ROCKET-AF investigators are standing by the results of the pivotal trial of rivaroxaban in patients with nonvalvular atrial fibrillation, even though the anticoagulation monitoring device used in the control group was later found to be defective.

Results of a new analysis “are consistent with the overall trial findings and indicate that possible malfunction of the point-of-care device used for INR [international normalized ratio] measurement in the ROCKET AF trial that potentially led to lower INR values than would be obtained by laboratory testing did not have any significant clinical effect on the primary efficacy and safety outcomes in the trial,” the study investigators concluded in a letter to the editor of the New England Journal of Medicine published online on Feb. 3 that included the new analysis ( 2016 Feb 3;doi: 10.1056/NEJMc1515842 ).

On the same day, the BMJ published a feature on that very topic, saying that “doctors and scientists are calling for an independent investigation” and questioning the validity of the ROCKET-AF (Rivaroxaban Once Daily Oral Direct Factor Xa Inhibition Compared With Vitamin K Antagonism for Prevention of Stroke and Embolism Trial in Atrial Fibrillation) results until such an analysis is completed.

“A falsely low reading could mean that patients had their warfarin dose unnecessarily increased, leading to a greater risk of bleeding. In terms of the trial results, it could make rivaroxaban [Xarelto] seem safer than it was in terms of the risk of bleeding and throws doubt on outcomes used to support the use of the world’s best-selling new oral anticoagulant,” Dr. Deborah Cohen , associate editor of BMJ, wrote ( 2016;352:i575 doi: 10.1136/bmj.i575 ).

At the center of this controversy is the Alere INRatio Monitor System, which was used in ROCKET-AF, the pivotal trial comparing the safety and efficacy of rivaroxaban with warfarin for stroke prevention in patients with nonvalvular atrial fibrillation, to measure the INR in the patients receiving warfarin ( N Engl J Med. 2011;365:883-91 ). That device was recalled by the Food and Drug Administration and European regulators in 2014, 4 years after ROCKET-AF’s completion, because “it may provide an INR result that is lower than expected [compared with the] result obtained using a laboratory INR method,” particularly in patients with medical conditions such as “anemia, conditions associated with elevated fibrinogen levels, or unusual bleeding or bruising.”

In the new analysis, Dr. Manesh R. Patel of the Duke Clinical Research Institute in Durham, N.C., and his coinvestigators reviewed the medical records and any on-trial adverse events of the participants, looking for conditions that were identified in the recall of the INRatio device. They compared major efficacy and safety outcomes in the patients in the overall population (14,236 patients) first with those without any recall conditions (8,942, 63%), and second with those who had recall conditions (5,294, 37%).

Neither analysis revealed any significant differences in safety and efficacy outcomes from the overall population, with all groups showing the noninferiority of rivaroxaban to warfarin in preventing strokes, similar rates of overall bleeding, lower rates of fatal and intracranial bleeding, and higher rates of gastrointestinal bleeding. However, in the patients with any recall conditions, all measures of bleeding were higher in both the warfarin- and rivaroxaban-treated groups. Furthermore, the risk of major bleeding was higher in the rivaroxaban patients than in the warfarin patients, with a hazard ratio of 1.18 (P = .04).

“This finding does not support the hypothesis that device malfunction led to an increased risk of bleeding in the warfarin group of the trial,” Dr. Patel and his colleagues said.

But this conclusion is not shared by all. In the BMJ investigative article, former team leader in the FDA’s Cardiovascular & Renal Drug Products Division, Dr. Thomas Marciniak is quoted as saying that he would not rely on any reanalyses done by Duke, Johnson & Johnson, or the FDA, and that releasing the full datasets would be “the only solution that would lead to unbiased analyses.”

Furthermore, Dr. Harlan Krumholz , professor of medicine (cardiology) and director of the Center for Outcomes Research and Evaluation at Yale University, New Haven, Conn., told the BMJ that ROCKET-AF “should be considered of uncertain validity until a more-thorough review can be done,” and that there should be “an investigation by an independent group of experts to quickly determine if there are grounds for retraction.”

In the middle ground is Dr. Sanjay Kaul, who served on the FDA advisory panel that evaluated rivaroxaban for its atrial fibrillation indication in 2010. He was reassured by the investigators’ analysis, while still calling for an independent review. “It was claimed [by the BMJ] that underestimation of INR by the recalled device could have resulted in erroneously increasing warfarin dose and associated bleeding, thereby making rivaroxaban appear safer relative to warfarin, he said. But in the analysis, “although bleeding was increased nearly twofold in both treatment groups in the subset with recalled conditions, the HR of 1.18 for rivaroxaban contradicts this claim.”

Because a legitimate question regarding reliability of INR measurements has been raised, it would seem prudent to reassess safety and efficacy data, said Dr. Kaul, professor of medicine at the University of California, Los Angeles, in an interview. “Ideally, this should be done by an independent party as was done for the RECORD trial of rosiglitazone by Duke investigators who were not involved in the clinical trial. Personally, I am not sure if the benefit-risk balance will be materially altered. Potential overestimation of safety of rivaroxaban related to spuriously low INR with warfarin will likely be counterbalanced by underestimation of efficacy.” If there are data from INR assessments performed centrally in a core lab, independent investigators could use those to reassess efficacy and safety, he continued. “At least the investigators should use the core lab INR to verify their assumptions in the research letter published in NEJM that conditions of recall were truly associated with low INR values. Otherwise, their results are open to question.”

Of note, Dr. Robert M. Califf , who was a ROCKET-AF study cochair and is now FDA Deputy Commissioner for Medical Products and Tobacco, was not involved in the analysis. Dr. Califf is awaiting confirmation from Congress on his nomination to head the FDA.

Dr. Kaul has stock interest in Johnson & Johnson, which sponsored ROCKET-AF, and serves as a consultant to Boehringer Ingelheim and Bristol-Myers Squibb.