RIVER
Guimarães HP, Lopes RD, de Barros e Silva PGM, et al. Rivaroxaban in patients with atrial fibrillation and a bioprosthetic mitral valve. N Engl J Med. 2020;383(22):2117-2126.
Clinical Question
Is Rivaroxaban (Xarelto) noninferior to Warfarin for prevention of thromboembolic events among patients with Atrial Fibrillation/Atrial Flutter and a bioprosthetic Mitral Valve.
Bottom Line
Rivaroxaban is non-inferior (p<0.001) to warfarin for prevention of the composite outcome of death, major cardiovascular events (stroke, TIA, systemic embolization, valve thrombosis, or hospitalization for heart failure), and major bleeding at 12 months in patients with atrial fibrillation/flutter and a bioprosthetic mitral valve.
Major Points
Patients with afib and a bioprosthetic mitral valve require long-term anticoagulation. At the moment the only approved agent for these patients is warfarin. Warfarin is a difficult medication to manage due to a very narrow therapeutic index requiring frequent INR monitoring and also has many medication and food interactions. The direct oral anticoagulants (DOACs) have been shown to be noninferior (and in some cases superior) to warfarin for the prevention of systemic embolization in atrial fibrillation in several studies over the last 15 years (ARISTOTLE, ENGAGE AF-TIMI 48, RE-LY, ROCKET-AF).
Specifically, the ROCKET AF trial showed that rivaroxaban was noninferior to warfarin for anticoagulation in nonvalvular afib and the ARISTOTLE and ENGAGE-AF-TIMI 48 trials both included small numbers of patients with bioprosthetic mitral valves (131 out of 21,105 patients for ENGAGE-AF-TIMI and 31 out of 18,201 for ARISTOTLE). The RIVER trial sought to extend the work of those prior trials by running an RCT that specifically looked at this population of interest. In the interest of attempting DOACs in valvular afib it was reasonable to start with bioprosthetic mitral valves because there IS limited data from those larger trials of safety and efficacy and because bioprosthetic valves, in general, are considered less thrombogenic than mechanical valves (for example the INR target for bioprosthetic valves is 2.0-3.0 rather than 2.5-3.5 used for mechanical valves).
The Rivaroxaban for Valvular Heart Disease and Atrial Fibrillation (RIVER) trial compared rivaroxaban to warfarin for the prevention of stroke/systemic embolization and major bleeding. The study used a time-to-event analysis with a prespecified end point of 1 year and a non-inferiority margin of 8 days (in favor of warfarin). This design is similar to Kaplan-Meier mortality analysis where the primary outcome is how long the mean patient would have before experiencing the primary outcome event. In this case the primary outcome event was a composite of death, major bleeding, and major cardiovascular event (defined as stroke, TIA, systemic embolization, valve thrombosis, or heart-failure hospitalization).
The primary outcome event occurred at a mean of 347.5 days in the rivaroxaban and 340.1 days in the warfarin group. This is a between group difference of 7.4 days (in favor of rivaroxaban) with a 95% CI of -1.6 - 26.3 days. This had a p-value for non inferiority of <0.001 (and p-value for superiority of 0.1).
Important secondary outcomes included a composite of death from cardiovascular cause or thromboembolic events with a hazard ratio (HR) 0.65 (95% CI: 0.35 - 1.2) and total incidence of stroke with HR 0.25 (95% CI: 0.07 - 0.88). Major bleeding had a HR of 0.54 (95% CI: 0.21 -1.35). All secondary outcomes favored rivaroxaban although most did not reach statistical significance.
This study adds to the growing literature of the safety and efficacy of DOACs for anticoagulation in atrial fibrillation. This study pushes into the territory of VALVULAR atrial fibrillation for which VKAs such as warfarin remain the standard of care. It is likely that over time the DOACs will be shown to be safe and efficacious in the remaining areas of valvular atrial fibrillation (ie. bioprosthetic aortic valves and mechanical mitral/aortic valves). The development of reversal agents (eg andexanet) and the approaching time for generic availability (eg 2024 for rivaroxaban) of these drugs will likely lead to the phasing out of warfarin for most patients with atrial fibrillation.
Guidelines
No guidelines have been updated to reflect the results of this trial although the 2019 update to the ACC/AHA/HRS guideline for the management of patients with atrial fibrillation (link) acknowledges that based on the small numbers of patients with bioprosthetic valves in ENGAGE-AF-TIMI and ARISTOTLE that DOACs might be noninferior to VKAs although further studies were needed.
Design
- Multicenter, randomized, non-inferiority, open-label (with blind adjudication) clinical trial.
- N=1005
- Rivaroxaban n=500
- Warfarin n=505
- Setting: 49 centers in Brazil
- Follow-up: 12 months (0.6% lost to follow-up)
- Analysis: Intention-to-treat (although also analyzed as per-protocol and as-treated for validity), non inferiority (80% power to detect non-inferiority margin of 8 days).
- Primary outcome:
- Mean time-to-event (Kaplan-Meier analysis) for primary outcome event
- Primary outcome event: composite of death, major cardiovascular event (stroke, TIA, systemic embolization, valve thrombosis, hospitalization for CHF), or major bleeding
- Secondary outcomes:
- Composite of death from cardiovascular causes or thromboembolic events
- Incidence of total stroke
- Valve thrombosis
- Safety outcomes:
- Major bleeding
Population
Inclusion Criteria
- Adults (≥18 years of age) AND
- Permanent, paroxysmal, or persistent atrial fibrillation or atrial flutter AND
- Bioprosthetic mitral valve (Patients were eligible for in- clusion in the trial at any time at least 48 hours after undergoing mitral-valve surgery.) AND
- Were receiving (or planning to receive) oral anticoagulation for thromboembolism prophylaxis.
Exclusion Criteria
- Contraindication to either rivaroxaban or warfarin
- An extremely high risk of bleeding
- Transient atrial fibrillation caused by surgery
- The placement of mechanical valves.
Baseline Characteristics
Very similar between groups
- Mean age: 59.3 ± 12.1 years old
- Female sex: 60.4%
- BMI: 26 (IQR 23.2-29.7
- Current smokers: 3.8%
- Median creatinine: 0.9
- Median creatinine clearance: 77.5
- Mean CHA2DS2VASc score 2.6 ± 1.4
- Mean HAS-BLED score: 1.6 ± 0.9
Race or ethnicity
- White: 56.1%
- Black: 13.1%
- Multiracial: 29.5%
- Asian: 1.1%
Medical History
- Diabetes Mellitus: 13.7%
- Hypertension: 60.7%
- Percutaneous valve intervention: 7.5%
- Dyslipidema: 33.6%
- Stroke: 12.8%
- TIA: 2.5%
- CHF: 38.8%
Type of rhythm:
- Paroxysmal afib: 22.2%
- Permanent afib: 61.7%
- Persistent afib: 11.6%
- Atrial flutter: 4.3%
Interval between mitral-valve implantation and randomization:
- <3 month: 18.8%
- 3mo - <1 year: 16.8%
- 1 year - <5 years: 32.2%
- 5 years - <10 years: 30.6%
- Missing data: 1.4%
Interventions
- Randomization to warfarin or rivaroxaban groups in 1:1 ratio using permuted blocks with the use of a central concealed, web-based automated randomization system.
- Rivaroxaban - 20mg once daily (or 15mg once daily if CrCl 30-49)
- Warfarin - dose adjusted to maintain target INR 2.0 - 3.0 (INR was measured at least every 4 weeks)
Outcomes
Comparisons are Rivaroxaban vs Warfarin
Primary Outcomes
Time to composite outcome of death, major cardiovascular event (stroke, TIA, systemic embolization, valve thrombosis, heart failure hospitalization), and major bleeding:
- Intention to treat: mean of 347.5 days vs 340.1 days. Between group difference 7.4 days; 95% CI: -1.4 - 16.3 days. P<0.001 for non inferiority. P=0.1 for superiority,
- As treated: mean of 350.1 days vs 339.6 days. Between group difference 10.5 days; 95% CI: 1.9 - 19.1 days.
- Per-protocol: mean of 356.7 vs 347.1 days. Between group difference 9.6 days; 95% CI: 2.2 - 16.9 days.
Secondary Outcomes
Composite of death from cardiovascular causes or thromboembolic event
17 (3.4%) vs. 26 (5.1%). HR 0.65; 95% CI: 0.35 - 1.2
Incidence of total stroke
0.6% vs 2.4%. HR 0.25; 95% CI: 0.07 - 0.88.
Valve thrombosis
5 (1%) vs 3 (0.6%)
Safety Events
Major Bleeding
7 (1.4%) vs 13 (2.6%). HR 0.54; 95% CI: 0.21 - 1.35.
Clinically relevant nonmajor bleeding
4.8% vs 4.6%
Subgroup Analyses
Results for the primary outcome were generally consistent across most subgroups.
However, among the patients who underwent randomization up to 3 months after mitral-valve surgery, the mean time until a primary-outcome event was 348.6 days in the rivaroxaban group and 313.5 days in the warfarin group (difference, 35.1 days; 95% CI, 8.6 to 61.7). Similarly, in this subgroup, the incidence of a primary-outcome event was 6.4% in the rivaroxaban group and 18.9% in the warfarin group (hazard ratio, 0.31; 95% CI, 0.12 to 0.79). This indicates that this subgroup might see an even greater benefit from rivaroxaban.
Criticisms
- The study was not blinded (but WAS blindly adjudicated)
- Study used time-to-event analysis (Kaplan-Meier) for primary outcome. This was chosen because it is not dependent on number of events but is a different metric than used by comparable studies with larger sample sizes.
- Study used a large composite primary outcome: composite of death, major bleeding, and major cardiovascular events (stroke, TIA, systemic embolization, valve thrombosis, and heart failure hospitalization). This could have magnified effect size.
- Warfarin group was only therapeutic for 65.5% of the time (this does match the standard set by other similar trials like ROCKET-AF).
Funding
- Brazilian Ministry of Health (Programa de Apoio ao De- senvolvimento Institucional do Sistema Único de Saúde [PROADI-SUS])
- Bayer (developer of Xarelto and distributer outside of the US - Janssen distributes in the US)