Patent Foramen Ovale and Cryptogenic Stroke

Abstract & Commentary

By Michael H. Crawford, MD

Source: Handke M, et al. Patent foramen ovale and cryptogenic stroke in older patients. N Engl J Med. 2007;357: 2262-2268.

Patent Foramen Ovale (PFO) has been associated with cryptogenic stroke in young individuals, but the association is less clear in older individuals where other causes of stroke predominate. Thus, this study from Germany is of interest. Out of 596 patients admitted to their stroke center, 503 were able to undergo transesophageal echocardiography (TEE), and thus formed the study population. They also had head CT, MRI, or both and carotid artery duplex sonography. Cerebral angiography by MRI or contrast, before transarterial thrombolysis, was done in 231 (46%). In addition, transthoracic echo and ECG were done, and those suspected of having atrial fibrillation had 24-hour Holter monitoring. Before TEE, the TOAST criteria were applied to classify the patients into one of the 4 known-causes of strokes or cryptogenic stroke. TEE was done a median of 2 days after the stroke with contrast and Valsalva.

Results: Of the 503 patients, 26% were younger than 55 years of age. A cause of stroke could be identified before TEE in 55% of the patients. The rest were classified as cryptogenic (63% younger and 39% of the older patients). The revalence of PFO was significantly greater in patients with cryptogenic stroke vs known cause stroke in the younger (44% vs 14%, OR 4.7, P < .001) and older patients (28% vs. 12%, OR 2.9, P < .001), as was the prevalence of PFO with atrial septal aneurysm (13 vs 2%, OR 7.4 younger and 15 vs 4% or 3.9 in the older). Multivariate analysis showed that PFO was independently associated with cryptogenic stroke in both groups (OR 3.7 young and 3.0 older). Handke and colleagues concluded that PFO with paradoxical embolism is a cause of cryptogenic stroke at all ages.


Previous studies of the association between PFO and cryptogenic stroke have focused on younger patients because they have few confounding diseases that also are associated with stroke. Thus, when PFO is found in younger cryptogenic stroke patients, especially if atrial septal aneurysm is present as well, the decision to close it is easy to make. Older individuals present a challenge because of their higher frequency of alternative causes of stroke, as well as the paucity of data in older patients.

This study solves the latter problem by doing TEE in a large group of older patients with stroke, and finding an association with PFO in those classified as cryptogenic. Also, the stronger association of PFO plus atrial septal aneurysm was also observed in the older patients. In addition, the older patients with cryptogenic stroke and PFO were observed to have less thick aortic atherosclerotic plaques. Although the incidence of PFO decreases with age (34% < 30 years vs 20% > 80 years), the incidence of venous thromboembolism, right atrial pressure elevation, and larger PFOs increased with age. Thus, it is not surprising that PFO is associated with cryptogenic stroke in older patients.

The issue is what to do with this information therapeutically? Despite the issue that cryptogenic stroke prevalence was probably inflated in this study, since cerebral angiography was done in only half the patients, the incidence of PFO was < 50% in this group. So, there are many patients with some other unknown cause of stroke. Also, some patients with known cause stroke also had PFOs (12-14%). Just because you have a PFO, does this absolutely mean it is the cause of the cryptogenic stroke? Certainly not. Perhaps in those with PFO plus atrial septal aneurysm and cryptogenic stroke, causality is a better bet, since few with known cause stroke had this finding (2-4%). Until we know the results of the device vs medical therapy trials in cryptogenic stroke currently underway, what do we do? It seems reasonable, if cryptogenic stroke is diagnosed after a full evaluation including MR angiography, and TEE shows PFO plus atrial septal aneurysm, to close it.

With a PFO alone, the decision becomes more difficult. Clearly, if there is any history of deep venous thrombosis, or evidence of it on peripheral Doppler examination, then closure seems reasonable; although you could reasonably offer lifelong warfarin as an alternative. With PFO alone, you and the patient will have to decide between aspirin, warfarin, or closure. How to close the defect is also not entirely clear, and I don't know of any trials of minimally-invasive surgical closure vs percutaneous device closure. A recent young patient of mine opted for minimally-invasive surgery over a device whose long-term history is unknown. Surgical closure in the absence of pulmonary hypertension in a young person should have a mortality of zero. Although this study contributes important new information on the topic, the management of PFO in patients with cryptogenic stroke at all ages is still a challenge.