Effect of Long-term Continuous Cardiac Monitoring vs Usual Care on Detection of Atrial Fibrillation in Patients With Stroke Attributed to Large- or Small-Vessel Disease: The STROKE-AF Randomized Clinical Trial | Atrial Fibrillation | JAMA | JAMA Network
[Skip to Navigation]
Figure 1.  Patient Flow in a Study of Long-term Continuous Cardiac Monitoring vs Usual Care on Detection of Atrial Fibrillation in Patients With Stroke Attributed to Large- or Small-Vessel Disease
Patient Flow in a Study of Long-term Continuous Cardiac Monitoring vs Usual Care on Detection of Atrial Fibrillation in Patients With Stroke Attributed to Large- or Small-Vessel Disease

ICM indicates insertable cardiac monitor.

aSites were not required to provide screening logs during the recruitment phase; therefore, the number of patients initially screened and reasons preventing their enrollment are not available.

Figure 2.  Time to First Detection of Atrial Fibrillation at 12 Months in a Study of Long-term Continuous Cardiac Monitoring vs Usual Care on Detection of Atrial Fibrillation (AF) in Patients With Stroke Attributed to Large- or Small-Vessel Disease
Time to First Detection of Atrial Fibrillation at 12 Months in a Study of Long-term Continuous Cardiac Monitoring vs Usual Care on Detection of Atrial Fibrillation (AF) in Patients With Stroke Attributed to Large- or Small-Vessel Disease

At 6 months, the AF incidence was 7.9% in the ICM group vs 0.8% in the control group (hazard ratio, 9.9 [95% CI, 2.3-43.5]; P < .001). The median (interquartile range) time from randomization to AF detection was 99 (36.0-235.0) days for the insertable cardiac monitor (ICM) group and 181 (86.0-231.0) days for the control group. The median (interquartile range) observation time was 365 (365-365) days for all randomized patients for each group.

Figure 3.  Time to First Detection of Atrial Fibrillation at 12 Months in a Study of Long-term Continuous Cardiac Monitoring vs Usual Care in Patients With Stroke Attributed to Large- or Small-Vessel Disease
Time to First Detection of Atrial Fibrillation at 12 Months in a Study of Long-term Continuous Cardiac Monitoring vs Usual Care in Patients With Stroke Attributed to Large- or Small-Vessel Disease

The median (interquartile range) observation time was 365 (365-365) days for all randomized patients for each group. ICM indicates insertable cardiac monitor.

Figure 4.  Rate of Recurrent Stroke at 12 Months in a Study of Long-term Continuous Cardiac Monitoring vs Usual Care on Detection of Atrial Fibrillation in Patients With Stroke Attributed to Large- or Small-Vessel Disease
Rate of Recurrent Stroke at 12 Months in a Study of Long-term Continuous Cardiac Monitoring vs Usual Care on Detection of Atrial Fibrillation in Patients With Stroke Attributed to Large- or Small-Vessel Disease

The median (interquartile range) observation time was 365 (365-365) days for all randomized patients for each group. ICM indicates insertable cardiac monitor.

Table.  Baseline Characteristics of Participants in a Study of Long-term Continuous Cardiac Monitoring vs Usual Care on Detection of Atrial Fibrillation in Patients With Stroke Attributed to Large- or Small-Vessel Disease
Baseline Characteristics of Participants in a Study of Long-term Continuous Cardiac Monitoring vs Usual Care on Detection of Atrial Fibrillation in Patients With Stroke Attributed to Large- or Small-Vessel Disease
1.
Virani  SS, Alonso  A, Benjamin  EJ,  et al; American Heart Association Council on Epidemiology and Prevention Statistics Committee and Stroke Statistics Subcommittee.  Heart disease and stroke statistics-2020 update: a report from the American Heart Association.   Circulation. 2020;141(9):e139-e596. doi:10.1161/CIR.0000000000000757PubMedGoogle ScholarCrossref
2.
Kernan  WN, Ovbiagele  B, Black  HR,  et al; American Heart Association Stroke Council, Council on Cardiovascular and Stroke Nursing, Council on Clinical Cardiology, and Council on Peripheral Vascular Disease.  Guidelines for the prevention of stroke in patients with stroke and transient ischemic attack: a guideline for healthcare professionals from the American Heart Association/American Stroke Association.   Stroke. 2014;45(7):2160-2236. doi:10.1161/STR.0000000000000024PubMedGoogle ScholarCrossref
3.
Sanna  T, Diener  HC, Passman  RS,  et al; CRYSTAL-AF Investigators.  Cryptogenic stroke and underlying atrial fibrillation.   N Engl J Med. 2014;370(26):2478-2486. doi:10.1056/NEJMoa1313600PubMedGoogle ScholarCrossref
4.
Wolf  PA, Abbott  RD, Kannel  WB.  Atrial fibrillation as an independent risk factor for stroke: the Framingham study.   Stroke. 1991;22(8):983-988. doi:10.1161/01.STR.22.8.983PubMedGoogle ScholarCrossref
5.
January  CT, Wann  LS, Calkins  H,  et al.  2019 AHA/ACC/HRS focused update of the 2014 AHA/ACC/HRS Guideline for the Management of Patients With Atrial Fibrillation: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines and the Heart Rhythm Society.   J Am Coll Cardiol. 2019;74(1):104-132. doi:10.1016/j.jacc.2019.01.011PubMedGoogle ScholarCrossref
6.
Sposato  LA, Cipriano  LE, Saposnik  G, Ruíz Vargas  E, Riccio  PM, Hachinski  V.  Diagnosis of atrial fibrillation after stroke and transient ischaemic attack: a systematic review and meta-analysis.   Lancet Neurol. 2015;14(4):377-387. doi:10.1016/S1474-4422(15)70027-XPubMedGoogle ScholarCrossref
7.
Katz  JM, Eng  MS, Carrazco  C,  et al.  Occult paroxysmal atrial fibrillation in non-cryptogenic ischemic stroke.   J Neurol. 2018;265(10):2237-2242. doi:10.1007/s00415-018-8982-9PubMedGoogle ScholarCrossref
8.
Rabinstein  AA, Fugate  JE, Mandrekar  J,  et al.  Paroxysmal atrial fibrillation in cryptogenic stroke: a case-control study.   J Stroke Cerebrovasc Dis. 2013;22(8):1405-1411. doi:10.1016/j.jstrokecerebrovasdis.2013.05.013PubMedGoogle ScholarCrossref
9.
Rizos  T, Güntner  J, Jenetzky  E,  et al.  Continuous stroke unit electrocardiographic monitoring versus 24-hour Holter electrocardiography for detection of paroxysmal atrial fibrillation after stroke.   Stroke. 2012;43(10):2689-2694. doi:10.1161/STROKEAHA.112.654954PubMedGoogle ScholarCrossref
10.
Wachter  R, Gröschel  K, Gelbrich  G,  et al; Find-AFrandomised Investigators and Coordinators.  Holter-electrocardiogram-monitoring in patients with acute ischaemic stroke (Find-AFRANDOMISED): an open-label randomised controlled trial.   Lancet Neurol. 2017;16(4):282-290. doi:10.1016/S1474-4422(17)30002-9PubMedGoogle ScholarCrossref
11.
Bernstein  RA, Kamel  H, Granger  CB, Kowal  RC, Ziegler  PD, Schwamm  LH.  Stroke of known cause and underlying atrial fibrillation (STROKE-AF) randomized trial: design and rationale.   Am Heart J. 2017;190:19-24. doi:10.1016/j.ahj.2017.04.007PubMedGoogle ScholarCrossref
12.
Adams  HP  Jr, Bendixen  BH, Kappelle  LJ,  et al.  Classification of subtype of acute ischemic stroke: definitions for use in a multicenter clinical trial: TOAST: Trial of Org 10172 in Acute Stroke Treatment.   Stroke. 1993;24(1):35-41. doi:10.1161/01.STR.24.1.35PubMedGoogle ScholarCrossref
13.
Calkins  H, Hindricks  G, Cappato  R,  et al.  2017 HRS/EHRA/ECAS/APHRS/SOLAECE expert consensus statement on catheter and surgical ablation of atrial fibrillation: executive summary.   Heart Rhythm. 2017;14(10):e445-e494. doi:10.1016/j.hrthm.2017.07.009PubMedGoogle ScholarCrossref
14.
Friberg  L, Rosenqvist  M, Lip  GY.  Evaluation of risk stratification schemes for ischaemic stroke and bleeding in 182 678 patients with atrial fibrillation: the Swedish Atrial Fibrillation cohort study.   Eur Heart J. 2012;33(12):1500-1510. doi:10.1093/eurheartj/ehr488PubMedGoogle ScholarCrossref
15.
Buck  BH, Hill  MD, Quinn  FR,  et al. PERDIEM post-embolic rhythm detection with implantable versus external monitoring. Oral presentation at: 5th European Stroke Organisation Conference; May 23, 2019; Italy, Milan.
16.
Kirchhof  P, Benussi  S, Kotecha  D,  et al; ESC Scientific Document Group.  2016 ESC guidelines for the management of atrial fibrillation developed in collaboration with EACTS.   Eur Heart J. 2016;37(38):2893-2962. doi:10.1093/eurheartj/ehw210PubMedGoogle ScholarCrossref
17.
Implantable cardiac monitors to detect atrial fibrillation after cryptogenic stroke. National Institute for Health and Care Excellence. Published September 2, 2020. Accessed November 11, 2020. https://www.nice.org.uk/guidance/dg41
18.
Diener  HC, Sacco  RL, Easton  JD,  et al; RE-SPECT ESUS Steering Committee and Investigators.  Dabigatran for prevention of stroke after embolic stroke of undetermined source.   N Engl J Med. 2019;380(20):1906-1917. doi:10.1056/NEJMoa1813959PubMedGoogle ScholarCrossref
19.
Hart  RG, Connolly  SJ, Mundl  H.  Rivaroxaban for stroke prevention after embolic stroke of undetermined source.   N Engl J Med. 2018;379(10):987. doi:10.1056/NEJMoa1802686PubMedGoogle Scholar
20.
Fisher  CM.  Lacunar strokes and infarcts: a review.   Neurology. 1982;32(8):871-876. doi:10.1212/WNL.32.8.871PubMedGoogle ScholarCrossref
21.
Benavente  OR, Coffey  CS, Conwit  R,  et al; SPS3 Study Group.  Blood-pressure targets in patients with recent lacunar stroke: the SPS3 randomised trial.   Lancet. 2013;382(9891):507-515. doi:10.1016/S0140-6736(13)60852-1PubMedGoogle ScholarCrossref
22.
Ay  H, Oliveira-Filho  J, Buonanno  FS,  et al.  Diffusion-weighted imaging identifies a subset of lacunar infarction associated with embolic source.   Stroke. 1999;30(12):2644-2650. doi:10.1161/01.STR.30.12.2644PubMedGoogle ScholarCrossref
23.
Kirchhof  P, Blank  BF, Calvert  M,  et al.  Probing oral anticoagulation in patients with atrial high rate episodes: rationale and design of the non-vitamin K antagonist oral anticoagulants in patients with atrial high rate episodes (NOAH-AFNET 6) trial.   Am Heart J. 2017;190:12-18. doi:10.1016/j.ahj.2017.04.015PubMedGoogle ScholarCrossref
24.
Lopes  RD, Alings  M, Connolly  SJ,  et al.  Rationale and design of the apixaban for the reduction of thrombo-embolism in patients with device-detected sub-clinical atrial fibrillation (ARTESiA) trial.   Am Heart J. 2017;189:137-145. doi:10.1016/j.ahj.2017.04.008PubMedGoogle ScholarCrossref
Original Investigation
June 1, 2021

Effect of Long-term Continuous Cardiac Monitoring vs Usual Care on Detection of Atrial Fibrillation in Patients With Stroke Attributed to Large- or Small-Vessel Disease: The STROKE-AF Randomized Clinical Trial

Author Affiliations
  • 1Davee Department of Neurology, Feinberg School of Medicine of Northwestern University, Chicago, Illinois
  • 2Department of Neurology, Weill Cornell Medicine, New York, New York
  • 3Duke Clinical Research Institute, Duke University Medical Center, Durham, North Carolina
  • 4Cone Health Stroke Center, Greensboro, North Carolina
  • 5Guilford Neurologic Research Associates, Greensboro, North Carolina
  • 6Department of Neurology, North Shore University Hospital, Manhasset, New York
  • 7Department of Radiology, North Shore University Hospital, Manhasset, New York
  • 8Cardiac Rhythm and Heart Failure, Medtronic Inc, Minneapolis, Minnesota
  • 9Department of Neurology, Massachusetts General Hospital, Boston
JAMA. 2021;325(21):2169-2177. doi:10.1001/jama.2021.6470
Key Points

Question  Do insertable cardiac monitors (ICMs) detect more atrial fibrillation than usual care in patients with a recent ischemic stroke attributed to large-vessel or small-vessel disease?

Findings  This randomized clinical trial included 492 patients with stroke attributed to large- or small-vessel disease evaluated with an ICM or usual care. Over 12 months, atrial fibrillation was detected in 12.1% of patients in the ICM group vs 1.8% in the usual care group, a difference that was statistically significant.

Meaning  Among patients with stroke attributed to large- or small-vessel disease, AF was more commonly detected via ICM than usual care, but whether this is of clinical importance is not addressed by the study findings.

Abstract

Importance  Patients with ischemic stroke attributed to large- or small-vessel disease are not considered at high risk for atrial fibrillation (AF), and the AF incidence rate in this population is unknown.

Objectives  To determine whether long-term cardiac monitoring is more effective than usual care for AF detection in patients with stroke attributed to large- or small-vessel disease through 12 months of follow-up.

Design, Setting, and Participants  The STROKE-AF trial was a randomized (1:1), multicenter (33 sites in the US) clinical trial that enrolled 496 patients between April 2016 and July 2019, with primary end point follow-up through August 2020. Eligible patients were aged 60 years or older or aged 50 to 59 years with at least 1 additional stroke risk factor and had an index stroke attributed to large- or small-vessel disease within 10 days prior to insertable cardiac monitor (ICM) insertion.

Interventions  Patients randomized to the intervention group (n = 242) received ICM insertion within 10 days of the index stroke; patients in the control group (n = 250) received site-specific usual care consisting of external cardiac monitoring, such as 12-lead electrocardiograms, Holter monitoring, telemetry, or event recorders.

Main Outcomes and Measures  Incident AF lasting more than 30 seconds through 12 months.

Results  Among 492 patients who were randomized (mean [SD] age, 67.1 [9.4] years; 185 [37.6%] women), 417 (84.8%) completed 12 months of follow-up. The median (interquartile range) CHA2DS2-VASc (congestive heart failure, hypertension, age ≥75 years, diabetes mellitus, stroke or transient ischemic attack, vascular disease, age 65 to 74 years, sex category) score was 5 (4-6). AF detection at 12 months was significantly higher in the ICM group vs the control group (27 patients [12.1%] vs 4 patients [1.8%]; hazard ratio, 7.4 [95% CI, 2.6-21.3]; P < .001). Among the 221 patients in the ICM group who received an ICM, 4 (1.8%) had ICM procedure–related adverse events (1 site infection, 2 incision site hemorrhages, and 1 implant site pain).

Conclusions and Relevance  Among patients with stroke attributed to large- or small-vessel disease, monitoring with an ICM compared with usual care detected significantly more AF over 12 months. However, further research is needed to understand whether identifying AF in these patients is of clinical importance.

Trial Registration  ClinicalTrials.gov Identifier: NCT02700945

Introduction

Data from 2020 indicate that the annual incidence of new or recurrent stroke in the US was approximately 795 000, of which 23% were recurrent strokes.1 Most strokes were ischemic (87%),1 and 2014 data suggests a confirmed diagnosis of atrial fibrillation (AF) was the presumed cause in 12% of cases each year,2 although this may be an underestimate because it did not account for undiagnosed AF.3 The risk of stroke was highest among patients with AF who had a stroke or a high CHA2DS2-VASc (congestive heart failure, hypertension, age ≥75 years, diabetes mellitus, stroke or transient ischemic attack [TIA], vascular disease, age 65 to 74 years, sex category) score.4 The identification of AF in patients with stroke, whether it is the cause of the index stroke, unmasked by it, or coexistent, is therefore crucial for secondary stroke prevention with guideline-recommended oral anticoagulation (OAC) therapy, although the minimum burden of AF that requires anticoagulation is unknown.5

The use of prolonged cardiac monitoring has been proven to identify AF in a substantial proportion of patients with ischemic stroke and TIA.6 Observational studies of patients with other ischemic stroke subtypes have reported on newly diagnosed subclinical AF.7-10 However, there are scarce data on the yield of heart rhythm monitoring relative to routine follow-up across ischemic stroke subtypes. In particular, the optimum duration of cardiac monitoring and the rate of AF detection remain unknown in patients with a recent ischemic stroke classified as being due to a noncardioembolic etiology.

The Stroke of Known Cause and Underlying Atrial Fibrillation (STROKE-AF) trial assessed whether long-term cardiac monitoring with an insertable cardiac monitor (ICM) was more effective than usual poststroke care for AF detection in patients with stroke attributed to cervical or intracranial large-artery atherosclerosis or small-vessel occlusion (ie, large- or small-vessel disease). The hypothesis was that ICMs would detect more AF than usual care in this subset of patients with stroke who do not normally undergo prolonged cardiac monitoring.

Methods
Study Design

This study was a multicenter, randomized (1:1), parallel-group clinical trial comparing the rate of AF detection with an ICM vs a control group receiving usual care in patients with an index ischemic stroke classified by the enrolling investigator as being due to large-artery atherosclerosis (large-vessel disease) or small-vessel occlusion (small-vessel disease). The study was conducted in compliance with international ethical and scientific quality standards and the principles of the Declaration of Helsinki. The study design has been previously reported,11 and the trial protocol and statistical analysis plan are available in Supplement 1 and Supplement 2. The protocol was approved by all relevant institutional review boards and all patients provided written informed consent prior to initiation of any study-specific procedures. All protocol modifications made after study initiation are listed in eTable 1 in Supplement 3.

Participants

Patients were admitted to high-performing stroke centers where the treating physicians routinely categorize stroke types using the Trial of Org 10172 in Acute Stroke Treatment (TOAST) criteria.12 Investigators were instructed to assign stroke mechanism using the TOAST criteria as applied in clinical practice. A prespecified research instrument for TOAST classification was not required. Patients were eligible if they had an index stroke attributed by the investigators to large- or small-vessel disease within 10 days prior to ICM insertion. Patients were aged 60 years or older or aged 50 to 59 years with a documented medical history of at least 1 of the following additional stroke risk factors: congestive heart failure, hypertension, diabetes, ischemic stroke more than 90 days before the index stroke, or other ischemic vascular disease. Patients were excluded if they had cryptogenic stroke or embolic stroke of undetermined source, cardioembolic stroke, a history of documented AF or atrial flutter, or a known indication or contraindication for long-term OAC. To avoid confounding factors, patients were excluded if they had a medical history of untreated hyperthyroidism, myocardial infarction or cardiac surgery less than 1 month prior to index stroke, or mechanical heart valve or valvular disease requiring immediate surgery.

Randomization

Patients were randomized in a 1:1 ratio to the ICM or control group. SAS, version 9.4 (SAS Institute Inc), code was developed to generate an output of site-specific randomization schedules. The schedules consisted of randomly permuted blocks of 2 and 4. The block permutation kept the balance of participants receiving each group assignment and helped prevent the next group assignment from being guessed by site personnel. Sites were limited to 50 enrollments to reduce site-specific bias. The patients, investigative site personnel, and Medtronic personnel were blinded to the overall randomization sequence. However, due to the nature of the intervention, blinding on an individual patient level was not feasible. Aggregate efficacy results and adverse events, analyzed by randomized group, remained blinded to study personnel until completion of the primary end point analysis. The clinical events committee (CEC) had visibility to group assignment.

Interventions

Patients in the ICM group received an insertable cardiac monitor (REVEAL LINQ, Medtronic) within 10 days of the index stroke and after randomization. The follow-up period began on the day of randomization. Device insertion was performed in accordance with each site’s standard procedure practice and the instructions in the device’s clinician manual. Patients in the control group received usual care specific to each participating site.

Outcomes

The primary outcome was AF detection through 12 months between study groups, which included a subgroup analysis comparing patients with index stroke attributed to large-vessel disease vs small-vessel disease. The secondary outcome was AF detection between study groups through the duration of the study (up to 36 months). This analysis will therefore be performed at study end. Post hoc analyses included the following prespecified ancillary end points planned at 36 months that were explored at 12 months: comparison of incidence rates of AF detection at 6 months between study groups; the proportion of asymptomatic AF episodes in the ICM group (no event on the patient activation recorder); identification of the longest AF episode duration per patient in the ICM group; comparison of OAC use; and the incidence rate of recurrent stroke between study groups.11

Clinical and monitoring data were collected at baseline and 1, 6, and 12 months after randomization, and will continue at 6-month intervals up to 36 months or the end of ICM battery life. Additionally, patients randomized to the ICM group had data collected at device insertion and at 3 and 9 months. ICMs were programmed according to standardized requirements (eTable 2 in Supplement 3).

AF was defined as an episode of irregular heart rhythm, without detectable P waves, lasting more than 30 seconds and adjudicated by the CEC.13 In the control group, electrocardiogram or other cardiac rhythm monitoring was performed at the discretion of the treating physician. If an AF episode longer than 30 seconds was detected, source documentation was reviewed by the CEC. However, due to the ICM’s automatic detection algorithms, all detected episodes in the ICM group were at least 2 minutes in duration. Investigators were instructed to ascertain whether patients experienced a recurrent stroke or TIA at each follow-up visit and provide supporting details (eTable 3 in Supplement 3). All recurrent TIAs and strokes were determined by the stroke centers and confirmed by the CEC. Recurrent stroke was defined as any hemorrhagic or ischemic event with rapid onset of a focal or global neurological deficit or other neurological signs or symptoms consistent with stroke. TIA was defined as any new focal neurological deficit with rapid symptom resolution (usually 1-2 hours; always within 24 hours) and without tissue injury (based on neuroimaging). Stroke severity was assessed using the National Institutes of Health Stroke Scale.

Sample Size Calculation

The primary objective was to compare the incidence rates of AF detection through 12 months of follow-up between the ICM and control groups. A total of 23 AF events within 12 months of follow-up were required to demonstrate a statistically significant difference (P < .05) between the 2 groups with 85% power. This calculation was based on the assumptions that the AF incidence rate would be 8% in the ICM group and 2% in the control group (hazard ratio [HR] of 4.13), annual attrition rate of no greater than 10%, and annual crossover rate of 5%. The assumption for HR of of 4.13 was based on the results of previous studies, such as the CRYSTAL-AF trial.3 Due to population differences, we expected a lower rate of AF in the current trial than in the CRYSTAL-AF study. The sample size was estimated to require 486 randomized patients. To account for an assumed 2% attrition rate from enrollment to randomization, it was estimated that approximately 496 patients would be enrolled in the study.

Statistical Analysis

Statistical analyses were performed using SAS, version 9.4, and R, version 3.6.0 (R Foundation for Statistical Computing). The analysis set consisted of all randomized patients (Figure 1). Patients were analyzed according to their randomization group. Survival estimates are reported for each group as well as an HR estimate for the effect with corresponding 2-sided 95% CIs. The Cox models analyzed time to first AF episode through 12 months. Prespecified subgroup analyses of patients by stroke subtype (large-vessel disease and small-vessel disease) were performed. A Cox proportional hazards model was used to test the interaction term between group and subgroup of stroke subtype. The proportional hazards assumption of the Cox models was assessed by examining the group by log(time) interaction term; the interaction term was not significant in each model (P > .28) and, consequently, the proportional hazard assumption was met. Because time-to-event methods were used, there were missing data only for participants with no follow-up time from randomization. Otherwise, each participant had data to contribute to the analyses. To evaluate whether the effect of ICM monitoring varied among the enrolling sites, a post hoc mixed-effects model analysis of the primary outcome that treated study site as a random effect was performed. Because of the potential for type I error due to multiple comparisons, findings for analyses of post hoc end points should be interpreted as exploratory. Statistical significance was set at a 2-sided P value of .05 for all analyses.

Results
Study Population

Between April 2016 and July 2019, a total of 496 patients were enrolled at 33 sites in the US, of whom 492 were randomly assigned to the ICM (n = 242) or the control group (n = 250). All randomized patients were analyzed according to their randomization group (Figure 1). The median (interquartile range [IQR]) time from index stroke to randomization was 4 (2-6) days. Crossover occurred in 14 of 492 randomized patients: 6 assigned to receive an ICM did not and 8 assigned to receive usual care received an ICM. Device insertion was successful in all 223 patients who received an ICM; 219 (98.2%) underwent the procedure 10 days or less after randomization, with median (IQR) time from index stroke to ICM insertion of 5 (3-8) days. All patients were assessed for symptoms at each 6-month follow-up visit. Follow-up continued through August 2020 (from randomization to 12 months) for a mean (SD) duration of 331.4 (90.9) days.

In total, 417 patients (84.8%) completed 12 months of follow-up: 210 in the ICM group and 207 in the control group. Of the 17 patients who died, 2 met the primary end point prior to death (both in the ICM group). Of the 58 patients who left the study early, 1 in the ICM group met the primary end point prior to exiting. Reasons for incomplete follow-up are shown in Figure 1.

Compared with patients who completed the 12-month follow-up visit, baseline demographics of patients with incomplete follow-up (n = 75) were similar except for more frequent congestive heart failure (P = .04) (eTable 4 in Supplement 3). The Table shows baseline demographics and clinical characteristics for each group. The mean (SD) age of the population was 67.1 (9.4) years, 185 (37.6%) were women, and the median (IQR) CHA2DS2-VASc score was 5 (4-6). eTable 5 in Supplement 3 shows baseline characteristics according to TOAST subtype. Major echocardiographic features were similar between study groups (eTable 6 in Supplement 3).

Primary End Point

The incidence of AF at 12 months was 12.1% (n = 27) in the ICM group vs 1.8% (n = 4) in the control group (HR, 7.4 [95% CI, 2.6-21.3]; P < .001) (Figure 2). The median (IQR) time from randomization to AF detection was 99 (36-235) days for the ICM group and 181 (86-231) days for the control group. AF episodes detected in the control group resulted from a total of 76 electrocardiograms in 57 patients, 26 Holter monitors/event recorders in 25 patients, and 2 mobile cardiac telemetry devices in 2 patients.

Two additional patients in the control group had mention of possible AF episodes with insufficient evidence for the CEC to adjudicate these as confirmed AF. Results of a sensitivity analysis that added these 2 AF episodes to the control group still showed a significantly increased detection incidence of AF in the ICM group compared with the control group (12.1% [n = 27] vs 2.6% [n = 6]; HR, 4.9 [95% CI, 2.0-11.9]; P < .001). In the ICM group, all AF episodes contributing to the primary end point were detected by the ICM.

A subgroup analysis was performed to compare the detection incidence of AF between patients with an index stroke classified as large-vessel disease (284 patients [57.3%]) and small-vessel disease (208 patients [42.3%]) (Table). A Cox model that included a term for the interaction between stroke subtype and group showed that the interaction term as well as stroke subtype were not significant (P = .42 and P = .46), and only the group effect was significant (P = .009). AF incidence at 12 months was significantly higher in the ICM group vs the control group in patients with large-vessel (15 [11.7%] vs 3 [2.3%]; HR, 5.3 95% CI, 1.5-18.2]; P < .001) and small-vessel (12 [12.6%] vs 1 [1.0%]; HR, 13.8 [95% CI, 1.8-111.1]; P < .001) disease (Figure 3). Among patients in the ICM group, there was no significant difference in AF detection in participants with large- vs small-vessel stroke (15 [11.7%] vs 12 [12.6%]; HR, 0.9 [95% CI, 0.4-1.9]; P = .74).

The proportional hazard assumption was checked by examining the group by log(time) interaction term. The interaction term was not significant in each model (P > .28). Consequently, the proportional hazard assumption was met. The effect of ICM monitoring was evaluated among the 33 enrolling sites with a post hoc mixed-effects model analysis of the primary end point. Sites were random effects and their effect was not significant (P = .10).

Post Hoc End Points

At 6 months, the incidence of AF was 7.9% (n = 18) in the ICM group vs 0.8% (n = 2) in the control group (HR, 9.9 [95% CI, 2.3-43.5]; P = .002). Few AF episodes were detected in the ICM group during the first 30 days of follow-up (6 [2.6%]), suggesting that 30 days of continuous monitoring (eg, mobile telemetry) would have missed 78% (21 of 27) of the patients detected by the ICM at 12 months. The first AF episode detected in the ICM group was asymptomatic in 26 of 27 patients (96.3%).

Among patients with AF in the ICM group, the median (IQR) duration for the longest single episode of AF detected was 88 (10-526) minutes, and was longer in those with small-vessel vs large-vessel stroke (267 [49-438] vs 44 [8-526] minutes). Furthermore, 15 patients (55.5%) with AF in the ICM group had an episode longer than 1 hour. A histogram showing the duration of the longest AF episode per patient is provided in eFigure 1 in Supplement 3.

At 12 months, 38 patients (15.7%) in the ICM group and 14 (5.6%) in the control group were prescribed OAC (unadjusted odds ratio, 3.1 [95% CI 1.7-6.0]; P < .001). Among patients who were prescribed OAC, 20 patients in the ICM group and 11 patients in the control group did not have AF. In patients in whom AF was detected, OAC was prescribed more often in the ICM vs the control group (18 [7.4%] vs 3 [1.2%]; unadjusted odds ratio, 6.6 [95% CI, 1.9-22.8]; P < .001). Direct oral anticoagulants were the predominant type of OAC prescribed (43 of 52 [83%]). In addition, 1 patient in the ICM group underwent left atrial appendage occlusion and no patients underwent ablation. The incidence of recurrent ischemic and hemorrhagic stroke at 12 months was 7.2% (n = 16) in the ICM group and 9.8% (n = 23) in the control group (HR, 0.7 [95% CI, 0.4-1.4]; P = .30) (Figure 4). Classification of first recurrent ischemic stroke according to TOAST subtype is shown in eTable 7 in Supplement 3. Of the 16 recurrent strokes in the ICM group, only 1 occurred in a patient who had AF detected prior to the stroke, and this patient was not prescribed anticoagulants until after the recurrent stroke. Additionally, 1 patient without AF detected prior to their recurrent stroke was prescribed OAC prior to the recurrent event. Of the 23 recurrent strokes in the control group, there were no patients with AF detected and no patients who were prescribed OAC. The rate of recurrent TIA at 12 months was 1.8% (n = 4) in the ICM group and 0.4% (n = 1) in the control group (HR, 4.1 [95% CI, 0.5-36.8]; P = .21). Hemorrhagic stroke was reported in 1 patient in each group, and frequency of recurrent strokes were similar in the large- vs small-vessel subtypes (27 [10.3%] vs 12 [6.1%]; HR 1.7 [95% CI 0.9-3.4]; P = .12).

Adverse Events

Among the 221 patients in the ICM group who received an ICM, 4 (1.8%) had insertion procedure–related adverse events: 1 had site infection (0.5%), 2 had incision site hemorrhages (0.9%), and 1 had implant site pain (0.5%).

Discussion

In this randomized clinical trial, among patients with a recent ischemic stroke attributed to large- or small-vessel disease, monitoring with an ICM resulted in a significantly higher rate of AF detection compared with usual care at 12 months.

The results presented are comparable to a small observational study of a noncryptogenic ischemic stroke cohort that found AF detected in 6 of 51 patients (12%) after monitoring with an ICM for 13 months.7 Other studies using prolonged continuous cardiac monitoring (64 hours to 21 days)8-10 in patients with acute and nonacute ischemic stroke have shown higher rates of AF detection compared with short-term monitoring. However, those cohorts included unselected patients with ischemic stroke in their reported AF detection rates, making comparisons of incidence rates between studies unreliable. The recently completed but unpublished PERDIEM randomized trial defined clinically actionable AF as lasting at least 2 minutes and compared the rates of AF at 1 year with an ICM vs 4 weeks of external loop recording in 300 adults.15 However, patients were enrolled up to 6 months after ischemic stroke and, although all ischemic subtypes were eligible, patients with cryptogenic stroke made up approximately 70% of the sample. The present study, on the other hand, provided evidence of the higher yield of AF detection by ICM in patients with noncryptogenic stroke starting within 10 days of the index stroke.

However, it is important to note that in the absence of a comparator group of patients with similar characteristics but without stroke, it is not possible from this study design to determine whether the rate of AF detection was related to the stroke or represented background asymptomatic AF that would have been detected at a similar rate in the absence of stroke. This study also did not address whether identifying AF in this setting affects clinical outcomes.

The CRYSTAL-AF study showed a 12.4% AF detection rate using ICMs in the first year after cryptogenic stroke,3 a stroke subtype in which paroxysmal AF was strongly suspected. Results of that study led to guideline recommendations for long-term cardiac monitoring in cryptogenic stroke.5,16 Monitoring is now frequently performed in patients with embolic stroke of undetermined source,17 because empirical anticoagulation in these patients in the absence of ICM-detected AF has not been proven effective.18,19

In contrast to those trials, the present study recruited only patients with stroke attributed to large- or small-vessel disease, a population of patients with ischemic stroke for whom long-term cardiac monitoring data are sparse. Because these strokes are generally thought to be due to atherosclerosis or lipohyalinosis,20 long-term cardiac monitoring is neither currently recommended nor routinely performed. However, many recurrent strokes do not have the same mechanism as the initial stroke,21 so patients with an index stroke due to large- or small-vessel disease can still be at risk for future AF-related stroke, especially if they are only treated with antiplatelet therapy.

Although some of the index strokes in this study may have been due to undetected paroxysmal AF and mistakenly attributed to large- or small-vessel disease, this could not be determined. Prior studies have demonstrated that up to 1 in 6 patients presenting with classic lacunar syndromes have multiple infarctions demonstrated on diffusion-weighted imaging, consistent with a proximal embolic source.22 The importance of detecting AF is not to ascribe an alternative etiology to the index stroke, but rather that it represents a risk for future cardioembolic stroke, which may be more effectively prevented by OAC than antiplatelet therapy.

The minimum duration of AF needed to benefit from long-term OAC remains unknown, but, in this study, a post hoc analysis identifying the longest AF episode at 12 months showed that AF events frequently lasted more than 1 hour and all patients had a minimum CHA2DS2-VASc score of 2 (prior stroke). For patients with a clinical diagnosis of AF, the overall median CHA2DS2-VASc score of 5 observed in this population corresponds to an estimated stroke risk of 7.2% per year.14 Ongoing studies may shed light on the clinical significance of ICM-detected moderate-burden AF for primary stroke prevention.23,24

Limitations

This study has several limitations. First, the attribution of stroke mechanism is subjective and may have led to the enrollment of a population at higher risk of underlying embolism; however, this attribution reflects real-world practice and the sample in this trial is representative of patients for whom ICM decisions need to be made. Second, due to study design, neither patients, physicians, nor the committee members who adjudicated AF events or recurrent stroke events were blinded to randomization assignment. However, the primary end point of AF detection was adjudicated using objective criteria. Third, there was a higher frequency of congestive heart failure among patients who did not complete the 12-month follow-up. Given that this is a risk factor for the development of AF, the results may have underestimated the true incidence of AF. Fourth, the study was not powered to detect a significant difference in rates of recurrent stroke. Fifth, although the duration definition for AF in the study was 30 seconds, patients in the ICM group would not have their AF detected unless the episode persisted for at least 2 minutes due to the requirements of the automatic detection algorithm. However, this limitation makes estimates of the AF detection rate in the ICM group more conservative. Sixth, electrocardiogram monitoring follow-up in the control group was variable and limited, further reducing the ability to detect AF in these patients. This may reflect standard of care practice for patients with strokes of “known” origin attributed to large- and small-vessel disease in whom a cardioembolic etiology is deemed unlikely. Seventh, because TOAST classifications were applied as customary in clinical practice, these classifications may have greater interrater variation than if a validated formal TOAST classification algorithm was used. Eighth, because of the absence of a comparator group with similar characteristics but without a stroke, and the absence of clinical outcomes, this study can make no conclusions about the clinical value of such testing.

Conclusions

Among patients with stroke attributed to large- or small-vessel disease, monitoring with an ICM compared with usual care detected significantly more AF over 12 months. However, further research is needed to understand whether identifying AF in these patients is of clinical importance.

Back to top
Article Information

Corresponding Author: Richard A. Bernstein, MD, PhD, Davee Department of Neurology, Feinberg School of Medicine of Northwestern University, 211 E Ontario, Ste 905, Chicago IL 60026 (rbernste@nm.org).

Accepted for Publication: April 9, 2021.

Author Contributions: Drs Bernstein and Schwamm had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Drs Schwamm and Bernstein contributed equally to the manuscript.

Concept and design: Bernstein, Kamel, Granger, Katz, Alfaro Vives, Ziegler, Schwamm.

Acquisition, analysis, or interpretation of data: Bernstein, Kamel, Piccini, Sethi, Katz, Alfaro Vives, Ziegler, Franco, Schwamm.

Drafting of the manuscript: Bernstein, Alfaro Vives, Ziegler, Franco, Schwamm.

Critical revision of the manuscript for important intellectual content: Bernstein, Kamel, Granger, Piccini, Sethi, Katz, Alfaro Vives, Ziegler, Franco, Schwamm.

Statistical analysis: Alfaro Vives.

Obtained funding: Bernstein.

Administrative, technical, or material support: Ziegler, Franco.

Supervision: Schwamm.

Conflict of Interest Disclosures: Dr Bernstein reported receiving grants from Northwestern University during the conduct of the study and personal fees from Medtronic outside the submitted work, and performing consulting and steering committee work with Boehringer Ingelheim Pharmaceuticals related to Pradaxa, atrial fibrillation, and cryptogenic stroke and consulting and speaking for Abbott related to patent foramen ovale closure. Dr Kamel reported serving as a principal investigator for the National Institutes of Health–funded ARCADIA trial (NINDS U01NS095869), which receives in-kind study drug from the BMS-Pfizer Alliance for Eliquis and ancillary study support from Roche Diagnostics, serving as deputy editor for JAMA Neurology, serving as a steering committee member of Medtronic's Stroke-AF trial (uncompensated), and serving on an end point adjudication committee for a trial of empagliflozin for Boehringer Ingelheim. Dr Granger reported receiving personal fees from Medtronic during the conduct of the study and personal fees from Bayer, Boehringer Ingelheim, Boston Scientific, Bristol Myers Squibb, Janssen Pharmaceuticals, and Pfizer and research grants from Medtronic Foundation, Boehringer Ingelheim, Bristol Myers Squibb, Daiichi Sankyo, Janssen Pharmaceuticals, and Pfizer. Dr Piccini reported receiving grants from Abbott, Boston Scientific, Bayer, and AHA and personal fees from Biotronik, Medtronic, and Pfizer outside the submitted work. Dr Sethi reported receiving speaker fees and honorarium from Medtronic during the conduct of the study and honorarium from WL Gore outside the submitted work. Dr Katz reported receiving grants from Medtronic during the conduct of the study and grants from Siemens Healthineers and the National Institutes of Health/National Institute of Neurological Disorders and Stroke (NINDS) outside the submitted work. Dr Alfaro Vives reported being an employee of and shareholder in Medtronic outside the submitted work. Dr Ziegler reported being an employee of and shareholder in Medtronic outside the submitted work. Dr Franco reported receiving personal fees from and being an employee of and shareholder in Medtronic outside the submitted work. Dr Schwamm reported receiving grants from Medtronic as the national co–primary investigator of the Stroke-AF trial during the conduct of the study and grants from Genentech, which supplied alteplase free of charge to Massachusetts General Hospital as well as supplemental per-patient payments to participating sites in our NINDS-funded trial; grants from NINDS as the national co–primary investigator in a late-window thrombolysis trial (P50NS051343, MR WITNESS NCT01282242) and as the site primary investigator of StrokeNet Network NINDS (New England Regional Coordinating Center U24NS107243); and personal fees from Genentech as a scientific consultant regarding trial design and conduct (late window thrombolysis) and a member of steering committee (TIMELESS NCT03785678), LifeImage as a consultant on user interface design and usability for teleradiology, Penumbra as a member of a data safety monitoring board (MIND NCT03342664), and Diffusion Pharma as a member of a data and safety monitoring board (PHAST-TSC NCT03763929) outside the submitted work.

Funding/Support: This research was funded by Medtronic Inc.

Role of the Funder/Sponsor: The Steering Committee plus specific Medtronic personnel were involved in the design and conduct of the study; collection, management, and interpretation of the data; preparation, review, and approval of the manuscript; and decision to submit the manuscript for publication. Medtronic did not have the right to veto publication, and the decision regarding which journal the paper was submitted to was reached among the authors.

Group Information: The STROKE-AF Investigators are listed in Supplement 4.

Data Sharing Statement: See Supplement 5.

Additional Contributions: We thank Medtronic statistician Theodore Merriam, MS, for his contributions to the review and analysis of study data. Mr Merriam is employed by Medtronic and received compensation for his contributions.

References
1.
Virani  SS, Alonso  A, Benjamin  EJ,  et al; American Heart Association Council on Epidemiology and Prevention Statistics Committee and Stroke Statistics Subcommittee.  Heart disease and stroke statistics-2020 update: a report from the American Heart Association.   Circulation. 2020;141(9):e139-e596. doi:10.1161/CIR.0000000000000757PubMedGoogle ScholarCrossref
2.
Kernan  WN, Ovbiagele  B, Black  HR,  et al; American Heart Association Stroke Council, Council on Cardiovascular and Stroke Nursing, Council on Clinical Cardiology, and Council on Peripheral Vascular Disease.  Guidelines for the prevention of stroke in patients with stroke and transient ischemic attack: a guideline for healthcare professionals from the American Heart Association/American Stroke Association.   Stroke. 2014;45(7):2160-2236. doi:10.1161/STR.0000000000000024PubMedGoogle ScholarCrossref
3.
Sanna  T, Diener  HC, Passman  RS,  et al; CRYSTAL-AF Investigators.  Cryptogenic stroke and underlying atrial fibrillation.   N Engl J Med. 2014;370(26):2478-2486. doi:10.1056/NEJMoa1313600PubMedGoogle ScholarCrossref
4.
Wolf  PA, Abbott  RD, Kannel  WB.  Atrial fibrillation as an independent risk factor for stroke: the Framingham study.   Stroke. 1991;22(8):983-988. doi:10.1161/01.STR.22.8.983PubMedGoogle ScholarCrossref
5.
January  CT, Wann  LS, Calkins  H,  et al.  2019 AHA/ACC/HRS focused update of the 2014 AHA/ACC/HRS Guideline for the Management of Patients With Atrial Fibrillation: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines and the Heart Rhythm Society.   J Am Coll Cardiol. 2019;74(1):104-132. doi:10.1016/j.jacc.2019.01.011PubMedGoogle ScholarCrossref
6.
Sposato  LA, Cipriano  LE, Saposnik  G, Ruíz Vargas  E, Riccio  PM, Hachinski  V.  Diagnosis of atrial fibrillation after stroke and transient ischaemic attack: a systematic review and meta-analysis.   Lancet Neurol. 2015;14(4):377-387. doi:10.1016/S1474-4422(15)70027-XPubMedGoogle ScholarCrossref
7.
Katz  JM, Eng  MS, Carrazco  C,  et al.  Occult paroxysmal atrial fibrillation in non-cryptogenic ischemic stroke.   J Neurol. 2018;265(10):2237-2242. doi:10.1007/s00415-018-8982-9PubMedGoogle ScholarCrossref
8.
Rabinstein  AA, Fugate  JE, Mandrekar  J,  et al.  Paroxysmal atrial fibrillation in cryptogenic stroke: a case-control study.   J Stroke Cerebrovasc Dis. 2013;22(8):1405-1411. doi:10.1016/j.jstrokecerebrovasdis.2013.05.013PubMedGoogle ScholarCrossref
9.
Rizos  T, Güntner  J, Jenetzky  E,  et al.  Continuous stroke unit electrocardiographic monitoring versus 24-hour Holter electrocardiography for detection of paroxysmal atrial fibrillation after stroke.   Stroke. 2012;43(10):2689-2694. doi:10.1161/STROKEAHA.112.654954PubMedGoogle ScholarCrossref
10.
Wachter  R, Gröschel  K, Gelbrich  G,  et al; Find-AFrandomised Investigators and Coordinators.  Holter-electrocardiogram-monitoring in patients with acute ischaemic stroke (Find-AFRANDOMISED): an open-label randomised controlled trial.   Lancet Neurol. 2017;16(4):282-290. doi:10.1016/S1474-4422(17)30002-9PubMedGoogle ScholarCrossref
11.
Bernstein  RA, Kamel  H, Granger  CB, Kowal  RC, Ziegler  PD, Schwamm  LH.  Stroke of known cause and underlying atrial fibrillation (STROKE-AF) randomized trial: design and rationale.   Am Heart J. 2017;190:19-24. doi:10.1016/j.ahj.2017.04.007PubMedGoogle ScholarCrossref
12.
Adams  HP  Jr, Bendixen  BH, Kappelle  LJ,  et al.  Classification of subtype of acute ischemic stroke: definitions for use in a multicenter clinical trial: TOAST: Trial of Org 10172 in Acute Stroke Treatment.   Stroke. 1993;24(1):35-41. doi:10.1161/01.STR.24.1.35PubMedGoogle ScholarCrossref
13.
Calkins  H, Hindricks  G, Cappato  R,  et al.  2017 HRS/EHRA/ECAS/APHRS/SOLAECE expert consensus statement on catheter and surgical ablation of atrial fibrillation: executive summary.   Heart Rhythm. 2017;14(10):e445-e494. doi:10.1016/j.hrthm.2017.07.009PubMedGoogle ScholarCrossref
14.
Friberg  L, Rosenqvist  M, Lip  GY.  Evaluation of risk stratification schemes for ischaemic stroke and bleeding in 182 678 patients with atrial fibrillation: the Swedish Atrial Fibrillation cohort study.   Eur Heart J. 2012;33(12):1500-1510. doi:10.1093/eurheartj/ehr488PubMedGoogle ScholarCrossref
15.
Buck  BH, Hill  MD, Quinn  FR,  et al. PERDIEM post-embolic rhythm detection with implantable versus external monitoring. Oral presentation at: 5th European Stroke Organisation Conference; May 23, 2019; Italy, Milan.
16.
Kirchhof  P, Benussi  S, Kotecha  D,  et al; ESC Scientific Document Group.  2016 ESC guidelines for the management of atrial fibrillation developed in collaboration with EACTS.   Eur Heart J. 2016;37(38):2893-2962. doi:10.1093/eurheartj/ehw210PubMedGoogle ScholarCrossref
17.
Implantable cardiac monitors to detect atrial fibrillation after cryptogenic stroke. National Institute for Health and Care Excellence. Published September 2, 2020. Accessed November 11, 2020. https://www.nice.org.uk/guidance/dg41
18.
Diener  HC, Sacco  RL, Easton  JD,  et al; RE-SPECT ESUS Steering Committee and Investigators.  Dabigatran for prevention of stroke after embolic stroke of undetermined source.   N Engl J Med. 2019;380(20):1906-1917. doi:10.1056/NEJMoa1813959PubMedGoogle ScholarCrossref
19.
Hart  RG, Connolly  SJ, Mundl  H.  Rivaroxaban for stroke prevention after embolic stroke of undetermined source.   N Engl J Med. 2018;379(10):987. doi:10.1056/NEJMoa1802686PubMedGoogle Scholar
20.
Fisher  CM.  Lacunar strokes and infarcts: a review.   Neurology. 1982;32(8):871-876. doi:10.1212/WNL.32.8.871PubMedGoogle ScholarCrossref
21.
Benavente  OR, Coffey  CS, Conwit  R,  et al; SPS3 Study Group.  Blood-pressure targets in patients with recent lacunar stroke: the SPS3 randomised trial.   Lancet. 2013;382(9891):507-515. doi:10.1016/S0140-6736(13)60852-1PubMedGoogle ScholarCrossref
22.
Ay  H, Oliveira-Filho  J, Buonanno  FS,  et al.  Diffusion-weighted imaging identifies a subset of lacunar infarction associated with embolic source.   Stroke. 1999;30(12):2644-2650. doi:10.1161/01.STR.30.12.2644PubMedGoogle ScholarCrossref
23.
Kirchhof  P, Blank  BF, Calvert  M,  et al.  Probing oral anticoagulation in patients with atrial high rate episodes: rationale and design of the non-vitamin K antagonist oral anticoagulants in patients with atrial high rate episodes (NOAH-AFNET 6) trial.   Am Heart J. 2017;190:12-18. doi:10.1016/j.ahj.2017.04.015PubMedGoogle ScholarCrossref
24.
Lopes  RD, Alings  M, Connolly  SJ,  et al.  Rationale and design of the apixaban for the reduction of thrombo-embolism in patients with device-detected sub-clinical atrial fibrillation (ARTESiA) trial.   Am Heart J. 2017;189:137-145. doi:10.1016/j.ahj.2017.04.008PubMedGoogle ScholarCrossref
×