By Michael Crawford, MD

Professor of Medicine, Chief of Clinical Cardiology, University of California, San Francisco

Dr. Crawford reports no financial relationship relevant to this field of study.

SYNOPSIS: A large Phase IV registry study shows that rivaroxaban is associated with a very low incidence of major bleeding, death, or stroke. Also, adherence to therapy was much higher than observed in other studies with vitamin K antagonists.

SOURCE: Camm AJ, Amarenco P, Haas S, et al. XANTUS: A real-world, prospective, observational study of patients treated with rivaroxaban for stroke prevention in atrial fibrillation. Eur Heart J 2016;37:1145-1153.

Although approved for stroke prevention in non-valvular atrial fibrillation (NVAF), there has been controversy regarding the safety of rivaroxaban. Investigators in Europe conducted a prospective, observational outcome study of rivaroxaban use, which assessed its safety and efficacy in NVAF. The treating physician made decisions as to the dose of rivaroxaban. Follow-up was for one year or 30 days after the last dose of rivaroxaban if it ended earlier. The primary outcome was related to safety and included major bleeding and all-cause death. Secondary outcomes included thromboembolic (TE) events and non-major bleeding. Intracranial bleeding was included in TE and major bleeding events. A central committee adjudicated all endpoints. From 311 centers largely in Europe, Israel, and Canada, 6,784 patients were included, of whom 79% took rivaroxaban 20 mg/day, 21% 15mg/day, and 0.5% other doses. Mean patient age was 72 years, with 37% aged older than 75 years. Comorbidities were common, and 18% had new atrial fibrillation, 41% had paroxysmal atrial fibrillation, and 41% had persistent or permanent atrial fibrillation. Ninety-six percent of patients did not experience any major adverse events (death, stroke, bleeding). There were 2.1 major bleeding events per 100 patient-years; 1.9 died and 0.7 experienced a stroke. Adverse events were more common in patients with more comorbidities, especially reduced renal function despite rivaroxaban dose reduction. Major bleeding usually was treated conservatively, with few receiving non-specific reversal agents. At the end of the observation period, 20% had discontinued treatment mainly for apparent adverse events. The authors concluded that rate of stroke and major bleeding was low in a broad NVAF population treated with rivaroxaban.


Randomized, controlled trials (RCTs) are expensive, but necessary for new drug approval. They also are highly selective in patient enrollment to minimize the effects of other patient variables on the results. However, once a drug is approved, there is a need for real-world, less selective data to more clearly define the safety and efficacy of new drugs. Such studies are mandated by the European Medicines Agency and generally take two forms. Large administrative database studies often involve entire European countries. Denmark recently has conducted many such studies. They are very large and less expensive, but they are often retrospective and provide less detailed information. Registry studies provide more detailed information on smaller, more select populations. Also, registry studies require patient consent, so there are fewer cognitively impaired subjects as compared to administrative databases. This study of rivaroxaban is a registry study and is the first such study conducted on the new oral anticoagulants.

Patients in this registry study were lower risk than those in the Phase III ROCKET-AF trial that secured FDA approval for rivaroxaban; the mean CHADS2 score for this study was 2 and was 3.5 for ROCKET-AF. This is not unusual for registry studies. RCTs want to include patients more likely to experience events so researchers can detect differences more readily. This registry study featured several strengths. First, it was prospective. Second, it used uniform definitions of events. Third, there was an independent event adjudication committee. However, since patients have to agree to participate, there may be residual confounding. Of interest, two-thirds of patients screened enrolled, suggesting this was a fairly representative sample.

In general, the results showed a low incidence of adverse events in patients treated with rivaroxaban. Compared to ROCKET-AF, there were fewer strokes (0.7 vs. 1.7 per 100 patient-years), major bleeds (2.1 vs. 3.6), and major gastrointestinal bleeds (0.9 vs. 2.0). Fatal bleeds (0.2 vs. 0.2) and intracranial hemorrhage were similar (0.4 vs. 0.5). Also, persistence of therapy was quite good (80% at one year compared to 62% in similar studies with vitamin K antagonists). It is important to use drugs that patients will actually take to prevent strokes.

One concerning finding was a systematic under-dosing in the study; 15% of patients were on a lower-than-recommended dose of rivaroxaban for no apparent reason. This is not unique to this study and has occurred in studies with dabigatran and apixaban. This doesn’t surprise me as I have been tempted to use a lower dose in very old, frailer patients because I am afraid of bleeding but don’t have the courage to withhold anticoagulants. Also, I have seen patients lower their dose after the first bruise on their arm appears. Many have been scared by the commercials on television, both from pharmaceutical companies and trial lawyers. Clinicians must use this encouraging data to counter these forces and make sure appropriate patients with NVAF take these drugs at appropriate doses.