Author + information
- Received May 31, 2018
- Revision received November 28, 2018
- Accepted November 28, 2018
- Published online April 15, 2019.
- Andrew E. Noll, MDa,
- Joseph Adewumi, MDb,
- Ram Amuthan, MDb,
- Carl B. Gillombardo, MDb,
- Zariyat Mannan, MDb,
- Erich L. Kiehl, MDc,
- Ayman A. Hussein, MDc,
- Mina K. Chung, MDc,
- Oussama M. Wazni, MDc,
- Randall C. Starling, MD, MPHd,
- Edward G. Soltesz, MD, MPHe and
- Daniel J. Cantillon, MDc,∗ ()
- aDepartment of Cardiovascular Medicine, Cleveland Clinic Foundation, Cleveland, Ohio
- bDepartment of Internal Medicine, Cleveland Clinic Foundation, Cleveland, Ohio
- cDepartment of Cardiovascular Medicine, Electrophysiology Section, Cleveland Clinic Foundation, Cleveland, Ohio
- dDepartment of Cardiovascular Medicine, Kaufman Center for Heart Failure, Cleveland Clinic Foundation, Cleveland, Ohio
- eDepartment of Cardiothoracic Surgery, Cleveland Clinic Foundation, Cleveland, Ohio
- ↵∗Address for correspondence:
Dr. Daniel J. Cantillon, Department of Cardiovascular Medicine, Electrophysiology Section, Cleveland Clinic Foundation, 9500 Euclid Avenue, J2-2, Cleveland, Ohio 44195.
Objectives This study sought to describe the burden of atrial fibrillation (AF)/atrial flutter (AFL) in patients with left ventricular assist devices (LVAD) and to evaluate the impact of rhythm control strategies.
Background AF and AFL among patients with LVADs are poorly characterized.
Methods Retrospective multivariable survival analysis of all LVAD recipients at the Cleveland Clinic from January 1, 2004 to June 30, 2016 examining the association of death, thromboembolism, and major bleeding with AF/AFL and exposure to rhythm control measures.
Results Among 418 patients (median age: 58 [interquartile range: 50 to 67] years, 80% male) with median follow-up of 445 (interquartile range: 165 to 936) days, AF (n = 287 of 418, 69%) and AFL (n = 61 of 418, 15%) were highly prevalent. Patients with AF/AFL (n = 302 of 418, 72%) and without AF/AFL (n = 116 of 418, 28%) had similar mortality (39% vs. 38%; p = 0.88) and major bleeding (46% vs. 49%; p = 0.53); AF/AFL patients had fewer thromboembolic events (13% vs. 23%; p < 0.01). Paroxysmal or persistent AF/AFL was present in 238 patients (57%), and rhythm control exposure (n = 166, 70%) was not associated with decreased mortality (39% vs. 43%; p = 0.57), thromboembolism (13% vs. 17%; p = 0.41), or bleeding (49% vs. 39%; p = 0.16). In the multivariable survival analysis only prior valve surgery (hazard ratio: 2.0; 95% confidence interval: 1.3 to 3.0; p = 0.002) was associated with increased hazard; AF/AFL had no association with risk of death, thromboembolism, or bleeding.
Conclusions Though highly prevalent among LVAD patients, AF/AFL was not associated with increased mortality, thromboembolism, or bleeding, and among paroxysmal/persistent AF patients, rhythm control measures were not associated with improved outcomes.
- atrial arrhythmias
- atrial fibrillation
- atrial flutter
- left ventricular assist devices
- mechanical circulatory support
- survival analysis
Atrial fibrillation (AF) and atrial flutter (AFL) commonly coexist with advanced heart failure, affecting up to 50% of patients with New York Heart Association functional class IV heart failure and 32% to 52% of patients with left ventricular assist devices (LVAD) (1–4). Atrial tachyarrhythmias in heart failure patients have been associated with poor outcomes including death in several large studies, but the degree to which AF/AFL affects mortality and rates of thromboembolism and bleeding in patients with the most advanced heart failure including those on LVAD support is less clear (5–11). Whereas it has been observed that AF/AFL can impair hemodynamics in LVAD patients, especially by causing right heart failure, some prior retrospective studies have reported no effect of AF/AFL on mortality or adverse outcomes, and others have found an increased risk of death, thromboembolism, and bleeding (2–4,12–13).
Furthermore, the outcomes associated with rhythm control measures such as antiarrhythmic medications, electrical cardioversion, and catheter ablation for patients with AF/AFL on LVAD support have not been previously examined. Although randomized trials of amiodarone and dofetilide have not demonstrated a survival benefit in heart failure, the recent CASTLE-AF (Catheter Ablation Versus Standard Conventional Therapy in Patients With Left Ventricular Dysfunction and Atrial Fibrillation) study showed remarkable reductions in all-cause death and heart failure hospitalization (hazard ratio [HR]: 0.53 and 0.56, respectively), reigniting previously discouraged hopes that rhythm control of AF may improve hard outcomes in heart failure patients (14–16).
The aims of this study were to determine the prevalence of atrial arrhythmias in the LVAD population; to examine the effect of AF/AFL on mortality, thromboembolism, and bleeding outcomes; and to investigate the effect of rhythm control measures on these outcomes.
Study population and data collection
All adult patients undergoing LVAD placement at the Cleveland Clinic (Cleveland, Ohio) between January 1, 2004 and June 30, 2016 were identified through query of the institutional LVAD database. The institutional review board of the Cleveland Clinic Foundation approved the study protocol. Chart review of the electronic medical record was performed for all patients and clinical data recorded from the time of LVAD implantation to one of the following endpoints: transplantation; death; LVAD explantation; loss to follow-up (defined as lack of a clinical encounter in the final 12 months of the study in a patient alive with an LVAD); or end of the study period. Clinical data collected included demographics information, pre-LVAD medical history including atrial arrhythmias and cardiac implantable electronic devices, discharge date and medications, post-LVAD clinical events including new onset atrial arrhythmias, thromboembolic and bleeding events, antiarrhythmic and antithrombotic drug therapy, electrical cardioversions, and cardiac ablations.
AF/AFL and clinical data
AF/AFL was defined as clinically manifest arrhythmia documented in progress notes, electrocardiograms, or device interrogations and that required treatment or medical attention. Subclinical AF/AFL detected incidentally on CIED was not included. Thromboembolism was defined as ischemic stroke, transient ischemic attack, or peripheral arterial embolism. Major bleeding was defined according to the INTERMACS (Interagency Registry of Mechanically Assisted Circulatory Support) definition or intracranial hemorrhage. Exposure to rhythm control was defined as treatment with antiarrhythmic drugs, electrical cardioversion, implantable cardioverter-defibrillator shock that converted AF/AFL or AF/AFL ablation.
Continuous variables with normal distribution were expressed as mean ± SD and compared with the Student’s t-test; those with non-normal distributions were expressed as median with interquartile range (IQR) and compared with the Mann-Whitney U test. Categorical variables were expressed as percentages and compared using the Pearson chi-squared or Fisher exact test. Kaplan-Meier survival analysis was performed, stratified by presence or absence of AF/AFL and, among patients with paroxysmal or persistent AF, by exposure to rhythm control measures; right censoring was performed at the time of transplantation, LVAD explantation, or loss to follow-up. Survival curves were compared with the log-rank test and the proportional hazard assumption was confirmed with Schoenfeld residuals. Univariable Cox regression analysis was performed, modeling AF/AFL as a time-dependent covariate. In addition to AF/AFL as the exposure variable of interest, all other variables meeting p < 0.20 in univariable analysis were included as covariates in subsequent multivariable regression analysis. For all analyses, p < 0.05 was considered significant. Analyses were performed using R software version 3.4.4 (March 15, 2018 (R Foundation for Statistical Computing, Vienna, Austria).
Baseline characteristics and AF/AFL prevalence
During the study period, 418 unique patients (median age: 58 [IQR: 50 to 67] years, 80% male) underwent LVAD placement at our institution and were analyzed (Table 1). Only 1 patient underwent LVAD reimplantation after prior explantation, and thus the remainder were de novo implants. The majority of devices placed (82%) were HeartMate II (Abbott, North Chicago, Illinois); a minority (18%) were HeartWare (HeartWare, Framingham, Massachusetts); and a few (0.7%) were HeartMate III devices. AF/AFL was present in 302 patients (72%) and absent in 116 patients (28%). The median age was significantly higher in those with AF/AFL than in those without (59 [IQR: 52 to 67] years vs. 52 [IQR: 41 to 63] years; p < 0.00001). Male sex was more common in AF/AFL patients (83% vs. 71%; p = 0.005). There was no significant difference in major medical comorbidities, but in the AF/AFL group, moderate or severe left atrial enlargement was more prevalent (74% vs. 64%; p = 0.036), and cardiac resynchronization therapy was more common (49% vs. 30%; p = 0.0004).
Before LVAD placement, 240 patients (57%) had a history of AF/AFL; 83% had AF alone, 13% had both AF and AFL, and 4% had isolated AFL (Figure 1). After a median 445 (IQR: 165 to 936) days of follow-up, a total of 302 patients (72%) were affected by AF/AFL, 240 known pre-LVAD and 62 new diagnoses post-LVAD. Eighty percent had AF alone, 15% AF and AFL, and 5% AFL alone. The median time to new AF/AFL diagnosis after LVAD was 4 (IQR: 2 to 8) days. Five patients (1.7%) had atrial tachycardia in the absence of AF or AFL, and 3 patients (1.0%) had atrial tachycardia with concomitant AF or AFL. This diagnosis was based on clinical characteristics rather than electrophysiologic study, and it was typically referred to as “possible” or “probable” atrial tachycardia versus AFL in the chart; therefore, these patients were included in the AFL cohort. Of the 287 patients (95%) with AF, 150 (52%) were paroxysmal, 88 (31%) persistent, 35 (12%) permanent, and 14 (4.9%) of unknown type.
Outcomes according to presence or absence of AF/AFL
Over the follow-up period of 445 (IQR: 165 to 936) days, 161 patients (39%) died, 156 (37%) received a transplant, 11 (3%) underwent LVAD explantation, 24 (6%) were lost to follow-up, and 66 (16%) remained alive with an LVAD. Death occurred in 117 AF/AFL patients (39%) and 44 patients (38%) without AF/AFL, and this difference was not statistically significant (p = 0.88) (Table 2). In survival analysis, there was no difference in all-cause mortality between patients with or without AF/AFL (p = 0.81) or between patients with paroxysmal, persistent, or permanent AF (p = 0.77) (Figure 2). Among AF/AFL patients there were more cardiovascular deaths (33 of 117 [28%] vs. 9 of 44 [21%]) and fewer neurologic deaths (17 of 117 [15%] vs. 12 of 44 [30%]), but these differences did not reach statistical significance (p = 0.054 and p = 0.085, respectively). Arterial thromboembolism was more common in the group without AF/AFL, affecting 23% versus 13% of patients (p = 0.009), and this difference was primarily due to ischemic stroke, which occurred in 19% of patients without AF/AFL and 7.9% of patients with AF/AFL (p = 0.001). There was no significant difference in the incidence of thromboembolism between paroxysmal, persistent, or permanent AF (p = 0.095). The incidence rate of thromboembolism was also higher in patients without AF/AFL, 15.2 versus 9.2 events per 100 person-years, as was the rate of ischemic stroke, 12.9 versus 5.5 events per 100 person-years. Despite the increased rate of stroke in patients without AF/AFL, there was no difference in survival free from thromboembolism, probably because thromboembolism was outweighed by the higher rate of death in both groups (Online Figure 1A). In order to examine the cause of the lower rate of thromboembolism in AF/AFL patients, antithrombotic regimens were analyzed. At the time of discharge from the index hospitalization, more patients in the AF/AFL group were treated with a single antiplatelet drug plus an anticoagulant (237 of 260 [91%] vs. 78 of 99 [79%]) and fewer were treated with a single antiplatelet alone (11 of 260 [4.2%] vs. 9 of 99 [9.1%]) or an anticoagulant alone (6 of 260 [2.3%] vs. 7 of 99 [7.1%]; p = 0.004) (Table 1). Among patients with thromboembolic events, there was no difference in antithrombotic regimen at the time of event, median international normalized ratio (INR) at the time of event (1.8 [IQR: 1.3 to 2.3] vs. 2.2 [IQR: 1.7 to 2.7]; p = 0.72 [AF/AFL vs. no AF/AFL]), or average INR over the 4 weeks preceding the event (2.0 [IQR: 1.7 to 2.5] vs. 2.1 [IQR: 1.7 to 2.5]; p = 0.81 [AF/AFL vs. no AF/AFL]) (Online Table 1). Central nervous system thromboembolic events were more likely to be fatal or disabling in the group without AF/AFL (16 of 23 [70%] vs. 7 of 36 [19%]; p = 0.0001), whereas transient ischemic attack and nondisabling stroke were more common in the AF/AFL group (12 of 36 [33%] vs. 1 of 23 [4.3%] and 17 of 36 [47%] vs. 6 of 23 [26%], respectively).
Major bleeding was common and affected a similar number of patients in the AF/AFL and no AF/AFL groups (138 of 302 [46%] vs. 57 of 116 [49%]; p = 0.53) and in the paroxysmal, persistent, and permanent AF groups (70 of 150 [47%], 39 of 88 [44%], 14 of 35 [40%]; p = 0.78). There was no significant difference in the rates of intracranial hemorrhage or gastrointestinal bleeding between the AF/AFL and no AF/AFL groups or among the types of AF. The incidence rate of major bleeding was similar in patients with and without AF/AFL, 45.7 versus 45.1 events per 100 person-years. There was no difference in survival free from major bleeding between patients with and without AF/AFL (Online Figure 1B). Post-operative bleeding events (such as mediastinal bleeding requiring reoperation) were common, affecting 44 of 302 (13%) patients with AF/AFL and 17 of 116 (15%) without AF/AFL. When these events were excluded, there was still no difference in major bleeding between groups (94 of 302 [31%] vs. 40 of 116 [35%]; p = 0.51). There was no difference in anticoagulation regimen at the time of the bleeding event, median INR at the time of event (2.2 [IQR: 1.5 to 2.8] vs. 2.5 [IQR: 1.5 to 2.9]; p = 0.39 [AF/AFL vs. no AF/AFL]), or average INR over the 4 weeks preceding the event (2.2 [IQR: 1.7 to 2.6] vs. 2.0 [IQR: 1.6 to 2.4]; p = 0.93 [AF/AFL vs. no AF/AFL]) (Online Table 1).
Paroxysmal/persistent AF and rhythm control exposure
Among the 238 patients with paroxysmal or persistent AF, 166 (70%) were exposed to rhythm control. Of these 166 patients with rhythm control exposure, nearly all (n = 163, 98%) were treated with antiarrhythmic drugs, predominantly amiodarone (93%), and most were prescribed the medication on discharge from the index hospitalization (86%) (Online Table 2). Use of other antiarrhythmic drugs was uncommon (sotalol 3.6%, dofetilide 3.0%, dronedarone 0%, Class Ic agents 1.8%). Thirty-four patients (21%) underwent electrical cardioversion for AF, 4 (2.4%) had 1 or more implantable cardioverter-defibrillator shocks converting AF/AFL, 2 patients (1.2%) had a surgical AF ablation at the time of LVAD implantation, and no patients had catheter AF ablation. One patient had a surgical AV node ablation and was not included in the rhythm control group. There was no difference in post-discharge survival among patients with paroxysmal or persistent AF when stratified by exposure to rhythm control measures (p = 0.33) (Figure 3). Compared with those with exposure to rhythm control, patients without rhythm control exposure had similar rates of death (43% vs. 39%; p = 0.57), thromboembolism (17% vs. 13%; p = 0.41), and major bleeding (39% vs. 49%; p = 0.16) (Online Table 3). In subgroup analysis, rhythm control exposure was associated with more time in sinus rhythm (AF burden ≤25% in 117 of 163 [72%] vs. 39 of 67 [58%]; p = 0.045), but there was no difference in death, thromboembolism, or bleeding between patients who maintained sinus rhythm and those with a high burden of AF (Online Table 4).
Univariable and multivariable survival analysis
A Cox proportional hazards model was used to determine factors associated with mortality after LVAD implantation (Online Table 5); AF/AFL was modeled as a time-dependent covariate. Factors associated with increased hazard in the univariable analysis included age (HR: 1.02; 95% confidence interval [CI]: 1.01 to 1.03 per 1 year of age; p < 0.001), ischemic cardiomyopathy (HR: 1.4; 95% CI: 1.01 to 1.90; p = 0.047), prior valve surgery (HR: 1.7; 95% CI: 1.1 to 2.6; p = 0.01), and pre-LVAD antiarrhythmic drugs (HR: 1.4; 95% CI: 1.02 to 1.90; p = 0.039). In the multivariable analysis, only prior valve surgery (HR: 2.0; 95% CI: 1.3 to 3.0; p = 0.002) was associated with increased hazard; AF/AFL had no association with risk of death. Univariable and multivariable regression analyses on all pre-LVAD patient variables and discharge antithrombotic regimens were also performed for survival free from thromboembolic and bleeding events. In the multivariable analysis for survival free from thromboembolism, increased age (HR: 1.02; 95% CI: 1.004 to 1.03 per 1 year of age; p = 0.01) and discharge antithrombotic regimen consisting of dual antiplatelet therapy without an anticoagulant (HR: 11.5; 95% CI: 1.5 to 90.0; p = 0.02) were associated with increased risk of thromboembolism or death, and discharge on dual antiplatelet therapy plus an anticoagulant was associated with reduced hazard (HR: 0.12; 95% CI: 0.02 to 0.90; p = 0.04). Of note, only 2.3% of all patients were discharged on a regimen including dual antiplatelet therapy, and the most common regimen of a single antiplatelet drug plus an anticoagulant had no significant effect on the risk of thromboembolism or death. In the multivariable analysis for survival free from bleeding, increased age (HR: 1.03; 95% CI: 1.01 to 1.04 per 1 year of age; p < 0.001) was associated with increased hazard and history of hypertension (HR: 0.66; 95% CI: 0.46 to 0.93; p = 0.02) was associated with reduced hazard. AF/AFL was not associated with either thromboembolism or bleeding.
The principal finding of our analysis was that AF/AFL was highly prevalent among LVAD patients but not associated with increased mortality or adverse events. Exposure to rhythm control measures among patients with paroxysmal or persistent AF did not appear to improve clinical outcomes. We found a high prevalence of AF/AFL in the cohort, 57% before LVAD and 72% after LVAD implantation, similar to what has been previously reported (2–3). In our study, AF/AFL was not associated with increased risk of death before transplantation or LVAD explantation, and in multivariable models, AF/AFL had no significant effect on mortality, bleeding, or thromboembolism. These findings reinforce those of Hickey et al. (4) and Stulak et al. (13), who found no difference in mortality in cohorts of 249 and 389 patients, respectively. The results, however, stand in contrast to 2 other studies: Deshmukh et al. (3), which reported in a cohort of 331 patients reduced 1-year survival among those with pre-LVAD persistent AF (43% vs. 63%; p = 0.02); and Enriquez et al. (2), a study of 106 patients that reported an increased risk of death or hospitalization for heart failure associated with persistent AF (HR: 3.54; 95% CI: 1.52 to 8.25; p < 0.01); this hazard was primarily driven by heart failure hospitalization, although there was a trend toward increased death (HR: 2.65; 95% CI: 0.96 to 7.35; p = 0.06). The discrepancy between these studies and our own may be explained in part by older age and higher rates of comorbidities among persistent AF patients in the Enriquez et al. (2) and Deshmukh et al. (3) cohorts and differences in definitions of AF (e.g., Enriquez et al.  excluded post-operative AF occurring within 30 days of surgery, but this was included in the other cohorts). The lack of effect on mortality in our study suggests that other factors, such as patient characteristics, advanced heart failure, and complications of the LVAD itself are likely the more important determinants of survival than atrial arrhythmias are.
We found that thromboembolism was less common in AF/AFL patients than in those without AF/AFL (13% vs. 23%), and this was primarily driven by ischemic stroke (7.9% vs. 19%). Furthermore, strokes in AF/AFL patients were less likely to be fatal or disabling than strokes in patients without AF/AFL. These findings were unexpected and perhaps partly explained by different anticoagulation regimens between groups: AF/AFL patients were more likely to be discharged with an antiplatelet drug plus an anticoagulant (91% vs. 79%) and less likely to be discharged on an antiplatelet drug alone (4.2% vs. 9.1%) or an anticoagulant alone (2.3% vs. 7.1%). In multivariable analysis, discharge on dual antiplatelet therapy plus an anticoagulant was associated with improved survival free from thromboembolism or death whereas treatment with dual antiplatelet therapy without an anticoagulant was associated with harm. This lower rate of thromboembolism associated with antiplatelet plus anticoagulant therapy supports the 2013 International Society for Heart and Lung Transplantation Guidelines for Mechanical Circulatory Support recommendation for long-term therapy with warfarin (Class I, Level of Evidence: B) and aspirin (Class I, Level of Evidence: C) in LVAD patients (17). It should be noted, however, that at the time of thromboembolic events, antithrombotic regimens and level of anticoagulation as measured by INR were similar between groups, so there may be other unmeasured confounders that also contribute to the observed differences in the rate and severity of stroke.
Rhythm control exposure not associated with improved outcomes
To our knowledge, this is the first study to systematically investigate the effect of rhythm control measures on clinical outcomes among LVAD patients with AF. Hickey et al. (4) reported that amiodarone had a neutral effect on mortality after LVAD, but other rhythm control strategies were not included. Randomized studies of rhythm control for AF in heart failure patients have shown a survival benefit with catheter ablation but not with antiarrhythmic drugs, but these results cannot be extrapolated to LVAD patients who generally have more advanced heart failure and who may not be as susceptible to loss of atrial systole due to continuous LVAD support (5–7,15,16). In our cohort, we found that exposure to rhythm control measures among patients with paroxysmal or persistent AF was associated with a lower burden of AF but not with reduced mortality, thromboembolism, or bleeding. Nearly all of the 166 patients with rhythm control exposure were treated with antiarrhythmic medications (n = 163, 98%), mostly amiodarone (n = 154, 93%). A substantial minority of patients (n = 34, 21%) also underwent direct current cardioversion for AF/AFL, but other rhythm control measures such as AF ablation or implantable cardioverter-defibrillator shocks converting AF/AFL were rare. Although antiarrhythmic therapy, primarily with amiodarone, was not necessarily targeted at AF/AFL (e.g., amiodarone may have been used to treat ventricular tachycardia), the finding that this powerful antiarrhythmic had no effect on mortality or important clinical endpoints despite being associated with more time in sinus rhythm argues that a rhythm control strategy may not lead to improved outcomes in LVAD patients with AF/AFL. Whereas our data suggest that preventing AF/AFL is unlikely to reduce thromboembolic or bleeding complications or improve survival in LVAD patients, there may be circumstances in which rhythm control is desirable in order to address patient symptoms or treat tachycardia-mediated heart failure.
First, this was a single-center, nonrandomized, retrospective study and thus was subject to the well-described limitations of retrospective research, including possible bias from unmeasured confounders. Second, we considered “rhythm control exposure” to include treatment with antiarrhythmic medications that may not have been targeted at AF/AFL; many patients also had ventricular arrhythmias and it was often difficult to determine from chart review whether the antiarrhythmic medication was primarily used to treat the AF/AFL or the ventricular arrhythmia. Finally, the history of anticoagulant and antiplatelet treatment is difficult to precisely characterize in a nonrandomized, real-world population, and although the reported antithrombotic regimens accurately represent the medical therapy at the times of discharge and clinical events, there may have been deviations from these regimens in the intervening time periods that were not fully captured. Lastly, the delineation of atrial arrhythmia mechanisms was based entirely on clinical characteristics and electrocardiographic findings rather than invasive electrophysiologic testing, thus there was some uncertainty as to the diagnosis in 5 patients in the AF/AFL group who had “possible” or “probable” atrial tachycardia versus AFL.
Atrial arrhythmias are very common in the LVAD population, but they do not portend a negative prognosis in this group of patients that is at high risk of death, thromboembolism, and bleeding regardless of the presence or absence of AF/AFL. Rhythm control of AF/AFL is unlikely to improve outcomes in LVAD patients, but identification of specific clinical scenarios in which this strategy might be of benefit will require further study.
COMPETENCY IN MEDICAL KNOWLEDGE: To describe and contextualize AF/AFL in LVAD patients; to increase understanding of its optimal management; and to provide an example of and method for meaningful clinical research derived from local databases of real-world patient care experience.
TRANSLATIONAL OUTLOOK: As LVAD technology advances, the risk profiles of devices can be expected to improve and the requisite anticoagulation regimens will likely evolve; the effect of AF/AFL on LVAD patients will require continued study to ensure that the risk of bleeding and thromboembolic complications is minimized for all patients.
Dr. Starling is on the advisory panel (without honoraria) for Medtronic. Dr. Cantillon has received consulting fees from Abbott, Biosense, and Boston Scientific. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
All authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the JACC: Clinical Electrophysiology author instructions page.
- Abbreviations and Acronyms
- atrial fibrillation
- atrial flutter
- confidence interval
- hazard ratio
- international normalized ratio
- interquartile range
- left ventricular assist device
- Received May 31, 2018.
- Revision received November 28, 2018.
- Accepted November 28, 2018.
- 2019 American College of Cardiology Foundation
- Enriquez A.D.,
- Calenda B.,
- Gandhi P.U.,
- Nair A.P.,
- Anyanwu A.C.,
- Pinney S.P.
- Deshmukh A.,
- Kim G.,
- Burke M.,
- et al.
- Hickey K.T.,
- Garan H.,
- Mancini D.M.,
- et al.
- Wang T.J.,
- Larson M.G.,
- Levy D.,
- et al.
- Dries D.L.,
- Exner D.V.,
- Gersh B.J.,
- Domanski M.J.,
- Waclawiw M.A.,
- Stevenson L.W.
- Middlekauf H.R.,
- Stevenson W.G.,
- Stevenson L.W.
- Hottigoudar R.U.,
- Deam A.G.,
- Birks E.J.,
- McCants K.C.,
- Slaughter M.S.,
- Gopinathannair R.
- Marrouche N.F.,
- Brachmann J.,
- Andresen D.,
- et al.,
- for the CASTLE-AF Investigators