PFO-Stroke Link Takes Another Hit

Download this article's Factoid (PDF & PPT for Gold Subscribers)

Over one-third of recurrent strokes or transient ischemic attacks (TIAs) in patients with patent foramen ovale (PFO) are more likely to arise from concurrent conditions like atrial fibrillation or large artery disease, casting doubt on the causal role of the defect, according to results appearing online August 4, 2011, ahead of print in Stroke.

Researchers led by Heinrich P. Mattle, MD, of University Hospital at the University of Bern (Bern, Switzerland), looked at 308 patients with cryptogenic stroke and PFO treated at their institution from January 1994 to August 2000. Roughly half of the patients (n = 158) received medical therapy (oral anticoagulants or antiplatelets) and half received percutaneous closure (n = 150). Dr. Mattle and colleagues followed both groups for a mean of 8.7 years to see if PFO closure had an effect on the expected rate of recurrent events and to judge whether the rate of concurrent etiologies might have caused these events.

Baseline characteristics were similar between the medical and closure groups, with the exception of previous cerebrovascular events and large right-to-left shunt, which were more common in the closure group (P < 0.03 and P < 0.001, respectively).

Many Recurrent Events Not Cryptogenic

Over a mean follow-up of 8.1 years, 32 recurrent cerebrovascular events (13 strokes, 19 TIAs) occurred in the medical group, amounting to a rate of 20.3%. Sixteen of these patients died. The frequency of recurrent stroke or TIA did not differ with respect to antiplatelet or anticoagulant treatment. In the PFO closure group, there were 16 recurrent events (8 strokes, 8 TIAs), which worked out to an event rate of 10.7% over a mean follow-up of 9.2 years, with 7 deaths.

In the PFO closure group, more than half of recurrent events were judged to be cryptogenic, while this was true for almost two-thirds of recurrent cerebrovascular events in the medical treatment group. Of the remaining events, which were deemed more likely associated with concurrent etiologies, most were related to atrial fibrillation or small or large artery disease (table 1).

Table 1. Etiologies Associated with Recurrent Cerebrovascular Events


PFO Closure
(n = 150)

(n = 158)




Atrial Fibrillation



Large Artery Disease



Small Artery Disease



Antiphospholipid Antibody Disease


Cerebral Vasculitis


Thrombophilic Disorder


Patients with recurrent cerebrovascular events showed similar levels of concurrent etiology between the medical treatment group and the PFO closure group (38% vs. 44%; P = 0.68). But the rate of concurrent etiologies in the medical therapy group may, in fact, have been underestimated since half of the patients with a recurrent event did not have a complete etiologic work-up.

Among patients with recurrent stroke or TIA, concurrent etiologies were more frequent in those older than age 55 at the time of the index event than in younger patients (P = 0.018). Other factors, such as atrial septal aneurysm and large right-to-left shunt showed no interaction with concurrent conditions. In the PFO closure group, patients with recurrent events more frequently had hypercholesterolemia if they also had concurrent etiologies than if they did not (100% vs. 56%; P = 0.042).

“Concurrent etiologies are identified for a considerable proportion of recurrent ischemic events [, casting] doubt on the sole causal role of PFO in the case of stroke recurrence, and indicat[ing] that secondary prevention in patients with cryptogenic stroke and PFO should not be focused on PFO closure alone,” the authors conclude. In particular, they note, “given that concurrent stroke etiologies are not prevented by percutaneous device closure of the PFO, we would not discontinue antithrombotic treatment in patients who undergo PFO closure.”

CLOSURE Confirmation

In a telephone interview with TCTMD, Anthony J. Furlan, MD, of the Case Western Reserve University School of Medicine (Cleveland, OH), agreed with the findings, but went even further. “I think it’s often not clear that even the initial stroke is due to the PFO,” he said.

In general, Dr. Furlan said the results of the current study reinforce those of the CLOSURE I trial, the only randomized, controlled study to have looked at the issue of PFO closure and cryptogenic stroke. Presented in November 2010 at the annual American Heart Association Scientific Sessions, the results provided strong evidence that percutaneous device closure lacks superiority over medical therapy alone in preventing recurrent stroke and mortality in patients with PFO.

“CLOSURE I followed patients for only 2 years, but even with that, we found an alternative reason for recurrent events in 80% of patients,” said Dr. Furlan, the trial’s principal investigator. “It’s actually reassuring that we found in a randomized trial what people are finding in their individual experiences.”

According to Dr. Furlan, there are multiple problems inherent in determining the cause of cryptogenic stroke. “The core problems are cryptogenic stroke, which is a mixed group of etiologies, vs. how do you prove something is truly due to paradoxical embolism?” he said. “And then it’s another matter to say closing the hole is any better than antithrombotic therapy. Or it’s also possible, as the authors of this paper point out, that maybe they’ve got 2 problems. Maybe they do have paradoxical embolism, but 6 years later, things change and they’ve got something different.”

Performing the Right Tests

A key for clinicians is to perform the right tests up front, which seldom happens, Dr. Furlan noted. “For example,” he said, “one of the most grossly underdiagnosed conditions in this cryptogenic stroke population is probably occult atrial fibrillation. If you only do a 24-hour Holter monitor, you’re not going to find it, but nobody does a routine 30-day event monitor. Probably the causes of cryptogenic stroke are not that many, it just depends on how hard you look.”

None of this is to say that PFOs should never be closed for cryptogenic stroke, Dr. Furlan stressed. “In CLOSURE I, [we did not conclude] that paradoxical embolism wasn’t real, or that there are no patients that should undergo endovascular closure,” he said. “All we had was anecdotal criteria and no hard data. Now, my impression is that everyone seems to agree that we were probably closing too many of these.”

In the future, Dr. Furlan expressed hope that results of CLOSURE I will be combined with those of other ongoing trials investigating PFO closure and cryptogenic stroke. “If we start putting all these trials together, we’ll be able to refine our selection criteria, and not close all these holes, willy-nilly” he said. “That’s what I’m starting to see. I think clinicians in general are being more conservative.”


Mono M-L, Geister L, Galimanis A, et al. Patent foramen ovale may be causal for the first stroke but unrelated to subsequent ischemic events. Stroke. 2011;Epub ahead of print.



Related Stories:

Jason R. Kahn, the former News Editor of TCTMD, worked at CRF for 11 years until his death in 2014…

Read Full Bio
  • The study was partially funded by a grant from the Swiss National Research Foundation.
  • Dr. Mattle reports receiving research grants, speaker’s fees, and consulting honoraria from AGA Medical.
  • Dr. Furlan reports no relevant conflicts of interest.