In patients with nonvalvular atrial fibrillation (A-fib), permanent discontinuation and even temporary interruption of anticoagulant therapy substantially raise the odds of stroke or systemic embolism. But the risk is similar for patients receiving either the novel factor Xa inhibitor rivaroxaban or warfarin, according to a post hoc analysis of the ROCKET AF trial published in the February 12, 2013, issue of the Journal of the American College of Cardiology.

Investigators led by Manesh R. Patel, MD, of the Duke Clinical Research Institute (Durham NC), looked at 3 groups of trial participants who:

• Temporarily interrupted rivaroxaban or warfarin for more than 3 days (n = 3,734 for rivaroxaban, 4,511 for warfarin)
• Permanently discontinued the study drug early (n = 2,470 for rivaroxaban, 2,425 for warfarin)
• Completed the study without suspected stroke or systemic embolism (n = 4,587 for rivaroxaban, 4,652 for warfarin)

For the primary efficacy endpoint, a composite of all stroke and non-central nervous system (CNS) systemic embolism, no difference was seen between the study arms after either temporary interruption or early permanent discontinuation. However, in the cohort that completed the study and transitioned to open-label therapy, more events occurred in the rivaroxaban arm. Likewise, when patients in all 3 scenarios were combined, rivaroxaban was associated with increased events (table 1).

Table 1. Primary Efficacy Endpoint: Events per 100 Patient Years

The same pattern was seen for the aggregate of all thrombotic events (stroke, non-CNS embolism, MI, and vascular death) within 30 days, except there was no difference between rivaroxaban and warfarin when the 3 scenarios were combined.

Risk of major bleeding was similar for the 2 anticoagulants after temporary interruption (HR 1.02; 95% CI 0.59-1.77; P = 0.94) and early permanent discontinuation (HR 0.60; 95% CI 0.35-1.04; P = 0.067) but higher with rivaroxaban during the post-study transition period (HR 3.62; 95% CI 1.56-8.36; P < 0.0026). No difference was seen in the combined cohort.

ROCKET AF randomized 14,264 patients to fixed-dose rivaroxaban (20 mg or 15 mg daily) or dose-adjusted warfarin (target international normalized ratio [INR] 2.0-3.0). Results published in the New England Journal of Medicine in September 2011 showed rivaroxaban to be noninferior to warfarin for the primary efficacy endpoint of stroke or systemic embolism. A spike in thrombotic events with rivaroxaban at the end of the trial, however, raised concern about a possible ‘rebound’ effect with the new drug, prompting the US Food and Drug Administration to mandate a black box warning on the drug’s label about discontinuation.

‘Bad Things’ Occur Without Anticoagulation

What comes through most clearly from the new analysis, writes Matthew R. Reynolds, MD, MSc, of the Harvard Clinical Research Institute (Boston, MA), in an accompanying editorial, is that “bad things will happen to high-risk [A-fib] patients if they are left untreated with effective anticoagulant therapy for sustained periods.”

In a telephone interview with TCTMD, Michael D. Ezekowitz, MD, PhD, of Thomas Jefferson Medical College (Philadelphia, PA), agreed, pointing out that the event rate among patients who either interrupted or discontinued rivaroxaban was several times greater than that for the overall rivaroxaban cohort.

Rebound Not the Culprit

As for the higher event rate for rivaroxaban at the end of the study, Dr. Ezekowitz said he did not believe in a so-called rebound phenomenon. “The simple fact is that these patients were not protected and therefore events increased,” he emphasized.

Calling this “a technical error” in the trial protocol, Dr. Ezekowitz explained, “Patients randomized to warfarin just continued on warfarin, so there was no interruption in their protection, whereas patients randomized to rivaroxaban, which is a short-acting drug, experienced a gap in anticoagulant coverage between termination of rivaroxaban and the initiation of warfarin. Also [as the authors point out], it takes time to reach a therapeutic range with warfarin.”

This hiatus in protection is a problem inherent in any drug transition and not peculiar to rivaroxaban, he noted.

Dr. Reynolds describes this explanation, also put forward by the authors, as “fairly persuasive” but cautions that INRs were not carefully collected during the transition period, precluding adjustment for time in therapeutic range. Moreover, no information was provided about use of bridging therapies, and the excess bleeding is “counterintuitive. . . . The investigators will need to scrutinize this large trial database further to understand these issues more fully,” he concludes.

Dr. Ezekowitz, however, speculated that the excess bleeding might be due to the unpredictability of response to warfarin. Patients starting on the drug may have high or low INR, he noted, which would account for both high thrombotic and high bleeding rates—with those ‘conflicting’ events occurring in different individuals.

Transition Uncommon in Practice

In real-world practice, it would be unusual to convert a patient from rivaroxaban to warfarin, Dr. Ezekowitz observed, citing as possible reasons deterioration of renal function or an upcoming surgical procedure involving a mechanical heart valve or other artificial surface. In the latter instance, clinicians might be wary of rivaroxaban since there is little experience with it in that setting, he explained.

One strategy for minimizing the time a patient is unprotected during a transition between drugs is bridging, he said, which entails use of an alternative anticoagulant with a rapid onset and offset of action, such as heparin. “But in the new era, all of the novel anticoagulants including rivaroxaban have [those qualities], so the only reason to use bridging with another agent is if the patient develops some contraindication to rivaroxaban,” he added.

Instead, “the message here is—unless there are important extenuating circumstances—do your very best to keep patients on their assigned medications,” Dr. Ezekowtiz advised. “And if there is an interruption, keep it as short as possible before an elective surgery and re-initiate therapy as soon as possible after the surgical interruption.”

Study Details

The most common reasons for temporary interruption (median 6 days) of therapy were surgical or invasive procedures (38.2%) and adverse events (40.2%). Permanent discontinuation was often triggered by adverse events (39%) but also by suspicion that a primary endpoint event had occurred (12.9%), withdrawal of patient consent (27.4%), or the investigator’s decision (7.4%). 

For temporary interruptions, investigators were instructed to stop warfarin or placebo 4 days before elective procedures and rivaroxaban or placebo 2 days before elective procedures.

 

---

Sources:
1. Patel MR, Hellkamp AS, Lokhnygina Y, et al. Outcomes of discontinuing rivaroxaban compared with warfarin in patients with nonvalvular atrial fibrillation: analysis from the ROCKET-AF trail (Rivaroxaban Once-Daily, Oral, Direct Factor Xa Inhibition Compared with Vitamin K Antagonism for Prevention of Stroke and Embolism Trial in Atrial Fibrillation). J Am Coll Cardiol. 2013;61:651-658.

2. Reynolds MR. Discontinuation of rivaroxaban: Filling in the gaps. J Am Coll Cardiol. 2013;61:659-660.

 

 

Related Stories:

• ROCKET AF Published: Rivaroxaban Matches Warfarin in Treating A-Fib Patients 
• ROCKET AF: Stopping Rivaroxaban or Warfarin Equally Harmful in A-fib Patients
• Rivaroxaban Continues Emergence of New Class of Anticoagulants for A-Fib