Evaluating Rates of Recurrent Ischemic Stroke Among Young Adults With Embolic Stroke of Undetermined Source: The Young ESUS Longitudinal Cohort Study

Key Points

Question  What are the rates of recurrent ischemic stroke and new-onset atrial fibrillation (AF) in younger adults with embolic stroke of undetermined source (ESUS), and what factors are associated with these prognoses?

Findings  In this longitudinal cohort study of 535 patients, recurrent stroke risk was 1.9 per 100 patient-years. History of stroke or transient ischemic attack, diabetes, and coronary artery disease were associated with recurrence, and AF was detected in 2.8% of participants during follow-up.

Meaning  These findings suggest that young adults with ESUS may have a relatively low rate of subsequent ischemic stroke and new-onset AF compared with older adults.

Importance  Cryptogenic strokes constitute approximately 40% of ischemic strokes in young adults, and most meet criteria for the embolic stroke of undetermined source (ESUS). Two randomized clinical trials, NAVIGATE ESUS and RESPECT ESUS, showed a high rate of stroke recurrence in older adults with ESUS but the prognosis and prognostic factors among younger individuals with ESUS is uncertain.

Objective  To determine rates of and factors associated with recurrent ischemic stroke and death and new-onset atrial fibrillation (AF) among young adults.

Design, Setting, and Participants  This multicenter longitudinal cohort study with enrollment from October 2017 to October 2019 and a mean follow-up period of 12 months ending in October 2020 included 41 stroke research centers in 13 countries. Consecutive patients 50 years and younger with a diagnosis of ESUS were included. Of 576 screened, 535 participants were enrolled after 1 withdrew consent, 41 were found to be ineligible, and 2 were excluded for other reasons. The final follow-up visit was completed by 520 patients.

Main Outcomes and Measures  Recurrent ischemic stroke and/or death, recurrent ischemic stroke, and prevalence of patent foramen ovale (PFO).

Results  The mean (SD) age of participants was 40.4 (7.3) years, and 297 (56%) participants were male. The most frequent vascular risk factors were tobacco use (240 patients [45%]), hypertension (118 patients [22%]), and dyslipidemia (109 patients [20%]). PFO was detected in 177 participants (50%) who had transthoracic echocardiograms with bubble studies. Following initial ESUS, 468 participants (88%) were receiving antiplatelet therapy, and 52 (10%) received anticoagulation. The recurrent ischemic stroke and death rate was 2.19 per 100 patient-years, and the ischemic stroke recurrence rate was 1.9 per 100 patient-years. Of the recurrent strokes, 9 (64%) were ESUS, 2 (14%) were cardioembolic, and 3 (21%) were of other determined cause. AF was detected in 15 participants (2.8%; 95% CI, 1.6-4.6). In multivariate analysis, the following were associated with recurrent ischemic stroke: history of stroke or transient ischemic attack (hazard ratio, 5.3; 95% CI, 1.8-15), presence of diabetes (hazard ratio, 4.4; 95% CI, 1.5-13), and history of coronary artery disease (hazard ratio, 10; 95% CI, 4.8-22).

Conclusions and Relevance  In this large cohort of young adult patients with ESUS, there was a relatively low rate of subsequent ischemic stroke and a low frequency of new-onset AF. Most recurrent strokes also met the criteria for ESUS, suggesting the need for future studies to improve our understanding of the underlying stroke mechanism in this population.

Stroke is a leading cause of death and disability with an increasing global burden. In 2019, there were an estimated 12.2 million incident stroke cases worldwide (95% uncertainty interval, 11.0-13.6), representing a 17% decrease (95% uncertainty interval, 15-18) in age-standardized incidence but a 70% increase (95% uncertainty interval, 67-73) in the absolute number of incident strokes globally since 1990.¹ A particular concern is that the incidence of young-onset ischemic stroke has steadily increased since the 1980s compared with rates among older adults.² Compared with the 1990s, in 2013 there was a 37% increase in disability-adjusted life-years following ischemic strokes in this population.³ Hence, strokes in young adults pose a major socioeconomic impact, and there is an urgent need to improve primary and secondary prevention of ischemic stroke in young adults. Timely and accurate etiological investigations as well as understanding the pathophysiology behind index and recurrent strokes are of utmost importance for optimal secondary prevention.

About 30% to 40% of strokes in young adults^4-6 are neither lacunar nor associated with identifiable sources of embolism, such as atrial fibrillation (AF) or atherosclerotic stenosis of the intracranial or extracranial arteries. These ischemic strokes were formerly called cryptogenic.⁷ Based on defined diagnostic criteria, currently most cryptogenic strokes are classified as embolic stroke of undetermined source (ESUS).⁸ The clinical construct of ESUS was the basis for 2 large international randomized clinical trials: New Approach Rivaroxaban Inhibition of Factor Xa in a Global Trial versus ASA to Prevent Embolism in Embolic Stroke of Undetermined Source (NAVIGATE ESUS)⁹ and Dabigatran Etexilate for Secondary Stroke Prevention in Patients With Embolic Stroke of Undetermined Source (RESPECT ESUS).¹⁰ Both trials found a high rate of stroke recurrence (an annualized rate of approximately 5%) in patients with ESUS but failed to show benefit with direct oral anticoagulant therapy over aspirin in preventing recurrent ischemic stroke. In total, these 2 trials had a sample size of approximately 12 600, but only RESPECT ESUS allowed patients 50 years and younger to be enrolled. Therefore, only 4.8% of the total enrolled were 50 years and younger.

So far, only a few studies have investigated ESUS in young adults specifically, and these studies are limited by small sample size and lack of prospective follow-up.^6,11-13 In this prospective cohort study, we sought to address important gaps in knowledge regarding young patients with ESUS. Specifically, we characterized the clinical features of young patients with ESUS and determined the subtype of the recurrent stroke and the rates of ischemic stroke recurrence, death, and AF during a follow-up period of up to 18 months.

We conducted the Young ESUS international prospective longitudinal cohort study in 41 stroke research centers in 13 countries, obtained via invitations to sites involved in NAVIGATE ESUS. Written consent from participants and institutional review board approval from each participating site were obtained. The study followed the Consolidated Standards of Reporting Trials (CONSORT) reporting guideline.

Participants were enrolled from October 2017 to October 2019 and follow-up was completed in October 2020. Follow-up was conducted at 6, 12, and 18 months or at end of study, whichever was earlier, via a structured telephone interview that ascertained the following: whether the patient was alive or dead, whether the patient had had a recurrent ischemic stroke, and whether AF had been diagnosed in the interval. Information was collected to determine the modified Rankin Scale score using a validated 2-question telephone survey.¹⁴ Source documents were collected for primary outcomes of stroke and AF and assessed by 2 blinded adjudicators.

Patients 50 years and younger who had an ESUS within 60 days of screening were eligible for enrollment. The age cutoff for this study of 50 years was chosen to align with existing registries that defined young adults as those younger than 45 or 49 years. Race and ethnicity data were collected to be in concordance with the data collected in the NAVIGATE ESUS study. Participants self-reported race. To qualify as an ESUS, the stroke had to be nonlacunar, noncardioembolic (eg, AF, left ventricular thrombus, mechanical prosthetic valve), have patent extracranial arteries (less than 50% stenosis), and have no other specific cause of stroke found (eg, arterial dissection). AF was defined as AF on 12-lead electrocardiogram or episodes of AF lasting 6 minutes or longer. Intracranial imaging was optional, but if performed, evidence of 50% or greater stenosis was an exclusion criterion. After the qualifying stroke, at least 24 hours of cardiac rhythm monitoring were required to rule out AF, although investigators could choose to monitor for longer periods. Patients with implantable loop recorders were allowed in the study, as were those with patent foramen ovale (PFO). Echocardiography was required before enrollment.

The primary outcome was recurrent ischemic stroke (according to updated American Heart Association and American Stroke Association criteria)¹⁵ or death. Information on the recurrent stroke was collected to determine the stroke subtype. The secondary outcome was rate of new-onset AF during follow-up. During follow-up, there was no requirement for additional cardiac rhythm monitoring except when a participant had a stroke, in which case they would receive similar monitoring for detection of AF as prior to study entry. To reduce bias in ascertaining outcome events, 2 blinded neurologists centrally adjudicated all stroke, death, and AF events.

Characteristics of all participants were described and compared with baseline characteristics for selected prespecified subgroups. Descriptive statistics used mean (SD), median (IQR), or proportion as appropriate. Univariate comparisons were made using t tests for continuous variables and χ² tests for categorical variables, and we present nominal 2-sided P values. For comparisons within subgroups, we considered P values less than .01 to be significant to account for the multiple comparisons. Risk of recurrence was calculated as a percentage of participants with an event during the follow-up period. Owing to the small number of participants with an event, exact binomial confidence intervals were calculated. The event rate was calculated as the number of participants with an event divided by the total number of participant-years of follow-up.

To determine the predictors of recurrent stroke or death, the following risk factors were considered: previous stroke or transient ischemic attack, current tobacco use vs never or former tobacco use, age, history of diabetes, hypertension, presence of PFO, and aspirin use prior to index event. Predictors were considered in a univariable model using a modified Poisson regression model with robust error variances, and relative risk and associated 95% CIs are presented for each predictor. Predictors with an associated P value less than .10 from the univariable analysis were considered in the multivariable modified Poisson regression model with robust error variances, and their relative risk, associated 95% CIs, and P values are presented.

Attempts were made to address missing data by database queries and by communicating with sites to complete any pending data. No attempt was made to impute missing values, and only observed values were used in the analyses; therefore, some analyses did not include all enrolled patients. All participants were censored at their last follow-up. If a participant was lost to follow-up, all the information until their last follow-up was considered.

Of 576 patients screened, 535 were enrolled in the study. The mean (SD) age was 40.4 (7.3) years, and 297 (56%) participants were male. Most participants (271 [51%]) were from Western Europe. Thirty-five participants (6.5%) were Asian, 40 (7.5%) were Black, 74 (13.8%) were Latino, 379 (70.8%) were White, 15 were of another race or ethnicity (participants self-reported as “other”), and 15 participants (2.8%) did not report race or ethnicity data (Table 1). Patients were monitored for a median (IQR) 17.7 (16.2-18.3) months, and 15 (2.8%) were lost to follow-up at the end of 18 months (Figure).

The main baseline characteristics for the entire study population are summarized in Table 1. The most common vascular risk factors included current or former tobacco use (240 participants [45%]), hypertension (118 participants [22%]), and dyslipidemia (109 participants [20%]). Thirty-four patients (6%) had ipsilateral nonstenotic carotid plaques. All participants had extracranial vascular imaging, echocardiography, and initial cardiac rhythm monitoring as required by protocol. Twenty-one participants (4%) had implantable loop recorders. In addition, 493 (92%) had intracranial vascular imaging. Median (IQR) baseline National Institutes of Health Stroke Scale (NIHSS) score was 2 (1-6) at the time of enrollment. Intravenous alteplase was administered to 123 patients (23%), 65 (12%) received endovascular thrombectomy, and 38 (7%) received both intravenous alteplase and endovascular thrombectomy for index ESUS (Table 1).

Of the 535 enrolled, 354 participants (66%) had transthoracic echocardiograms with bubble studies. Of those who had transthoracic echocardiograms with bubble study, 177 (50%) had PFO identified. Transesophageal echocardiography was done in 226 participants (42%), 113 (50%) had a PFO identified, and 14 (3%) had aortic arch atheroma identified. The PFO status was unknown in 148 (28%) (Table 2).

At enrollment, 297 (56%) participants were receiving aspirin monotherapy, 60 (11%) clopidogrel monotherapy, 110 (21%) dual antiplatelet therapy, 52 (10%), and 1 patient was receiving ticagrelor anticoagulation. Angiotensin-converting enzyme inhibitor was the most commonly used antihypertensive in all regions, followed by calcium channel blockers. Statins were used in 354 participants (66%) (Table 3). Most participants (288 [54%]) were receiving aspirin monotherapy at 6 months, followed by clopidogrel (106 [20%]) and dual antiplatelet therapy (41 [8%]), and 57 participants (11%) were taking anticoagulation, usually a direct oral anticoagulant therapy (eTable in Supplement 1).

Recurrent strokes occurred in 14 patients, and 3 patients died during follow-up. The primary outcome, a composite of ischemic stroke or death, occurred in 16 participants (3.0%; 95% CI, 1.7-4.8), or 2.19 events per 100 patient-years. Stroke recurrence was 2.6% (14 of 535 participants; 95% CI, 1.4-4.4), or 1.9 events per 100 patient-years. AF was detected in 15 participants (2.8%; 95% CI, 1.6-4.6), and 1 patient with implantable loop recorder was diagnosed with AF at 18-month follow-up.

Of the 14 recurrent strokes, 2 (14%) were classified as cardioembolic (aortic valve mass—noted at the time of the index stroke but only thought to be important after the recurrent stroke—and left ventricular thrombus following Takotsubo cardiomyopathy). Two recurrent strokes were secondary to new diagnosis of dissection. One patient was diagnosed with mitochondrial encephalomyopathy, lactic acidosis, and stroke-like episodes at the time of recurrent stroke. The remaining 9 recurrent strokes (64%) met the criteria for ESUS.

In univariate analysis history of stroke or transient ischemic attack before enrollment, history of diabetes and history of coronary artery disease were associated with recurrent ischemic strokes. These remained significant in the multivariate analysis with history of coronary artery disease having the highest hazard ratio for recurrent events (hazard ratio, 10.19; 95% CI, 4.82-21.51) (Table 4).

To our knowledge, the Young ESUS longitudinal cohort study is the largest existing prospective registry for ESUS patients 50 years and younger. This study encompassed a wide spectrum of patients across multiple continents and yielded 3 main findings. First, the rate of recurrent ischemic stroke was lower than that of older patients with ESUS, most of the recurrent ischemic strokes met the criteria for ESUS, and no source of embolus was identified in two-thirds of participants despite repeated diagnostic workup at the time of recurrent stroke. Patients with a previous history of stroke, hypertension, or coronary artery disease were more likely to have a recurrent ischemic stroke or an ESUS. Second, recurrences were not associated with AF or presence of PFO. The rate of AF in this younger population was low, and even though the prevalence of PFO was high, it was not associated with an increase in risk of recurrent stroke. Third, presence of aortic arch atheroma and ipsilateral carotid plaques was low in this young population.

In RESPECT ESUS, patients 50 years and younger showed an annualized stroke rate of 3.5%.¹⁰ RESPECT ESUS required the presence of 1 or more additional vascular risk factors for younger patients, which allowed them to select a high-risk population but likely also introduced a selection bias, which could explain the observed difference in stroke recurrence rates. The annualized stroke rate of 1.9% we observed in our study is closer to the rates observed in previous retrospective studies of the young ESUS population (rate of recurrent stroke or transient ischemic attack of 1.09; 95% CI, 0.54-1.65 per 100 patient-years).¹² The predominance of embolic features in most of the recurrent strokes after ESUS in our analysis supports the validity of the ESUS construct in terms of embolism as the pathogenic mechanism. Persisting uncertainty regarding the embolic source after stroke recurrence in most of our patients suggests that the search for an underlying source may remain futile in many individuals with ESUS, given the limitations of widely used diagnostic testing at present. Consequently, addressing a specific embolic source by targeted antithrombotic stroke prevention remains an unresolved dilemma for many patients with ESUS.

When the ESUS construct was originally proposed, common thinking was that covert AF may frequently underlie ESUS and that anticoagulation may be beneficial for stroke prevention because of this association. This thinking may have led some clinicians to treat patients in the Young ESUS cohort with anticoagulation. However, the lack of success with anticoagulation in the NAVIGATE ESUS⁹ and RESPECT ESUS¹⁰ trials contradicts the assumption of an association between AF-related cardioembolism and ESUS. Even though AF-related recurrent ischemic strokes after ESUS are infrequent (9% in NAVIGATE ESUS⁹), these strokes can have severe outcomes in terms of disability and mortality compared with other recurrent stroke types.¹⁶ Therefore, although covert AF may not underlie most recurrent strokes after ESUS, its particularly grave outcomes warrant a more extensive search for AF in select populations. In current clinical practice, the use of implantable cardiac monitors is considered a highly effective tool for detecting episodes of AF. However, this is not without cost to the already strained health care system, and there is persistent equipoise about the importance of short duration AF episodes, the anticipated benefit in terms of ischemic stroke risk reduction, and the increase of bleeding events. The LOOP study¹⁷ showed implantable loop recorder screening to increase AF detection and anticoagulation initiation 3-fold, but there was no significant reduction in the risk of stroke or systemic arterial embolism. Even though this study was not conducted specifically in an ESUS population, these findings imply that not all AF is worth screening for, and not all screen-detected AF merits anticoagulation. Hence, developing clinical and investigative criteria to identify patients who would benefit from longer monitoring for AF is of utmost importance. A secondary analysis¹⁸ of the NAVIGATE ESUS⁹ trial showed that left atrial diameter and premature atrial contraction frequency predicted subsequent clinical AF. The ATTICUS study,¹⁹ which stopped after the interim analysis owing to futility, looked at an ESUS population with a high risk profile for cardiac thromboembolism (ie, left atrium larger than 45 mm, spontaneous echo contrast in left atrium appendage, left atrium appendage flow velocity 0.2 cm/s or greater, atrial high-rate episodes, CHA2DS2-Vasc [congestive heart failure, hypertension, age 75 years and older, diabetes, stroke or transient ischemic attack, vascular disease, age 65-74 years, and sex category] score 4 or greater, and PFO). To date, 23% of the study patients have developed AF during follow-up.¹⁹ Hence, further assessment of predictors of AF in the ESUS population is warranted and is the basis for the ongoing randomized clinical trial ARCADIA (NCT03192215).

PFO in patients with stroke may represent an incidental finding, a risk factor for stroke occurrence, or a robust cause. PFO is associated with ischemic stroke through several mechanisms. Most theories support paradoxical embolism, in situ thrombus formation, and arrhythmogenesis.²⁰ Presence of a PFO is considered a potential source of cardioembolism, especially in young adults, despite the difficulties confirming paradoxical embolism, in situ thrombosis, or transient atrial arrythmia.^20,21 In the Young ESUS cohort, the prevalence of PFO was higher than expected in the general population^21,22 and similar to that reported in young patients with cryptogenic stroke or ESUS.^11,20,23 The prevalence rate is more likely an underestimation owing to the underutilization of transesophageal echocardiography and transthoracic echocardiogram with bubble studies in the study, as these were not required by ESUS criteria. Compared with the non-PFO group, participants with PFO were less likely to have traditional vascular risk factors (ie, hypertension, diabetes mellitus, and hyperlipidemia) and had a lower body mass index (calculated as weight in kilograms divided by height in meters squared). However, presence of PFO was not associated with recurrent stroke. This could be because of the relatively shorter length of follow-up in our study or the relatively lower risk of recurrent stroke via PFO.²⁰ Percutaneous PFO closure is known to be beneficial and relatively safe for secondary stroke prevention among adults younger than 60 years with few conventional vascular risk factors who have undergone a thorough work-up of their stroke etiology, who are deemed cryptogenic except for the presence of PFO, and who may have large shunts or associated atrial septal aneurysms.²⁰

Hence, transesophageal echocardiography to identify PFO in young patients with ESUS could have more value than in older patients. At a minimum, young patients with ESUS should be investigated with a transthoracic echocardiogram with bubble study or a transcranial doppler to identify a potential pathogenic PFO or other right-to-left shunt.

The observed lower rates of aortic arch atheroma and ipsilateral carotid plaques in our study could be secondary to lower incidences of traditional vascular risk factors that contribute to plaque development and progression, or the lower sensitivity of traditional imaging modalities, such as ultrasonography, computed tomography angiogram, or magnetic resonance angiogram, in identifying clinically significant plaques. With current advances in imaging modalities, it may be prudent to evaluate whether assessment of vessels or vessel walls with high resolution imaging techniques would be beneficial in this population. Advances in diagnostic imaging of atherosclerotic plaques, both carotid and intracranial, may offer the potential for identifying the cause of stroke among some patients currently classified as having ESUS.

Limitations of this study include participation of selected high-volume stroke centers. A few sites in 1 geographic area may not accurately reflect the spectrum of stroke. However, the study sites represented many global regions, allowing characterization of young patients with ESUS around the world. Further limitations include selection bias owing to the informed consent process and the bias introduced by differential loss to follow-up. Even though patients required an electrocardiogram and 24-hour cardiac monitoring to enroll in the study, there were no study specific recommendations for further monitoring for AF after a recurrent stroke. This may have made some contribution to the low rates of AF detected during follow-up.

In conclusion, in this large cohort of young patients with ESUS, there was a relatively low rate of subsequent ischemic stroke compared with that in older patients with ESUS during midterm follow-up. There was a low frequency of new-onset AF. Most recurrent strokes also met the criteria for ESUS, suggesting the need for future studies to improve our understanding of underlying stroke mechanisms in this population.

Accepted for Publication: December 31, 2021.

Published Online: March 14, 2022. doi:10.1001/jamaneurol.2022.0048

Corresponding Author: Kanjana S. Perera, MBBS, Population Health Research Institute, Hamilton Health Sciences, DBCVSRI C4-104, 237 Barton St E, Hamilton, ON L8L 2X2, Canada (kanjana.perera@phri.ca).

Author Contributions: Dr Hart had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Perera, Hankey, Lee, Ameriso, Meseguer, Hart.

Acquisition, analysis, or interpretation of data: Perera, de Sa Boasquevisque, Rao-Melacini, Taylor, Cheng, Lee, Martí-Fàbregas, Ameriso, Field, Arauz, Coutts, Arnold, Mikulik, Toni, Mandzia, Veltkamp, Meseguer, Haeusler, Hart.

Drafting of the manuscript: Perera, Taylor, Mandzia, Hart.

Critical revision of the manuscript for important intellectual content: Perera, de Sa Boasquevisque, Rao-Melacini, Cheng, Hankey, Lee, Martí-Fàbregas, Ameriso, Field, Arauz, Coutts, Arnold, Mikulik, Toni, Mandzia, Veltkamp, Meseguer, Haeusler, Hart.

Statistical analysis: Perera, Rao-Melacini.

Obtained funding: Perera, Hart.

Administrative, technical, or material support: Perera, Cheng, Lee, Haeusler.

Supervision: Martí-Fàbregas, Arnold, Mikulik, Meseguer, Hart.

Conflict of Interest Disclosures: Dr Perera reported grants from Bayer during the conduct of the study and personal fees from Abbott and nonfinancial support from Bayer outside the submitted work. Dr Hankey reported personal fees from Bayer for consulting about stroke prevention in atrial fibrillation, from Bristol Myers Squibb for serving as a member of the Steering Committee of A Study on BMS-986177 for the Prevention of a Stroke in Patients Receiving Aspirin and Clopidogrel (AXIOMATIC-SSP) trial of milvexian, and from the American Heart Association for serving as an Associate Editor of Circulation outside the submitted work; Dr Hankey was also a member of the steering committee of the Rivaroxaban Versus Aspirin in Secondary Prevention of Stroke and Prevention of Systemic Embolism in Patients With Recent Embolic Stroke of Undetermined Source (NAVIGATE ESUS) trial comparing rivaroxaban vs aspirin in patients with Embolic Stroke of Undetermined Source (ESUS). Dr Ameriso reported grants from Bayer during the conduct of the study and grants from Boehringer Ingelheim and personal fees from Abbott outside the submitted work. Dr Field reported personal fees from Servier Canada Speaker’s Bureau outside the submitted work. Dr Arnold reported personal fees from Amgen, Bayer, Bristol Myers Squibb, Daiichi Sankyo, Medtronic, Novartis, and Pfizer Honoraria for scientific advisory boards and from AstraZeneca, Bayer, Covidien, Medtronic, and Sanofi Honoraria for lectures during the conduct of the study as well as grants from the Swiss National Science Foundation and the Swiss Heart Foundation outside the submitted work. Dr Mikulik reported grants from the COST Association, the Ministry of Education, Youth and Sports of the Czech Republic INTER-EXCELLENCE INTER-COST; the Ministry of Education, Youth and Sports of the Czech Republic National Program of Sustainability II; and the state budget of the Czech Republic outside the submitted work. Dr Toni reported personal fees from Abbott, Bayer, Boehringer Ingelheim, Daiichi Sankyo, Medtronic, and Pfizer outside the submitted work. Dr Mandzia reported personal fees from Bayer outside the submitted work. Dr Veltkamp reported grants from Imperial College London RV, is an investigator of the Imperial BRC, and is partially funded by the European Union’s Horizon 2020 research and innovation program under grant agreement No. 754517 during the conduct of the study; Dr Veltkamp also reports grants from Bayer, Boehringer, Bristol Myers Squibb, Pfizer, Biogen, and Daiichi Sankyo and nonfinancial support from Medtronic, personal fees from Javelin, Alexion, Bayer, and AstraZeneca outside the submitted work; and is the coordinating investigator of the European Union’s Horizon 2020 research and innovation program. Dr Haeusler reported personal fees from Bayer during the conduct of the study and personal fees from Abbott, Amarin, AstraZeneca, Bayer, Biotronik, Boehringer Ingelheim, Bristol Myers Squibb, Daiichi Sankyo, Edwards Lifesciences, Medtronic, Pfizer, Premier Research, Sun Pharma, and W. L. Gore & Associates outside the submitted work. Dr Hart reported grants from Bayer during the conduct of the study. No other disclosures were reported.

Funding/Support: This study was funded by the Population Health Research Institute and in part by Bayer in the first 2 years of the study.

Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Group Information: A complete list of the members of the Young ESUS Investigators appears in Supplement 2.