Rivaroxaban raises GI bleeding risk vs warfarin: ROCKET-AF analysis

Emma Hitt

October 24, 2012

Atlanta, GA - Patients with atrial fibrillation (AF) receiving anticoagulant therapy are more likely to experience gastrointestinal (GI) bleeding when treated with rivaroxaban (Xarelto, Bayer) than with warfarin, according to a new analysis of data from ROCKET-AF [1].

"Compared with warfarin, the risk of GI bleeding is increased, but the incidence of life-threatening or fatal GI bleeding is lower with rivaroxaban," Dr Christopher Nessel (Johnson & Johnson, Raritan, NJ) said in an interview. "A careful benefit/risk assessment is needed prior to prescribing rivaroxaban for high-risk patients."

Nessel reported the study findings in an oral presentation October 22, 2012 at CHEST 2012 , the annual meeting of the American College of Chest Physicians. The study was published online in Chest, with Nessel as first author, to coincide with his presentation.

The current analysis examined the incidence and outcomes of GI hemorrhage among 14 264 patients with nonvalvular AF enrolled in the ROCKET-AF study. The patients were randomized to either rivaroxaban or dose-adjusted warfarin. All GI bleeding events were recorded during treatment and for two days after the last dose was administered. Severity of bleeding was defined by a corresponding drop in hemoglobin or transfusion of >2 units of red cells.

Data analysis revealed that the composite principal safety end point for GI bleeding events (upper GI, lower GI, and rectal bleeding) occurred more frequently in patients receiving rivaroxaban than warfarin (3.61% vs 2.60% per year, respectively; hazard ratio 1.39; 95% CI 1.19-1.61). Both major (2.00% vs 1.24%) and nonmajor, clinically relevant (1.75% vs 1.39%) bleeding occurred more frequently on rivaroxaban than on warfarin.

Patients who experienced major GI bleeding were more likely to have experienced GI bleeding in the past, to have mild anemia or a lower creatinine clearance, or to be previous or current smokers or older than patients who did not experience a GI bleed (n=13 552) during this trial. They were less likely to be female and to have previously experienced a stroke or transient ischemic attack.

The incidence of those who experienced the most severe bleeding, as measured by transfusion of at least 4 units, however, was similar between both treatment groups (49 in the rivaroxaban group and 47 in the warfarin group). Six patients developed fatal bleeding: one from the rivaroxaban group and five patients from the warfarin group.

Data may give clinicians pause when considering rivaroxaban

"The data presented extend the observations from the ROCKET AF clinical study," Nessel said. "Specifically, the analyses identified characteristics of nonvalvular-AF patients that may predispose them to the occurrence of GI hemorrhage. The data also indicated that the overall fatality rates for bleeds of this nature are very low."

Independent commentator Dr James Wisler (Duke University Medical Center, Durham, NC) pointed out that this study underscores the importance of carefully evaluating patients individually when considering use of the newer oral anticoagulants.

"The decision regarding which anticoagulant to use for a given patient is complex, and risks and benefits need to be considered thoughtfully," he said. He added that the results of this study may give some physicians pause about initiating a newer anticoagulant such as rivaroxaban in a given patient with atrial fibrillation and an unfavorable risk profile, such as those with a previous gastrointestinal bleed.

"While the previously published results of ROCKET-AF suggested the risk profiles were quite similar between rivaroxaban and warfarin, these results demonstrate there is indeed a subpopulation of patients who may be better served with warfarin as opposed to rivaroxaban."

According to Wisler, both this analysis as well as the initial ROCKET-AF study demonstrate that rivaroxaban is associated with less episodes of severe or fatal bleeding events, despite an increase in major and clinically relevant nonmajor bleeding observed in the specific subgroup of this study. "Currently it is unclear why this discrepancy exists," he added. 

He recommended that clinicians take a careful history to assess bleeding risk factors in all patients in whom a provider is considering initiation of a newer anticoagulant such as rivaroxaban.

"While perhaps more convenient and efficacious, certain patient populations such as that evaluated in this study may receive net harm from these newer agents."

Nessel is an employee of Johnson & Johnson. Disclosures for the coauthors are listed in the paper.Wisler reports no relevant financial relationships.


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.
Post as: