ATLAS shrugged off: FDA panel rejects rivaroxaban for ACS patients
 |
During an all-day discussion on May 23, the panel committee weighed multiple issues surrounding the design and the execution of the trial, with particular focus on dosing, labeling and rivaroxaban’s use with a thienopyridine.
About the two rivaroxaban doses assessed in the trial, there was a “strong consensus” among panelists about the benefits of evaluating both 2.5 mg BID and 5 mg BID, as noted by acting committee chairperson Philip Sager, MD, of the FDA’s CRDAC. Specifically, Debra Madden of Newtown, Conn., who served as the patient representative on the panel, commended the study’s sponsor for “asking a question with clinical relevance to patients by evaluating and recognizing that a higher dose may be too much for patients.”
Sanjay Kaul, MD, director of the cardiovascular diseases fellowship training program at Cedars-Sinai Heart Institute in Los Angeles, added the caveat that due to the “missing data, it remains to be seen how the 2.5 mg BID dose group truly did.” For myriad reasons, the panel resoundingly supported a greater reliance on intention-to-treat (ITT) trials, as opposed to the modified intention-to-treat evaluation used in this trial. While the FDA had initially recommended an ITT analysis, the agency eventually accepted the modification from the sponsor while the trial was still ongoing.
About the missing data, Kaul expressed his “concern” about the rate, compared with the APPRAISE-2 trial. While the sensitivity analysis conducted by the sponsor of this trial seemed to quell some of the panel’s concerns, there was consensus that the missing data coupled with a lack of robustness of the endpoints was “problematic,” summed Sager.
Missing data is “a significant issue in this trial,” said Thomas R. Fleming, MD, a biostatistician at the University of Washington in Seattle, adding that the optimal time to address patient withdrawal is during the trial. “Thus, a plan for following up with patients who withdraw from trials needs to be proactively addressed at the design protocol stage, and not retrospectively.”
Kaul and Steven Nissen, MD, of the Cleveland Clinic, tipped their hats at their negative votes during the panel discussion when they said the bleeding risks were too strong to overcome the mortality benefits. Also, Mori Krantz, MD, director of the Colorado Prevention Center at University of Colorado in Denver, spoke to the devastating impact that stroke on patients.
The FDA is not required to take the recommendations of the panel, but it most often does.