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 . 5 , N O . 1 , 2 0 1 9

C R OWN CO P Y R I G H T ª 2 0 1 9 P U B L I S H E D B Y E L S E V I E R O N B E H A L F O F 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 . A L L R I G H T S R E S E R V E D .

VT MAPPING AND ABLATION

Combined Endocardial-Epicardial VersusEndocardial Catheter Ablation Alonefor Ventricular Tachycardia inStructural Heart DiseaseA Systematic Review and Meta-Analysis

Jorge Romero, MD,a Roberto C. Cerrud-Rodriguez, MD,a Luigi Di Biase, MD, PHD,a Juan Carlos Diaz, MD,a

Isabella Alviz, MD,a Vito Grupposo, RT,a Luis Cerna, MD,a Ricardo Avendano, MD,a Usha Tedrow, MD, MS,b

Andrea Natale, MD,c Roderick Tung, MD,d Saurabh Kumar, BSC(MED)/MBBS, PHDe

ABSTRACT

ISS

Fro

Me

Bri

Ar

Sch

Re

Ste

EP

OBJECTIVES This study sought to determine whether combined endocardial-epicardial (endo-epi) ablation was

superior to endocardial only ablation in patients with scar-related ventricular tachycardia (VT).

BACKGROUND Limited single-center studies suggest that combined endo-epi ablation strategy may be superior to

endocardial ablation (endo) alone in patients with nonischemic cardiomyopathy (NICM) and arrhythmogenic right

ventricular cardiomyopathy (ARVC), and ischemic cardiomyopathy (ICM).

METHODS A systematic review of Medline, Cochrane, and Embase databases was performed for studies that reported

outcomes comparing endo-epi with endo VT ablation alone.

RESULTS Seventeen studies consisting of 975 patients were included (mean 56 � 10 years of age; 79% male; NICM in

36.6%; ICM in 32.8%; and ARVC in 30.6%). After a mean follow-up of 27 � 21 months, endo-epi ablation was associated

with a 35% reduction in risk of VT recurrence compared with endocardial ablation alone (risk ratio [RR]: 0.65; 95%

confidence interval [CI]: 0.55 to 0.78; p < 0.001). Sensitivity analysis showed lower risk of VT recurrence in ICM (RR:

0.43; 95% CI: 0.28 to 0.67; p ¼ 0.0002) and ARVC (RR: 0.59; 95% CI: 0.43 to 0.82; p ¼ 0.0002), with a nonsignificant

trend in NICM (RR: 0.87; 95% CI: 0.70 to 1.08; p ¼ 0.20). Endo-epi, compared with endo ablation, was associated with

reduced all-cause mortality (RR: 0.56; 95% CI: 0.32 to 0.97; p ¼ 0.04). Acute procedural complications were higher with

the endo-epi approach (RR: 2.62; 95% CI: 0.91 to 7.52; p ¼ 0.07).

CONCLUSIONS This meta-analysis suggests that a combined endo-epi ablation is associated with a lower risk of

VT recurrence and subsequent mortality than endo only VT ablation in patients with scar-related VT. Procedural

complications, however, are higher with the endo-epi approach. (J Am Coll Cardiol EP 2019;5:13–24)

Crown Copyright © 2019 Published by Elsevier on behalf of the American College of Cardiology Foundation.

All rights reserved.

N 2405-500X/$36.00 https://doi.org/10.1016/j.jacep.2018.08.010

m the aArrhythmia Services, Department of Medicine, Montefiore-Einstein Center for Heart and Vascular Care, Montefiore

dical Center, Albert Einstein College of Medicine, Bronx, New York; bCardiac Arrhythmia Service, Heart and Vascular Center,

gham and Women’s Hospital Heart and Vascular Center, Harvard Medical School, Boston, Massachusetts; cTexas Cardiac

rhythmia Institute, St. David’s Medical Center, Austin, Texas; dCenter for Arrhythmia Care, Heart and Vascular Center, Pritzker

ool of Medicine, The University of Chicago Medicine, Chicago, Illinois; and the eDepartment of Cardiology, Westmead Applied

search Centre, University of Sydney, Sydney, New South Wales, Australia. Dr. Di Biase is a consultant for Biosense Webster,

reoataxis, Boston Scientific, and St. Jude Medical; and has received speaker and travel honoraria from Medtronic, Atricure,

iEP, and Biotronik. Dr. Tedrow is a consultant for Boston Scientific and Abbott; and has received research funding from

ABBR EV I A T I ON S

AND ACRONYMS

AADs = antiarrhythmic drugs

ARVC = arrhythmogenic right

ventricular cardiomyopathy

CA = catheter ablation

endo-epi = endocardial-

epicardial

SHD = structural heart disease

VT = ventricular tachycardia

VA = ventricular arrhythmia

Biosense W

Medtronic,

Tung has r

contents of

All authors

institutions

visit the JA

Manuscript

Romero 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 . 5 , N O . 1 , 2 0 1 9

Endo-Epicardial vs. Endocardial Only VT Ablation J A N U A R Y 2 0 1 9 : 1 3 – 2 4

14

V entricular tachycardia (VT) is animportant cause of morbidity andmortality, regardless of the under-

lying arrhythmogenic substrate (1). Recur-rent defibrillator shocks for VT areassociated with poor quality of life andincreased mortality (2,3). Although antiar-rhythmic drugs (AADs) can reduce recur-rences of ventricular arrhythmia (VA),adjunctive percutaneous catheter ablation(CA) may be more efficacious in reducingVA burden (4). In some situations, the

modest efficacy of CA for VT is improved with epicar-dial mapping and ablation, especially in disease sub-strates such as arrhythmogenic right ventricularcardiomyopathy (ARVC), basolateral phenotype ofnonischemic idiopathic dilated cardiomyopathy(NICM), Brugada phenotypes, and in some situationsin ischemic cardiomyopathy (ICM) (5,6). Some studieshave shown outcomes from combined endocardial-epicardial (endo-epi) ablation are superior to thosefrom endocardial alone (endo) ablation for VT (6,7).However, epicardial access carries a risk of proceduralcomplications that include inadvertent right ventric-ular free-wall puncture with subsequent bleedingand injury to abdominal viscera or blood vessels,liver, and diaphragm (8–10). These risks, coupledwith an unestablished benefit compared to an endo-cardial only approach, have limited the widespreaduse of a combined endo-epi ablation as standardtreatment for VT. This meta-analysis sought to deter-mine whether endo-epi ablation was superior to endoVT ablation alone in patients with structural heartdisease (SHD) with regard to acute procedural failure,complications, VT recurrence, and mortality.

SEE PAGE 25

METHODS

The present meta-analysis was performed accordingto Cochrane Collaboration and Preferred ReportingItems for Systematic Reviews and Meta-Analyses(PRISMA) criteria (11). This meta-analysis was regis-tered in PROSPERO database (Epicardial-endocardial

ebster and Abbott. Dr. Grupposo is an employee of CAS Biosense

and Biosense Webster. Dr. Natale is a consultant for Biosense We

eceived research grants from Abbott. All other authors have repo

this paper to disclose. Francis Marchlinski, MD, served as Guest

attest they are in compliance with human studies committe

and Food and Drug Administration guidelines, including patien

CC: Clinical Electrophysiology author instructions page.

received May 14, 2018; revised manuscript received August 13, 2

[epi-endocardial] ablation versus endocardial abla-tion for the management of ventricular tachycardia instructural heart disease: a systematic review andmeta-analysis [CRD42018084653]; National Institutefor Health Research, University of York, York, UnitedKingdom).

SEARCH STRATEGY. We searched PubMed, Embase,and Cochrane Central Register of Clinical Trials(Cochrane Library; issue 02, 2017) databases fromJanuary 1996 through January 2018 to identify studiesthat reported clinical outcomes of endo-epi versusendo ablation alone in the management of VT. Weused the search terms (“ventricular tachycardia” OR“ventricular arrhythmias” OR “VT”) AND (“endocar-dial ablation” OR “epicardial access” OR “epicardialablation” OR “epi-endocardial ablation” OR “radio-frequency ablation” OR “catheter ablation” OR“radiofrequency ablation”). No language restrictionwas applied. The reference lists of identified articleswere also reviewed to identify additional citations.

ELIGIBILITY CRITERIA. Studies with the followingpatient characteristics were considered eligible: 1)those that included patients with SHD-related VTundergoing endo ablation and/or endo-epi ablation;and 2) those that reported clinical outcomes of VTrecurrence and/or total mortality and/or acute pro-cedural complications and/or acute procedural failureand compared endo-epi with endo only VT ablation.Abstracts, case reports, conference presentations,editorials, reviews, and expert opinions wereexcluded from our analysis.

PRIMARY AND SECONDARY OUTCOMES. The pri-mary outcomes of this study were individual end-points of VT recurrence and all-cause mortalityduring follow-up after ablation. Secondary outcomeswere acute procedural failure and acute proceduralcomplications.

VT recurrence was defined as a documented recur-rence of any VT during the follow-up period of eachstudy. All-causemortality was defined as the death of apatient due to any cause within the follow-up periodafter the VT procedure. Acute procedural complica-tions included the following: procedure-related

Webster; and has financial relationships with Abbott,

bster, St. Jude/Abbott, Medtronic, and Biotronik. Dr

rted that they have no relationships relevant to the

Editor for this paper.

es and animal welfare regulations of the authors’

t consent where appropriate. For more information,

018, accepted August 15, 2018.

FIGURE 1 Study Selection

Four of the 17 selected papers were split into 2 each (1,16–18), because each included 2

distinct patient populations (ICM, NICM and ARVD/C). ARVC ¼ arrhythmogenic right ven-

tricular cardiomyopathy; ICM¼ ischemic cardiomyopathy;NICM¼ ischemic cardiomyopathy.

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 . 5 , N O . 1 , 2 0 1 9 Romero et al.J A N U A R Y 2 0 1 9 : 1 3 – 2 4 Endo-Epicardial vs. Endocardial Only VT Ablation

15

complications such as vascular injuries, pericardialeffusion, cardiac tamponade, electromechanicaldissociation, atrioventricular block, phrenic nerveinjury, acute stroke, and death. Finally, acute proce-dural failure was defined as inducibility of any VT atthe end of the procedure.

DATA EXTRACTIONS AND QUALITY APPRAISAL.

Three investigators (J.R., R.C.R., and J.C.D.) inde-pendently screened all titles and abstracts andmanually searched the full text versions of all rele-vant studies that fulfilled inclusion criteria. Refer-ence lists in the retrieved articles were independentlyreviewed for further identification of potentiallyrelevant studies. Disagreements were resolved byconsensus after discussion (J.R. and S.K.). Weextracted characteristics of each study, includingmethodology and baseline patient characteristics, VTrecurrence, all-cause mortality according to ablationapproach (endo-epi vs. endo alone), acute proceduralcomplications, and acute procedural failure rate ofeach approach. If the above-mentioned informationwas not readily available in the written article, theprincipal investigator of that particular study wascontacted to supply pertinent information.

QUALITY ASSESSMENT. The quality and reporting ofthe studies were assessed using the Newcastle-Ottawa Scale (12). Studies were subsequentlydivided into 1 of 3 categories: 1) high-quality studiesreceived 6 to 9 points; 2) satisfactory quality studiesreceived 3 to 5 points; and 3) unsatisfactory qualitystudies received 0 to 2 points.

STATISTICAL ANALYSIS. Descriptive statistics arepresented for the number of cases for dichotomousand categorical variables. Statistical analysis wasperformed in line with recommendations from theCochrane Collaboration and PRISMA guidelines, us-ing Review Manager version 5.3 software (RevMan,Cochrane Collaboration, 2014). Heterogeneity wasassessed using I2 statistics (13). Data were summa-rized across groups using the Mantel-Haenszel riskratio (RR) fixed-effect model if I2 <25. We consideredI2 #25% as low and I2 $75% as high. The random-effects model was used if I2 >25% (14). Publicationbias was estimated visually by funnel plots, as pre-viously described (14,15).

RESULTS

A total of 4,933 studies were identified by usingspecified search criteria (Figure 1). After we evaluatedthese studies in detail, 17 relevant studies wereincluded. Review of these studies allowed 21 cohorts

of patients to be separated, because 4 of the 17 studieswere split into 2 studies for the purpose of analysis, asthey included distinct data sets for ICM, NICM, andARVC populations. Studies by Tung et al. (1), Naka-hara et al. (16), and Vergara et al. (17) were analyzedas 2 separate cohorts as they included distinct ICMand NICM populations. The study by Berte et al. (18)was also analyzed as 2 separate cohorts as thatstudy also included NICM and ARVC populations.

CHARACTERISTICS OF INCLUDED STUDIES. Ourmeta-analysis included studies assessing outcomesin patients with ICM (1,7,16,17,19,20), NICM(1,16–18,21–25), and ARVC (5,6,18,26–28). Nine pro-spective studies included patients assigned sequen-tially to treatment cohorts of endo-epi or endo onlyablation. Eleven retrospective studies obtained theirinformation from registries or hospital databases andtheir analyses were performed by chart review.Finally, the study by Izquierdo et al. (7) included aretrospective study of patients in which endo ablation

TABLE 1 Baseline Characteristics of Included Studies

First Author(Ref. #), Year Type of Study Mean Age n Males

AblationApproach

ArrhythmogenicSubstrate

Epi-Endo Endo ARVC NICM ICM

Acosta et al. (19), 2016 Prospective, single-center, not randomized 67.4 90 83 24 66 NA NA 90

Arya et al. (21), 2010 Retrospective, single-center 56.8 13 9 4 9 NA 13 NA

Bai et al. (6), 2011 Prospective, multicenter, not randomized 35.5 49 33 26 23 49 NA NA

Berte et al. (18), 2016 Prospective, multicenter, not randomized 59 13 12 8 5 NA 13 NA

Berte et al. (18), 2016 Prospective, multicenter, not randomized 48 7 4 4 3 7 NA NA

Di Biase et al. (20), 2012 Prospective, multicenter, not randomized 62 92 74 43 49 NA NA 92

Izquierdo et al. (7), 2015 Retrospective, prospective, nonrandomized, single-center 66.5 50 47 15 35 NA NA 50

Killu et al. (22), 2016 Retrospective, single-center 53.7 105 83 80 25 NA 105 NA

Mussigbrodt et al. (26), 2017 Retrospective, single-center 53.2 45 30 22 23 45 NA NA

Nakahara et al. (16), 2010 (a) Retrospective, single-center 69 17 16 3 14 NA NA 17

Nakahara et al. (16), 2010 (b) Retrospective, single-center 59 16 13 6 10 NA 16 NA

Philips et al. (5), 2012 Retrospective, multicenter 38 87 45 23 64 87 NA NA

Piers et al. (24), 2013 Retrospective, single-center 60 45 34 30 15 NA 45 NA

Santangeli et al. (23), 2010 Prospective, multicenter, not randomized 50.4 22 17 13 9 NA 22 NA

Santangeli et al. (28), 2015 Retrospective, single-center 39 62 45 39 23 62 NA NA

Soejima et al. (25), 2004 Prospective, single-center, not randomized 54 28 24 6 22 NA 28 NA

Tung et al. (1), 2013 (a) Retrospective, single-center 68.5 72 66 21 51 NA NA 72

Tung et al. (1), 2013 (b) Retrospective, single-center 57.5 64 51 29 35 NA 64 NA

Vergara et al. (17), 2012 (a) Prospective, single-center, not randomized 63.6 36 NP 9 27 NA NA 36

Vergara et al. (17), 2012 (b) Prospective, single-center, not randomized 63.6 14 NP 11 3 NA 14 NA

Wei et al. (27), 2017 Retrospective, single-center 39.9 48 33 17 31 48 NA NA

TABLE 1 Continued

First Author(Ref. #), Year

Mean Follow-UpTime (months)

MeanLVEF (%)

n PatientsWith ICD

MeanProcedureTime forEndo (h)

MeanProcedureTime for

Epi-Endo (h)

MeanFluoroscopyTime for

Endo (min)

MeanFluoroscopyTime for

Epi-Endo (min)

AverageCumulativeRF Time forEndo (min)

AverageCumulativeRF Time for

Epi-Endo (min)

Acosta et al. (19), 2016 22.5 33 80 NP NP NP NP 25.5 25.9

Arya et al. (21), 2010 23 33.1 13 NP NP NP NP NP NP

Bai et al. (6), 2011 36.5 55 49 3.9 5.3 51.3 65.9 20.2 25.8

Berte et al. (18), 2016 26 41 13 NP NP NP NP NP NP

Berte et al. (18), 2016 30 49 7 NP NP NP NP NP NP

Di Biase et al. (20), 2012 22 25.5 92 3.6 4.8 32 38 39 74

Izquierdo et al. (7), 2015 15 31 50 5.5 6 NP NP 19 16

Killu et al. (22), 2016 12.2 40 NP 6.2 6.9 76.8 81.1 NP NP

Mussigbrodt et al. (26), 2017 31.3 56 NP NP NP NP NP NP NP

Nakahara et al. (16), 2010 (a) 15 21 17 NP NP NP NP NP NP

Nakahara et al. (16), 2010 (b) 15 27 16 NP NP NP NP NP NP

Philips et al. (5), 2012 88.3 NP 82 NP NP NP NP NP NP

Piers et al. (24), 2013 25 44 30 NP NP NP NP NP NP

Santangeli et al. (23), 2010 20 34.3 18 NP NP NP NP NP NP

Santangeli et al. (28), 2015 56 NP 58 NP NP NP NP 80.6 148.7

Soejima et al. (25), 2004 11.1 30 20 NP NP NP NP NP NP

Tung et al. (1), 2013 (a) 12 30 69 5.3 6.8 40 91 NP NP

Tung et al. (1), 2013 (b) 12 32 58 1.5 6 78 85 NP NP

Vergara et al. (17), 2012 (a) 13.4 32.2 NP NP NP NP NP NP NP

Vergara et al. (17), 2012 (b) 13.4 35.6 NP NP NP NP NP NP NP

Wei et al. (27), 2017 71.4 51.4 11 NP NP NP NP NP NP

Berte et al. 2016 (b): this section of the study included NICM patients only. Berte et al. 2016 (c): this section of the study included ARVC patients only. For clarification purposes, there is NO Berte et al. 2016(a). Nakahara et al. 2010 (a): this section of the study included ICM patients only. Nakahara et al. 2010 (b): this section of the study included NICM patients only. Tung et al. 2013 (a): this section of the studyincluded ICM patients only. Tung et al. 2013 (b): this section of the study included NICM patients only. Vergara et al. 2012 (a): this section of the study included ICM patients only. Vergara et al. 2012 (b): thissection of the study included NICM patients only.

ARVC ¼ arrhythmogenic right ventricular dysplasia/cardiomyopathy; Endo ¼ endocardial-only ablation; n ¼ number of patients included in the study; Endo-epi ¼ epicardial-endocardial ablation;ICD ¼ implantable cardioverter-defibrillator; ICM ¼ ischemic cardiomyopathy; LVEF ¼ left ventricular ejection fraction; NA ¼ not applicable; NICM ¼ nonischemic cardiomyopathy; NP ¼ not presented.

Romero 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 . 5 , N O . 1 , 2 0 1 9

Endo-Epicardial vs. Endocardial Only VT Ablation J A N U A R Y 2 0 1 9 : 1 3 – 2 4

16

FIGURE 2 Study Outcomes Comparing Endocardial-Epicardial Versus Endocardial Only Groups

(A) Freedom from VT recurrence; (B) overall survival; (C) acute procedural success; and (D) acute procedural complications. ARR ¼ absolute risk

reduction; CI ¼ confidence interval; RR ¼ risk ratio; VT ¼ ventricular tachycardia.

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 . 5 , N O . 1 , 2 0 1 9 Romero et al.J A N U A R Y 2 0 1 9 : 1 3 – 2 4 Endo-Epicardial vs. Endocardial Only VT Ablation

17

had been used. Analysis then compared historicaloutcomes to those from a recent cohort of patientswho were prospectively assigned to the endo-epiablation group.

A total of 975 patients (79.5% male; mean 56 � 10years of age) comprised the final populationanalyzed. Study characteristics, mean follow-up pe-riods, and patient characteristics are described inTable 1. Results are summarized graphically inFigure 2.

Proportions of patients with distinct arrhythmo-genic substrates were as follows: 36.6% NICM; 32.8%ICM; and 30.6% ARVC. Of 975 patients, 44.4% under-went epi-endo CA, whereas 55.6% received endo CAalone. Most studies compared combined endo-epiwith endo ablation alone in terms of VT recurrence,

all-cause mortality, acute procedural failure, and/oracute procedural complications. Studies that did notinclude 1 or more of these outcomes in their analysiswere not included in the analysis, as outcome datawere missing. In all studies, AAD medical therapyfailed.

QUALITY ASSESSMENT AND PUBLICATION BIAS.

Funnel plots did not suggest publication bias forcomparison of VT recurrence and all-cause mortalityin all groups between epi-endo CA versus endo CA(Online Figure S1). All studies included in this meta-analysis had good methodological quality, indicatinga low risk of bias (Online Table S1). All 17 studies wereclassified as being high quality, based on theNewcastle-Ottawa Scale, using 9 different parameters.

FIGURE 3 VT Recurrence by Subgroups

Forest plot reports VT recurrence by subgroup of ICM (top panel), NICM (middle panel), and ARVC (lower panel). Diamonds indicates overall summary estimates for the

analysis: width of the diamond represents 95% CI; width of the shaded square represents the size of the population (fixed effects model was used in the analysis).

ARVC ¼ arrhythmogenic right ventricular cardiomyopathy; CI ¼ confidence interval; M-H ¼ Mantel–Haenszel.

Romero 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 . 5 , N O . 1 , 2 0 1 9

Endo-Epicardial vs. Endocardial Only VT Ablation J A N U A R Y 2 0 1 9 : 1 3 – 2 4

18

IMPACT ON VT RECURRENCE. Overall, there was astatistically significant reduction in any VT recur-rence in the endo-epi ablation group compared withthat in the endo only ablation group, with a relativerisk reduction (RRR) of 35% (RR: 0.65; 95% confidenceinterval [CI]: 0.55 to 0.78; p < 0.00001). No significantheterogeneity was observed between trials (Figures 2and 3).

In subgroup analysis, ICM and ARVC groupsshowed a statistically significant lower risk of VTrecurrence with endo-epi than with endo only abla-tion (RRR: 0.43; 95% CI: 0.28 to 0.67; p ¼ 0.0002; andRRR: 0.59; 95% CI: 0.43 to 0.82; p ¼ 0.0002,

respectively) (Figure 3), whereas the NICM subgroupshowed a trend toward benefit in VT recurrence (RR:0.87; 95% CI: 0.70 to 1.07; p ¼ 0.20) (Figure 3). TheICM group had an absolute risk reduction (ARR) of20%, whereas the ARVC group had an ARR of 23%.

IMPACT ON ALL CAUSE MORTALITY. When we com-bined all groups, endo-epi ablation was associatedwith a statistically significant lower all-cause mor-tality than endo only ablation (RRR: 0.56; 95% CI:0.32 to 0.97; p ¼ 0.04). A subgroup difference in all-cause mortality was not found, likely due to smallnumbers of studies (Figures 2 and 4).

FIGURE 4 All-Cause Mortality by Subgroups

Forest plot reports all-cause mortality by subgroups of ICM (top panel), NICM (middle panel), and ARVC (lower panel). Diamonds indicate overall summary estimate for

the analysis: width of the diamond represents the 95% CI; width of the shaded square represents the size of the population (fixed effects model used in analysis).

Abbreviations as in Figure 1.

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 . 5 , N O . 1 , 2 0 1 9 Romero et al.J A N U A R Y 2 0 1 9 : 1 3 – 2 4 Endo-Epicardial vs. Endocardial Only VT Ablation

19

IMPACT ON ACUTE PROCEDURAL SUCCESS FOR ALL

GROUPS. There were no significant differences be-tween the rates of acute procedural success in theendo-epi procedure and those in the endo ablationgroup (RRR: 1.06; 95% CI: 0.97 to 1.15; p ¼ 0.21)(Figures 2 and 5).

IMPACT ON ACUTE PROCEDURAL COMPLICATIONS

FOR ALL GROUPS. There was a trend toward greatrisk of complications in the endo-epi ablation groupcompared to those in the endo only ablation group(RRR: 2.62; 95% CI: 0.91 to 7.52; p ¼ 0.07) (Figures 2and 6). Of note, the complications most commonlyassociated with epicardial access were pericardialeffusion, pericardial tamponade, and right ventriclepuncture (Online Table S2).

TIMING OF ENDO-EPI APPROACH. Patients includedin the endo-epi group consisted of those in whomprimary endo-epi ablation was performed and thosein whom endo-epi ablation was attempted after

failure of endo ablation. However, sensitivity analysisfound that the beneficial effect of endo-epi was pre-served regardless of timing (Online Table S3, OnlineFigure S2).

DISCUSSION

Recurrent VT is a major cause of morbidity and mor-tality in patients with SHD, and CA has beendemonstrated to be superior to AAD in VT recurrenceand VT control (29). Although contemporary ablationstrategies (e.g., substrate modification) have demon-strated clinical outcomes that are improved comparedto those in conventional VT mapping and ablation, VTrecurrence is still a major issue in patients undergo-ing endocardial ablation (30). In some clinical sce-narios, additional epicardial ablation yields improvedoutcomes due to concealed substrate for VT in theepicardium, particularly in conditions such asthe inferolateral subtype of idiopathic dilated

FIGURE 5 Acute Procedural Success by Subgroup

Forest plot reports acute procedural failure by subgroups of ICM (top panel), NICM (middle panel), and ARVC (lower panel). Diamonds indicate overall summary

estimate for the analysis: width of the diamond represents the 95% CI;width of the shaded square represents the size of the population (fixed effects model was used in

the analysis). Abbreviations as in Figures 1 and 4.

Romero 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 . 5 , N O . 1 , 2 0 1 9

Endo-Epicardial vs. Endocardial Only VT Ablation J A N U A R Y 2 0 1 9 : 1 3 – 2 4

20

cardiomyopathy, ARVC, and transmural post-infarctsubstrates. Whether the combined endo-epi ablationuniversally yields outcomes that are improvedcompared with those of routine endo ablation alone iscontroversial. In this meta-analysis, we compared theoutcomes of endo-epi with those of endo ablation onVT recurrence, all-cause mortality, acute proceduralfailure, and acute procedural complications. Weincluded a total of 975 patients undergoing ablationof VT with a mean follow-up of 27 � 21 months. Majorfindings of this study are the following. 1) Combinedendo-epi ablation is associated with a 35% relativerisk reduction in VT recurrence compared to endoonly ablation; a significant relative risk reduction inVT recurrence was noted in ICM (57%) and ARVC(41%) patients, with a trend toward reduction in VTrecurrence in the NICM subgroup. 2) Combined

endo-epi ablation was associated with a lower mor-tality rate than endo ablation alone, with an RRR of44% and an ARR of 5.9%. 3) Acute procedural failurewith endo-epi was not significantly different fromthat with endo only ablation; however, proceduralcomplications were higher in the endo-epi procedurethan in the endo only ablation strategy.

In current clinical practice, the epicardial mappingand ablation procedure is usually reserved for patientswith VT in whom prior endo procedures have failedand patients with morphological criteria for epicardialVT on 12-lead electrocardiography or specificmyocardial substrates such as inferolateral subtype ofNICM, ARVC, Brugada syndrome, and some ICM sub-strates. Results of this meta-analysis have reinforcedthe concept that a combined endo-epi approach,compared to an endo-only ablation, is associated with

FIGURE 6 Acute Procedural Complications by Subgroup

Forest plot reports acute procedural complications by subgroups of ICM (top panel), NICM (middle panel), and ARVC (lower panel). Diamonds indicate overall summary

estimate for the analysis: width of the diamond represents the 95% CI; width of the shaded square represents the size of the population (random effects model was

used as I2 >25%). Abbreviations as in Figures 1 and 4.

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 . 5 , N O . 1 , 2 0 1 9 Romero et al.J A N U A R Y 2 0 1 9 : 1 3 – 2 4 Endo-Epicardial vs. Endocardial Only VT Ablation

21

significant benefit in ARVC patients, consistent withprior small studies. These findings are not surprising,as ARVC substrate is diffuse, manifesting itselfinitially in the subepicardium and eventually pro-gressing to the subendocardium (5,6). Adequate con-trol of VT recurrence in ARVC is critically dependenton targeting both the endocardial and the epicardiallow-voltage areas: the endo component is crucial fortargeting the perivalvular substrate, and the epicardialcomponent is used to ablate any remaining sponta-neous or inducible VTs (28). An ablation strategy tar-geting all regions of substrate may be necessary owingto the potential for multiple re-entrant VTs, as suc-cessful ablation of the clinical VT alone may not beassociated with improved long-term freedom from VTrecurrence (5). Because the RV wall may be thickenedin some of these patients, a combined endo-epi

approach is more likely to achieve full-thicknessablation with subsequent elimination of all thearrhythmogenic substrate (5,6). However, it must benoted that there appears to be a negative correlationbetween the endocardial scar area and the presenceof epicardial arrhythmogenic substrate (i.e., exten-sive endocardial scar area is associated with limitedepicardial substrate) (31). This fact can be used todecide between the endo only or the endo-epi abla-tion procedure as a first-line approach. Of note, inpatients with advanced ARVC, the RV is dilated, andthe risk of inadvertently perforation during epicar-dial access is higher.

Evidence supporting routine endo-epi ablation inICM is less compelling. VT circuits in ICM are gener-ally thought to be subendocardial, easily accessible,and targeted with endocardial mapping and ablation

Romero 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 . 5 , N O . 1 , 2 0 1 9

Endo-Epicardial vs. Endocardial Only VT Ablation J A N U A R Y 2 0 1 9 : 1 3 – 2 4

22

(32,33). However, 34% to 75% of ICM patients mayexhibit epicardial substrate (19). Recent studies havedescribed epicardial substrates in ICM that aremore frequently found in patients with inferiorwall myocardial infarction (34). Advances in phar-macological and/or mechanical reperfusion therapyhave led to higher rates of nontransmural necrosisconsisting of fibrous tissue and bundles of myocytessurrounded by complex border zones that merge withnormal myocardium in a patchy pattern (33). “No-reflow” may increase the complexity of the VT sub-strate, with multiple slow conduction channelslocated intramurally or epicardially. To reach theseareas, a combined approach may be needed, as ven-tricular muscle thickness could limit appropriate en-ergy delivery from the endocardium. This couldexplain why a combined endo-epi ablation would bebeneficial in ICM (20). Nevertheless, it must be notedthat a significant proportion of patients who undergoepicardial mapping do not exhibit epicardialarrhythmogenic substrate (19). Use of an imagingmodality such as late gadolinium enhancement car-diac magnetic resonance (LGE-CMR) to identify pa-tients with a classic transmural MI (defined ashyperenhancement of >75% of wall thickness) wouldbe of immense help in determining which individualpatients with ICM would benefit from an endo-epi CAas the first approach. A recent study showed thatthose patients with transmural myocardial infarctionas defined by LGE-CMR criteria and with an epicardialscar area >14 cm2 were more likely to have anepicardial arrhythmogenic substrate (35). In thecontext of the limitations hitherto described, ourstudy findings are consistent with a benefit fromendo-epi ablation in ICM. A randomized study isneeded to corroborate our findings.

No statistically significant differences in VT recur-rence were observed when the endo-epi approach wascompared with the endo approach in NICM. Therewas, however, a nonsignificant trend towarddecreased risk of VT recurrence. A number of factorsmay explain the lack of statistically significant posi-tive effect in NICM. First, a major confounder is thevariable effect of epicardial ablation among a hetero-geneous group of disorders classified under the NICM“umbrella,” such as myocarditis, genetic heart dis-ease, and possibly undiagnosed sarcoid or left-dominant ARVC. Second, some NICM substrates arepurely mid-myocardial, for which endo-epi ablationmay not be successful. Furthermore, only the basal-lateral phenotype is associated with epicardialarrhythmogenic substrate that would benefit from theendo-epi approach, whereas the basal septal subtypewould not. Third, despite the presence of epicardial

substrate, ablation may not be possible due to closeproximity to the phrenic nerve, coronary arteries, orbasal epicardial fat, limiting radiofrequency ablationdepth. Fourth, NICM is known to have fewer ablationtargets (e.g., fractionated, late or delayed potentials,conducting channels) than ICM and ARVC (16).

Subanalysis of NICM type by scar location (e.g., byusing imaging techniques such as LGE-CMR to detectepicardial scar) may help differentiate the benefit ofthe endo-epi ablation approach. It has been suggestedthat the addition of LGE-CMR to endo bipolar voltagemapping (i.e., image integration) could improvedetection of an epicardial substrate for VT in NICM(36). Unfortunately, this information was not avail-able in any of the included studies.

The current findings do suggest that endo-epiablation is associated with a statistically significantreduction in all-cause mortality (RR: 0.56; 95% CI:0.32 to 0.97; p ¼ 0.04) compared with endo ablationalone. This reduction in mortality may be due to acombination of factors such as decrease in lethal VAand decrease in number of ICD shocks. As expected,acute procedural complications were higher in theendo-epi group than in the endo ablation group, butthe difference was not significant. This is likelyentirely related to epicardial access complications(Online Table S2). Whether the benefit of VT recur-rence and/or mortality negates the increased risk ofcomplications needs further study in a prospectiverandomized study.

Findings of the present meta-analysis suggest thatit may be reasonable to consider a combined endo-epiablation to patients with VT due to structural heartdisease, especially if imaging demonstrates the pres-ence of an epicardial scar. It must be noted that allstudies included in this meta-analysis came fromhigh-volume centers for epicardial access, hencecaution is advised when generalizing these conclu-sions in centers that have low-volume or are lessexperienced with epicardial ablation.STUDY LIMITATIONS. Our meta-analysis has limita-tions which must be considered when analyzing ourresults. Most of the studies included for review wereretrospective, single-center studies, which are subjectto confounding factors and bias. The few prospectivestudies that were included were not randomizedcontrolled trials. Nonetheless, endo-epi CA for VT is acomplex procedure, generally performed in high-volume, tertiary care referral centers with a limitednumber of patients. As such, a multicenter random-ized trial, although ideal, would be difficult to un-dertake, and our meta-analysis offers the bestevidence to date for the positive impact of a combinedapproach in ablation of VT associated with SHD.

PERSPECTIVES

COMPETENCY IN MEDICAL KNOWLEDGE: No firm evi-

dence in support of the endo-epi ablation of VT compared with

the endo approach for prevention of recurrent VT has been

published previously. AADs are frequently unsuccessful in con-

trolling VT in patients with defibrillators, resulting in recurrent

shocks. Shocks are associated with decreased quality of life and

increased mortality. Given the overall mortality benefit and the

decrease in VT recurrence seen in certain patient populations

(ICM and ARVC), it is reasonable to propose the endo-epi abla-

tion approach as the first-line treatment in properly selected

patients in whom recurrent VT secondary to ICM or ARVC is not

well controlled with optimal medical therapy, as well as in those

in whom there is a high chance of having an epicardial substrate.

TRANSLATIONAL OUTLOOK: The endo-epi approach may be

considered a first-line option in the ablation of VT in patients

with structural heart disease due to its benefit in reducing VT

recurrence and possible reduction in mortality. Procedural com-

plications, however, are higher.

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 . 5 , N O . 1 , 2 0 1 9 Romero et al.J A N U A R Y 2 0 1 9 : 1 3 – 2 4 Endo-Epicardial vs. Endocardial Only VT Ablation

23

CONCLUSIONS

In the present meta-analysis, a combined endo-epiapproach is superior to endo CA alone in terms ofreduction in VT recurrence and reduction in overallmortality in patients with SHD. Within each type ofheart disease (ICM, NICM, ARVC), there appeared tobe a mortality difference favoring the endo-epi abla-tion approach. An endo-epi approach also favored areduction in VT recurrence in ICM and ARVC patients,with a statistical trend in NICM patients. Complica-tions were higher with an endo-epi approach. Thecombined endo-epi approach may enhance clinicaloutcomes of CA for VT and needs to be tested in arandomized trial.

ADDRESS FOR CORRESPONDENCE: Assoc. Prof.Saurabh Kumar, Cardiology Department, WestmeadHospital, Westmead Applied Research Centre, CornerHawkesbury and Darcy Roads, Westmead, New SouthWales–2145, Australia. E-mail: Saurabh.kumar@health.nsw.gov.au.

RE F E RENCE S

1. Tung R, Michowitz Y, Yu R, et al. Epicardialablation of ventricular tachycardia: an institutionalexperience of safety and efficacy. Heart Rhythm2013;10:490–8.

2. Moss AJ, Zareba W, Hall WJ, et al. Prophylacticimplantation of a defibrillator in patients withmyocardial infarction and reduced ejection frac-tion. N Engl J Med 2002;346:877–83.

3. Poole JE, Johnson GW, Hellkamp AS, et al.Prognostic importance of defibrillator shocks inpatients with heart failure. N Engl J Med 2008;359:1009–17.

4. Santangeli P, Muser D, Maeda S, et al.Comparative effectiveness of antiarrhythmic drugsand catheter ablation for the prevention ofrecurrent ventricular tachycardia in patients withimplantable cardioverter-defibrillators: a system-atic review and meta-analysis of randomizedcontrolled trials. Heart Rhythm 2016;13:1552–9.

5. Philips B, Madhavan S, James C, et al. Outcomesof catheter ablation of ventricular tachycardia inarrhythmogenic right ventricular dysplasia/car-diomyopathy. Circ Arrhythm Electrophysiol 2012;5:499–505.

6. Bai R, Di Biase L, Shivkumar K, et al. Ablationof ventricular arrhythmias in arrhythmogenic rightventricular dysplasia/cardiomyopathy: arrhythmia-free survival after endo-epicardial substrate basedmapping and ablation. Circ Arrhythm Electrophysiol2011;4:478–85.

7. Izquierdo M, Sanchez-Gomez JM, Ferrero deLoma-Osorio A, et al. Endo-epicardial versus only-endocardial ablation as a first line strategy for thetreatment of ventricular tachycardia in patients

with ischemic heart disease. Circ Arrhythm Elec-trophysiol 2015;8:882–9.

8. Bradfield JS, Tung R, Boyle NG, Buch E,Shivkumar K. Our approach to minimize risk ofepicardial access: standard techniques with theaddition of electroanatomic mapping guidance.J Cardiovasc Electrophysiol 2013;24:723–7.

9. Kumar S, Bazaz R, Barbhaiya CR, et al. “Needle-in-needle” epicardial access: Preliminary observa-tions with a modified technique for facilitatingepicardial interventional procedures. HeartRhythm 2015;12:1691–7.

10. Scanavacca MI, Venancio AC, Pisani CF, et al.Percutaneous transatrial access to the pericardialspace for epicardial mapping and ablation. CircArrhythm Electrophysiol 2011;4:331–6.

11. Knobloch K, Yoon U, Vogt PM. Preferredreporting items for systematic reviews and meta-analyses (PRISMA) statement and publicationbias. J Craniomaxillofac Surg 2011;39:91–2.

12. Wells GS, O’Connell B, Peterson D, Welch J,Losos V, Tugwell M. The Newcastle-Ottawa Scale(NOS) for assessing the quality of nonrandomisedstudies in meta-analyses. Available at: https://www.researchgate.net/publication/261773681_The_Newcastle-Ottawa_Scale_NOS_for_Assessing_the_Quality_of_Non-Randomized_Studies_in_Meta-Analysis.Accessed September 17, 2018.

13. Moher D, Liberati A, Tetzlaff J, Altman DPG.Preferred reporting items for systematic reviewsand meta-analyses: the PRISMA statement. Int JSurg 2010;8:336–41.

14. DerSimonian R, Laird N. Meta-analysis in clin-ical trials. Control Clin Trials 1986;7:177–88.

15. Egger M, Davey Smith G, Schneider M,Minder C. Bias in meta-analysis detected by asimple, graphical test. BMJ 1997;315:629–34.

16. Nakahara S, Tung R, Ramirez RJ, et al. Char-acterization of the arrhythmogenic substrate inischemic and nonischemic cardiomyopathy impli-cations for catheter ablation of hemodynamicallyunstable ventricular tachycardia. J Am Coll Cardiol2010;55:2355–65.

17. Vergara P, Trevisi N, Ricco A, et al. Late po-tentials abolition as an additional technique forreduction of arrhythmia recurrence in scar relatedventricular tachycardia ablation. J CardiovascElectrophysiol 2012;23:621–7.

18. Berte B, Sacher F, Venlet J, et al. VT recurrenceafter ablation: incomplete ablation or diseaseprogression? A multicentric European study.J Cardiovasc Electrophysiol 2016;27:80–7.

19. Acosta J, Fernandez-Armenta J, Penela D,et al. Infarct transmurality as a criterion for first-line endo-epicardial substrate-guided ventriculartachycardia ablation in ischemic cardiomyopathy.Heart Rhythm 2016;13:85–95.

20. Di Biase L, Santangeli P, Burkhardt DJ, et al.Endo-epicardial homogenization of the scar versuslimited substrate ablation for the treatment ofelectrical storms in patients with ischemic cardio-myopathy. J Am Coll Cardiol 2012;60:132–41.

21. Arya A, Bode K, Piorkowski C, et al. Catheterablation of electrical storm due to monomorphicventricular tachycardia in patients with non-ischemic cardiomyopathy: acute results and itseffect on long-term survival. Pacing Clin Electro-physiol 2010;33:1504–9.

Romero 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 . 5 , N O . 1 , 2 0 1 9

Endo-Epicardial vs. Endocardial Only VT Ablation J A N U A R Y 2 0 1 9 : 1 3 – 2 4

24

22. Killu AM, Mulpuru SK, Al-Hijji MA, et al. Out-comes of combined endocardial-epicardial abla-tion compared with endocardial ablation alone inpatients who undergo epicardial access. Am JCardiol 2016;118:842–8.

23. Santangeli P, Di Biase L, Lakkireddy D, et al.Radiofrequency catheter ablation of ventriculararrhythmias in patients with hypertrophic cardio-myopathy: safety and feasibility. Heart Rhythm2010;7:1036–42.

24. Piers SR, Leong DP, van Huls van Taxis CF,et al. Outcome of ventricular tachycardia ablationin patients with nonischemic cardiomyopathy: theimpact of noninducibility. Circ Arrhythm Electro-physiol 2013;6:513–21.

25. Soejima K, Stevenson WG, Sapp JL, Selwyn AP,Couper G, Epstein LM. Endocardial and epicardialradiofrequency ablation of ventricular tachycardiaassociated with dilated cardiomyopathy: theimportance of low-voltage scars. J Am Coll Cardiol2004;43:1834–42.

26. Mussigbrodt A, Efimova E, Knopp H, et al.Should all patients with arrhythmogenic rightventricular dysplasia/cardiomyopathy undergoepicardial catheter ablation? J Interv Card Elec-trophysiol 2017;48:193–9.

27. Wei W, Liao H, Xue Y, et al. Long-term out-comes of radio-frequency catheter ablation onventricular tachycardias due to arrhythmogenic

right ventricular cardiomyopathy: a single centerexperience. PLoS One 2017;12:e0169863.

28. Santangeli P, Zado ES, Supple GE, et al. Long-term outcome with catheter ablation of ventricu-lar tachycardia in patients with arrhythmogenicright ventricular cardiomyopathy. Circ ArrhythmElectrophysiol 2015;8:1413–21.

29. Sapp JL, Wells GA, Parkash R, et al. Ventriculartachycardia ablation versus escalation of antiar-rhythmic drugs. N Engl J Med 2016;375:111–21.

30. Briceno DF, Romero J, Villablanca PA, et al.Long-term outcomes of different ablation strate-gies for ventricular tachycardia in patients withstructural heart disease: systematic review andmeta-analysis. Europace 2018;20:104–15.

31. Berruezo A, Acosta J, Fernandez-Armenta J, et al.Safety, long-term outcomes and predictors of recur-rence after first-line combined endoepicardial ven-tricular tachycardia substrate ablation inarrhythmogenic cardiomyopathy. Impact ofarrhythmic substrate distribution pattern. A prospec-tive multicentre study. Europace 2017;19:607–16.

32. Tschabrunn CM, Roujol S, Nezafat R, et al.A swine model of infarct-related reentrant ven-tricular tachycardia: electroanatomic, magneticresonance, and histopathological characterization.Heart Rhythm 2016;13:262–73.

33. Arenal A, Perez-David E, Avila P, et al.Noninvasive identification of epicardial ventricular

tachycardia substrate by magnetic resonance-based signal intensity mapping. Heart Rhythm2014;11:1456–64.

34. Ling Z, Hari A, Tandri H. VT ablation: newdevelopments and approaches. Curr Treat OptionsCardiovasc Med 2014;16:297.

35. Soto-Iglesias D, Acosta J, Penela D, et al. Im-age-based criteria to identify the presence ofepicardial arrhythmogenic substrate in patientswith transmural myocardial infarction. HeartRhythm 2018;15:814–21.

36. Mukherjee RK, Whitaker J, Williams SE,Razavi R, O’Neill MD. Magnetic resonanceimaging guidance for the optimization of ven-tricular tachycardia ablation. Europace 2018;20:1721–32.

KEY WORDS arrhythmogenic rightventricular cardiomyopathy, catheterablation, endocardial ablation, epicardialablation, ischemic cardiomyopathy,nonischemic cardiomyopathy, structuralheart disease, ventricular tachycardia, VTrecurrence

APPENDIX For supplemental figures and ta-bles, please see the online version of this paper.