Medicare beneficiaries with nonvalvular atrial fibrillation who received 20 mg per day of rivaroxaban had statistically significant increases in intracranial hemorrhage and major extracranial bleeding compared with those who received 150 mg of dabigatran twice daily, according to a retrospective cohort study.

Lead researcher David J. Graham, MD, MPH, of the U.S. Food and Drug Administration (FDA)’s Center for Drug Evaluation and Research, and colleagues published their results online Oct. 3 in JAMA Internal Medicine.

The FDA and Centers for Medicare & Medicaid Services funded the study through its SafeRx project.

The FDA has approved dabigatran (Pradaxa) and rivaroxaban (Xarelto) for stroke prevention in patients with nonvalvular atrial fibrillation. They are oral anticoagulants that do not require therapeutic monitoring.

In 2014, rivaroxaban was used two to three times more often than dabigatran in U.S. patients with atrial fibrillation, according to the researchers. They said that was in part due to the FDA receiving a large number of postmarketing case reports that claimed dabigatran was associated with bleeding risks.

For this analysis, the researchers compared the standard doses of each medication in Medicare beneficiaries who were at least 65 years old. The adults initiated treatment with dabigatran or rivaroxaban from Nov. 4, 2011, to June 30, 2014.

The analysis included 52,240 patients treated with dabigatran and 66,651 patients treated with rivaroxaban. Patients were excluded if they were enrolled in Medicare Parts A, B and D for less than six months, were younger than 65 years old, received prior treatment with warfarin or another oral anticoagulant, resided in a skilled nursing facility or received hospice care when they filled their first dabigatran or rivaroxaban prescription.

During the follow-up period, there were 846 deaths, 306 thromboembolic strokes, 176 intracranial bleeding events and 1,209 major extracranial bleeding events.

The researchers found that rivaroxaban use was associated with a statistically nonsignificant reduction in thromboembolic stroke, statistically nonsignificant increase in mortality and statistically significant increases in intracranial hemorrhage, major extracranial bleeding and major gastrointestinal bleeding.

They also found that there was a statistically significant increase in all hospitalized extracranial bleeding events with rivaroxaban and no difference in acute MI risk between the medications.

Among patients who were at least 75 years old and had a CHADS₂ score greater than 2, rivaroxaban use was associated with significantly increased mortality compared with dabigatran use.

After adjustments, patients who received rivaroxaban had 1.8 fewer cases of thromboembolic stroke, 2.3 excess cases of ICH, 13.0 excess cases of major extracranial bleeding, 9.4 excess cases of major gastrointestinal bleeding and 3.1 excess deaths per 1000 person-years compared with those who received dabigatran. The researchers said the net increase in intracranial hemorrhage exceeded the net reduction in thromboembolic stroke for patients taking rivaroxaban.

The researchers cited a few limitations of the study, including that the mean duration of follow-up was less than four months. The observational design also meant the study could have been subject to residual confounding, which could lead to biased estimates of risk. In addition, the study only analyzed patients who were 65 or older, so the results might not be generalizable to younger patients.

“The greater anticoagulant effect observed with rivaroxaban treatment may be due to the higher dose required for once-daily dosing,” the researchers wrote. “A contribution to this effect by off-label use of standard-dose rivaroxaban in patients with impaired renal function cannot be excluded.”