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

ª 2 0 1 9 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

Electrophysiologic Substrate, Safety,Procedural Approaches, and Outcomesof Catheter Ablation for VentricularTachycardia in Patients AfterAortic Valve Replacement

Jackson J. Liang, DO, Simon A. Castro, MD, Daniele Muser, MD, David F. Briceno, MD, Yasuhiro Shirai, MD,Andres Enriquez, MD, Ramanan Kumareswaran, MD, Pasquale Santangeli, MD, PHD, Erica S. Zado, PA-C,Jeffrey S. Arkles, MD, Robert D. Schaller, DO, Gregory E. Supple, MD, David S. Frankel, MD,Saman Nazarian, MD, PHD, Michael P. Riley, MD, PHD, Fermin C. Garcia, MD, David Lin, MD, Sanjay Dixit, MD,David J. Callans, MD, Francis E. Marchlinski, MD

ABSTRACT

ISS

Fro

syl

Bio

rep

All

sti

the

Ma

OBJECTIVES This study sought to investigate the substrate, procedural strategies, safety, and outcomes of catheter

ablation (CA) for ventricular tachycardia (VT) in patients with aortic valve replacement (AVR).

BACKGROUND VT ablation in patients with AVR is challenging, particularly when mapping and ablation in the peri-

aortic region are necessary.

METHODS We identified consecutive patients with mechanical, bioprosthetic, and transcatheter AVR who underwent

CA for VT refractory to antiarrhythmic drugs and analyzed VT substrate, approach to LV access, complications, and long-

term outcomes.

RESULTS Overall, 29 patients (87% men, mean age 67.9 � 9.8 years, left ventricular ejection fraction 39 � 10%)

with prior AVR (13 mechanical, 15 bioprosthetic, 1 transcatheter AVR) underwent 40 ablations from 2004 to 2016.

Left-sided mapping/CA was performed in 27 patients (36 procedures). Access was retrograde aortic in 11 procedures

(all bioprosthetic), transseptal in 24 (13 mechanical; 10 bioprosthetic; 1 transcatheter AVR), or transventricular septal

in 1. Periaortic bipolar or unipolar scar was detected in all 24 patients in whom detailed periaortic mapping was

performed. Clinical VT circuit(s) involved the periaortic region in 10 patients (34%), 2 (7%) had bundle branch re-entry

VT, and 17 (59%) had substrate unrelated to AVR. There were 2 major complications (both related to vascular access).

Only 2 patients (9.1%) had VT recurrence. Over median follow-up of 12.8 months, 11 patients died (none as a result of

recurrent VT).

CONCLUSIONS Whereas most patients undergoing CA for VT after AVR had VT from substrate unrelated to AVR,

periaortic scar is universally present and bundle branch re-entry can be the VT mechanism. CA can be safely performed

with excellent long-term VT elimination. (J Am Coll Cardiol EP 2019;5:28–38) © 2019 by the American College of

Cardiology Foundation.

N 2405-500X/$36.00 https://doi.org/10.1016/j.jacep.2018.08.008

m the Electrophysiology Section, Cardiovascular Division, Hospital of the University of Pennsylvania, Philadelphia, Penn-

vania. Dr. Arkles provides consulting services to Biosense Webster. Dr. Nazarian provides consulting services to Siemens,

sense Webster, and CardioSolv; and has received research grants from Biosense Webster and ImriCor. All other authors have

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

authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ in-

tutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit

JACC: Clinical Electrophysiology author instructions page.

nuscript received April 9, 2018; revised manuscript received August 13, 2018, accepted August 14, 2018.

AB BR E V I A T I O N S

AND ACRONYM S

AAD = antiarrhythmic drugs

AVR = aortic valve

replacement

BBR = bundle branch re-entry

CA = catheter ablation

ICD = implantable

cardioverter-defibrillator

ICE = intracardiac

echocardiography

IQR = interquartile range

LV = left ventricle

LVOT = left ventricular

outflow tract

NIPS = noninvasive

programmed stimulation

TAVR = transcatheter aortic

valve replacement

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 Liang et al.J A N U A R Y 2 0 1 9 : 2 8 – 3 8 VT Ablation After Aortic Valve Replacement

29

S ustained monomorphic ventricular tachycardia(VT) occurs most commonly in patients withstructurally abnormal hearts in the presence

of scar due to prior myocardial infarction or nonische-mic cardiomyopathy. Catheter ablation (CA) is aneffective treatment option to eliminate VT and pre-vent implantable cardioverter-defibrillator (ICD)shocks. When VT occurs after valve surgery, themechanism of VT (i.e., underlying ischemic cardio-myopathy and prior infarction or nonischemic cardio-myopathy) may be unrelated to aortic valvereplacement (AVR). Additionally, VT can also origi-nate from perivalvular substrate related to the priorAVR or due to bundle branch re-entry (BBR) VT.AVR-related VT has been reported to present in abimodal manner—either immediately post-operatively or years later (1). Ablation in patientswith prior AVR is feasible, but it may be technicallycomplicated when access to the left ventricle (LV) isnecessary, particularly when the VT substrate in-volves the periaortic region. Small series have re-ported on the safety and outcomes after VT ablationin patients, but data remain limited.

We report our institution’s experience with VTablation in patients with prior AVR. Specifically, wedescribe patient characteristics, electrophysiologicfindings including VT mechanism and underlying VTsubstrate, ablation techniques, and approach to LVaccess, safety, and long-term outcomes.

METHODS

PATIENT POPULATION. We screened our in-stitution’s VT ablation database to identify all pa-tients with prior AVR who were treated with VTablation. Of the 3,367 VT and premature ventricularcomplex ablation procedures performed at the Uni-versity of Pennsylvania between January 1, 2004 andDecember 31, 2015, we identified a total of 40 VTablation procedures performed in 29 unique patientswho had undergone prior AVR. Per institutionalguidelines of the University of Pennsylvania HealthSystem, all patients provided written informed con-sent both for CA and for their anonymized medicalinformation to be included in research studies.

ELECTROPHYSIOLOGIC STUDY, MAPPING, AND

ABLATION. For patients with mechanical AVR orthose who were on warfarin for other reasons,warfarin was held and therapeutic unfractionatedheparin infusion was initiated when the internationalnormalized ratio became subtherapeutic, targeting agoal-activated partial thromboplastin time of 56 to80 s. After providing informed consent, patients un-derwent electrophysiologic study and ablation in the

post-absorptive state. When feasible, beta-blocker, calcium channel blocker, and anti-arrhythmic medications were discontinuedfor at least 5 half-lives before the study, andintravenous antiarrhythmic drugs (AAD) werestopped 12 h before the procedure. Conscioussedation was utilized preferentially; generalanesthesia was induced prior to obtainingepicardial access or when necessary for pa-tient comfort or stability. In each case an 8-FAcuNav (Siemens Medical Solutions, Moun-tain View, California) or 10-F SoundStar(Biosense Webster, Diamond Bar, California)intracardiac echocardiography (ICE) probewas advanced into the right atrium and rightventricle to define anatomy, facilitate map-ping, and assess contact during ablation.Electroanatomical mapping (Carto, BiosenseWebster) was performed during sinus orpaced rhythm to define areas of low voltage

and abnormal electrograms consistent with scar (withbipolar and unipolar voltage cutoffs of <1.5 mVand <8.3 mV, respectively) (2,3). When assessingconfluent bipolar and unipolar voltage regions basedon amplitude (including for assessment of “peri-aortic” scar), we excluded areas <1.0 cm of the aorticand mitral valve annuli (as defined on ICE) from themeasurement. The periaortic region was defined asbeing involved with VT in cases where abnormalelectrograms and critical sites responsible for VT wereidentified in the periaortic region within 2 cm of theaortic valve annulus—either with activation orentrainment mapping when possible, or with sub-strate characterization including abnormal electro-grams and concordant pace maps for unstable VT.Intravenous heparin was administered to maintain anactivated clotting time above 300 s prior to LV access.In patients with a mechanical aortic valve, the valvewas not crossed in a retrograde fashion to gain accessto the LV endocardium for mapping. Otherwise, LVaccess was obtained either via transseptal, retrogradeaortic, or transventricular septal approach, per theoperator’s discretion. Long sheaths were occasionallyused per the operator’s discretion in patients inwhom a retrograde aortic approach was chosen,mainly to improve support to permit cathetermanipulation and achieve stability during ablationwithin the LV.

Mapping and ablation of VT was performed as wehave previously described (2–4). Briefly, when sus-tained VT was not present spontaneously or withcatheter manipulation, programmed electrical stim-ulation or burst pacing from multiple sites andisoproterenol infusion (2 to 20 mg/min) were

VT =

Liang 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

VT Ablation After Aortic Valve Replacement J A N U A R Y 2 0 1 9 : 2 8 – 3 8

30

performed in an attempt to provoke ventricular ar-rhythmias. The mechanism of VT (re-entrant vs.automatic/triggered) was determined via pacingmaneuvers during sustained VT. In the case ofautomatic/triggered VT, detailed activation mappingand pace-mapping were performed to approximatethe site of origin. In patients with re-entrant VT,entrainment mapping and ablation was performed inour standard fashion as we have previouslydescribed targeting a goal of complete non-inducibility of VT (4). Ablation in the LV wasdelivered using an open-irrigated 3.5-mm tip cath-eter (Navistar Thermocool, Biosense Webster) withpower 30 to 45 W and temperature limit 42�C toachieve impedance drops of 10 to 15 U. Duringablation in the periaortic region, power was gener-ally set at 20 W to begin and titrated up to 40 to45 W, targeting impedance drops of 10 to 15 U or10% to 12%. Ablation of automatic/triggered VT wasdeemed successful with immediate suppression andin the absence of spontaneous/inducible prematureventricular complexes/VT after repeating the in-duction protocol with isoproterenol and/or ventric-ular pacing. A minimum of 60-min waiting periodwas observed after successful elimination of theventricular arrhythmia. Ablation for re-entrant VTwas considered successful when clinical VT wasrendered noninducible after ablation. At the end ofthe case, arterial hemostasis was achieved usingPerclose ProGlide closure device (Abbott Vascular,Santa Clara, California) or manual compression, andvenous hemostasis was achieved with manualcompression.

POST-ABLATION MANAGEMENT. Following the pro-cedure, patients were monitored on telemetry forrecurrent VT. Heparin infusion was reinitiated with nobolus 4 h after hemostasis was achieved, and warfarinwas typically restarted the day after the procedure.For patients in whom clinical VT was unable to beinduced at the end of the ablation procedure and didnot occur spontaneously post-ablation, noninvasiveprogrammed stimulation (NIPS) was typically per-formedwithin several days of ablation (5). For patientsin whom clinical or nonclinical VT was inducible atNIPS, repeat ablation, AAD adjustment, or ICDreprogramming was performed per physician’sdiscretion based on inducible VT information. Thedecision to continue, decrease, or discontinue AADincluding amiodarone after ablation was left to thediscretion of the physician and was based on the pre-sumed likelihood of ablation success, as indexed byinducibility of clinical and nonclinical VT at the end ofthe procedure and at NIPS (4).

CLINICAL FOLLOW-UP. For patients who chose tofollow-up at the Penn Arrhythmia Center, our routinepractice is to arrange an outpatient follow-upappointment 6 weeks following ablation and then at3- to 6-month intervals thereafter. For patients whochose to continue their clinical follow-up elsewhere,we attempted to contact either the patient or refer-ring cardiologists at 6- to 12-month intervals andreviewed ICD interrogations to determine arrhythmiarecurrence. For patients who were lost to follow-upfrom our institution before 2014, vital status wasalso assessed by querying the Social Security DeathIndex for mortality data available up to February1, 2014.

STATISTICAL ANALYSIS. Continuous variables wereexpressed as mean � SD if normally distributed ormedian (interquartile range [IQR]: 25th, 75th percen-tile) if not normally distributed. All continuous vari-ables were tested for normal distribution using the 1-sample Kolmogorov–Smirnov test. Categorical datawere expressed as counts and percentages. Survivalcurves were generated by the Kaplan–Meier method.All the analyses were performed using IBM SPSSsoftware version 23.0 (Armonk, New York).

RESULTS

Between 2004 and 2016, 29 patients (87% men, meanage 67.9 � 9.8 years) with prior AVR underwent 40 VTablation procedures. Baseline characteristics at timeof VT ablation for all patients can be found in Table 1.A total of 10 patients (34%) had at least 1 prior failedVT ablation attempt before being referred to ourcenter (range 1 to 4 prior failed attempts elsewhere).Over the entire study period, 7 patients (24%)underwent $2 ablation procedures at our institution(range 2 to 4 ablations: 2 procedures in 4 patients, 3procedures in 2 patients, and 4 procedures in 1 pa-tient). The type of AVR was mechanical AVR in 13,bioprosthetic AVR in 15, and transcatheter aorticvalve replacement (TAVR) in 1. Mean LVEF was 40 �10% and 12 patients (41%) had LV ejectionfraction #35%. Overall, there were 13 patients (45%)with only ischemic cardiomyopathy and a history ofmyocardial infarction, 10 patients (34%) with onlynonischemic cardiomyopathy (1 with cardiacsarcoidosis diagnosed with cardiac magnetic reso-nance imaging and positron emission tomography,and 9 with idiopathic dilated or valvular cardiomy-opathy), and 6 patients (21%) with mixed ischemic/nonischemic cardiomyopathy (with scar on imagingor voltage abnormalities on electroanatomical map-ping in regions not related to prior known myocardial

TABLE 2 Technical and Electrophysiologic Characteristics of

VT Ablations

Sedation approach

Conscious sedation only 24 (60.0)

General anesthesia 16 (40.0)

Total procedure time, min

Median 295 (205.5, 360.0)

Mean 309.5 � 125.9

Total fluoroscopy time, min

Median 52.4 (31.9, 61.0)

Mean 50.6 � 25.3

Total radiofrequency ablationduration, min

Median 56.5 (20.2, 71.0)

Mean 59.3 � 49.7

Number of radiofrequency ablationlesions delivered

Median 44 (15.3, 64.5)

Mean 41.6 � 27.6

Approach for access to LV (n ¼ 36)

Retrograde aortic 11 (30.6)

Transseptal 24 (66.7)

Transventricular septal 1 (2.8)

Number of spontaneous/inducible VT 2.7 � 2.0

Mean VT cycle length, ms 423 � 11

Periaortic bipolar scar present 20 (83.3)

Periaortic unipolar scar present 24 (100.0)

Clinical VT involved region of periaortic scar 13 (44.8)

VT mechanism

Re-entrant VT mechanism 24 (82.8)

Myocardial re-entrant VT only 22

BBR VT only 1

Both myocardial and BBR VT 1

Triggered/automatic VT 7 (24.1)

Both re-entry and triggered/automatic 2 (6.9)

Acute VT ablation outcomes

Complete noninducibility achievedat end of ablation

24 (58.5)

Clinical VT not inducible at end of ablation 41 (100.0)

Induction not attempted after ablation 4 (9.8)

Major complication 2 (4.9)

NIPS performed before hospital discharge 16 (39.0)

Complete noninducibility at NIPS 7 (43.8)

Clinical VT not inducible at NIPS 3 (18.8)

Values are n (%), mean � SD, and median (interquartile range: 25th, 75thpercentile).

BBR ¼ bundle branch re-entry; NIPS ¼ noninvasive programmed stimulation;other abbreviations as in Table 1.

TABLE 1 Baseline Clinical Characteristics of Patients With Prior

AVR Undergoing VT Ablation (N ¼ 29)

Male 26 (89.7)

Age, yrs 67.9 � 9.8

Mechanical AVR type 13 (44.8)

Bioprosthetic AVR type 15 (51.7)

Transcatheter AVR type 1 (3.4)

Interval between AVR and initial VTepisode, days

1,548(221, 3,806)

Interval between AVR and VT ablation, days 269(82, 703)

Ischemic cardiomyopathy 13 (44.8)

Nonischemic cardiomyopathy 10 (34.5)

Mixed ischemic/nonischemic cardiomyopathy 6 (20.7)

Mean LV ejection fraction, % 39.0 � 10.2

Hypertension 15 (51.7)

Diabetes mellitus 4 (13.8)

History of stroke or transient ischemic attack 6 (20.7)

History of atrial fibrillation 17 (58.6)

History of myocardial infarction 21 (72.4)

History of prior coronary artery bypass surgery 12 (41.4)

Implantable cardioverter defibrillator present 24 (82.8)

CRT-D present 10 (34.5)

Prior VT ablation, % 10 (34.5)

Medications at time of ablation

Beta-blocker 26 (89.7)

ACE inhibitor or angiotensin receptor blocker 18 (62.1)

Aldosterone antagonist 6 (20.7)

Diuretic 20 (69.0)

Warfarin 19 (65.5)

Digoxin 6 (20.7)

Any AAD 25 (86.2)

Amiodarone 21 (72.4)

Prior AAD use

Failed any AAD 27 (93.1)

Mean number of failed AADs 1.63 � 1.0

Failed amiodarone 25 (86.2)

Values are n (%), mean � SD, or median (interquartile range: 25th, 75thpercentile).

AAD¼ antiarrhythmic drug; ACE ¼ angiotensin converting enzyme; AVR ¼ aorticvalve replacement; CRT-D ¼ cardiac resynchronization therapy with defibrillator;IQR ¼ interquartile range; LV ¼ left ventricular; 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 Liang et al.J A N U A R Y 2 0 1 9 : 2 8 – 3 8 VT Ablation After Aortic Valve Replacement

31

infarction). A total of 12 patients (41%) had priorcoronary artery bypass graft surgery. An ICD andbiventricular ICD were present in 24 patients (83%)and 10 (35%), respectively, at time of VT ablation.Twenty-seven patients (93%) had failed $1 AAD, 25(86%) of whom had failed amiodarone. At the time ofVT ablation, 25 (86%) were taking $1 AAD and 21(72.4%) were taking amiodarone.

The first episode of sustained VT in 3 patientspreceded the AVR surgery and these patientscontinued to have persistent VT after AVR for whichVT ablation was performed. Meanwhile, in theremaining 26 patients, VT began after AVR surgery(median 65.5 months after AVR; IQR 11.5, 132.2

months). Median time interval between the last AVRsurgery and first VT ablation procedure was 9.0months (IQR: 2.7, 23.4 months). In 4 patients, VTablation was performed within 7 days of the AVRsurgery due to refractory VT in the post-operativesetting.

ELECTROPHYSIOLOGIC STUDY. Procedural charac-teristics can be found in Table 2. Conscious sedation

FIGURE 1 Approaches to Access to the LV for CA After AVR

(A) Right anterior oblique (top) and left anterior oblique (bottom) views demonstrating catheter ablation of the left ventricle (LV) through a

sheath (arrow pointing at ablation catheter) inserted from the right femoral artery via a retrograde aortic approach. (B) Right anterior oblique

(top) and left anterior oblique (bottom) views demonstrating catheter ablation of the LV through a sheath (arrow pointing at ablation catheter)

inserted from the right femoral vein via a transseptal approach. (C) Right anterior oblique (top) and left anterior oblique (bottom) views

demonstrating catheter ablation in the LV through a sheath (arrow pointing at ablation catheter) inserted from the right internal jugular vein

via a radiofrequency wire-facilitated transventricular septal approach in a patient with mechanical mitral and aortic valves. AVR ¼ aortic valve

replacement; CA ¼ catheter ablation.

Liang 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

VT Ablation After Aortic Valve Replacement J A N U A R Y 2 0 1 9 : 2 8 – 3 8

32

only was used in 24 procedures (60%), whereas gen-eral anesthesia was used in 16 (40%). Left-sidedmapping and/or ablation was performed in 27 pa-tients (36 total procedures). Of these, access was ob-tained via a retrograde aortic approach in 11procedures (31%) (all in patients with bioprostheticAVR), transseptal in 24 (67%) (13 mechanical AVR; 10bioprosthetic AVR; 1 TAVR), or transventricular septalin 1 (3%) in a patient with dual mechanical aortic andmitral valves (Figure 1). There was only 1 patient (withseptal substrate) who had different LV access ap-proaches at different ablation procedures. A trans-septal approach was used at the first ablation, and aretrograde aortic approach was instead used at thesecond ablation procedure to facilitate simultaneousunipolar ablation from both sides of the septum.There were no cases in this series in whom simulta-neous transseptal and retrograde accesses were

obtained. There was 1 patient with a bioprostheticAVR in whom the catheter could not be advancedthrough the aortic valve via a retrograde aorticapproach, thus a transseptal approach was usedinstead. Patients had a mean 2.7 � 2.0 spontaneous orinducible VT and mean cycle length was 423 � 111 ms.Of the 11 patients in whom a retrograde aorticapproach was utilized, 81-cm SL-1 sheaths (St. JudeMedical, St. Paul, Minnesota) were used in 3, 65-cmArrow sheaths (Teleflex Inc., Wayne, Pennsylvania)was used in 2, a 21-cm Arrow sheath was used in 1,and short 11-cm Arrow sheaths were used in theremaining 6 cases.

Endocardial voltage mapping with adequate sam-pling density of the periaortic region (within 1 cmbeneath the aortic valve) to identify the presence ofscar was performed in 24 patients (82.8%). Periaorticbipolar or unipolar voltage abnormalities (using

FIGURE 2 Periaortic Bipolar and Unipolar Voltage Abnormalities

Left ventricular electroanatomic maps depicting bipolar (Bi) (top) and unipolar (Uni) (bottom) of 3 representative patients who had evidence of

both bipolar (<1.5 mV) and unipolar (<8.3 mV) voltage abnormalities. Red dots represent regions where ablation was performed.

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 Liang et al.J A N U A R Y 2 0 1 9 : 2 8 – 3 8 VT Ablation After Aortic Valve Replacement

33

voltage cutoffs of <1.5 mV for bipolar and <8.3 mVfor unipolar mapping) with abnormal electrogramsextending beyond 1 cm from the aortic valveannulus were detected in all 24 of these patients(Figure 2). Bipolar periaortic voltage abnormalitieswere seen in 20 patients, whereas 4 additional pa-tients had normal bipolar but abnormal unipolarvoltages in the periaortic region. These periaorticvoltage abnormalities persisted after lowering theunipolar voltage scar cutoff to <6.0 mV (6). In theremaining 5 patients, the clinical VT were unrelatedto the periaortic region and the presence of peri-aortic scar could not be excluded due to inadequatesampling of the region. Clinical VT was felt to orig-inate from or involve regions of periaortic scar in 13patients, as confirmed by pace mapping andentrainment mapping (for re-entrant VT) or pacemapping and activation mapping (for automatic/triggered VT) (Figure 3).

Twenty-four patients had VT due to re-entry(myocardial re-entrant VT in 22, BBR VT only in 1,and both myocardial re-entrant VT and BBR VT in 1).Seven patients had VT due to triggered/automaticmechanism (successfully ablated from the periaorticLV in 3, aortic cusp region in 1, posteromedial

papillary muscle in 1, anterolateral papillary musclein 1, and apical septum in 1). Two patients had VT ofboth re-entrant and triggered/automatic mechanismsthat were targeted. One patient underwent 3 abla-tions over the course of 30 months for 2 different VTmechanisms (BBR VT in the first 2 ablations andperiaortic scar-related re-entrant VT in the thirdablation). Of the 4 patients who underwent VT abla-tion within 7 days of AVR, 1 patient had atypical BBRVT and the other 3 patients had myocardial VT notoriginating from the periaortic region. Of the 10 pa-tients with NICM, VT was targeted from the periaorticregion in 7 (70%). VT was targeted from regions ofprior infarction in all 13 patients with ICM. Of the 6patients with mixed ICM/NICM, in 3 (50%), the VTsubstrate was related to prior infarction, whereas inthe remaining 3, it was unrelated (basal septal sub-strate in 2, BBR VT in 1).

One patient underwent epicardial mapping andablation during 1 of his procedures. In that patient,epicardial access was obtained using percutaneoussubxiphoid puncture. Of note, due to prior cardiacsurgery, the patient had surgical adhesions and therewas difficulty in passing the guidewire freely in theepicardial space. As such, only the epicardial aspect

FIGURE 3 Clinical VT Morphology and Pace Map of VT Originating From Periaortic Region

Clinical 12-lead ventricular tachycardia (VT) morphology (A) and best pace maps (B) of the clinical VT in a patient with nonischemic cardio-

myopathy and VT originating from the periaortic region. Bipolar (C) and unipolar (D) electroanatomic maps (superior view) showing site of best

pace map (yellow star). The red dots are sites where ablation lesions were delivered to eliminate the clinical VT. Abbreviations as in Figure 2.

Liang 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

VT Ablation After Aortic Valve Replacement J A N U A R Y 2 0 1 9 : 2 8 – 3 8

34

of the septum and free wall of the LV to the superiorand lateral margin were able to be mapped in thispatient. Ablation was performed on the LV basal freewall 3 to 4 cm lateral to the left anterior descendingartery.

ABLATION. Details about the procedural aspects ofthe VT ablation procedures are listed in Table 2. Themedian total procedure time was 360 (IQR: 206, 296)min and median fluoroscopy time was 52.4 (IQR: 31.9,61) min. A median of 44 (IQR: 15.25, 64.5) radio-frequency lesions were delivered per procedure(median total radiofrequency ablation duration of56.5 [IQR: 20.2, 71.0] min per procedure). At the endof the first ablation procedure, all 29 patients werenoninducible for clinical VT, whereas 10 patients(34.5%) were inducible only for nonclinical VT at theend of the procedure. Overall (including repeat pro-cedures), repeat induction was attempted after abla-tion in 37 procedures (92.5%). Of these, completenoninducibility was achieved acutely after 24 pro-cedures and only nonclinical VT remained inducibleat the end of the remaining 13 procedures. NIPS wasperformed after 16 of the ablation procedures (mean2.6 � 1.2 days post-ablation), and post-ablation NIPSshowed no inducible VT in 7, inducible nonclinical VTonly in 6, and inducible clinical VT in 3.

After last ablation, 21 patients (72.4%) were dis-charged on AAD, 16 (55.2%) of whom were dischargedon amiodarone. Among the 21 patients who weretaking amiodarone prior to last ablation, amiodaronedosage at time of discharge was either decreased ordiscontinued altogether in 13 (61.9%).

Follow-up data regarding VT recurrence status af-ter hospital discharge post-ablation was available in22 patients. Long-term freedom from recurrent VTwas achieved in 16 patients (72.7%) after a singleprocedure. Meanwhile, allowing for multiple pro-cedures, only 2 patients (9.1%) experienced VTrecurrence after their last ablation (1,617 days and 177days after last ablation). Figure 4 shows the Kaplan-Meier survival curve for VT recurrence after a single(Figure 4A) and multiple (Figure 4B) procedures. Vitalstatus was able to be assessed in 26 patients, 11(42.3%) of whom have died since their last ablation,as assessed by a combination of clinical medical re-cords and querying of the Social Security Death In-dex, over a median follow-up duration of 12.8 (IQR:2.2, 36.5) months for mortality. Cause of death wasnoncardiac in 1 patient, cardiac in 3, and unknown inthe remaining 7. All 3 cardiac deaths were due toprogressive heart failure rather than ventricular ar-rhythmias. Figure 4C shows the Kaplan-Meier survivalcurve for overall survival after last VT ablation. AAD

FIGURE 4 Freedom From Recurrent VT and Overall Survival

Kaplan-Meier survival analysis showing survival free from ventricular tachycardia (VT) recurrence after the first (A) and last (B) ablation

procedure and cumulative survival (C) of the study population. *Of note, the initial number at risk is 28 because 1 patient had recurrent VT

immediately after the initial ablation requiring repeat ablation.

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 Liang et al.J A N U A R Y 2 0 1 9 : 2 8 – 3 8 VT Ablation After Aortic Valve Replacement

35

use at last follow-up was able to be assessed in 24patients. Of these, 15 (62.5%) remained on any AAD atlast follow-up, 9 (37.5%) of whom remained onamiodarone.

COMPLICATIONS. All patients survived to hospitaldischarge after VT ablation. Overall, there were 2 (5%)major complications related to the ablation proced-ure, both of which were vascular access complica-tions: right common iliac artery dissection in 1 patient(treated conservatively) and femoral pseudoaneur-ysm in another (requiring surgical repair). There wereno cerebral or systemic noncerebral arterial emboliccomplications. Pre- and post-ablation transthoracicechocardiography showed no change in aortic valvefunction after ablation and no worsening of aorticregurgitation. Importantly, in cases where retrogradeaortic access was attempted, careful intracardiacechocardiography assessment of aortic valve gradi-ents and degree of regurgitation with color flowDoppler was performed immediately before retro-grade access and at the end of the procedure toensure no change in aortic valve function or wors-ening of aortic regurgitation.

DISCUSSION

The major findings of our study is that in patientswith prior AVR referred for VT ablation, ablation canbe performed safely with excellent freedom from VTduring long-term follow-up. Regarding the underly-ing substrate, the presence of abnormal electro-grams in the periaortic region with bipolar or

unipolar voltage abnormalities consistent with scarare universally present, but these regions of scar arenot always related to the clinical VT circuits. In fact,in the majority of patients in our series, the VTsubstrate was usually unrelated to the AVR, such asscar from prior myocardial infarction or anotherunderlying nonischemic process (i.e., cardiacsarcoidosis).

SUBSTRATE AND VT MECHANISMS. Eckart et al. (1)previously reported mechanisms of VT in 20 pa-tients with prior valve surgery, 12 of whom had priorAVR and 2 more with dual aortic and mitral valvesurgery. In their study, the investigators excludedpatients with prior myocardial infarction. Theydemonstrated that VT related to valve surgery pre-sented in a bimodal manner—occurring either dayspost-surgery or years later. In their series, VT in 70%of patients was attributed to scar-related re-entry,whereas 10% had BBR VT. In their study (whichincluded 3 patients with lone mitral valve surgery), 9of 14 patients (64%) had periaortic bipolar scar, andonly 8 of 14 patients had bipolar voltage abnormal-ities seen adjacent to the valve that had been oper-ated on. Their results are slightly different than ourexperience, because we detected either bipolar orunipolar scar in all 24 patients who underwentdetailed voltage mapping of the periaortic region—even among several patients whose VT was presumedto be unrelated to the prior AVR. In recent years, ithas become well accepted that unipolar voltagemapping can be helpful in identifying mid-myocardial or epicardial substrate in patients

Liang 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

VT Ablation After Aortic Valve Replacement J A N U A R Y 2 0 1 9 : 2 8 – 3 8

36

without endocardial scar (3). Thus, our findingswould suggest that aortic valve surgery can injure themyocardium in the left ventricular outflow tract(LVOT) resulting in electrophysiologic scar, which inrare cases can be the substrate for re-entrant VT.Nagashima et al. (7) have reported that in patientswith VT from the LVOT in the absence of structuralheart disease, the presence of abnormal, low-voltageelectrograms in the periaortic region consistent withscar can be the underlying VT substrate. In their se-ries, the inducibility of multiple VT morphologiessuggested a re-entrant mechanism associated withperiaortic scar. Interestingly, all 8 patients withmultiple VT morphologies in whom ICE was used todefine the aortic annulus in their study had evidenceof periaortic scar (bipolar voltages <1.5 mV) at both1.0 and 1.5 cm from the AV annulus, whereas peri-aortic bipolar voltages at these sites were >1.5 mV inthe 2 patients with single VT morphologies as well asall patients with idiopathic premature ventricularcomplexes (7).

ACCESS TO THE LV. CA of VT in patients with priorAVR can be technically difficult, especially in patientswith periaortic VT substrate. In patients with me-chanical AVR, retrograde aortic access is prohibitedand a transseptal approach must be taken, as waspreviously described by Yamada et al. (8). Meanwhile,in the presence of bioprosthetic AVR or TAVR, it re-mains unclear whether retrograde aortic access isfeasible and can be safely performed. Srivatsa et al.(9) recently reported a case of periaortic ventriculararrhythmias in a patient with prior TAVR that weresuccessfully eliminated with ablation utilizing aretrograde aortic approach using the Stereotaxismagnetic navigational system (Stereotaxis Inc., St.Louis, Missouri). In our series, ablation was safelyperformed via transseptal access in patients withmechanical AVR and TAVR, whereas retrograde ac-cess was performed in several patients with bio-prosthetic AVR without any evidence of acutedamage to the valve or worsening of aortic regurgi-tation. There was only 1 patient with a bioprostheticAVR in whom retrograde aortic access failed due toinability to cross the valve, and a transseptalapproach was therefore necessary to access the LV. Inanother patient, the bioprosthetic AVR was unable tobe crossed in the standard fashion (with the cathetertip looped), but the LV was able to be accessedretrogradely when the catheter tip was straightenedand carefully passed directly through the aortic valveunder ICE guidance. In 1 patient with dual mechanicalaortic and mitral valves (“no-access LV”), access was

successfully obtained via transventricular septalapproach facilitated by a radiofrequency wire, as re-ported by Santangeli et al. (10).

The use of ICE allows for visualization of theaortic valve to assess for calcification as well asdegree of aortic stenosis and insufficiency beforethe decision is made to access the LV via a retro-grade versus transseptal approach. In cases wherethere is any concern about bioprosthetic valvefunction, a transseptal approach should be consid-ered. In cases where retrograde access is pursued,ICE (including with CARTOSOUND, BiosenseWebster) can aid in visualizing the aortic cusps tofacilitate passage of the ablation catheter throughthe valve into the LV. Additionally, ICE can visu-alize the ascending aorta to assess for calcificationand atheroma.

Mechanical hemodynamic support devices areoften utilized in high-risk patients undergoing VTablation, either prophylactically or as rescue therapyafter acute hemodynamic decline (11–14). We did notuse hemodynamic support devices in any proced-ures in this series. The presence of a mechanicalaortic valve is an absolute contraindication for theuse of the Impella (Abiomed, Danvers, Massachu-setts) device, and whereas the uncomplicated use ofImpella after bioprosthetic AVR has been reported,the safety of doing so remains unknown (15). Forthese patients, if mechanical hemodynamic supportis felt to be appropriate, extracorporeal membraneoxygenation or TandemHeart (Cardiac Assist Inc.,Pittsburgh, Pennsylvania) may be more suitablealternatives.

SAFETY AND EFFICACY. When VT ablation was per-formed by experienced operators, under ICE guid-ance, it can be safely performed with low rates ofserious complications. As previously reported to bethe most common complication after VT ablation,vascular access complications comprised the only twomajor complications in our series (16). It is importantto note that we did not attempt retrograde aortic ac-cess in any patients with mechanical AVR or TAVR.Theoretically, a retrograde aortic approach in patientswith a bioprosthetic aortic valve could damage thevalve leaflets and result in aortic regurgitation if notperformed carefully. Thus, ablation and cathetermanipulation across a bioprosthetic valve should al-ways be done carefully under ICE guidance and op-erators should perform imaging with transthoracicechocardiogram and/or ICE both pre- and post-ablation to evaluate for iatrogenic aortic regurgita-tion or damage to the aortic valve apparatus.

PERSPECTIVES

COMPETENCY IN MEDICAL KNOWLEDGE: In patients with

AVR and VT who undergo CA, periaortic voltage abnormalities on

electroanatomic mapping are universally present. The VT sub-

strate can involve the periaortic region or be unrelated to prior

AVR, and BBR can be the VT mechanism in some patients. CA can

be safely performed with excellent long-term VT elimination. A

retrograde approach can be utilized safely under ICE guidance in

patients with bioprosthetic AVR, whereas a transseptal approach

is necessary in those with mechanical AVR.

TRANSLATIONAL OUTLOOK: Larger series are necessary to

confirm the safety and efficacy of VT ablation in patients after

AVR and to further define the periaortic substrate in these pa-

tients. Further studies should also examine the safety and

feasibility of nontraditional methods to achieve LV access, as in

patients with mechanical mitral and aortic valves (“no-entry”

LV).

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 Liang et al.J A N U A R Y 2 0 1 9 : 2 8 – 3 8 VT Ablation After Aortic Valve Replacement

37

Furthermore, in no patients did we perform ablationin the aortic cusp region, so whether aortic cuspablation in patients after AVR can be safely performedremains unclear.

Of note, Whitman et al. (17) recently reported highrates (58%) of new silent cerebral emboli detected onpost-procedural brain magnetic resonance imagingin a small prospective series, including a 63% inci-dence of cerebral emboli in patients in whom aretrograde aortic approach was performed. Althoughwe did not perform brain imaging before and afterablation in patients to rule out silent cerebralemboli, there were no clinically significant cerebro-vascular or systemic noncerebral arterial emboliassociated with any of the ablation procedures.Whereas the clinical significance and persistence ofsilent cerebral emboli detected shortly after theprocedure remain unknown, we are unable toquantify the risk of such events in patients with AVRundergoing VT ablation, especially those with abioprosthetic valve in whom a retrograde aorticapproach is performed.

STUDY LIMITATIONS. This was a retrospective,single-center study and is thus subject to bias. Theprocedures were performed by experienced operatorsat a high-volume VT ablation center, thus safety andefficacy outcomes may not be generalizable to lessexperienced centers. Our institution is a tertiary carereferral center and due to concern for damaging theaortic valve in patients with prior AVR, many pro-viders may consider CA to be a treatment of last resortin this patient population. Thus, these patients arefrequently referred to us late in the course of theirdisease, which may affect likelihood of success withCA. Also, because we are a tertiary care center, manypatients were referred urgently for their VT ablationand chose to follow-up after the procedure with theirlocal providers. As such, long-term follow-up toassess for VT recurrence was unavailable for 7 pa-tients after last ablation. We were unable to ascertaincause of death in 7 patients who were followed upelsewhere and are therefore unable to excludewhether their deaths were related to recurrent ven-tricular arrhythmias. Finally, in our series we identi-fied 9 patients who had VT that we thought to be dueto a triggered/automatic VT mechanism, in 3 of whomVT was successfully ablated from the periaortic re-gion. As discussed previously (7), “idiopathic” VTfrom the LVOT in patients with structurally normalhearts has been demonstrated in some cases toactually originate from regions of periaortic scar andbe due to reentrant mechanisms. Even though we

were able to successfully eliminate VT with focalablation in the LVOT region in all patients with trig-gered/automatic VT, we were unable to completelyexclude micro–re-entry as the underlying VTmechanism.

CONCLUSIONS

In patients with prior AVR and symptomatic VT,catheter ablation is a safe and effective treatmentoption. A retrograde approach can be utilized safelyunder ICE guidance in patients with bioprostheticAVR, whereas a transseptal approach is necessaryin those with mechanical AVR. Although the ma-jority of patients referred for VT ablation after AVRhad VT from substrate unrelated to prior AVR, asignificant percentage (41.4%) had VT involvingperiaortic substrate or BBR. Periaortic bipolar orunipolar scar was universally present in patientsafter AVR in patients with adequate periaorticsampling density. Catheter ablation can be safelyperformed in patients after AVR with excellentlong-term VT elimination allowing for multipleablations.

ADDRESS FOR CORRESPONDENCE: Dr. Francis E.Marchlinski, Electrophysiology Section, Cardiovas-cular Division, Hospital of the University of Pennsyl-vania, 9 Founders Pavilion, 3400 Spruce Street,Philadelphia, Pennsylvania 19104. E-mail: Francis.Marchlinski@uphs.upenn.edu.

Liang 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

VT Ablation After Aortic Valve Replacement J A N U A R Y 2 0 1 9 : 2 8 – 3 8

38

RE F E RENCE S

1. Eckart RE, Hruczkowski TW, Tedrow UB,Koplan BA, Epstein LM, Stevenson WG. Sustainedventricular tachycardia associated with correctivevalve surgery. Circulation 2007;116:2005–11.

2. Marchlinski FE, Callans DJ, Gottlieb CD, Zado E.Linear ablation lesions for control of unmappableventricular tachycardia in patients with ischemicand nonischemic cardiomyopathy. Circulation2000;101:1288–96.

3. HutchinsonMD,GerstenfeldEP,DesjardinsB, etal.Endocardial unipolar voltage mapping to detectepicardial ventricular tachycardia substrate in pa-tients with nonischemic left ventricular cardiomyop-athy. Circ Arrhythm Electrophysiol 2011;4:49–55.

4. Liang JJ, Yang W, Santangeli P, et al. Amio-darone discontinuation or dose reductionfollowing catheter ablation for ventricular tachy-cardia in structural heart disease. J Am Coll CardiolEP 2017;3:503–11.

5. Frankel DS, Mountantonakis SE, Zado ES, et al.Noninvasive programmed ventricular stimulationearly after ventricular tachycardia ablation topredict risk of late recurrence. J Am Coll Cardiol2012;59:1529–35.

6. Santangeli P, Marchlinski FE. Substrate map-ping for unstable ventricular tachycardia. HeartRhythm 2016;13:569–83.

7. Nagashima K, Tedrow UB, Koplan BA, et al.Reentrant ventricular tachycardia originating from

the periaortic region in the absence of overtstructural heart disease. Circ Arrhythm Electro-physiol 2014;7:99–106.

8. Yamada T, McElderry HT, Doppalapudi H,Kay GN. Successful radiofrequency catheter abla-tion of ventricular tachycardia originating fromunderneath the mechanical prosthetic aortic valve.Pacing Clin Electrophysiol 2008;31:618–20.

9. Srivatsa UN, Nordsieck EJ, Pezeshkian N,Yang Y, Southard J. Ablation of ventriculartachycardia from the aortic root after trans-catheter aortic valve replacement. HeartRhythmCase Rep 2018;4:201–3.

10. Santangeli P, Shaw GC, Marchlinski FE. Radi-ofrequency wire facilitated interventricular septalaccess for catheter ablation of ventricular tachy-cardia in a patient with aortic and mitral me-chanical valves. Circ Arrhythm Electrophysiol2017;10. pii:e004771.

11. Baratto F, Pappalardo F, Oloriz T, et al. Extra-corporeal membrane oxygenation for hemody-namic support of ventricular tachycardia ablation.Circ Arrhythm Electrophysiol 2016;9. pii:e004492.

12. Reddy YM, Chinitz L, Mansour M, et al.Percutaneous left ventricular assist devices inventricular tachycardia ablation: multicenterexperience. Circ Arrhythm Electrophysiol 2014;7:244–50.

13. Lu F, Eckman PM, Liao KK, et al. Catheterablation of hemodynamically unstable ventriculartachycardia with mechanical circulatory support.Int J Cardiol 2013;168:3859–65.

14. Mathuria N, Wu G, Rojas-Delgado F, et al.Outcomes of pre-emptive and rescue use ofpercutaneous left ventricular assist device inpatients with structural heart disease under-going catheter ablation of ventricular tachy-cardia. J Interv Card Electrophysiol 2017;48:27–34.

15. Thomas MP, Altman A, Magovern GJ Jr.,Moraca RJ. Insertion of an Abiomed Impella(R) leftventricular assist device following bioprostheticaortic valve placement. J Card Surg 2013;28:469–71.

16. Pothineni NV, Deshmukh A,Padmanabhan D, et al. Complication rates ofventricular tachycardia ablation: comparison ofsafety outcomes derived from administrativedatabases and clinical trials. Int J Cardiol 2015;201:529–31.

17. Whitman IR, Gladstone RA, Badhwar N, et al.Brain emboli after left ventricular endocardialablation. Circulation 2017;135:867–77.

KEY WORDS aortic valve replacement,cardiomyopathy, catheter ablation, safety,ventricular tachycardia