Coronary Artery Disease in Adults Undergoing Percutaneous Patent Foramen Ovale Closure Following Cryptogenic Stroke

Background. A patent foramen ovale (PFO) is found in nearly half of patients with cryptogenic stroke. Little guidance on the use or necessity of coronary angiography at the time of percutaneous PFO closure exists. We aimed to characterize the presence of coronary artery disease (CAD) in patients undergoing PFO closure following a cryptogenic stroke. Methods. A retrospective analysis of consecutive patients who underwent concurrent percutaneous PFO closure and coronary angiography was performed. Patients were ≥40 years of age and had a preceding diagnosis of cryptogenic stroke with no known CAD. Visual analysis of coronary angiograms was performed. Results. Of 180 patients, 8 (4%) had severe CAD, 15 (8%) had moderate CAD, 32 (18%) had mild CAD, and 12 (7%) had luminal irregularities. Of those with moderate-to-severe CAD, 9 (5%) had proximal disease and 9 (5%) had multivessel disease. Of those with moderate-to-severe CAD, 8 (35%) underwent further risk stratification with stress testing, 4 (17%) were medically managed, and 1 (4%) underwent concurrent angioplasty. Conclusions. Coronary angiography identified a low prevalence of CAD in patients with cryptogenic stroke undergoing PFO closure, suggesting that coronary angiography is not routinely indicated in patients undergoing PFO closure.

J INVASIVE CARDIOL 2021 October 15 (Ahead of Issue).

Key words: PFO closure, stroke

Cryptogenic stroke, characterized as a cerebrovascular event in the absence of any identified etiology, represents approximately 30%-40% of all cerebral ischemic events.¹ Paradoxical embolism via a patent foramen ovale (PFO) has been cited as an important mechanism for cryptogenic strokes in adults aged 18-60 years.^2,3 A PFO is estimated to affect nearly half of patients with cryptogenic stroke.^4-6

Following the initial introduction of percutaneous PFO closure in 1992, its adoption for secondary stroke prevention following a cryptogenic stroke has since been supported by an extensive evidence base of randomized controlled trials (RCTs).^7-11 The peri- and postoperative management of patients undergoing PFO closure nonetheless remains nuanced. Having a patient’s coronary anatomy defined prior to a percutaneous procedure that carries a risk of emergent surgical intervention can be of value. Although rare, with a reported incidence of 0.7%-1.2%, PFO device embolization during closure remains a clinical problem that will grow alongside the increasing magnitude of new operators.¹² Post PFO closure, patients may experience chest pain, representing a diagnostic and management challenge for clinicians. In 2007, Wahl and colleagues identified incidental coronary artery disease (CAD) in 29% of patients undergoing percutaneous PFO closure with concurrent coronary angiography.¹³ The rationale for coronary angiography in patients with cryptogenic stroke undergoing PFO closure is multifactorial. Patients have already experienced a vascular neurologic event with a potential for atherosclerotic disease in the coronary vascular bed, may require urgent surgery following device implantation, and may experience chest pain after implantation. In addition, with PFO closure indicated in adults up to age 60, patients may have concurrent risk factors for cardiovascular disease.

There are currently no guideline indications for which patients undergoing PFO closure should undergo concurrent coronary angiography. In 2010, our institution adopted a policy of performing routine coronary angiography at the time of percutaneous PFO closure in all patients who were ≥40 years old following a cryptogenic stroke. We hypothesized that patients with CAD may require additional diagnostic and management considerations that could be guided by the status of their coronary arteries. The aim of our study was to determine the prevalence of CAD in patients undergoing transcatheter PFO closure following a cryptogenic stroke and explore the impact of CAD diagnosis on clinical management.

Study design and participants. A retrospective, single-center database of patients who underwent percutaneous PFO closure following cryptogenic stroke at the Peter Munk Cardiac Centre, Toronto General Hospital in Toronto, Ontario was developed. Consecutive patients who were ≥40 years old, diagnosed with a cryptogenic stroke, and underwent PFO closure from January 1, 2010 (marking the onset of an institutional policy to perform coronary angiography in all patients undergoing PFO closure aged ≥40 years) through December 31, 2017 were included. A diagnosis of cryptogenic stroke was established by a stroke neurologist and required neuroimaging evidence of a stroke on brain computed tomography (CT) or magnetic resonance imaging. Patients were excluded from the analysis if they underwent PFO closure for any other indication or had a documented history of established CAD (based on prior percutaneous coronary intervention or coronary artery bypass graft surgery). Evidence of a PFO was confirmed by transesophageal echocardiography with a bubble study prior to the procedure. Informed consent for the procedure was obtained.

PFO closure was performed with 1 of 4 device types: Amplatzer PFO occluder or the Amplatzer septal occluder (Abbott Vascular); the Gore Helex septal occluder (WL Gore Medical); or the CardioSeal/StarFlex occluder (NMT Medical). Procedures were performed under conscious sedation, usually with fluoroscopic guidance alone. Each patient underwent a coronary angiogram prior to device implantation. After implantation, right atrial angiography was performed to verify positioning. Intracardiac echocardiography was used as necessary.  Patients were routinely discharged the same day after the procedure. At discharge, all patients were treated with dual-antiplatelet therapy (acetylsalicylic acid and clopidogrel) for a minimum of 6 months. Follow-up consisted of an outpatient clinical appointment between 3-6 months accompanied by a transthoracic echocardiogram (TTE) with saline contrast.

A retrospective chart review was undertaken for all patients in the database using structured data extraction forms. Abstracted data from the clinical records included baseline demographics (age, gender, body mass index), a history of established CAD, and presence of cardiovascular risk factors (hypertension, diabetes, dyslipidemia, family history of early CAD, cigarette smoking), pre- and postprocedural symptoms (ie, chest pain, dyspnea, palpitations, headache/migraine, and syncope), along with imaging and procedural information. The University Health Network research ethics board approved this study.

Coronary angiography. Coronary angiograms were independently reviewed using syngo Dynamics imaging storage system (Siemens Healthineers) by 2 experienced interventional cardiologists (AM, RGR) who were blinded to patients’ medical histories and clinical outcomes. In patients with evidence of CAD, each lesion was characterized by presence and location using the SYNTAX I (Synergy Between Percutaneous Coronary Intervention With Taxus and Cardiac Surgery) scoring system.¹⁴ Each stenosis was visually graded and categorized as luminal irregularities (<10%), mild CAD (10%-39%), moderate CAD, (40%-69%), or severe CAD (≥70%). The highest grade documented determined the degree of CAD. In the event of a discrepancy, a third interventional cardiologist (EMH) reviewed the angiograms and determined the final CAD grade. Interobserver agreement of CAD classification by the 2 reviewers was assessed using the kappa statistic.

Moderate CAD was defined as the presence of at least 1 lesion with a stenosis between 40%-69% and severe CAD as the presence of at least 1 lesion with stenosis ≥70%.^15,16 Moderate and severe CAD was considered proximal in patients who had lesions in the upper right coronary artery (SYNTAX segment 1); left main coronary artery (SYNTAX segment 5); or either of its proximal branches, the left anterior descending artery (SYNTAX segment 6) or circumflex artery (SYNTAX segment 11). Multivessel disease was considered present if at least 2 of the major epicardial arteries (right coronary artery, left main coronary artery, left anterior descending, circumflex, or posterior descending arteries) had evidence of moderate-to-severe CAD.

Statistical analysis. Statistical analyses were performed using R statistical software, version 3.6.0.¹⁷ Categorical data were summarized using counts and frequencies and compared using the Chi-square test or Fisher’s exact test, as appropriate. Continuous data were summarized using means and standard deviations and compared using the 1-way ANOVA with equal variance assumption. P-values of ≤.05 were considered to indicate statistical significance. For analysis, we grouped patients with mild CAD and luminal irregularities in one group, and patients with moderate and severe CAD in a second group.

Of 212 patients who underwent PFO closure during this time period for a cryptogenic stroke, 180 underwent concurrent coronary angiography and were included in the study (Figure 1). Of the 32 excluded patients, 12 (5.6%) had established pre-existing CAD and 20 (9.4%) did not have a concurrent coronary angiogram. Of these 20 patients, 8 had their coronary anatomy defined at a prior study or had non-invasive cardiac imaging, while the remaining 12 were subject to operator practice variability, particularly in the initial phase of the policy implementation whereby younger patients without cardiovascular risk factors did not undergo angiography. There were no statistically significant differences in the baseline characteristics of patients with and without coronary angiography (Table 1). Interobserver variability was assessed between the 2 independent cardiologists performing the coronary angiographic assessments. Cohen’s kappa coefficient showed excellent agreement between assessments (k=0.91; P<.05).

Baseline patient characteristics according to CAD status are presented in Table 2. There was no difference in the degree of preprocedural clinical symptoms or in baseline medications (antithrombotic agents and statins). Normal coronary arteries and no luminal irregularities were found in 125 patients (69%), while 55 patients (31%) had at least mild CAD. Mild, moderate, and severe CAD were reported in 32 (18%), 15 (8%), and 8 patients (4%), respectively (Figure 2). Of those with moderate-to-severe CAD, 9 (5%) had proximal disease and 9 (5%) had multivessel disease. The severity of CAD increased in concordance with mean age (Figure 3). Statistically significant differences amidst the CAD groups were found for age, body mass index, hypertension, diabetes mellitus, and migraines. Of the 23 patients found to have moderate-to-severe CAD, 8 (35%) underwent further risk stratification with stress testing, 4 (17%) were managed medically, and 1 (4%) underwent synchronous angioplasty with PFO closure (Figure 2). The median duration for follow-up was 136 days (interquartile range, 99-336 days). Post procedure, the proportion of patients with and without CAD who reported cardiac symptoms was comparable.

Procedural safety. PFO device closure was performed using the Amplatzer PFO Occluder, Gore Helex Septal Occluder, Amplatzer Septal Occluder, and the CardioSeal/StarFlex Occluder in 170 patients (94.4%), 5 patients (2.8%), 3 patients (1.7%), and 2 patients (1.1%), respectively. Femoral arterial access was used to perform coronary angiography in the majority of patients, with the exception of 4 patients in whom radial arterial access was used. The femoral site was closed using a ProGlide suture (Abbott Medical), as was the venous access. No device embolizations occurred. The average total contrast volume was 174 mL of iodixanol (GE Healthcare), average procedural radiation dose was 6916 cGy•cm², average procedural time was 40 minutes, and there were no acute coronary vascular events. Two patients developed vascular access complications (2 right femoral artery lacerations) and were successfully treated by vascular surgery without sequalae. Four patients developed arrhythmias (1 atrial fibrillation/flutter, 1 atrial fibrillation, 1 supraventricular tachycardia, and 1 bradycardia) and required intraoperative medical therapy. No acute renal injury or urgent dialysis was documented in the postoperative period.

This single-center retrospective study assessed outcomes in adults ≥40 years old who underwent concurrent coronary angiography at the time of percutaneous PFO closure following a diagnosis of cryptogenic stroke. Moderate or severe CAD was found in 12% of patients. Half of those with moderate or severe CAD were subject to further diagnostic testing or pharmacologic intervention. Pre- and postprocedure cardiac symptoms of chest pain and dyspnea were not discriminative surrogates for the presence of CAD.

In 2016 and 2018, the United States Food and Drug Administration approved the Amplatzer PFO occluder and Gore Cardioform device for PFO closure after a cryptogenic stroke in patients 18-60 years of age.¹⁸ The landmark RESPECT, CLOSE, and REDUCE RCTs affirmed the superiority of percutaneous PFO closure with antiplatelet therapy over antiplatelet therapy alone in the prevention of recurrent ischemic stroke, marking the end of over 20 years of clinical uncertainty that has since resulted in an increase in the volume of percutaneous PFO closures.^7-10,19 To date, there has been only 1 preceding study examining subclinical CAD prevalence in a PFO population with routine coronary angiography. In 2007, Wahl et al conducted an analysis of 494 asymptomatic patients in men >40 years old, women >50 years old, or those with a particular risk pattern.¹³ Their cohort had a prevalence of cardiovascular risk factors nearly equivalent to our cohort. Significant CAD (defined as a stenosis ≥50%) was found in 53 patients (9%), while 19 patients (3%) had advanced CAD (triple-vessel disease or involvement of the proximal left anterior descending artery). With similar angiographic criteria applied, the CAD prevalence in their PFO population is consistent with our findings.¹³

Development of cardiac symptoms after PFO closure is not infrequent. Nearly 12% of our cohort experienced chest pain before PFO closure, and 20% reported chest pain post closure. Cardiac symptoms after PFO closure are common, often occurring at the first follow-up visit.²⁰ In a previous study of patients undergoing percutaneous closure of PFOs and atrial septal defects, 20% of patients reported chest pain in the follow-up period.²¹ The crossover in symptomology between ischemic and non-ischemic etiologies renders management of chest pain post PFO closure challenging; as such, additional investigations are occasionally employed to identify treatable cardiac pathologies contributing to a patient’s symptoms. In addition to the routine TTE, functional testing such as CT coronary angiography and myocardial perfusion imaging can be considered.^22,23 These modalities can non-invasively determine a patient’s CAD status and inform cardiac risk prognostication.

Drawbacks to concurrent angiography include an additional vascular access. The increased contrast volume, prolongation in procedure times, and 2 cases (1%) of arterial vascular access injury (both treated successfully with surgery) observed in our study may be attributed to the use of the femoral approach, which was prevalent in the timeframe during which the data were collected. Contemporary use of radial-first access is associated with an improved safety profile and would thus minimize the increase in procedure time and vascular complications associated with additional coronary angiography.²⁴ We purposely did not perform non-invasive risk stratification in patients undergoing PFO closure before the procedure, with the knowledge that we would be performing angiography at the time of the procedure. Similarly, our approach obviated the need for diagnostic investigations post procedure in patients presenting with chest pain of an ischemic nature. The use of angiography also confirmed the absence of CAD prior to the implantation of a costly cardiac device.

Of the 12.8% of patients with moderate-to-severe CAD, 65% (8% of the total cohort) were managed medically, of which half had no change in their medications. This highlights the use of pharmacologic therapy (predominantly statins) for ongoing risk modification after stroke. A remaining 35% of patients with moderate-to-severe CAD (4% of the total cohort) underwent risk stratification with non-invasive imaging, while only 1 patient (0.6% of the total cohort) underwent concurrent angioplasty. These findings suggest that the overall impact of CAD detection with coronary angiography in this patient cohort was minimal and affected only a small percentage of patients.

There is no current consensus on which patients should undergo concurrent coronary angiography when undergoing transcatheter PFO closure following a cryptogenic stroke. Based on the findings of this retrospective cohort, there does not appear to be a role for routine coronary angiography in the absence of significant symptoms. An important balance between the value of diagnosing CAD in this population and the risks of coronary angiography is to be weighed on an individual patient basis. We hope this experience informs other interventional cardiology programs who face similar questions. We have always believed that a thorough clinical history to identify symptoms suggestive of coronary ischemia is warranted. In patients deemed at intermediate risk of CAD based on the presence of concomitant cardiovascular risk factors or suggestive ischemic symptoms, it may be worthwhile to consider a standard approach of non-invasive functional testing prior to PFO closure, as noted by current guidelines.²⁵ Although our registry did not have the capacity to determine if patients with moderate-to-severe CAD underwent revascularization beyond their immediate follow-up at our institution, further follow-up data on this cohort may be valuable in ascertaining their long-term clinical outcomes (eg, major cardiovascular events and revascularization rates).

Study limitations. This study was performed in a single quaternary institution, rendering the study cohort susceptible to referral bias. Baseline characteristics of our study sample were nonetheless comparable to those reported in previous studies.¹³ Despite the intended consecutive nature of our patient sampling, not all patients underwent concurrent coronary angiography; however, there were no statistically significant differences in baseline characteristics in those who underwent angiography vs those who did not. Furthermore, clinical data were derived from medical records, which may be subject to missing or inconsistently reported data. We used 40% instead of 50% as the cut-off between mild and moderate intraluminal coronary stenosis, initially in an attempt to be more sensitive in capturing CAD; however, it is recognized in the literature that most clinicians use 50% to denote a moderately stenosed coronary lesion. This definition may have thus resulted in slight overestimation of the CAD prevalence. Our registry did not have the capacity to determine whether patients with significant CAD underwent revascularization beyond their follow-up at our site given that their ongoing care was managed by their referring cardiologists in peripheral geographic territories. Lastly, determination of CAD was based on visual analysis by 2 observers in lieu of an objective quantification methodology; however, our overall estimates of CAD were comparable to those of previously reported studies.¹³

This study reports on the status of CAD in patients with cryptogenic strokes who underwent combined percutaneous PFO closure and coronary angiography. Based on these data, the overall prevalence of CAD is low. There does not appear to be a role for routine coronary angiography at the time of PFO closure in the absence of significant cardiovascular symptoms. These findings can be used to inform upcoming studies on perioperative CAD management and the eventual development of a coronary assessment algorithm for patients undergoing PFO transcatheter closure following cryptogenic stroke. It may be worthwhile to explore the role of functional testing in symptomatic patients at intermediate risk prior to undergoing PFO closure. With emerging acceptance of percutaneous PFO closure as a management strategy for the risk reduction of recurrent strokes in patients diagnosed with cryptogenic stroke, these findings can be used to inform forthcoming PFO closure periprocedural guideline development.

Acknowledgments. This work is supported by the Peter Munk Chair in Structural Heart Disease Intervention. We recognize Ms Louise Pei and Yvonne Bach for their contributions in data collection.