Intended for healthcare professionals


Data on trial of anticoagulant is to be reanalyzed after discovery that investigators used faulty device

BMJ 2015; 351 doi: (Published 03 December 2015) Cite this as: BMJ 2015;351:h6431
  1. Deborah Cohen
  1. 1The BMJ

Data from the key phase III trial that allowed the oral anticoagulant rivaroxaban on to the market for the prevention of ischemic stroke in non-valvular atrial fibrillation is being reanalyzed because of the use of a faulty device.

The European Medicines Agency has also told The BMJ that it has launched an investigation into the trial.

The ROCKET-AF trial, published in the New England Journal of Medicine in 2011, compared rivaroxaban, which is manufactured by Bayer and marketed in the United States by Janssen, part of Johnson and Johnson, as Xarelto, with warfarin in more than 14 000 people.1

The trial used the INRatio device, made by Alere, to determine whether, and by how much, patients’ doses of warfarin should be adjusted to keep the drug in the right therapeutic range. Too much warfarin and patients are at an increased risk of bleeding; too little and they are at risk of clots.

However, The BMJ has learnt that the point of care device used in the trial to monitor the international normalized ratio (INR) of warfarin in participants who took it has been subject to a recall. In December 2014 a recall notice said that certain INRatio devices could deliver INR results that were “clinically significantly lower” than a laboratory INR method. It said that Alere had received 18 924 reports of malfunctions, including 14 serious injuries. A falsely low reading could mean that participants had their warfarin unnecessarily adjusted and that too much could be given. Such adjustment could have led to an increased risk of bleeding in participants in the warfarin arm of the trial.

In terms of the trial results, it could make rivaroxaban seem better than it was at reducing the risk of bleeding.

The BMJ alerted Janssen and Bayer about the issue more than two months ago, and the US independent watchdog Project on Government Oversight has since reported that the trial data were being reanalyzed by Duke Clinical Research Institute at Duke University, Durham, North Carolina.

In September this year a spokesman for Johnson and Johnson said that they “were unaware of this recall.” He added, “We take this seriously and have reached out to the manufacturer for more information.”

The Project on Government Oversight also reported that the 2014 recall notice did not require that Alere retrieve the devices from patients. Instead, the recall notice said that patients with specific medical conditions should stop using the recalled devices and that other patients should have their INR readings verified by laboratory tests. Alere told the Project on Government Oversight that it was working on a software upgrade to fix the inaccuracies.

Even before the recall, the ROCKET-AF trial had been criticized for the relatively short time that participants taking warfarin were in the correct therapeutic range of the drug. The implication of the criticism was that patients’ use of warfarin was not adequately controlled, which, if correct, would have the effect of making rivaroxaban seem safer and more effective.

At the time that the US Food and Drug Administration (FDA) approved rivaroxaban, one of the advisory committee members, the cardiologist Steven Nissen of the Cleveland Clinic, Cleveland, Ohio, said that the trial’s approach to warfarin treatment “was a fatal flaw in the study design.” Nissen also pointed out that other clinical trials of anticoagulants had managed warfarin more effectively.

An FDA reviewer also said in a memo to the FDA that the “poor warfarin control” in the ROCKET-AF trial “biased the study in favor of rivaroxaban.”

Other FDA advisory committee members, however, commended the fact that ROCKET-AF used warfarin as it would be used in the real world.

Robert Califf, President Barack Obama’s nominee for the head of the FDA, who co-chaired the executive committee of the trial, said that the researchers “gave warfarin not only in an acceptable way, we gave it in a commendable way during this trial.”

In response to The BMJ’s questions to Janssen and Bayer about the concerns raised, a spokesperson for Janssen said, “The FDA considered all input received from the independent reviewers and approved Xarelto based on the positive benefit-risk profile observed in the 14 264 patient ROCKET-AF study.

“We’re pleased with the FDA’s decision to approve this medicine that has helped millions of people around the world with non-valvular atrial fibrillation reduce their risk of stroke. Additionally, as a complement to our phase III studies, we’ve evaluated more than 91 000 patients across our six indications in global real-world settings, and study after study continues to confirm the positive benefit-risk profile.”

Bayer has not responded to The BMJ’s request for a comment.


Cite this as: BMJ 2015;351:h6431



View Abstract