Thromboembolic Events 7–11 Years after Catheter Ablation of Atrial Fibrillation VAN NAM TRAN, M.D.,* ELENA TESSITORE, M.D.,* PASCALE GENTIL-BARON, R.N.,* ANNE-SOPHIE JANNOT, M.D.,† HENRI SUNTHORN, M.D.,* HARAN BURRI, M.D.,* FRANC ¸ OIS MACH, M.D.,* and DIPEN SHAH, M.D.* From the *Cardiology Division, University Hospital of Geneva, Geneva, Switzerland; and †Epidemiology Service, University Hospital of Geneva, Geneva, Switzerland
Background: The results of catheter ablation of atrial fibrillation (AF) beyond 6 years remain unknown. The goal of this study is to assess the risk of thromboembolic events (TEs) and outcomes of AF ablation at long-term follow-up (FU). Methods: All patients who had AF ablation from 2002 until 2005 in our center were contacted for a FU including a questionnaire, cardiac rhythm monitoring, and transthoracic echocardiography. Results: Among the 264 eligible patients, 164 (62%) completed the study. The mean FU was 9.1 years (7.7–10.5). Seven patients had a TE during FU (event ratio 0.41 per 100 patient years [PY]) and their mean CHA2DS2-VASc score was 3.1 ± 1.3. Two patients died from stroke (0.14 per 100 PY) and five of the seven were considered in sinus rhythm (SR) and were off anticoagulation at the time of event. Prior to ablation, 13 patients had history of TE, and only one had a TE during FU. Overall, 14 deaths were documented (0.58 per 100 PY). Stable SR was present in 111 patients (68% of 164 patients) after 1.5 ± 0.6 procedures/patient. Univariate analysis showed that dyslipidemia (odds ratio [OR] = 2.95, P = 0.003), CHA2DS2-VASc ࣙ2 (OR = 3.22, P = 0.001), and amiodarone (OR = 5.64, P < 0.001) were predictors of long-term recurrence. Multivariate analysis showed that only CHA2DS2-VASc ࣙ2 (OR = 2.67, P = 0.023) and amiodarone (OR = 4.62, P = 0.001) were predictors. Conclusions: Our study shows low TE rates 9 years after ablation of AF that are lower than published data for AF patients with anticoagulation only. AF patients with a CHA2DS2-VASc ࣙ2 should, however, be maintained on anticoagulation. (PACE 2015; 00:1–8) atrial fibrillation ablation, very long-term follow-up, thromboembolic events
Introduction Atrial fibrillation (AF) remains the most commonly encountered arrhythmia.1 Rhythm control improves quality of life and left ventricular function and may have the additional benefit of avoiding oral anticoagulation (OAC). Even when sinus rhythm (SR) can be restored with antiarrhythmic drugs (AADs), their use may be limited due to serious side effects.
This study was funded by an unrestricted grant from Biosense Webster. Dr. Shah received research grants from Biosense Webster, St. Jude Medical, and Bard. He is a consultant to Biosense Webster and St. Jude. He has received speaking honorariums from Biosense Webster, Biotronik, St. Jude Medical, and Endosense. Address for reprints: Dipen Shah, M.D., F.H.R.S., Cardiology Division, University Hospital of Geneva, Rue Gabrielle-PerretGentil 4, 1211 Geneva 14, Switzerland. Fax: 41 22 3727229; e-mail: [email protected] Received July 22, 2014; revised December 11, 2014; accepted December 21, 2014. doi: 10.1111/pace.12588
Ablation as rhythm control for AF was first reported in 1998.2 Many studies have reported the superiority of ablation over AAD in maintaining SR at 1-year follow-up (FU).1,3 . However, the success rate for maintaining SR at 5 years and beyond remains uncertain because long-term rhythm monitoring (RM) is challenging and recurrences are often silent.4 The long-term impact of ablation on the progression of the arrhythmic disease over time is unknown. Perhaps as a consequence of these deficiencies, clear recommendations for long-term FU and thromboembolic event (TE) prevention are presently not available.1,5,6 Therefore, the goals of our study are to assess the long-term evolution of TEs after catheter ablation, as well as to identify predictors of late recurrence. Method After approval from our Ethics Committee, every patient who underwent first AF ablation from January 2002 until December 2005 at our center was recruited. A questionnaire was sent to document current medications, symptoms, complications (including recurrences), and status
©2015 Wiley Periodicals, Inc. PACE, Vol. 00
2015
1
TRAN, ET AL.
of FU. After informed consent was obtained, an evaluation was set up either with their cardiologist or at our center. Patients refusing to participate were excluded. Patients lost to FU were defined as patients who could not be contacted (loss of FU with their caregiver; contact information not valid anymore such as phone, mail, or email; and new address not found with local cantonal [government] authorities or on the Internet). Evaluation included RM with either 24-, 48-hour or 7-day monitoring or with a pacemaker when applicable along with transthoracic echocardiography. The choice of heart monitoring modality was at the discretion of the patients’ cardiologist. If the studies were already available within the preceding 6 months, their results were valid for the study. Recurrence of AF, typical flutter, or newly diagnosed atypical flutter or atrial tachycardia was defined by a run longer than 30 seconds in accordance with the expert consensus recommendations.5 Baseline characteristics (BC) were collected at the time of the initial ablation.
and 2 years. Successful ablation was defined by the absence of recurrent atrial arrhythmia longer than 30 seconds beyond the 3-month blanking period. In case of recurrence of atrial arrhythmias, a repeat ablation was advised. Redo Ablation Strategy In case of recurrent AF, pulmonary vein reconduction was eliminated to achieve stable isolation. For persistent AF, LA linear lesions were also verified and ablation was performed to achieve complete conduction block. Supplementary ablation targeting complex fractionated LA electrograms was performed if AF persisted despite conduction block or if conduction block could not be achieved. In case of atypical flutter or atrial tachycardia, linear left and/or right atrial lesions were completed to terminate macroreentrant arrhythmias and block conduction through critical isthmuses. Narrow slow conducting isthmuses were “focally” ablated whereas ablation targeting the earliest activation was performed to eliminate nonreentrant atrial tachycardias10 . After redo ablations, FU was similar to that after index ablation.
Preablation Preparation OAC (vitamin K antagonist) was maintained at least 4 weeks prior to ablation with a therapeutic international normalized ratio (INR) and discontinued 2–4 days before admission. Continuous IV heparin (partial thromboplastin time between 50 seconds and 60 seconds) was administered until 4–6 hours before the procedure; AADs were discontinued 48–72 hours before the intervention. Transesophageal echocardiography was performed to exclude a left atrial (LA) thrombus.
Statistical Analysis A χ 2 test or Fisher’s exact test was used to compare the BC. For each of them, odds ratios (OR) and confidence intervals were estimated. All BC with a P value estimate below 0.15 were included in a multivariate logistic model with arrhythmia recurrence at FU as the dependent variable. Statistical analyses were performed using R 2.13.2 software.
Ablation Procedure Catheter ablation was performed as previously described.7,8 Every pulmonary vein was isolated by segmental ablation.9 Supplementary linear LA lesions were delivered for nonparoxysmal AF. The cavotricuspid isthmus was ablated with the end point of bidirectional block when typical flutter was induced or documented before. Ablation was performed with temperaturecontrolled irrigated radiofrequency delivery (target temperature 50°C, 45–60-second applications, 35–40 W power).
Results From April 2011 to December 2013, 264 patients were contacted and 164 completed the study (62%). Of the remaining 100 patients, 48 declined to participate (18%), 38 were lost to FU (28 in Switzerland and 10 abroad, 14%), and 14 (5%) died before the start of the study. BC of the 164 patients are summarized in Table I. Mean FU was 9.1 years (range from 7.7 years to 10.5 years), which represents 1,492 patient years (PY). The mean age of the study population at FU was 66.0 ± 9.3 years. Initially, 72% of patients had paroxysmal AF; the mean CHA2DS2-VASc score was 1.3 ± 1.2 and 27% had a history of structural heart disease. A single antiarrhythmic agent had been tried in 52% of patients, two in 29%, three or more in 6%, and none in 13%. BC were also compared with those of the 100 patients who did not terminate the study. There
Postprocedural Care Continuous IV heparin followed by lowmolecular-weight heparin was administered until the INR reached a value between 2 and 3 and was maintained for at least 3 months. RM was advised at 1 month, 3 months, 6 months, 1 year,
2
2015
PACE, Vol. 00
LATE EMBOLI AFTER ATRIAL FIBRILLATION ABLATION
Table I. Baseline Patients’ Characteristics Baseline Characteristics Mean follow-up (years) Age >55 years (%) Male (%) Paroxysmal AF (%) Stroke before initial ablation (%) CHA2DS2-VASc 30% (%) Hypertension (%) Dyslipidemia (%) Smoking (%) Vascular disease† (%) LV ejection fraction ࣙ55% (%) Left atrial size ࣘ20 cm2 (%) Amiodarone (%) Flecainide (%)
Patients with Follow-Up (Study Population) (n = 164)
Patients Loss of Follow-Up (n = 100)
9.1 (7.7–10.5) 97 (59) 133 (81) 118 (72) 13 (8) 104 (63) 44 (27) 17 5 2 20 53 (32) 69 (42) 71 (43) 20 (12) 58 (35) 136 (83) 82 (50) 102 (62) 88 (53)
None 47 (47) 79 (79) 68 (68) 6 (6) 59 (59) 17 (17) 4 0 3 10 33 (33) 27 (27) 39 (39) 18 (18) 28 (28) 74 (74) 46 (46) 52 (52) 37 (37)
P Value
0.40 0.92 0.85 0.73 0.80 0.17
0.58 0.06 0.96 0.16 0.63 0.49 0.96 0.90 0.20
AF = atrial fibrillation; BMI = body mass index; CAD = coronary artery disease; CHA2DS2-VASc = congestive heart failure, hypertension, age ࣙ75, diabetes, stroke, vascular disease (†defined as any one of myocardial infarction, peripheral vascular disease, or aortic plaque), age: 65–74 years., sex (female); CP = cardiomyopathy; LV = left ventricle.
(n = 1), septic shock (n = 1), and unknown cause (n = 3) (rate 0.58 per 100 PY). At time of final FU, RM was performed with 24 hours, 48 hours, and 7 days cardiac monitoring in 60%, 7%, and 30%, respectively. In 3%, the AF tracking function of either pacemaker or defibrillator was used. Stable SR was maintained in 68% (111 of 164 patients). Among them, 68% (75 of 111 patients) required only a single procedure, whereas 5% (n = 6) needed ࣙ3 ablations to achieve SR (ratio procedures/patients, 1.5 ± 0.6). At final FU, 89% were off OAC and 94% discontinued AAD. Among the 164 patients, 33% (n = 54) had recurrence of arrhythmia within the first 20 months postablation (Fig. 1). BC were analyzed to identify predictors of arrhythmia recurrence at late FU. Univariate analysis showed that patients in arrhythmia significantly more often had dyslipidemia (OR = 2.95 [1.49; 5.94], P = 0.003), CHA2DS2VASc ࣙ2 (OR = 3.22 [1.63; 6.39], P = 0.001), and amiodarone treatment (OR = 5.64 [2.34; 13.60], P = 0.001). Neither type of AF, number of ablations, age, sex, and hypertension nor LA
was no statistically significant difference, even though the latter were noted to have less systemic hypertension (P = 0.06; Table I). During FU, seven patients had a TE (rate 0.41 per 100 PY) 114 ± 69 months after the index procedure. Of them, two died as a consequence of stroke (one was in SR but under OAC due to previous history of stroke and one in paroxysmal AF with subtherapeutic INR; rate 0.14 per 100 PY). Six patients were considered in SR and five were off OAC at time of event. The mean CHA2DS2VASc score was 3.1 ± 1.3. The LA was markedly dilated in four patients (Table II). There were 13 patients with a history of TE prior to ablation (mean CHA2DS2-VASc score 3.2 ± 1.0), 10 of whom did not have arrhythmia recurrence. One patient had a TE during late FU (without evidence of arrhythmia recurrence), 12 were event free, and eight of them were off OAC (Table III). Fourteen deaths were reported. The causes of death were ischemic stroke (n = 2), congestive heart failure (n = 1), sudden cardiac death (n = 1), neoplasia (n = 3), plane crash (n = 1), horseriding accident (n = 1), aortic aneurysm rupture
PACE, Vol. 00
2015
3
TRAN, ET AL.
Table II. Thromboembolic Events (TE) during Follow-Up (n = 7)
Initial AF
First/Last Procedures
Prior TE
Paroxysmal
None
Paroxysmal Paroxysmal Persistent
None None July 2005
Persistent Paroxysmal
None None
March 2004/ November 2007 January 2004 November 2005 November 2005/ May 2006 October 2005 July 2007
Paroxysmal
1982
February 2004
Rhythm at Time of Event
TE Date
CHA2DS2VASc Score
LA Size at Follow-Up
June 2007
SR off OAC
2
N
January 2011 January 2009 March 2009
SR off OAC SR off OAC SR on OAC
3 1 5
D D D
November 2012 Death from ischemic stroke after 2008 Death from ischemic stroke after 2006
SR off OAC SR off OAC
2 3
D Unknown
Paroxysmal AF subtherapeutic INR
6
Unknown
N: ࣘ20 cm2 . Of the seven patients, six were in SR and off OAC. One was in AF but with subtherapeutic INR. Two died from ischemic stroke. Mean CHA2DS2-VASc score: 3.1 ±1.7. Time of occurrence of TE after ablation: 114 ± 69 months. AF = atrial fibrillation; D = dilated; INR = international normalized ratio; LA = left atrial; OAC = oral anticoagulation; SR = sinus rhythm.
Table III. Patients with History of Stroke Prior to Ablation (n = 13)
1 2 3 4 5 6 7 8 9 10 11 12 13
Initial AF
Rhythm at Follow-Up
OAC Status
CHA2DS2VAsc
CHA2DS2-VAsc without Stroke
Initial LA Size
LA Size at Follow-Up
Stroke after First Ablation
Paroxysmal Paroxysmal Paroxysmal Paroxysmal Paroxysmal Paroxysmal Paroxysmal Persistent Persistent Persistent Persistent Persistent Persistent
SR SR SR SR SR Persistent SR SR Persistent SR Persistent SR SR
Off Off Off Off Off On ? Off On Off On Off On
3 2 2 3 3 4 5 4 4 3 2 2 4
1 0 0 1 1 2 3 2 2 1 0 0 2
N D N N D D N D D D D N D
N N D D D D N D D N ? N D
0 0 0 0 0 0 0 0 0 0 0 0 1
Three patients were in persistent AF. Ten were in SR (eight of them were off OAC, one of them had recurrent stroke). Mean CHA2DS2-VASc score: 3.2 ±1.0. If stroke was removed, the mean CHA2DS2-VASc score would be: 1.2 ±1.0 (see text). AF = atrial fibrillation; D = dilated; LA = left atrial; N = normal; OAC = oral anticoagulation; SR = sinus rhythm.
CHA2DS2-VASc ࣙ2 (OR = 2.67 [1.14; 6.25], P = 0.023) and amiodarone (OR = 4.62 [1.85; 11.58], P = 0.001). Among the 53 patients with recurrent arrhythmia at final FU (32% of 164 patients), 37 had initially paroxysmal AF. Over the course of FU,
size at baseline was predictive of late recurrence (Table IV). In a multivariate logistic model including hypertension, structural heart disease, dyslipidemia, CHA2DS2-VASc ࣙ2, and amiodarone, the only characteristics that remained significant were
4
2015
PACE, Vol. 00
LATE EMBOLI AFTER ATRIAL FIBRILLATION ABLATION
Figure 1. Survival analysis using Kaplan-Meier estimates of the study population of 164 individuals; 54 patients (33%) had recurrence of arrhythmia during the first 20 months following the initial ablation.
19 (51%) progressed to persistent AF; however, no characteristics were identified which could predict AF progression. Eight patients had a pacemaker for sick sinus or tachy-brady syndrome and six with atrioventricular node ablation. Two patients had an intracardiac defibrillator for secondary prevention of malignant ventricular arrhythmias (dilated cardiomyopathy).
Thromboembolic Risk The seven TEs reported during FU in our study occurred 114 ± 69 months after the initial ablation in patients at a high risk of stroke (CHA2DS2-VASc score of 3.1 [±1.2]) and given this late occurrence, they are not related to the procedure itself. This translates to a TE rate of 0.41 in 100 PY after ablation (and a mortality rate induced by TE of 0.14 in 100 PY). In comparison, the TE rate in a population with AF treated with OAC (and heterogeneously also with antiarrhythmic and rate controlling drugs) was estimated at 1.56 in 100 PY in a study conducted by Go et al.19 One can argue that our group of patients was selective, as 63% had a CHA2DS2-VASc score
Background: The results of catheter ablation of atrial fibrillation (AF) beyond 6 years remain unknown. The goal of this study is to assess the risk of thromboembolic events (TEs) and outcomes of AF ablation at long-term follow-up (FU). Methods: All patients who had AF ablation from 2002 until 2005 in our center were contacted for a FU including a questionnaire, cardiac rhythm monitoring, and transthoracic echocardiography. Results: Among the 264 eligible patients, 164 (62%) completed the study. The mean FU was 9.1 years (7.7–10.5). Seven patients had a TE during FU (event ratio 0.41 per 100 patient years [PY]) and their mean CHA2DS2-VASc score was 3.1 ± 1.3. Two patients died from stroke (0.14 per 100 PY) and five of the seven were considered in sinus rhythm (SR) and were off anticoagulation at the time of event. Prior to ablation, 13 patients had history of TE, and only one had a TE during FU. Overall, 14 deaths were documented (0.58 per 100 PY). Stable SR was present in 111 patients (68% of 164 patients) after 1.5 ± 0.6 procedures/patient. Univariate analysis showed that dyslipidemia (odds ratio [OR] = 2.95, P = 0.003), CHA2DS2-VASc ࣙ2 (OR = 3.22, P = 0.001), and amiodarone (OR = 5.64, P < 0.001) were predictors of long-term recurrence. Multivariate analysis showed that only CHA2DS2-VASc ࣙ2 (OR = 2.67, P = 0.023) and amiodarone (OR = 4.62, P = 0.001) were predictors. Conclusions: Our study shows low TE rates 9 years after ablation of AF that are lower than published data for AF patients with anticoagulation only. AF patients with a CHA2DS2-VASc ࣙ2 should, however, be maintained on anticoagulation. (PACE 2015; 00:1–8) atrial fibrillation ablation, very long-term follow-up, thromboembolic events
Introduction Atrial fibrillation (AF) remains the most commonly encountered arrhythmia.1 Rhythm control improves quality of life and left ventricular function and may have the additional benefit of avoiding oral anticoagulation (OAC). Even when sinus rhythm (SR) can be restored with antiarrhythmic drugs (AADs), their use may be limited due to serious side effects.
This study was funded by an unrestricted grant from Biosense Webster. Dr. Shah received research grants from Biosense Webster, St. Jude Medical, and Bard. He is a consultant to Biosense Webster and St. Jude. He has received speaking honorariums from Biosense Webster, Biotronik, St. Jude Medical, and Endosense. Address for reprints: Dipen Shah, M.D., F.H.R.S., Cardiology Division, University Hospital of Geneva, Rue Gabrielle-PerretGentil 4, 1211 Geneva 14, Switzerland. Fax: 41 22 3727229; e-mail: [email protected] Received July 22, 2014; revised December 11, 2014; accepted December 21, 2014. doi: 10.1111/pace.12588
Ablation as rhythm control for AF was first reported in 1998.2 Many studies have reported the superiority of ablation over AAD in maintaining SR at 1-year follow-up (FU).1,3 . However, the success rate for maintaining SR at 5 years and beyond remains uncertain because long-term rhythm monitoring (RM) is challenging and recurrences are often silent.4 The long-term impact of ablation on the progression of the arrhythmic disease over time is unknown. Perhaps as a consequence of these deficiencies, clear recommendations for long-term FU and thromboembolic event (TE) prevention are presently not available.1,5,6 Therefore, the goals of our study are to assess the long-term evolution of TEs after catheter ablation, as well as to identify predictors of late recurrence. Method After approval from our Ethics Committee, every patient who underwent first AF ablation from January 2002 until December 2005 at our center was recruited. A questionnaire was sent to document current medications, symptoms, complications (including recurrences), and status
©2015 Wiley Periodicals, Inc. PACE, Vol. 00
2015
1
TRAN, ET AL.
of FU. After informed consent was obtained, an evaluation was set up either with their cardiologist or at our center. Patients refusing to participate were excluded. Patients lost to FU were defined as patients who could not be contacted (loss of FU with their caregiver; contact information not valid anymore such as phone, mail, or email; and new address not found with local cantonal [government] authorities or on the Internet). Evaluation included RM with either 24-, 48-hour or 7-day monitoring or with a pacemaker when applicable along with transthoracic echocardiography. The choice of heart monitoring modality was at the discretion of the patients’ cardiologist. If the studies were already available within the preceding 6 months, their results were valid for the study. Recurrence of AF, typical flutter, or newly diagnosed atypical flutter or atrial tachycardia was defined by a run longer than 30 seconds in accordance with the expert consensus recommendations.5 Baseline characteristics (BC) were collected at the time of the initial ablation.
and 2 years. Successful ablation was defined by the absence of recurrent atrial arrhythmia longer than 30 seconds beyond the 3-month blanking period. In case of recurrence of atrial arrhythmias, a repeat ablation was advised. Redo Ablation Strategy In case of recurrent AF, pulmonary vein reconduction was eliminated to achieve stable isolation. For persistent AF, LA linear lesions were also verified and ablation was performed to achieve complete conduction block. Supplementary ablation targeting complex fractionated LA electrograms was performed if AF persisted despite conduction block or if conduction block could not be achieved. In case of atypical flutter or atrial tachycardia, linear left and/or right atrial lesions were completed to terminate macroreentrant arrhythmias and block conduction through critical isthmuses. Narrow slow conducting isthmuses were “focally” ablated whereas ablation targeting the earliest activation was performed to eliminate nonreentrant atrial tachycardias10 . After redo ablations, FU was similar to that after index ablation.
Preablation Preparation OAC (vitamin K antagonist) was maintained at least 4 weeks prior to ablation with a therapeutic international normalized ratio (INR) and discontinued 2–4 days before admission. Continuous IV heparin (partial thromboplastin time between 50 seconds and 60 seconds) was administered until 4–6 hours before the procedure; AADs were discontinued 48–72 hours before the intervention. Transesophageal echocardiography was performed to exclude a left atrial (LA) thrombus.
Statistical Analysis A χ 2 test or Fisher’s exact test was used to compare the BC. For each of them, odds ratios (OR) and confidence intervals were estimated. All BC with a P value estimate below 0.15 were included in a multivariate logistic model with arrhythmia recurrence at FU as the dependent variable. Statistical analyses were performed using R 2.13.2 software.
Ablation Procedure Catheter ablation was performed as previously described.7,8 Every pulmonary vein was isolated by segmental ablation.9 Supplementary linear LA lesions were delivered for nonparoxysmal AF. The cavotricuspid isthmus was ablated with the end point of bidirectional block when typical flutter was induced or documented before. Ablation was performed with temperaturecontrolled irrigated radiofrequency delivery (target temperature 50°C, 45–60-second applications, 35–40 W power).
Results From April 2011 to December 2013, 264 patients were contacted and 164 completed the study (62%). Of the remaining 100 patients, 48 declined to participate (18%), 38 were lost to FU (28 in Switzerland and 10 abroad, 14%), and 14 (5%) died before the start of the study. BC of the 164 patients are summarized in Table I. Mean FU was 9.1 years (range from 7.7 years to 10.5 years), which represents 1,492 patient years (PY). The mean age of the study population at FU was 66.0 ± 9.3 years. Initially, 72% of patients had paroxysmal AF; the mean CHA2DS2-VASc score was 1.3 ± 1.2 and 27% had a history of structural heart disease. A single antiarrhythmic agent had been tried in 52% of patients, two in 29%, three or more in 6%, and none in 13%. BC were also compared with those of the 100 patients who did not terminate the study. There
Postprocedural Care Continuous IV heparin followed by lowmolecular-weight heparin was administered until the INR reached a value between 2 and 3 and was maintained for at least 3 months. RM was advised at 1 month, 3 months, 6 months, 1 year,
2
2015
PACE, Vol. 00
LATE EMBOLI AFTER ATRIAL FIBRILLATION ABLATION
Table I. Baseline Patients’ Characteristics Baseline Characteristics Mean follow-up (years) Age >55 years (%) Male (%) Paroxysmal AF (%) Stroke before initial ablation (%) CHA2DS2-VASc 30% (%) Hypertension (%) Dyslipidemia (%) Smoking (%) Vascular disease† (%) LV ejection fraction ࣙ55% (%) Left atrial size ࣘ20 cm2 (%) Amiodarone (%) Flecainide (%)
Patients with Follow-Up (Study Population) (n = 164)
Patients Loss of Follow-Up (n = 100)
9.1 (7.7–10.5) 97 (59) 133 (81) 118 (72) 13 (8) 104 (63) 44 (27) 17 5 2 20 53 (32) 69 (42) 71 (43) 20 (12) 58 (35) 136 (83) 82 (50) 102 (62) 88 (53)
None 47 (47) 79 (79) 68 (68) 6 (6) 59 (59) 17 (17) 4 0 3 10 33 (33) 27 (27) 39 (39) 18 (18) 28 (28) 74 (74) 46 (46) 52 (52) 37 (37)
P Value
0.40 0.92 0.85 0.73 0.80 0.17
0.58 0.06 0.96 0.16 0.63 0.49 0.96 0.90 0.20
AF = atrial fibrillation; BMI = body mass index; CAD = coronary artery disease; CHA2DS2-VASc = congestive heart failure, hypertension, age ࣙ75, diabetes, stroke, vascular disease (†defined as any one of myocardial infarction, peripheral vascular disease, or aortic plaque), age: 65–74 years., sex (female); CP = cardiomyopathy; LV = left ventricle.
(n = 1), septic shock (n = 1), and unknown cause (n = 3) (rate 0.58 per 100 PY). At time of final FU, RM was performed with 24 hours, 48 hours, and 7 days cardiac monitoring in 60%, 7%, and 30%, respectively. In 3%, the AF tracking function of either pacemaker or defibrillator was used. Stable SR was maintained in 68% (111 of 164 patients). Among them, 68% (75 of 111 patients) required only a single procedure, whereas 5% (n = 6) needed ࣙ3 ablations to achieve SR (ratio procedures/patients, 1.5 ± 0.6). At final FU, 89% were off OAC and 94% discontinued AAD. Among the 164 patients, 33% (n = 54) had recurrence of arrhythmia within the first 20 months postablation (Fig. 1). BC were analyzed to identify predictors of arrhythmia recurrence at late FU. Univariate analysis showed that patients in arrhythmia significantly more often had dyslipidemia (OR = 2.95 [1.49; 5.94], P = 0.003), CHA2DS2VASc ࣙ2 (OR = 3.22 [1.63; 6.39], P = 0.001), and amiodarone treatment (OR = 5.64 [2.34; 13.60], P = 0.001). Neither type of AF, number of ablations, age, sex, and hypertension nor LA
was no statistically significant difference, even though the latter were noted to have less systemic hypertension (P = 0.06; Table I). During FU, seven patients had a TE (rate 0.41 per 100 PY) 114 ± 69 months after the index procedure. Of them, two died as a consequence of stroke (one was in SR but under OAC due to previous history of stroke and one in paroxysmal AF with subtherapeutic INR; rate 0.14 per 100 PY). Six patients were considered in SR and five were off OAC at time of event. The mean CHA2DS2VASc score was 3.1 ± 1.3. The LA was markedly dilated in four patients (Table II). There were 13 patients with a history of TE prior to ablation (mean CHA2DS2-VASc score 3.2 ± 1.0), 10 of whom did not have arrhythmia recurrence. One patient had a TE during late FU (without evidence of arrhythmia recurrence), 12 were event free, and eight of them were off OAC (Table III). Fourteen deaths were reported. The causes of death were ischemic stroke (n = 2), congestive heart failure (n = 1), sudden cardiac death (n = 1), neoplasia (n = 3), plane crash (n = 1), horseriding accident (n = 1), aortic aneurysm rupture
PACE, Vol. 00
2015
3
TRAN, ET AL.
Table II. Thromboembolic Events (TE) during Follow-Up (n = 7)
Initial AF
First/Last Procedures
Prior TE
Paroxysmal
None
Paroxysmal Paroxysmal Persistent
None None July 2005
Persistent Paroxysmal
None None
March 2004/ November 2007 January 2004 November 2005 November 2005/ May 2006 October 2005 July 2007
Paroxysmal
1982
February 2004
Rhythm at Time of Event
TE Date
CHA2DS2VASc Score
LA Size at Follow-Up
June 2007
SR off OAC
2
N
January 2011 January 2009 March 2009
SR off OAC SR off OAC SR on OAC
3 1 5
D D D
November 2012 Death from ischemic stroke after 2008 Death from ischemic stroke after 2006
SR off OAC SR off OAC
2 3
D Unknown
Paroxysmal AF subtherapeutic INR
6
Unknown
N: ࣘ20 cm2 . Of the seven patients, six were in SR and off OAC. One was in AF but with subtherapeutic INR. Two died from ischemic stroke. Mean CHA2DS2-VASc score: 3.1 ±1.7. Time of occurrence of TE after ablation: 114 ± 69 months. AF = atrial fibrillation; D = dilated; INR = international normalized ratio; LA = left atrial; OAC = oral anticoagulation; SR = sinus rhythm.
Table III. Patients with History of Stroke Prior to Ablation (n = 13)
1 2 3 4 5 6 7 8 9 10 11 12 13
Initial AF
Rhythm at Follow-Up
OAC Status
CHA2DS2VAsc
CHA2DS2-VAsc without Stroke
Initial LA Size
LA Size at Follow-Up
Stroke after First Ablation
Paroxysmal Paroxysmal Paroxysmal Paroxysmal Paroxysmal Paroxysmal Paroxysmal Persistent Persistent Persistent Persistent Persistent Persistent
SR SR SR SR SR Persistent SR SR Persistent SR Persistent SR SR
Off Off Off Off Off On ? Off On Off On Off On
3 2 2 3 3 4 5 4 4 3 2 2 4
1 0 0 1 1 2 3 2 2 1 0 0 2
N D N N D D N D D D D N D
N N D D D D N D D N ? N D
0 0 0 0 0 0 0 0 0 0 0 0 1
Three patients were in persistent AF. Ten were in SR (eight of them were off OAC, one of them had recurrent stroke). Mean CHA2DS2-VASc score: 3.2 ±1.0. If stroke was removed, the mean CHA2DS2-VASc score would be: 1.2 ±1.0 (see text). AF = atrial fibrillation; D = dilated; LA = left atrial; N = normal; OAC = oral anticoagulation; SR = sinus rhythm.
CHA2DS2-VASc ࣙ2 (OR = 2.67 [1.14; 6.25], P = 0.023) and amiodarone (OR = 4.62 [1.85; 11.58], P = 0.001). Among the 53 patients with recurrent arrhythmia at final FU (32% of 164 patients), 37 had initially paroxysmal AF. Over the course of FU,
size at baseline was predictive of late recurrence (Table IV). In a multivariate logistic model including hypertension, structural heart disease, dyslipidemia, CHA2DS2-VASc ࣙ2, and amiodarone, the only characteristics that remained significant were
4
2015
PACE, Vol. 00
LATE EMBOLI AFTER ATRIAL FIBRILLATION ABLATION
Figure 1. Survival analysis using Kaplan-Meier estimates of the study population of 164 individuals; 54 patients (33%) had recurrence of arrhythmia during the first 20 months following the initial ablation.
19 (51%) progressed to persistent AF; however, no characteristics were identified which could predict AF progression. Eight patients had a pacemaker for sick sinus or tachy-brady syndrome and six with atrioventricular node ablation. Two patients had an intracardiac defibrillator for secondary prevention of malignant ventricular arrhythmias (dilated cardiomyopathy).
Thromboembolic Risk The seven TEs reported during FU in our study occurred 114 ± 69 months after the initial ablation in patients at a high risk of stroke (CHA2DS2-VASc score of 3.1 [±1.2]) and given this late occurrence, they are not related to the procedure itself. This translates to a TE rate of 0.41 in 100 PY after ablation (and a mortality rate induced by TE of 0.14 in 100 PY). In comparison, the TE rate in a population with AF treated with OAC (and heterogeneously also with antiarrhythmic and rate controlling drugs) was estimated at 1.56 in 100 PY in a study conducted by Go et al.19 One can argue that our group of patients was selective, as 63% had a CHA2DS2-VASc score
