Advertisement

Mortality and cerebrovascular events after radiofrequency catheter ablation of atrial fibrillation

      Background

      Atrial fibrillation (AF) is associated with a significant increase in the risk of stroke and mortality. It is unclear whether maintaining sinus rhythm (SR) after radiofrequency ablation (RFA) is associated with an improvement in stroke risk and survival.

      Objective

      The purpose of this study was to determine whether SR after RFA of AF is associated with an improvement in the risk of cerebrovascular events (CVEs) and mortality during an extended 10-year follow-up.

      Methods

      RFA was performed in 3058 patients (age 58 ± 10 years) with paroxysmal (n = 1888) or persistent AF (n = 1170). The effects of time-dependent rhythm status on CVEs and cardiac and all-cause mortality were assessed using multivariable Cox models adjusted for baseline and time-dependent variables during 11,347 patient-years of follow-up.

      Results

      Independent predictors of a higher arrhythmia burden after RFA were age (estimated beta coefficient [β] = 0.017 per 10 years, 95% confidence interval [CI] 0.006–0.029, P = .003), left atrial (LA) diameter (β = 0.044 per 5-mm increase in LA diameter, 95% CI 0.034–0.055, P <.0001), and persistent AF (β = 0.174, 95% CI 0.147–0.201, P <.0001). CVEs and cardiac and all-cause mortality occurred in 71 (2.3%), 33 (1.1%), and 111 (3.6%), respectively. SR after RFA was associated with a significantly lower risk of cardiac mortality (hazard ratio [HR] 0.41, 95% CI 0.20–0.84, P = .015). There was not a significant reduction in all-cause mortality (HR 0.86, 95% CI 0.58–1.29, P = .48) or CVEs (HR 0.79, 95% CI 0.48–1.29, P = .34) in patients who remained in SR after RFA.

      Conclusion

      Maintenance of SR after RFA is associated with a reduction in cardiovascular mortality in patients with AF.

      Abbreviations:

      AF (atrial fibrillation), CI (confidence interval), CVE (cerebrovascular event), HR (hazard ratio), LA (left atrium), LVEF (left ventricular ejection fraction), OSA (obstructive sleep apnea), RFA (radiofrequency ablation), SR (sinus rhythm)

      Keywords

      Introduction

      Atrial fibrillation (AF) is associated with an increase in mortality and cerebrovascular events (CVEs).
      • Benjamin E.J.
      • Wolf P.A.
      • D’Agostino R.B.
      • Silbershatz H.
      • Kannel W.B.
      • Levy D.
      Impact of atrial fibrillation on the risk of death: the Framingham Heart Study.
      • Piccini J.P.
      • Hammill B.G.
      • Sinner M.F.
      • Jensen P.N.
      • Hernandez A.F.
      • Heckbert S.R.
      • Benjamin E.J.
      • Curtis L.H.
      Incidence and prevalence of atrial fibrillation and associated mortality among Medicare beneficiaries, 1993–2007.
      However, whether the presence of AF is a surrogate for underlying comorbidities or whether AF is the causal factor is not clear. If the latter were true, maintenance of sinus rhythm (SR) would improve survival and freedom from CVEs.
      Radiofrequency catheter ablation (RFA) of AF has evolved as an effective treatment modality to eliminate AF. Whether elimination of AF by RFA is associated with an improvement in survival remains unclear. Recurrences of AF after RFA are not uncommon and often may prompt subsequent repeat ablation procedures, resetting the follow-up period. Time spent in SR, or conversely the AF burden, is a more realistic and accurate indicator of actual clinical outcomes of RFA for AF.
      The purpose of this study was to assess the arrhythmia burden and its predictors after RFA of AF and to determine whether maintenance of SR after RFA is associated with an improvement in survival and thromboembolic risk in a large cohort of patients with AF followed over 10 years.

      Methods

      Study subjects

      The subjects of this study were 3058 consecutive patients who underwent RFA to eliminate AF between 2001 and 2011. There were 2261 men (74%) and 797 women (26%). Mean age was 58 ± 10 years. AF was paroxysmal in 1888 patients (62%) and persistent in 1170 (38%) patients (Table 1). Patients with long-standing persistent AF were grouped as persistent AF.
      Table 1Clinical characteristics of the study patients
      No. of patients3058
      Age (years)58 ± 10
      Male (%)2261 (74%)
      Type of atrial fibrillation
       Paroxysmal1888 (62%)
       Persistent1170 (38%)
      Left ventricular ejection fraction0.561 ± 0.09
      Left atrial diameter (mm)43.3 ± 6.8
      Hypertension (%)1486 (49%)
      Chronic kidney disease (%)34 (1%)
      Diabetes mellitus (%)339 (11%)
      Obstructive sleep apnea (%)501 (16%)
      Cerebrovascular events (%)189 (6%)
      Chronici obstructive pulmonary disease (%)166 (5%)
      Congestive heart failure (%)378 (12%)
      CHADS2 score
       01042 (36%)
       11150 (39%)
       2533 (18%)
       3156 (5%)
       441 (1%)
       58 (0.3%)
      Data are given as mean ± SD or no. (%).

      Study protocol

      The study protocol was approved by the Institutional Review Board. Medical records and the latest follow-up through local physician and/or referring physician were reviewed for all patients. A questionnaire on rhythm status, clinical events, and antiarrhythmic and anticoagulant therapy was sent to all patients. Any episode of atrial arrhythmia that lasted >30 seconds indicated a recurrence.

      Electrophysiologic study and RFA

      All patients provided informed written consent. RFA was performed as described previously.
      • Oral H.
      • Chugh A.
      • Yoshida K.
      • et al.
      A randomized assessment of the incremental role of ablation of complex fractionated atrial electrograms after antral pulmonary vein isolation for long-lasting persistent atrial fibrillation.
      Briefly, after the transseptal puncture, intravenous heparin was administered to maintain an activated clotting time of 300–350 seconds. Catheter navigation was performed with fluoroscopic guidance and a 3-dimensional electroanatomic mapping system (CARTO, Biosense Webster, Diamond Bar, CA; or NavX, St. Jude Medical, Minneapolis, MN). Over the last decade, the ablation strategy evolved from segmental ostial pulmonary vein isolation to left atrial circumferential ablation to antral pulmonary vein isolation. Ablation of complex fractionated atrial electrograms and linear ablation was performed at the discretion of the operator. As catheter technology evolved, ablation initially was performed with a 4-mm-tip catheter, then with an 8-mm-tip catheter, and most recently with an open irrigation-tip catheter.

      Post-RFA follow-up

      After ablation, all patients were observed during an overnight hospital stay. All patients were seen in an outpatient clinic 3 months after the procedure and every 6–12 months thereafter, by either 1 of the investigators or their local referring physician. Patients were instructed to call a dedicated nurse whenever they experienced symptoms. At 6–12 months after RFA and whenever they reported symptoms, rhythm status was monitored by serial ECGs, extended Holter monitors, and event recorders. Anticoagulation was maintained for at least 3 months after RFA.
      Deaths were ascertained through hospital records as well as using the Social Security Death Index. Of those patients who died, death certificates could be obtained for all but 5 (95%) for whom consent from family could not be obtained. The time and cause of death were adjudicated after reviewing the death certificates and medical records. Every effort was made to obtain the latest records on the rhythm and clinical status prior to death. In 3 patients the final rhythm status could not be determined. For these 3 patients, there were 1669, 552, and 295 days with missing information on rhythm status.
      The cause of death was defined as cardiac when the terminal event was congestive heart failure, cardiogenic shock, fatal arrhythmia, cardiopulmonary arrest, or myocardial infarction. The cause of death was considered to be arrhythmic if there was sudden collapse, unexplained death during sleep, or spontaneous and rapid loss of circulation and respiration in the absence of a preceding medical condition. Cerebrovascular deaths were defined as terminal events due to ischemic or hemorrhagic stroke.
      • Steinberg J.S.
      • Sadaniantz A.
      • Kron J.
      • et al.
      Analysis of cause-specific mortality in the Atrial Fibrillation Follow-up Investigation of Rhythm Management (AFFIRM) study.
      CVEs were defined as ischemic, or hemorrhagic and non–central nervous system thromboembolic events. The data were censored on January 2, 2013.

      Rhythm status

      To assess the effect of maintenance of SR on overall mortality, first the presence of atrial arrhythmias (AF and atrial flutter/tachycardia) was assessed. The use of anticoagulant and/or antiarrhythmic drugs at any point during follow-up also was determined. The types of follow-up encounters were procedure (15.2%), survey (14.4%), phone call (8.3%), clinic visit (56.2%), and other (5.8%). Each time interval between encounters was assigned a rhythm status as SR or AF (including all atrial arrhythmias) based on the current and previous encounter (Figure 1). Whenever a recurrence was suspected based on patient-reported symptoms, rhythm status was always confirmed with an event monitor, Holter monitor, or ECG. The proportion of time spent in SR was estimated by dividing the time spent in SR by the total follow-up period for each patient.
      Figure thumbnail gr1
      Figure 1Schematic representation of various patient scenarios with regard to documented and allocated rhythm. Allocated rhythm was determined as the documented rhythm carried backward from each clinic visit, phone call, survey, or procedure (P). Dark bars and light bars indicate time in atrial fibrillation (AF) and time in sinus rhythm (SR), respectively.

      Statistical analysis

      Descriptive statistics were calculated based on complete case data. Continuous variables were compared using the Student t test. Missing data were handled by multiple imputations (m = 5 imputations). Variables with missing data were left ventricular ejection fraction (LVEF) in 38 (1.2%) and left atrial (LA) size in 156 (5.1%).

      Proportion of time spent in AF

      Linear regression was used to test for the predictors of the proportion of time in AF. Parameter estimates were obtained by combining the results from the 5 imputations. Beta coefficients (β) reflect the increase in AF burden for 1-unit increase in the independent variable. The β coefficients for LA size, LVEF, and age represent the increase in AF burden for a 5-, 10-, and 10-unit change, respectively. Models were built using the Least Absolute Shrinkage and Selection Operator (LASSO) and the best subset methods on each imputation after considering consistency of the results between imputations.

      Time-varying data

      The rhythm during an interval was assumed to be the rhythm assessed at the end of the interval. Rhythm status was missing for 19 intervals in 19 patients (3 in the last interval among patients who died). The missing intervals were omitted from the analysis.

      Survival analysis

      All-cause and cardiac mortality and CVEs were the outcome measures. Survival analysis, including Kaplan-Meier plots and Cox regression, was used to investigate predictors of time from the first procedure to study end-points. Cox regression models were fit to assess the effects of baseline covariates. The strategy was to fit a parsimonious base model for each outcome. Because of colinearity among the covariates, both the LASSO and the best subset selection methods were used for variable selection. With the LASSO method, bootstrap samples were used to ensure robustness. For the best subset method, parameter estimates were obtained by using imputations. After the base model was fit, the proportional hazards property was tested, and none of the covariates showed significant nonproportional hazards.
      Rhythm status at each time point was added to the base model for each outcome measure. Antiarrhythmic and anticoagulant drugs were entered as time-dependent variables. In addition, the effect of a 90-day blanking period was assessed, first to exclude deaths or events that might have been due to technical factors (and focus on the potential effect of maintenance of SR) and then to include all deaths or events to evaluate the overall effect of the procedure. A 2-sided P <.05 indicated significance (SAS, SAS Institute Inc., Cary, NC).

      Results

      RFA for AF

      In 3058 patients, a total of 4922 RFA procedures were performed. There were 1.55 ± 0.78 and 1.71 ± 0.85 procedures per patient with paroxysmal and persistent AF, respectively (P <.0001). Follow-up was 7390 and 3957 person-years for patients with paroxysmal (range 0.03–11.91 years, median 3.43 years) and persistent AF (range 0.01–11.23 years, median 2.81 years), respectively. The proportion of time in SR after RFA was higher in patients with paroxysmal AF (0.74 ± 0.34) than in those with persistent AF (0.52 ± 0.38, P <.0001). Time in SR had a bimodal distribution in both groups, with clustering either below 0.20 or above 0.80 (Appendix, Appendix). An antiarrhythmic drug was concomitantly administered in 1351 of 1888 patients (72%) with paroxysmal AF and in 860/1170 patients (74%) with persistent AF (P = .24) at some point during follow-up.
      At last follow-up, 1489 of 1888 patients with paroxysmal AF (78.9%) were in SR, including 29.9% who were taking an antiarrhythmic drug. Among patients with persistent AF, 691 of 1170 (59.1%) were in SR, including 35.4% who were taking an antiarrhythmic drug.

      Predictors of recurrence

      Single variable predictors of a higher arrhythmia burden after RFA included older age, female gender, obstructive sleep apnea (OSA), persistent AF, larger LA, chronic kidney disease, lower LVEF, hypertension, and diabetes mellitus (Figure 2).
      Figure thumbnail gr2
      Figure 2Single variable predictors of the proportion of time spent in atrial fibrillation (AF). In order of appearance: LA = left atrial diameter (per 5 mm); EF = left ventricular ejection fraction (per 10%); CKD = chronic kidney disease (estimated glomerular filtration rate <30 m:/min); CVA = cerebrovascular accident; OSA = obstructive sleep apnea; DM = diabetes mellitus; HTN = hypertension; CAD = coronary artery disease; Age (per 10 years). CI = confidence interval.
      In multivariable analysis, the independent predictors were older age (estimated β = 0.017 per 10 years, 95% confidence interval [CI] 0.006–0.029, P = .0035), larger LA size (β = 0.044 per 5-mm increase in LA diameter, 95% CI 0.034–0.055, P <.0001), and persistent AF (β = 0.174, 95% CI 0.147–0.201, P <.0001). Based on standardized beta coefficients (std β), persistent AF (vs paroxysmal AF) had the largest effect on arrhythmia burden after RFA (std β = 0.23), followed by LA size (std β = 0.16) and age (std β = 0.05).

      Mortality after RFA

      During 11,347 patient-years of follow-up after RFA, there were 111 deaths among 3058 patients (3.6%). All-cause mortality was higher in patients with persistent AF (54/1170 [4.6%]) than in those with paroxysmal AF (57/1888 [3.0%], log-rank P <.001; Figure 3). Cardiac mortality also was higher in patients with persistent AF (18/1170 [1.53%]) than in those with paroxysmal AF (15/1888 [0.79%], log-rank P = .011; Figure 4).
      Figure thumbnail gr3
      Figure 3Adjusted Kaplan-Meier curve for all-cause mortality by paroxysmal vs persistent atrial fibrillation (AF). Estimates are adjusted using Cox regression and presented for the following covariate values: all-cause death (age 58 years, ejection fraction 56.27%, and no diabetes or coronary artery disease [CAD]).
      Figure thumbnail gr4
      Figure 4Adjusted Kaplan-Meier curve for cardiac mortality. Estimates are adjusted using Cox regression and presented for the following covariates: Cardiac death (age 58 years, EF 56.27%, no CAD or obstructive sleep apnea). Abbreviations as in .
      The cause of death was cardiac in 33 patients (30%), noncardiac in 73 (66%), and unknown in 5 (4%). Cardiac causes of death included 13 arrhythmic (12%) and 20 nonarrhythmic (18%). The noncardiac causes of death were cancer in 37 patients (33%), ischemic CVE in 3 (3%), hemorrhagic CVE in 10 (9%), pulmonary in 10 (9%), and other causes in 13 (12%).
      There were 6 deaths within 90 days after RFA, including ischemic CVE in 1 patient (day 52), hemorrhagic CVE in 1 (day 8), sudden cardiac death in 1 who was taking dronaderone (day 12), atrio-esophageal fistula in 1 (day 12), and congestive heart failure in 2 (days 16 and 76).

      Predictors of mortality

      In single variable models, age, coronary artery disease, hypertension, diabetes mellitus, OSA, LVEF, LA size, and persistent AF were associated with an increased risk of all-cause mortality (P <.05; Table 2). The strongest predictors of all-cause mortality were coronary artery disease (hazard ratio [HR] 3.20, 95% CI 2.11–4.85, P <.0001), diabetes mellitus (HR 2.32, 95% CI 1.43–3.78, P = .0007), and age (HR 1.93 per 10 years, 95% CI 1.55–2.40, P <.0001).
      Table 2Predictors of all-cause mortality, cardiac mortality, and CVEs based on single-variable Cox regression models
      All-cause mortalityCardiac mortalityCVEs
      HR (95% CI)HR (95% CI)HR (95% CI)
      P valueP valueP value
      Age
      Hazard ratio (HR) and confidence interval (CI) are for a 10-year increase in age and a 10% increase in left ventricular ejection fraction (LVEF).
      1.93 (1.55–2.40)1.60 (1.09–2.35)1.39 (1.08–1.79)
      P <.0001P = .02P = .01
      LVEF
      Hazard ratio (HR) and confidence interval (CI) are for a 10-year increase in age and a 10% increase in left ventricular ejection fraction (LVEF).
      0.65 (0.54–0.77)0.48 (0.37–0.62)0.87 (0.67–1.14)
      P <.0001P <.0001P = .31
      LA size
      HR and CI are for a 5-mm increase in left atrial (LA) size.
      1.27 (1.11–1.46)1.33 (1.03–1.71)1.16 (0.97–1.38)
      P = .0007P = .03P = .10
      Gender1.36 (0.90–2.04)1.06 (0.48–2.36)1.30 (0.77–2.18)
      P = .15P = .88P = .33
      Persistent AF1.89 (1.30–2.74)2.37 (1.19–4.73)0.862 (0.52–1.43)
      P = .0009P = .01P = .57
      CAD3.20 (2.11–4.85)4.71 (2.31–9.58)2.15 (1.22–3.81)
      P <.0001P <.0001P = .0084
      HTN1.86 (1.28–2.72)1.39 (0.70–2.76)1.32 (0.83–2.11)
      P = .0013P = .34P = .25
      DM2.32 (1.43–3.78)2.83 (1.22–6.54)0.92 (0.40–2.12)
      P = .0007P = .015P = .84
      OSA1.71 (1.05–2.79)2.93 (1.35–6.34)1.17 (0.60–2.28)
      P = .03P = .007P = .65
      History of CVE0.76 (0.31–1.86)0.50 (0.07–3.65)4.15 (2.31 -7.47)
      P = .55P = .49P <.0001
      Chronic kidney disease1.03 (0.14–7.36)3.55 (0.48–26.03)3.22 (0.79–13.14)
      P = .98P = .21P = .10
      AF = atrial fibrillation; CAD = coronary artery disease; CVE = cerebrovascular event; DM = diabetes mellitus; HTN = hypertension; OSA = obstructive sleep apnea.
      low asterisk Hazard ratio (HR) and confidence interval (CI) are for a 10-year increase in age and a 10% increase in left ventricular ejection fraction (LVEF).
      HR and CI are for a 5-mm increase in left atrial (LA) size.
      The individual predictors of excess cardiac mortality were age, LVEF, LA size, persistent AF, diabetes mellitus, coronary artery disease, and OSA. The strongest predictors of cardiac mortality were LVEF (HR 0.48 per 10%, 95% CI 0.37–0.62, P <.001), coronary artery disease (HR 4.71, 95% CI 2.31–9.58, P <.0001), OSA (HR 2.9, 95% CI 1.3–6.3, P <.007), and diabetes mellitus (HR 2.93, 95% CI 1.35–6.34, P = .007).
      On multivariable analysis, older age, lower LVEF, diabetes mellitus, coronary artery disease, and persistent AF (borderline significant) were predictive of all-cause mortality (Online Supplementary Table 1). Lower LVEF, coronary artery disease, and OSA were significant predictors of cardiac mortality.

      SR and mortality after RFA

      Maintenance of SR after RFA was associated with a significantly lower risk of cardiac mortality (HR 0.41, 95% CI 0.20–0.84, P = .015), irrespective of the blanking period or antiarrhythmic or anticoagulant drug use (Table 3). The hazard ratio for all-cause mortality in patients with SR after RFA was not significant (HR 0.86, 95% CI 0.58–1.29, P = .48).
      Table 3SR vs AF as a time-dependent predictor of all-cause mortality, cardiac mortality, and CVEs, adjusted for baseline covariates
      All-cause mortality (N = 111)Cardiac mortality (N = 33)CVE (N = 71)
      HR (95% CI)HR (95% CI)HR (95% CI)
      P valueP valueP value
      Rhythm (with 90-day blanking period)HRSRvs.AF = 0.86HRSRvs.AF = 0.41HRSRvs.AF = 0.79
      (0.58–1.29)(0.20–0.84)(0.48–1.29)
      P = .48P = .015P = 0.34
      Rhythm (without 90-day blanking period)HRSRvs.AF = 0.80HRSRvs.AF = 0.37HRSRvs.AF = 0.82
      (0.54–1.19)(0.19–0.76)(0.51–1.33)
      P = .27P = .006P = .42
      Rhythm (AAD/AC) (with 90-day blanking period)HRSRvs.AF = 0.86HRSRvs.AF = 0.44HRSRvs.AF = 0.76
      (0.57–1.29)(0.21–0.91)(0.46–1.27)
      P = .46P = .028P = .30
      Rhythm (AAD/AC) (without 90-day blanking period)HRSRvs.AF = 0.80HRSRvs.AF = 0.40HRSRvs.AF = 0.80
      (0.53–1.19)(0.20–0.82)(0.49–1.31)
      P = .26P = .01P = .38
      AAD = antiarrhythmic drug; AC = anticoagulant drug; AF = atrial fibrillation; CI = confidence interval; CVE = cerebrovascular event; HR = hazard ratio; SR = sinus rhythm.
      Repeat RFA, as a time-dependent variable, was associated with a reduction in all-cause mortality (HR 0.45, 95% CI 0.30–0.67, P = .0001). There also was a trend toward a decrease in cardiovascular mortality (HR 0.52, 95% CI 0.24–1.04, P = .076).

      CVEs after RFA

      There were 71 thromboembolic and CVEs in 3058 patients (2.3%). There were 29 transient ischemic CVEs (41%), 18 ischemic CVEs (25%), 12 hemorrhagic CVEs (17%), 8 CVEs of unknown etiology (11%), and 4 noncerebrovascular events (6%). The prevalence of thromboembolic events was similar among patients with persistent (49/1888 [2.6%]) and paroxysmal AF (22/1170 [1.9%], P = .57; Figure 5).
      Figure thumbnail gr5
      Figure 5Adjusted Kaplan-Meier curve for cerebrovascular events. Estimates and abbreviations as in Figure 3.

      Predictors of CVEs

      Significant single variable predictors of CVEs after RFA included age (HR 1.4 per 10 years, 95% CI 1.1–1.8, P = .01), coronary artery disease (HR 2.2, 95% CI 1.2–3.8, P = .008), and history of a CVE (HR 4.2, 95% CI 2.3–7.5, P <.0001; Table 2). On multivariable analysis, coronary artery disease and history of CVE were the only significant predictors of CVEs (Online Supplementary Table 1).
      Maintenance of SR after RFA was not associated with a significant reduction in CVEs (HR 0.82, 95% CI 0.51–1.33, P = .42), irrespective of the blanking period or antiarrhythmic or anticoagulant drug use (Table 3). There was no association between repeat procedures and CVEs (HR 0.74, 95% CI 0.45–1.22, P = .24).

      Discussion

      Main findings

      The main findings of this study are as follows. (1) Arrhythmia burden has a bimodal distribution, with the majority of patients having an AF burden <20% or >80% after RFA. (2) The independent predictors of a higher arrhythmia burden after RFA are persistent AF, larger LA size, and older age. (3) During 10-year follow-up, the total mortality after RFA is 3.6%, and both all-cause and cardiac mortality are higher in patients with persistent AF than in those paroxysmal AF. (4) Older age, lower LVEF, diabetes mellitus, coronary artery disease, and persistent AF are significant predictors of all-cause mortality, whereas lower LVEF, coronary artery disease, and OSA are significant predictors of cardiac mortality. (5) Absence of symptomatic or documented recurrences of AF after RFA is associated with a significant 60% lower risk of cardiac mortality, irrespective of the blanking period or antiarrhythmic or anticoagulant drug use. (6) CVEs occur in a similar proportion of patients with paroxysmal and persistent AF (~2%), and ~20% are hemorrhagic. (7) Histories of CVE and coronary artery disease are significant predictors of CVEs. (8) Maintenance of SR after RFA is not associated with a significant decrease in the risk of CVEs, irrespective of the blanking period or antiarrhythmic or anticoagulant drug use.
      This is the first study that systematically investigated the relationship among arrhythmia burden, maintenance of SR as time-dependent variables, and clinical outcomes after RFA in a large cohort of patients with paroxysmal or persistent AF during 10 years of follow-up. Maintenance of SR after RFA of AF was associated with a significant 60% reduction in cardiac mortality.

      Arrhythmia burden

      Previous studies have focused on recurrences of AF as an end-point for evaluating long-term outcomes. Recurrences can occur months to years after RFA, and many patients may experience extended periods in SR. In fact, short or infrequent recurrences of AF may represent a significant improvement for many patients. A significant reduction in arrhythmia burden after RFA represents a more accurate and clinically relevant outcome for many patients with AF than complete freedom from recurrent AF.
      In this study, patients with paroxysmal AF and patients with persistent AF were in SR during 74% and 52% of the 10-year follow-up period, respectively. After the most recent RFA, ~80% and 60% of patients with paroxysmal and persistent AF were in SR, respectively, including ~30% of the patients who were receiving an antiarrhythmic drug. These results highlight the favorable outcomes in the majority of patients with AF long after RFA; however, they also point to the need to further improve the efficacy of RFA, particularly in patients with persistent AF.

      SR and mortality after RFA

      Based on an intention-to-treat analysis, there was no mortality benefit in patients randomized to rhythm control in the AFFIRM trial.
      • Wyse D.G.
      • Waldo A.L.
      • DiMarco J.P.
      • Domanski M.J.
      • Rosenberg Y.
      • Schron E.B.
      • Kellen J.C.
      • Greene H.L.
      • Mickel M.C.
      • Dalquist J.E.
      • Corley S.D.
      A comparison of rate control and rhythm control in patients with atrial fibrillation.
      However, rhythm control was attempted only with antiarrhythmic drugs, the proarrhythmic effects of which may have negated any beneficial effects on survival.
      • Corley S.D.
      • Epstein A.E.
      • DiMarco J.P.
      • et al.
      Relationships between sinus rhythm, treatment, and survival in the Atrial Fibrillation Follow-Up Investigation of Rhythm Management (AFFIRM) Study.
      Furthermore, SR could be maintained in only <30% of the patients. Post hoc analyses from the AFFIRM and DIAMOND trials suggested a survival benefit in patients who remained in SR.
      • Corley S.D.
      • Epstein A.E.
      • DiMarco J.P.
      • et al.
      Relationships between sinus rhythm, treatment, and survival in the Atrial Fibrillation Follow-Up Investigation of Rhythm Management (AFFIRM) Study.
      • Pedersen O.D.
      • Bagger H.
      • Keller N.
      • Marchant B.
      • Kober L.
      • Torp-Pedersen C.
      Efficacy of dofetilide in the treatment of atrial fibrillation-flutter in patients with reduced left ventricular function: a Danish investigations of arrhythmia and mortality on dofetilide (diamond) substudy.
      Prior randomized studies demonstrated superior efficacy of RFA in maintaining SR than antiarrhythmic drug therapy in patients with AF.
      • Oral H.
      • Pappone C.
      • Chugh A.
      • et al.
      Circumferential pulmonary-vein ablation for chronic atrial fibrillation.
      • Wilber D.J.
      • Pappone C.
      • Neuzil P.
      • et al.
      Comparison of antiarrhythmic drug therapy and radiofrequency catheter ablation in patients with paroxysmal atrial fibrillation: a randomized controlled trial.
      The ongoing CABANA trial (NCT00911508) will address the effect of RFA on clinical outcomes. However, the results will not be available for several years.
      In this study, overall mortality was low compared to predicted mortality in the general population with AF,
      • Piccini J.P.
      • Hammill B.G.
      • Sinner M.F.
      • Jensen P.N.
      • Hernandez A.F.
      • Heckbert S.R.
      • Benjamin E.J.
      • Curtis L.H.
      Incidence and prevalence of atrial fibrillation and associated mortality among Medicare beneficiaries, 1993–2007.
      which is consistent with prior reports (0%–6.5%).
      • Nademanee K.
      • Schwab M.C.
      • Kosar E.M.
      • Karwecki M.
      • Moran M.D.
      • Visessook N.
      • Michael A.D.
      • Ngarmukos T.
      Clinical outcomes of catheter substrate ablation for high-risk patients with atrial fibrillation.
      • Pappone C.
      • Rosanio S.
      • Augello G.
      • et al.
      Mortality, morbidity, and quality of life after circumferential pulmonary vein ablation for atrial fibrillation: outcomes from a controlled nonrandomized long-term study.
      • Bunch T.J.
      • May H.T.
      • Bair T.L.
      • Weiss J.P.
      • Crandall B.G.
      • Osborn J.S.
      • Mallender C.
      • Anderson J.L.
      • Muhlestein B.J.
      • Lappe D.L.
      • Day J.D.
      Atrial fibrillation ablation patients have long-term stroke rates similar to patients without atrial fibrillation regardless of CHADS2 score.
      • Ouyang F.
      • Tilz R.
      • Chun J.
      • Schmidt B.
      • Wissner E.
      • Zerm T.
      • Neven K.
      • Kokturk B.
      • Konstantinidou M.
      • Metzner A.
      • Fuernkranz A.
      • Kuck K.H.
      Long-term results of catheter ablation in paroxysmal atrial fibrillation: lessons from a 5-year follow-up.
      • Hunter R.J.
      • McCready J.
      • Diab I.
      • et al.
      Maintenance of sinus rhythm with an ablation strategy in patients with atrial fibrillation is associated with a lower risk of stroke and death.
      • Sonne K.
      • Patel D.
      • Mohanty P.
      • et al.
      Pulmonary vein antrum isolation, atrioventricular junction ablation, and antiarrhythmic drugs combined with direct current cardioversion: survival rates at 7 years follow-up.
      Possible explanations include selection of healthier and younger individuals with fewer comorbidities. More vigorous follow-up after RFA may have led to better medical care.
      A novel finding of this study is that maintenance of SR after RFA was associated with a 60% decrease in cardiovascular mortality. AF can contribute to excess cardiac mortality by increasing the risk of congestive heart failure and coronary heart disease.
      • Benjamin E.J.
      • Wolf P.A.
      • D’Agostino R.B.
      • Silbershatz H.
      • Kannel W.B.
      • Levy D.
      Impact of atrial fibrillation on the risk of death: the Framingham Heart Study.
      • Krahn A.D.
      • Manfreda J.
      • Tate R.B.
      • Mathewson F.A.
      • Cuddy T.E.
      The natural history of atrial fibrillation: incidence, risk factors, and prognosis in the Manitoba Follow-Up Study.
      However, a significant improvement in all-cause mortality was not observed. The lack of effect on all-cause mortality may be related to a smaller overall effect of SR on this outcome and possible competing causes of death. A larger sample size might have provided sufficient power to detect a significant difference in all-cause mortality.

      Predictors of mortality

      Excess mortality associated with AF can be attributed to an increase in the risk of CVEs, congestive heart failure, and complications related to therapy.
      • Benjamin E.J.
      • Wolf P.A.
      • D’Agostino R.B.
      • Silbershatz H.
      • Kannel W.B.
      • Levy D.
      Impact of atrial fibrillation on the risk of death: the Framingham Heart Study.
      Furthermore, some of the risk factors for AF, such as hypertension and age, are more prevalent in patients with coronary artery and peripheral vascular disease. Therefore, comorbidities associated with AF may also increase the risk of mortality. In this study, age, ejection fraction, diabetes mellitus, persistent AF, coronary artery disease, OSA, and history of CVE were predictors of excess CVEs and of all-cause and cardiac mortality. However, SR remained as a significant and strong predictor of improved cardiac survival after correcting for baseline covariates on multivariable analysis. This finding challenges the notion that AF is simply another surrogate for underlying disease and other comorbidities. Rather, it suggests that elimination of AF by RFA is likely to improve survival.
      Proarrhythmia due to antiarrhythmic therapy has been implicated in the increased mortality associated with a rhythm control strategy in prior studies.
      • Wyse D.G.
      • Waldo A.L.
      • DiMarco J.P.
      • Domanski M.J.
      • Rosenberg Y.
      • Schron E.B.
      • Kellen J.C.
      • Greene H.L.
      • Mickel M.C.
      • Dalquist J.E.
      • Corley S.D.
      A comparison of rate control and rhythm control in patients with atrial fibrillation.
      • Corley S.D.
      • Epstein A.E.
      • DiMarco J.P.
      • et al.
      Relationships between sinus rhythm, treatment, and survival in the Atrial Fibrillation Follow-Up Investigation of Rhythm Management (AFFIRM) Study.
      An association with antiarrhythmic or anticoagulant use was not observed in this study. However, both the event rate and the number of patients who were on long-term antiarrhythmic drug therapy were small.
      A reduction in all-cause mortality was observed after repeat procedures. These findings suggest that maintenance of SR after RFA may have a significant impact on long-term outcomes despite the need for several procedures to achieve SR. Further studies are needed to evaluate the role of repeat ablations on long-term outcomes after RFA.

      CVEs and rhythm status

      The modest reduction in the risk of CVEs with maintenance of SR after RFA was not statistically significant in the present study, regardless of concomitant anticoagulant therapy. However, there were few CVEs (2.3%) in this study. Healthier patients with a lower risk of stroke may have been included in this study; ~75% of patients had a CHADS2 score of 0 or 1. Therefore, the magnitude of benefit from SR on CVEs in these patients is likely to be small. The combination of small net effect and low event rates would require a larger sample size to detect small differences in outcomes. It should be noted that prior reports from large registries that reported a beneficial effect of RFA on the risk of stroke made comparisons to historical cohorts whose risk of stroke appears to be much higher than the risk reported in recent studies. This most likely is due to patient selection and improvements in contemporary care of patients with AF.
      • Bunch T.J.
      • May H.T.
      • Bair T.L.
      • Weiss J.P.
      • Crandall B.G.
      • Osborn J.S.
      • Mallender C.
      • Anderson J.L.
      • Muhlestein B.J.
      • Lappe D.L.
      • Day J.D.
      Atrial fibrillation ablation patients have long-term stroke rates similar to patients without atrial fibrillation regardless of CHADS2 score.
      • Oral H.
      • Chugh A.
      • Ozaydin M.
      • et al.
      Risk of thromboembolic events after percutaneous left atrial radiofrequency ablation of atrial fibrillation.

      Previous studies

      Prior reports from several registries assessed the effect of RFA on mortality in patients with AF.
      • Nademanee K.
      • Schwab M.C.
      • Kosar E.M.
      • Karwecki M.
      • Moran M.D.
      • Visessook N.
      • Michael A.D.
      • Ngarmukos T.
      Clinical outcomes of catheter substrate ablation for high-risk patients with atrial fibrillation.
      • Pappone C.
      • Rosanio S.
      • Augello G.
      • et al.
      Mortality, morbidity, and quality of life after circumferential pulmonary vein ablation for atrial fibrillation: outcomes from a controlled nonrandomized long-term study.
      • Bunch T.J.
      • Crandall B.G.
      • Weiss J.P.
      • May H.T.
      • Bair T.L.
      • Osborn J.S.
      • Anderson J.L.
      • Muhlestein J.B.
      • Horne B.D.
      • Lappe D.L.
      • Day J.D.
      Patients treated with catheter ablation for atrial fibrillation have long-term rates of death, stroke, and dementia similar to patients without atrial fibrillation.
      In a few of those reports, comparisons were made to historical cohorts whose mortality rate was likely to be higher than that of the patients included in the studies, as discussed earlier. Furthermore, rhythm status, based on recurrence of AF, was determined as a 1-time terminal event and was used to predict mortality using Cox regression and Kaplan-Meier analysis.
      • Hunter R.J.
      • McCready J.
      • Diab I.
      • et al.
      Maintenance of sinus rhythm with an ablation strategy in patients with atrial fibrillation is associated with a lower risk of stroke and death.
      However, rhythm status often is dynamic after RFA in patients with AF. In a Cox model, the rhythm status should be evaluated for each person at risk at the time of each death event. Using a future event (rhythm status after RFA) to predict another future event (mortality) does not satisfy the requirements of survival analysis and Cox model. In this study, the dynamic nature of the rhythm status was recognized and was analyzed as a time-dependent variable.

      Study limitations

      Arrhythmia burden was assessed by careful and frequent monitoring and assessment of patient encounters, which also prompted ECG monitoring. It would have been ideal to use an implantable loop recorder; however, this was not feasible in >3000 patients followed for >10 years. Therefore, recurrences of AF may have been underestimated. Stringent criteria were used when adjudicating rhythm status for each interval. This may have led to an underestimation of amount of time spent in SR, which in turn could underestimate the beneficial effects of SR after RFA. Modes of death were determined based on death certificates and medical records and therefore are limited by the accuracy of the information provided.

      Conclusion

      Maintenance of SR after RFA is associated with a reduction in cardiac mortality. Whether maintenance of SR would also lower the risk of all-cause mortality and CVEs, particularly in the absence of concomitant anticoagulant therapy, remains to be determined in large and long-term studies using continuous monitoring of the rhythm status.

      References

        • Benjamin E.J.
        • Wolf P.A.
        • D’Agostino R.B.
        • Silbershatz H.
        • Kannel W.B.
        • Levy D.
        Impact of atrial fibrillation on the risk of death: the Framingham Heart Study.
        Circulation. 1998; 98: 946-952
        • Piccini J.P.
        • Hammill B.G.
        • Sinner M.F.
        • Jensen P.N.
        • Hernandez A.F.
        • Heckbert S.R.
        • Benjamin E.J.
        • Curtis L.H.
        Incidence and prevalence of atrial fibrillation and associated mortality among Medicare beneficiaries, 1993–2007.
        Circ Cardiovasc Qual Outcomes. 2012; 5: 85-93
        • Oral H.
        • Chugh A.
        • Yoshida K.
        • et al.
        A randomized assessment of the incremental role of ablation of complex fractionated atrial electrograms after antral pulmonary vein isolation for long-lasting persistent atrial fibrillation.
        J Am Coll Cardiol. 2009; 53: 782-789
        • Steinberg J.S.
        • Sadaniantz A.
        • Kron J.
        • et al.
        Analysis of cause-specific mortality in the Atrial Fibrillation Follow-up Investigation of Rhythm Management (AFFIRM) study.
        Circulation. 2004; 109: 1973-1980
        • Wyse D.G.
        • Waldo A.L.
        • DiMarco J.P.
        • Domanski M.J.
        • Rosenberg Y.
        • Schron E.B.
        • Kellen J.C.
        • Greene H.L.
        • Mickel M.C.
        • Dalquist J.E.
        • Corley S.D.
        A comparison of rate control and rhythm control in patients with atrial fibrillation.
        N Engl J Med. 2002; 347: 1825-1833
        • Corley S.D.
        • Epstein A.E.
        • DiMarco J.P.
        • et al.
        Relationships between sinus rhythm, treatment, and survival in the Atrial Fibrillation Follow-Up Investigation of Rhythm Management (AFFIRM) Study.
        Circulation. 2004; 109: 1509-1513
        • Pedersen O.D.
        • Bagger H.
        • Keller N.
        • Marchant B.
        • Kober L.
        • Torp-Pedersen C.
        Efficacy of dofetilide in the treatment of atrial fibrillation-flutter in patients with reduced left ventricular function: a Danish investigations of arrhythmia and mortality on dofetilide (diamond) substudy.
        Circulation. 2001; 104: 292-296
        • Oral H.
        • Pappone C.
        • Chugh A.
        • et al.
        Circumferential pulmonary-vein ablation for chronic atrial fibrillation.
        N Engl J Med. 2006; 354: 934-941
        • Wilber D.J.
        • Pappone C.
        • Neuzil P.
        • et al.
        Comparison of antiarrhythmic drug therapy and radiofrequency catheter ablation in patients with paroxysmal atrial fibrillation: a randomized controlled trial.
        JAMA. 2010; 303: 333-340
        • Nademanee K.
        • Schwab M.C.
        • Kosar E.M.
        • Karwecki M.
        • Moran M.D.
        • Visessook N.
        • Michael A.D.
        • Ngarmukos T.
        Clinical outcomes of catheter substrate ablation for high-risk patients with atrial fibrillation.
        J Am Coll Cardiol. 2008; 51: 843-849
        • Pappone C.
        • Rosanio S.
        • Augello G.
        • et al.
        Mortality, morbidity, and quality of life after circumferential pulmonary vein ablation for atrial fibrillation: outcomes from a controlled nonrandomized long-term study.
        J Am Coll Cardiol. 2003; 42: 185-197
        • Bunch T.J.
        • May H.T.
        • Bair T.L.
        • Weiss J.P.
        • Crandall B.G.
        • Osborn J.S.
        • Mallender C.
        • Anderson J.L.
        • Muhlestein B.J.
        • Lappe D.L.
        • Day J.D.
        Atrial fibrillation ablation patients have long-term stroke rates similar to patients without atrial fibrillation regardless of CHADS2 score.
        Heart Rhythm. 2013; 10: 1272-1277
        • Ouyang F.
        • Tilz R.
        • Chun J.
        • Schmidt B.
        • Wissner E.
        • Zerm T.
        • Neven K.
        • Kokturk B.
        • Konstantinidou M.
        • Metzner A.
        • Fuernkranz A.
        • Kuck K.H.
        Long-term results of catheter ablation in paroxysmal atrial fibrillation: lessons from a 5-year follow-up.
        Circulation. 2010; 122: 2368-2377
        • Hunter R.J.
        • McCready J.
        • Diab I.
        • et al.
        Maintenance of sinus rhythm with an ablation strategy in patients with atrial fibrillation is associated with a lower risk of stroke and death.
        Heart. 2012; 98: 48-53
        • Sonne K.
        • Patel D.
        • Mohanty P.
        • et al.
        Pulmonary vein antrum isolation, atrioventricular junction ablation, and antiarrhythmic drugs combined with direct current cardioversion: survival rates at 7 years follow-up.
        J Interv Card Electrophysiol. 2009; 26: 121-126
        • Krahn A.D.
        • Manfreda J.
        • Tate R.B.
        • Mathewson F.A.
        • Cuddy T.E.
        The natural history of atrial fibrillation: incidence, risk factors, and prognosis in the Manitoba Follow-Up Study.
        Am J Med. 1995; 98: 476-484
        • Oral H.
        • Chugh A.
        • Ozaydin M.
        • et al.
        Risk of thromboembolic events after percutaneous left atrial radiofrequency ablation of atrial fibrillation.
        Circulation. 2006; 114: 759-765
        • Bunch T.J.
        • Crandall B.G.
        • Weiss J.P.
        • May H.T.
        • Bair T.L.
        • Osborn J.S.
        • Anderson J.L.
        • Muhlestein J.B.
        • Horne B.D.
        • Lappe D.L.
        • Day J.D.
        Patients treated with catheter ablation for atrial fibrillation have long-term rates of death, stroke, and dementia similar to patients without atrial fibrillation.
        J Cardiovasc Electrophysiol. 2011; 22: 839-845

      Linked Article

      • ERRATUM
        Heart RhythmVol. 12Issue 1
        • Preview
          In the article titled: Mortality and cerebrovascular events after radiofrequency catheter ablation of atrial fibrillation by Hamid Ghanbari, Kazım Bas¸er, Krit Jongnarangsin, Aman Chugh, Brahmajee K. Nallamothu, Brenda W. Gillespie, Hatice Duygu Bas¸er, Arisara Swangasool, Thomas Crawford, Rakesh Latchamsetty, Eric Good, Frank Pelosi Jr, Frank Bogun, Fred Morady, Hakan Oral, that published in the September issue of HeartRhythm journal (2014; 11: 1503-1511) there was a typo in the name of Arisara Suwanagool, MD.
        • Full-Text
        • PDF