Rivaroxaban Cuts Stroke Risk in Sinus-Rhythm Heart Failure: COMMANDER-HF

May 31, 2019

Patients in the COMMANDER-HF trial, who had heart failure with reduced ejection fraction (HFrEF) and were in sinus rhythm, showed a steep decline in risk for neurologic events over 4 years after a direct oral anticoagulant (DOAC) was added to their standard meds, compared with similar patients who had no DOAC added.

The trial's post hoc analysis suggests that low-dose rivaroxaban (Xarelto, Bayer/Janssen Pharmaceuticals) protects against some strokes or transient ischemic attacks (TIA) in such patients without posing an untoward bleeding risk.

But also, researchers say, it supports a role for the CHA2DS2VASc stroke risk prediction tool outside its more familiar realm of atrial fibrillation (AF).

The risk for stroke or TIA fell a significant 31% in patients taking rivaroxaban 2.5 mg twice daily on top of background antiplatelet therapy without an associated jump in risk for major bleeding, although minor bleeds were more common than in the control group.

A CHA2DS2VASc score greater than 4 was seen to "identify and capture nearly half of stroke events, and use of rivaroxaban appreciably reduced the number needed to treat to less than 100 per year in this high-risk subset," said Muthiah Vaduganathan, MD, MPH, Brigham and Women's Hospital, Boston.

This points to a need for further studies to determine "whether select populations with heart failure and reduced ejection fraction in sinus rhythm perhaps can be identified by use of traditional risk scores such as the CHA2DS2VASc score," he concluded in his presentation of the analysis here at European Society of Cardiology Heart Failure 2019.

The previously reported COMMANDER-HF had in its primary analysis failed to show an advantage for rivaroxaban for the end point of death, myocardial infarction (MI), or stroke, so cautions are in order in interpreting the post hoc analysis of the negative trial.

But stroke or TIA was at least a component of the primary end point, and the findings are largely consistent with other studies of antithrombotic therapy in heart failure, so "we believe it's a true signal," Vaduganathan told theheart.org | Medscape Cardiology.

COMMANDER HF had randomly assigned 5022 patients with recent exacerbation of chronic HFrEF and coronary artery disease but no AF to the low-dose rivaroxaban regimen or placebo.

In addition to the neutral clinical-event outcome, the principal safety end point of fatal bleeding or bleeding into a critical space was not significantly different between the groups. Nor was the risk for major bleeding by International Society of Thrombosis and Hemostasis (ISTH) criteria.

But in an exploratory analysis, rivaroxaban had been associated with a significant 34% decline in risk for stroke or TIA.

Patients like those in the study face a risk for stroke or TIA approaching that seen in heart-failure patients with AF," Vaduganathan observed when presenting the analysis. And in COMMANDER-HF, "nearly half of first primary neurological events were either disabling or fatal."

The cumulative risk for stroke was close to 2% at 1 year, but it rose early, plateaued at about 6 months, and stayed elevated.

Because the DOACs are not approved for patients in sinus rhythm with HFrEF, in practice such patients would be treated if an approved indication, such as AF, is identified on further workup, Mandeep R. Mehra, MD, Brigham and Woman's Hospital in Boston, told theheart.org | Medscape Cardiology.

"What we don't know is the rate of intermittent, paroxysmal, or silent atrial fib in these patients," said Mehra, who is senior author on the analysis. "You know, technologies like the Apple Watch might actually yield some benefit in heart failure patients if they show atrial fib, if you apply them in a high-risk population."

On the other hand, "we don't understand why patients with heart failure have strokes," observed John J.V. McMurray, MB, MD, University of Glasgow, United Kingdom, as discussant after the analysis was presented.

Strokes in patients with HF were once thought to be primarily caused by thromboembolism originating in a low-ejection-fraction left ventricle, but more recently it's been appreciated that ejection fraction and stroke risk are not closely associated, he noted.

"Maybe we should be thinking about that, and be trying hard to detect atrial fibrillation," because that's already a target for oral anticoagulation, McMurray agreed.

Earlier studies of oral anticoagulation in sinus-rhythm heart failure, he pointed out, have not shown a good balance of risk and benefit. That's largely because in such patients "the risk of stroke is usually relatively low and the risk of bleeding is significant, at least with conventional anticoagulants given at the conventional dose."

He proposed a few explanations for why the risk–benefit equation in COMMANDER-HF was more favorable than in earlier trials. For example, their fairly high rate of stroke over the follow-up suggests that the study population was at higher risk than in some trials.

Or it may be that the study's safety end-point definition for bleeding — that is fatal bleeding or bleeding into a critical space — was less expansive than the "major bleeding" end point used in most past trials, McMurray said.

"But I do agree with the COMMANDER-HF investigators in that we all need to think of better ways to identify high-risk patients, and target this treatment to those patients." Maybe CHA2DS2VASc risk stratification would be effective, or another risk tool, "but that's where we need to go."

In the post hoc analysis, the 31% drop in risk for stroke or TIA over a mean of 21 months was driven by the reduction in ischemic strokes, which made up the overwhelming majority of events.

Hazard Ratio (HR) for Neurologic End Points, Rivaroxaban vs Placebo, in COMMANDER-HF
End Point HR (95% CI)
Any stroke or TIA 0.69 (0.50–0.95)
Ischemic stroke 0.64 (0.43–0.95)
Hemorrhagic stroke 0.74 (0.25–2.13)
TIA 0.77 (0.34–1.75)

That worked out to a number needed to treat of 164 to prevent 1 stroke or TIA. It fell to less than 100 for patients with a baseline CHA2DS2VASc score greater than 4.

Stroke or TIA Incidence per 100 Patient-Years by CHA2DS2VASc Stratum
End Point Placebo, Rate Rivaroxaban, Rate NNT
Overall 1.9 1.29 164
CHA2DS2VASc ≤4 1.44 1.13 316
CHA2DS2VASc >4 2.56 1.52 96
NNT=number needed to treat to prevent one end point.

Indeed, the HR for stroke or TIA for every CHA2DS2VASc point increase was 1.29 (95% CI, 1.13 - 1.48; P < .001), with history of stroke among the strongest predictors at HR 2.35 (95% CI, 1.39 - 3.98; P = .002).

"I think the really critical finding in this is that these scores like CHA2DS2VASc had the exact same performance quality as in atrial fib," Mehra said.

"So in other words, these scores are not just the domain of atrial fib, they're the domain of vascular disease and heart failure, and I think you can apply them more broadly."

COMMANDER HF was funded by Janssen. Vaduganathan discloses serving on advisory boards for Amgen, AstraZeneca, Bayer, and Baxter Healthcare; and on a clinical end-point committee for Novartis. Mehra has previously disclosed being a consultant for Abbott, Portola, Bayer, and Xogenex; a trial steering committee member for Medtronic and Janssen; a scientific advisory board member for NuPulseCV and FineHeart; and a data safety monitoring board member for Mesoblast; and receiving travel support from Abbott. McMurray has previously disclosed his institution paid for his participation in clinical trial committees by AbbVie, AstraZeneca, Amgen, Bayer, Bristol-Myers Squibb, Dalcor, GlaxoSmithKline, Merck, Novartis, Resverlogix, Stealth, and Theracos; that these sponsors funded some costs to attend meetings; and it paid for attendance by Novartis and Sanofi-Aventis advisory boards.

European Society of Cardiology Heart Failure (ESC-HF) 2019: Abstract 648. Presented May 26, 2019.

Follow Steve Stiles on Twitter: @SteveStiles2. For more from theheart.org | Medscape Cardiology, follow us on Twitter and Facebook.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.