J Interv Card Electrophysiol DOI 10.1007/s10840-014-9879-x
Detection of occult atrial fibrillation by pacemaker interrogation in cryptogenic stroke Ghassan Moubarak & Ruben Tamazyan & Philippe Garcon & Claire Join Lambert & Marie Bruandet & Serge Cazeau & Mathieu Zuber
Received: 6 December 2013 / Accepted: 28 January 2014 # Springer Science+Business Media New York 2014
Abstract Purpose Atrial fibrillation (AF) is a major cause of ischemic strokes, and it is assumed that occult intermittent episodes of AF are responsible for some of the seemingly cryptogenic strokes. Cardiac pacemakers feature rhythm diagnostic capabilities and data storage. We investigated whether pacemaker memory interrogation led to identification of undetected AF episodes prior to cryptogenic strokes. Methods The study enrolled all patients admitted between June 2010 and July 2013 for an acute cryptogenic stroke and who were implanted with a permanent pacemaker. Patients with a history of AF and a history of stroke with an identifiable origin were excluded. Pacemaker memories were interrogated to determine the presence of AF prior to the stroke and its temporal relationship with the stroke. Results Fourteen patients (nine men and five women) with a median (interquartile range) age of 84.5 (82.25–87.5) years were included. Median CHADS2 and CHA2DS2-VASc scores were 2 (1–2.75) and 3.5 (3–4), respectively. Pacemaker memories were activated in 13 patients with atrial arrhythmia
detection based on an atrial cutoff rate in 8 patients and on the detection of atrial rate acceleration in 5 patients. Electrograms were available for review in 10 patients. Unknown AF or atrial flutter was diagnosed previous to the stroke in six (43 %) patients. Four patients experienced more than one arrhythmia episode. The last episode occurred in the 48 h prior to stroke in three patients and in the previous 4 weeks in five patients. Anticoagulation was started after the stroke in all of these six patients. Conclusions Pacemaker interrogation has a high diagnostic yield in seemingly cryptogenic stroke, with frequent detection of occult AF. The causal link between AF and stroke is convincingly reinforced by their close temporal proximity, and anticoagulation is warranted in this clinical situation.
The authors state that they have full control of all primary data and that they agree to allow the journal to review their data if requested.
Up to 40 % of ischemic strokes have a negative diagnostic workup [1] and are termed cryptogenic. Atrial fibrillation (AF) is the cause of 25 % of all ischemic strokes [1] with an attributable risk of stroke due to AF increasing with age [2], and it is assumed that occult intermittent episodes of AF are responsible for some of the seemingly cryptogenic strokes. Detection of paroxysmal episodes is important because thromboembolic risk is similar than for continuous AF [3] and requires anticoagulation. The detection rate of occult AF after cryptogenic stroke by cardiac rhythm monitoring is influenced by the duration of monitoring [4–6]. A main limitation of these techniques (Holters, 30-day external event recorders, or implantable loop recorders) is that they are initiated after the stroke, so that no firm conclusion can be drawn regarding the presence of AF before the stroke. Cardiac pacemakers
G. Moubarak (*) : S. Cazeau Department of Pacing and Interventional Electrophysiology, Groupe Hospitalier Paris Saint-Joseph, 185, rue Raymond Losserand, 75014 Paris, France e-mail: [email protected] R. Tamazyan : C. Join Lambert : M. Bruandet : M. Zuber Department of Neurology, Groupe Hospitalier Paris Saint-Joseph, Paris, France R. Tamazyan : C. Join Lambert : M. Bruandet Paris Descartes University, Paris, France P. Garcon Department of Cardiology, Groupe Hospitalier Paris Saint-Joseph, Paris, France
Keywords Pacemaker . Atrial fibrillation . Cryptogenic stroke
1 Introduction
J Interv Card Electrophysiol
feature rhythm diagnostic capabilities and data storage and might be a useful tool in diagnosing intermittent AF. The objective of this observational study was to investigate whether interrogation of pacemaker memory could lead to identification of undetected AF episodes prior to cryptogenic strokes.
2 Methods
2.3 Pacemaker interrogation Pacemaker memories of all eligible patients were systematically interrogated. When available, stored data were retrieved to determine the presence of AF or atrial flutter prior to the stroke, and electrograms (EGMs) were reviewed to verify the appropriateness of automatic arrhythmia classification. The number and duration of arrhythmic episodes and their temporal relationship with the stroke were analyzed. The study was approved by our local ethics committee.
2.1 Patient population The study included all patients admitted to our stroke unit between June 2010 and July 2013 for an ischemic cryptogenic stroke and who were implanted with a permanent dual chamber pacemaker. Strokes were diagnosed using computed tomography (magnetic resonance imaging being contraindicated by the presence of the pacemaker) and classified as cryptogenic according to the TOAST criteria [7]. Patients with transient ischemic attacks or a history of stroke of determined origin were excluded, as were patients with single chamber pacemakers (due to absence of atrial diagnostics) or a history of AF.
2.2 Diagnostic workup Minimal diagnostic workup consisted of daily electrocardiograms, 48-h continuous bedside rhythm monitoring in the stroke unit, duplex ultrasound examination of supra-aortic arteries, and transthoracic echocardiogram. Additional investigations were performed on an individual basis. Table 1 Type of atrial arrhythmia detection and availability of electrograms
Patient number
3 Results 3.1 Patient population Among the 1,179 patients with ischemic stroke admitted to the stroke unit during the study period, 14 (1.2 %) presented a cryptogenic origin and were implanted with a pacemaker. There were nine men and five women with a median (interquartile range) age of 84.5 (82.25–87.5) years. Twelve patients were aged 75 years or older. A history of stroke of undetermined origin, hypertension, and diabetes were present in two, eight, and two patients, respectively. Median CHADS2 and CHA2DS2-VASc scores were 2 (1–2.75) and 3.5 (3–4), respectively. Eleven patients were on aspirin, including two also on clopidogrel, and none were on anticoagulants. Infarcts occurred in the carotid territory in 12 patients (including middle cerebral artery territory in 10) and in the vertebrobasilar territory in 2 patients. Median National Institute of Health Stroke Score (NIHSS) at admission was 2.5 (1.25–5.5). Echocardiography revealed mildly reduced left
Pacemaker memories activated
Atrial arrhythmia detection
Electrograms available
Based on acceleration of atrial rate >175 bpm >180 bpm >160 bpm >170 bpm Based on acceleration of atrial rate
Yes Yes Yes No Yes Yes
>180 bpm >175 bpm None >175 bpm Based on acceleration of atrial rate Based on acceleration of atrial rate Based on acceleration of atrial rate >175 bpm
No Yes No Yes No Yes Yes Yes
Patients with detected AF 1 Yes 2 Yes 3 Yes 4 Yes 5 Yes 6 Yes Patients without detected AF 7 Yes 8 Yes 9 No 10 Yes 11 Yes 12 Yes 13 Yes 14 Yes
J Interv Card Electrophysiol Table 2 Patient risk factors for thromboembolism Patient number
Age
Patients with detected AF 1 85 2 86 3 100 4 84 5 90 6 88 Patients without detected AF 7 91 8 73 9 58 10 83 11 82 12 86 13 84 14 81
Sex
CHADS2
CHA2DS2-VASc
M M F F M F
3 3 2 2 1 2
4 5 4 4 3 4
M M M F M M M F
3 1 0 1 1 2 2 5
5 3 0 3 3 3 3 7
strokes, five in middle cerebral artery topography and one with a large pontine paramedian stroke. There was neither hemorrhagic transformation nor multiple infarcts. Four patients experienced more than one arrhythmia episode (Table 3). The last episode occurred in close temporal proximity to the stroke, in the previous 48 h in three patients and in the previous 4 weeks in five patients. Anticoagulation was started after the stroke in all of these six patients (three with vitamin K antagonists and three with dabigatran).
4 Discussion
ventricular ejection fraction in three patients (45 % in two and 50 % in one) without akinetic or hypokinetic segments and a dilated left atrium in four patients. 3.2 Pacemaker interrogation Indications for pacing were atrioventricular block, sinus node disease, vasovagal syncope, and unknown in six, five, one, and two patients, respectively. Pacemaker memories were activated in 13 patients with EGMs available for review in 10 patients. The type of atrial arrhythmia detection for each pacemaker is presented in Table 1; it was based on an atrial cutoff rate ranging from 160 to 180 bpm in eight patients and on the detection of atrial rate acceleration in five patients. Unknown AF or atrial flutter was diagnosed previous to the stroke in six (43 %) patients, five of whom had EGMs available for review. Individual patient risk factors for thromboembolism are presented in Table 2. The six patients with atrial arrhythmias presented with cortical and large subcortical
This study suggests that AF could be a major cause of seemingly cryptogenic stroke in patients implanted with a cardiac pacemaker, as occult AF was discovered in nearly half of them prior to the stroke. The incidence of paroxysmal AF undetected by short-term monitoring in patients with cryptogenic stroke is still unknown, and studies are ongoing to precise this clinically important point [8]. Pacemaker memory offers the possibility of long-term rhythm monitoring. AF incidence is high in pacemaker patients and is associated with stroke risk [9–13], but these findings have been obtained from patients who mostly did not experience stroke since their pacemaker implantation. Determining the individual incident risk of stroke in case of AF detection in this patient population is difficult. The ASSERT study [11] suggested that episodes as short as 6 min were associated with an increased risk of stroke. Our study explores a different clinical situation, in which we searched for AF prior to a well-documented stroke. Usually, when AF is detected during diagnostic workup after an unexplained stroke, it is inferred that occult AF might have occurred previously, and anticoagulation is started. This assumption may not always be true, as an episode of AF detected after the stroke might well be the first ever experienced by the patient and even be in some cases a consequence and not the cause of the stroke [14]. Pacemaker interrogation is a powerful tool to ascertain or exclude AF episodes prior to stroke. In a subgroup analysis of
Table 3 Patients with atrial fibrillation prior to stroke: arrhythmia description Patient number
Number of episodes
Type of arrhythmia
First recorded episode
Last episode before stroke
Duration of last episode before Duration of longest stroke episode
1 2 3 4 5 6
1 1 9 13 >10 2
AF AF AF AF AF Atrial flutter
6 weeks 1 day 4 months 13 months 4 months 6 weeks
6 1 7 2 1 2
7 h 25 min 30 min
Detection of occult atrial fibrillation by pacemaker interrogation in cryptogenic stroke Ghassan Moubarak & Ruben Tamazyan & Philippe Garcon & Claire Join Lambert & Marie Bruandet & Serge Cazeau & Mathieu Zuber
Received: 6 December 2013 / Accepted: 28 January 2014 # Springer Science+Business Media New York 2014
Abstract Purpose Atrial fibrillation (AF) is a major cause of ischemic strokes, and it is assumed that occult intermittent episodes of AF are responsible for some of the seemingly cryptogenic strokes. Cardiac pacemakers feature rhythm diagnostic capabilities and data storage. We investigated whether pacemaker memory interrogation led to identification of undetected AF episodes prior to cryptogenic strokes. Methods The study enrolled all patients admitted between June 2010 and July 2013 for an acute cryptogenic stroke and who were implanted with a permanent pacemaker. Patients with a history of AF and a history of stroke with an identifiable origin were excluded. Pacemaker memories were interrogated to determine the presence of AF prior to the stroke and its temporal relationship with the stroke. Results Fourteen patients (nine men and five women) with a median (interquartile range) age of 84.5 (82.25–87.5) years were included. Median CHADS2 and CHA2DS2-VASc scores were 2 (1–2.75) and 3.5 (3–4), respectively. Pacemaker memories were activated in 13 patients with atrial arrhythmia
detection based on an atrial cutoff rate in 8 patients and on the detection of atrial rate acceleration in 5 patients. Electrograms were available for review in 10 patients. Unknown AF or atrial flutter was diagnosed previous to the stroke in six (43 %) patients. Four patients experienced more than one arrhythmia episode. The last episode occurred in the 48 h prior to stroke in three patients and in the previous 4 weeks in five patients. Anticoagulation was started after the stroke in all of these six patients. Conclusions Pacemaker interrogation has a high diagnostic yield in seemingly cryptogenic stroke, with frequent detection of occult AF. The causal link between AF and stroke is convincingly reinforced by their close temporal proximity, and anticoagulation is warranted in this clinical situation.
The authors state that they have full control of all primary data and that they agree to allow the journal to review their data if requested.
Up to 40 % of ischemic strokes have a negative diagnostic workup [1] and are termed cryptogenic. Atrial fibrillation (AF) is the cause of 25 % of all ischemic strokes [1] with an attributable risk of stroke due to AF increasing with age [2], and it is assumed that occult intermittent episodes of AF are responsible for some of the seemingly cryptogenic strokes. Detection of paroxysmal episodes is important because thromboembolic risk is similar than for continuous AF [3] and requires anticoagulation. The detection rate of occult AF after cryptogenic stroke by cardiac rhythm monitoring is influenced by the duration of monitoring [4–6]. A main limitation of these techniques (Holters, 30-day external event recorders, or implantable loop recorders) is that they are initiated after the stroke, so that no firm conclusion can be drawn regarding the presence of AF before the stroke. Cardiac pacemakers
G. Moubarak (*) : S. Cazeau Department of Pacing and Interventional Electrophysiology, Groupe Hospitalier Paris Saint-Joseph, 185, rue Raymond Losserand, 75014 Paris, France e-mail: [email protected] R. Tamazyan : C. Join Lambert : M. Bruandet : M. Zuber Department of Neurology, Groupe Hospitalier Paris Saint-Joseph, Paris, France R. Tamazyan : C. Join Lambert : M. Bruandet Paris Descartes University, Paris, France P. Garcon Department of Cardiology, Groupe Hospitalier Paris Saint-Joseph, Paris, France
Keywords Pacemaker . Atrial fibrillation . Cryptogenic stroke
1 Introduction
J Interv Card Electrophysiol
feature rhythm diagnostic capabilities and data storage and might be a useful tool in diagnosing intermittent AF. The objective of this observational study was to investigate whether interrogation of pacemaker memory could lead to identification of undetected AF episodes prior to cryptogenic strokes.
2 Methods
2.3 Pacemaker interrogation Pacemaker memories of all eligible patients were systematically interrogated. When available, stored data were retrieved to determine the presence of AF or atrial flutter prior to the stroke, and electrograms (EGMs) were reviewed to verify the appropriateness of automatic arrhythmia classification. The number and duration of arrhythmic episodes and their temporal relationship with the stroke were analyzed. The study was approved by our local ethics committee.
2.1 Patient population The study included all patients admitted to our stroke unit between June 2010 and July 2013 for an ischemic cryptogenic stroke and who were implanted with a permanent dual chamber pacemaker. Strokes were diagnosed using computed tomography (magnetic resonance imaging being contraindicated by the presence of the pacemaker) and classified as cryptogenic according to the TOAST criteria [7]. Patients with transient ischemic attacks or a history of stroke of determined origin were excluded, as were patients with single chamber pacemakers (due to absence of atrial diagnostics) or a history of AF.
2.2 Diagnostic workup Minimal diagnostic workup consisted of daily electrocardiograms, 48-h continuous bedside rhythm monitoring in the stroke unit, duplex ultrasound examination of supra-aortic arteries, and transthoracic echocardiogram. Additional investigations were performed on an individual basis. Table 1 Type of atrial arrhythmia detection and availability of electrograms
Patient number
3 Results 3.1 Patient population Among the 1,179 patients with ischemic stroke admitted to the stroke unit during the study period, 14 (1.2 %) presented a cryptogenic origin and were implanted with a pacemaker. There were nine men and five women with a median (interquartile range) age of 84.5 (82.25–87.5) years. Twelve patients were aged 75 years or older. A history of stroke of undetermined origin, hypertension, and diabetes were present in two, eight, and two patients, respectively. Median CHADS2 and CHA2DS2-VASc scores were 2 (1–2.75) and 3.5 (3–4), respectively. Eleven patients were on aspirin, including two also on clopidogrel, and none were on anticoagulants. Infarcts occurred in the carotid territory in 12 patients (including middle cerebral artery territory in 10) and in the vertebrobasilar territory in 2 patients. Median National Institute of Health Stroke Score (NIHSS) at admission was 2.5 (1.25–5.5). Echocardiography revealed mildly reduced left
Pacemaker memories activated
Atrial arrhythmia detection
Electrograms available
Based on acceleration of atrial rate >175 bpm >180 bpm >160 bpm >170 bpm Based on acceleration of atrial rate
Yes Yes Yes No Yes Yes
>180 bpm >175 bpm None >175 bpm Based on acceleration of atrial rate Based on acceleration of atrial rate Based on acceleration of atrial rate >175 bpm
No Yes No Yes No Yes Yes Yes
Patients with detected AF 1 Yes 2 Yes 3 Yes 4 Yes 5 Yes 6 Yes Patients without detected AF 7 Yes 8 Yes 9 No 10 Yes 11 Yes 12 Yes 13 Yes 14 Yes
J Interv Card Electrophysiol Table 2 Patient risk factors for thromboembolism Patient number
Age
Patients with detected AF 1 85 2 86 3 100 4 84 5 90 6 88 Patients without detected AF 7 91 8 73 9 58 10 83 11 82 12 86 13 84 14 81
Sex
CHADS2
CHA2DS2-VASc
M M F F M F
3 3 2 2 1 2
4 5 4 4 3 4
M M M F M M M F
3 1 0 1 1 2 2 5
5 3 0 3 3 3 3 7
strokes, five in middle cerebral artery topography and one with a large pontine paramedian stroke. There was neither hemorrhagic transformation nor multiple infarcts. Four patients experienced more than one arrhythmia episode (Table 3). The last episode occurred in close temporal proximity to the stroke, in the previous 48 h in three patients and in the previous 4 weeks in five patients. Anticoagulation was started after the stroke in all of these six patients (three with vitamin K antagonists and three with dabigatran).
4 Discussion
ventricular ejection fraction in three patients (45 % in two and 50 % in one) without akinetic or hypokinetic segments and a dilated left atrium in four patients. 3.2 Pacemaker interrogation Indications for pacing were atrioventricular block, sinus node disease, vasovagal syncope, and unknown in six, five, one, and two patients, respectively. Pacemaker memories were activated in 13 patients with EGMs available for review in 10 patients. The type of atrial arrhythmia detection for each pacemaker is presented in Table 1; it was based on an atrial cutoff rate ranging from 160 to 180 bpm in eight patients and on the detection of atrial rate acceleration in five patients. Unknown AF or atrial flutter was diagnosed previous to the stroke in six (43 %) patients, five of whom had EGMs available for review. Individual patient risk factors for thromboembolism are presented in Table 2. The six patients with atrial arrhythmias presented with cortical and large subcortical
This study suggests that AF could be a major cause of seemingly cryptogenic stroke in patients implanted with a cardiac pacemaker, as occult AF was discovered in nearly half of them prior to the stroke. The incidence of paroxysmal AF undetected by short-term monitoring in patients with cryptogenic stroke is still unknown, and studies are ongoing to precise this clinically important point [8]. Pacemaker memory offers the possibility of long-term rhythm monitoring. AF incidence is high in pacemaker patients and is associated with stroke risk [9–13], but these findings have been obtained from patients who mostly did not experience stroke since their pacemaker implantation. Determining the individual incident risk of stroke in case of AF detection in this patient population is difficult. The ASSERT study [11] suggested that episodes as short as 6 min were associated with an increased risk of stroke. Our study explores a different clinical situation, in which we searched for AF prior to a well-documented stroke. Usually, when AF is detected during diagnostic workup after an unexplained stroke, it is inferred that occult AF might have occurred previously, and anticoagulation is started. This assumption may not always be true, as an episode of AF detected after the stroke might well be the first ever experienced by the patient and even be in some cases a consequence and not the cause of the stroke [14]. Pacemaker interrogation is a powerful tool to ascertain or exclude AF episodes prior to stroke. In a subgroup analysis of
Table 3 Patients with atrial fibrillation prior to stroke: arrhythmia description Patient number
Number of episodes
Type of arrhythmia
First recorded episode
Last episode before stroke
Duration of last episode before Duration of longest stroke episode
1 2 3 4 5 6
1 1 9 13 >10 2
AF AF AF AF AF Atrial flutter
6 weeks 1 day 4 months 13 months 4 months 6 weeks
6 1 7 2 1 2
7 h 25 min 30 min
