Will Rivaroxaban ROCKET Into The Atrial Fibrillation World?... Or Sputter Out?

Written By

Jill S. Borchert, PharmD, BCPS, FCCP

Reviewed By

Henry I. Bussey,Pharm.D., FCCP, FAHA
Katie Kiser, Pharm.D., BCPS


Patel MR, Mahaffey KW, Garg J, et al. Rivaroxaban versus warfarin in nonvalvular atrial fibrillation. N Engl J Med 2011 Aug 10.

The wait for alternatives to warfarin for stroke prevention in atrial fibrillation has been a long one. Patients and clinicians have long noted the limitations of warfarin, with dietary restrictions, drug-interactions and frequent monitoring.  Last year, dabigatran became available as an alternative based on the results of the RE-LY trial.1 Recently, the results of the ROCKET-AF trial provide us with insight into another potential alternative, the oral factor Xa inhibitor rivaroxaban.2  Rivaroxaban is currently FDA-approved for prevention of thromboembolism in hip and knee replacement surgery.3

The Rivaroxaban Once Daily Oral Direct Factor Xa Inhibition Compared with Vitamin K Antagonism for Prevention of Stroke and Embolism Trial in Atrial Fibrillation (ROCKET-AF) was a multi-center, randomized, double-blind, double-dummy study comparing oral rivaroxaban to dose-adjusted warfarin in moderate-to-high risk patients with nonvalvular AFib. The ROCKET-AF trial included over 14,000 patients assigned to either rivaroxaban, dosed as a 20 mg daily or adjusted to 15 mg daily for patients with a creatinine clearance of 30-49 ml/min, or warfarin titrated to a goal INR of 2.0-3.0. Moderate-to-high risk was defined by a history of stroke or transient ischemic attack or a CHADS2 score of at least 2 (i.e. two or more of the following: heart failure, ejection fraction ≤35%, hypertension,age ≥75, diabetes). To limit the number of patients on the lower end of the stroke risk cut off, the protocol only allowed up to 10% of patients in each region to meet the minimum criteria of CHADS2 score of 2 and the remainder of the patients must have had a score of 3 or higher. Investigators hypothesized that rivaroxaban would be noninferior to warfarin in the composite primary outcome of stroke (ischemic and hemorrhagic) and systemic embolism.  The pre-specified primary analysis was a per-protocol analysis including patients with data regarding event rates who took at least one dose of study medication and without major protocol violations. Authors pre-specified that they would conduct a superiority analysis if noninferiority was met.  A post hoc intention-to-treat analysis was also performed.

The inclusion criteria resulted in a patient population that was 40% women with a median age of 73. Approximately 80% of patients had persistent AFib and 62-63% of subjects were previously on vitamin K antagonists before enrollment. The mean CHADSscore was 3.5 with 13.0% of the population with a score of 2, 86.9% with a score of 3 or higher and 43.4% with a score of 4 or higher. INR mean time in therapeutic range was 55%.

After median follow-up of almost two years, the results regarding the event rates per patient-year for the primary outcome measure (stroke and systemic embolism) were:








Primary Analysis


(188 events)

(241 events)

< 0.001 for noninferiority

Post hoc intention-to-treat



< 0.001 for noninferiority


(269 events)

(306 events)

0.12 for superiority




0.02 for on-treatment superiority analysis

The rate of major and non-major clinically relevant bleeding was not statistically significantly different in the rivaroxaban and warfarin groups however intracranial hemorrhage and fatal bleeding were significantly lower in the rivaroxaban group. Major GI bleeding was more common in the rivaroxaban group when compared to warfarin.





Drug Favored

Major and clinicallysignificant non-major bleeding

14.9% per patient-year

14.5% per patient-year



Major bleeding

3.6% per patient-year

3.4% per patient-year



Drop in hemoglobin of 2 g/dL or more

2.8% per patient-year

2.3% per patient-year



Fatal bleeding

0.2% per patient-year

0.5% per patient-year



Intracranial hemorrhage

0.5% per patient-year

0.7% per patient-year



Major GI bleeding





So, what do these numbers mean? We can conclude that rivaroxaban is noninferior to warfarin, but how do we make decisions in individual patients? To a clinician, it would be nice to know how outcomes varied based on individual time in therapeutic range (iTTR).  In other words, how do patients on warfarin with poor INR control fare and how do those with good INR control fare in comparison to rivaroxaban? While authors reported that there was no difference in the effect of rivaroxaban when the time in therapeutic range by center (TTRc) were placed in into quartiles (centers with poor anticoagulation control versus centers with better anticoagulation control), no data on iTTR is available. Good INR control by center does not necessarily mean good INR control for most patients at that center as a few well controlled patients can skew the data. Studies confirm that thromboembolic and bleeding event rates are lower for patients with stable INRs within the therapeutic range for 6 months or 12 months versus those who are not stable.4,5 Data on outcomes by iTTR quartiles would help the clinician weigh clinical decisions of switching patients either currently well-controlled or poorly controlled on warfarin to rivaroxaban. We may also consider the time in therapeutic range in this trial (55%) is lower than other studies in atrial fibrillation. How does this compare to patients in your anticoagulation clinic or patients in other trials? Other considerations may be individual patient risk factors. Patients in ROCKET-AF and RE-LY on rivaroxaban and dabigatran, respectively had higher rates of GI bleeding than those on warfarin.1,2  Therefore, careful consideration of an individual patient’s GI risk might lead to a continuation of warfarin.

As clinicians, we often encounter patients who will not (or perhaps cannot) take warfarin. We may soon have two alternatives to warfarin for stroke prevention in atrial fibrillation. Reflecting on the lack of head-to-head studies comparing rivaroxaban to dabigatran, how can we approach making a decision between the two agents? A big consideration will be the stroke risk of the individual patient. The patient population in ROCKET-AF (rivaroxaban) had a higher stroke risk overall than the RE-LY trial (dabigatran).1,2 Almost one-third of patients in the RE-LY trial had a CHADSscore of 0 or 1, where current ACCP guidelines recommend aspirin or warfarin as appropriate treatment options.4 In patients who have a higher stroke risk, we might favor rivaroxaban based on the study population. Others may argue that dabigatran at the 150 mg twice daily dose was superior to warfarin in the superiority analysis, while rivaroxaban was not superior to warfarin; however, without a head-to-head trial it is difficult to make any conclusions about superiority – especially given the differences in patient populations. Another consideration is cost. Given that both rivaroxaban and dabigatran were noninferior to warfarin, our best bet is sticking with warfarin and doing our best to keep the INR well-controlled. But, what would you do if a patient can't or won't take warfarin? If I were forced to make a decision today, I would go with rivaroxaban (over dabigatran) because the patient population studied had a higher stroke risk than the populations in the dabigatran studies. What would you do and why? What other factors would you consider?



<p>For those precepting students, being able to discuss RELY, ROCKET-AF, and now ARISTOTLE present wonderful teaching points in evaluating the medical literature and practicing EBM.&nbsp; I agree with the author in the sense that dabigatran vs rivaroxaban is an "apples to oranges" comparison but it is one that clinicians "may" be forced to make (if rivaroxaban is approved) and I thank her for her wonderful commentary.</p><p>Personally, I struggle accepting the "practical" trial explanations regarding TTR in this trial - perhaps I am too spoiled having only experienced pharmacist-run OAC settings where studies indicate that TTR is higher or equivalent to usual care (as has been investigated by one of the reviewers).&nbsp; I am also slightly concerned that there was such a lag time between announcement of the results (Nov 2010) and the publication almost 10 months later.&nbsp; Given the evidence I think I have to reserve rivaroxaban only to a patient population that would have difficulty with adherence to a twice daily medication (dabigatran) and that is having difficulty maintaining INR values (regardless of CHADS2 risk score).</p><p>However, now that ARISTOTLE has been published our choices may include three new therapies as opposed to the two described above.&nbsp; Could "the golden mean of anticoagulation" be here with apixaban instead?&nbsp; Personally I look forward to having many more of these conversations in the future.</p>