In July 2016, the hand-held, point-of-care INRatio device (Alere, Inc.) was recalled because of the potential to generate inaccurate results. This device was used to monitor warfarin therapy in the control group of the pivotal ROCKET-AF trial, which provided the primary data to support the 2011 approval of the blood-thinner rivaroxaban (Xarelto, Janssen). Because of the concern about the INRatio device, the FDA’s Center for Drug Evaluation and Research has completed a series of analyses to assess the affect that this faulty monitoring device may have had on the ROCKET-AF study results.
The agency has determined that effects of the device on strokes or bleeding, including intracranial bleeding, were minimal. The FDA concluded that rivaroxaban is a safe and effective alternative to warfarin in patients with nonvalvular atrial fibrillation.
The ROCKET-AF trial was a randomized, double-blind, event-driven, confirmatory study of rivaroxaban (n = 7,081) versus dose-adjusted warfarin (n = 7,090) in patients with nonvalvular atrial fibrillation. The study’s primary objective was to demonstrate that the efficacy of rivaroxaban for the intended indication––a reduction in the rate of stroke and systemic embolism––was noninferior to that of dose-adjusted warfarin. The proposed primary efficacy endpoint was the time to the first occurrence of stroke (any type) or non-CNS systemic embolism. The dosage of rivaroxaban was 20 mg once daily for most patients. The dosage was adjusted to 15 mg daily for patients with creatinine clearance (CrCL) of 30 mL/min to 50 mL/min; those with CrCL of less than 30 mL/min were excluded.
All study sites were instructed to use only the INRatio device to adjust the warfarin dose. Approximately 87% of patients in the warfarin arm had at least one same-day pair of international normalized ratio (INR) assays read with the INRatio device at the study site (point-of-care) as well as with a laboratory-based device at Duke University near the end of the trial. Most of these patients had two such pairs of INR readings.
The study’s primary endpoint was achieved in 3.8% of the rivaroxaban group and in 4.3% of the warfarin group (hazard ratio [HR], 0.88). The rates of all-cause mortality were 8.8% and 9.4% in the two groups, respectively. Major bleeding occurred in 3.6% of rivaroxaban-treated patients and in 3.5% of warfarin-treated patients (HR, 1.04).
After the INRatio device was recalled, the FDA modeled the affect of the device on clinical outcomes in the ROCKET-AF trial. The agency estimated that if a more-accurate device had been used, the rate of major bleeding in the warfarin arm might have been 7% to 10% less than observed, with correspondingly small changes in the rivaroxaban-versus-warfarin HR that would disfavor rivaroxaban. In addition, it was estimated that the HRs for life-threatening or fatal bleeding and hemorrhagic stroke, which both favored rivaroxaban numerically, would be minimally increased, but both would still favor rivaroxaban. In addition, the HR for ischemic stroke would be moderately changed in favor of rivaroxaban.