Rivaroxaban Again Linked to More Bleeding vs Other DOACs

In the absence of head-to-head RCTs, experts say existing observational data should inform decision-making.

Rivaroxaban Again Linked to More Bleeding vs Other DOACs

Rivaroxaban (Xarelto; Bayer/Janssen) carries a significantly higher risk of gastrointestinal bleeding—including major GI bleeding—than do other direct oral anticoagulants (DOACs) across a real-world population taking the drugs for various indications, observational data from Iceland confirm.

Specifically, GI bleeding was 46% more common with rivaroxaban than with apixaban (Eliquis; Bristol-Myers Squibb). Risk trended similarly compared with dabigatran (Pradaxa; Boehringer Ingelheim), but the difference didn’t reach statistical significance.

Lead author Arnar B. Ingason, MD (University of Iceland, Reykjavik), in an interview with TCTMD, said that their study was borne out of his PhD research on the efficacy and safety of oral anticoagulation.

“We had this theory that rivaroxaban would have potentially higher risks of GI bleeding because it’s given as a once-daily dose, compared to the other two drugs, which are given twice daily. Theoretically, this should cause greater variance in drug plasma concentration, making these patients more susceptible to bleeding,” he explained.

In their paper, published today in Annals of Internal Medicine, the researchers suggest another possibility: that the convenience of rivaroxaban may inadvertently lead to harm. “Alternatively, the increased bleeding risk associated with rivaroxaban may be due to better adherence . . . , leading to more bleeding events and potentially fewer thromboembolic events,” they note.

This isn’t the first time rivaroxaban’s bleeding risks have been highlighted. Back in 2015, for instance, two BMJ papers raised concerns in contrast to warfarin. In 2017, Neena S. Abraham, MD (Mayo Clinic, Phoenix, AZ), who led of the BMJ-published studies, and colleagues followed with a comparison of DOACs in a US population reaching nearly 375,000 patients. “Our data revealed apixaban had the most favorable GI bleeding profile and rivaroxaban the worst,” Abraham told TCTMD in an email, noting that similar results from others have followed.

“I don’t think further investigation is needed on this topic. The data from the last 4 years overwhelmingly show that rivaroxaban is most likely to cause gastrointestinal bleeding. This finding is consistent among all age groups,” she stressed.

Craig January, MD, PhD (University of Wisconsin, Madison), who led the 2019 American College of Cardiology/American Heart Association atrial fibrillation (AF) guideline update, similarly said that “this story with rivaroxaban is not new.” This study, like others on the topic, is observational and thus calls attention to the lack of randomized controlled trials comparing different DOACs to each other. And while it has many strengths, including its thoroughness, one weakness is that it focuses on a “very homogenous, Icelandic population,” he added.

Major GI Bleeds More Frequent

Ingason et al analyzed results for new users of apixaban (n = 2,157), dabigatran (n = 494), and rivaroxaban (n = 3,217) at Landspítali–The National University Hospital of Iceland and four regional hospitals from 2014 to 2019. Manually confirming data by chart review, they identified 241 GI bleeds, of which 56% originated from the lower GI tract and 30% from the upper. Six in 10 of these bleeds were considered ISTH major (leading to hemoglobin decrease of ≥ 20 g/L, transfusion of ≥ 2 units of red blood cells, symptomatic bleeding in a closed compartment, or death due to bleeding).

Per 100 person-years, rivaroxaban was associated with higher risks of both overall and major GI bleeding compared with apixaban. The same patterns were seen with dabigatran, though the confidence intervals were wider, raising the “possibility of a null effect,” the researchers say.

Risk With Rivaroxaban vs Other DOACs


Per 100


HR (95% CI)

GI Bleeding









1.42 (1.04-1.93)

1.63 (0.91-2.92)

Major GI Bleeding









1.50 (1.00-2.24)

1.39 (0.67-2.88)

Sensitivity analyses including only AF patients showed increased risk of overall GI bleeding with rivaroxaban compared with both apixaban (HR 1.40; 95% 1.01-1.94) and dabigatran (HR 2.04; 95% CI 1.17-3.55).

For upper GI bleeding in particular, dabigatran was linked to the lowest risk. For lower GI bleeding, the lowest risk was seen with apixaban.

Time to Take Note?

For clinicians, then, there is the question of what to do—or not do—with this information.

Abraham says the message is clear: “It is time for cardiologists to choose a better choice when prescribing a direct oral anticoagulant to their atrial fibrillation patients, and that choice is not rivaroxaban.” For patients already on rivaroxaban, she suggested clinicians switch them to apixaban.

As to whether the evidence should influence decision-making, Ingason said, “I think so, yes. I think definitely it can help when trying to choose the correct oral anticoagulation for patients, especially those at high risk of GI bleeding,” such as those with a history or underlying disease that makes them extra vulnerable to having bleeds.

At their advent, DOACs were compared against warfarin in pivotal trials. What hasn’t happened with the same rigor are head-to-head trials among the newer drugs. “That is one of the main reasons why we did this study. There’s a knowledge gap when we’re comparing different DOACs,” said Ingason, though he isn’t optimistic these will happen. For one thing, such trials are expensive, he noted. Also, because “now it’s kind of debatable which drug is the best, I’m not sure the pharmaceutical companies want to take the risk of being a lesser DOAC medication.”

Nor was January hopeful that such trials will occur, given that they are “hard to fund.” Moreover, guidelines are unlikely to change in their absence, he predicted.

Already, though, the data on rivaroxaban have “affected some prescribing patterns,” said January, who noted that apixaban as of now is the most widely used. “If I’m asked the question which [non-vitamin K antagonist oral anticoagulant] do I use, my answer to that is a very clear ‘apixaban.’ I don’t mince words here.” His reasons include bleeding risk with rivaroxaban as well as a “fair amount” of gastroesophageal reflux disease-like symptoms with dabigatran, he noted.

Still, rivaroxaban’s once-daily pill remains attractive to many clinicians and patients, said January. “It would be nice to have all of these drugs available once a day in sort of a sustained release fashion, so it wasn’t just one big bolus after they swallow them.”

In Canada, two randomized controlled trials led by the Ottawa Hospital Research Institute, COBBRA and COBBRA-AF, are currently recruiting patients for a comparison of bleeding with rivaroxaban versus apixaban.

Caitlin E. Cox is News Editor of TCTMD and Associate Director, Editorial Content at the Cardiovascular Research Foundation. She produces the…

Read Full Bio
  • Ingason AB, Hreinsson JP, Ágústsson AS, et al. Rivaroxaban is associated with higher rates of gastrointestinal bleeding than other direct oral anticoagulants: a nationwide propensity score–weighted study. Ann Intern Med. 2021;Epub ahead of print.

  • Ingason, January, and Abraham report no relevant conflicts of interest.