May 6th, 2012

Patent Foramen Ovale — Are the Choices Patently Obvious?


CardioExchange editor John Ryan interviews Shelby Kutty, of the University of Nebraska College of Medicine, about his recent state-of-the-art paper in JACC. In the article, Kutty and his colleagues explore what we know and what we don’t know about patent foramen ovale.


About a quarter of adults have a patent foramen ovale (PFO). When this normal anatomic variant for infants persists into adulthood, it can be associated with cryptogenic stroke, transient ischemic attack, migraines, and other conditions. Available management strategies include observation, medical therapy, percutaneous intervention, and surgery. Here are some highlights from Kutty’s article:

1. PFO is more common in patients with cyptogenic stroke than in the general population (roughly 50% vs. 20%).

2. In some series, patients with an atrial septal defect or PFO who had suffered multiple recurrent events and then underwent transcatheter defect closure experienced a reduction in the recurrence rate of focal neurological events.

3. The FDA has not approved any currently available PFO-closure devices. Closures are performed off-label using devices approved for other indications.

4. The recently published CLOSURE I randomized trial failed to demonstrate the superiority of PFO device closure plus medical therapy (6 months of aspirin and clopidogrel, followed by 18 months of aspirin) over best medical therapy (24 months of warfarin or aspirin, or a combination).


Ryan: What precisely is the link between PFO and cryptogenic stroke?

Kutty: A stroke may be labeled “cryptogenic” at the time of diagnosis because no clear cause has been identified. Paradoxical embolization of a small venous thrombus through a PFO is one of many causes of cryptogenic stroke. The diagnosis is one of exclusion, as in nearly all cases the thrombus has already moved through the PFO by the time the stroke occurs. Other causes of cryptogenic stroke should be actively considered. If none are found, it may be reasonable to consider paradoxical embolus as a potential cause.

Ryan: Why have randomized trials failed to show definitively whether PFO device closure works or, at the very least, to point to a definitive practice?

Kutty: I think the studies have actually shown that PFO closure, followed by relatively benign antiplatelet therapy, is about as effective as taking powerful anticoagulation drugs such as warfarin. Closing the PFO was not “better” at preventing later strokes, so the trial designs have made some of the trials appear to be failures, but the data tend to show that PFO closure may be a reasonable alternative to lifelong anticoagulation. Interestingly, the causes of later recurrent stokes were often identifiable, so most of them were not “cryptogenic.” It’s not clear whether better screening would have identified the causes before the trial participants were enrolled.

Ryan: Is the frequent off-label use of closure devices a good thing? And what’s the practice in Europe, where regulations are more lenient?

Kutty: I still feel that in selected patients who have truly had a cryptogenic stroke, PFO closure is one alternative for trying to prevent stroke recurrence. I believe we went through a period of over-enthusiasm, when some were suggesting PFO closure as first-line treatment rather than carefully considering it as one option for particular patients. This enthusiasm still seems to be prevalent in Europe.

Ryan: Do you expect the new generation of anticoagulants to improve medical therapy and, specifically, to change the options available to patients with a PFO?

Kutty: We’ll have to see how safe and well tolerated the new medications are. Remember that for a paradoxical embolus to occur, there must first be a clot in the venous system and then the potential for that clot to embolize through a PFO. Eliminating either the clot or the PFO can be protective. Many patients find it difficult to take powerful, expensive medications every day. If PFO closure plus milder, less expensive antiplatelet drugs such as aspirin are as effective as powerful, expensive new medications, some patients may opt to undergo PFO closure and not worry so much about the medications. The REDUCE trial of the Helex device is designed to test whether PFO closure significantly improves outcomes (evident strokes and, also, silent MRI changes) over simple antiplatelet medications alone (which many feel are as good as more-powerful anticoagulants for cryptogenic stroke).

Ryan: If a 30-year-old woman on birth control with a family history of stroke and an atrial septal aneurysm >3 cm comes into your clinic, what do you offer her?

Kutty: I think the birth control pill has a stronger association with possible stroke than the PFO does. The data are not consistent with respect to aneurysms associated with PFO, and I don’t personally consider those to be a large additional risk factor. I would recommend stopping the pill and no treatment for the PFO. Aspirin may be reasonable, since it is cheap and low risk and may have a significant beneficial effect, as I’ve noted. But I don’t currently recommend even considering PFO closure in the absence of a likely paradoxical embolic event.

Ryan: On her way out, the woman realizes she forgot to tell you that she has migraines with aura. Do you ask her to sit back down to discuss PFO closure?

Kutty: I consider PFO closure to be only a last option as a treatment for migraines. I would want a migraine expert to explore all other possibilities first. If the migraines cannot be controlled after trying several other better-proven therapies, PFO closure may be considered after a long review with the patient. Experience and data have shown that PFO closure can have dramatic effects in some individuals, but overall it is no better than placebo.

Share your views about Dr. Kutty’s review of the evidence base on PFO and about how to manage patients with a PFO in clinical practice. 

11 Responses to “Patent Foramen Ovale — Are the Choices Patently Obvious?”

  1. Very nice review. Thank you.

  2. Steven Greer, MD says:

    Neurologist S. Claiborne Johnston, M.D., Ph.D. of UCSF reviews the NMT Medical plagued CLOSURE I trial testing the hypothesis that early closure of patent foramen ovales (PFO’s) reduces the incidence of cryptogenic stroke caused by paradoxical emboli. The study was redesigned in midcourse due to slow enrollment. Insurance reimbursement for off-label PFO closure was diverting eligible patients. In addition, some of the patients highest at risk were excluded from the trial.

    Dr. Johnston discusses what clinicians should take away from the flawed data and mentions a second similar trial with results expected this year. The larger topic of medical device approval in the U.S. is discussed.

  3. I was confused about the last comment by Dr. Kutty: “Experience and data have shown that PFO closure can have dramatic effects in some individuals, but overall it is no better than placebo.” So where does this leave us… and if it is really no better than placebo… where does anecdotal evidence fit?

    • Shelby Kutty, MD says:

      Dr. Krumholz,
      Great question. A pathophysiological relationship between PFO closure and migraine relief (for example, an association of the completeness of PFO closure/absence of residual shunt with migraine relief) has not yet been shown. It can be assumed based on the results of observational studies and anecdotal evidence that, some patients with migraine/migraine with aura who undergo PFO closure for another reason than migraine may experience some improvement in migraine symptoms. But, migraine symptoms can be spontaneously variable and with a placebo response rate around 20-30% in contemporary clinical trials. So, the real benefit of PFO closure as adjunct or alternative to medical treatment of migraine is unknown. Clearly, further data from randomized studies are necessary.

  4. Although the data are not consistent in the association of PFO and atrial septal aneurysm, atrial fibrillation and the latter are so associated. How extensively do we monitor( eg,implantable loop recorders) in patients with cryptogenic stroke?
    I agree with Harlan- how does one identify “some” individuals and dramatic effects?

    • Shelby Kutty, MD says:

      Dr. Prada, see my comment to Dr. Krumholz. Unfortunately, based on available PFO-migraine data, we are currently unable to identify the subset of patients who may have symptom relief. Hopefully the results of the 2 PFO-migraine trails will shed more light into this in the coming years.
      Agree an association of atrial fibrillation with atrial septal aneurysm has been reported. In all reports, however, the number of patients with atrial septal aneurysm, atrial fibrillation, plus embolic events was very small, and a relationship between the three has not been shown. Since atrial fibrillation could be an independent risk factor for embolic events, it is reasonable to investigate an elderly patient, but the evidence is weak.

  5. Joao Ferreira, MD says:

    Dear Dr. Kutty, Dear Members:
    1. From clinical standpoints is there any evidence or expert opinion to use aspirin alone vs warfarin or a combination?
    2. Once we detect a PFO, anti-platelet and/or hipocoagulation should be offered to the patient regardless of foramen size, comorbidities or symptoms?
    With regards
    João Pedro

  6. Shelby Kutty, MD says:

    Dr. Ferreira,
    1. Though an apparent benefit of medical therapy after a cryptogenic stroke has been shown in numerous studies, the best medical treatment remains controversial. Because aspirin has been shown to be as good as warfarin in prevention of stroke recurrence (and it is safe and inexpensive), we use aspirin alone in most patients. Moreover, we have noted a very strong dislike of warfarin in our patients and seen complications from aggressive anticoagulation regimens in some patents.
    2. We do not consider PFO a pathological finding in the absence of comorbidities/other specific clinical reasons, and offer no treatment for an incidentally detected PFO.

  7. David Powell , MD, FACC says:

    Xavier: I generally do outpatient telemetry for 3 weeks to look for occult afib in cryptogenic stroke. I now consider episodes of at least 30 sec to warrant anticoagulation. As we perform in- hospital TEEs fairly frequently for cryptogenic stroke, I have 2 questions (whose answers may be unknown): (1) If a TEE “shortly” after a CVA shows neither thrombus nor stigmata for atrial stasis, does that exclude afib as an etiology..and the integrally related (2) how long after left atial embolic stroke will such TEE findings persist. I vaguely recall a study suggesting early (within 48 hrs) TEE to increase sensitivity for thrombus/smoke (in the absence of known afib).

  8. Marcio Jablonka, MD says:

    I have a patient with past history of occipital stroke, and diagnosis of PFO 5 years ago. After that she has been kept on warfarin, and there has been no issues controlling her INR. She is currently asymptomatic. Is the evidence supporting a change to aspirin due to less risk of bleeding and similar outcomes to warfarin therapy?
    thank you