The Cardiovascular and Renal Drugs Advisory Committee voted 11 to 5 not to recommend the new drug application of King Pharmaceuticals' CorVue (binodenoson) 250 mg for approval by the FDA.
In December 2008, King submitted a new drug application to the FDA for the cardiac pharmacologic stress SPECT imaging agent intended for use in patients with or at risk for coronary artery disease who are unable to undergo a cardiac exercise stress test.
In three double-blind, phase III trials assessing the efficacy of the drug, researchers reported that the studies failed to meet their original primary endpoints. Patients in VISION 301 (357) and a second study VISION 302 (419) were randomly assigned to receive either adenosine or binodenoson with myocardial perfusion imaging (MPI). Patients in VISION 305 (433) were randomly assigned to receive either adenosine or binodenoson for MPI in various combinations and sequences.
Initially, the researchers touted success upon meeting its endpoint at the 2008 American College of Cardiology (ACC) meeting in Chicago. The efficacy analyses were based on comparison of the extent and severity of ischemia, expressed as summed difference scores (SDS). The mean paired SDS difference of the CorVue and adenosine (Adenocard from Astellas Pharma) images was -0.09 in VISION 302 and -0.68 in VISION 305, and the 95 percent confidence intervals were well within the pre-specified 1.5 SDS units; only 3 percent of the images were determined to be sufficiently discordant. In addition, there were no reported cases of atrioventricular block.
However, the FDA panelists expressed concern over changed endpoints. Researchers originally defined success in the study as the concordance of results from MPI tests as measured by the lower limit of the 95 percent confidence interval for the kappa statistic exceeding 0.61.
According to filed data, the limits of the kappa statistic were lowered to 0.15 for study 301, 0.27 for study 302 and 0.36 for study 305. The modified trial criteria required a 95 percent confidence interval of the mean difference scores fall between the interval of -1.5 and 1.5, and the upper limit of the 95 percent confidence interval in patients with extreme discordance fall less than 10 percent. Despite the panel's concerns, the researchers reported that all three studies met the requirements for approval.