More Bleeding Linked To Rivaroxaban In Observational Study

–Compared to dabigatran, rivaroxaban was associated with more intracranial and other serious bleeds.

A new observational study finds that patients who take rivaroxaban (Xarelto, Johnson & Johnson) may be more likely to have serious bleeding events than patients who take dabigatran (Pradaxa, Boehringer Ingelheim).

In a paper published in JAMA Internal Medicine, David Graham and colleagues at the FDA performed a retrospective analysis of 118,000 Medicare patients with atrial fibrillation who started therapy with either rivaroxaban or dabigatran. They found that rivaroxaban was associated with statistically significant increases in intracranial hemorrhage, resulting in an excess of 2.3 cases per 1,000 person-years after an adjustment for baseline differences in the groups. Rivaroxaban was also linked to significant excesses in major extracranial bleeding (an excess of 13 cases), including an increase in major GI bleeds (9.4 cases). There was a non-significant reduction in thromboembolic stroke associated with rivaroxaban, resulting in 1.8 fewer cases. There was no significant difference in mortality, but the authors noted a significant increase in the subgroup of patients 75 or older with a CHADS2 score over 2.

The authors noted that in 2014 rivaroxaban was used 2-3 times more often than dabigatran in US patients with AF. They speculated that this might be due to “prescriber misperceptions about bleeding risks with dabigatran, arising from US Food and Drug Administration receipt of a large number of postmarketing case reports following its approval. Ironically, we found substantially higher bleeding risks with use of rivaroxaban than dabigatran.”

Dabigatran and rivaroxaban are both new oral anticoagulants (NOACs) that have been the subject of controversy and criticism. Dabigatran has been plagued by reports of bleeding complications since shortly after its approval. More recently, there have been reports critical of the pivotal rivaroxaban ROCKET AF trial because of the use of a defective monitoring device in the trial.

The authors acknowledged that a “randomized direct comparison between dabigatran and rivaroxaban would be optimal,” but that “such a study is unlikely to be undertaken by the manufacturer of either product.”

Outside experts expressed concern about conclusions derived from observational studies based on Medicare claims. Comparative effectiveness studies based on claims do not provide “very strong evidence,” said Harlan Krumholz (Yale University). Sanjay Kaul (Cedars-Sinai) warned that “drawing strong inferences from a claims database is a slippery slope.”

In an editor’s note, on the other hand, Rita Redberg (UCSF) and Anna Parks offer strong support for the study’s real world application:

“The additional information should lead us to prescribe dabigatran over rivaroxaban for patients with atrial fibrillation. Also, it can help inform further investigation of NOAC monitoring and tailored dosing, as others have previously recommended for dabigatran. In addition, knowing that a patient prescribed rivaroxaban is at comparatively increased risk of bleeding might encourage clinicians to more vigilantly identify and mitigate modifiable risk factors. Finally, it offers important guidance to consumers on relative efficacy and safety profiles that drive NOAC selection, particularly for those at greatest risk of hemorrhage.”

I asked Redberg if she had any concerns about basing recommendations on observational studies. “Clearly I would prefer RCT data,” she responded. “But currently there are multiple NOACs on the market and there are no RCT that compare them , and I know of none in the works. Thus, the best data we will get right now to help inform patient care is from high quality observational studies.”


  1. If these NOACs are competing with each other for the market, oughtn’t they be required to show head-to-head evidence of respective effectiveness/safety? The public interest seems to be getting short shrift.

    Why not prescribe apixaban over dabigatran? It has shown itself indirectly, vs warfarin, to be less prone to cause bleeds.

    I would further ask what proportion of prospective dabigatran subjects dropped out of trials – before being counted – because of GI complaints other than bleeds. The product is acidified to enhance absorption. Trouble, uncounted.

Leave a Reply

%d bloggers like this: