was associated with significantly lower rates of bleeding.

Tailored to your role

Want this tailored to you?

Add your specialty or training stage to get role-specific takeaways and more questions.

Personalize this analysis