ATLAS ACS trial in hot seat at FDA panel, mainly about missing data

Twitter icon
Facebook icon
LinkedIn icon
e-mail icon
Google icon
dark - 28.02 Kb

SILVER SPRINGS, Md.—The FDA’s Cardiovascular and Renal Drugs Advisory Committee, in a May 23 meeting to assess new indications for rivaroxaban (Xarelto, Bayer HealthCare/Janssen Pharmaceuticals), kicked off with the ATLAS ACS trial investigators and sponsor addressing the FDA’s trial concerns, specifically related to patient withdrawal from the trial and missing data.  

The committee is reviewing a supplemental new drug application for new indications for rivaroxaban: reducing the risk of thrombotic cardiovascular events in patients with acute coronary syndrome (STEMI, non-STEMI or unstable angina) in combination with aspirin alone or with aspirin plus clopidogrel or ticlopidine. The FDA's priority review for this indication is based on the results of the ATLAS ACS trial, which randomized 15,526 patients.

In its briefing documents released prior to the panel meeting, FDA’s Karen A. Hicks, MD, recommended approval of Xarelto  (2.5 mg po BID) to reduce the risk of cardiovascular events in patients with ACS for the indications above. However, she added that “[s]hould additional information become evident in the next two months that affect overall trial interpretability, I may choose to reconsider this recommendation.”

In his opening remarks, Norman Stockbridge, MD, PhD, director of the division of cardiovascular and renal products, Office of Drug Evaluation at the FDA, reviewed some of the FDA’s concerns with the trial:

  • If a trial with multiple doses can be appropriately powered;
  • Differing patient populations, on whom the likely effects of the drug would be quite different and are readily distinguishable at baseline;
  • Trial execution, including “now missing data”; and
  • Trial results, according to varying interpretations of net clinical benefit.

Finally, Stockbridge questioned whether one study without a significant p-value suffices in this setting, which he defined as the “most difficult matter.”

In his review of the trial’s efficacy data, ATLAS ACS principal investigator C. Michael Gibson, MD, of Beth Israel Deaconess Medical Center in Boston, addressed some of the concerns about the patient follow-up. The follow-up across the three study arms was approximately 85 percent. Specifically, there were 1,296 patients (8.3 percent of the trial participants) who withdrew consent early.

Based on these withdrawals, the investigators initiated a two-phase plan in February 2011 to ascertain vital status on these subjects at the study end. The investigators were given permission to follow up with 399 patients, 183 of whom signed permission status and only 177 were reached for vitals status—all of whom were alive.  

Echoing the sentiments of his fellow panel members about the missing data, Scott Emerson, MD, PhD, from the department of biostatistics at the University of Washington in Seattle, said, “It’s not the question about what happened with the data we have, but the bigger question is what happened in the patients with the missing data.”

Assuming the “worse-case scenario” that all patients whom the investigators could not reach had died, Gibson said that the mortality rate in the trial would have been similar to the placebo arm of the trial. However, because they were able to track when the patients withdrew, and most mortalities in the trial occurred in the first two weeks, that would be “unlikely,” he noted.

When questioned by the FDA’s Robert Temple, MD, deputy director for clinical science at the Center for Drug Evaluation and Research, whether they sought to find out anything about the patients through public records, Paul Burton, MD, PhD, vice president of cardiovascular department at Janssen Research & Development, said that they did not, per the trial’s predesignated consent withdrawal initiative.

Also, Steven Nissen, MD, of the Cleveland Clinic, questioned why only clopidogrel and ticlopidine were the antiplatelet option, and why prasugrel (Effient, Daiichi Sankyo/Lilly) was not used in the trial. If approved, should the rivaroxaban label be indicated for use with prasugrel, he asked. Gibson responded that the “prasugrel label suggests that it should not be used with an anticoagulant, which is why the decision was made not include it in the trial. So, we don’t have any data available on the use with prasugrel.”

Also responding, Eugene Braunwald, MD, of the TIMI Study Group at Harvard Medical School, said this was a “key question [among the study investigators], and one we discussed a lot.