Introduction: A patent foramen ovale (PFO) may coexist with other potential embolic sources (PESs) in patients with embolic stroke of undetermined source (ESUS), leading to difficulty in attributing the stroke to either the PFO or other PESs. We aimed to investigate the prevalence and predictors of concomitant PESs in ESUS patients with PFOs. Methods: A retrospective cohort study was conducted in a tertiary stroke centre. Consecutive patients with ESUS and a concomitant PFO admitted between 2012 and 2021 were included in the study. Baseline characteristics and investigations as a part of stroke workup including echocardiographic and neuroimaging data were collected. PESs were adjudicated by 2 independent neurologists after reviewing the relevant workup. Results: Out of 1,487 ESUS patients, a total of 309 patients who had a concomitant PFO with mean age of 48.8 ± 13.2 years were identified during the study period. The median Risk of Paradoxical Embolism (RoPE) score for the study cohort was 6 (IQR 5–7.5). Of the 309 patients, 154 (49.8%) only had PFO, 105 (34.0%) patients had 1 other PES, 34 (11.0%) had 2 PES, and 16 (5.2%) had 3 or more PES. The most common PESs were atrial cardiopathy (23.9%), left ventricular dysfunction (22.0%), and cardiac valve disease (12.9%). The presence of additional PESs was associated with age ≥60 years (p < 0.001), RoPE score ≤6 (p ≤0.001), and the presence of comorbidities including diabetes mellitus (p = 0.004), hypertension (p≤ 0.001), and ischaemic heart disease (p = 0.011). Conclusion: A large proportion of ESUS patients with PFOs had concomitant PESs. The presence of concomitant PESs was associated with older age and a lower RoPE score. Further, large cohort studies are warranted to investigate the significance of the PES and their overlap with PFOs in ESUS.

© 2022 The Author(s). Published by S. Karger AG, Basel

Introduction

Embolic stroke of undetermined source (ESUS) accounts for about 9–25% of all ischaemic strokes [1, 2]. ESUS is defined as a non-lacunar brain infarct in the absence of extracranial or intracranial atherosclerosis, major-risk cardioembolic sources, and any other specific cause of strokes [3]. It, therefore, comprises a heterogeneous subgroup of sources of ischaemic stroke with an unknown cause despite extensive investigation. Potential embolic sources (PESs) such as atrial cardiopathy, arterial atherosclerotic plaques, left ventricular (LV) dysfunction, cardiac valvular disease, and cancer are proposed to cause ESUS [4-6]. Another example of PES is a patent foramen ovale (PFO). The foramen ovale is a conduit between the right and left atria which typically closes anatomically by about 12 months of age. A PFO occurs when the foramen ovale persists and is found to be present in 27% of patients who underwent transoesophageal echocardiogram (TEE) in the NAVIGATE-ESUS trial, which is a prevalence higher than in the general population [7, 8]. PFOs are typically asymptomatic but may lead to paradoxical embolization due to the passage of thrombi from systemic venous return through the inter-atrial septum into the systemic circulation, predisposing to embolic strokes.

Given the high prevalence of a PFO in ESUS of 27% [7, 8], one significant clinical challenge is to determine whether the PFO is incidental or causative in the setting of ischaemic strokes [9]. This is important as PFO closure is beneficial in patients with PFO-attributed strokes in recent randomised controlled trials [10], but it is invasive and is associated with post-procedural atrial fibrillation and periprocedural risks [11]. Predictive clinical scores and factors have been identified to help predict the significance of PFOs, such as the Risk of Paradoxical Embolism (RoPE) score based on age, neuroimaging, and the presence of vascular risk factors [12]. PFO morphology, for example, the size of the right to left shunt (RLS), may also affect the risk of stroke [13]. However, there remains a lack of guidance and consensus on optimal patient selection for PFO closure [14].

In ESUS, the PFO may also coexist with other PES, and previous studies have found more than 65% of ESUS patients had >1 PES [4]. However, the prevalence and overlap of additional PESs in ESUS-PFO are unknown. Therefore, we aimed to perform a cohort study investigating the prevalence of PESs in patients with ESUS and PFOs, comparing those with PFO only and those with additional PESs in terms of clinical characteristics, RoPE score, and PFO shunt size.

MethodsStudy Design and Population

This was a retrospective cohort study performed at a single comprehensive stroke centre. Consecutive adult patients with ESUS-PFO who were admitted from 2012 to 2021 were identified. Inclusion criteria for these patients were as follows: (1) ESUS with (2) PFOs identified on transcranial doppler (TCD) bubble study, transthoracic echocardiogram (TTE), or TEE. ESUS was diagnosed according to the consensus criteria, defined by the cryptogenic stroke/ESUS International Working Group [3]. All patients included had neuroimaging (computed tomography or magnetic resonance imaging) and vascular studies (computed tomography angiography, magnetic resonance angiography, or transcranial and extracranial Doppler sonographies), at least 24-h of inpatient cardiac telemetry or Holter monitoring and TTE performed. Echocardiographs were read by trained cardiologists in accordance with American Society of Echocardiography-European Association of Cardiovascular Imaging (ASE-EACVI) guidelines. The study protocol was reviewed and approved by National Healthcare Group Domain Specific Review Board, approval number 2021/00623, and exemption for consent taking was given in view of the retrospective nature of the cohort and use of de-identified data.

Data Collection

Clinical and imaging data were collected from the institution’s electronic medical record and TTE data from the electronic echocardiography database. Data collected include patient demographics, medical comorbidities, and laboratory results.

The PESs were divided into five categories, namely atrial cardiopathy, LV dysfunction, cardiac valve disease, arterial atherosclerosis, and cancer, adapted from a prior study by Ntaios et al. [15]. The definitions of the PESs are detailed in the supplementary material (see www.karger.com/doi/10.1159/000527791 for all online suppl. material). The presence of each PES was determined from the clinical information on electronic medical records and investigations after adjudication by 2 independent neurologists.

The 10-point RoPE score has been proposed to represent the probability of PFO being causative of the stroke, as opposed to it being incidental, and those with higher RoPE scores have a higher likelihood of stroke attributed to PFO [11, 12]. We stratified our patients into high RoPE scores (>6) and low RoPE scores (≤6) consistent with the previous literature [16]. Patients who were detected to have a RLS on TCD were stratified into 2 groups of large and small shunts based on the Spencer Logarithmic Scale (SLS). Spencer grading of 1 (1–10 microbubbles) or 2 (11–30 microbubbles) represented a small shunt, while Spencer grading of 3 (31–100 microbubbles), 4 (101–300 microbubbles), or 5 (>300 microbubbles) represented a large shunt. This was consistent with the previous literature reporting that Spencer grades 3–5 have higher positive predictive value in the detection of large and functional RLS [17].

Statistical Analysis

Means and standard deviations were reported for normally distributed continuous variables, median and interquartile range (IQR) for non-normally distributed variables, while frequencies and percentages were presented for categorical variables. Continuous variables were compared using the t test, while categorical variables were analysed using the χ2 test or Fisher’s exact test as appropriate. Patients with PFO only and those with one or more additional PESs were compared in terms of age, sex, SLS grading, RoPE scores, comorbidities, and laboratory findings. A multivariable logistic regression model adjusting for age, sex, diabetes mellitus, and hypertension was employed, and results are presented as adjusted odds ratios (ORs) with 95% confidence intervals. All data processing was performed through IBM SPSS statistics version 28.0 (IBM Corp., version 28.0 Armonk, NY), where a p value of <0.05 was considered statistically significant.

ResultsBaseline Characteristics

Out of 1,487 ESUS patients, a total of 309 patients who had a concomitant PFO were identified. A summary of the patients’ clinical characteristics is shown in Table 1. The patients had a mean age of 48.8 years (standard deviation, 13.2), and 212 patients (68.6%) were male. The median RoPE score for the study cohort was 6 (IQR 5–7.5), and 163 patients (52.8%) had a large shunt (Spencer grade ≥3). The median National Institute of Health Stroke Score (NIHSS) of the patients was 5 (IQR 2–15). Slightly more than half of the patients (n = 161, 52.1%) had a cortical infarct on brain imaging.

Table 1.

Demographics and baseline clinical characteristics of the study cohort (n = 309)

Overlapping PESs

Amongst 309 patients, 154 (49.8%) did not have any other PESs beside PFOs, 105 patients (34.0%) had 1 other PES only, 34 patients (11.0%) had 2 other PESs, 13 patients (4.2%) had 3 other PESs, and 3 patients (1.0%) had 4 other PESs. In the PFO cohort, 74 patients (23.9%) had concomitant atrial cardiopathy, 68 patients (22.0%) had concomitant LV dysfunction, 40 patients (12.9%) had either concomitant mitral, aortic, or both mitral and aortic valve disease, 19 patients (6.1%) had concomitant cancer, and 22 patients (7.1%) had non-stenosing atherosclerotic plaques identified. LA cardiopathy, LV dysfunction, and cardiac valve disease were PESs that had the highest degrees of overlap (Fig. 1).

Fig. 1.

Illustration of PESs in ESUS patients with PFO. Created with BioRender.com.

Associations with PESs

Patients were more likely to have other PESs beside PFOs if they were older (age ≥60 years) (OR 3.91, 95% CI 2.07–7.38, p ≤ 0.001) and had a RoPE score ≤6 (OR 3.02, 95% CI 1.90–4.79, p ≤ 0.001) (Table 2). The presence of other comorbidities including hypertension (OR 2.30, 95% CI 1.40–3.77, p ≤ 0.001), diabetes mellitus (OR 2.87, 95% CI 1.37–6.01, p = 0.004), and ischaemic heart disease (OR 5.81, 95% CI 1.27–26.67, p = 0.011) was also individually significantly associated with other PESs. A shunt size of Spencer grade ≤2 was not significantly associated with the presence of other PESs (OR 1.04, 95% CI 0.67–1.63, p = 0.862) (Table 2). On multivariable analysis, this association remained significant for age (adjusted OR 3.03, 95% CI 1.56–5.85, p = 0.001) and RoPE scores (adjusted OR 2.39, 95% CI 1.38–4.13, p = 0.002) (Table 3).

Table 2.

Characteristics of ESUS-PFO patients with or without at least 1 other PES

Table 3.

Multivariable logistic regression analysis adjusting for age, sex, diabetes mellitus, and hypertension

Discussion

In this cohort of ESUS-PFO patients, we demonstrated that (1) more than half of the patients had at least 1 PES other than PFO, (2) atrial cardiopathy and LV dysfunction were the PES with the highest prevalence, and (3) patients with additional PESs were more likely to be older, have more medical comorbidities, and a lower RoPE score.

There is an increased prevalence of PFO in ESUS patients than in the general population [7, 8]. However, one-third of the PFOs in ESUS patients are incidental rather than truly pathological [18]. PFO closure has been performed to reduce recurrent stroke incidence in patients with PFO. However, a meta-analysis of 5 RCTs comprising 3,627 patients showed that PFO closure only led to treatment benefit in patients with high-risk PFO characteristics, not those with low-risk PFO characteristics [19]. Furthermore, PFO closure is associated with procedural risks [11]. It is therefore important to improve patient selection for PFO closure by determining whether PFOs are causative of stroke rather than incidental. The PASCAL criteria, comprising of high-risk PFO features on TEE, and the RoPE score, classified patients by the probability that the stroke was causally related to PFOs, which correlated with the magnitude of risk reduction for recurrent stroke with device closure [20]. Taken together with the results of previous RCTs, it has been shown that in well-selected PFO patients without other causes of ischaemic stroke and medium- to high-risk PFOs, the PFO is likely causative of stroke [19]. Therefore, it has been proposed PFO-associated strokes should not be categorized within ESUS but rather a distinct entity [21]. Our finding that ESUS-PFO patients with at least 1 PES other than PFO had a lower RoPE score supports the proposal that PFO patients with other PES had lower probability of having PFO-associated strokes. However, the PASCAL criteria were not evaluated in our study due to limitations in TEE data available and will require further research and validation. More than half of our ESUS-PFO patients had additional PESs, which may also be contributory or causative of ESUS, and may require different secondary prevention strategies.

The striking prevalence of additional PESs in ESUS-PFO patients being present in more than one-third of our study population is in line with previous studies such as an analysis by the NAVIGATE-ESUS trial, where 41% of ESUS patients had >1 PESs and the most common PESs were atrial cardiopathy (37%) and LV dysfunction (36%) in overall ESUS patients [15]. Atrial cardiopathy is a proposed pathophysiological concept of abnormal atrial substrates and functions [22]. Incidence of severe left atrial enlargement, a reliable marker of atrial cardiopathy, was higher in ESUS patients than those with non-cardioembolic strokes [23], and patients with likely pathological PFOs based on the RoPE score were less likely to have atrial cardiopathy [24]. A post hoc analysis of the NAVIGATE-ESUS trial found that administration of rivaroxaban in patients with atrial cardiopathy and ESUS reduced the risk of recurrent strokes, supporting the conclusion that atrial cardiopathy is potentially causative of ESUS [25]. This hypothesis is also currently under testing in the ARCADIA trial [26]. Thus, atrial cardiopathy may represent an independent cause of ESUS, making the PFO finding incidental. PFO has also been shown to be associated with atrial tachyarrhythmia due to effects on the atrial septum, thus, it may be synergistic with atrial cardiopathy, leading to the stroke [27].

Development of one PES may predispose to another, leading to stroke independent of PFO. For example, worsening LV compliance in LV dysfunction results in LA enlargement [28], which may explain the coexistence of cardiac-related PESs in our study. PFO rarely causes LV dysfunction, although large-volume shunts can lead to LV and RV dysfunction in case reports [29]. We found that 22.0% of ESUS-PFO patients had LV dysfunction, which likely developed independently of PFOs. Further research into the interaction of PFOs and other PES, and relative benefits of anticoagulation with or without PFO closure in patients with multiple PESs are needed.

In this study, we found that PFO patients with additional PESs were older and had more comorbidities. Similarly, a 15-year observational study of 169 ESUS patients found that PFO is the only common PES in ESUS patients <50 years old, compared with older patients who significantly more likely had multiple PES [30]. American Academy of Neurology guidelines suggest that PFO closure probably reduces recurrent stroke risk in select patients <60 years with embolic-appearing infarction [31], further supporting a potential role of the PESs in patient selection. In addition to older age, ESUS-PFO patients in our study with diabetes mellitus, hypertension, and ischaemic heart disease were more likely to have additional PESs, cardiovascular comorbidities which are more prevalent in older individuals [30] and associated with atrial cardiopathy and LV dysfunction [32]. Thus, in younger ESUS-PFO patients, it might be beneficial to explore defining other features for pathogenic PFOs without other PESs, to optimise patient selection for PFO closure.

Age and absence of several cardiovascular comorbidities are key variables in the RoPE score [12]. Studies such as a recent pooled analysis comprising 2,289 patients from three randomized trials have validated that the RoPE score reliably discriminates patients with ESUS likely attributable to PFOs, and whom will more likely benefit from PFO closure [16]. Interestingly, we found that ESUS-PFO patients without other PES were more likely to have a RoPE score >6, which supports our finding that the absence of an additional PES suggests the PFO to be more likely causative of stroke. The RoPE validation study performed by Strambo et al. [33] found that the additional variables such as absence of LV hypertrophy and atherosclerosis were also predictive of pathogenic PFO. Additionally, a 2021 multicentre study by Ntaios et al. [34] demonstrated an inverse relationship between the presence of carotid atherosclerosis, a PES, and likely pathogenic PFO in patients with ESUS. Therefore, it would be interesting to study the addition of PESs to RoPE to improve its predictive value for pathogenic PFO.

The relationship between the RLS size and stroke risk has been mixed in the previous literature. A pooled analysis of 2 prospective observational studies and 2 RCTs showed that a large PFO was not independently associated with recurrent strokes after adjusting for various factors [35]. In contrast, a small shunt was a significant predictor of recurrent strokes in patients with PFO and cryptogenic strokes, only in patients with RoPE >6 and not ≤6 [36]. In our study, we found no significant association of the shunt size with likelihood of having additional PESs in ESUS-PFO patients. Thus, further studies to elucidate the significance of the shunt size in predicting whether the PFO is pathogenic are required.

Limitations

We acknowledge that this study has several limitations. Firstly, in view of the retrospective nature of the study, we could only demonstrate associations and not causality. Secondly, the PES was defined according to Ntaois et al. [15]. However, there are several recently identified PESs such as new potential biomarkers of atrial cardiopathy such as NT-proBNP [26] and left atrial strain [37] that were not included in this retrospective cohort, which might have underestimated the number and overlap of PESs in ESUS-PFO patients. Lastly, the PFO was diagnosed in this cohort either by TCD bubble study, TTE, or TEE. Not all patients had a confirmatory diagnosis by TEE, which is the gold standard evaluation, and therefore, PASCAL classification could not be reliably calculated in this study. Nonetheless, TCD bubble study has been shown to be an appropriate screening modality for PFOs with good sensitivity and specificity [38].

Conclusion

We demonstrated in our cohort that more than half of ESUS-PFO patients have other PESs, and those with PFO alone were more likely to be <60 years old, and have a RoPE score of >6. The absence of additional PESs may be predictive of pathogenic PFO and may supplement the validated RoPE score. Further studies are warranted to investigate the significance of the PES and their overlap with PFOs, and RCTs are needed to study the benefit versus the risk of invasive PFO closure in patients with PFOs only and no additional PESs.

Statement of Ethics

The study protocol was reviewed and approved by National Healthcare Group Domain Specific Review Board, approval number 2021/00623, and exemption for consent taking was given in view of the retrospective nature of the cohort and use of de-identified data.

Conflict of Interest Statement

The authors declare that they have no conflict of interest.

Funding Sources

Benjamin YQ Tan was supported by the MOH Healthcare Research Scholarship, National Medical Research Council, Singapore.

Author Contributions

Keith Zhi Xian Toh, Ming Yi Koh, Jamie SY Ho, Kathleen Hui Xin Ong, Yong Qin Lee, Xintong Chen, Jun Tao Fang, Elliot Yeung Chong, Isis Claire ZY Lim, Yao Hao Teo, and Yao Neng Teo contributed to data collection, data curation, and data analysis. Keith Zhi Xian Toh, Ming Yi Koh, and Jamie SY Ho contributed to writing original draft. Christopher YK Chua, Yinghao Lim, Bernard PL Chan, Vijay K Sharma, Leonard LL Yeo, Ching Hui Sia, and Benjamin YQ Tan contributed to study conceptualization, editing manuscript, and supervision of study.

Data Availability Statement

All data generated or analysed during this study are included in this article. Further enquiries can be directed to the corresponding author.

This article is licensed under the Creative Commons Attribution-NonCommercial 4.0 International License (CC BY-NC). Usage and distribution for commercial purposes requires written permission. Drug Dosage: The authors and the publisher have exerted every effort to ensure that drug selection and dosage set forth in this text are in accord with current recommendations and practice at the time of publication. However, in view of ongoing research, changes in government regulations, and the constant flow of information relating to drug therapy and drug reactions, the reader is urged to check the package insert for each drug for any changes in indications and dosage and for added warnings and precautions. This is particularly important when the recommended agent is a new and/or infrequently employed drug. Disclaimer: The statements, opinions and data contained in this publication are solely those of the individual authors and contributors and not of the publishers and the editor(s). The appearance of advertisements or/and product references in the publication is not a warranty, endorsement, or approval of the products or services advertised or of their effectiveness, quality or safety. The publisher and the editor(s) disclaim responsibility for any injury to persons or property resulting from any ideas, methods, instructions or products referred to in the content or advertisements.