New PFO Guidelines: No Routine Offer of Closure for Cryptogenic Stroke, Says American Academy of Neurology

Taking a strong stance, new recommendations released today by the American Academy of Neurology (AAN) advise against the routine recommendation of patent foramen ovale (PFO) closure for cryptogenic stroke, exhorting physicians to educate patients as to just how common PFOs are and how rarely they contribute to a subsequent stroke. 

The practice advisory comes as the US Food and Drug Administration (FDA) is poised to make a decision on the Amplatzer PFO closure device (St. Jude Medical) following a split vote by its advisory panel in May. As previously reported by TCTMD, the agency’s Circulatory System Devices Panel voted 15 “yes” to 1 “no” on a question of device safety, 9 to 7 on the question of effectiveness, and 11 to 5 on the question of whether the device’s potential benefits outweigh the risks.

The position of the AAN document, led by Steven R. Messé, MD (Perelman School of Medicine at the University of Pennsylvania, Philadelphia), is far less equivocal. “Clinicians should not routinely offer percutaneous PFO closure to patients with cryptogenic ischemic stroke outside of a research setting,” the advisory reads. “In rare circumstances, such as recurrent strokes despite adequate medical therapy with no other mechanism identified, clinicians may offer the Amplatzer PFO occluder if it is available.” This last recommendation, however, is given a strength of evidence grade of C.

New Evidence, No Big Change in Guidance

The practice advisory is an update to earlier recommendations released in 2004, which concluded that the optimal therapy for secondary stroke prevention in patients with PFO was unknown. For their update, an AAN committee reviewed new published studies with the specific aim of answering two questions relating to the role of percutaneous PFO closure versus medical therapy alone and the role of anticoagulation versus antiplatelet therapy in preventing recurrent stroke.

After an extensive review of the literature, Messé et al considered the CLOSURE I trial, the PC Trial, and the RESPECT trial to answer the percutaneous closure question as well as a recently published meta-analysis combining the three. All three trials, they note, failed to meet their primary endpoints, although the meta-analysis (as well as their own analysis, which looked only at the two Amplatzer trials) favored closure. Here, however, the number of patients lost to follow-up or crossing over between arms in the trials introduced a risk of bias that offset the overall findings.

“If the meta-analysis of these trials or the individual trials themselves had shown a huge benefit to closure, the guideline would have come to the conclusion that it’s effective,” one of the co-authors on the practice advisory, Gary S. Gronseth, MD (University of Kansas, Kansas City, KS), told TCTMD. “But unfortunately, the individual studies are negative and when you meta-analyze them, there’s a small difference that is statistically significant. But you compare that to the risk of bias in the studies, and a lot of uncertainly remains.”

Bad Timing?

Commenting on the AAN guidance to TCTMD, John Carroll, MD (University of Colorado School of Medicine, Aurora, CO), pointed to the fact that while the FDA panel had access to the extended, complete RESPECT data set, the AAN document did not. “The only thing that was in the public domain was the abstract that was presented at TCT,” he said, “and they didn't even comment on that within the advisory, because that would be inappropriate based on an abstract presentation.”

Still, said Carroll, why rush out this document now, while an FDA decision is still pending and with the knowledge that the extended RESPECT analysis has been done?

“That’s why I say that it's strange that they’ve published this now, rather than wait for a period of time when the practice advisory could be inherently more useful and aligned with the more recent data, and certainly aligned if there is FDA approval. . . . They’ve already waited 12 years [to update] this advisory. In my opinion, it would be prudent to have waited 6 more months.”

Asked about the timing, Gronseth told TCTMD it just came down to where it fell within the review cycle across the range of stroke treatments for which the AAN issues guidance. “There was no conscious effort to do it now rather than last year,” he said. “It’s been in the works literally for several years.”

But Gronseth, who was himself an investigator on RESPECT, points out that additional follow-up from that trial doesn’t change the fact that there are two sources of bias in RESPECT, the cross-overs/loss to follow-up, and how events were adjudicated. “It’s not just getting to statistical significance for the differences,” he said, “but is the difference large enough to overcome the risk of bias in the studies?”

Expert Input

Carroll also pointed to the make-up of the writing committee, which was comprised of five neurologists and two cardiologists, both of whom list cardiac imaging as their area of specialization. By contrast, on the FDA advisory panel all of the interventional cardiologists—but none of the neurologists—voted yes on the question of device whether they believed the device was effective at preventing stroke.

“They didn’t have an interventional cardiologist in the AAN writing group, and I find that odd, to have a proclamation on the value of a procedure without anyone on the group having performed the procedure,” Carroll said.

Gronseth disagreed, saying excluding interventional cardiologists was “not a conscious effort.”

Moreover, he continued, “the counter argument to that is [the make-up of the writing group] doesn’t change the evidence. The evidence is what the evidence is.” He added that interventional cardiologist could likely add some valuable insights, particularly with regard to procedure complications, “but that doesn’t ultimately change whether the evidence of whether something is effective or not or, as in this case, if it’s ambiguous.”

Sunichi Homma, MD (Columbia University Medical Center, New York, NY), one of the two cardiologists on the AAN document, has spent decades studying the link between PFO and stroke. It was Homma who spoke during the open public comment period during the FDA advisory panel hearing for the Amplatzer.

Speaking with TCTMD this week, Homma reiterated that he hoped the FDA would take into account the AAN position before delivering its decision on Amplatzer. “I personally think, as many others do, that just thinking that a patient who happens to have a stroke [just] because the PFO is there, without thinking of other potential reasons mechanisms, is too simplistic a way to think,” he said. “I do think there are some patients who do have stroke because of their PFO, [but] I don't think we've really done enough to identify those very clearly such that placing a device in all the patients who happen to have a PFO and cryptogenic stroke . . . is in the public interest.”

He points out that others who spoke during the public comment period urged the FDA to consider closing PFOs in everyone found to have one, something Homma characterized as “upsetting.” Approximately 75 million people have a PFO in the United States alone, he points out. “It’s like having an ear crease.”

Indeed, the AAN statement takes pains to remind physicians to explain to their patients just how common an anomaly a PFO truly is, and that patients’ understanding of risk and ambiguity need to be managed appropriately. “Informing patients about the commonness of PFO within the general population and the difficulty in determining whether their PFOs caused their symptoms will assist in selecting an appropriate decision reference frame,” the statement reads.

Indeed, Homma made the point that despite years of interest in PFO, interventional cardiologists are not the best choice of physician when it comes to explaining options for stroke care and prevention. “Stroke neurologists who specialize in this are most knowledgeable,” he said. “We stress in the article that appropriate work up and screening needs to be done by physicians who actually can recommend the potential choices for therapy, rather than by a physician whose choice of therapy that can be offered is very, very limited and knowledge of the tests for stroke is also very limited.”

Cardiologists treating a patient for acute MI, he pointed out, “don’t make a judgement on how subarachnoid hemorrhage should be treated.”

The FDA’s Next Move

Carroll, for his part, predicts that the FDA will make a decision on the Amplatzer occluder before the end of October. He stressed, however, that while he is a member of the RESPECT steering committee, he has had no direct communication with the FDA. Yet “there don’t seem to be any major glitches that have been identified,” he added.

The AAN document mentions that three clinical trials of PFO closure for cryptogenic stroke are still ongoing, including the REDUCE trial with the Gore device.

As to the second question—that of anticoagulants versus antiplatelets in this population—the writing group concluded that aspirin or other antiplatelets are likely the better choice than warfarin or heparin, unless a patient is already taking an anticoagulant for another indication or in rare circumstances when the stroke occurs in a patient already taking an antiplatelet. Whether one of the new oral anticoagulants will be useful in this setting needs further study.

Asked if he worries the guideline document will be out of date—or out of step with the FDA—should the Amplatzer be approved, Gronseth pointed out that the writing group chose their words carefully.

“To a certain extent we immunized the guidelines, because we say ‘if available,’ and there are exceptions if the device is available,” Gronseth said. Currently, he noted, “most people in the country who are getting PFOs closed are getting devices designed to close atrial septal defects, not PFOs. So the recommendation itself is that it should only be done in unusual circumstances . . .  but only if available, so if it becomes available then the unusual circumstances would apply.” 


  • Messe SR, Gronseth G, Kent DM, et al. Practice advisory : Recurrent stroke with patent foramen ovale (update of practice guidelines). Neurology. 2016;87:1-7. 


  • Messé reports funding from the NIH, consulting for GlaxoSmithKline, and research funding from GlaxoSmithKline and WL Gore & Associates as an investigator for the Gore REDUCE trial, and as a subinvestigator for the CLOSURE-I trial.  
  • Carroll reports receiving personal and institutional fees for serving on the RESPECT steering committee and has consulted for St. Jude Medical. 
  • Gronseth reports being an investigator for the RESPECT trial, serving as an associate editor for Neurology and as an editorial advisory board member of Neurology Now, and receiving compensation from the AAN. 
  • Homma reports receiving financial compensation for serving on scientific advisory boards for Boehringer Ingelheim, Bristol-Meyers Squibb, Daiicho Sankyo, and Pfizer; has served on the data safety monitoring board (DSMB) for the RESPECT trial and on the editorial board for Circulation; and has received honoraria from Lankenau Medical Center and research support from the NIH. 

Related Stories:


Shelley Wood is the Editor-in-Chief of TCTMD and the Editorial Director at CRF. She did her undergraduate degree at McGill…

Read Full Bio