A subscription to JoVE is required to view this content. Sign in or start your free trial.
Review Article
This review aims to summarize the current diagnostic strategies and therapeutic interventions for patent foramen ovale (PFO)-associated cryptogenic stroke (CS). We comprehensively evaluate the underlying pathogenic mechanisms, clinical assessment tools, and ongoing controversies regarding interventional closure versus antithrombotic therapy, proposing mechanism-based hypotheses to guide future clinical research.
Patent foramen ovale (PFO) is an important cause of cryptogenic stroke (CS), and its pathogenic mechanisms involve paradoxical embolism, in situ thrombus, hemodynamic changes caused by atrial remodeling, and high-risk anatomical features such as atrial septal aneurysm. Clinical assessment requires multidisciplinary evaluation of anatomical high-risk features, stroke imaging characteristics, and Risk of Paradoxical Embolism (RoPE) score to determine treatment plans. Currently, interventional closure and oral antithrombotic therapy are recommended, but the determination of interventional indications and the choice between anticoagulants and antiplatelet drugs remain controversial. From the perspective of pathological mechanisms, patients with PFO-related CS may have venous emboli caused by right-to-left shunting, and anticoagulant therapy can more precisely intervene in the chain reaction of embolism by inhibiting venous thrombus formation and promoting the dissolution of in situ thrombi. Based on the PICO principle, this review summarizes currently published studies on PFO, including clinical trials, guideline statements, and position papers. It summarizes the diagnosis and treatment of PFO and proposes mechanism-based hypotheses for validation in future large-scale clinical studies.
During the fetal period, the foramen ovale is a physiological opening in the atrial septum of the heart. In most people, ~2 months after birth, the primary and secondary septa approach each other and fuse, closing the foramen ovale. If by age 3, the primary and secondary septa have not fused, leaving a gap between them, it is called patent foramen ovale (PFO). The prevalence of PFO in the general population ranges from 14.7% to 31.3%, with ~25% of adults having PFO1.
PFO is usually asymptomatic but may lead to paradoxical embolism and thereby contribute to cryptogenic stroke (CS). While PFO has historically been recognized as a risk factor primarily in young adults, it is important to note that the risk of venous thrombosis—the prerequisite for paradoxical embolism—increases exponentially with age and comorbid illness. Consequently, PFO represents a minor risk factor in younger individuals but may constitute a major risk factor for CS in elderly patients and those with significant underlying disease. In a 2004 clinical series of 402 consecutive patients with ischemic cardioembolic stroke, patent foramen ovale and atrial septal aneurysm were identified as the cardiac source of embolism in only two patients2.
With further research on PFO, it has been found to be associated with various diseases such as migraine, syncope, and acute myocardial infarction3, and is especially closely related to the pathogenesis of CS. In 2021, Pristipino et al. provided evidence-based recommendations for decompression sickness, migraine, arterial hypoxia syndrome, and certain high-risk clinical conditions in patients with PFO4. Patients diagnosed with CS after standard assessment account for approximately 30% to 40% of ischemic strokes5. Accumulating evidence suggests that PFO may constitute a potential risk factor for ischemic stroke6. Effectively preventing or treating PFO-related CS and identifying PFOs likely to cause CS are issues that require attention. This article systematically reviews the pathophysiological basis, diagnostic criteria, and therapeutic advances in PFO-associated CS, highlighting the potential advantages of anticoagulant therapy and its mechanistic rationale. By comparing divergent recommendations across international guidelines, this review provides evidence-based recommendations for clinical decision-making.
Relevant studies were identified through searches of electronic databases, including PubMed, Embase, and the Cochrane Library, according to the PICO principle. Search terms included "patent foramen ovale," "cryptogenic stroke," "PFO closure," "anticoagulation," "antiplatelet therapy," "risk stratification," "guidelines," and "position statement." Study types considered for inclusion comprised randomized controlled trials, observational studies, meta-analyses, clinical guidelines, expert consensus statements, and position papers related to PFO diagnosis and management. Studies were excluded if they were case reports without generalizable clinical data, non-English and non-Chinese publications without available translations, or studies addressing PFO in clinical contexts unrelated to stroke or systemic embolism. The status of PFO diagnosis and treatment was summarized with a focus on pathogenesis and pathophysiology.
Access restricted. Please log in or start a trial to view this content.
Mechanism by which PFO leads to CS
During the embryonic period, before the pulmonary circulation is formed, the foramen ovale serves as the only valve between the right and left atria of the human heart, allowing continuous communication through the right-to-left circulatory pathway. Postnatally, the establishment of pulmonary circulation alters the interatrial pressure gradient: left atrial pressure increases while right atrial pressure simultaneously decreases. Both hemodynamic changes act in concert to passively appose the foramen ovale valve against the interatrial septum, resulting in functional closure. Current understanding suggests that ischemic stroke associated with PFO can be mediated by three underlying mechanisms: paradoxical embolism, in situ thrombosis, and hemodynamic factors7.
Paradoxical embolism constitutes a primary mechanism underlying ischemic stroke in PFO patients, with its incidence positively correlating with the PFO diameter. The larger the PFO diameter, the larger the embolic material that can pass through, resulting in a higher incidence of paradoxical embolism8. Failure of normal postnatal fusion between the septum primum and septum secundum leaves a patent channel measuring 1–6 mm in width and approximately 7 mm in length. When right atrial pressure exceeds left atrial pressure, the septum primum on the left side thins out and is pushed open due to the pressure gradient, thereby creating a right-to-left shunt (RLS). This serves as the pathological basis for paradoxical embolism. Such paradoxical embolism is a potential risk factor for migraine and ischemic stroke. In particular, actions such as performing Valsalva maneuvers, coughing, or defecating, which can increase intra-abdominal pressure, can exacerbate RLS and further cause various symptoms9. Emboli originating from the venous circulation or right atrium can circumvent the pulmonary vasculature and gain direct access to the arterial system through the foramen ovale. Their subsequent migration to the cerebral circulation induces a stroke, a phenomenon termed paradoxical cerebral embolism10.
Abnormal embolism occurs not only due to the presence of PFO but also due to the presence of a thrombus. A hypercoagulable state of blood easily leads to venous thrombosis and in situ thrombus formation at the site of PFO11. In situ thrombosis—the development of a blood clot directly within the foramen ovale—has been proposed as a potential mechanism, but it is exceedingly rare in clinical practice, if it occurs at all, and its contribution to systemic embolism is considered negligible compared with paradoxical embolism. Nonetheless, a hypercoagulable state, frequently observed in stroke patients with PFO and often driven by prothrombin gene mutations, remains an important contributing factor to overall thromboembolic risk12. Additionally, the reason why paradoxical embolism rarely occurs in patients with a normally closed foramen ovale is that lung tissues, such as pulmonary epithelial cells and pulmonary vascular endothelial cells, produce tissue plasminogen activator (t-PA)13,14, a serine protease that cleaves plasminogen into plasmin and thereby promotes the dissolution of thrombi.
During normal pulmonary circulation, venous blood is pumped into the lungs via the pulmonary valve and enters the pulmonary artery. The t-PA produced by lung tissue can dissolve small thrombi in venous blood, reducing the risk of stroke from venous thrombosis. After oxygenation in the lungs, pulmonary venous return delivers oxygenated blood to the left atrium. It then proceeds to the left ventricle, from which it is pumped throughout the systemic vasculature15. In the presence of a PFO resulting in RLS, a fraction of venous or right atrial blood circumvents the pulmonary circulation and is shunted directly into the arterial system. The blood does not encounter t-PA from lung tissue, meaning that existing thrombi are not dissolved and may repeatedly circulate in the bloodstream, which is one of the risk factors for stroke. Moreover, PFO has the potential to modify the electrophysiological properties of the left atrium. Studies have confirmed that myocardial cells in patients with an abnormal interatrial septum are more fragile than those without such abnormalities16, making PFO patients more susceptible to arrhythmias. If atrial fibrillation occurs, atrial contractions weaken, leading to blood stasis and increasing the risk of thrombus formation.
Finally, the atrial structure interacts with hemodynamics. Changes in atrial structure lead to alterations in hemodynamics, which further exacerbate these changes. Typically, right-to-left shunting through a PFO occurs only intermittently under specific triggering conditions. If Left Atrial Enlargement (LAE) occurs, it may become a persistent bidirectional shunt, leading to hemodynamic disorders in the left atrium and promoting thrombus formation17.
Compared to healthy individuals, patients with PFO have longer blood retention times in the left atrium. Patients with PFO combined with stroke have even longer blood retention times than asymptomatic patients. This may be attributed to the PFO, which generates a vortex of blood entering the left atrium via the foramen ovale, disrupting the normally orderly vortex system of the left and right pulmonary veins and thus prolonging blood retention time. This could also be a potential mechanism by which PFO leads to CS.
Based on the pathophysiological mechanisms and theoretical foundations of PFO leading to CS, this review proposes that thromboprophylaxis in such patients may favor anticoagulation therapy over antiplatelet therapy. This view is informed by landmark randomized trials and comparative analyses evaluating PFO closure, anticoagulation, and antiplatelet therapy for the prevention of recurrent stroke18,19.
Clinical diagnosis of PFO
Current research indicates that the PFO channel length spans 3–18 mm (mean, 8 mm), while its diameter measures 1–19 mm (mean, 4.9 mm), with both dimensions exhibiting an age-dependent increase20. With broader access to diagnostic technologies, PFO is being diagnosed more accurately and promptly in clinical settings. Current diagnosis of PFO primarily relies on imaging techniques, especially echocardiography, which can effectively identify and assess structural abnormalities in the heart and observe hemodynamic characteristics, providing crucial diagnostic information for physicians. There are several main diagnostic methods.
Contrast-enhanced Transcranial Doppler (c-TCD) is a noninvasive diagnostic modality that uses intravenously administered contrast agents to enhance ultrasonic detection of cerebral hemodynamics through the transcranial acoustic windows. It has high safety and can indirectly confirm the existence and size of PFO by monitoring whether the contrast agent ultimately appears in the cerebral blood flow after being injected into the vein21. cTCD has high sensitivity in detecting PFO, especially during Valsalva maneuvers. It is a safe examination method, but it cannot provide anatomical details of the PFO and is somewhat dependent on the operator's skills.
Transthoracic Echocardiography (TTE) is a technique that uses ultrasound to examine the heart through the chest wall, assessing cardiac structure and function. Although it can be used to detect PFO, its sensitivity is low. Contrast-enhanced TTE (cTTE), which enhances cardiac ultrasound imaging by injecting ultrasound contrast agents, can improve the sensitivity of PFO detection. Most secondary hospitals and above can perform this procedure, making it widely applicable. It is a relatively simple and safe method for diagnosing PFO, particularly suitable for the initial screening of PFO22. However, it is difficult to conduct further accurate assessment, to detect smaller PFO, and to evaluate the anatomical details of PFO.
Enhanced transesophageal echocardiography (Contrast-enhanced TEE, cTEE) is a semi-invasive technique that involves placing an ultrasound probe in the esophagus to assess cardiac structure and function. Because of its proximity to the heart, it provides clearer images than TTE, thereby establishing its status as an important benchmark for identifying a PFO. It can clearly display the location, size, and other anatomical features of PFO, not only for diagnosing PFO but also guiding the closure of PFO23. However, limitations include the invasive nature of the examination, patient discomfort, and the need for professional operation and interpretation.
Right heart contrast echocardiography (RHCE), which can dynamically display whether RLS occurs in congenital heart disease after contrast agent enhancement, effectively evaluates whether RLS occurs and the degree of shunt in patients with PFO under stress, is a noninvasive examination, but it requires intravenous injection of contrast agent and has high technical requirements for the operator24.
Cardiovascular magnetic resonance (CMR) is an advanced imaging modality that noninvasively assesses cardiac structure and function, eliminating the need for contrast agents. CMR can assess chamber sizes, myocardial motion, and hemodynamics25. However, CMR equipment is expensive and has limited availability, and it is contraindicated in some patients with metal implants. The advantages and disadvantages of various diagnostic methods are shown in Table 1. In summary, diagnosing PFO requires combining a medical history with echocardiographic examinations. When evaluating the relationship between PFO and CS, multidisciplinary teams (including neurologists, cardiologists, and imaging experts) often need to discuss and determine the correlation between them.
Clinical evaluation of PFO causing CS
According to the 2019 Global Burden of Disease study, potentially modifiable risk factors accounted for 87.9% of disability-adjusted life years (DALYs) attributable to ischemic stroke26. These potentially modifiable risk factors can be reduced through primary prevention, thereby reducing the DALYs associated with ischemic stroke. The timely recognition and stratification of patient-specific risk factors underpin the primary prevention of ischemic stroke. Currently, up to half of CS cases may be related to PFO27. Data based on four major trials (RESPECT, CLOSE, DEFENSE-PFO, REDUCE) in 2021 indicated that, for patients aged ≤60 years with cryptogenic stroke, PFO closure may be considered to prevent recurrence, provided that a comprehensive evaluation has been performed and the procedural benefits and risks—particularly atrial fibrillation (AF)—have been thoroughly discussed28. However, closure or drug treatment of PFO unrelated to stroke occurrence may fail to reduce stroke recurrence and may expose patients to unnecessary risks. It is paramount to establish a causal attribution between the patient’s PFO and the index stroke, and to identify high-risk PFO phenotypes to guide subsequent therapeutic strategies.
For PFO patients, imaging features of associated cryptogenic stroke can guide clinical assessment. Among CS patients harboring a PFO, studies suggest the following imaging features are observed: multiple small cortical infarcts on T2WI or DWI, fewer lesions in subcortical/cortico-subcortical regions, and predominantly posterior circulation involvement29.
Furthermore, high-risk PFO anatomy includes atrial septal aneurysm (ASA), prominent Eustachian valve or Chiari network, and large PFO (primary-secondary septum distance ≥ 2 mm during Valsalva)30RLS at rest, high volume RLS (≥ 20 microbubbles), long PFO channel (≥ 10 mm), and excessive septal mobility. These are considered high-risk anatomical features for stroke in PFO patients. The presence of these factors increases the risk of recurrent stroke31,32.
Kent developed the Risk of Paradoxical Embolism (RoPE) scale in 201333, which includes six factors: age, history of hypertension, history of diabetes, history of stroke or transient ischemic attack (TIA), smoking status, and presence of cortical infarction on imaging. The total score ranges from 0 to 10 points and can predict the association between PFO and CS before TEE. He et al. proposed the PFO judgment formula (Hr-PFOJ) in 2018, which is based on clinical presentation, neuroimaging characteristics, and laboratory test results. It comprehensively evaluates patient history, infarct pattern, platelet function, coagulation function, and hemodynamics to assess the correlation between stroke and PFO. This formula is more comprehensive than the RoPE scale and has higher sensitivity and specificity in predicting PFO-related strokes. Generally, after assessment, if the PFO diameter is ≥ 2 mm, transcatheter PFO closure is recommended. However, based on these scales, some patients with PFO have a smaller hole size than the current treatment threshold for interventional closure and still carry a risk of causing ischemic stroke.
For such patients, particular attention should be paid, given the high risk of CS recurrence. Prevention before CS occurs is crucial. There are also studies showing that hyperhomocysteine (Hcy) can impair vascular endothelial function and lead to a hypercoagulable state in the blood, increasing the risk of thrombosis34. In PFO patients, elevated Hcy levels heighten CS risk and guide clinical evaluation. Such patients also need to be carefully screened. In 2024, Chiriac et al. proposed a new risk stratification algorithm34. First, the RoPE score was used for preliminary screening. For patients at high risk on preliminary screening, more detailed evaluations were conducted, including imaging such as cTEE or CMR to assess the anatomical characteristics of PFO. Blood indicators, such as Hcy levels, were detected. Patients were stratified into high-, intermediate-, and low-risk groups based on a comprehensive assessment to optimize the clinical management of PFO.
Based on the above considerations, clinicians can assess the likelihood that PFO causes CS using available scales. Nevertheless, individuals with minor atrial septal defects (<2 mm) who are ineligible for PFO closure remain susceptible to CS. This review suggests that a persistent PFO may be an independent risk factor for CS. These patients should regularly visit hospitals as they age to evaluate whether further interventional closure or medication is needed to prevent CS.
Treatment strategies for PFO
Because RLS caused by PFO may promote CS, interventional closure and drug therapy are currently used in clinical practice to eliminate thrombi or prevent CS. In a long-term follow-up study spanning up to 15 years, Wahl et al.35 monitored 308 patients with cryptogenic stroke and concurrent PFO. Their findings revealed that the incidence of recurrent stroke, TIA, or peripheral embolism was significantly reduced in the PFO closure cohort (11%) compared to the medical therapy alone group (21%). Notably, at 10-year follow-up, patients who underwent PFO closure also demonstrated a significant reduction in overall mortality compared with those receiving drug treatment alone, underscoring the durable protective effect of interventional closure. Additional trials comparing PFO interventional closure with antiplatelet therapy for prevention of recurrent stroke have shown a lower recurrence rate of ischemic stroke after interventional closure than after drug treatment alone18. However, not all PFO patients should undergo interventional treatment, which may waste medical resources and increase patient care costs.
First, clinicians can assess whether PFO is the main cause of CS and decide whether to perform PFO interventional closure. Although risk scores such as RoPE and PASCAL may help identify patients most likely to benefit, they should not be used as absolute gatekeeping criteria that deprive patients of interventional closure when a causal link between PFO and the index event is plausible. Indeed, recent evidence suggests that PFO interventional closure is safe, with a very low procedural risk36, and may be considered in selected patients in whom the PFO may have contributed to the event, as well as in individuals with demonstrably high-risk PFO anatomy after individualized evaluation. The failure rate of PFO interventional closure is relatively low. The success rate of most procedures is above 90%. The main reasons for interventional failure include occluder-related complications and individual differences among patients. Especially for elderly patients or those with severe conditions, post-procedural care should be strengthened. The latest European guidelines36 also suggest employing both the RoPE score and the PASCAL classification system simultaneously; this strategy helps more accurately evaluate the causal association between PFO and CS and identify the most appropriate patients for PFO closure. Regarding minimally invasive alternatives, the NobleStitch suture-mediated percutaneous closure system has been available for over 15 years. While it eliminates the need for a permanent metallic implant, clinical experience has revealed limitations, including a relatively high rate of residual right-to-left shunt, procedural complications such as septal tear, and long-term outcomes that remain inferior to those of established device-based closure systems37. Its role in clinical practice, therefore, remains limited and requires further evaluation before broader adoption can be recommended.
Regarding patients with PFO-associated CS, the optimal pharmacotherapy remains a matter of debate. The main options are antiplatelet and anticoagulant drugs. The 2018 guidelines38 note that for patients under 60 years old whose CS is considered related to PFO, anticoagulation therapy is weakly recommended compared to using antiplatelet therapy alone. In the 2019 expert consensus39, when patients have contraindications to or refuse interventional closure and do not have high bleeding risk, vitamin K antagonists should be considered for anticoagulation. In a 2019 European position paper on PFO patient management, Pristipino et al.40 highlighted a trade-off: oral anticoagulants outperform antiplatelet drugs in stroke prevention but significantly increase the likelihood of major bleeding. However, these conclusions are based on evidence with extremely high uncertainty.
In Shariat's study41, an 18-month follow-up was conducted on patients who were randomly allocated to receive either aspirin or warfarin, and ultimately found no difference in recurrence rate of ischemic events, mortality, and adverse reactions between PFO patients taking aspirin and warfarin. In the 2020 AAN42, it was also mentioned that for patients with CS accompanied by PFO, anticoagulation therapy may demonstrate comparable efficacy to antiplatelet therapy in preventing stroke recurrence. Additionally, recent comparative studies between new oral anticoagulants (NOACs) and antiplatelet agents indicate that for patients with cryptogenic embolic stroke and concurrent PFO, the rate of stroke recurrence is similar whether they are treated with dabigatran or aspirin43. European guidelines recommend that for patients without other obvious causes of stroke besides PFO, the comparative efficacy and safety of anticoagulation versus antiplatelet therapy remain inconclusive.
Although current guidelines and some studies suggest that anticoagulation therapy and antiplatelet therapy do not significantly differ in cardiovascular disease incidence among CS patients with PFO, from a pathophysiological perspective, the presumed embolic source is often venous thrombosis rather than platelet-rich arterial thrombosis. Anticoagulation therapy may therefore have a stronger mechanistic rationale than antiplatelet therapy. Recent studies have shown that for CS patients, the oral warfarin anticoagulation group had lower mortality rates and recurrence rates of cerebral infarction than those taking aspirin orally. However, using anticoagulants also increases the risk of bleeding. Chen et al.'s study44 included 375 CS patients with PFO who were randomly assigned to dabigatran or aspirin. After two years of follow-up, it was found that dabigatran was more effective in preventing CS combined with PFO without increasing the risk of bleeding complications. There are also reports indicating that NOAC treatment for CS may be as effective as warfarin and may decrease the occurrence of intracranial hemorrhage. With the increased use of NOACs such as rivaroxaban, apixaban, and dabigatran in clinical practice, which have helped reduce bleeding risk, the role of anticoagulation therapy has received increasing attention.
Post-2022 evidence has further examined the role of NOACs in specific PFO subgroups. The ATTICUS randomized trial was designed to evaluate apixaban versus aspirin in patients with embolic stroke of undetermined source and additional cardioembolic risk factors45. Further studies are needed to determine whether patients with high-risk PFO anatomy, such as a large shunt or concurrent atrial septal aneurysm, derive greater benefit from NOAC therapy than those with low-risk anatomical features. Nonetheless, the optimal NOAC agent and duration of treatment for PFO-associated CS remain to be determined in prospective randomized trials specifically designed for this subgroup. For safety outcomes, antiplatelet drugs are generally associated with the lowest bleeding risk, while interventional closure and anticoagulants carry different procedure-related and bleeding-related risks. Therefore, when preventing CS caused by PFO, the relatively high risk of bleeding may limit the use of anticoagulant drugs.
Since individual differences, genetic predispositions, and comorbidities can all affect coagulation balance, clinicians should comprehensively assess patients' thrombotic and bleeding risks and help formulate safe and effective anticoagulation plans. This includes adjusting doses and enhancing follow-ups to avoid bleeding events as much as possible46. Follow-up content should cover multiple aspects, including patient treatment progress, dynamic assessment of thrombotic and bleeding risks, monitoring of adverse drug reactions, medication compliance, and key indicators such as coagulation tests and liver and kidney function. If bleeding events still occur, symptomatic treatment should be administered based on the severity of bleeding, with comprehensive, refined, and dynamic treatment guidance to ensure the safety and effectiveness of anticoagulation therapy.
Additionally, Ghannam et al.47 analyzed seven randomized controlled trials involving 14,804 patients, with 7,406 receiving anticoagulation therapy and 7,398 receiving antiplatelet therapy. For the secondary prevention of stroke in Embolic Strokes of Undetermined Source (ESUS) patients, empirical anticoagulation did not outperform antiplatelet therapy regarding the rates of recurrent ischemic stroke, major hemorrhage, or mortality, thus emphasizing the necessity for personalized therapeutic strategies. Furthermore, the risk of ischemic stroke recurrence was comparable between the two pharmacological regimens across all subgroups stratified by age (< 75 vs. ≥ 75), sex, LAE, and atrial cardiomyopathy. However, a distinct benefit was observed in patients presenting with both ESUS and PFO; in this medically treated cohort, anticoagulation conferred a greater advantage over antiplatelet agents in reducing ischemic stroke.
Based on the above RCT and meta-analysis results, for PFO patients who have experienced CS, given the high recurrence rate of CS, they should be actively evaluated for interventional closure or anticoagulant and antiplatelet drugs for secondary prevention. With the increasing use and research on NOACs, bleeding events during anticoagulation can be better avoided and managed, and anticoagulation therapy remains a potential option for selected PFO patients. This review recommends individualized treatment choices for stroke patients with PFO. Recent meta-analyses related to PFO treatment are summarized in Table 2.
Access restricted. Please log in or start a trial to view this content.
In patients with PFO, the assessment of CS risk requires integrating anatomical characteristics, imaging features, and the RoPE scale, and interventional closure or pharmacologic therapy should be initiated promptly. Patients who have experienced CS need to consider the possibility of paradoxical embolism caused by PFO. After multidisciplinary evaluation, patients who meet the criteria should be prioritized for interventional closure, while others may be considered for pharmacologic therapy and regular follow-up. Althoug...
Access restricted. Please log in or start a trial to view this content.
The authors have no conflicts of interest to declare.
The authors thank the neurology and cardiology medical teams of The Fifth Affiliated Hospital of Dali University, Baoshan People's Hospital, for their assistance during manuscript preparation. This work was supported by the Special Research Project of the Center for Capacity Building and Continuing Education, National Health Commission (Grant No. GWJJZX20251005014).
Access restricted. Please log in or start a trial to view this content.
Access restricted. Please log in or start a trial to view this content.