Medical Device Daily Washington Editor
GAITHERSBURG, Maryland — NMT Medical (Boston) and AGA Medical (Plymouth, Minnesota), two firms in the business of making closure devices for patent foramen ovales (PFOs), or holes in the heart, have long sought to persuade FDA to allow them to market their products without undergoing randomized, controlled clinical trials (RCTs).
The agency once again put the question to its circulatory systems advisory committee Friday, but this third attempt to nudge the panel into giving FDA a green light to allow non-RCT trials for this indication failed — largely because a link between PFOs and stroke is not definitive.
PFOs occur naturally during pre-natal life, but they may persist into adulthood in as many as three in 10 cases. This hole between the left and right atria can allow unfiltered blood back into circulation, raising the risk for strokes.
Donna Buckley, a reviewer at the Center for Devices and Radiological Health, noted that the agency has not yet approved a device to repair a PFO, although the two companies were previously granted the humanitarian device exemption (HDE) protocol. The HDE is used only when the annual rate of diagnosis is less than 4,000 per year, and FDA informed the companies last autumn that the total diagnoses exceeded this number.
Both NMT and AGA then withdrew their exemptions before the agency formally pulled the HDEs.
Buckley indicated that “single-arm studies are ongoing . . . to allow continued patient access” to the therapies.
AGA’s current PFO closure device, the Amplatazer, is currently in a Phase III trial that commenced last October, and in that study, the firm proposes a link between PFOs and migraine headaches. NMT is enrolling patients in a trial for its Starflex closure device under the hypothesis that PFO closure can prevent stroke.
The circulatory systems advisory panel addressed this question in 1997, when the panel concluded that RCTs were necessary for this application and that assessment should include freedom from recurrent stroke.
In a September 2002 meeting, the panel came to the same conclusion about RCTs, but amended the recommended endpoints to stroke and death at two years.
NMT and AGA are not the first to complain about the difficulty of enrolling a Web-educated public in a clinical study control arm.
In a Feb. 22 meeting of the orthopedic and rehabilitative devices advisory committee, the Corin Group (Cirincester, UK) expressed the same frustration regarding its trial for the Cormet 2000 hip resurfacing system (Medical Device Daily, Feb. 26).
That panel recommended that FDA give the Cormet the nod, with panelists acknowledging this difficult “reality” of modern clinical trial conduct.
Of the ongoing efforts to roll out Phase III trials with controls, Buckley remarked that “enrollment in those studies has been very slow,” in part because patients want to close the PFO.
Buckley said that another factor is that the use of warfarin is “incompatible with the active lifestyle of younger patients.” Other factors include physician bias, which may be the predilection of interventionists to intervene, and the use of other devices on an off-label basis.
The devices most commonly used for PFO include those to treat atrial and ventricular septal defects.
Billy Dunn, a stroke neurologist at the Center for Drug Evaluation and Review, referred to the uncertainty of the connection between PFO and stroke in noting that between 35% and 40% of ischemic strokes in adults younger than 55 are of unknown origin, or cryptogenic. He noted that the “prevalence of PFO is higher in individuals with cryptogenic stroke than in individuals with a stroke of a known cause,” but medical science has yet to discern whether “other co-variates may account” for this artifact.
Regarding the hypothesized connection between PFO and stroke, Dunn remarked that the relationship “is uncertain and, as you can see, the data are somewhat conflicted.”
During the public hearing, Steven Kittner, MD, speaking on behalf of the American Academy of Neurology (St. Paul, Minnesota), disclosed that he is a site principal investigator for one of the trials, and he opposed the petition.
He argued that off-label use to treat PFO “is really the big gorilla in the room,” and that while FDA has no way to control off-label use, the Centers for Medicare & Medicaid Services’ reimbursement mechanism “would be a way to move these trials forward,” presumably a reference to the coverage with evidence development mechanism now in use at CMS.
Larry Latson, MD, director of the pediatric and congenital cardiac catheterization lab at the Cleveland Clinic (Cleveland), explained that “after strokes, patients are terribly frightened” and “basic personality traits dictate that they don’t want to give up control.”
The typical patient does not want invasive procedure unless the doctor can guarantee the outcome and only if there is no alternative, but most patients also prefer to avoid a lifetime drug regime. Attitudes toward medications “often do change over time with poor experiences,” and patients see closing the hole as a desirable alternative “if the procedure is not too bad,” according to Latson.
The fact that the PFO closure procedure can be done as an outpatient procedure lowers the resistance.
The panel discussed the possibility of the sponsors pooling their controls, but the counter-arguments included the contorted mathematical adjustments needed to make sense of the data, due ot the different approaches to closing the PFO, and the inherent differences in efficacy and safety endpoints.
Panelist Richard Ringel, MD, an associate professor of cardiology at Johns Hopkins University (Baltimore), said that off-label device use is necessary because in the world of pediatric cardiology, “we use devices that have never been in a randomized controlled trial and never will be.”
He was the sole member of the panel to advocate the non-RCT approach for this indication, but after the panel deliberated the questionable link between PFO and stroke, he joined the others in a unanimous vote to recommend that FDA maintain its requirement for RCTs.