Abstract
Objective: Atrial fibrillation (AF) may be present within a subset of patients with presumed cryptogenic TIA or stroke and remains undetected by standard diagnostic methods. We hypothesized that AF may be an under-recognized mechanism for cryptogenic TIA/stroke.
Methods: A consecutive series of 56 patients with cryptogenic TIA/stroke was analyzed after diagnostic evaluation and Mobile Cardiac Outpatient Telemetry (MCOT) for up to 21 days. Demographic, radiographic, echocardiographic, and MCOT results were reviewed. Predictors of AF detection by MCOT were determined by univariate analysis including Student t test and Fisher exact tests and multivariate analysis.
Results: The median MCOT monitoring duration was 21 (range 5–21) days resulting in an AF detection rate of 23% (13/56). AF was first detected after a median of 7 (range 2–19) days of monitoring. Twenty-seven asymptomatic AF episodes were detected in the 13 patients, of which 85% (23/27) were <30 seconds and the remaining 15% (4/27) were 4–24 hours in duration. Diabetes was predictive of AF detection by both univariate (p = 0.024) and multivariate analysis (OR 6.15; 95% CI 1.16 to 32.73; p = 0.033).
Conclusions: There is a high rate of atrial fibrillation (AF) detection by Mobile Cardiac Outpatient Telemetry (21 days) in patients with cryptogenic TIA/stroke that may be related to extended monitoring duration, patient selection, and inclusion of all new onset AF episodes. Brief AF episodes (<30 seconds) may be biomarkers of more prolonged and clinically significant AF.
Glossary
- AF=
- atrial fibrillation;
- ASA=
- atrial septal aneurysm;
- AV=
- Mobitz type 2 second degree atrioventricular block;
- DWI=
- diffusion weighted imaging;
- ECG=
- electrocardiogram;
- ELR=
- event-loop recording;
- HR=
- heart rate;
- MCOT=
- Mobile Cardiac Outpatient Telemetry;
- NSVT=
- nonsustained ventricular tachycardia;
- PAC=
- premature atrial complex;
- PAF=
- paroxysmal atrial fibrillation;
- PFO=
- patent foramen ovale;
- PSVT=
- paroxysmal supraventricular tachycardia;
- TEE=
- transesophageal echocardiography;
- TOAST=
- Trial of Org 10172 in Acute Stroke Treatment;
- TTE=
- transthoracic echocardiography.
Patients experiencing a TIA/stroke frequently have no determined etiology after standard diagnostic evaluation. Previous reports show that 36% of stroke survivors are classified as cryptogenic and 19% are cardioembolic based upon Trial of Org 10172 in Acute Stroke Treatment (TOAST) classification.1–3 Patients initially diagnosed with cryptogenic stroke and TIA of undetermined etiology subsequently can be found to have atrial fibrillation (AF), suggesting that improved efforts to detect AF in this subgroup are warranted. Unrecognized AF may be present in patients with cryptogenic TIA/stroke and represent an alternative stroke etiology.
We sought to determine the percent of patients presenting with cryptogenic TIA/stroke who have unrecognized new-onset AF using a 21-day Mobile Cardiac Outpatient Telemetry (MCOT) unit.
METHODS
Patients.
This retrospective analysis was approved by the Institutional Review Board of Allegheny General Hospital. From January 2006 to May 2007, a total of 372 patients was admitted with ischemic stroke and 68 (18%) were found to be cryptogenic in etiology. Of these 68 patients, 56 (82%) met criteria for inclusion in our study and were reviewed. In this study, the term cryptogenic TIA/stroke refers to stroke of undetermined etiology based upon TOAST criteria and TIAs of undetermined etiology. Clinical, radiographic, echocardiography, and MCOT data were collected and reviewed. Inclusion criteria were 1) age greater than 18 years; 2) ischemic stroke or TIA within the last 3 months; and 3) diagnosis of cryptogenic TIA/stroke. TIA was defined as sudden-onset focal neurologic symptoms or signs that resolved within 24 hours and was not associated with high-intensity abnormality in the diffusion-weighted sequence. TIA symptoms and signs included hemiplegia/hemiparesis, monoplegia/monoparesis, aphasia, transient monocular blindness, vertigo, dysarthria, and isolated sensory symptoms. The exclusionary criteria were 1) history of AF; 2) admission electrocardiogram (ECG), inpatient cardiac telemetry monitoring, or 24-hour Holter data that demonstrated AF prior to initiation of MCOT; and 3) prothrombotic state (defined as a mutation in the Factor V Leiden gene, prothrombin G20210A gene mutation, antiphospholipid antibody syndrome, or methylenetetrahydrofolate reductase deficiency). Twelve patients were excluded from analysis for the following reasons: 1) denial of insurance coverage for MCOT; 2) death prior to MCOT; and 3) lack of phone access necessary for MCOT.
Patients with patent foramen ovale (PFO) and atrial septal aneurysm (ASA) were included when they had no evidence of deep venous thrombosis by duplex ultrasound study and a normal hypercoagulability profile.
Clinical data included demographic information, stroke risk factors, cardiac history, TOAST diagnosis, and discharge disposition. TOAST diagnosis was assigned by an independent reviewer blinded to MCOT results (D.G.W.). Radiographic data included MRI or CT evidence of cortical vs subcortical infarction, and vascular territory of infarction. All patients underwent MRA or CTA to exclude large vessel disease as the cause of TIA/stroke. All patients underwent transthoracic echocardiography (TTE) with or without transesophageal echocardiography (TEE) and the following data were reviewed: ejection fraction, presence of PFO and ASA, mitral regurgitation, spontaneous echocontrast, left atrial diameter, and presence of aortic arch atheroma.
MCOT data included duration of MCOT monitoring, detection of AF events, heart rate during AF, associated symptoms during AF, and detection of other significant dysrhythmias. AF was defined by loss of p wave activity, irregular baseline undulations (f waves), and RR interval variability. All new-onset AF episodes detected by the MCOT arrhythmia detection algorithm and confirmed by manual review, including AF episodes <30 seconds, were included.
Mobile cardiac outpatient telemetry monitoring.
The MCOT system is a Food and Drug Administration–approved device that provides prolonged cardiac outpatient telemetry monitoring for up to 21 days. The MCOT system is an autotriggered device that uses an event detection algorithm (Mortara) well suited to capture asymptomatic paroxysmal events, including intermittent AF. The event detection algorithm uses RR interval variability and QRS morphology analysis to detect all possible AF events, which are then transmitted to a physician for manual review and confirmation based on standard diagnostic criteria (absence of p wave activity and irregular RR interval). The MCOT system has been tested against the standard Massachusetts Institute of Technology–Beth Israel Hospital and American Heart Association arrhythmia databases with reported 99% sensitivity and 96% positive predictive value for AF lasting >30 seconds, consistent with American National Testing standards (American National Standards Institute/Association for the Advancement of Medical Instrumentation EC 57:1998). Arrhythmia detection algorithms of all cardiac monitoring devices including MCOT have reduced sensitivity for AF episodes <30 seconds and require manual review of all possible AF events. Arrhythmias that may be misdiagnosed as AF by the MCOT arrhythmia detection algorithm include short episodes of paroxysmal supraventricular tachycardia and accelerated junctional tachycardia and were excluded by manual review.
The MCOT system included a home Internet base unit and a pocket-sized wireless recorder/transmitter personal data assistant (PDA) worn by the patients (CardioNet, San Diego, CA).4 The device included three cutaneous lead sensors that recorded two channels of ECG allowing for continuous recording of rhythm strips and automatic wireless transmission via an integrated cellular modem from the PDA.5 Transmission of abnormal ECGs was based upon a standardized arrhythmia detection algorithm and predetermined physician monitoring thresholds and response parameters, or when the patient activated the PDA screen to report symptoms. The MCOT system has no limitations on storage time (24 hours × 21 days) for individual arrhythmia events as compared with autotriggered memory loop recorders. Parameters for event-triggered recording were as follows: severe bradycardia (RR interval >1.71 s); severe pause (RR interval >4 s); moderate pause (RR interval 2–4 s); and no trigger for ST shift. The MCOT algorithm determines probable AF based upon RR interval variability and QRS morphology. The algorithm has the ability to reject ventricular-generated arrhythmias that may be misinterpreted as AF. The algorithm also classifies AF events into severe AF (AF with high heart rate [HR] >130 bpm or RR interval <461 msec) or AF (HR <130 bpm or RR >461 msec).
Statistical analysis.
Descriptive and frequency statistical analyses were performed and comparisons were made with SPSS (version 14.0, SPSS Inc.). Univariate analysis of baseline characteristics for detection vs non-detection of AF by MCOT was performed by the Fisher exact test for categorical variables and Student t test for continuous variables. A logistic regression model was constructed analyzing variables with a p value <0.20 to assess for independent predictors of AF detection by MCOT. A p value <0.05 was considered significant.
RESULTS
Baseline clinical, radiographic, and echocardiographic variables for primary analysis are presented in table 1. MRI evidence of diffusion weighted imaging (DWI)–positive cerebral infarcts was found in 87% (48/55) of patients; 85% (41/48) of these DWI-positive infarcts were cortical in location. TIA was the index event in 14% of patients. All patients underwent either MRA or CTA (extracranial and intracranial), of whom 30% (17/56) had large vessel stenosis (>50%) or occlusion that was not in the arterial distribution of the qualifying stroke or TIA. We found no relation between DWI- positive cortical infarction and AF detection by MCOT (p = 0.653).
Table 1 Baseline clinical, radiographic, and echocardiographic variables for primary analysis (n = 56 except where specified)
All patients underwent either TTE or TEE. TEE was performed in 52% (29/56) of patients and the remaining patients underwent TTE. In our study, PFO with or without associated ASA was found in 24% (13/55) of the cohort and had no relationship to AF detection by MCOT (p = 0.808). Additional echocardiographic variables evaluated in this cohort included 1) left atrial enlargement (>4 cm) in 15% (8/53) of patients; 2) 2–4+ mitral regurgitation in 2% (1/56); and 3) left ventricular dysfunction in 5% (3/56). Left atrial enlargement, mitral regurgitation, and left ventricular dysfunction were not associated with AF detection. Patients had an electrocardiogram 93% (52/56) or a 24-hour Holter study 7% (4/56) and inpatient telemetry monitoring for greater than 24 hours 86% (48/56) that documented normal sinus rhythm and no episodes of AF prior to undergoing MCOT.
All 56 patients with cryptogenic TIA/stroke completed MCOT monitoring for up to 21 days resulting in AF detection in 23% (13/56) of patients (table 2). The AF detection rate was 5.3% (3/56) for AF (>30 seconds). The median duration from index event (TIA/stroke) to initiation of MCOT monitoring was 20 (range 1–122) days. The median monitoring duration was 21 (range 5–21) days and 7 (range 2–19) days of monitoring were required to detect the first episode of AF (table 2). Twenty-seven asymptomatic AF episodes were detected among the 13 patients. Most AF episodes 85% (23/27) were <30 seconds and occurred in 10 patients. The remaining AF episodes 15% (4/27) were 4–24 hours in duration and occurred in 3 patients. There were no AF episodes >30 seconds and <4 hours (figure).
Table 2 Mobile Cardiac Outpatient Telemetry (MCOT) results (n = 56)
Figure Duration and distribution of 27 atrial fibrillation (AF) events detected by Mobile Cardiac Outpatient Telemetry (n = 13)
Brief and prolonged AF events were measured in seconds and hours.
A univariate analysis of clinical and echocardiographic predictors of AF detection by MCOT is presented (table 3). Only patients with a history of diabetes mellitus were found to be at potentially higher risk of developing AF by MCOT by both univariate (p = 0.024) (table 3) and multivariate analysis (OR 6.15; 95% CI 1.16 to 32.73; p = 0.033) (table 4).
Table 3 Univariate analysis of predictors of atrial fibrillation (AF) detection by Mobile Cardiac Outpatient Telemetry (MCOT)
Table 4 Multivariate analysis of predictors of atrial fibrillation detection by Mobile Cardiac Outpatient Telemetry
MCOT in patients with cryptogenic TIA/stroke frequently revealed other significant dysrhythmias, including 1) nonsustained ventricular tachycardia (NSVT) (9%; 5/56); 2) paroxysmal supraventricular tachycardia (PSVT) (29%; 16/56); 3) Mobitz II second-degree AV block (7%; 4/56); and 4) premature atrial complexes (PAC) (18%; 10/56).
The quantitative burden of PACs and PSVT was not measured in this study. Patients found to have NSVT were referred for cardiac evaluations.
DISCUSSION
The main finding of this study is a high rate of asymptomatic AF detection (23%) by MCOT in patients with cryptogenic TIA/stroke, which may be related to three factors: 1) increased duration of cardiac monitoring (21 days) in comparison to prior studies; 2) monitoring of a rigidly defined subset of patients with cryptogenic TIA/stroke with no history of AF; and 3) event detection based upon a standardized automated arrhythmia detection algorithm and continuous review of monitoring data. Monitoring duration required to detect the first AF episode was 7 (range 2–19) days, suggesting that prolonged cardiac monitoring (>7 days) may be warranted for patients with cryptogenic TIA/stroke. We also found that AF episode duration varied significantly from <30 seconds to >4 hours.
The population prevalence of AF in patients aged >50 years is 2.4%.6 AF is estimated to cause approximately 10% of all ischemic strokes or approximately half of all cardioembolic strokes.7,8 Approximately 25% of patients with AF have intermittent AF.6,9 AF is an independent risk factor for ischemic stroke, which increases in prevalence with age. The annual risk of stroke secondary to AF was reported to be 1.5% in patients aged 50–59 years and 23.5% in ages 80–89 years.10
ECG and 24-hour Holter monitoring were reported recently to have a 1.3% (2/149) rate of AF detection in patients with TIA or ischemic stroke and no history of AF.11 AF was detected in 5.7% (5/88) of unselected patients with TIA/stroke with a normal ECG and a 24-hour Holter monitor study by event-loop recording (ELR) for 7 days; this study did not report the AF detection rate by ELR in cryptogenic TIA/stroke and included patients with a history of AF 4.7% (7/149).12 The detection rate for AF >30 seconds by MCOT was 5.3% (3/56) and cannot be directly compared with the ELR detection rate because of significant differences in patient selection and monitoring methodology. A retrospective review of patients referred for known or suspected dysrhythmias undergoing autotriggered memory loop recorders demonstrated asymptomatic AF in 8.7% (52/600).13 A prospective study of patients with syncope/presyncope and severe palpitations undergoing MCOT showed an asymptomatic AF detection rate of 23% (31/134).5 Currently, there are no published reports of controlled prospective studies evaluating specific prolonged arrhythmia monitoring methods in cryptogenic stroke.
Intermittent AF is characterized by recurrent, frequently asymptomatic episodes of AF that self-terminate and may last seconds to hours occurring over several years.14 Paroxysmal atrial fibrillation (PAF) has been defined as self-terminating AF episodes >30 seconds.14 A recent systematic analysis of five prospective cohort studies of occult AF detection by various continuous cardiac monitoring methods showed no consistent established definition of AF.15 We measured the frequency and distribution of all observed new-onset AF episodes in patients with cryptogenic TIA/stroke and found a high rate of AF episodes (<30 seconds). The natural history of these brief AF episodes remains poorly understood. We suggest that brief AF events (<30 seconds) may be important biomarkers associated with more prolonged AF episodes of sufficient duration to result in cardiogenic embolization in patients otherwise presumed to have cryptogenic TIA/stroke. The Canadian Registry of Atrial Fibrillation reported that 24.7% of 757 patients with PAF eventually developed chronic AF over 5 years of follow-up, suggesting that episodes of PAF increase in duration and become persistent over time.16 The duration of continuous AF necessary for thrombus formation in the left atrial appendage has been reported variably to be 48 hours; however, TEE-based studies have shown shorter intervals.17,18
Efforts to detect intermittent AF by prolonged cardiac monitoring of presumed cryptogenic TIA/stroke are warranted because of the potential therapeutic benefit of anticoagulation, although this remains unproven. The Stroke Prevention in Atrial Fibrillation study reported a similar annualized recurrent stroke rate during aspirin therapy in patients with paroxysmal (3.2%) and permanent AF (3.3%).9 Those with prior TIA or stroke have a rate of subsequent stroke of 10–12% per year when treated with aspirin; these patients benefit substantially from adjusted-dose oral anticoagulation.19,20
Prior to the initiation of MCOT, 82.1% (46/56) of our patients were receiving antiplatelet medication, 14.3% (8/56) were receiving warfarin, and 3.6% (2/56) were receiving both antiplatelet medication and warfarin. MCOT results altered patient management in the 13 patients found to have new onset AF by MCOT. Five patients had their antiplatelet medication changed to warfarin, 6 patients were maintained on the warfarin they were taking prior to MCOT, and 2 patients were maintained on antiplatelet medication. The purpose of our study was to demonstrate that AF could be detected by prolonged monitoring in patients with cryptogenic stroke/TIA. The optimal management of patients found to have brief runs of AF has not been established. Medicare reimbursement for MCOT (21 days) was $1,124. Based upon an AF detection rate of 23%, the cost would be $4,841 per new case of AF detected in cryptogenic TIA/stroke.
A potential limitation of this study was the absence of an age-matched control group without a history of TIA/stroke. In addition, not all patients underwent a TEE in this cohort. This issue remains a subject of debate in the literature and there are no standardized guidelines with regard to use of TTE vs TEE in the diagnostic evaluation of unselected patients with stroke. No specific TTE or TEE parameter has been shown to be independently predictive of cardiac thromboembolism in patients with AF when corrected for clinical risk factors.14,21 There are reports showing that TEE does not increase the yield of detection for high risk cardiac sources of emboli requiring anticoagulation.22,23 Conversely, a more recent prospective study of 231 patients showed that 16% of patients with a normal TTE had a high risk source of emboli on a TEE.24 Despite these limitations, this initial study reveals that a significant proportion of patients with stroke of undetermined etiology was found to have AF.
Footnotes
-
e-Pub ahead of print on September 24, 2008, at www.neurology.org.
Disclosure: The authors report no disclosures.
Received December 21, 2007. Accepted in final form June 3, 2008.
REFERENCES
Letters: Rapid online correspondence
REQUIREMENTS
If you are uploading a letter concerning an article:
You must have updated your disclosures within six months: http://submit.neurology.org
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.
