Patent foramen ovale closure for stroke prevention: A myriad of unanswered questions

Migraines, including those with aura, are common and generally benign, affecting between 10% and 15% of the population. Patent foramen ovale (PFO) is also common, found in 20% to 25% of the population. This anatomic anomaly has been associated with stroke in young individuals, and, more recently, with migraines with aura. Spurred by advances in endovascular therapy, PFO closure has become easier and more readily available. In the hands of experienced interventionalists, endovascular PFO closure can be accomplished with very low risk.

Dr Meier reports on the largest series of PFO closures using the Amplatzer® PFO Occluder. In 396 patients treated with this device, complete closure was achieved in 91% of patients at 6 months. This was accomplished with a remarkable 1-year event rate of 1%. The outcome “events” in this study included recurrent ischemic stroke, transient ischemic attack, other peripheral embolism, or device problems. A 0.5% incidence of thrombi on the devices was noted on follow-up echocardiography and treated with anticoagulation. Based on these results, the author concludes that PFO closure should be performed for primary prevention of stroke with the added benefit of reducing migraines.

Dr Meier’s report has a few limitations because it leaves out important information regarding outcome measures, methods, and inclusion criteria, including indications for treatment, definitions of outcome events, independent neurological adjudication of postprocedural outcomes, and initial and recurrent event adjudication. Considerably more studies need to be conducted before it is possible to conclude that PFO closure is a safe, effective, and appropriate procedure for primary or secondary prevention of stroke and migraines. As of yet, no randomized trials have shown that PFO closure reduces the risk of stroke or migraines. In fact, in a large randomized trial for migraine prevention, PFO closure demonstrated no benefits over sham therapy, but was associated with increased complications.¹ Although the design of this trial has been highly criticized, it remains the only prospective randomized trial to evaluate PFO closure as a means of migraine prevention. All of the other available data come from retrospective series without any standardized outcome measures or definitions of migraine.

Based on the available data, is it truly prudent to surgically treat a disorder that can be treated medically, especially when surgical intervention exposes patients to the risk of stroke, vascular access site complications, and infections, and requires them to take potent antithrombotic treatments and anticoagulants? Furthermore, although there is a link between migraine with aura and PFO, the nature of that link is unclear. Migraines, including those with aura, are a genetic disorder that predisposes individuals to headaches and local brain dysfunction. PFO and migraines may simply share a genetic locus and be unrelated pathophysiologically. Even if they are related, the benefits on migraines with aura reported in Dr Meier’s study may be a placebo effect or attributable to the medical regimen used in these patients. If one posits that PFO is causal in some cases of migraine, then R-L shunt and paradoxical embolism are the likely mechanisms. Is it not possible then that the benefit from PFO closure stems from treatment with antithrombotics? Perhaps the migraineurs with PFO get headaches because of microscopic emboli that are easily abolished with aspirin or clopidogrel. In such cases, antithrombotic therapy rather than PFO closure is the effective treatment. Without a randomized trial, we cannot know the answer with certainty.

Data regarding stroke prevention and PFO closure on the natural history of cryptogenic stroke in young individuals with PFO are lacking. The risk of recurrent stroke in these patients is low, and in the largest randomized trial of stroke prevention that included patients with PFO, the recurrent stroke rates were no different between those who had a PFO and those whose strokes were cryptogenic.² These results raise the possibility that patients with PFO may have an alternative cause of stroke, such as a hypercoagulable state that would not be ameliorated by PFO closure, but may be treated with antithrombotic or anticoagulant therapy.

PFO closure certainly may reduce the risk of stroke in some patients, and it is logical to reason that it should be considered in those who have recurrent events despite medical therapy or in those who have undergone an extensive evaluation by a stroke neurologist to exclude other causes of stroke. Regarding primary prevention, there are no data to guide treatment. The risk of stroke in such patients is unknown, but is likely very low and may be lower than the risk of the procedure. The authors themselves point out that the current clinical trials will have only a few years of follow-up and will not show benefit for PFO closure because “events often take decades to happen.” For this very reason, large, randomized, prospective studies are needed.

On a final note, although Dr Meier’s facility and experience with PFO closure is remarkable (1000 patients with procedural durations as short as 15 minutes), procedural outcome data from 1 center are not sufficient to arrive at a conclusion to this issue. The results could be unique due to random factors or because of more specific factors, such as experience with the procedure, anesthetics, anticoagulants, venipuncture techniques, and patient population. Also, more data are needed to understand if 1 device is better than another, especially if closure rates are as different as the authors suggest. In the United States, an interventionalist needs training on as few as 3 PFO closure cases to be allowed to begin treating patients; thus, it is unrealistic to expect that the results that Dr Meier and colleagues achieved can be generalized.