ATLAS Data Not Strong Enough To Support Xarelto’s Use In ACS, FDA Panel Says

Six of 11 members of the Cardiovascular and Renal Drugs Advisory Committee voted against rivaroxaban’s approval for acute coronary syndromes, saying concerns about missing data in the pivotal trial raised doubts about the efficacy results and could not support creating a new standard of care by adding the Factor Xa inhibitor to dual-antiplatelet therapy.

Concerns about missing data in the pivotal ATLAS trial led the majority of FDA’s Cardiovascular and Renal Drugs Advisory Committee to recommend against approval of Johnson & Johnson/Bayer HealthCare AG’s Factor Xa inhibitor Xarelto (rivaroxaban) for combination use with dual-antiplatelet therapy for treatment of acute coronary syndromes.

In their May 23 meeting, the committee voted 6-4, with one abstention, that rivaroxaban should not be approved for use in the treatment of ACS. Panel members voting against approval said there were too many questions about the efficacy and safety data from the 15,000-patient ATLAS trial due to the large amount of missing follow-up data, including vital status, on patients who discontinued the study.

Given the uncertainties as to whether the pivotal trial results could be trusted, panelists said the prospect of adding rivaroxaban to current dual-antiplatelet therapy (aspirin plus a thienopyridine) in ACS raised deep concerns given all the drugs’ propensity to increase bleeding.

“What we’re really saying if we say ‘yes’ is that patients with ACS should be treated with three drugs, not two, and the burden of proof here, particularly given the safety issues related to this class of drugs … [means that] we really want to have very compelling evidence,” said temporary voting member Steven Nissen, Cleveland Clinic Foundation. Nissen said he believed the ATLAS trial design was flawed and had the effect of encouraging patients to withdraw from study without being subject to proper follow-up.

Temporary voting member Sidney Wolfe, Public Citizen Health Research Group, said the level of incomplete data in ATLAS was worrisome, amounting to shaky evidence upon which to approve a drug that would become the standard of care in treating ACS.

“The implications of approving another drug … are that it would become the standard for everyone,” Wolfe said, “and this seems like a major, major leap forward given the fragility of the data.”

Even committee members who favored approval did so with some reservations about the missing data points and the establishment of a new standard of care.

The extent of missing data raised a cloud over the ATLAS results that is difficult to overcome, conceded temporary voting member Dean Follmann, National Institute of Allergy and Infectious Diseases. Follmann nevertheless favored approval, believing that more focused sensitivity analyses could overcome the data gaps.

Committee member Mori Krantz, University of Colorado, voted for approval but said he had a great deal of uncertainty about how rivaroxaban would be used and generalized in clinical practice. “This is not just a simple paradigm shift. This is really looking at going to triple therapy, which is really a three-headed monster in many ways. I think what you’re going to see in practice, if this is not done carefully with the proper labeling and secondary studies, is really dramatic magnification of bleeding and perhaps minimization of the efficacy benefits.”

The panel’s split vote was a reflection of the internal dissent within FDA on the rivaroxaban application. Clinical reviewer Karen Hicks recommended in favor of approval, while her medical team leader, Thomas Marciniak, argued that the amount of missing data essentially invalidated the study’s efficacy results.

Disputing The Effects Of Missing Data

J&J’s Janssen Pharmaceuticals Inc. division is seeking approval of rivaroxaban to reduce the risk of thrombotic CV events in patients with acute ACS – defined as ST elevation myocardial infarction (STEMI), non-STEMI or unstable angina – in combination with aspirin alone or aspirin plus clopidogrel or ticlopidine. The proposed dose is 2.5 mg twice daily.

If approved, the ACS indication would be the third for rivaroxaban. The drug gained its initial approval in July 2011 for prevention of deep vein thrombosis and pulmonary embolism after hip or knee replacement surgery (Also see "Xarelto Approval For DVT Bodes Well For Coming Indications" - Pink Sheet, 5 Jul, 2011.). In November, it added a claim for stroke risk reduction in patients with atrial fibrillation (Also see "Bayer/J&J's Xarelto Approval For Stroke Prevention Sets Up Marketing Battle With Pradaxa" - Pink Sheet, 4 Nov, 2011.).

In the ATLAS trial, ACS patients were randomized to rivaroxaban 2.5 mg twice daily, 5 mg twice daily or placebo; patients were stratified based upon whether they received concomitant aspirin alone (Stratum 1) or aspirin with a thienopyridine (Stratum 2). FDA has deemed Stratum 2 to be the most clinically relevant for U.S. ACS patients. Overall, rivaroxaban was effective in reducing the risk of the primary endpoint, a composite of cardiovascular death, non-fatal myocardial infarction and non-fatal stroke, in men and women, albeit with an increased risk of bleeding.

The issues surrounding missing data were a key focus of the advisory committee discussion, as they were in the FDA’s background briefing documents (Also see "Xarelto Acute Coronary Syndromes Claim Faces Data Quality Hurdle" - Pink Sheet, 21 May, 2012.).

Across all subjects in the trial, 2,402 (15.5%) discontinued the study prematurely, including 1,294 (8%) who withdrew consent. There were more than 1,000 subjects at the end of study with unknown vital status.

The high rate of missing data raised concerns as to whether there was informative censoring, meaning that patients who left study were different than those who stayed on treatment. Informative censoring can have the effect of making a drug look more efficacious than it really is because individuals who did poorly on treatment and withdrew are not included in the analysis.

J&J and ATLAS study officials described steps taken to try to obtain vital status follow-up in subjects who had withdrawn consent. They presented analyses showing that subjects who withdrew consent clinically resembled those who survived and did not disproportionately experience events that could lead to death in the period prior to their withdrawal of consent. The primary endpoint event rate in the rivaroxaban patients during the off-treatment period would have had to double, and the CV death rate would have had to increase four-fold, in order to overturn the positive efficacy results, the sponsor said. [Editor’s note: The previous sentence was updated on May 30, 2012 to correctly reflect the results of the sponsors’ sensitivity analysis.]

However, Cardio-Renal Medical Team Leader Thomas Marciniak was unconvinced, arguing that the study’s results could not be trusted due to the extent of the missing data. He asserted that the higher rate of bleeding observed with rivaroxaban, coupled with the incomplete follow-up rates, could explain the endpoint differences favoring the drug.

He also took issue with the sponsors’ failure to obtain vital status for all study subjects and how the sponsors counted deaths in some cases. ATLAS was a major CV outcomes trial, he noted, and one statistic that should be readily ascertainable is how many patients died and how many patients lived.