Abstract
Background and Purpose—Ticagrelor is an effective antiplatelet therapy among patients with atherosclerotic disease and, therefore, could be more effective than aspirin in preventing recurrent stroke and cardiovascular events among patients with embolic stroke of unknown source (ESUS), which includes patients with ipsilateral stenosis <50% and aortic arch atherosclerosis.
Methods—We randomized 13 199 patients with a noncardioembolic, nonsevere ischemic stroke or high-risk transient ischemic attack to ticagrelor (180 mg loading dose on day 1 followed by 90 mg twice daily for days 2–90) or aspirin (300 mg on day 1 followed by 100 mg daily for days 2–90) within 24 hours of symptom onset. In all patients, investigators informed on the presence of ipsilateral stenosis ≥50%, small deep infarct <15 mm, and on cardiac source of embolism detected after enrollment or rare causes, which allowed to construct an ESUS category in all other patients with documented brain infarction. The primary end point was the time to the occurrence of stroke, myocardial infarction, or death within 90 days.
Results—ESUS was identified in 4329 (32.8%) patients. There was no treatment-by-ESUS category interaction (P=0.83). Hazard ratio in ESUS patients was 0.87 (95% confidence interval, 0.68–1.10; P=0.24). However, hazard ratio was 0.51 (95% confidence interval, 0.29–0.90; P=0.02) in ESUS patients with ipsilateral stenosis <50% or aortic arch atherosclerosis (n=961) and 0.98 (95% confidence interval, 0.76–1.27; P=0.89) in the remaining ESUS patients (n=3368; P for heterogeneity =0.04).
Conclusions—In this post hoc, exploratory analysis, we found no treatment-by-ESUS category interaction. ESUS subgroups have heterogeneous response to treatment (Funded by AstraZeneca).
Clinical Trial Registration—URL: http://www.clinicaltrials.gov. Unique identifier: NCT01994720.
Introduction
Ischemic strokes of unknown cause or cryptogenic strokes represent ≤40% of ischemic strokes.1,2 Cryptogenic stroke is a highly heterogeneous group with multiple potential pathologies as mechanisms of stroke.3–5 Magnetic resonance imaging has led to an understanding of the patterns of embolic infarction, whether artery-to-artery emboli from large artery atherosclerosis or cardioembolic from or via the heart. The concept of embolic stroke of unknown source (ESUS) has been proposed to subclassify cryptogenic stroke patients with an embolic stroke mechanism, regardless of the underlying pathology.6 The ESUS construct is based on patients with documented brain infarction and no ipsilateral extra- or intracranial stenosis ≥50%, no small deep infarct <15 mm, no cardiac source of embolism (eg, atrial fibrillation), or no other definite rare cause.6 These patients may benefit from anticoagulant rather than antiplatelet therapy, which is under evaluation in 2 large randomized control trials (clinicaltrials.gov NCT02313909 and clinicaltrials.gov NCT02239120). If these trials are neutral, as were previous comparison between oral anticoagulant and aspirin,7 then there is a rationale to examine what types of antiplatelet therapy is best in this patient subset.
A proportion of patients with ESUS have atherosclerotic disease, including ipsilateral stenosis <50% or aortic arch atherosclerosis or atherosclerotic disease in other vascular beds. There is no convincing evidence that oral anticoagulants are superior to aspirin in this setting. Antiplatelet treatment is intended to limit thrombosis superimposed on ulcerated atherosclerotic plaque and subsequent distal embolization to the brain.8 If the mechanism of stroke among ESUS patients is actually artery-to-artery emboli, antiplatelet therapy may be the ideal stroke prevention therapy. New antiplatelet agents that are potentially more effective than aspirin and other currently recommended antiplatelet agents (including in cryptogenic strokes) could be good candidates to reduce the recurrence risk.
Ticagrelor has been shown to be more effective than clopidogrel,9 and then placebo,10 on a background of aspirin among patients with coronary atherosclerotic disease. In the SOCRATES trial (Acute Stroke or Transient Ischemic Attack Treated With Aspirin or Ticagrelor and Patient Outcomes; clinicaltrials.gov NCT01994720), ticagrelor 90 mg BID was evaluated against aspirin 100 mg/d among patients with TIA and minor stroke of noncardioembolic origin that occurred within the previous 24 hours; it found a nonsignificant 11% relative risk reduction in the composite of stroke, myocardial infarction (MI), and death.11 Investigators systematically used the ASCOD (Atherosclerosis, Small vessel disease, Cardiac pathology, Other, Dissection) phenotyping system to assign stroke etiology,12 which provided the data to define an ESUS subgroup among the SOCRATES trial patients. Our post hoc, exploratory hypothesis was that all or some patients with ESUS would show greater benefit from ticagrelor compared with aspirin.
Material and Methods
Patients
Patients ≥40 years were randomized in the SOCRATES trial within 24 hours of symptom onset of a nonsevere ischemic stroke (National Institutes of Health Stroke Scale score ≤5) or high-risk transient ischemic attack (high-risk was defined by an ABCD2 score ≥4 (stroke risk score ranging from 0 to 7 with higher scores indicating higher risk of stroke) or an ipsilateral atherosclerotic stenosis).11,13 Among major exclusion criteria were cardiac embolism.
Collection of Data
In addition to collection of demographic and clinical characteristics,13 the ASCOD grading system was used in all patients at visit 1 (enrollment/randomization), but investigators could complete the investigations at any time during hospitalization, at premature treatment discontinuation visit, and at visit 5 (end of treatment at 90 days).12 The burden of atherosclerotic disease (A of ASCOD) was scored as follows: if no atherosclerotic disease was found, patients were graded A0; if there was a ≥50% ipsilateral stenosis of extracranial or intracranial arteries or a mobile thrombus in the aortic arch, they were graded A1 (likely causal); if there was a <50% ipsilateral stenosis of extracranial or intracranial arteries or an aortic arch plaque >4 mm in thickness without mobile thrombus, they were graded A2 (causal relationship possible but uncertain); and if there was only a plaque without stenosis or a stenosis contralateral to the cerebral infarct or a concomitant coronary or peripheral arterial disease, they were graded A3 (unlikely causal). If either intracranial arteries or extracranial arteries were not evaluated, then the evaluation was considered insufficient to grade atherosclerosis, and the patient was coded A9. A similar phenotyping system was used for small vessel disease (S),12 cardiac pathology (C), other causes (O), and dissection (D); however, the latter 3 classifications applied to few enrolled patients, particularly because known cardioembolic strokes were explicitly excluded by study entry criteria. For more details, see online-only Data Supplement.
Definition of the ESUS Subgroup
Based on baseline ASCOD assessment by investigators (Table I in the online-only Data Supplement), we applied ESUS criteria6: (1) exclusion of patients with ipsilateral stenosis ≥50% (A1; patients with aortic plaques with thrombus were included because they belong to the ESUS construct6), (2) exclusion of patients with ipsilateral small deep infarct <15 mm (S1 and S2), and (3) exclusion of patients with cardiac source of embolism (C1) or with other definite cause (O1) or dissection (D1). A 24-hour cardiac rhythm monitoring to exclude atrial fibrillation could not be completed before the decision to enroll the patient in SOCRATES, which makes our ESUS subgroup slightly contaminated by undetected atrial fibrillation. These patients were kept in the analysis. We also excluded patients randomized with a TIA because the ESUS construct requires brain imaging evidence of infarction. Patients who had no check for extracranial or intracranial stenosis (A9) were also excluded. The degree of extracranial and intracranial stenosis was not adjudicated and was reported by investigators using the methods used in clinical practice for ASCOD grading (any narrowing of the arterial lumen >3 mm). Data from ultrasound/Doppler, computed tomography, magnetic resonance, or x-ray angiography was accepted.
Primary End Point
The primary end point was a composite of stroke, MI, and death evaluated at 90 days postrandomization. The first secondary end point was ischemic stroke. Among other secondary end points were all strokes and net clinical benefit (stroke, MI, death, or life-threatening bleeding).
Statistical Analysis
All analyses were based on the intention-to-treat principle and considered exploratory and hypothesis generating. The population of primary interest was the ESUS group. The second population of interest was the ESUS subgroups: patients with ipsilateral stenosis <50% or aortic arch atherosclerosis (ESUS subclinical atherosclerosis) compared with patients without these conditions (ESUS other). The time from randomization to the first occurrence of any event for a given end point was compared using a Cox proportional hazards model. Interactions between treatment assignment and ESUS characteristics were evaluated by including terms for treatment, subgroup, and treatment-by-subgroup interaction in the Cox model.
Results
Among 13 199 patients enrolled between January 7, 2014, and October 29, 2015, at 674 sites in 33 countries, 6589 were assigned to ticagrelor 90 mg twice daily and 6610 to aspirin 100 mg once daily. Based on ASCOD assessment of the SOCRATES cohort (Table I in the online-only Data Supplement), we identified 4329 patients fulfilling the definition of ESUS (Figure 1). Table 1 shows the baseline characteristics of patients on ticagrelor or aspirin according to ESUS or non ESUS groups. Because before inclusion in the trial a 24-hour holter monitoring could not been performed, 100 (2.3%) ESUS patients (49 in the ticagrelor group and 58 in the aspirin group) and 238 (2.7%) non-ESUS patients (137 in the ticagrelor group and 101 in the aspirin group) had atrial fibrillation diagnosed with the 90-day follow-up.
Baseline Characteristics of Patients With ESUS and Patients With Non-ESUS
Embolic stroke of unknown source (ESUS) construct from the SOCRATES cohort. The ESUS construct basically excludes (1) patients with ipsilateral atherosclerotic stenosis ≥50% of extra- or intracranial arteries, (2) patients with ipsilateral symptomatic small deep infarct <15 mm in diameter, (3) patients with covert cardiac source of embolism, (4) other rare definite causes or arterial dissection and also it mandates that extra- and intracranial arteries should be checked for presence of atherosclerotic stenosis and that a transthoracic echocardiography be performed. *See Table I in the online-only Data Supplement.
Figure 2 shows Kaplan–Meier curves among patients with ESUS or non-ESUS, assigned to ticagrelor or aspirin. The primary end point occurred in 126/2126 (5.9%) ESUS patients on ticagrelor and in 150/2203 (6.8%) patients on aspirin (hazard ratio [HR], 0.87; 95% confidence interval [CI], 0.68–1.10; P=0.24), whereas in the non-ESUS group (N=8870), the primary end point occurred in 7.1% of patients on ticagrelor and 7.9% of patients on aspirin (HR, 0.89; 95% CI, 0.77–1.04; P=0.15). There was no evidence for a treatment-by-ESUS group interaction (P=0.83). None of the secondary end points showed significant heterogeneity regarding treatment effect across ESUS and non-ESUS groups (Table 2). In a sensitivity analysis, exclusion of patients with TIA from the Non-ESUS group (the ESUS definition requires the presence of brain infarction) did not change the results (Table II in the online-only Data Supplement).
Outcomes in Patients With ESUS or in Patients With Non-ESUS on Ticagrelor or Aspirin*
Embolic stroke of unknown source (ESUS) and non-ESUS groups. A, Probability of the primary outcome: stroke, myocardial infarction, or death in ticagrelor and aspirin groups; B, probability of ischemic stroke in ticagrelor and aspirin groups. ASA indicates acetylsalicylic acid (ie, aspirin); CI, confidence interval; HR, hazard ratio; and %KM, Kaplan–Meier percentage.
Within the patient group with ESUS, there was evidence of a treatment interaction according to the presence of proven atherosclerotic disease as a possible mechanism of stroke (P for heterogeneity =0.04). There were 961 patients classified as ESUS subclinical atherosclerosis (with ipsilateral atherosclerotic stenosis of extracranial or intracranial artery <50% or severe aortic arch atherosclerotic plaque) and 3368 as ESUS other. A primary end point occurred in 3.7% in the ticagrelor group and 7.0% in the aspirin group (HR, 0.51; 95% CI, 0.29–0.90; P=0.02) in ESUS subclinical atherosclerosis patients; in contrast, the primary outcome occurred in 6.6% in the ticagrelor group and 6.7% in the aspirin groups (HR, 0.98; 95% CI, 0.76–1.27; P=0.89) among ESUS other patients (Figure 3).
Effect of ticagrelor and aspirin in embolic stroke of unknown source (ESUS) and non-ESUS groups and in ESUS subcategories. CI indicates confidence interval; and HR, hazard ratio.
Discussion
Consistent with overall result of the trial,11 we found that ticagrelor was not superior to aspirin in reducing the risk of stroke, MI, and death among patients who fulfilled the criteria for ESUS. There was no significant treatment-by-ESUS group interaction. However, we found that within the ESUS patients, a subgroup with atherosclerotic disease showed a 49% relative reduction in recurrence with ticagrelor compared with aspirin alone (HR, 0.51; 95% CI, 0.29–0.90; P=0.02) and only 2% relative reduction in the rest of ESUS group (P for heterogeneity =0.04). Atherosclerotic disease meant ipsilateral stenosis <50% or aortic arch atherosclerosis. This suggests that nonstenotic atherosclerosis (stenosis <50%) or aortic arch atherosclerosis may be causative of ESUS and responsive to antiplatelet treatment more potent than aspirin.
ESUS patients seem to be a group of patients with heterogeneous causes that include ipsilateral atherosclerotic stenosis of cerebral arteries <50% (intracranial or extracranial) or severe aortic arch atherothrombotic disease, cardiac abnormalities deemed to be at minor or moderate risk (eg, patent foramen ovale, atrial septal aneurysm, left ventricle ejection fraction >35%, enlarged atrium >20 cm2, spontaneous echo-contrast of the left atrium, burst of atrial fibrillation lasting <6 minutes), and undetected paroxysmal atrial fibrillation.2,3,6,12,14,15 The overlap between pathologies underlying cryptogenic stroke is large.16 Large (>3 mm thick but <50% stenosis) atherosclerotic plaques were considerably more common ipsilateral than contralateral in ESUS patients, suggesting a causal role,17 as was the case in cryptogenic stroke patients with aortic arch atherosclerosis >4 mm thick.4,5,18 As shown in Figure 2A, the overall ESUS group, including the ESUS subgroup with atherosclerotic disease, has a lower 90-day risk than non-ESUS group, as it was already observed for cryptogenic stroke in several cohorts.19–21
Our result suggests that among patients fulfilling the criteria for ESUS,6 there may have been a different treatment response according to the underlying pathology, particularly in the subgroup with substantial atherosclerotic disease, known to be at high risk of recurrence.18 As we have previously reported, this subgroup might benefit more from ticagrelor than from aspirin treatment.22 This finding would need to be confirmed in a randomized trial comparing ticagrelor and aspirin in ESUS subclinical atherosclerosis group.
The strength of our analysis is that the ASCOD grading system12 was included in the study design a priori. Using the ASCOD system, investigators described the underlying pathology (Table I in the online-only Data Supplement). We could, thus, select patients fulfilling the criteria for ESUS. The limitation was that 3194 patients were classified as A9 (unknown atherosclerotic disease status) meaning that intracranial or extracranial arteries have not been assessed for the presence of ipsilateral stenosis, which is mandatory in the ESUS construct. A 24-hour holter ECG monitoring has to be performed to exclude atrial fibrillation before diagnosing ESUS, which could not be done prior to randomization in the SOCRATES trial, given its specific time window for enrollment. Hence, during the 90-day follow-up, 2.3% of patients in the ESUS group and 2.7% of patients in the non-ESUS group were diagnosed with atrial fibrillation, which contaminated a little our ESUS group subcategorization. We also did not mandate systematically performing a 24-hour holter ECG monitoring after inclusion. One other limitation was that transthoracic echocardiography was not mandatory in the SOCRATES trial; however, because investigators had to exclude patients with known cardiac source of embolism, it is unlikely that many patients with cardiac structural abnormalities or thrombus were included in the trial.
In conclusion, in this exploratory analysis, ticagrelor was not superior to aspirin in preventing stroke, MI, and death at 90 days among ESUS patients with minor ischemic stroke. However, we found a significant 49% relative risk reduction with ticagrelor relative to aspirin in ESUS patients with substantial arterial atherosclerosis (ipsilateral carotid stenosis <50% or aortic arch atherothrombosis), suggesting that treatment response may vary across ESUS subgroups.
Sources of Funding
The study was supported by AstraZeneca.
Disclosures
Dr Amarenco reports receipt of research grant support and lecture fees from Pfizer, Sanofi, Bristol-Myers-Squibb, Merck, AstraZeneca, and Boehringer-Ingelheim, consultancy fees from Pfizer, Bristol-Myers-Squibb, Merck, Boehringer-Ingelheim, AstraZeneca, Bayer, Daiichi-Sankyo, Lundbeck, Edwards, Boston Scientific, Kowa, GSK, and Fibrogen, lecture fees from Bayer, Boston Scientific, and St-Jude Medical, and research grants from the French government. Dr Albers reports equity interest from iSchemaView and consultant fees from Lundbeck, Covidien, Johnson and Johnson, Biogen, and Astra Zeneca. Dr Denison reports being full-time employee at AstraZeneca R&D. Dr Easton receives research grant support from AstraZeneca for the SOCRATES trial (NCT01994720) and from NIH/NINDS/Sanofi as Co-PI on the POINT trial (U01 NS062835-01A1); POINT received free study drug and placebo from Sanofi (NCT00991029). He is a consultant to Boehringer-Ingelheim and Bristol-Myers Squibb for the planning and conduct of the RE-SPECT ESUS trial (NCT02239120) and the phase 2 PARFAIT trial, respectively. Dr Evans is a statistical consultant to AstraZeneca. Dr Held reports being full-time employee at AstraZeneca R&D. Dr Hill reports receipt of research grant support to the University of Calgary from Medtronic LLC, Bayer Canada, Boehringer-Ingelheim Canada, and Hoffmann-LaRoche Canada; research grant support from Alberta Innovates Health Solutions, Canadian Institutes for Health Research, and the Heart & Stroke Foundation of Canada; speaker fees and honoraria from Bristol-Myers-Squibb-Pfizer, Boehringer-Ingelheim, Bayer, and Medtronic. Dr Jonasson reports being full-time employee at AstraZeneca R&D. Dr Kasner reports research grant support from AstraZeneca, Bayer, Bristol-Myers-Squibb, WL Gore, and Acorda; consultancy fees from Merck, Boehringer-Ingelheim, Daiichi-Sankyo, Abbvie, and Johnson & Johnson. Dr Ladenvall reports being an employee at AstraZeneca R&D. Dr Minematsu reports honoraria for oral presentations from Bayer, Otsuka, Boehringer-Ingelheim, AstraZeneca, Pfizer, Mitsubishi Tanabe, Japan Stryker, Kowa, Nihon Medi-Physics, Bristol-Myers-Squibb, Sawai, Sumitomo Dainippon, Medico’s Hirata, Daiichi-Sankyo, Astellas, Kyowa Hakko Kirin, Sanofi, MSD, Eisai, and Towa. Dr Molina serves in the Steering Committee of CLOTBUST-ER trial (Cerevast); SOCRATES (AstraZeneca), IMPACT-24b (Brainsgate), REVASCAT (Fundació Ictus Malaltia Vascular). He has received honoraria for participation in clinical trials, contribution to advisory boards, or oral presentations from Astra Zeneca: Boehringer-Ingelheim, Daiichi-Sankyo, Bristol-Myers-Squibb, Covidien, Cerevast, Brainsgate. Dr Molina has no ownership interest and does not own stocks of any pharmaceutical or medical device company. Dr Wang reports research grant support from AstraZeneca. Dr Wong reports honoraria as a member of a steering committee for Johnson & Johnson, Astra Zeneca, and Bayer; honoraria for participation in clinical trials, contributions to advisory boards, or oral presentations from Bayer, Sanofi-Aventis, Bristol-Myers Squibb, Boehringer-Ingelheim, and Pfizer. Dr Johnston reports acting as a consultant to AstraZeneca during the planning of the trial, and now his institution has received research support for its conduct.
Footnotes
Guest Editor for this article was Seemant Chaturvedi, MD.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.117.017217/-/DC1.
- Received March 1, 2017.
- Revision received April 28, 2017.
- Accepted May 25, 2017.
- © 2017 American Heart Association, Inc.
References
Jump to
This Issue
Article Tools
Share this Article
- Ticagrelor Versus Aspirin in Acute Embolic Stroke of Undetermined SourcePermalink:
