Sinus rhythm restoration and arrhythmia noninducibility are major predictors of arrhythmia-free outcome after ablation for long-standing persistent atrial fibrillation: A prospective

Sinus rhythm restoration and arrhythmia noninducibilityare major predictors of arrhythmia-free outcome afterablation for long-standing persistent atrial fibrillation:A prospective studyMartin Fiala, MD, PhD,*† Veronika Bulková, MSc, PhD,‡ Libor Škňouřil, MD,†

Renáta Nevřalová, MD,† Ondřej Toman, MD, PhD,* Jaroslav Januška, MD, PhD,†

Jindřich Špinar, MD, PhD,* Dan Wichterle, MD, PhD§

From the *Department of Internal Medicine and Cardiology, University Hospital, Brno, Czech Republic,†Department of Cardiology, Hospital Podlesí, Třinec, Czech Republic, ‡Department of Cardiology andAngiology, St. Anne’s University Hospital and International, Clinical Research Center, Brno, Czech Republicand §Department of Cardiology, Institute for Clinical and Experimental Medicine, Prague, Czech Republic.

BACKGROUND The impact of restoring sinus rhythm (SR) by initialablation in patients with long-standing persistent atrial fibrillation(LSPAF) is not fully established.

OBJECTIVE The purpose of this study was to investigate theprognostic value of SR restoration at the initial procedure andarrhythmia noninducibility at the final repeat procedure for long-term outcome.

METHODS A total of 203 patients (22% female; age 59 � 9 years)underwent stepwise catheter ablation for LSPAF.

RESULTS The procedural end-point of SR restoration was achievedin 50% of patients. During follow-up (median 48 months) and after1.7 procedures per patient, 72% of patients were free fromarrhythmia off antiarrhythmic drugs. Failure to restore SR wasindependently predicted by left atrial (LA) long-axis diameterZ68mm (relative risk [RR] 1.55, P¼ .03], proportion of high-voltage LAsites o20% (RR 1.62, P ¼ .02), and left atrial appendage (LAA)atrial fibrillation cycle length (AFCL) o155 ms (RR 1.5, P ¼ .05).Arrhythmia recurrence after the initial procedure was predicted bySR nonrestoration (RR 2.99, P o.000001) and LAA AFCL Z155 ms(RR 1.90, P ¼ .0002). Arrhythmia recurrence after the finalprocedure was predicted by SR nonrestoration at the initialprocedure (RR 2.83, P ¼ .0007), persistent AF duration Z24

This study was supported by grants from the Czech Ministry of HealthNR 9143-3/2007 and NS/9684-4/2008. Address reprint requestsand correspondence: Dr. Martin Fiala, Department of Cardiology andAngiology, University Hospital, Jihlavská 53, Brno, 659 91, CzechRepublic. E-mail address: [email protected].

1547-5271/$-see front matter B 2015 Heart Rhythm Society. All rights reserved.

months (RR 2.74, P ¼ .002), LAA outflow velocity o40 cm/s (RR2.21, P¼ .006), and LAA AFCLZ155 ms (RR 1.92, P¼ .02). In 115patients with repeat procedure(s), failure to achieve arrhythmianoninducibility at the final procedure (19% of patients) wasassociated with arrhythmia recurrence (RR 8.9, P o .000001).

CONCLUSION SR restoration at the initial procedure and arrhyth-mia noninducibility at the last repeat procedure were majorpredictors of arrhythmia-free outcome after ablation for LSPAF.

KEYWORDS Atrial fibrillation; Long-standing persistent atrialfibrillation; Sinus rhythm restoration; Noninducibility; Outcome

ABBREVIATIONS AAD ¼ antiarrhythmic drug; AF ¼ atrialfibrillation; ANOVA ¼ 1-way analysis of variance; AT ¼ atrialtachycardia; CI ¼ confidence interval; CL ¼ cycle length; CS ¼coronary sinus; CTI ¼ cavotricuspid isthmus; LA ¼ left atrium;LAA ¼ left atrial appendage; LMI ¼ lateral mitral isthmus;LSPAF ¼ long-standing persistent atrial fibrillation; LV ¼ leftventricle; NT-proBNP¼ N-terminal prohormone of brain natriureticpeptide; PV ¼ pulmonary vein; RA ¼ right atrium; RR ¼ relativerisk; SR ¼ sinus rhythm

(Heart Rhythm 2015;12:687–698) I 2015 Heart Rhythm Society. Allrights reserved.

IntroductionPredictors of successful catheter ablation for persistent atrialfibrillation (AF) were investigated primarily in mixedpopulations of subjects having AF continuous for 7 days toyears.1–6 Limited data are available for patients with purely

long-standing persistent AF (LSPAF). Arrhythmogenic sub-strate progression as a result of continuous AF durationinversely relates with the success rate of restoring sinusrhythm (SR) by ablation and favorable long-term out-come.2,7 Prior studies in LSPAF patients did not confirm aprognostic value of SR restoration by ablation.8,9

The present study sought to identify predictors of (1)failure to restore SR by ablation, (2) long-term AF or atrialtachycardia (AT) recurrence after the initial ablation, and (3)long-term AF/AT recurrence after the final (ie, initial orrepeat) ablation in consecutive subjects with LSPAF.

http://dx.doi.org/10.1016/j.hrthm.2015.01.004

Heart Rhythm, Vol 12, No 4, April 2015688

MethodsStudy populationThis single-center study analyzed a prospective registry of all203 consecutive patients who underwent their first AF ablationbetween July 2006 and December 2011 and met all of thefollowing criteria: (1) age 18–80 years; (2) symptomaticLSPAF lasting 412 months without intervening SR; (3)refractory to oral amiodarone; and (4) resistant to electricalcardioversion or recurrence within 7 days after cardioversion.Patients with tachycardia-induced cardiomyopathy, which wassuspected clinically after careful review of medical records,were eligible for the study. Patients with left ventricular (LV)dysfunction and significant coronary artery disease or sus-pected idiopathic dilated cardiomyopathy were not included.Ethical approval by the Institutional Review Board wasobtained, and all patients gave written informed consent.

Preablation examinationBefore the initial procedure, transthoracic/transesophagealechocardiography and blood sampling for N-terminal pro-hormone of brain natriuretic peptide (NT-proBNP) levelwere performed. Standardized left atrial (LA) and LVdiameters were measured. LV ejection fraction was assessedby the Simpson method from the cycle exhibiting maximumLV filling, and left atrial appendage (LAA) outflow velocitywas defined as the peak value within the 30-secondrecording.

Electrophysiologic study, electroanatomicmapping, and ablationPeriprocedural anticoagulation regimen, mapping/ablationtechnology, and strategy were described in detail else-where.10 In brief, point-by-point ablation with an irrigated-tip catheter (NaviStar ThermoCool, Biosense Webster,Diamond Bar, CA) navigated in preacquired 3-dimensionalelectroanatomic maps (CARTO, Biosense Webster) of theLA and right atrium (RA) was performed off warfarin, withclass I or III antiarrhythmic drugs (AADs) discontinuedexcept for amiodarone, target activated clotting time of 300–400 seconds, irrigation at 20 mL/min, temperature 421C, andpower limit of 35 W (20–25 W inside the coronarysinus [CS]).

Mandatory wide-area pulmonary vein (PV) isolation andablation of the LA roof, lateral mitral isthmus (LMI), andcavotricuspid isthmus (CTI) were performed in all patients.When AF continued, a stepwise approach (consisting ofCS ablation/isolation and additional linear or clusterelectrogram-guided ablation) was successively used in theLA and RA. Ablation at the high interatrial septum and alongthe Bachmann bundle was deferred as the last step in order tominimize the risk of LAA activation delay or isolation.

The stepwise ablation strategy was pursued until AFconverted directly into SR or into AT, which was sub-sequently identified and ablated by activation and entrain-ment mapping. If ablation failed to restore SR, intravenouspropafenone, overdrive pacing, and electrical cardioversion

were successively applied. At the end of the procedure, PVisolation was confirmed by the circular catheter (Lasso,Biosense Webster), and efforts were made to achievebidirectional conduction block (if not already present) atthe CTI, LMI, and roof line. LA volumes and bipolar voltagecharacteristics were derived from electroanatomic maps asdescribed previously.11 The dominant atrial fibrillation cyclelength (AFCL) in the LAA was assessed by averagingmultiple consecutive cycles.3,12

Postablation managementLow-molecular-weight heparin was administered after theprocedure until the resumption of warfarin treatment. Class Ior III AADs were discontinued at the 3-month visit in casesof uneventful follow-up. Warfarin was stopped after 6months in patients with stable SR, preserved LAA function(outflow velocity Z40 cm/s), and absence of other con-ditions requiring permanent anticoagulation.

At repeat procedures, PVs were reisolated and incidentalgaps in linear lesions at the roof line, LMI and CTI wereclosed, even if passive in the mechanism of recurrentarrhythmia. Arrhythmia noninducibility was tested by atrialincremental pacing up to 300 bpm for at least 5 times aftereach arrhythmia termination in all patients. Isoproterenol andadenosine were used in 63 and 43 patients who were notinducible by pacing.

Follow-upPatients were seen in the outpatient department every 3months during the first year and every 6 months thereafter.ECG documentation consisted of 12-lead ECG and 24-hourHolter monitoring before each visit. In addition, 3-weektranstelephonic episodic ECG monitoring was performedtwice per year up to 2 years and annually thereafter.Arrhythmia recurrence after the 3-month blanking periodincluded ECG-documented AF/AT as well as palpitationssuggestive of AF/AT lasting430 seconds unless arrhythmiawas objectively excluded. “Good arrhythmia control” statusat the end of follow-up was defined as AF/AT absence in thepast 6 months and r1 electrical/pharmacologic cardiover-sion in the past 2 years.

Statistical analysisContinuous variables are expressed as mean� SD or medianwith interquartile range. Categorical variables are expressedas percentages. The relationship of continuous variables wasanalyzed by Pearson correlation, and differences betweensubgroups of patients were compared by 1-way or multi-factorial analysis of variance (ANOVA). Association of allavailable clinical and procedure-related characteristics with 3end-points (SR nonrestoration by ablation, and AF/ATrecurrence after the initial ablation and after the finalablation) were investigated by linear or Cox proportionalhazards regression models, as appropriate. For this purpose,continuous variables were dichotomized at the median orother clinically meaningful value. All factors, which were

Table 1 Baseline patient characteristics

Variable Value

Females 45 (22%)Age (years) 59 � 9 (32–79)Total AF duration (months) Median 60; IQR 34–96 (13–504)Persistent AF duration (months) Median 26; IQR 17–46 (13–254)DC cardioversion failed torestore SR

96 (47%)

Amiodarone treatment present 70 (34%)ACEI/ARB treatment 138 (68%)LV ejection fraction r40% 39 (19%)Hypertension 123 (61%)Diabetes mellitus 29 (14%)Stroke/transient ischemicattack

26 (13%)

Vascular disease 14 (7%)Mitral regurgitation (grade) 1.3 � 0.8 (0–4)Hypothyroidism 19 (9%)Bronchopulmonary disease 27 (14%)CHADS2 score 1.2 � 1.0 (0–4)CHA2DS2-VASc score 1.8 � 1.4 (0–6)Body surface area (m2) 2.1 � 0.2 (1.5–2.7)Body mass index (kg/m2) 30.9 � 4.5 (19.4–43.8)LAA peak outflow velocity(cm/s)

44 � 20 (13–105)

LA long-axis diameter (mm) 68 � 7 (52–90)LA short-axis diameter (mm) 47 � 6 (31–66)LA anteroposterior diameter(mm)

48 � 6 (34–69)

LV ejection fraction (%) 53 � 10 (25–67)LVED diameter (mm) 53 � 7 (37–74)LVES diameter (mm) 38 � 8 (21–67)NT-proBNP (pg/mL) Median 903; IQR 601–1407 (82–

7447)

Data are given as mean � SD or median with interquartile range (IQR)and minimum–maximum range (in parentheses) or counts with percentages.

ACEI ¼ angiotensin-converting enzyme inhibitor; AF ¼ atrial fibrilla-tion; ARB ¼ angiotensin II receptor blocker; DC ¼ direct current; LA ¼ leftatrium; LAA ¼ left atrial appendage; LV ¼ left ventricle; LVED ¼ leftventricular end-diastolic; LVES¼ left ventricular end-systolic; NT-proBNP¼N-terminal prohormone of brain natriuretic peptide; SR ¼ sinus rhythm.

689Fiala et al Prognostic Role of Atrial Fibrillation Termination

univariately associated (P o .10) with the end-points wereanalyzed multivariately by the stepwise forward method.Kaplan–Meier curves were used to display cumulativearrhythmia-free survival. P o .05 was considered signifi-cant. All analyses were performed using STATISTICAsoftware (version 12 , Statsoft Inc, Tulsa, OK, USA).

Table 2 Parameters at initial procedure

Parameter Value

LA mapping points 212 � 41 (128–387)LA volume (mL) 178 � 37 (102–297)Proportion of LA points o0.2 mV (%) 26 � 18 (2–86)Proportion of LA points 41.0 mV (%) 19 � 12 (1–71)LA maximum voltage (mV) 5.3 � 2.4 (1.2–13.8)LAA AFCL (ms) 159 � 27 (106–270)Procedural time (minutes) 321 � 54 (180–430)Fluoroscopy time (minutes) 24 � 8 (10–57)RF energy application time (minutes) 130 � 33 (46–198)

Data are given shown as mean � SD and minimum–maximum range (inparentheses).

AFCL ¼ atrial fibrillation cycle length; LA ¼ left atrial; LAA ¼ left atrialappendage; RF ¼ radiofrequency.

ResultsInitial and repeat ablation resultsBaseline clinical and initial mapping characteristics aresummarized in Tables 1 and 2. Patients had continuous AFof median duration 42 years, and 19% presented withsignificant LV systolic dysfunction. They had severelyremodeled LA with median volume �180 mL and extensiveareas of scarring as assessed by electroanatomic mapping.

Initial ablation terminated AF (ie, converted eitherdirectly into SR or into AT) in 138 patients (68%). SR wasrestored by ablation (ie, directly from AF or from inter-mediate AT) in 101 patients (50%; Figure 1). Of the

remaining 102 patients with failed SR restoration by ablationitself, SR was restored by intravenous propafenone/over-drive pacing (n¼ 10) or by electrical cardioversion (n¼ 86),or SR could not be achieved at all (n ¼ 7). The sites of AFtermination were in the LA (n ¼ 109), including PV antra(n ¼ 27), CS (n ¼ 19), and RA (n ¼ 10). SR was restoreddirectly from AF in 22 patients (11%) or via intermediate ATin 79 patients (39%); by ablation of a localized source in theLA (n ¼ 64), including PV antra (n ¼ 16), RA (n ¼ 9), andCS (n ¼ 6); or by ablation of a perimitral AT (n ¼ 14) orroof-dependent AT (n ¼ 8). Roof-line and LMI blocks werecompleted in 178 (88%) patients and 176 patients (87%),respectively.

One repeat procedure was performed in 115 patients(57%), 2 repeat procedures in 33 (16%), and 3 repeatprocedures in 3 (1%), accounting for 1.74 procedures perpatient. Recurrent AF or AT, noninducibility status, andoutcomes of repeat procedures are shown in Figure 1. Atfirst, second, and third repeat procedure, ablation restored SRin 83%, 85%, and 100% patients, respectively. Macroreen-trant AT, localized AT, or both types of AT were found in28%, 49%, and 23% of patients with restored SR at firstrepeat procedure, respectively, and in 29%, 65%, and 6% ofpatients with restored SR at second repeat procedure,respectively. In total, 56 macroreentrant ATs (23 roof-dependent ATs, 33 perimitral ATs) and 82 localized ATs(8 from PV antra, 49 in extra-PV LA areas, 13 within CS,and 12 in RA) were targeted at first repeat ablation, and 13macroreentrant ATs (8 roof-dependent, 5 peri-mitral), and 26localized ATs (16 in extra-PV LA areas, 7 within CS, and 3in RA) were ablated at second repeat procedure. Only asingle localized AT source from LA/CS was found at allthird repeat procedures.

At the last repeat procedure, SR was restored in 102patients (89%). Noninducibility testing was sequentiallyperformed by incremental pacing in all 102 patients,isoproterenol infusion in 63 (62%) of the 93 patients withnoninducibility by pacing, and adenosine bolus in 43 of the63 patients (68%) with noninducibility by pacing andisoproterenol. Pharmacologic challenge did not identifyany inducible patient other than those inducible by pacing,so noninducibility was achieved in 93 patients (81%).

1st Ablation

2nd Ablation

SRR+ (n=101)

AF/AT Reabl –n=9 (2)(9%)

AF/AT Reabl +

n=46(45.5%)

NoAF/ATn=46

(45.5%)

AF/AT Reabl –n=5 (1)(12%)

AF/AT Reabl +

n=5 (12%)

NoAF/ATn=31 (76%)

AF/AT Reabl –n=1 (1)(20%)

AF/AT Reabl +

n=2 (40%)

NoAF/AT

n=2(40%)

3rd Ablation NI+ (n=4) NI– (n=1)

AF/AT Reabl +

n=1(25%)

NoAF/AT

n=3 (75%)

AF/AT Reabl +

n=1(100%)

NoAF/AT

n=1(100%)

AF/AT Reabl –

n=1(100%)

NI+ (n=1) NI– (n=1)

NI+ (n=1)

NoAF/AT

n=2 (100%)

4th Ablation NI+ (n=1)

ArrhythmiaAF 13AT 28

AF 2AT 3

NI+ (n=41) NI– (n=5)

AF 19 (35%)AT 36 (65%)

Arrhythmia AT 1AF 3 (23%)AT 10 (77%)AT 4 AF 1 AT 1

AF 1 (14%)AT 6 (86%)Arrhythmia AT 1AT 1 AT 3 (100%)

AF 4AT 5

AF 1AT 4

AT 1

AF 1

1st Ablation

2nd Ablation

SRR– (n=102)

AF/AT Reabl –n=21 (9)(20%)

AF/AT Reabl +

n=69(68%)

NoAF/ATn=12(12%)

AF/AT Reabl –n=7 (1)(17%)

AF/AT Reabl +

n=10(24%)

NoAF/ATn=25(59%)

AF/AT Reabl –n=8 (5)(30%)

AF/AT Reabl +

n=16(59%)

NoAF/AT

n=3(11%)

3rd Ablation NI+ (n=10)

NoAF/ATn=10

(100%)

AF/AT Reab l+

n=1(13%)

NoAF/AT

n=8(100%)

AF/AT Reabl –n=4 (1)(50%)

NI+ (n=8) NI– (n=8)

NoAF/AT

n=1(100%)

4th Ablation NI+ (n=1)

AF 1 (14%)AT 6 (86%)

ArrhythmiaAF 15AT 27

AF 16AT 11

NI+ (n=42) NI– (n=27)

AF 46 (51%)AT 44 (49%)

ArrhythmiaAF 13 (32%)AT 28 (68%)

AF 1 (14%)AT 6 (86%)Arrhythmia AT 1AT 5 (100%)

NoAF/AT

n=3(37%)

AF 15AT 6

AF 2AT 5

AF 2AT 8

AF 4AT 4

AF 4AT 4

AF 1AT 7

AT 4

Heart Rhythm, Vol 12, No 4, April 2015690

Figure 2 Predictors of study end-points. Association of baseline and procedure-related factors with study end-points is plotted as relative risks with 95%confidence intervals. Only predictors with Po .10 for at least 1 of the 3 end-points are included. AF ¼ atrial fibrillation; AT ¼ tachycardia; CL¼ cycle length;LA ¼ left atrial; LAA ¼ left atrial appendage; LVES ¼ left ventricular end-systolic; SR ¼ sinus rhythm.

691Fiala et al Prognostic Role of Atrial Fibrillation Termination

Five major complications without permanent sequelae (allduring initial ablation) included 2 transient ischemic attacks,cardiac tamponade from steam pop, hemothorax from sub-clavian vein bleeding, and transitory femoral neurapraxia.

Long-term maintenance of sinus rhythmDuring the total median follow-up of 48 months (interquar-tile range 36–60 months, range 6–80 months) and lastprocedure median follow-up of 38 months (interquartilerange 25–49 months, range 6–74 months), 3 deaths (1.5%)not related to the ablation procedure occurred after 13, 6, and76 months since the initial procedure due to hemorrhagicstroke on optimal warfarin, newly diagnosed pancreascancer, and pneumonia in the setting of pulmonary hyper-tension, respectively.

After the initial and the last procedures, 58 patients (29%)and 146 patients (72%) were AF/AT-free off class I/III AADs,respectively. After the last ablation, only 20 patients (10%)continued to have persistent (n¼ 14) or paroxysmal (n¼ 6) AF/AT and declined further procedures (Figure 1). Good arrhythmiacontrol at the end of follow-up was present in 183 (90%)patients, including 150 patients (74%) without class I/III AADs.

Prediction of SR nonrestoration andAF/AT recurrenceUnivariate analysis of all factors for all end-points issummarized in Appendix Tables 1–3 and in Figure 2. On

Figure 1 Scheme of ablation procedures, end-points, and outcomes. Two panelspatients with sinus rhythm restoration by ablation (SRRþ); (B) for the rest of potachycardia (AT) or pharmacologic/electrical cardioversion of atrial fibrillation (AFfurther organized according to noninducibility (NI) data at repeat procedures. Patidistributed into 3 outcome levels: no AF/AT recurrence (No AF/AT), AF/AT recurfurther repeat procedure (AF/AT Reabl–). Bold numbers in parentheses denote “Rremaining “Reabl–“ patients attained good arrhythmia control. Recurrent arrhythnumbers summarize the counts and proportions of recurrent AF/AT at each level

multivariate analysis, only dilated LA in the long axis(relative risk [RR] 1.55, 95% confidence interval [CI]1.03–2.34, P ¼ .03), low proportion of high-voltage LAsites (RR 1.62, 95% CI 1.08–2.43, P ¼ .02), and shortdominant AFCL in LAA (RR 1.5, 95% CI 1.00–2.27, P ¼.05) were independent predictors of failure to restore SR(Appendix Table 1).

SR nonrestoration (RR 2.99, 95% CI 2.10–4.25,P o .000001) and short dominant LAA AFCL (RR 0.53,95% CI 0.38–0.74, P ¼ .0002) were the only independentpredictors of AF/AT recurrence after the initial procedure(Figures 3A and 4A, and Appendix Table 2). For LAAAFCL, this could be interpreted inversely as long CL beingpredictive for AF/AT recurrence (RR 1.90, 95% CI 1.35–2.66, P ¼ .0002).

Four baseline factors—SR nonrestoration (RR 2.83, 95%CI 1.56–5.15, P ¼ .0007), short dominant LAA AFCL (RR0.52, 95% CI 0.30–0.91, P ¼ .02), long duration of LSPAF(RR 2.74, 95% CI 1.46–5.14, P ¼ .002), and low LAAoutflow velocity (RR 2.21, 95% CI 1.25–3.91, P ¼ .006)—were independent predictors of AF/AT recurrence after thelast procedure (Figures 3B and 4B, and Appendix Table 3).Again for LAA AFCL, this could be interpreted inversely aslong CL being predictive for AF/AT recurrence (RR 1.92,95% CI 1.10–3.34, P ¼ .02).

The dominant impact of restored SR and LAA AFCL isbest demonstrated by clinical outcome in patients from the

are composed according to ablation success at the index procedure: (A) forpulation (SRR–), that is, those with overdrive pacing termination of atrial)/AT (n ¼ 96) or failure to achieve sinus rhythm at all (n ¼ 7). Flowchart isents with completed (SRRþ, NIþ) and failed (SRR–, NI–) end-points wererence with repeat ablation (AF/AT Reablþ), and AF/AT recurrence with noeabl–” patients with clinically significant AF/AT recurrences, whereas themias in each outcome subgroup are shown in the light gray frames. Boldof the study.

Figure 3 Cumulative arrhythmia-free survival. Population is stratified according to sinus rhythm (SR) restoration at initial ablation. Kaplan–Meier curves aredisplayed for the outcome after the initial (left) and the last (right) procedure.

Heart Rhythm, Vol 12, No 4, April 2015692

most diverse subgroups. In patients with SR restoration andLAA AFCL o155 ms (n ¼ 43), the AF/AT recurrence ratewas only 35% after the initial procedure and decreasedfurther to 5% after the final procedure. In contrast, 97% and44% of patients with SR nonrestoration and LAA AFCLZ155 ms (n ¼ 39) experienced AF/AT recurrence after theinitial and the final procedure, respectively.

Figure 5 shows how the magnitude of LA remodeling interms of LA volume (ANOVA P ¼ .03), proportions of LApoints o0.2 mV (ANOVA P ¼ .0001) or 41.0 mV(ANOVA P ¼ .004), and LAA outflow velocity (ANOVAP ¼ .11) correlated with combination of 3 factors—dichotomized LAA AFCL at baseline, SR restoration at theinitial procedure, and AF/AT-free outcome after singleprocedure. Overall, the strongest correlation was foundbetween LAA AFCL and proportions of LA points o0.2mV (Pearson R ¼ 0.42, P o .000001).

In a separate analysis of 115 patients with repeatprocedure(s), failure to achieve AF/AT noninducibility at

Figure 4 Cumulative arrhythmia-free survival. Population is stratified according(CL) at initial ablation. Kaplan–Meier curves for all combinations of stratifiers are dLAA ¼ left atrial appendage.

the last repeat procedure (19% of patients) was stronglyassociated with AF/AT recurrence (RR 8.9, 95% CI 95%4.0–19.6, P o .000001). Outcome of patients with restoredSR but AF/AT inducibility was comparable to those inwhom AF/AT was not terminated at all during the last repeatprocedure (Figure 6).

DiscussionThis study sought to identify the baseline clinical andprocedure-related predictors of AF/AT recurrence afterablation for LSPAF in patients resistant to amiodarone andfailed electrical cardioversion. Because of possible differ-ences in atrial remodeling, arrhythmogenic substrate, SRrestoration rates, and clinical outcome, it may not beadequate to extrapolate the results of risk stratification fromearlier studies, including shorter-lasting persistent AF ontothe subjects with LSPAF.2,7 The impact of ablation-inducedAF termination on clinical outcome has been controversial

to sinus rhythm (SR) restoration and dominant atrial fibrillation cycle lengthisplayed for the outcome after the initial (left) and the last (right) procedure.

Figure 5 Indices of left atrial (LA) remodeling categorized by left atrial appendage (LAA) atrial fibrillation cycle length (AF CL) and outcome measures.Magnitude of LA remodeling is shown for 8 combinations of 3 binary factors: dichotomized LAAAFCL at baseline, restoration of sinus rhythm (SR) by ablationat the initial procedure, and atrial fibrillation/atrial tachycardia (AF/AT)–free outcome after single procedure. Presence and absence of individual factors isindicated by colored þ/– signs, respectively. ANOVA ¼ analysis of variance; CI ¼ confidence interval.

693Fiala et al Prognostic Role of Atrial Fibrillation Termination

for all types of persistent AF, let alone LSPAF; therefore,consensus on the ideal procedure end-point has not yet beenreached. In addition, the extent of atrial ablation, which is

Figure 6 Cumulative arrhythmia-free survival after the last repeatprocedure. Subpopulation (n ¼ 115) is stratified according to sinus rhythm(SR) restoration at the last repeat procedure. Subgroup with SR restoration isfurther stratified according to arrhythmia inducibility. Kaplan–Meier curvesare displayed for all 3 subgroups. AF ¼ atrial fibrillation; AT ¼ atrialtachycardia.

often required to terminate persistent AF, has been repeatedlyquestioned, although recent studies (including reports from thispopulation) showed that benefits from restored SR in terms ofhemodynamic, functional, and quality-of life-improvementoutweigh possible deleterious effects of extensive persistentAF/LSPAF ablation.10,13–15 The major findings of this studyare as follows. (1) Nonrestoration of SR by ablation at the initialprocedure is the dominant factor associated not only withclinical failure of the primary procedure but also with AF/ATrecurrence after subsequent repeat procedure(s). (2) Failure toachieve noninducibility of AF/AT at the last repeat procedure isalso strongly predictive for AF/AT recurrence, with a riskcomparable to that associated with AF/AT nontermination.

Prediction of SR restoration at initial procedureThere is growing body of evidence that SR restoration is thelikely fundamental procedural end-point of persistent AF/LSPAF ablation, which limits the risk of AF/AT recurrenceto the recovery of existing source(s) and/or to the develop-ment of new source(s) due to natural disease progression orproarrhythmic effect of ablation lesions. In prior studies ofmixed persistent AF populations or small series of selectedpatients with LSPAF, this end-point was predicted by shorterduration of continuous AF,1 smaller LA size,5 better LAA

Heart Rhythm, Vol 12, No 4, April 2015694

mechanical function,16,17 or preserved LV ejection frac-tion.18 We confirmed lower SR restoration rate in patientswith shorter LAA AFCL, which had been earlier explainedby an inverse relationship between AFCL and the number ofdistinct AF sources.1,3,5,18,19 Surprisingly, we further foundthat the long-axis LA diameter outweighed other parametersof LA size, including LA volume. We can speculate that such“tall” LA shapes may be susceptible to ablation failurebecause of the greater mass of potentially arrhythmogenicatrial myocardium between the isolated venous compartmentand the mitral annulus. Finally, association between thelower proportion of high-voltage LA sites and failure torestore SR is another new finding suggesting that LAscarring may be a more relevant marker of the LA substratethan simple LA volume.

Prediction of AF/AT recurrencesTwo large studies of LSPAF patients failed to confirm clinicalbenefit from restored SR.8,9 However, AF termination wasattempted in only 37 of 205 patients with failed cardioversionafter PV isolation, and the data on SR restoration rate were notprovided in 1 study.8 In the second study using PV isolationand complex fractionated electrogram ablation, the low SRrestoration rate (14%) offered limited statistical power todetect the association with outcome.9 Our observation is inaccordance with the results of studies in patients having allforms of persistent AF2,4,6,18,12–22 and confirms that thepredictive value of restored SR at initial ablation extendsbeyond repeat procedures.4,5,22

In patients with persistent AF, prior investigations ofpreprocedural factors identified smaller LA size,2,22 shortercontinuous AF duration,1,2,4,8,21 younger age,5,23 male gender,4

absence of congestive heart failure,4 absence of hypertension,5

higher LAA outflow velocity,15 LA systolic strain, and LAAwall velocity22 as the predictors of arrhythmia-free outcomeafter the initial or repeat procedure(s). In our study, SRrestoration at the initial procedure overpowered baseline clinicalfactors on multivariate analysis, which suggests that even anadvanced arrhythmogenic substrate can be abolished by exten-sive ablation. The impact of longer duration of continuous AFor lower LAA outflow velocity—the expected natural predic-tors of worse outcome8,17—emerged only after repeat proce-dures when the most refractory AF patients were unmasked.

Previous studies of persistent AF ablation linked longerAFCL with AF termination and, to some extent, with betterclinical outcome.1,5 However, recent observation of the bestsingle-procedure outcome in persistent AF patients withbaseline AFCL between 155 and 200 ms compared to thosewith shorter but also with longer AFCL suggests a morecomplex relationship between AFCL and outcome meas-ures.24 Based on prior data, the authors speculated that verylong AFCL resulted from extensive atrial fibrosis and, there-fore, was associated with worse clinical results.25,26 Theapparent paradox between positive impact of longer AFCLon SR restoration and its negative impact on arrhythmia-freeoutcome in our study is not in conflict with these observations.

Longer LAA AFCLmay facilitate the search for persistent AFsources and improve acute ablation efficacy. Beyond certainlimits, however, and more likely in LSPAF, very long LAAAFCL as a marker of advanced LA remodeling may competewith procedural success and result in poorer clinical outcome.From the clinical perspective, such patients with long LAAAFCL should be differentiated from less diseased subjects inwhom long LAA AFCL may have reflected a rather localizedsubstrate within PV antra with decremental conduction to arelatively healthy LA body.3

Noninducibility of AF/AT after repeat procedureThe end-point of arrhythmia noninducibility at the end of thefinal repeat procedures was a very strong (the only statisti-cally independent) predictor of clinical outcome. Using bothSR restoration and arrhythmia noninducibility at the initialprocedure was considered largely unrealistic because of timeconstraints in such a technically demanding procedure. Thatis why noninducibility was systematically tested only atrepeat procedures when the recurrent arrhythmias weregenerally simpler and the ablation less extensive.

Study limitationsPostablation arrhythmia-free survival may be affected byunderestimating asymptomatic recurrences. However, per-sistent AT/AF as the dominant mode of recurrent arrhythmiawas well recognized by instructed patients, and recurrencesdisclosed by longer-term ECG monitoring correlated sur-prisingly well with subjective rhythm perception.

SR restoration during procedures might be consideredrandom (ie, not related to a specific source ablation).However, incidental arrhythmia termination was highlyunlikely because prior attempts at direct current cardiover-sion either failed to restore SR in nearly half or restored onlyshort-lived SR in the other half of patients.

Noninducibility testing did not include pharmacologicchallenge in 32% of the patients who were noninducible withincremental pacing. However, following noninducibility byatrial pacing, pharmacologic challenge was of no help inunmasking residual AF/AT sources.

Finally, this was a single-center study of LSPAF patientsablated by experienced operators. Therefore, the general-izability of our results should be interpreted with caution.

ConclusionThis study demonstrated the role of SR restoration by theinitial ablation and noninducibility at repeat procedures forlong-term freedom from arrhythmia recurrences in patientswith LSPAF, suggesting that extensive atrial ablation shouldbe completed with strict procedural end-points to minimizethe risk of proarrhythmia and enhance the maintenance ofstable SR associated with possible benefits in hemodynam-ics, functional status, and quality of life.

Appendix ASee Table E1–E3

Table E1 Predictors of SR nonrestoration by ablation at initial procedure

Risk factor

SR nonrestorationrate in low-riskgroup

SR nonrestorationrate in high-riskgroup

Univariate risk of SRnonrestoration

Multivariate risk of SRnonrestoration

N/total % N/total % RR 95% CI P value RR 95% CI P value

Females 80/158 51% 22/45 49% 0.97 0.61–1.54 .88Age Z60 years 59/113 52% 43/90 48% 0.92 0.62–1.35 .66Total AF duration Z60 months 52/101 51% 50/102 49% 0.95 0.64–1.41 .80Persistent AF duration Z24 months 36/87 41% 66/116 57% 1.36 0.92–2.01 .12Amiodarone treatment present 69/133 52% 33/70 47% 0.91 0.61–1.36 .65ACEI/ARB treatment 33/65 51% 69/38 50% 0.98 0.65–1.49 .94LV ejection fraction r40% 85/164 52% 17/39 44% 0.85 0.53–1.37 .51Hypertension 36/80 45% 66/123 54% 1.19 0.80–1.76 .39Diabetes 84/174 48% 18/29 62% 1.36 0.73–2.55 .33Stroke/transient ischemic attack 92/177 52% 10/26 38% 0.78 0.46–1.33 .36Vascular disease 95/189 50% 7/14 50% 0.99 0.46–2.14 .99Mitral regurgitation grade Z1.5 70/129 54% 32/74 43% 0.81 0.54–1.20 .28Pulmonary disease 86/176 49% 16/27 59% 1.26 0.67–2.35 .48Body surface area Z2.1 m2 39/90 43% 63/113 56% 1.28 0.87–1.89 .21Body mass index Z30 kg/m2 44/95 46% 58/108 54% 1.16 0.79–1.71 .46LAA peak outflow velocity o40 cm/s 46/100 46% 56/103 54% 1.18 0.80–1.75 .40LA long-axis diameter Z68 mm 38/101 38% 63/101 62% 1.66 1.11–2.48 .01 1.55 1.03–2.34 .03LA short-axis diameter Z47 mm 45/98 46% 56/104 54% 1.17 0.79–1.73 .43LA anteroposterior diameter Z48 mm 39/93 42% 63/110 57% 1.36 0.92–2.01 .12LV ejection fraction o57% 57/110 52% 45/93 48% 0.93 0.63–1.38 .73LVED diameter Z52 mm 39/93 42% 63/110 57% 1.36 0.92–2.01 .12LVES diameter Z36 mm 46/94 49% 55/107 51% 1.05 0.71–1.56 .80NT-proBNP Z900 pg/mL 51/99 52% 51/104 49% 0.95 0.64–1.41 .80CARTO LA volume Z175 mL 46/101 46% 56/102 55% 1.21 0.82–1.79 .35Proportion of LA points o0.2 mV Z25% 47/106 44% 55/97 57% 1.29 0.87–1.91 .21Proportion of LA points 41.0 mV o20% 33/85 39% 69/118 58% 1.47 1.00–2.18 .05 1.62 1.08–2.43 .02LA maximum voltage o5.0 mV 46/102 45% 56/101 55% 1.23 0.83–1.82 .30LAA AFCL o155 ms 39/97 40% 63/106 59% 1.47 0.99–2.19 .05 1.50 1.00–2.27 .05SR nonrestoration at first ablation NA NA NA NA NA NA NA NA NA NA

ACEI¼ angiotensin-converting enzyme inhibitor; AF¼ atrial fibrillation; AFCL¼ atrial fibrillation cycle length; ARB¼ angiotensin II receptor blocker; CI¼confidence interval; LA¼ left atrial; LAA¼ left atrial appendage; LV¼ left ventricle; LVED¼ left ventricular end-diastolic; LVES¼ left ventricular end-systolic;NA = not applicable; NT-proBNP ¼ N-terminal prohormone of brain natriuretic peptide; RR ¼ relative risk; SR ¼ sinus rhythm.

695Fiala et al Prognostic Role of Atrial Fibrillation Termination

Table E2 Predictors of AF/AT recurrence after initial procedure

AF/AT recurrence after firstablation in low-risk group

AF/AT recurrence after firstablation in high-risk group

Univariate risk of AF/AT recurrenceafter first ablation

Multivariate risk of AF/AT recurrenceafter first ablation

Risk factor N/total % N/total % RR 95% CI P value RR 95% CI P value

Females 109/158 69% 34/45 76% 1.17 0.80–1.72 .42Age Z60 years 77/113 68% 66/90 73% 1.06 0.76–1.48 .72Total AF duration Z60 months 61/101 60% 82/102 80% 1.54 1.10–2.14 .01Persistent AF duration Z24 months 54/87 62% 89/116 77% 1.41 1.00–1.97 .05Amiodarone treatment present 87/133 65% 56/70 80% 1.33 0.95–1.86 .10ACEI/ARB treatment 45/65 69% 98/138 71% 1.02 0.71–1.45 .92LV ejection fraction r40% 120/164 73% 23/39 59% 0.70 0.45–1.10 .12Hypertension 53/80 66% 90/123 73% 1.19 0.85–1.67 .31Diabetes mellitus 120/174 69% 23/29 79% 1.26 0.80–1.96 .32Stroke/transient ischemic attack 124/177 70% 19/26 73% 1.15 0.71–1.86 .57Vascular disease 135/189 71% 8/14 57% 0.77 0.38–1.57 .48Mitral regurgitation grade Z1.5 91/129 71% 52/74 70% 0.96 0.68–1.34 .80Bronchopulmonary disease 120/176 68% 23/27 85% 1.58 1.01–2.47 .05Body surface area Z2.1 m2 65/90 72% 78/113 69% 0.93 0.67–1.29 .66Body mass index Z30 kg/m2 68/95 72% 75/108 69% 0.96 0.69–1.33 .81LAA peak outflow velocity o40 cm/s 65/100 65% 78/103 76% 1.33 0.96–1.86 .09LA long-axis diameter Z68 mm 64/101 63% 78/101 77% 1.45 1.04–2.01 .03LA short-axis diameter Z47 mm 71/98 72% 71/104 68% 0.92 0.66–1.28 .61LA anteroposterior diameter Z48 mm 56/93 60% 87/110 79% 1.64 1.17–2.29 .004LV ejection fraction o57% 76/110 69% 67/93 72% 1.13 0.81–1.57 .46LVED diameter Z52 mm 69/93 74% 74/110 67% 0.94 0.68–1.31 .73LVES diameter Z36 mm 70/94 74% 71/107 66% 0.91 0.65–1.26 .55NT-proBNP Z900 pg/mL 70/99 71% 73/104 70% 0.99 0.71–1.38 .96CARTO LA volume Z175 mL 66/101 65% 77/102 75% 1.34 0.96–1.86 .08Proportion of LA points o0.2 mV Z25% 68/106 64% 75/97 77% 1.55 1.11–2.15 .01Proportion of LA points 41.0 mV o20% 52/85 61% 91/118 77% 1.54 1.10–2.17 .01LA maximum voltage o5.0 mV 65/102 64% 78/101 77% 1.46 1.05–2.03 .02LAA AFCL o155 ms 77/97 79% 66/106 62% 0.65 0.47–0.91 .01 0.53 0.38–0.74 .0002SR nonrestoration at first ablation 54/101 53% 89/102 87% 2.62 1.86–3.70 o.000001 2.99 2.10–4.25 o.000001

ACEI ¼ angiotensin-converting enzyme inhibitor; AF ¼ atrial fibrillation; AFCL ¼ atrial fibrillation cycle length; ARB ¼ angiotensin II receptor blocker; AT ¼ atrial tachycardia; CI ¼ confidence interval; LA ¼ leftatrial; LAA ¼ left atrial appendage; LV ¼ left ventricle; LVED ¼ left ventricular end-diastolic; LVES ¼ left ventricular end-systolic; NT-proBNP ¼ N-terminal prohormone of brain natriuretic peptide; RR ¼ relative risk;SR ¼ sinus rhythm.

Heart

Rhythm,Vol12,

No4,

April2015696

Table E3 Predictors of AF/AT recurrence after last procedure

Risk factor

AF/AT recurrence after lastablation in low–risk group

AF/AT recurrence after lastablation in high–risk group

Univariate risk of AF/AT recurrenceafter last ablation

Multivariate risk of AF/AT recurrenceafter last ablation

N/total % N/total % RR 95% CI P value RR 95% CI P value

Females 37/158 23% 17/45 38% 1.96 1.10–3.49 .02Age Z60 years 30/113 27% 24/90 27% 1.05 0.61–1.80 .85Total AF duration Z60 months 15/101 15% 39/102 38% 2.96 1.63–5.37 .0004Persistent AF duration Z24 months 13/87 15% 41/116 35% 2.65 1.42–4.94 .002 2.74 1.46–5.14 .002Amiodarone treatment present 32/133 24% 22/70 31% 1.31 0.76–2.26 .33ACEI/ARB treatment 15/65 23% 39/138 28% 1.22 0.67–2.21 .52LV ejection fraction r40% 47/164 29% 7/39 18% 0.63 0.28–1.39 .25Hypertension 14/80 18% 40/123 33% 1.94 1.06–3.57 .03Diabetes mellitus 44/174 25% 10/29 34% 1.37 0.69–2.72 .37Stroke/transient ischemic attack 47/177 27% 7/26 27% 0.93 0.42–2.06 .86Vascular disease 53/189 28% 1/14 7% 0.22 0.03–1.62 .14Mitral regurgitation grade Z1.5 29/129 22% 25/74 34% 1.66 0.97–2.83 .06Bronchopulmonary disease 46/176 26% 8/27 30% 1.23 0.58–2.61 .59Body surface area Z2.1 m2 27/90 30% 27/113 24% 0.76 0.45–1.30 .32Body mass index Z30 kg/m2 23/95 24% 31/108 29% 1.21 0.70–2.08 .49LAA peak outflow velocity o40 cm/s 19/100 19% 35/103 34% 2.29 1.31–4.02 .004 2.21 1.25–3.91 .006LA long–axis diameter Z68 mm 21/101 21% 32/101 32% 1.69 0.97–2.94 .06LA short–axis diameter Z47 mm 24/98 24% 29/104 28% 1.19 0.69–2.05 .53LA antero–posterior diameter Z48 mm 20/93 22% 34/110 31% 1.59 0.92–2.77 .10LV ejection fraction o57% 32/110 29% 22/93 24% 0.82 0.47–1.41 .47LVED diameter Z52 mm 27/93 29% 27/110 25% 0.86 0.51–1.47 .59LVES diameter Z36 mm 31/94 33% 21/107 20% 0.57 0.33–1.00 .05NT–proBNP Z900 pg/mL 25/99 25% 29/104 28% 1.25 0.73–2.14 .41CARTO LA volume Z175 mL 25/101 25% 29/102 28% 1.21 0.71–2.06 .49Proportion of LA points o0.2 mV Z25% 21/106 20% 33/97 34% 2.03 1.17–3.51 .01Proportion of LA points 41.0 mV o20% 18/85 21% 36/118 31% 1.56 0.89–2.75 .12LA maximum voltage o5.0 mV 23/102 23% 31/101 31% 1.45 0.84–2.49 .18LAA AFCL o155 ms 31/97 32% 23/106 22% 0.64 0.37–1.09 .10 0.52 0.30–0.91 .02SR nonrestoration at first ablation 16/101 16% 38/102 37% 2.78 1.55–4.99 .000006 2.83 1.56–5.15 .0007

ACEI ¼ angiotensin-converting enzyme inhibitor; AF ¼ atrial fibrillation; AFCL ¼ atrial fibrillation cycle length; ARB ¼ angiotensin II receptor blocker; AT ¼ atrial tachycardia; CI ¼ confidence interval; LA ¼ leftatrial; LAA ¼ left atrial appendage; LV ¼ left ventricle; LVED ¼ left ventricular end-diastolic; LVES ¼ left ventricular end-systolic; NT-proBNP ¼ N-terminal prohormone of brain natriuretic peptide; RR ¼ relative risk;SR ¼ sinus rhythm.

697Fiala

etal

PrognosticRole

ofAtrialFibrillation

Termination

Heart Rhythm, Vol 12, No 4, April 2015698

References1. Matsuo S, Lellouche N, Wright M, et al. Clinical predictors of termination and

clinical outcome of catheter ablation for persistent atrial fibrillation. J Am CollCardiol 2009;54:788–795.

2. O'Neill MD, Wright M, Knecht S, et al. Long-term follow-up of persistent atrialfibrillation ablation using termination as a procedural endpoint. Eur Heart J2009;30:1105–1112.

3. Drewitz I, Willems S, Salukhe TV, Steven D, Hoffmann BA, Servatius H, Bock K,Aydin MA, Wegscheider K, Meinertz T, Rostock T. Atrial fibrillation cycle lengthis a sole independent predictor of a substrate for consecutive arrhythmias in patientswith persistent atrial fibrillation. Circ Arrhythm Electrophysiol 2010;3:351–360.

4. Rostock T, Salukhe TV, Steven D, et al. Long-term single- and multiple-procedure outcome and predictors of success after catheter ablation for persistentatrial fibrillation. Heart Rhythm 2011;8:1391–1397.

5. Heist EK, Chalhoub F, Barrett C, Danik S, Ruskin JN, Mansour M. Predictors ofatrial fibrillation termination and clinical success of catheter ablation of persistentatrial fibrillation. Am J Cardiol 2012;15:545–551.

6. Rostock T, Salukhe TV, Hoffmann BA, Steven D, Berner I, Müllerleile K,Theis C, Bock K, Servatius H, Sultan A, Willems S. Prognostic role of subsequentatrial tachycardias occurring during ablation of persistent atrial fibrillation: aprospective randomized trial. Circ Arrhythm Electrophysiol 2013;6:1059–1065.

7. Haïssaguerre M, Hocini M, Denis A, et al. Driver domains in persistent atrialfibrillation. Circulation 2014;130:530–538.

8. Tilz RR, Rillig A, Thum AM, Arya A, Wohlmuth P, Metzner A, Mathew S,Yoshiga Y, Wissner E, Kuck KH, Ouyang F. Catheter ablation of long-standingpersistent atrial fibrillation: 5-year outcomes of the Hamburg sequential ablationstrategy. J Am Coll Cardiol 2012;60:1921–1929.

9. Elayi CS, Di Biase L, Barrett C, et al. Atrial fibrillation termination as aprocedural endpoint during ablation in long-standing persistent atrial fibrillation.Heart Rhythm 2010;9:1216–1223.

10. Fiala M, Wichterle D, Bulková V, Škňouřil L, Nevřalová R, Toman O, Dorda M,Januška J, Špinar J. A prospective evaluation of hemodynamics, functional status,and quality of life after radiofrequency catheter ablation of long-standingpersistent atrial fibrillation. Europace 2014;16:15–25.

11. Fiala M, Wichterle D, Chovančík J, Bulková V, Wojnarová D, Nevřalová R,Januška J. Left atrial voltage during atrial fibrillation in paroxysmal and persistentatrial fibrillation patients. Pacing Clin Electrophysiol 2010;33:541–548.

12. Haïssaguerre M, Sanders P, Hocini M, Hsu LF, Shah DC, Scavée C, TakahashiY, Rotter M, Pasquié JL, Garrigue S, Clémenty J, Jaïs P. Changes in atrialfibrillation cycle length and inducibility during catheter ablation and their relationto outcome. Circulation 2004;109:3007–3013.

13. Jones DJ, Haldar SK, Hussain W, Sharma R, Francis DP, Rahman-Haley SL,McDonagh TA, Underwood SR, Markides V, Wong T. A randomized trial toassess catheter ablation versus rate control in the management of persistent atrialfibrillation in heart failure. J Am Coll Cardiol 2013;61:1894–1903.

14. Hunter RJ, Berriman TJ, Diab I, et al. A randomized controlled trial of catheterablation versus medical treatment of atrial fibrillation in heart failure (theCAMTAF Trial). Circ Arrhythm Electrophysiol 2014;7:31–38.

15. Bulková V, Fiala M, Havránek S, Šimek J, Škňouřil L, Januška J, Špinar J,Wichterle D. Improvement in quality of life after catheter ablation for paroxysmalversus long-standing persistent atrial fibrillation: a prospective study with 3-yearfollow-up. J Am Heart Assoc 2014;3:e000881.

16. Kumagai K, Sakamoto T, Nakamura K, Hayano M, Yamashita E, Oshima S. Pre-procedural prediction of termination of persistent atrial fibrillation by catheterablation as an indicator of reverse remodeling of the left atrium. Circ J 2013;77:1416–1423.

17. Combes S, Jacob S, Combes N, Karam N, Chaumeil A, Guy-Moyat B, Treguer F,Deplagne A, Boveda S, Marijon E, Albenque JP. Predicting favorable outcomesin the setting of radiofrequency catheter ablation of long-standing persistent atrialfibrillation: a pilot study assessing the value of left atrial appendage peak flowvelocity. Arch Cardiovasc Dis 2013;106:36–43.

18. Faustino M, Pizzi C, Capuzzi D, Agricola T, Costa GM, Flacco ME, Marzuillo C,Nocciolini M, Capasso L, Manzoli L. Impact of atrial fibrillation terminationmode during catheter ablation procedure on maintenance of sinus rhythm. HeartRhythm 2014;11:1528–1535.

19. Haïssaguerre M, Lim KT, Jacquemet V, Rotter M, Fang L, Hocini M, Matsuo S,Knecht S, Jaïs P, Virag N. Atrial fibrillatory cycle length: computer simulationand potential clinical importance. Europace 2007;9(Suppl 6):64–70.

20. Ammar S, Hessling G, Reents T, Paulik M, Fichtner SA, Schön P, Dillier R,Kathan S, Jilek C, Kolb C, Haller B, Deisenhofer I. Importance of sinus rhythm asendpoint of persistent atrial fibrillation ablation. J Cardiovasc Electrophysiol2013;24:388–395.

21. Komatsu Y, Taniguchi H, Miyazaki S, Nakamura H, Kusa S, Uchiyama T, KakitaK, Kakuta T, Hachiya H, Iesaka Y. Impact of atrial fibrillation termination onclinical outcome after ablation in relation to the duration of persistent atrialfibrillation. Pacing Clin Electrophysiol 2012;35:1436–1443.

22. Miyazaki S, Taniguchi H, Komatsu Y, Uchiyama T, Kusa S, Nakamura H,Hachiya JH, Isobe M, Hirao K, Iesaka Y. Sequential biatrial linear defragmenta-tion approach for persistent atrial fibrillation. Heart Rhythm 2013;10:338–346.

23. Yoshida K, Rabbani AB, Oral H, Bach D, Morady F, Chugh A. Left atrial volumeand dominant frequency of atrial fibrillation in patients undergoing catheter ablationof persistent atrial fibrillation. J Interv Card Electrophysiol 2011;32:155–161.

24. Ammar S, Hessling G, Paulik M, Reents T, Dillier R, Buiatti A, Semmler V, KolbC, Haller B, Deisenhofer I. Impact of baseline atrial fibrillation cycle length onacute and long-term outcome of persistent atrial fibrillation ablation. J Interv CardElectrophysiol 2014;41:253–259.

25. Mahnkopf C, Badger TJ, Burgon NS, Daccarett M, Haslam TS, Badger CT,McGann CJ, Akoum N, Kholmovski E, Macleod RS, Marrouche NF. Evaluationof the left atrial substrate in patients with lone atrial fibrillation using delayed-enhanced MRI: implications for disease progression and response to catheterablation. Heart Rhythm 2010;7:1475–1481.

26. Yokokawa M, Latchamsetty R, Good E, Crawford T, Jongnarangsin K, Pelosi FJr, Bogun F, Oral H, Morady F, Chugh A. The impact of age on the atrialsubstrate: insights from patients with a low scar burden undergoing catheterablation of persistent atrial fibrillation. J Interv Card Electrophysiol 2012;34:287–294.

CLINICAL PERSPECTIVESIn light of increasing evidence of significant hemodynamic and functional improvement from restored SR in patients withpersistent AF/LSPAF, this study encourages efforts to pursue strict electrophysiologic end-points during extensive ablationfor LSPAF. This may improve long-term SR maintenance despite the ubiquitous phenomenon of recovery of the existingAF sources or proarrhythmic effect of prior ablation. Restoration of SR can be understood as functionally effectiveelimination of all currently active AF sources and perpetuators, which not only improves the arrhythmia-free outcome of thefirst ablation but also increases the efficacy of repeat ablation through simpler recurrent arrhythmias that are usually easierto ablate than recurrent AF. Arrhythmia noninducibility, which often is an unrealistic end-point of the initial ablation, playsa major role in predicting outcome after repeat procedures with independent predictive power on top of arrhythmiatermination. For this reason, we believe that the initial anatomically guided ablation steps in patients with LSPAF must befollowed by an individually tailored ablation in order to achieve arrhythmia termination and noninducibility. A lesssystematic approach with vaguely defined or easier to achieve electrophysiologic end-points, mainly relying on electricalcardioversion and postablation reverse remodeling, may provide a suboptimal outcome. Our approach should beinvestigated head to head in future randomized studies using novel sophisticated AF mapping tools that offer a moreselective approach to LSPAF sources. These techniques may better preserve LA mechanical function and enhance globalhemodynamic and functional benefits if comparable ablation efficacy could be achieved with significantly reducedradiofrequency energy delivery.