J A C C : C L I N I C A L E L E C T R O P H Y S I O L O G Y V O L . 1 , N O . 3 , 2 0 1 5

ª 2 0 1 5 B Y T H E A M E R I C A N C O L L E G E O F C A R D I O L O G Y F O U N D A T I O N

P U B L I S H E D B Y E L S E V I E R I N C .

I S S N 2 4 0 5 - 5 0 0 X / $ 3 6 . 0 0

h t t p : / / d x . d o i . o r g / 1 0 . 1 0 1 6 / j . j a c e p . 2 0 1 5 . 0 2 . 0 1 8

Catheter Ablation for Atrial Fibrillationin Heart Failure PatientsA Meta-Analysis of Randomized, Controlled Trials

Shadi Al Halabi, MD, MPH,* Mohammed Qintar, MD,* Ayman Hussein, MD,* M. Chadi Alraies, MD,yDavid G. Jones, MD,zx Tom Wong, MD,zx Michael R. MacDonald, MD,k Mark C. Petrie, MD,{ Daniel Cantillon, MD,*Khaldoun G. Tarakji, MD, MPH,* Mohamed Kanj, MD,* Mandeep Bhargava, MD,* Niraj Varma, MD,*Bryan Baranowski, MD,* Bruce L. Wilkoff, MD,* Oussama Wazni, MD,* Thomas Callahan, MD,* Walid Saliba, MD,*Mina K. Chung, MD*

ABSTRACT

Fro

of

da

kCUn

co

ha

Ma

OBJECTIVES The objective of the study was to compare rate control versus atrial fibrillation (AF) catheter ablation

strategies in patients with AF and heart failure (HF).

BACKGROUND Rhythm control with antiarrhythmic drugs (AADs) is not superior to rate control in patients with HF and

AF, but AF ablation may be more successful at achieving rhythm control than are AADs. However, risks for both ablation

and AADs are probably higher and success rates lower in patients with HF.

METHODS We conducted a meta-analysis of trials that randomized HF patients (left ventricular ejection fraction

[LVEF] <50%) with AF to a rate control or AF catheter ablation strategy and reported changes in LVEF, quality of life,

6-min walk test, or peak oxygen consumption. Study quality and heterogeneity were assessed through the use of Jadad

scores and Cochran’s Q statistics, respectively. Mantel-Haenszel relative risks and mean differences were calculated

through the use of random effect models.

RESULTS Four trials (N ¼ 224) met inclusion criteria; 82.5% (n ¼ 185) had persistent AF. AF ablation was asso-

ciated with an increase in LVEF (mean difference, 8.5%; 95% confidence interval [CI]: 6.4% to 10.7%; p < 0.001)

compared with rate control. AF ablation was superior in improving quality of life by Minnesota Living With Heart

Failure (MLWHF) questionnaire scores (mean difference, –11.9; 95% CI: –17.2 to 6.6; p < 0.001). Peak oxygen

consumption and 6-min walk distance increased in AF ablation compared with rate-control patients (mean difference,

3.2; 95% CI: 1.1 to 5.3; p ¼ 0.003; mean difference, 34.8; 95% CI: 2.9 to 66.7; p ¼ 0.03, respectively). In the

persistent AF subgroup, LVEF and MLWHF were significantly improved with AF ablation. Major adverse event

rates (risk ratio: 1.3; 95% CI: 0.4 to 3.9; p ¼ 0.64) were not significantly different. No significant heterogeneity

was evident.

CONCLUSIONS In patients with HF and AF, AF catheter ablation is superior to rate control in improving LVEF, quality

of life, and functional capacity. Before accepting a rate-control strategy in HF patients with persistent or drug-refractory

AF, consideration should be given to AF ablation. (J Am Coll Cardiol EP 2015;1:200–9) © 2015 by the American College

of Cardiology Foundation.

m the *Department of Cardiovascular Medicine, Heart and Vascular Institute, Cleveland Clinic, Cleveland, Ohio; yDepartment

Cardiology, University of Minnesota, Minneapolis, Minnesota; zRoyal Brompton and Harefield National Health Service Foun-

tion Trust, London, United Kingdom; xNational Heart and Lung Institute, Imperial College London, London, United Kingdom;

hangi, National Hospital, Glasgow, Scotland, United Kingdom; and the {Golden Jubilee National Hospital, Glasgow, Scotland,

ited Kingdom. Dr. Cantillon has received consulting/honoraria from St. Jude Medical. Dr. Tarakji has received speaker/

nsulting honorarium from Medtronic and Spectranetics. Dr. Chung was supported by NIH grant R01HL111314. All other authors

ve reported that they have no relationships relevant to the contents of this paper to disclose.

nuscript received February 11, 2015; accepted February 26, 2015.

AB BR E V I A T I O N S

AND ACRONYM S

AAD = anti-arrhythmic drug

AF = atrial fibrillation

HF = heart failure

LVEF = left ventricular

ejection fraction

MLWHF = Minnesota Living

With Heart Failure

NYHA = New York Heart

Association

PVI = pulmonary vein isolation

J A C C : C L I N I C A L E L E C T R O P H Y S I O L O G Y V O L . 1 , N O . 3 , 2 0 1 5 Al Halabi et al.J U N E 2 0 1 5 : 2 0 0 – 9 Meta-Analysis: RCTs of AF Catheter Ablation Versus Rate Control for AF in HF

201

A trial fibrillation (AF) and heart failure (HF)are 2 common cardiac conditions associatedwith substantial morbidity, mortality, and

cost on health care systems (1–4). The 2 conditionsfrequently coexist, and each may promote the other.AF is present in up to 50% of patients with HF (5).AF in HF patients is associated with increased hospi-tal stay, stroke, and mortality (6–8). This may be atleast partially attributed to the hemodynamic effectsof AF caused by loss of atrial contraction along withirregular and/or rapid ventricular rates, which canlead to left ventricular dysfunction and decreasedcardiac output (9,10).

Rhythm control with anti-arrhythmic drugs (AADs)has failed to be superior to rate control in patientswith HF and AF in terms of cardiovascular mortalityor worsening of HF (11). The risk of adverse eventsassociated with AADs and their limited efficacy inrestoring sinus rhythm have triggered an increasedinterest in AF catheter ablation (12,13).

Several observational studies of AF catheter abla-tion in patients with HF have reported that mainte-nance of sinus rhythm by catheter ablation cansignificantly improve cardiac function (14–16).

The aim of our study was to determine if AF cath-eter ablation is superior to rate control in patientswith AF and HF. We performed a meta-analysisof randomized, controlled trials that comparedAF catheter ablation with rate control in patientswith HF and AF.

METHODS

This meta-analysis of clinical trials was performedaccording to the guidelines of the Preferred ReportingItems for Systematic Reviews and Meta-Analyses(PRISMA) (17).

LITERATURE REVIEW. Relevant studies were selectedby searching PubMed, Medline, Embase, Central,ClinicalTrials.gov, The Cochrane Library, and ISIWeb of Science (January 1980 to February 2015). Thesearch was independently conducted by 2 in-vestigators (S.A. and M.Q.). Search key terms wereatrial fibrillation, persistent atrial fibrillation, pul-monary venous isolation, catheter ablation, heartfailure, left ventricular dysfunction, impaired leftventricular systolic function, reduced left ventricularsystolic function, low ejection fraction, heart failurewith reduced ejection fraction, functional capacity,and quality of life. Bibliographies of retrieved studieswere hand-searched to identify relevant studies.

SELECTION CRITERIA AND QUALITY ANALYSIS.

Prospective randomized, controlled trials published

in English that compared rate control with AFcatheter ablation were included.

Trials that included patients with leftventricular ejection fraction (LVEF) <50%,randomized to a rate-control strategy or AFcatheter ablation and that reported at leastone of the studied outcomes, were included.Studies had to fulfill the following criteriafor inclusion: 1) prospective randomized,controlled trial design; 2) patients enrolledwith LVEF <50% and history of AF; 3)randomization to AF catheter ablation versusrate control (pharmacologic or atrioventric-

ular [AV] node ablation with biventricular or rightventricular pacing); and 4) study follow-up of at least6 months. Studies that used AADs as rhythm controlor that did not report at least one outcome of interestwere excluded.

Quality of the studies was assessed independentlyby 2 reviewers (S.A. and M.Q.). Jadad’s method wasused to assess quality of the studies (18). The itemsassessed were blinding, randomization, and descrip-tion of withdrawals or dropouts. All included studieshad a quality score of 3. No studies attempted double-blinding. Therefore, weighting of the results was notperformed.

DATA EXTRACTION AND OUTCOME MEASURES.

Data were independently extracted by 2 inves-tigators (S.A. and M.Q.), and discrepancies wereresolved by unanimous consensus. Extracted dataincluded number of patients in each interventionarm, characteristics of included patients, procedurecharacteristics, LVEF, quality-of-life parameters,and complications. Primary authors were con-tacted when data on studied outcomes were notreported.

The primary clinical endpoint was change in LVEFafter 6 months. Secondary endpoints were MinnesotaLiving with Heart Failure (MLWHF) questionnairescores, 6-min walk test distance, and peak oxygenconsumption. Complications, adverse events, anddeaths were also summarized.

Major adverse events were defined as death,intracranial hemorrhage, cardiac tamponade, peri-cardial effusion, pneumothorax, hemothorax, deepvenous thrombosis or pulmonary embolism, sepsis,or pulmonary vein stenosis (>50%) requiring inter-vention. Procedural complications were defined asany complication that occurred in the 30-day post-procedural period. AF ablation was defined as pul-monary vein isolation with or without additionalsubstrate modification and excluding AV junctionablation. Rate control was defined as the use of

FIGURE 1 Preferre

(PRISMA) Diagram

Search strategy resu

Al Halabi et al. J A C C : C L I N I C A L E L E C T R O P H Y S I O L O G Y V O L . 1 , N O . 3 , 2 0 1 5

Meta-Analysis: RCTs of AF Catheter Ablation Versus Rate Control for AF in HF J U N E 2 0 1 5 : 2 0 0 – 9

202

pharmacological therapy or AV junction ablation forrate control.

STATISTICAL ANALYSIS. The results are presentedas mean difference for continuous outcome measureswith 95% confidence intervals (CIs). Random-effectmodels were used for all reported outcomes (19).

Heterogeneity among studies was assessed withthe inconsistency index (I2) statistic, which rangesfrom 0% to 100% and is defined as the percentageof the observed inter-trial variability that is due toheterogeneity rather than chance for each outcome(I2 >60% denotes significant heterogeneity) (20).Potential publication bias was evaluated by meansof Begg’s funnel plot method (21).

To further detect any clinical heterogeneity,several sensitivity analyses were performed for theLVEF and MLWHF outcomes:

d Reporting Items for Systematic Reviews and Meta-Analyses

lts and exclusion steps are shown.

1. Trials including only patients with persistent AFwere analyzed, excluding a trial that includedpatients with both persistent and paroxysmal AF.

2. Trials that used only pharmacologic rate controlwere analyzed, excluding a study that used AVnode ablation with biventricular pacing as a rate-control strategy.

3. One trial assessed LVEF by means of 2 methods,with different results. Both results were used in asensitivity analysis.

4. The LVEF for inclusion in 3 trials was <40% andin 1 trial was <50%. The trial with LVEFcriterion of <50% was excluded as a sensitivityanalysis.

5. All trials had >80% of patients free of AF afterablation, except 1 trial, which had >50% ofpatients free of AF. This trial was excluded as asensitivity analysis.

All statistical analyses were performed with the useof REVMAN software version 5.3. Two-tailed proba-bility values of <0.05 were considered significant.

RESULTS

STUDY SELECTION. Of 1,144 papers originally retrievedby searching the databases, 4 met the inclusioncriteria (Figure 1).

CHARACTERISTICS OF INCLUDED STUDIES AND

PATIENTS. The 4 randomized, controlled trials(RCTs) were published between 2008 and 2014 andinvolved 224 patients (22–25). Three studies wereconducted in Europe and one in both Europe and theUnited States. All trials were published in English.

The mean age of patients included in the trialsranged from 57 to 63 years. The proportion of men inthe studies was 89%. Ischemic cardiomyopathy wasthe most common etiology for HF in the includedpatients. Three of the trials included patients withonly persistent AF (23–25), whereas 1 trial includedboth paroxysmal and persistent AF (22). All but 39of the included 224 patients had persistent AF. Themean duration of persistent AF was >1 year. Themean LVEF of the included patients was 26.1%, andall patients had New York Heart Association (NYHA)functional classification of II or III. Further patientcharacteristics are listed in Table 1.

All of the included trials were of high quality ($3/5)according to the Jadad quality assessment score.None of the included trials attempted double-blinding. Dropouts and withdrawals were describedappropriately in the included trials. The blankingperiod ranged from 2 to 3 months. The percentage ofpatients requiring repeat procedures ranged from

TABLE 1 Patient Demographics

Khan et al.(22), 2008

Jones et al.(24), 2013

Macdonald et al.(23), 2010

Hunter et al.(25), 2014

Mean age, yrs 60.5 � 8.0 63.0 � 9.5 63.3 � 7.5 57.4 � 11.0

Randomized arm(AF catheter ablation/rate control)

41/40 26/26 22/19 26/24

Mean follow-up, months 6 12 $6 6

Male/female, n 74/7 45/7 32/9 48/2

Coronary artery disease, % 70.5 46.0 51.4 25.9

Persistent AF, % 48.5 100.0 100.0 100.0

Duration of continuous AF, months N/A 23.5 � 25.5 53.3 � 42.7 24 (range, 12–48)

Time since HF diagnosis, months N/A 58 � 60 N/A 26.3 (range, 14–56)

Time since first AF diagnosis, months 47.4 � 31.1 51 � 60 N/A N/A

Body mass index, kg/m2 N/A 29.4 � 4.6 30 � 5.6 N/A

Hypertension, % N/A 33.3 61.0 N/A

Diabetes mellitus, % N/A 23.5 26.8 N/A

Prior TIA/stroke, % N/A 11.8 9.8 N/A

Chronic lung disease, % N/A 13.7 22.0 N/A

NYHA functional class, % 2.48 � 0.5

I 0 0 0

II 52 10 45

III 2 or 3, 100% 48 90 55

IV 0 0 0

Cardiomyopathy, % N/A

Ischemic 32.5 48.6 26

Non-ischemic 67.5 51.4 74

QRS duration at ECG, ms 91 � 9.5 116 � 20 103 � 14 N/A

Baseline medical therapy, % N/A

Amiodarone N/A 12 N/A

ACE I/ARB 100 98 95

Beta-blockers 100 92 88

Aldosterone antagonist 100 in NYHA functional class III 36.5 31.6

Digoxin N/A 54 51.3

LVEF at baseline, % 28.0 � 7.5 23.5 � 7.8 17.7 � 6.9 32.7 � 10.0

Left atrial diameter, cm 4.8 � 0.6 4.8 � 0.7 N/A 5.1 � 1.0

Baseline 6-min walk, m 275 � 49 413.5 � 94.0 333.7 � 121 N/A

Baseline pro–B-type natriuretic peptide, pg/ml N/A 347.5 � 309.0 2,216.5 � 1,951.0 500 � 1,255

MLWHF score at baseline 89.0 � 11.4 45.5 � 22.0 57.4 � 20.1 44.4 � 89.0

Peak VO2 at baseline N/A 17.25 � 5.00 N/A 20.8 � 13.0

AF-free survival after AF catheter ablation, % 88 88 50 81

Prior antiarrhythmic drugs, % N/A N/A

AF catheter ablation group 34.6 53.8

Rate-control group 34.6 41.7

Values are mean � SD or median (interquartile range), unless otherwise stated.

AF ¼ atrial fibrillation; ACE ¼ angiotensin-converting enzyme; ARB ¼ angiotensin receptor blocker; ECG ¼ electrocardiography; HF ¼ heart failure; LVEF ¼ left ventricularejection fraction; MLWHF ¼ Minnesota Living With Heart Failure; N/A ¼ not applicable; NYHA ¼ New York Heart Association; TIA ¼ transient ischemic attack; VO2 ¼ oxygenconsumption.

J A C C : C L I N I C A L E L E C T R O P H Y S I O L O G Y V O L . 1 , N O . 3 , 2 0 1 5 Al Halabi et al.J U N E 2 0 1 5 : 2 0 0 – 9 Meta-Analysis: RCTs of AF Catheter Ablation Versus Rate Control for AF in HF

203

19.5% to 53.7%. Only 1 study had cross-over ofpatients, and intention-to-treat analysis was used(Table 2).

OUTCOMES. LVEF. Data for LVEF were availablefrom all included trials. There was no significant het-erogeneity (I2 ¼ 0%) nor detectable publication bias.

AF catheter ablation compared with rate controlwas associated with an 8.5% increase in LVEF at 6 to12 months (mean difference, 8.53; 95% CI: 6.4 to 10.7;

p < 0.001). The improvement in LVEF in the AFcatheter ablation arm compared with rate control wasevident in all of the included trials (Figure 2A).Quality of life and functional capacity measures.Catheter ablation was superior to rate-control strat-egy in improving quality of life. Data on MLWHF wereavailable from all the included trials. Across theincluded trials, there was no evidence of significantheterogeneity (I2 ¼ 8%) or publication bias. There wasa significant improvement in MLWHF questionnaire

Al Halabi et al. J A C C : C L I N I C A L E L E C T R O P H Y S I O L O G Y V O L . 1 , N O . 3 , 2 0 1 5

Meta-Analysis: RCTs of AF Catheter Ablation Versus Rate Control for AF in HF J U N E 2 0 1 5 : 2 0 0 – 9

204

scores in the AF catheter ablation interventiongroup versus that in the rate-control group (meandifference, –11.9; 95% CI: –17.2 to –6.6; p < 0.001)(Figure 2B).

Data on 6-min walk tests were available fromthree prospective clinical trials. Across the includedtrials, there was evidence of moderate heterogeneity(I2 ¼ 45%). Significant improvement in performanceon 6-min walk tests was observed in patientsundergoing AF catheter ablation compared with arate-control strategy (mean difference, 34.8; 95% CI:2.9 to 66.7; p ¼ 0.03) (Figure 2C).

Data on peak oxygen consumption (VO2) wereavailable from 2 prospective clinical trials. Acrossthe included trials, there was no evidence of signifi-cant heterogeneity (I2 ¼ 0%). Peak VO2 significantlyincreased in AF catheter ablation compared with rate-control patients (mean difference, 3.2; 95% CI: 1.1 to5.3; p ¼ 0.003) (Figure 2D).Sens i t iv i ty ana lyses . One trial included patientswith both paroxysmal and persistent AF, in additionto the use of AV node ablation with biventricularpacing as a rate-control strategy (22). On exclusion ofthat trial, AF catheter ablation was still associatedwith significant improvements in LVEF (mean dif-ference, 7.8; 95% CI: 4.2 to 11.5; p < 0.001) andMLWHF (mean difference, –11.7; 95% CI: –20.3 to –3.2;p ¼ 0.007).

One trial assessed LVEF by means of radionuclideventriculography and cardiovascular magnetic reso-nance (CMR) (23). Both results were used in separateanalyses, both yielding statistically significant in-creases in LVEF in the AF catheter ablation group

TABLE 2 Intervention Characteristic

Khan et al. (22) Jones e

Blanking period, months 2

Frequency of monitoring, months 2, 3*, 6 2, 3,

Modality of assessing heartrhythm

Loop recorder 48-h HoltExisting impl

AAD strategy after ablation AAD for 2 months AAD stopped

No. of patients undergoingrepeat procedures, n (%)

8 (19.5) 5 (2

Cross-over None 2

Ablation strategy of AF PVI � Linear lesionsand sourcesof complexfractionatedelectrograms

PVI � Lineleft atr

fracelectr

Cardiovcavotricu

ab

Follow-up, months 6 1

*Only in the atrial fibrillation (AF) catheter ablation group. †One patient in the AF cathetewas used.

AAD ¼ antiarrhythmic drug; PVI ¼ pulmonary vein isolation.

versus the rate-control group. The same trial in-cluded patients with LVEF <50%, whereas theother three trials included patients with LVEF <40%.It also reported 50% of patients free of AF afterablation, whereas the other trials had >80% ofpatients free of AF after ablation. On exclusion of thattrial (23), AF catheter ablation was still associatedwith significant improvements in LVEF (mean dif-ference, 8.8; 95% CI: 6.3 to 11.3; p < 0.001) andMLWHF (mean difference, –13.8; 95% CI: –19.3 to 8.3;p < 0.001).Procedura l compl i cat ions and adverse events .We were able to obtain previously unpublisheddata on complications and incorporate them into ouranalysis from 2 RCTs (23,24). Two strokes, 4cardiac tamponades, and 1 pericardial effusion wereseen in the AF catheter ablation, culminating in aprocedural complication rate of 6.3%. Details ofcomplications and adverse events are listed inTable 3.

Overall, no statistically significant difference wasfound in the major adverse event rates between theAF catheter ablation (7.2%) versus the rate-control(4.6%) arms (RR: 1.3; 95% CI: 0.4 to 3.9; p ¼ 0.64)(Figure 3A). On examining all reported adverseevents, there was a trend toward more events in theAF catheter ablation (18.9%) compared with the rate-control arm (12.0%), but this did not reach statisticalsignificance (RR: 1.6; 95% CI: 0.8 to 3.0; p ¼ 0.17)(Figure 3B).

No difference in HF re-admissions was seenbetween the 2 treatment groups (RR: 1.4; 95% CI: 0.4to 4.6; p ¼ 0.56) (Figure 3C).

t al. (24) Macdonald et al. (23) Hunter et al. (25)

2 3 3

6, 12 3*, 6 1, 3, 6

er monitoranted devices

24-h Holter monitor 48-h Holter monitor

after ablation Amiodarone for 3 months AAD stopped after ablation

0.1) 6 (30.0) 14 (53.7)

† None None

ar lesions �ial complextionatedograms �ersion andspid isthmuslation

PVI � Linear lesions andsources of complex

fractionatedelectrograms �Cardioversion �cavotricuspid

isthmus ablation

PVI with ablation ofcomplex or fractionatedelectrograms � Linearlesions � Cavotricuspid

isthmus ablation

2 6 6–12

r ablation group and one patient in the rate-control group. Intention-to-treat analysis

FIGURE 2 Changes in Functional Outcomes

A

B

C

D

(A) Change in left ventricular ejection fraction (LVEF). (B) Change in Minnesota Living With Heart Failure (MLWHF). (C) Change in 6-min walk test distance. (D) Change in

peak oxygen consumption (VO2). Mean difference and 95% confidence intervals (CIs) in studies comparing atrial fibrillation catheter ablation with rate control in patients

with heart failure. IV ¼ inverse variance; SD ¼ standard deviation.

J A C C : C L I N I C A L E L E C T R O P H Y S I O L O G Y V O L . 1 , N O . 3 , 2 0 1 5 Al Halabi et al.J U N E 2 0 1 5 : 2 0 0 – 9 Meta-Analysis: RCTs of AF Catheter Ablation Versus Rate Control for AF in HF

205

DISCUSSION

We present the first meta-analysis of high-qualityprospective, RCTs comparing AF catheter ablationversus rate-control strategies in patients with HFand AF.

Previous meta-analyses have relied mainly ondata from observational studies in addition to asmall number of RCTs (26–28). This resulted in het-erogeneity that was highly significant, raising un-certainties about the consistency of the combinedstudies. The main outcomes studied were LVEF,

B-type natriuretic peptide, and complications; allconcluded improvements in LVEF of 11% to 13%.Through collaboration with primary investigatorsof the included RCTs, our current meta-analysisprovides similar conclusions regarding improve-ment in LVEF in addition to new insight into sub-jective and objective quality-of-life measures andcomplication rates, including heart failure read-missions. It also provides conclusions that are basedon a larger number of RCTs, culminating in re-sults with trivial heterogeneity among the majorityof reported outcomes, which suggests strong

TABLE 3 Adverse Events

Complications

Khan et al. (22) Jones et al. (24) Macdonald et al. (23) Hunter et al. (25)

AF CatheterAblation

RateControl

AF CatheterAblation

RateControl

AF CatheterAblation

RateControl

AF CatheterAblation

RateControl

Major

Death – – 1† – – – – 1

Stroke – – – 1 1 – 1 –

Intracranial hemorrhage – – – – – – – 1

Myocardial infarction – – – 1 – – – –

Cardiac tamponade – – 1 – 2 – 1 –

Pericardial effusion 1 – – – – – –

Pneumothorax – 1 – – – – –

Minor

Pulmonary vein stenosis 2* – – – – – – –

Temporary worsening of heart failure 1 – 2 3 3 1 – –

Groin bleeding 3 – 1 – – – – –

Pocket hematoma – 2 – – – – – –

Chest infection – – 1 – – – – –

Ventricular lead dislodgment – 2 – – – – – –

*Mild asymptomatic pulmonary vein stenosis that did not require intervention. †Death occurred 11 months after ablation from progressive worsening of heart failure andend-stage lung disease.

Al Halabi et al. J A C C : C L I N I C A L E L E C T R O P H Y S I O L O G Y V O L . 1 , N O . 3 , 2 0 1 5

Meta-Analysis: RCTs of AF Catheter Ablation Versus Rate Control for AF in HF J U N E 2 0 1 5 : 2 0 0 – 9

206

homogeneity in clinical and methodological charac-teristics of the included RCTs.

We sought evidence from RCTs in an attempt tolimit the influence of selection bias and control forunmeasured confounders (29). We also targeted thespecific patient population of HF with reduced LVEFand documented AF instead of including HF withpreserved LVEF. At least 6 months of follow-up waschosen partially to control for differences in the useof AAD after AF catheter ablation (30).

A significant improvement in LVEF was seen withthe AF catheter ablation strategy in comparison to therate-control approaches. The trend in improvementof LVEF was evident among all of the included RCTs.Of note, 22% to 50% of patients were still in AFafter AF catheter ablation, which suggests a role forreduction of AF burden rather than AF cure, lead-ing to the aforementioned benefits. Anotherpossibility is that patients may have had betterfollow-up and potentially better overall medicalcare in the AF catheter ablation group (Table 2).Several other observational studies have alsoshown improvement in LVEF in ablation comparedwith rate-control arms (14–16). This highlights theimportance of preserving or restoring the atrialcontribution to cardiac hemodynamics, because theAF catheter ablation rhythm control strategy pro-vides additional benefits over simple control ofrapid ventricular rate in patients with HF in AF.

Novel findings from this meta-analysis includedthe marked improvements in the quality of life,6-min walk test performance, and peak VO2 in the AF

catheter ablation group. The improvement in peakVO2 carries particular significance, given that it isa well-established prognostic indicator in HF withpotential impact on survival and hospital stay (31–34).The change in distance walked on the 6-min walktest has also been established as an independentpredictor of survival (35). In one of the trials in whichhalf of patients in the AF catheter ablation remainedin AF, there was no difference in the 6-min walktest outcome between the 2 groups (23). This trialled to the moderate heterogeneity observed of thisoutcome. Nevertheless, the improvements in qualityof life observed despite similar HF re-admissionrates between the 2 groups are consistent with theoverall improvements observed in the 6-minwalk test and peak VO2 (31). The concordant find-ings in these 3 outcomes provide both subjectiveand objective evidence for the benefits of catheterablation for AF on quality of life.

Although the inconsistency index statistic mea-sure of heterogeneity does not capture differences inmethodology between AF ablation approaches (e.g.,pulmonary vein isolation [PVI] and various substrateablation approaches) as well as between rate-controlapproaches (e.g., AV nodal–blocking medicationsversus AV junction ablation), we performed sensi-tivity analyses, including repeating meta-analysesof LVEF and MLWHF without the 1 study that usedAV junction ablation for rate control (22) and foundsimilar results.

Catheter ablation is an invasive procedure, whichmay provide benefits but also carries well-known

FIGURE 3 Adverse Outcomes

B

C

A

(A) Major adverse events. (B) All adverse events. (C) Heart failure re-admissions. Risk ratios and 95% confidence intervals (CI) in studies comparing atrial fibrillation

catheter ablation with rate control in patients with heart failure. M-H ¼ Mantel-Haenszel.

J A C C : C L I N I C A L E L E C T R O P H Y S I O L O G Y V O L . 1 , N O . 3 , 2 0 1 5 Al Halabi et al.J U N E 2 0 1 5 : 2 0 0 – 9 Meta-Analysis: RCTs of AF Catheter Ablation Versus Rate Control for AF in HF

207

risks, such as stroke, pericardial tamponade, pericar-ditis, bleeding, pulmonary stenosis, atrio-esophagealfistula, and even death, risks that may be evenhigher in a population with HF. The AF catheterablation peri-procedural major complication rate inthis study was 6.3%. Despite the structural changesseen in patients with reduced LVEF, the reportedAF catheter ablation complication rate in this meta-analysis is comparable to, though perhaps slightlyhigher than, AF catheter ablation complication ratesreported in a large prospective study (5.2%) and arecent meta-analysis in patients with structurallynormal hearts (2.7% to 3.5%) (36,37). The trend to-ward higher adverse event rates compared with therate-control group (p ¼ 0.17) suggests there maybe a price paid for improving the other outcomesmeasured in this study. Given the overall low adverseevent rates, our study was probably under-powered to

detect statistically significant differences in adverseevents. These risks are put in context to the rate-control adverse event rates in this study of 12.5% inthe pharmacological rate-control studies and 10% inthe study using AV node ablation and biventricularpacing. Cross-over occurred in one trial in only 2patients, making it less likely that cross-overs canexplain the comparability in the incidence of majorcomplications. It is perhaps more plausible that AFablation was associated with relatively low ratesof major adverse event complications, such as deathor stroke (<1%), in these studies (38).

The majority (82.5%) of patients in the currentstudy had persistent AF; therefore, results maynot be easily extrapolated to a population withparoxysmal AF. However, this high proportion ofpatients with persistent AF, as well as the sensitivityanalyses demonstrating similar efficacy of AF catheter

PERSPECTIVES

COMPETENCY IN MEDICAL KNOWLEDGE 1:

Selecting the strategy to address AF in heart failure

should consider several factors, including the type of

AF and patient preferences.

COMPETENCY IN MEDICAL KNOWLEDGE 2:

A strategy of AF catheter ablation may lead to

improvement of AF with improvements in left ventric-

ular function and functional capacity.

TRANSLATIONAL OUTLOOK: The mortality rate

associated with AF catheter ablation versus rate-

control strategy is an important potential area of

research.

Al Halabi et al. J A C C : C L I N I C A L E L E C T R O P H Y S I O L O G Y V O L . 1 , N O . 3 , 2 0 1 5

Meta-Analysis: RCTs of AF Catheter Ablation Versus Rate Control for AF in HF J U N E 2 0 1 5 : 2 0 0 – 9

208

ablation compared with rate control for persistent AF,supports attempting an AF catheter ablation approachin this population despite persistence of AF. In currentclinical practice, patients with HF and persistent AFmay be more readily relegated to a rate-controlapproach. Our findings suggest that significant addi-tional benefit might be achieved by pursuing a rhythmcontrol approach with the use of AF catheter ablation.A well-designed, adequately powered, RCT seemswarranted to provide further data on safety, efficacy,survival, and functional capacity outcomes associatedwith catheter ablation in patients with HF and AF.Subanalyses of patients with AF and HF in the ongoingCABANA trial of AF ablation versus antiarrhythmicdrug use will also be of interest.

STUDY LIMITATIONS. First, patients were only fol-lowed up to 1 year, therefore limiting the assessmentof long-term outcomes. Second, the number of pa-tients included in these RCTs remains small. Thus,the included RCTs were not powered to assess hardoutcomes, such as mortality. Third, there may havebeen bias in selecting healthier patients, who weresuitable candidates for an invasive strategy, for theserandomized trials. A fourth limitation is the fact thatthe power of funnel plots in detecting publicationbias increases with the inclusion of more studies;however, we were limited by the available trials in theliterature. Nevertheless, the mean LVEF of patientswas <35% in all studies, NYHA functional class waspredominantly 2 to 3, and complication and adverseevent rates were at the higher end of that reported inthe literature, which suggests that these patientsmight be representative of a HF population at large.Last, PVI alone was not used in the majority of cases;after PVI, additional atrial substrate modification wasused to varying degrees in each study. However,these approaches reflect contemporary local practicefor catheter ablation of persistent AF.

CONCLUSIONS

An AF catheter ablation strategy in patients with AFand HF results in improved LV function, functionalcapacity, HF symptoms, and quality of life comparedwith a rate-control strategy. Patients with HFmay be athigher risk of complications with interventional ap-proaches for both rhythm and rate-control strategies.However, our analyses suggest that before accepting arate-control strategy in HF patients with persistent ordrug-refractory AF, an individualized approach shouldbe pursued, including consideration to performing AFcatheter ablation in appropriately selected patients.

REPRINT REQUESTS AND CORRESPONDENCE: Dr.Mina K. Chung, Department of Cardiovascular Medi-cine, Heart and Vascular Institute, Cleveland Clinic,9500 Euclid Avenue, J2-2, Cleveland, Ohio 44195.E-mail: chungm@ccf.org.

RE F E RENCE S

1. Chen LY, Shen WK. Epidemiology of atrialfibrillation: a current perspective. Heart Rhythm2007;4:S1–6.

2. Kim MH, Johnston SS, Chu BC, et al.Estimation of total incremental health care costs inpatients with atrial fibrillation in the United States.Circ Cardiovasc Qual Outcomes 2011;4:313–20.

3. Levy D, Kenchaiah S, Larson MG, et al. Long-term trends in the incidence of and survival withheart failure. N Engl J Med 2002;347:1397–402.

4. Berry C, Murdoch DR, McMurray JJ. Economics ofchronic heart failure. Eur J Heart Fail 2001;3:283–91.

5. Maisel WH, Stevenson LW. Atrial fibrillationin heart failure: epidemiology, pathophysiology,

and rationale for therapy. Am J Cardiol 2003;91:2D–8.

6. Stevenson WG, Stevenson LW, Middlekauff HR,et al. Improving survival for patients with atrialfibrillation and advanced heart failure. J Am CollCardiol 1996;28:1458–63.

7. Dries DL, Exner DV, Gersh BJ, et al. Atrialfibrillation is associated with an increased risk formortality and heart failure progression in patientswith asymptomatic and symptomatic left ventric-ular systolic dysfunction: a retrospective analysisof the SOLVD trials, Studies Of Left VentricularDysfunction. J Am Coll Cardiol 1998;32:695–703.

8. Wang TJ, Larson MG, Levy D, et al. Temporalrelations of atrial fibrillation and congestive heart

failure and their joint influence on mortality: the Fra-mingham Heart Study. Circulation 2003;107:2920–5.

9. Daoud EG, Weiss R, Bahu M, et al. Effect of anirregular ventricular rhythm on cardiac output. AmJ Cardiol 1996;78:1433–6.

10. Clark DM, Plumb VJ, Epstein AE, Kay GN. He-modynamic effects of an irregular sequence ofventricular cycle lengths during atrial fibrillation.J Am Coll Cardiol 1997;30:1039–45.

11. Roy D, Talajic M, Nattel S, et al. Rhythm controlversus rate control for atrial fibrillation and heartfailure. N Engl J Med 2008;358:2667–77.

12. Corley SD, Epstein AE, DiMarco JP, et al.Relationships between sinus rhythm, treatment,

J A C C : C L I N I C A L E L E C T R O P H Y S I O L O G Y V O L . 1 , N O . 3 , 2 0 1 5 Al Halabi et al.J U N E 2 0 1 5 : 2 0 0 – 9 Meta-Analysis: RCTs of AF Catheter Ablation Versus Rate Control for AF in HF

209

and survival in the Atrial Fibrillation Follow-upInvestigation of Rhythm Management (AFFIRM)study. Circulation 2004;109:1509–13.

13. Van Gelder IC, Hagens VE, Bosker HA, et al.A comparison of rate control and rhythm control inpatients with recurrent persistent atrial fibrilla-tion. N Engl J Med 2002;347:1834–40.

14. Chen MS, Marrouche NF, Khaykin Y, et al.Pulmonary vein isolation for the treatment ofatrial fibrillation in patients with impaired systolicfunction. J Am Coll Cardiol 2004;43:1004–9.

15. Hsu LF, Jais P, Sanders P, et al. Catheterablation for atrial fibrillation in congestive heartfailure. N Engl J Med 2004;351:2373–83.

16. Sacher F, Corcuff JB, Schraub P, et al. Chronicatrial fibrillation ablation impact on endocrine andmechanical cardiac functions. Eur Heart J 2008;29:1290–5.

17. Liberati A, Altman DG, Tetzlaff J, et al.The PRISMA statement for reporting systematicreviews and meta-analyses of studies that eval-uate health care interventions: explanation andelaboration. J Clin Epidemiol 2009;62:e1–34.

18. Jadad AR, Moore RA, Carroll D, et al. Assessingthe quality of reports of randomized clinical trials:Is blinding necessary? Control Clin Trials 1996;17:1–12.

19. DerSimonian R, Laird N. Meta-analysisin clinical trials. Control Clin Trials 1986;7:177–88.

20. Deeks JJ, Higgins JPT, Altman DG. Chapter 9:Analysing data and undertaking meta-analyses. In:Higgins JPT, Green S, editors. Cochrane Handbookfor Systematic Reviews of Interventions Version5.1.0 [updated March 2011]. The Cochrane Col-laboration, 2011. Available at: http://www.cochrane-handbook.org. Accessed February 2,2015.

21. Begg CB, Mazumdar M. Operating character-istics of a rank correlation test for publication bias.Biometrics 1994;50:1088–101.

22. Khan MN, Jais P, Cummings J, et al. Pulmo-nary-vein isolation for atrial fibrillation in patientswith heart failure. N Engl J Med 2008;359:1778–85.

23. MacDonald MR, Connelly DT, Hawkins NM,et al. Radiofrequency ablation for persistent atrialfibrillation in patients with advanced heart failureand severe left ventricular systolic dysfunction: arandomised controlled trial. Heart 2011;97:740–7.

24. Jones DG, Haldar SK, Hussain W, et al.A randomized trial to assess catheter ablationversus rate control in the management of persis-tent atrial fibrillation in heart failure. J Am CollCardiol 2013;61:1894–903.

25. Hunter RJ, Berriman TJ, Diab I, et al.A randomized controlled trial of catheter ablationversus medical treatment of atrial fibrillation inheart failure (the CAMTAF trial). Circ ArrhythmElectrophysiol 2014;7:31–8.

26. Wilton SB, Fundytus A, Ghali WA, et al. Meta-analysis of the effectiveness and safety of catheterablation of atrial fibrillation in patients with versuswithout left ventricular systolic dysfunction. Am JCardiol 2010;106:1284–91.

27. Dagres N, Varounis C, Gaspar T, et al. Catheterablation for atrial fibrillation in patients with leftventricular systolic dysfunction: a systematic re-view and meta-analysis. J Card Fail 2011;17:964–70.

28. Anselmino M, Matta M, D’Ascenzo F, et al.Catheter ablation of atrial fibrillation in patientswith left ventricular systolic dysfunction: a sys-tematic review and meta-analysis. Circ ArrhythmElectrophysiol 2014;7:1011–8.

29. Odgaard-Jensen J, Vist GE, Timmer A, et al.Randomisation to protect against selection biasin healthcare trials. Cochrane Database Syst Rev2011;(4):MR000012.

30. Leong-Sit P, Roux JF, Zado E, et al. Antiar-rhythmics after ablation of atrial fibrillation(5A study): six-month follow-up study. CircArrhythm Electrophysiol 2011;4:11–4.

31. Flynn KE, Lin L, Moe GW, et al. Relationshipsbetween changes in patient-reported health sta-tus and functional capacity in outpatients withheart failure. Am Heart J 2012;163:88–94.e3.

32. De Groote P, Dagorn J, Soudan B, et al. B-typenatriuretic peptide and peak exercise oxygenconsumption provide independent information forrisk stratification in patients with stable conges-tive heart failure. J Am Coll Cardiol 2004;43:1584–9.

33. Corra U, Mezzani A, Bosimini E, Giannuzzi P.Cardiopulmonary exercise testing and prognosisin chronic heart failure: a prognosticating algo-rithm for the individual patient. Chest 2004;126:942–50.

34. Swank AM, Horton J, Fleg JL, et al. Modestincrease in peak VO2 is related to better clinicaloutcomes in chronic heart failure patients: resultsfrom heart failure and a controlled trial to inves-tigate outcomes of exercise training. Circ HeartFail 2012;5:579–85.

35. Passantino A, Lagioia R, Mastropasqua F,Scrutinio D. Short-term change in distance walkedin 6 min is an indicator of outcome in patients withchronic heart failure in clinical practice. J Am CollCardiol 2006;48:99–105.

36. Bohnen M, Stevenson WG, Tedrow UB, et al.Incidence and predictors of major complicationsfrom contemporary catheter ablation to treatcardiac arrhythmias. Heart Rhythm 2011;8:1661–6.

37. Gupta A, Perera T, Ganesan A, et al. Compli-cations of catheter ablation of atrial fibrillation: asystematic review. Circ Arrhythm Electrophysiol2013;6:1082–8.

38. Scheinman MM, Huang S. The 1998 NASPEprospective catheter ablation registry. Pacing ClinElectrophysiol 2000;23:1020–8.

KEY WORDS atrial fibrillation, catheterablation, heart failure, pulmonary veinisolation