But physicians who work for J&J said the adverse events observed in the study wouldn't be mirrored in the real world because patients generally don't just stop taking a blood thinner cold-turkey as they did in the study. In the "real world," patients are transitioned more seamlessly onto a different anticoagulant, the company said.
An FDA reviewer and panelists were also concerned that warfarin wasn't administered ideally in the trial, causing rivaroxaban to appear more effective than it may have been had warfarin been dosed properly. Just 56 percent of patients who received warfarin in the ROCKET-AF were in the ideal INF range of 2 to 3 during the study, which the panel criticized as being way too low.
"INR control in ROCKET was worse than in other recent trials, which might have biased the overall results in favor of rivaroxaban," said FDA reviewer Dr. Martin Rose, who authored a highly critical review of rivaroxaban that was released several days before Thursday's meeting.
One of those recent trials is the RE-LY trial, which found that the direct thrombin inhibitor dabigatran (Pradaxa) was more effective in preventing strokes in high-risk patients than warfarin.
But panel chairman Dr. Michael Lincoff, another cardiologist from the Cleveland Clinic, said the spotty warfarin dosing in ROCKET-AF was acceptable because it was similar to how warfarin is used outside of clinical trials.
Dabigatran was a constant topic of conversation at the meeting, with panelists and the FDA making comparisons between rivaroxaban and dabigatran, despite the fact that no trial has yet been conducted pitting the two anticoagulants head-to-head.
"Any current expert will say it's important to do head-to-head clinical trials," said Dr. Robert Califf, director of Duke University's Translational Medicine Institute and a principal investigator in the ROCKET-AF trial.
Some panelists questioned the need for another blood thinner that was shown to be noninferior to warfarin, when dabigatran was shown to be superior to warfarin.
But others at the meeting -- including a patient representative -- said more oral anticoagulants are needed because warfarin is not always well-tolerated and dabigatran can cause gastrointestinal side effects.
"All afib patients are aware of the three drugs we need to take," said patient representative Debra McCall of California. "A drug to regulate rhythm, a drug to regulate rate, and a blood thinner. We have options for rate and rhythm, but we don't have many options for blood thinners."
Following the committee vote the American Heart Association released a statement noting it will review its guidelines for anticoagulation therapy in patients with nonvalvular atrial fibrillation if the FDA does decide to approve the drug for stroke prevention in that population.
"For the millions of patients with atrial fibrillation, stroke is a real health threat, and the emerging studies of new anticoagulant drugs like rivaroxaban have been very encouraging," said Dr. Gordon Tomaselli, president of the American Heart Association.
The FDA is expected to make an approval decision on rivaroxaban by Nov. 4. The agency isn't required to follow the advice of its advisory committees, but it often does.