Much discussion during Wednesday’s advisory panel meeting on whether to approve Bayer/J&J anticoagulant Xarelto (rivaroxaban) for a new use centered around a pivotal trial called ATLAS. Advisors complained of missing data and inconsistent record keeping (click here for a recap). A final decision is expected in late June.

In this Q&A, Dr. Steven Nissen, a panel member and chairman of the department of cardiovascular medicine at the Cleveland Clinic Foundation, explains why he voted against approving Xarelto for use in acute coronary syndrome (ACS) patients already taking dual antiplatelet therapy, and how the problems plaguing ATLAS are disturbingly common.

MM&M: What was it about the benefit-risk profile that swayed you to vote that way?

Dr. Steven Nissen: Marginal statistical evidence of benefit; a ‘p’ value of 0.039 for benefit at the target dose of 2.5mg b.i.d.; increased risk of intracranial hemorrhage and life-threatening bleeding; and a study that had issues with respect to the quality of study conduct, particularly the high rate of withdrawal of consent that undermined the ability to interpret the results of the study.

The sponsor has vowed to supply the FDA with the data that the panelists said was missing. What will the companies have to show for you to be convinced?

Another trial, which is what I said.

The panelists seemed taken aback by Bayer/Janssen’s data omission. What was it about the companies’ preparation for the meeting that inspired such passion among the panel members?

Look, this is not personal. So I wouldn’t make those sorts of comments. I would just simply say that we were not convinced that the benefits exceeded the risks. And we wanted more information. And I will leave it at that.

Much discussion was focused around, as you said earlier, issues that related to data collection, maybe some would say poor data collection. The rejection of the drug, albeit a narrow one, was almost a condemnation of that practice. Do you agree or not?

Well, the problem of patients withdrawing consent or incomplete ascertainment of outcomes is a problem that extends across the entire clinical trial enterprise.  And certainly the committee was making a generalized statement that we are worried about it, and that we think it needs to be addressed in this application, and in future applications for drugs.

And another data question: Three study sites were eliminated due to misconduct, with the company claiming it couldn’t find the subjects because privacy laws barred them from doing things like checking death notices or contacting them. The FDA person said that if the local laws are getting in the way of clarity, don’t have trials there. Increased use of international locations is a growing trend. Can the FDA really bar their use and kind of impose safeguards, or will it need to just fall in line with the data it’s given?

I think that that is the choice of the sponsor. The FDA…the companies are taking their own risks if they go to places that they are not confident of the quality of the data. It is their risk to take. It is not the FDA’s job to tell them not to take the risk. But companies need to know that if they don’t supply data that the FDA or its advisers believe is reliable, then it may impact approvability.

And one final question: Do you think that drug will eventually reach the market for the indication in ACS?

I have no way to know. I think they need another trial and if it’s successful then it will convince many of us. If it’s unsuccessful, it won’t make it. It’s just speculation entirely.