Rivaroxaban: It's Not Time to Cut the Rope, Yet

John Mandrola, MD


February 09, 2016

Recent controversy surrounding the pivotal ROCKET AF[1] trial deserves your attention. Use of a faulty device for INR testing in the warfarin arm has raised concerns over the validity of the study, which found rivaroxaban (Xarelto, Bayer/Johnson & Johnson) noninferior to warfarin for the prevention of stroke or embolism in patients with nonvalvular atrial fibrillation (AF).

Any uncertainty over ROCKET AF is significant, because it was the only study the FDA considered for approval of rivaroxaban for stroke prevention in nonvalvular AF. The once-daily and well-tolerated rivaroxaban was the most-prescribed direct oral anticoagulant drug in 2014.[2]

In this post, I will review these concerns and make the case that's it's too early to abandon rivaroxaban for stroke prevention in patients with AF.

The Issue

The new concern with ROCKET AF is an extension of its initial criticism—namely, that inadequacy of the warfarin comparator arm stacked the deck in favor of rivaroxaban. At the time of FDA approval, some experts cited the time in therapeutic range of only 55% in patients in the warfarin arm as a reason to doubt rivaroxaban's relative efficacy and safety.

Now we learn that patients in the warfarin arm of ROCKET AF had their INRs adjusted using a point-of-care device that is under a class I FDA recall. That device can provide spuriously low INR results in patients with certain conditions, including abnormal hematocrit levels or any illness known to increase fibrinogen levels, such as infections or inflammatory conditions. This is big, because falsely low INR readings could lead to updosing of warfarin and possible overanticoagulation.

This news sent the ROCKET AF researchers back to the patient-level data. In a research letter[3] to the New England Journal of Medicine, they reported that 37% of patients in the warfarin arm had a condition that could have been affected by the faulty INR device. After performing an extensive subgroup analysis, the investigators upheld the main results of the trial: rivaroxaban remained noninferior to warfarin without an increase in major bleeding.

The reanalysis did not convince everyone. In an article[4] published in the BMJ, Deborah Cohen, a BMJ associate editor, made the case that ROCKET AF needs an independent analysis. Cohen cited other experts who called for release of the patient-level data. Former FDA cardiovascular and renal drug reviewer, Thomas Marcinicak, said that he would not rely on any reanalyses done by Duke, Johnson & Johnson, or the FDA. Dr Harlan Krumholz (Yale Open Data Project) called this "an ideal situation for data sharing," adding that it seems "imprudent to allow those who conducted the trial to be the only ones who can touch the data."

Rivaroxaban is manufactured by Bayer Healthcare and marketed in the United States by Janssen, part of Johnson & Johnson. Dr Nihar Desai, also of the Yale Open Data Access project, explained to me via email that he requested access to the data from Bayer but was told that because "Xarelto was approved in the US and the EU before January 1, 2014 . . . it is not in the current scope of clinical-trial data sharing according to our transparency policy." The 2014 date is when the European Federation of Pharmaceutical Industries and Associations (EFPIA) and the Pharmaceutical Research and Manufacturers of America (PhRMA) principles on responsible clinical trial data sharing went into effect.


To the Bayer corporation, I repeat my message from a previous column: If you have a great result, doesn't validation make it greater?

Why is transparency time sensitive? Surely there is enough uncertainty in this case to warrant independent verification.

That said, there are four reasons to remain somewhat confident in rivaroxaban.

The first line of reasoning comes from the post hoc analysis of ROCKET AF.[3] When the research team tallied bleeding events in patients with any condition that could be affected by the recall, they found major bleeding trended higher with rivaroxaban. That's reassuring, because if faulty INR tests affected the results, you would have expected recall conditions would have caused higher rates of bleeding in the warfarin group.

The second line of reasoning concerns the results of performing cardioversion and AF ablation with periprocedural rivaroxaban. I chose these two procedures because both are known to require adequate anticoagulation. If rivaroxaban lacked efficacy, we would likely see a signal of harm when it was used before and after these procedures.

In the X-Vert trial,[5] a prospective RCT comparing rivaroxaban and warfarin in 1504 patients with AF undergoing elective cardioversion, researchers reported an incidence of stroke of 0.51% in the rivaroxaban group and 1.02% in the warfarin group. In a post hoc analysis of the ROCKET AF, Piccini et al[6] compared outcomes in patients who underwent cardioversion or AF ablation and found no significant difference in long-term outcomes. Recently, Vamos et al[7] performed a systematic review and meta-analysis of 16 studies (7400 patients) comparing rivaroxaban and warfarin for periprocedural anticoagulation for AF ablation. They reported similar efficacy and safety of rivaroxaban relative to warfarin.

Taken together, these studies provide a measure of confidence about rivaroxaban use in anticoagulation-dependent procedures. These data also confirm my 5-year personal experience with rivaroxaban, which has been favorable.

A third line of reasoning comes from the results of real-world data. In the industry-funded XANTUS trial,[8] a prospective observational study of patients treated with rivaroxaban for stroke prevention, investigators enrolled 6784 AF patients from 311 centers in Europe, Canada, and Israel. Characteristics of the cohort approximated everyday clinical practice. Over an average follow-up of 329 days, rates of major bleeding (2.1 per 100 patient-years), stroke (0.7 per 100 patient-years), and death (1.9 per 100 patient years) were either equivalent to or less than those reported in ROCKET AF.

The final reason I will continue to use rivaroxaban—at least, in the short term—is its once-daily dosing. The literature is replete with studies confirming what any doctor or patient knows: adherence is greater with daily dosing. Adherence is an important issue with any drug, but it's especially relevant with direct-acting anticoagulants, because, unlike warfarin, we don't have a reliable way to test drug adherence.

Yes, the 5-year experience with rivaroxaban has been positive. There are reasons to believe in this drug. But uncertainty over the pivotal trial will remain until that data set is verified by independent scientists. Otherwise, rivaroxaban for stroke prevention will always have an asterisk.



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.