INTRODUCTION — Options for prevention of recurrent stroke in patients with an embolic ischemic stroke that is attributed to a patent foramen ovale (PFO) include medical therapy with antithrombotic agents and closure of the defect by percutaneous device or rarely using a surgical approach.

This topic will review the medical and interventional options for secondary stroke prevention in patients with an embolic stroke presumed to be associated with a PFO. Cryptogenic stroke is reviewed in detail separately. (See "Cryptogenic stroke".)

The risk of stroke related to atrial septal abnormalities and indications for treating atrial septal defects in adults are discussed elsewhere. (See "Atrial septal abnormalities (PFO, ASD, and ASA) and risk of cerebral emboli in adults" and "Indications for closure and medical management of atrial septal defects in adults".)

Despite a possible association of migraine with right-to-left cardiac shunts, PFO closure is not an effective treatment for migraine. This is reviewed elsewhere. (See "Preventive treatment of episodic migraine in adults", section on 'Other interventions not recommended'.)

APPROACH TO EVALUATION AND TREATMENT — Closure of a PFO may prevent paradoxical embolism and thereby reduce the risk of recurrent stroke. While some results are conflicting, evidence from randomized controlled trials on balance suggests that PFO closure is effective in reducing the risk of recurrent stroke for select patients with PFO-associated stroke. (See 'Patient selection for PFO closure' below and 'Benefit' below.)

Patient selection for PFO closure — Mounting evidence suggests that percutaneous PFO closure is more effective for preventing recurrent ischemic stroke than antiplatelet therapy alone for highly selected patients who have an apparently embolic (seemingly cryptogenic) ischemic stroke and a PFO [1-4]. (See 'Percutaneous closure of PFO' below.)

●PFO associated stroke – Most patients with PFO with all of the following characteristics are potential candidates for percutaneous PFO closure:

•Age ≤60 years

•Embolic stroke topography

•No other evident source of stroke despite a comprehensive evaluation (see 'Exclusion of other sources of ischemic stroke' below)

•A possible, probable, or definite likelihood that the stroke was causally related to the PFO according to the PFO-associated stroke causal likelihood (PASCAL) classification system (table 1), which incorporates the Risk of Paradoxical Embolism (RoPE) score and high-risk features of the PFO on echocardiogram (table 2) (calculator 1) (see 'Is PFO the most likely stroke mechanism?' below)

•No concurrent indication for anticoagulation (see 'Special considerations in anticoagulated patients' below)

For most patients without a concurrent indication for anticoagulation who are ≤60 years of age with a possible, probable, or definite likelihood by PASCAL that the stroke was causally associated with the PFO, we suggest percutaneous PFO device closure in addition to antiplatelet therapy. (See 'Percutaneous closure of PFO' below.)

For patients age >60 years with a PASCAL determination of possible, probable, or definite causal association, the benefit of PFO device closure is unknown since these patients were excluded from the randomized trials of PFO device closure (see 'Benefit' below). Furthermore, the patient-level meta-analysis demonstrated that risk of atrial fibrillation after PFO closure increases with age [5]. Thus, decisions about PFO closure for such patients should be based upon individual patient characteristics.

•Concurrent indication for anticoagulation – If otherwise indicated, PFO device closure may be temporarily deferred for patients with a concurrent indication for short-term anticoagulation (≤1 year); the benefit of PFO device closure is uncertain for patients with a concurrent indication for long-term anticoagulation, as discussed below. (See 'Special considerations in anticoagulated patients' below.)

•Concurrent indication for cardiac surgery – For rare patients aged ≤60 years who meet criteria for PFO device closure but have a concurrent indication for cardiac surgery (eg, for valve surgery), surgical closure of PFO for secondary stroke prevention is appropriate. (See 'Surgical closure of PFO' below.)

•Percutaneous PFO closure not feasible – For rare patients who have a PFO that is not amenable to percutaneous closure but meet PASCAL criteria for PFO causal association, the benefit of surgical closure is uncertain. In the absence of a concurrent indication for cardiac surgery, we suggest against surgical closure.

●PFO unlikely to be associated with stroke – For patients who are unlikely to have a PFO-associated stroke by PASCAL classification (table 1), we suggest against percutaneous PFO device closure. (See 'Benefit' below.)

Note that the evidence from a 2021 meta-analysis [5] supporting the use of the PASCAL classification was published after the 2018 European position paper on the management of patients with PFO [6], the 2019 Clinical Commissioning Policy for PFO closure from the National Health Service in England [7], the 2020 American Academy of Neurology practice advisory [8], and the 2021 stroke prevention guideline from the American Heart Association/American Stroke Association [9]. Therefore, recommendations from these guidelines did not incorporate the PASCAL classification.

Exclusion of other sources of ischemic stroke — Patients considered for PFO closure should have a comprehensive evaluation by both a stroke neurologist and a cardiologist to ensure that other causes of ischemic stroke are excluded and that PFO-associated stroke (ie, paradoxical embolism through a PFO) is the most likely mechanism [10]. (See "Cryptogenic stroke", section on 'Evaluation and diagnosis'.)

By the TOAST classification (table 3), which is the most common system used in clinical practice to classify stroke subtypes, cryptogenic stroke (or stroke of undetermined origin in TOAST terminology) was defined as brain infarction that is not attributable to a source of definite cardioembolism, large artery atherosclerosis, small artery, or other determined etiology disease despite extensive vascular, cardiac, and serologic evaluation. In this classification, PFO was not recognized as a causative mechanism of ischemic stroke. (See "Cryptogenic stroke", section on 'Classification'.)

However, it has been proposed that patients with an embolic stroke who have a medium- or high-risk PFO and who have no other identified stroke etiology except for a PFO should be recognized as having a PFO-associated stroke [11]. For the purposes of this topic review, we define a PFO-associated embolic infarct and no other evident source of stroke as an ischemic stroke that has been thoroughly evaluated and is characterized by the following features [12]:

●No large vessel stenosis (≥50 percent) or occlusion in the territory of the infarct

●No evidence of occult atrial fibrillation and no other high-risk cardioembolic source (table 4)

●No radiographic acute lacunar infarction (ie, a small [≤1.5 cm] deep perforator infarct) and no clinical lacunar stroke syndrome (ie, hemiparesis/plegia, hemianesthesia without cortical signs) if imaging shows no infarct (see "Lacunar infarcts", section on 'Clinical features')

The evaluation should include neuroimaging of the brain to exclude lacunar infarction or a nonischemic brain lesion as the cause of the symptoms, and intracranial and extracranial neurovascular imaging to exclude stroke caused by large artery atherosclerosis, dissection, or other vasculopathy. There should be no evidence of atrial fibrillation on a 12-lead electrocardiogram (ECG) and on 24-hour cardiac monitoring. Ambulatory cardiac monitoring for several weeks (eg, 30 days) is warranted for adult patients over age 40 years with a cryptogenic ischemic stroke or cryptogenic transient ischemic attack (TIA) if no atrial fibrillation is detected by ECG and 24-hour monitoring [13].

Hematologic testing to exclude hypercoagulable states (eg, antiphospholipid syndrome and hyperhomocysteinemia) is indicated for patients with stroke being considered for PFO closure.

Given the high prevalence of PFO in the general population and the low risk of stroke related to PFO, there is always some degree of uncertainty about the causal relationship between PFO and an embolic-appearing ischemic stroke with no other evident stroke mechanism despite a comprehensive evaluation [14]. The possibility that the PFO is an "innocent bystander" and that another mechanism is responsible for the stroke is particularly applicable to older patients and to those with known risk factors for stroke (eg, hypertension, hypercholesterolemia, smoking) [15-17]. Causality can best be inferred in younger patients with no other apparent etiology for stroke [18], particularly if DVT is present (as a potential source for paradoxical emboli).

Evaluation for PFO — Transthoracic echocardiography (TTE), transesophageal echocardiography (TEE), and transcranial Doppler (TCD), in conjunction with agitated saline contrast (a "bubble study"), can all detect a right-to-left shunt associated with a PFO. Among these methods, only TEE allows visualization of the site and size of the shunt (eg, PFO, atrial septal defect). (See "Cryptogenic stroke", section on 'Cardiac and aortic evaluation'.)

●TTE or TEE is performed with intravenous injection of agitated saline contrast at rest, with Valsalva, and with cough; the study is considered positive if microbubbles (typically three or more) appear in the left heart within three cardiac cycles of bubbles filling the right atrium [19]. Multiple agitated saline contrast injections may be required to identify (or exclude) a shunt via the PFO. In some cases, intermittent flow through the PFO is also visualized by color Doppler. In some circumstances, the presence of an intracardiac shunt is best detected before administering sedation (typically used during TEE) as the patient may not be able to perform an adequate Valsalva following sedation.

●TCD, insonating the middle cerebral artery through the temporal bone window, is performed with intravenous injection of agitated saline contrast at rest, with Valsalva, and with cough; the study is considered positive for a right-to-left shunt if bubbles are detected in the middle cerebral artery.

TTE is often used as the initial study because it is better tolerated than TEE and is more widely available than TEE or TCD. However, TTE is less sensitive compared with TEE for the detection of PFO and does not adequately demonstrate its anatomic features. Therefore, patients who are candidates for PFO closure should have a TEE either as the initial study or as a follow-up study to TTE or TCD, even when a preceding TTE is negative or nondiagnostic for PFO. In general, candidates for PFO closure are selected patients who are ≤60 years of age with an embolic-appearing stroke or a TIA who have no evident source of stroke or TIA other than a PFO, despite a comprehensive evaluation. For patients who are not candidates for PFO closure, a follow-up TEE is generally not necessary when TTE or TCD is negative for a right-to-left shunt.

All patients who are candidates for PFO closure should have preprocedural imaging with TEE, including those with features suggesting PFO detected by TTE or TCD. (See 'Preprocedural imaging' below.)

The diagnostic evaluation of PFO is discussed in detail separately. (See "Patent foramen ovale", section on 'Diagnosis'.)

Is PFO the most likely stroke mechanism? — For patients with an embolic stroke without an evident cause, we recommend using the PASCAL system (table 1) along with the RoPE score (table 2) (calculator 1) to classify the causal association of PFO with the stroke and to guide decision-making for PFO device closure. In the setting of a PFO-associated embolic infarct and no other evident source of stroke despite a comprehensive evaluation in a patient ≤60 years of age, it is reasonable to conclude that paradoxical embolism through a PFO is the most likely stroke mechanism, and that PFO closure is warranted. (See 'PASCAL classification' below and 'RoPE score' below.)

Aspects of the evaluation particularly relevant to establishing the likelihood of a PFO-associated stroke include the history, presence of embolic stroke, evaluation for venous thromboembolism, and an assessment of PFO stroke risk, as reviewed in the sections that follow.

History — The history should include specific questions that define the circumstances immediately preceding the event. Was the patient doing something that might increase right-to-left shunt flow through a PFO, such as straining, coughing vigorously, or lifting or pushing a heavy object? Was there any risk for deep venous thrombosis such as prolonged immobility (eg, postoperative status, sitting in a cramped airline seat with legs dependent and knees flexed), dehydration, or hypercoagulability?

Is embolic stroke present? — Embolism is especially likely when neuroimaging reveals infarcts in multiple vascular territories or a single wedge-shaped infarct involving cortex and the underlying subcortical white matter. With embolic stroke, the neurologic deficit is typically maximal from the onset, with a tendency to improve quickly. (See "Overview of the evaluation of stroke", section on 'Determining a presumptive diagnosis of stroke subtype'.)

Whether or not an embolic stroke is related to PFO or to another mechanism cannot be determined reliably based upon neuroimaging features alone. However, certain findings may be more suggestive of PFO. In an analysis of data from subjects with cryptogenic stroke and PFO (n = 1141) or no PFO (n = 1539) in the Risk of Paradoxical Embolism (RoPE) study database, characteristics associated with a significantly higher prevalence of PFO were a large index stroke (odds ratio [OR] 1.36), index stroke seen on imaging (OR 1.53), and superficial (ie, involving the cerebral or cerebellar cortex) stroke location (OR 1.54) [20].

Evaluation for venous thromboembolism — We suggest performing an evaluation for DVT in all patients with PFO-associated embolic infarct and no other evident source of stroke as identification of DVT or other evidence of venous thromboembolism can help strengthen the clinical inference of paradoxical embolism and also has important implications for therapy (including identifying an indication for anticoagulation, which impacts the timing and need for PFO device closure). Ideally, the evaluation should be done within two to three days of stroke onset, before venous thrombosis develops secondary to stroke-related immobility [11].

Studies using radiograph venography or magnetic resonance venography to examine patients with ischemic cerebral events or other suspected embolic events have found variable rates of proximal leg or pelvic deep venous thrombosis (10 to 22 percent) [21-23]. Among those with documented deep vein thrombus, most had no symptoms or signs of venous thrombosis. There are several potential explanations for the relatively low rates of proximal deep vein thrombosis in patients with PFO-associated embolic infarct and no other evident source of stroke:

●Evaluation for deep vein thrombosis was incomplete in these studies since venography evaluated either the lower extremities or pelvic veins, but not both.

●The emboli consist of platelet fibrin particles that normally circulate in the systemic venous bed and are too small to be seen with conventional testing. These particles are removed by the efficient lytic system of the lungs; however, when shunted across a PFO or ASD, they may travel unlysed into the cerebral circulation [24].

●The clot forms in the heart on a device lead, at the edges or in the tunnel of the PFO, or in an ASA.

●The embolic source is in the systemic arterial circulation. (See "Atrial septal abnormalities (PFO, ASD, and ASA) and risk of cerebral emboli in adults", section on 'Left-sided sources'.)

●Deep vein thrombi have resolved (lysed, completely embolized, or recanalized) during the delay between onset of the embolic event and venography.

●Proximal deep vein thrombi may be lysed in some patients treated with intravenous thrombolytic therapy [23].

PFO risk assignment — Among patients with an embolic stroke topography, a PFO, and no alternative cause, features associated with medium- to high-risk PFO include factors that increase right-to-left shunt flow (eg, large PFO size, chronic right atrial hypertension, or a Valsalva maneuver at onset of stroke) and concomitant pulmonary embolism or deep venous thrombosis preceding the ischemic stroke [11]. Individual patient characteristics (age and vascular risk factors) are also important in the assessment of PFO risk. The RoPE score (table 2) and the PASCAL classification (table 1) are useful tools to estimate the likelihood of a causal relationship between PFO and embolic stroke for individual patients, as discussed below. (See 'RoPE score' below and 'PASCAL classification' below and 'Benefit' below.)

PFO size is generally inferred from the degree of right-to-left shunting apparent on echocardiography with agitated saline contrast, based on the number of microbubbles appearing in a single frame in the left atrium either spontaneously or after a provocative (Valsalva) maneuver within three cardiac cycles after opacification of the right atrium [25,26]. Although there is some variability among trials, the shunt size can be graded as follows [4]:

●Grade 0, no microbubbles

●Grade 1 (small), 1 to 5 microbubbles

●Grade 2 (moderate), 6 to 25 microbubbles

●Grade 3 (large), more than 25 microbubbles

In one small study, PFO measured ante mortem by assessment of jet size on color Doppler TEE and number of microbubbles on contrast TEE correlated well with autopsy finding; patients with a PFO >10 mm at autopsy had large shunts with >25 microbubbles on contrast TEE [27].

Thrombus trapped in a PFO is rarely encountered [28-31], but this finding supports a very high risk-PFO that may warrant surgical treatment [11].

RoPE score — The Risk of Paradoxical Embolism (RoPE) score, as shown in the table (table 2) and calculator (calculator 1), estimates the probability that a PFO is incidental or pathogenic in a patient with a seemingly cryptogenic stroke [32]. The PFO-attributable fraction of stroke derived from the RoPE score (table 5) varies widely and decreases with age and the presence of vascular risk factors. High RoPE scores, as found in younger patients who lack vascular risk factors and have a cortical infarct on neuroimaging, suggest pathogenic, higher risk PFOs. In contrast, low RoPE scores, as found in older patients with vascular risk factors, suggest incidental, lower-risk PFOs. The RoPE score is a major component of the PASCAL classification system, which provides additional discrimination, as described below.

PASCAL classification — We recommend using the PASCAL classification system to guide informed decision-making for PFO device closure. The PASCAL classification system estimates the probability that stroke is associated with a PFO in patients with embolic infarct topography and without other major sources of ischemic stroke [11]. The classification is based upon the RoPE score combined with anatomic and clinical factors (shunt size, presence or absence of atrial septal aneurysm and/or venous thromboembolism) and categorizes the likelihood that the stroke is caused by a PFO as unlikely, possible, probable, highly probable, or definite, as shown in the table (table 1).

Exclusions to device closure — Exclusions to percutaneous device closure include the presence of an inferior vena cava filter, elevated bleeding risk or coagulopathy, and vascular, cardiac, or PFO anatomy that is unsuitable for device placement.

Special considerations in anticoagulated patients — Patients with venous thromboembolism that is provoked by a known event or an identifiable transient risk factor are generally treated with anticoagulation for 3 to 12 months. In such cases, PFO device closure, if otherwise indicated, can be postponed until anticoagulation is stopped. For patients assigned to PFO closure in the RESPECT trial, the incidence of venous thromboembolism more than 30 days postprocedure was higher among those with a history of overt deep vein thrombosis than for those without such a history [1].

For patients with an embolic-appearing stroke who have an indication for chronic anticoagulation (eg, unprovoked or recurrent deep venous thrombosis), the benefit of PFO closure is uncertain. For such patients, we suggest individualized, multidisciplinary, shared decision-making that accounts for the risks of thrombosis, embolism, and intervention in determining whether to proceed with chronic anticoagulation alone or to also perform percutaneous PFO closure.

Informed decision-making — Consideration of PFO closure, including benefits, risks, and alternative treatment options must be discussed with the patient by the neurologist and cardiologist. The patient should understand the immediate and long-term potential benefits and risks of treatment options (including decreased risk of recurrent stroke and increased risk of atrial fibrillation with PFO percutaneous device closure) in order to make an appropriately informed decision that accounts for their own values and preferences.

PERCUTANEOUS CLOSURE OF PFO

Benefit — Mounting evidence suggests that PFO percutaneous device closure is more effective than antiplatelet therapy alone for reducing the risk of recurrent stroke in select patients aged ≤60 years with an embolic-appearing ischemic stroke who have a PFO with a right-to-left interatrial shunt and who have no other identified stroke cause or mechanism. The risk of recurrent stroke with device closure is reduced by approximately 60 percent compared with medical therapy (eg, from about 5 percent to 2 percent during a three- to six-year period) [33]; the corresponding number needed to treat to prevent one recurrent stroke during this period is approximately 30. The patients most likely to benefit may be those with a large right-to-left interatrial shunt and/or an associated atrial septal aneurysm (ASA), characteristics that suggest an increased risk for paradoxical embolism.

Randomized controlled trials of percutaneous PFO closure have all found point estimates suggesting that PFO closure is more effective than medical therapy for reducing recurrent stroke rates. These results were not statistically significant by intention-to-treat analyses in the first three trials (CLOSURE I [34], PC [35], and RESPECT [36]), but were significant in later trials (RESPECT extended follow-up [1], REDUCE [4], and CLOSE [2]). The trials that found clear benefit for PFO device closure were likely positive because of several factors. First, these latter trials enrolled subjects when off-label PFO closure had waned to some extent, and thus it is possible that patients who were more likely to benefit were included in these studies. Furthermore, to varying degrees for the individual studies, they included a requirement for neuroimaging confirmation of stroke prior to enrollment, excluded lacunar infarcts, provided longer follow-up, and selected patients with PFO features (ie, large shunt size or presence of an associated ASA) that may portend an increased risk of paradoxical embolism. Meta-analyses of closure trials consistently show that patients in the PFO closure groups had increased rates of newly detected atrial fibrillation compared with the medical therapy groups [33,37]. (See 'Adverse effects' below.)

●Meta-analyses – In a 2018 meta-analysis that included four trials (PC [35], RESPECT extended follow-up [1], REDUCE [4], and CLOSE [2]) with 2531 subjects and follow-up ranging from 3.2 to 5.9 years, PFO closure reduced the risk of recurrent stroke from 5.1 percent with medical therapy to 1.8 percent (absolute risk reduction [ARR] 3.3 percent, 95% CI 6.2-0.4 percent) [33,38]. Based on these data, the number needed to treat (NNT) with PFO device closure to prevent one recurrent stroke was approximately 30. Similarly, a separate 2018 meta-analysis that included the same trials, with follow-up ranging from 2.1 to 5.3 years, found that PFO closure reduced the risk of recurrent stroke from 4.1 percent with medical therapy to 1.2 percent (ARR 3.1 percent, 95% CI 5.1-1.0 percent) [37]. Both meta-analyses excluded the CLOSURE I trial because it used the STARFlex PFO closure device, which was associated with higher complication and lower procedural success rates than the PFO closure devices used in the other trials, and is no longer available [33,37]. Other meta-analyses of PFO closure have generally reported similar findings [39-47].

A 2021 meta-analysis of individual patient data (n = 3740) from six randomized controlled trials, with a median follow-up of 57 months, found that PFO closure reduced the annualized incidence of stroke compared with medical therapy alone (0.47 versus 1.09 percent, adjusted hazard ratio [HR] 0.41, 95% CI 0.28-0.60) [5]. Importantly, the risk reduction for recurrent stroke with PFO closure varied among subgroups with different probabilities that the stroke was causally related to the PFO, as determined by the Risk of Paradoxical Embolism (RoPE) score (table 2) and a modified PFO-associated stroke causal likelihood (PASCAL) classification (table 1). For the risk of recurrent ischemic stroke with PFO closure compared with medical therapy alone, patients with high RoPE score (ie, higher risk PFO) had a HR of 0.21 (95% CI 0.11-0.42), while those with a low RoPE score (ie, lower risk PFO) had a HR of 0.61 (95% CI 0.37-1.00). However, patients categorized by the PASCAL classification as probable, possible, or unlikely to have a PFO-associated stroke had HRs of 0.10 (95% CI 0.03-0.35), 0.38 (95% CI 0.22-0.65), and 1.14 (95% CI 0.53-2.46), respectively. Thus, the PASCAL score provides greater discrimination of who is likely to benefit from closure and by what degree, compared with the RoPE score alone. (See 'RoPE score' above and 'PASCAL classification' above.)

●Individual trials – The individual trials used the broad term "cryptogenic" rather than "PFO-associated" stroke but aimed to enroll the latter, and reported the following results:

•In the CLOSURE I trial, 909 adult patients ≤60 years old with a PFO and cryptogenic stroke or transient ischemic attack (TIA) were randomly assigned either to PFO device closure (n = 447) or to medical therapy (n = 462) [34]. Patients in the device group were treated with the STARFlex PFO closure device and received aspirin plus clopidogrel for six months followed by aspirin alone; those in the medical therapy group were treated with aspirin or warfarin or both. The primary endpoint was a composite of stroke or TIA at two years plus 30-day mortality and neurologic mortality beyond 30 days. At two years, by intention-to-treat analysis, there was no significant difference between device closure and medical therapy in the rates of the primary endpoint (5.5 versus 6.8 percent, hazard ratio [HR] 0.78, 95% CI 0.45-1.35), stroke (2.9 versus 3.1 percent), or TIA (3.1 versus 4.1 percent). Major vascular complications were significantly more frequent with device closure (3.2 versus 0 percent), as was atrial fibrillation (5.7 versus 0.7 percent), most of which was periprocedural.

•The PC trial randomly assigned 414 adults (<60 years of age) with PFO and ischemic stroke, TIA, or a peripheral embolic event to treatment with the Amplatzer PFO Occluder or medical therapy [35]. After a mean follow-up of four years, the composite primary endpoint of death, nonfatal stroke, TIA, or peripheral embolism for the intention-to-treat cohort occurred in 7 of 204 patients (3.4 percent) in the device closure group and 11 of 210 patients (5.2 percent) in the medical therapy group; the difference was not statistically significant (HR 0.63, 95% CI 0.24-1.62). Serious adverse events were slightly more frequent in the device closure group (21.1 percent versus 17.6 percent), including a nonsignificantly higher rate of new-onset atrial fibrillation in the device closure group (2.9 versus 1.0 percent).

•In the RESPECT trial, 980 patients (age 18 to 60 years) with a PFO and cryptogenic ischemic stroke were randomly assigned to receive treatment with the Amplatzer PFO Occluder or medical therapy [36]. The primary endpoint was a composite of recurrent nonfatal ischemic stroke, fatal ischemic stroke, or early death after randomization. The trial results were analyzed after reaching the target of 25 primary endpoint events; all 25 events were nonfatal ischemic strokes. The mean follow-up was approximately 2.6 years, and the primary endpoint for the intention-to-treat cohort occurred in 9 of 499 patients (1.8 percent) in the closure group and 16 of 481 patients (3.3 percent) in the medical therapy group, a difference that was not statistically significant (0.66 versus 1.38 events per 100 patient-years, HR 0.49, 95% CI 0.22-1.11).

A later RESPECT publication reported outcomes at median follow-up of 5.9 years [1]. By intention-to-treat analysis, recurrent ischemic stroke was less frequent in the closure group compared with the medical therapy group (18 versus 28 events, 0.58 versus 1.07 events per 100 patient-years, HR 0.55, 95% CI 0.31-0.99). However, the dropout rate was higher and treatment exposure lower in the medical therapy group, leading to an unequal exposure to the risk of outcome events among the two groups.

•The REDUCE trial randomly assigned 664 patients 18 to 59 years of age with cryptogenic embolic-appearing ischemic stroke and PFO with a right-to-left shunt demonstrated by means of transesophageal echocardiography [4,48]. Patients were randomly assigned to PFO closure combined with antiplatelet therapy or treatment with antiplatelet therapy alone in a 2:1 ratio. During a median follow-up of 5.0 years, clinical ischemic stroke by intention-to-treat analysis occurred in fewer patients in the PFO closure group compared with the antiplatelet-only group (8 of 441 versus 12 of 223 patients, 1.8 versus 5.4 percent, HR 0.31, 95% CI 0.13-0.76).

•The CLOSE trial enrolled patients 16 to 60 years of age with recent cryptogenic stroke attributed to PFO who had an associated ASA or large interatrial shunt on echocardiography [2]. Patients were randomly assigned in a 1:1:1 ratio to PFO closure plus antiplatelet therapy, antiplatelet therapy alone, or oral anticoagulation, with the exception that patients with contraindications to PFO device closure or to anticoagulation were assigned to alternative non-contraindicated treatment or to antiplatelet therapy. The main arms of the trial (n = 473) compared PFO closure with antiplatelet therapy; at a mean follow-up of 5.3 years, there were no recurrent strokes among 238 patients in the PFO closure group compared with 14 strokes among 233 patients the antiplatelet-only group (HR 0.03, 95% CI 0.0-0.26).

●Trial limitations – There are important limitations of these trials that lower confidence in the results or result in important questions about generalizability. As examples:

•All of these trials utilized open label endpoint ascertainment, which increases the risk of bias.

•The number of primary events was relatively low, with a total of 52 events in the CLOSURE I trial [34], 18 in the PC trial [35], 46 in the RESPECT trial [1], 18 clinical events in the REDUCE trial [4], and 14 events in the treatment groups comparing of PFO closure with antiplatelet treatment in the CLOSE trial [2].

•The duration of follow-up in the CLOSURE I trial (two years) and the primary analysis of the RESPECT trial (2.6 years) was not long enough to demonstrate benefit (eg, the trials of endarterectomy for asymptomatic carotid disease would not have demonstrated benefit at only two years).

•There was slow enrollment in most of these trials and suspicion that patients at high risk of recurrent embolism were disproportionately treated outside of the trials with PFO closure, particularly for the earlier trials.

•The trials primarily enrolled patients under 60 years of age, and the safety and efficacy of closure remains unclear in patients older than age 60, who are likely to be at increased risk of atrial fibrillation as a complication of PFO closure.

Adverse effects — New onset atrial fibrillation is the most common adverse effect of PFO device closure. Several meta-analyses have confirmed that PFO closure increases the risk of atrial fibrillation or atrial flutter [5,33,37,39-41,44,49], with a risk difference (ie, absolute risk increase) of 3.4 percent in one of the meta-analyses [37]. Of note, atrial fibrillation is a frequent complication of cardiac surgery, which is relevant for patients undergoing cardiac surgery with concomitant PFO closure. (See "Atrial fibrillation and flutter after cardiac surgery".)

One concern about new onset atrial fibrillation is its potential long-term impact on the risk of stroke and other embolic events, but limited information is available on the clinical course of postprocedural atrial fibrillation. The management of new-onset atrial fibrillation after PFO closure (whether by device or surgery) is similar to that after cardiac surgery generally as reviewed in detail separately. (See "Atrial fibrillation and flutter after cardiac surgery".)

Other complications, all rare, include hematoma at the puncture site, device migration, device embolization, device erosion, and device thrombosis with possible and recurrent ischemic stroke. (See 'Recurrent ischemic stroke' below.)

Device erosion is rare (0.2 to 0.3 percent of cases) after device closure of ASDs, in most cases occurring during the first six months after implantation [50,51], and may be rarer with device closure of PFOs. Device erosion can lead to cardiac perforation with pericardial effusion, cardiac tamponade, and fistula formation or may rarely create an atrial septal defect [52,53].

Preprocedural imaging — Patients with an embolic infarct and no other evident source of stroke who are being considered for PFO closure should undergo transesophageal echocardiography (TEE) to confirm that the intracardiac shunt is caused by a PFO, to define atrial septal anatomy (including thickness of rims around the PFO) and suitability for device closure, and to exclude other causes of embolic stroke (eg, intracardiac thrombus, mass or vegetation) or shunt [19]. The atrial septum is carefully examined to determine whether there are one or more concomitant ASDs and/or an atrial septal aneurysm (defined a redundant mobile interatrial septal tissue in the region of the fossa ovalis with phasic excursion of at least 10 to 15 mm). The length of the PFO tunnel is also assessed. If a PFO is accompanied by one or more secundum-type ASDs, the location and size of these defects are examined to determine whether all the defects can be closed percutaneously by one or two devices, and whether a surgical approach might be preferred.

Procedure — Percutaneous PFO closure should be performed using an approved PFO closure device. Access to the right atrium is established via the right femoral vein and the PFO is crossed with a guidewire or catheter under fluoroscopic and echocardiographic (either TEE or intracardiac) guidance [50]. After the left atrium is accessed, an exchange-length stiff guidewire is advanced into a pulmonary vein. Balloon sizing may be used to determine the size of the device (typically twice the size of the balloon-stretched diameter of the defect). After the balloon is withdrawn, the delivery system is advanced into the left atrium over the guidewire. The device and the delivery system are flushed prior to insertion and the catheters aspirated to avoid air embolism. The left-sided occluder is opened in the left-atrium and retracted against the interatrial septum before the right-sided occluder is opened. After device position is confirmed by echocardiography, the closure device is released from the delivery system. Echocardiography is performed after device release to assess for residual shunting and presence of any complications.

The echocardiographic guidance is achieved by intracardiac echocardiography (ICE) or TEE. ICE, performed via a second venous access to the right atrium, is generally preferred as it avoids the general anesthesia and intubation required for TEE [54]. When TEE is not used, the percutaneous closure procedure (with ICE) is generally performed with conscious sedation.

Antithrombotic therapy during the procedure — Patients undergoing percutaneous device closure routinely receive antithrombotic therapy prior to, during, and following the procedure, though specific regimens vary. As an example, in the CLOSE trial, all patients undergoing percutaneous PFO closure received clopidogrel 300 mg, low molecular weight heparin, or continuation of their prior antiplatelet therapy before the procedure [2]. During the procedure unfractionated heparin 100 international unit/kg (up to 10,000 international unit) was administered intravenously.

MEDICAL TREATMENT

General measures — Patients with PFO who have an ischemic stroke or transient ischemic attack (TIA) should be treated with all appropriate risk reduction strategies, most importantly antithrombotic therapy (see 'Antithrombotic therapy' below). Other measures include lifestyle modification (diet and exercise), blood pressure reduction, and statins (if indicated). (See "Overview of secondary prevention of ischemic stroke" and "Long-term antithrombotic therapy for the secondary prevention of ischemic stroke".)

Antithrombotic therapy — For patients who have undergone percutaneous or surgical PFO closure who have no concurrent indication for anticoagulation, we suggest long-term antiplatelet therapy rather than no treatment or anticoagulation. Patients are treated with antiplatelet therapy whether or not the PFO is successfully closed.

Following PFO closure, we treat with aspirin 75 to 81 mg/day plus clopidogrel 75 mg/day for three months, followed by continued aspirin therapy (75 to 81 mg/day). In the CLOSE trial, the antithrombotic regimen after percutaneous PFO closure was aspirin 75 mg/day plus clopidogrel 75 mg/day for three months [2]. From the fourth month, patients were treated with aspirin alone, clopidogrel alone, or the combination product aspirin-extended-release dipyridamole.

For most patients with a PFO-associated embolic infarct and no other evident source of stroke who do not undergo PFO closure, we suggest antithrombotic therapy with antiplatelet agents rather than anticoagulation. However, we suggest anticoagulation therapy rather than antiplatelet agents for patients with expected high risk of venous thromboembolism, as discussed below. Although the comparative effectiveness of different types of antithrombotic therapy for secondary stroke prevention among patients with a PFO-associated ischemic stroke or TIA is uncertain (see 'Comparative studies' below), high-quality data from randomized trials have established that aspirin is effective for ischemic stroke prevention, as discussed elsewhere. (See "Long-term antithrombotic therapy for the secondary prevention of ischemic stroke", section on 'Aspirin'.)

Continued anticoagulation is used for patients with a PFO-associated stroke who have a concurrent indication, such as acute deep venous thrombosis, pulmonary embolism, other venous thromboembolism (VTE), or a hypercoagulable state. Management of these conditions, including the duration of anticoagulation, is discussed separately. (See "Overview of the treatment of lower extremity deep vein thrombosis (DVT)" and "Treatment, prognosis, and follow-up of acute pulmonary embolism in adults" and "Placement of vena cava filters and their complications".)

Comparative studies — The evidence comparing the benefit of antiplatelet therapy with warfarin for stroke prevention among patients with a PFO and ischemic stroke comes mainly from nonrandomized studies or randomized trials with important limitations, as illustrated by the following observations:

●In a 2015 meta-analysis of individual participant data from 12 observational studies involving 2385 medically treated patients with cryptogenic stroke and PFO, there was no significant difference between treatment with oral anticoagulation compared with antiplatelet therapy for the composite outcome of recurrent stroke, TIA, or death (9 versus 10 percent, adjusted hazard ratio 0.76, 95% CI 0.52-1.12) and no difference for the outcome of recurrent stroke alone (4 versus 5 percent, adjusted hazard ratio 0.75, 95% CI 0.44-1.27) [55].

●In an updated meta-analysis of combined data from four trials (PICSS [56], CLOSE [2], NAVIGATE ESUS [57], and RESPECT ESUS [58]) that included patients with PFO and cryptogenic stroke who were randomly assigned to treatment with anticoagulant or antiplatelet therapy, the risk of recurrent ischemic stroke was similar for anticoagulation versus antiplatelet therapy (odds ratio 0.70, 95% CI 0.43-1.14) [58]. However, confidence in this result is limited by imprecision due to the small number of outcome events and wide confidence interval.

Additional preventive measures — Certain general measures may be beneficial independent of the therapy chosen for the atrial septal abnormality. Since embolic material originates most commonly in lower extremity veins, patients at risk should avoid sitting for extended periods of time with knees flexed and legs dependent or the legs crossed, and should avoid prolonged passive standing. Risk is implicit during long airplane flights. For long-distance travelers with individual risk factors for VTE, we suggest frequent ambulation and calf exercises, avoidance of dehydration or sedatives, and graduated compression stockings to reduce the risk of travel-associated VTE. These measures are particularly important in patients in whom deep vein thrombosis was identified at the time of the initial cerebrovascular event. Recommendations for prevention of venous thromboembolism are discussed in greater detail separately. (See "Prevention of venous thromboembolism in adult travelers".)

RECURRENT ISCHEMIC STROKE — As with any stroke, patients who have a recurrent ischemic stroke after PFO closure should have another comprehensive evaluation to determine the stroke mechanism, including assessment of the PFO closure device for defects, device thrombosis, and residual shunt. Recurrent ischemic stroke may occur in patients with a PFO, regardless of whether the PFO was closed, due to mechanisms unrelated to paradoxical embolism, such as cardiogenic embolism, large artery atherosclerosis, small artery disease, and other determined stroke etiologies. In a minority of patients with PFO closure, a residual shunt persists, allowing continued potential risk for paradoxical embolism [59-63]. Alternatively, thrombus may spontaneously form on or adjacent to the PFO device or in the left atrium due to stagnant blood flow [64], particularly given the possible increased risk of atrial arrhythmias (mainly atrial fibrillation) in patients with PFO and/or atrial septal aneurysm [65]. This risk may be augmented after PFO closure [34,35], especially in the first few weeks after device implantation.

Recurrent stroke should be treated according to the underlying mechanism, if it can be identified:

●If the recurrence occurs in a patient who has not had their PFO closed, and the PFO still appears to be the most likely cause of stroke, we suggest PFO closure.

●For patients on antiplatelet therapy who have a recurrent PFO-associated stroke (regardless of PFO closure status) and no atrial fibrillation on reevaluation with long-term cardiac monitoring, options include continuing the same antiplatelet agent or switching to another antiplatelet regimen. For patients with recurrent embolic stroke of undetermined source, switching to empiric anticoagulant therapy is also a reasonable option. These issues are discussed in detail elsewhere. (See "Cryptogenic stroke", section on 'Embolic stroke of undetermined source'.)

●In rare cases, recurrent thrombus formation on the PFO closure device despite anticoagulant therapy may require device removal [66].

SURGICAL CLOSURE OF PFO — For rare patients aged ≤60 years with a PFO-associated stroke and no other evident source of stroke despite a comprehensive evaluation who have a concurrent indication for cardiac surgery (eg, an indication for valve surgery rather than a transcatheter procedure), surgical closure of PFO via standard or minimally invasive (including robotic) techniques for secondary stroke prevention after PFO-associated stroke is appropriate.

The reported efficacy of surgical closure of a PFO in patients with prior cerebrovascular ischemic events has been variable [15,67-69], and randomized trials comparing surgical PFO closure with percutaneous closure or with medical therapy have not been performed.

Rates of recurrent cerebrovascular events following surgical closure have ranged from 7 to 14 percent at one to two years [15,67]. Similar to findings from the randomized controlled trials for device closure of PFO, these events are likely due to mechanisms unrelated to paradoxical embolization, as illustrated by a report of 91 patients (mean age 44 years) with one or more cerebrovascular ischemic events who underwent surgical PFO closure [67]. The overall freedom from an ischemic episode at one and four years was 93 and 83 percent, respectively. The recurrent events were transient ischemic attacks (there were no cerebral infarctions), one of which was attributed to giant cell arteritis. Transesophageal echocardiography showed that the closures were intact in all patients, implying that paradoxical embolization was not the cause of the ischemic events.

In patients with high cardiovascular risk and an incidentally discovered PFO, surgical closure may actually increase the risk of postoperative stroke. This conclusion comes from a retrospective study of over 13,000 adults without a prior diagnosis of PFO or atrial septal defect who had cardiothoracic surgery [70]. A PFO was detected intraoperatively in 2277 patients, and closure was performed at the discretion of the surgeon in 28 percent. Using propensity-matched analysis, the risk of perioperative stroke was significantly higher in patients who had surgical PFO closure than in those who did not (2.8 versus 1.2 percent; odds ratio 2.47, 95% CI 1.02-6.0). There was no difference between the two groups in long-term survival. The uncontrolled retrospective design and small number of events limit the strength of this study.

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Stroke in adults".)

INFORMATION FOR PATIENTS — UpToDate offers two types of patient education materials, "The Basics" and "Beyond the Basics." The Basics patient education pieces are written in plain language, at the 5^th to 6^th grade reading level, and they answer the four or five key questions a patient might have about a given condition. These articles are best for patients who want a general overview and who prefer short, easy-to-read materials. Beyond the Basics patient education pieces are longer, more sophisticated, and more detailed. These articles are written at the 10^th to 12^th grade reading level and are best for patients who want in-depth information and are comfortable with some medical jargon.

Here are the patient education articles that are relevant to this topic. We encourage you to print or e-mail these topics to your patients. (You can also locate patient education articles on a variety of subjects by searching on "patient info" and the keyword(s) of interest.)

●Basics topic (see "Patient education: Patent foramen ovale (The Basics)")

SUMMARY AND RECOMMENDATIONS

●Overview of treatment decisions – For patients with an ischemic stroke and a patent foramen ovale (PFO), treatment options include antithrombotic therapy and/or PFO closure, along with other stroke prevention therapies. (See 'Percutaneous closure of PFO' above and 'Medical treatment' above and 'Surgical closure of PFO' above.)

●Benefit of PFO closure – PFO device closure is more effective than medical therapy alone for select patients aged ≤60 years with a PFO-associated stroke (ie, a nonlacunar ischemic stroke in the setting of a PFO with a right-to-left interatrial shunt and no other source of stroke despite a comprehensive evaluation). The risk of recurrent stroke with device closure is reduced by approximately 60 percent compared with medical therapy, from about 5 percent to 2 percent, during a three- to six-year period. For patients age >60 years, the benefit of PFO device closure is unknown since these patients were predominantly excluded from the randomized trials of PFO device closure. (See 'Benefit' above.)

●Determining the likelihood that PFO is the cause of the stroke – A causal association between a PFO and stroke (ie, a PFO-associated stroke) is thought likely in patients with PFO who meet all of the following criteria (see 'Approach to evaluation and treatment' above):

•Under 60 years of age

•Embolic stroke topography (see 'Is embolic stroke present?' above)

•No other evident source of stroke despite a comprehensive evaluation (see 'Exclusion of other sources of ischemic stroke' above)

•A possible, probable, or definite likelihood that the stroke was causally related to the PFO according to the PFO-associated stroke causal likelihood (PASCAL) classification system (table 1), which is used with the Risk of Paradoxical Embolism (RoPE) score (table 2) (see 'PASCAL classification' above)

●Decision-making for PFO closure – A decision to close the PFO is based in part upon the patient's age and the likelihood that the PFO was causally associated with the stroke.

•PFO-associated stroke – For most patients without a concurrent indication for anticoagulation who are ≤60 years of age with a possible, probable, or definite likelihood by PASCAL (table 1) that the PFO was causally associated with the stroke, we suggest percutaneous PFO device closure in addition to antiplatelet therapy (Grade 2B). For patients who are >60 years of age, we use an individualized treatment approach to PFO closure based upon patient preferences and other risk factors and comorbidities.

-Concurrent indication for anticoagulation – If otherwise indicated, PFO device closure may be temporarily deferred for patients with an indication for short-term anticoagulation. The benefit of PFO device closure is uncertain for patients with an indication for long-term anticoagulation; an individualized approach with shared decision-making is appropriate. (See 'Special considerations in anticoagulated patients' above.)

-Concurrent indication for cardiac surgery – For rare patients who have a concurrent indication for cardiac surgery (eg, valve surgery) and meet PASCAL criteria for PFO causal association, surgical PFO closure is appropriate.

-Percutaneous PFO closure not feasible – For rare patients who have a PFO that is not amenable to percutaneous closure but meet PASCAL criteria for PFO causal association, the benefit of surgical closure is uncertain. In the absence of a concurrent indication for cardiac surgery, we suggest against surgical closure (Grade 2C).

•PFO unlikely to be associated with stroke – For patients who do not meet criteria for a causal association of PFO with stroke (ie, association unlikely by PASCAL (table 1)), we suggest against PFO closure (Grade 2C).

•Recurrent stroke on antiplatelet therapy – For patients with a previous PFO-associated stroke treated with antiplatelet therapy but not PFO closure who have a recurrent embolic-appearing stroke with no other identified cause, we suggest PFO closure in addition to antithrombotic therapy (Grade 2C). Switching antithrombotic therapy (another antiplatelet agent or anticoagulant therapy) is a reasonable alternative in such patients.

●Choice of antithrombotic agent – Recommendations for antithrombotic therapy vary based on the presence of and/or risk for venous thromboembolism.

•Following PFO closure – For patients who have undergone percutaneous or surgical PFO closure who have no concurrent indication for anticoagulation, we suggest long-term antiplatelet therapy rather than no treatment or anticoagulation (Grade 2C). Patients are treated with antiplatelet therapy whether or not the PFO is successfully closed. (See 'Antithrombotic therapy' above.)

•Without PFO closure – For patients with stroke causally associated to PFO who have not undergone PFO closure, treatment with antiplatelet therapy is recommended, which is consistent with the recommendations for cryptogenic stroke. (See 'Antithrombotic therapy' above and "Cryptogenic stroke", section on 'Secondary prevention'.)

•Concurrent indication for anticoagulation – For patients with stroke causally associated with PFO with concurrent indication for anticoagulation, standard recommendations for anticoagulation apply, whether or not PFO closure has been performed. In this setting, anticoagulation is generally the sole antithrombotic therapy required. Recommendations are provided in separate topic reviews. (See "Overview of the treatment of lower extremity deep vein thrombosis (DVT)" and "Selecting adult patients with lower extremity deep venous thrombosis and pulmonary embolism for indefinite anticoagulation".)

ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges Naser M Ammash, MD, and Robert S Schwartz, MD, and Joseph K Perloff, MD who contributed to earlier versions of this topic review.