Contrary to expectations, vitamin K antagonists (VKAs) reduced the risk for ischemic stroke and death, compared with the factor Xa inhibitor rivaroxaban, (Xarelto, Janssen) in patients with rheumatic heart disease and atrial fibrillation (AFib), in the INVICTUS trial.
Patients receiving a VKA, typically warfarin, had a 25% lower risk for the primary outcome – a composite of stroke, systemic embolism, myocardial infarction, or death from vascular or unknown causes outcome – than receiving rivaroxaban (hazard ratio, 1.25; 95% confidence interval, 1.10-1.41).
This difference was driven primarily by a significant reduction in the risk for death in the VKA group, and without a significant increase in major bleeding, reported Ganesan Karthikeyan, MD, from the All India Institute of Medical Sciences in New Delhi.
“VKA should remain the standard of care for patients with rheumatic heart disease and atrial fibrillation,” he concluded in a hotline session at the annual congress of the European Society of Cardiology.
The study, simultaneously published in the New England Journal of Medicine, is the first randomized controlled trial to assess anticoagulant therapy in patients with rheumatic heart disease and AFib.
“Who could have possibly guessed these results? Certainly not me,” said invited discussant Renato D. Lopes, MD, MHS, PhD, Duke Clinical Research Institute, Durham, N.C. “To me, this is one more classical example of why we need to do randomized trials, since they are the only reliable way to determine treatment effects and drive clinical practice.”
Rheumatic heart disease affects over 40 million people, mainly living in low- and low- to middle-income countries. About 20% of symptomatic patients have AF and an elevated stroke risk, but previous AFib trials excluded these patients, Dr. Karthikeyan noted.
INVICTUS was led by the Population Health Research Institute in Hamilton, Ont., and enrolled 4,565 patients from 24 countries in Africa, Asia, and Latin America who had rheumatic heart disease, AFib or atrial flutter, and an increased stroke risk caused by any of the following: CHA2DS2VASc score of 2 or more, moderate to severe mitral stenosis (valve area ≤ 2.0 cm2), left atrial spontaneous echo contrast, or left atrial thrombus.
Participants were randomly assigned to receive rivaroxaban, 20 mg once daily (15 mg/d if creatinine clearance was 15-49 mL/min), or a VKA titrated to an international normalized ratio (INR) of 2.0-3.0.
Warfarin was used in 79%-85% of patients assigned to VKA, with the percentage varying between visits. The INR was in therapeutic range in 33.2% of patients at baseline, 65.1% at 3 years, and 64.1% at 4 years.
During an average follow-up of 3.1 years, the primary outcome occurred in 446 patients in the VKA group (6.49% per year) and 560 patients in the rivaroxaban group (8.21% per year). The restricted mean survival time for the primary outcome was 1,675 vs. 1,599 days, respectively (difference, –76 days; 95% CI, –121 to –31 days; P for superiority < .001).
The rate of stroke or systemic embolism was similar between the VKA and rivaroxaban groups (75 vs. 94 events), although ischemic strokes were significantly lower with VKA (48 vs. 74 events).
No easy explanation
Deaths were significantly lower with VKA than rivaroxaban, at 442 versus 552 (restricted mean survival time for death, 1,608 vs. 1,587 days; difference, −72 days; 95% CI, –117 to –28 days).
“This reduction is not easily explained,” Dr. Karthikeyan acknowledged. “We cannot explain this reduction by the reduction in stroke that we saw because the number of deaths that are prevented by VKA are far larger than the number of strokes that are prevented. Moreover, the number of deaths were mainly heart failure or sudden deaths.”
Numbers of patients with major bleeding were also similar in the VKA and rivaroxaban groups (56 vs. 40 patients; P = .18), although numbers with fatal bleeding were lower with rivaroxaban (15 vs. 4, respectively).
By design, there were more physician interactions for monthly monitoring of INR in the VKA group, “but we do not believe such a large reduction can be explained entirely by increased health care contact,” he said. Moreover, there was no significant between-group difference in heart failure medications or hospitalizations or the need for valve replacement.
Almost a quarter (23%) of patients in the rivaroxaban group permanently discontinued the study drug versus just 6% in the VKA group.
Importantly, the mortality benefit emerged much later than in other trials and coincided with the time when the INR became therapeutic at about 3 years, Dr. Karthikeyan said. But it is unknown whether this is because of the INR or an unrelated effect.
More physician contact
Following the presentation, session cochair C. Michael Gibson, MD, Baim Institute for Clinical Research, Harvard Medical School, Boston, questioned the 23% discontinuation rate for rivaroxaban. “Is this really a superiority of warfarin or is this superiority of having someone come in and see their physician for a lot of checks on their INR?”
In response, Dr. Karthikeyan said that permanent discontinuation rates were about 20%-25% in shorter-duration direct oral anticoagulant trials, such as RELY, ROCKET-AF, and ARISTOLE, and exceeded 30% in ENGAGE-AF with 2.8 years’ follow-up.
“So, this is not new,” he said, adding that 31.4% of rivaroxaban patients did so for valve replacement surgery and subsequently received nonstudy VKA.
Dr. Lopes said it is important to keep in mind that INVICTUS enrolled a “very different population” that was younger (mean age, 50.5 years), was much more often female (72.3%), and had fewer comorbidities than patients with AFib who did not have rheumatic heart disease in the pivotal trials.
“It will be interesting to see the treatment effect according to mitral stenosis severity, since we had about 30% with mild mitral stenosis and additionally 18% of patients without mitral stenosis,” he added.
Co–principal investigator Stuart J. Connolly, MD, from the Population Health Research Institute, said physician contacts may be a factor but that the mortality difference was clear, highly significant, and sufficiently powered.
“What’s amazing is that what we’re seeing here is something that hasn’t been previously described with VKA or warfarin, which is that it reduces mortality,” he said in an interview.
Rivaroxaban has never been shown to reduce mortality in any particular condition, and a meta-analysis of other novel oral anticoagulants shows only a small reduction in mortality, caused almost completely by less intracranial hemorrhage than warfarin, he added. “So, we don’t think this is a problem with rivaroxaban. In some ways, rivaroxaban is an innocent bystander to a trial of warfarin in patients with rheumatic heart disease and atrial fibrillation.”
Dr. Connolly said more work is needed to explain the findings and analyses are planned to see which patients are at highest risk for death as well as looking at the relationship between INR control and outcomes.
“We need to do more research on what it is about VKA that could explain this,” he said. “Is it affecting the myocardium in some way, is it preventing fibrosis, is there some off target effect, not on the anticoagulation system, that could explain this?”
Athena Poppas, MD, chief of cardiology at Brown University, Providence, R.I., and past president of the American College of Cardiology, said “INVICTUS is an incredibly important study that needed to be done.”
“The results – though disappointing and surprising in some ways – I don’t think we can explain them away and change what we are doing right now,” she said in an interview.
Although warfarin is a cheap drug, Dr. Poppas said, it would be tremendously helpful to have an alternative treatment for these patients. Mechanistic studies are needed to understand the observed mortality advantage and low bleeding rates but that trials of other novel anticoagulants are also needed.
“But I’m not sure that will happen,” she added. “It’s unlikely to be industry sponsored, so it would be a very expensive lift with a low likelihood of success.”
In an editorial accompanying the paper, Gregory Y.H. Lip, MD, University of Liverpool (England), pointed out that observational data show similar or even higher risks for major bleeding with rivaroxaban than with warfarin. “To improve outcomes in these patients, we therefore need to look beyond anticoagulation alone or beyond a type of anticoagulation drug per se. Indeed, a one-size-fits-all approach may not be appropriate.”
The study was funded by an unrestricted grant from Bayer. Dr. Karthikeyan and Dr. Poppas reported no relevant conflicts of interest.
A version of this article first appeared on.