AngelMed Cardiac Monitor Panned At FDA Panel Over Clinical Trial Shortcomings

Trial shortfalls including the early termination of patient enrollment and failure to meet an efficacy endpoint sank Angel Medical Systems Inc.'s implantable cardiac monitor at an FDA Circulatory System Devices advisory panel meeting.The 12-member panel unanimously voted that the evidence presented by the company for its AngelMed Guardian System did not show the device is effective or that its benefits outweigh its risks during a March 15 meeting. Meanwhile, an eight-member majority voted that the sponsor had not proven device safety.The Guardian system is a first-of-its-kind implanted device about the size of a pacemaker that performs an electrocardiogram every 90 seconds and sends an alarm to a patient-carried unit if the results suggest an imminent heart attack. Angel Medical is seeking an indication for use in either heart attack survivors or those who are scheduled for coronary artery bypass surgery with diabetes, renal failure or a high risk for future cardiac events. The device has been on the market in Europe since 2010.This company's trial randomized 907 high-risk subjects who had experienced a heart attack or acute coronary syndrome event. Patients in the study were all implanted with the AngelMed Guardian system, and then assigned to have the alerting feature of the device either turned on or off for a six-month period to assess whether the alerts reduced the rate of death, new heart attacks or thrombotic coronary occlusions.Angel Medical said the system could speed treatment and ensure peace of mind for patients who may not experience traditional heart attack symptoms, which is more of the case for women.“At the end of the treatment period, patients consistently vote with their feet and opt for re-implantation,” said trial investigator Mitchell Krucoff, Duke University Medical Center. Around 93% of patients in the trial opted to have the Guardian’s battery replaced at the end of the six-month trial, which allowed them to continue using it.Many Questions UnansweredBut while panelists agreed the device offered exciting potential, they said the trial raised too many questions for them to recommend PMA approval.First, the trial did not meet its primary efficacy endpoint of reducing death, heart attacks or thrombotic coronary occlusion. Specifically, there was only one cardiac or unexplained death during the trial period in the control group, while there were three such deaths in the treatment group. Fourteen members of the control group experienced a new heart attack, while 10 members of the treatment group did. The number of patients seeking treatment within two hours of an event was higher for the treatment group than the control group, but FDA and panelists said that data was muddied because Angel Medical changed the maximum window of time measured between a Guardian-detected event and confirmation of a thrombotic event at a medical facility from seven to 90 days after the data was unblinded.Another concerning factor was that the sponsor opted to stop enrolling patients after registering around 1,000 individuals, despite FDA approving a protocol that would include 1,200 patients. Krucoff says Angel Medical stopped enrollment because the analysis did not need more patients, so recruiting more people would have raised ethical issues. However, FDA views the change as a significant protocol violation.“FDA did not ask the manufacturer to stop enrollment,” said Zhiheng Xu, a biostatistician in FDA’s device center. “The manufacturer did it on its own.”The trial report showed numerous additional data-handling issues, such as excluding certain times patients’ alarms sounded without much explanation, that panelists and FDA said made the results very hard to interpret.“We can’t know true negatives or false negatives, and without false negatives, we can’t really know false positives,” said Kimberly Selzman, a reviewer in the FDA’s division of cardiovascular devices. “If a patient had a cardiac event and the device didn’t detect it, it would probably be missed unless the patient presented with symptoms,” she noted.“We are not statistical purists, and we’ve learned not to obsess about individual margins when we’re looking at data in the context of possible clinical benefits,” said panel chair Richard Page, University of Wisconsin School of Public Health. But, Page said, the Guardian had not shown enough of a clinical benefit to outweigh the statistical problems.In addition, Page expressed concerns that if the Guardian failed, patients could end up in a worse place than they would be without it. “I hear it gives patients peace of mind, and I worry about that peace of mind,” he said, speculating patients with Guardian implants might ignore cardiac signals if the device didn’t sound an alert. He also noted that the number of events identified in the trial was “small and inconclusive.”The panelists were also troubled by the adverse events, such as infections, that occurred during the trial. There were 364 adverse events during the study in total, 31 of which were specifically related to the system. These included 11 infections, four patients with pain at the implant site, four with lead migration, and other patients with erosion, loss of sensation, or general unspecified complications.These rates are similar to those for pacemakers, Angel Medical said, but panelists highlighted a key difference. “If I really knew how effective this was, I’d have little trouble recommending approval … but I’m very concerned we might have to implant this in a hundred people to detect one event,” said JoAnne Lindenfeld, Vanderbilt Health and Vascular Institute.“The study had a lot of imperfections,” Krucoff conceded. He noted that he thought the worst among them was the “too short” six-month follow-up time. But, he said, the device could provide a crucial window into patient cardiac events that currently go undetected.Michael Carome, medical director at consumer watchdog group Public Citizen, also spoke in strong opposition to approval of the device. Carome said the trial results didn’t establish safety or efficacy, and also didn’t show any clinically significant reduction in deaths.