CHICAGO—Rivaroxaban was non-inferior to warfarin for the prevention of stroke and non-CNS systemic embolism for patients with non-valvular atrial fibrillation, but did not achieve superiority by the intention-to-treat analysis, according to the late-breaking ROCKET AF trial, which was presented today at the American Heart Association (AHA) Scientific Sessions.
In October 2006, the FDA issued a black box warning for warfarin due to a growing understanding of its hazards in routine clinical practice. The requirement of INR monitoring has relegated millions of individuals to no therapy or ineffective therapy because of a lack of access to monitoring, explained Elaine M. Hylek, MD, from Boston University School of Medicine, who commented on ROCKET AF at this morning’s press conference. “These complications have led to an intense search for alternatives to warfarin with a more predictable dose response that obviates the need for monitoring,” she said.
|Slideshow | ROCKET AF|
|Kenneth W. Mahaffey, MD and Keith AA Fox, MD on behalf of the ROCKET AF investigators
In ROCKET AF (Rivaroxaban Once-daily oral direct factor Xa inhibition Compared with vitamin K antagonism for prevention of stroke and Embolism Trial in Atrial Fibrillation), investigators randomized 14,269 patients with AF and a history of stroke or at least two risk factors at 1,215 participating sites in 45 countries to receive either rivaroxaban (20 mg daily) or dose-adjusted warfarin (target INR range 2.5) in a double-blind fashion.
The primary efficacy endpoint was all adjudicated strokes (ischemic and hemorrhagic) and systemic embolic events. The primary analysis—measuring what happened to the study participants while they were actually taking the drug—was based on establishing non-inferiority in the per-protocol population.
The median age of the study population was 73 years, 40 percent were female, 63 percent had heart failure, 90 percent had hypertension, 40 percent had diabetes and 55 percent had a prior stroke or TIA. Also, the intrinsic stroke risk of enrolled patients was “high,” with a majority with a CHADS2 score of more than three, said co-principal investigator Robert M. Califf, MD, from Duke University in Durham, N.C., who presented the trial.
He noted that this was a very high-risk patient population who were older and had multiple co-morbidities. Hylek concurred, comparing the baseline characteristics to previous trials, including SPORTIF III/IV, ACTIVE W and RE-LY.
In the primary analysis, those taking rivaroxaban compared with warfarin had fewer strokes and emboli to other parts of the body. With rivaroxaban, there were 1.71 events per 100 patient-years (188 patients) and with warfarin 2.16 (241 patients). Major bleeding complications were comparable in both treatment groups, occurring at a rate of 3.60 per 100 patient-years (395 patients) with rivaroxaban and 3.45 per 100 patient-years (386 patients) with warfarin.
The full intention-to-treat analysis, which counts all events from the time of randomization until study completion, regardless of whether participants were taking medication found fewer strokes and emboli occurred in participants assigned to rivaroxaban (296 patients) versus warfarin (306 patients). However, this difference fell short of statistical significance for declaring rivaroxaban superior to warfarin.
Death from any cause occurred slightly less frequently among rivaroxaban patients at a rate of 4.52 per 100 patient-years (582 patients) versus 4.91 per 100 patient-years (632 patients) among those on warfarin.
Hylek pointed out that intention-to-treat analyses avoid the bias associated with non-random loss of participants, but may bias toward equivalence due to non-adherence. She said that “per-protocol” or “on-treatment” analysis is important to confirm non-inferiority of a primary intention-to-treat analysis.
In this trial, Hylek said that the difference between the intention-to-treat and per-protocol analysis may be accounted for by poor adherence, which “raises concerns about the relevance of the per-protocol analysis to real-world practice, particularly for a drug with a half-life of five to 13 hours versus 20 to 60 hours for warfarin.” Therefore, she said the intention-to-treat result showing no significant superiority is more likely to reflect the actual difference in effectiveness between these treatments. She also noted that the overall safety profile was “less clear.”
Califf said that the study of