PFO Closure to Prevent Recurrent TIA and Stroke

What is the evidence base for patent foramen ovale closure?

By Michael H. Salinger, MD, FACC, FSCAI; Mehmet Cilingiroglu, MD, FACC, FESC, FSCAI; Justin P. Levisay, MD, FACC, FSCAI; and Ted E. Feldman, MD, FACC, FESC, FSCAI

It has been 133 years since the German anatomist J.Cohnheim first postulated a relationship between patent foramen ovale (PFO) and the death of a young woman from nonhemorrhagic stroke.1 There has been ongoing debate and investigation regardingPFO and what has been termed “cryptogenic stroke.”Parties to this debate have included neurologists and cardiologists.More recently, nonphysician groups have entered into the debate in the form of health insurance companies. As we look forward to the presentation of the first studies from randomized prospective studies on percutaneous closure of PFOs, we will explore the evidence base that has already evolved regarding PFO closure and the prevention of recurrent TIA and stroke. The outline that this article will follow will expand on the following points: (1) PFO is highly associated with cryptogenic stroke; (2) thromboembolism is more commonly fatal when associated with PFO; (3) PFO prevalence decreases with age, possibly due to the attrition of people with PFO; (4) PFO is seen in 45% of travelers with stroke; (5) warfarin seems ineffective for stroke prevention;(6) recurrent stroke appears to be diminished byPFO closure; and (7) randomized trials of closure versus medical therapy are ongoing.

The association between PFO and cryptogenic stroke has been identified in multiple epidemiological studies.2-11Although the prevalence of PFO in the general population ranges from 15% to 25%, evaluations of patients with cryptogenic stroke have shown a PFO prevalence of 40%to 60%. The relationship of PFO and cryptogenic stroke is most apparent when evaluating patients younger than 55years. Additional data exist to support the relationship ofPFO in older patients as well.12 Two population-based studies of 1,072 and 585 patients, respectively, have failed to demonstrate a relationship between PFO and stroke.13,14 However, these studies are underpowered given that there are only 50,000 estimated cases of cryptogenic stroke per year the United States. Based on available census data, that computes to an incidence of one case per 6,000 individuals per year. Study populations of many thousands would be needed to evaluate the relationship between cryptogenic stroke and PFO in the general population.

The relationship between PFO and stroke has been shown in several small populations with increased thrombotic risk. One such group includes travelers who remain seated for long periods of time. In one observational study of 338 subjects presenting with a first-time cerebral ischemic event, the frequency of PFO among the patients with a history of recent travel was 45% compared to an11% incidence in subjects with cerebral events and no history of travel.15 In a study of pulmonary embolism patients, those with PFOs had a six-fold greater incidence of stroke, a 15-fold increase in the incidence of peripheral embolism, and a doubling of the incidence of death compared to the cohort of pulmonary embolism patients without PFOs.16

Dramatic visual demonstration of the relationship between PFO and stroke has come from the operating room, echocardiography laboratory (Figure 1), and autopsy suite. Casts of lower extremity or pelvic veins have been seen on echocardiograms in many cases and have been recovered at autopsy in some.

An association between PFO and mortality has also been made. Autopsy data of nearly 1,000 normal hearts from the Mayo Clinic reported an overall 27% incidence of PFO; however, when stratified by age, an interesting pattern evolves.17 The incidence of PFO in patients younger than 30 years was 34%; in patients aged 31 to 80 years, thePFO incidence decreases to 25%; and in patients aged 81to 99 years, the incidence of PFO was 20%. Because PFO is not known to close spontaneously with age, this observation suggests that PFO is associated with ongoing excess mortality with increasing age. A systematic review of autopsies that were performed on 3,430 geriatric patients in Switzerland showed a surprisingly low (0.2%) prevalence of PFO.18 This suggests that with advancing age,people with PFO are more likely to die when they have other associated cardiovascular problems.

Despite the data, some still debate whether the relationship between PFO and cryptogenic stroke is only incidental rather than pathologic. A recent statistical analysis of the relationship between PFO and cryptogenic stroke was conducted by applying Bayes’ theorem to 23case-controlled studies. These investigators found that in cryptogenic stroke patients younger than 55 years who had both a PFO and an atrial septal aneurysm, the probability that the PFO was incidental to the stroke was less than one in 10.19

Multiple studies have evaluated long-term medical therapies for patients with cryptogenic stroke, including antiplatelet therapy with aspirin, persantine, or clopidogrel,or anticoagulation therapy with warfarin. A number of nonrandomized, prospective follow-up studies have suggested that there is no benefit associated with anticoagulation therapy over antiplatelet therapy.20-23 The PICSS study remains the only randomized controlled trial comparing aspirin to warfarin as medical therapy for PFO-associated stroke.24 In the PICSS subgroup with transesophageal echocardiography demonstrated PFO, the 2-year recurrent stroke/transient ischemic attack (TIA) rates were similar in the warfarin and aspirin cohorts (16.5% vs 13.2%; P = .65).This suggests that neither aspirin nor warfarin is an effective therapy. There are only limited data regarding the natural history of PFO-associated cryptogenic stroke. In one small study of 33 patients, the annual recurrent event rate without treatment was 16%.25 When viewed in the context of this small natural history study, the PICSS data might lead one to conclude that both warfarin and aspirin are equally ineffective.

The problems associated with long-term warfarin anticoagulation therapy are well known. Contemporary warfarin therapy is associated with a 5% to 15% incidence of intracranial hemorrhage26 and, at any given time, < 25%of patients on warfarin are actually in the prescribed therapeutic range.27

Evidence for a decrease in the recurrent stroke and TIA rates in patients with a history of cerebral events who subsequently undergo PFO closure can be found in a large number of nonrandomized single-center experiences.20,28-43 In the majority of these studies, the event rate in the year before closure is compared to the event rate in the year after closure. The event rates before closure range from 2% to 26%, and the range of recurrent events after closure is between 1% and 2.5%. For example,Windecker followed 308 patients with cryptogenic stroke; 150 patients were treated with percutaneous closure using a variety of devices, and 158 patients were treated with medical therapy that was equally divided between antiplatelet and anticoagulant therapy.28 At 4 years, the cohort of patients who received devices (n = 122) and had complete closure at 6 months had a significantly lower recurrent event rate compared to the medical therapy group (6.5% vs 22.2%; P = .04). The importance of device design and the ability to achieve complete closure shows the potential efficacy of device therapy and is an important consideration as randomized data becomes available in the future. In one experience (n = 128) using the Helexdevice (W. L. Gore & Associates, Flagstaff, AZ), the complete closure rate at a mean of 21 months follow-up was90%, and the recurrent stroke rate was 0%.44 These two reports are typical of many that are all highly concordant.

The strongest evidence supporting the hypothesis thatPFO closure is effective in reducing recurrent ischemic events comes from several meta-analyses (Figure 2).45Landzberg and Khairy reported a compilation of 20 studies that included 2,250 patients with an adjusted 1-year stroke and TIA rate of 7.07% before closure and 2.71% in the year after closure.46 Homma and Sacco provided a meta-analysisof 26 studies involving 2,534 patients with recurrent event rates per 100 patient years of 5.55% in the surgical closure group, 4.86% in the medical therapy group, and 2.95% in the percutaneous closure group.2 Furthermore, Wöhrle reported an analysis of 20 studies involving 3,014 patients.47The 1-year recurrence rate for stroke or TIA was 5.6% in the medical therapy group and 1.3% in the device closure group. In these three analyses involving thousands of patients, the proportional decrease in recurrent cerebral events with PFO closure as compared to medical therapy was 62%, 39%, and 75%, respectively.45

These meta-analyses are limited by methodological issues that are inherent when analyzing nonrandomized data across different patient cohorts. It is possible that some patients that were considered to have TIA may have had complex migraines rather than TIA. Many of the reports in these meta-analyses included use of older devices with lower complete occlusion rates. Although the effects of some of the methodological issues are not predictable, some of the issues may strengthen the conclusion that outcome differences between medical therapy and closure in these analyses are indeed genuine. For example,older devices with lower closure rates did not diminish the consistency of the reductions in recurrent stroke or TIA in these studies.

Randomized trial data are forthcoming but may not be immediately conclusive. There is some potential for selection biases because patients with high-risk characteristics may not be as likely to be submitted to randomization,but rather are preferentially treated with device closure outside of the randomized trials.Incomplete closure rates or thrombosis issues that are unique to specific devices may also leave questions unanswered. Also, inclusion of migraine patients who appear to have TIA may skew outcomes. Finally, payment decisions by third-party payers in the form of payment refusals may limit the inclusion of younger patients in whom the relationship between PFO and cryptogenic stroke appears to be the most robust.

Recently, in the form of a company press release, it was announced that the 960-patient, NMT Medical-sponsored(Boston, MA) CLOSURE I trial failed to meet its primary endpoint.48 Preliminary results indicate that device closure with the StarFlex device (NMT Medical) provided a small but not statistically significant benefit over current medical therapy. The group who received theStarFlex device had low complication and thrombus formation rates. Complete closure rates in the trial were reported to be 86.5%. Details of the study are expected to be presented at 2010 American Heart Association meeting later this fall.

The RESPECT trial (sponsored by AGA MedicalCorporation, Plymouth, MN) is structured as a 900-patient, device superiority trial and is currently enrolling.There is also the 600-patient REDUCE trial, which is studying the Helex device. The REDUCE trial will be unique in the manner in which the protocol standardizes the medical therapy in the cohort not receiving a device. TheREDUCE trial uses brain imaging for both selection and outcome endpoints, and patients are required to have clear imaging-based evidence of embolic stroke to be included. The use of brain imaging has the potential to provide hard data endpoints upon which firm conclusions might be better drawn. Furthermore, a 2:1 randomization scheme may assist in the pace of enrollment.

In 2004, in the absence of controlled randomized data,the American Academy of Neurology concluded that there was insufficient evidence to evaluate the efficacy of surgical or endovascular closure for PFOs to prevent recurrent stroke or TIA.49 IN 2009, the American College of Cardiology issued a statement calling for the completion of randomized trials and asking involved parties to support patient referral to ongoing randomized trials.50In view of the controversy, completion of randomized trials appears to be necessary to validate the clinical use ofPFO device closure to prevent recurrent stroke and TIA.

It is a common perception among clinicians that there are no data to support PFO closure for prevention of recurrent TIA or stroke. As we have presented, there are indeed numerous studies with thousands of patients at multiple centers using a variety of devices and closure techniques with highly consistent outcomes, all supporting the conclusion that PFO closure reduces recurrent events in cryptogenic stroke patients.

Although data from the first completed randomized trial will be presented later this year, there are methodological issues that may cloud the conclusions once these data are released. These include potential patient selection biases in which patients with high-risk characteristics may not be enrolled in the trials but rather are preferentially treated outside of the randomized trial. Incomplete closure rates or thrombosis rates that are unique to specific devices may also contribute to uncertainty over these results, as well as the variability of medical therapy and the inadvertent inclusion of patients with migraine rather than TIA. Ongoing randomized trials including only brain imaging-positive patients and uniform medical therapy in the nondevice controls may provide clearer endpoints upon which firmer conclusions might be drawn.

In the absence of randomized data, we currently have numerous studies with thousands of subjects upon which to base conclusions and direct the care of our patients.These results are highly consistent and concordant and support the observation that PFO closure reduces recurrent cerebral embolic events in cryptogenic stroke.

Michael H. Salinger, MD, FACC, FSCAI, is an interventionalcardiologist at NorthShore University HealthSystems inEvanston, Illinois; and Associate Professor of Medicine at theUniversity of Chicago Pritzker School of Medicine in Chicago,Illinois. He has disclosed that he holds no financial interest inany product or manufacturer mentioned herein. Dr. Salingermay be reached at (847) 570-2250;

Mehmet Cilingiroglu, MD, FACC, FESC, FSCAI, is an interventionalcardiology advanced fellow in structural heartdisease at the NorthShore University HealthSystems inEvanston, Illinois. Financial interest disclosure informationwas not available at the time of publication.

Justin P. Levisay, MD, FACC, FSCAI, is an interventionalcardiologist at NorthShore University HealthSystems inEvanston, Illinois; and an instructor at the University ofChicago Pritzker School of Medicine in Chicago, Illinois.Financial interest disclosure information was not availableat the time of publication.

Ted E. Feldman, MD, FACC, FESC, FSCAI, is Director of theCardiac Catheterization Laboratory at Evanston Hospital inEvanston, Illinois. He has disclosed that he received researchgrant funding from Gore, and is a consultant to Gore andCoherex. Dr. Feldman may be reached at (847) 570-2250;


Contact Info

For advertising rates and opportunities, contact:
Craig McChesney

Stephen Hoerst

Charles Philip

About Cardiac Interventions Today

Cardiac Interventions Today (ISSN 2572-5955 print and ISSN 2572-5963 online) is a publication dedicated to providing comprehensive coverage of the latest developments in technology, techniques, clinical studies, and regulatory and reimbursement issues in the field of coronary and cardiac interventions. Cardiac Interventions Today premiered in March 2007 and each edition contains a variety of topics in a flexible format, including articles covering various perspectives on current clinical topics, in-depth interviews with expert physicians, overviews of available technologies, industry news, and insights into the issues affecting today's interventional cardiology practices.