Repairing a common heart anomaly — a patent foramen ovale (PFO) — does not reduce the risk of a stroke or transient ischemic attack (TIA) and could in fact lead to serious complications, according to a large, multicenter randomized study. The findings, published in the March 15 New England Journal of Medicine, is the death knell for the company sponsoring the study, NMT Medical, which closed its doors following the results of the study. But it may not stop the procedure from being done as there are many similar devices on the market, which are used off-label and are being studied to repair PFO in patients with cryptogenic, or unexplained, stroke.
PATENT FORAMEN OVALES are documented in up to 56 percent of patients with cryptogenic stroke.
PFOs are common in general with one in four people showing evidence of these small holes at autopsy. The reason that closure with a percutaneous device is recommended is that PFOs are documented in as many as 56 percent of patients with cryptogenic stroke. In addition, it is theoretically possible for an embolism from the venous system to work its way from the right to left atrium and into the systemic circulation through a PFO (paradoxical embolism).
The team of stroke neurologists and cardiologists designed the current trial — CLOSURE 1 (Evaluation of the STARFlex Septal Closure System in Patients with a Stroke and/or Transient Ischemic Attack due to Presumed Paradoxical Embolism through a Patient Foramen Ovale) — to test the use of the device to close the PFO versus standard medical therapy to prevent a subsequent stroke on the heels of the first one.
PFOs do not cause cardiac problems and generally come to light following a stroke or TIA. The hypothesis going into the study was that closing the hole would reduce the risk of a second stroke or TIA by two-thirds.
DR. ANTHONY J. FURLAN: “Closing these holes does not appear to be any more effective at preventing stroke or TIA than traditional preventive medical therapies. Too many of these holes are being closed.”
Anthony J. Furlan, MD, the Gilbert Humphrey Professor of Neurology, chairman of the department of neurology and co-director of the Neurological Institute at University Hospitals Case Medical Center in Cleveland, and colleagues from 87 centers in North America enrolled 909 patients in the open-label randomized trial, with half getting closure of the PFO and the others getting medical treatment to reduce the risk for a second stroke. The primary outcome for the CLOSURE 1 trial was a stroke or TIA up to two years after the initial stroke, death from any cause within the first 30 days of the start of the intervention, and death from neurological causes between day 31 until the end of the two year follow-up.
The patients, between the ages of 18 and 60, were identified for the study if they had a diagnosis of cryptogenic stroke and a PFO. They were randomized to get preventive treatment with the STARflex septal closure system plus antiplatelet therapy or medical therapy alone (aspirin or warfarin).
Dr. Furlan noted that the original design of the study included 1,600 patients but the investigators were not able to meet that goal. They had trouble with recruitment for the study as many interventional cardiologists are closing holes off-label after a stroke or TIA, he said.
At the end of two years, 2.9 percent of patients who had closure of the hole and 3.1 percent of patients with medical therapies suffered a stroke (p=0.79). There were similarly low numbers for TIA, with 3.1 percent for closure patients and 4.1 percent for those who received medical therapy. There were no deaths in the primary outcome.
Thirteen of the 447 patients (3.2 percent) who had the closure device suffered major vascular complications. Atrial fibrillation was significantly more frequent in the closure group with 23 patients (5.7 percent) versus three out of 462 patients (0.7 percent) in the medical arm of the study. Atrial fibrillation occurred within 30 days of the procedure in 14 of the 23 patients (61 percent) and the problem persisted in six patients.
“Closing these holes does not appear to be any more effective at preventing stroke or TIA than traditional preventive medical therapies,” said Dr. Furlan. “Too many of these holes are being closed.”
Dr. Furlan said he worries that the financial incentive — the procedure costs $10,000 and insurance companies reimburse for it — is driving the use of the device for stroke prevention. The FDA has not approved any of these devices for this indication, he noted.
Previously, the FDA approved PFO closure under a Humanitarian Device Exemption (HDE), which is applied to devices that would be used in fewer than 4,000 patients a year. However, Dr. Furlan said that no one was keeping count of the number of closures done off label.
Midway through the study the FDA actually withdrew the HDE exemption with hopes that it would help with recruitment but Dr. Furlan said it did not because of off-label closures. “CLOSURE 1 has been criticized as underpowered but we saw absolutely no difference in the recurrent stroke rate which indicates that any benefit of device is likely to be very small and require more than 4,000 patients to demonstrate in a randomized trial,” said Dr. Furlan.
There are at least half a dozen companies with different types of devices that close holes, and some of these trials are still under way or are in the process of analyzing results. Other trials have closed because of slow patient recruitment. Neurologists are running all of these studies. “CLOSURE 1 may put the brakes on closing holes but it is not the final answer,” Dr. Furlan said. “We need to wait for the other trial results for comparison.”
The Boston-based makers of the STARflex device, tested in the current trial, filed for bankruptcy last year. The company had no hand in the design or the actual testing and analysis in the study.
Dr. Furlan said that some patients may derive a stroke prevention benefit from closure but “we do not know who the best candidates would be. The problem is there are many causes of cryptogenic stroke and only a small subset have true paradoxical embolism even if a PFO is present. Even in refined subgroups, the challenge will be to demonstrate that device closure is superior to medical therapy alone.”
Marc Fisher, MD, professor of neurology at the University of Massachusetts Medical School and editor of Stroke, was watching the results of the study as an independent member of an adjudication committee. “We have had no data on the use of these devices,” said Dr. Fisher. “This study shows us that there is a low risk for stroke with a PFO. A lot of people are being closed just because doctors think it needs to be done. But now we have evidence that doing these device closures is not indicated in most patients.
“We also have to be concerned with the development of atrial fibrillation and major vascular complications with this device,” he added.
Dr. Fisher said that devices are approved if it can be shown that it is equivalent to what is on the market. “The big issue in the stroke field is that we need evidence-based data showing that the devices are effective,” said Dr. Fisher.
In an editorial in the same issue of the journal, S. Claiborne Johnston, MD, PhD, professor of neurology and epidemiology and associate vice chancellor of research and director of the University of California, San Francisco Clinical and Translational Science Institute, said that the American Heart Association and the American College of Cardiology “have discouraged use of these devices to close a patent foramen ovale as a means of preventing recurrent stroke, yet this practice has remained fairly constant over the past several years.”
He added that the trial “provides the best evidence available regarding the role of closure in stroke prevention and should not be ignored.”
DR. MARC FISHER: “This study shows us that there is a low risk for stroke with a PFO. A lot of people are being closed just because doctors think it needs to be done. But now we have evidence that doing these device closures is not indicated in most patients.”
Dr. Johnston also discussed the estimated numbers of closures — 80,000 patients — in the same nine-year period it took to recruit patients for the CLOSURE 1 study. He said that would amount to $400 million “spent on a procedure that had no apparent benefit, to say nothing of the potential clinical risks involved.”
He agrees that the studies under way should continue to see whether the findings are replicated with different devices. “…it is reasonable to complete these other trials given the remaining uncertainty.”
He and others, including Dr. Furlan, agree closures should be limited to these studies and should not be reimbursed outside of a clinical trial. He believes that limiting reimbursement to studies will also help with recruitment. “Trials need to be completed more quickly so that better decisions can be made sooner,” he said, adding “the costs to society of unproven interventions are also extremely high.”
Listen here as Dr. Anthony Furlan, the Gilbert Humphrey Professor of Neurology, chairman of the department of neurology and co-director of the Neurological Institute at University Hospitals Case Medical Center in Cleveland, discusses the findings of the CLOSURE 1 trial and why he think using devices to close the patent foramen ovale should be restricted to participants in clinical trials: http://bit.ly/dy2KLx