Clinical implications of left ventricularassist device implantation in patientswith an implantable cardioverter-defibrillator

Fanny Boudghène-Stambouli & Stéphane Boulé & Céline Goéminne & Edward Botcherby &

Christelle Marquié & Claude Kouakam & Laurence Guédon-Moreau & Guillaume Schurtz &

Pascal de Groote & Nicolas Lamblin & Marie Fertin & Emmanuel Robin &

François Brigadeau & Didier Klug & Dominique Lacroix & Jonathan Meurice &

Ludivine Wissocque & André Vincentelli & Salem Kacet

Received: 26 July 2013 /Accepted: 17 October 2013 /Published online: 13 December 2013# Springer Science+Business Media New York 2013

AbstractPurpose This study aims to study the clinical implications ofthe concomitant use of a left ventricular assist device (LVAD)and an implantable cardioverter-defibrillator (ICD).Methods In this retrospective study, all patients whounderwent LVAD (Heart Mate II) implantation withconcomitant ICD therapy at our institution between June2007 and August 2012 were included. We sought toinvestigate (1) the electromagnetic interference betweenLVAD and ICD telemetry, (2) the effect of LVAD implantation

on right ventricular (RV) lead parameters and (3) theventricular tachyarrhythmias (VAs) that occur post-LVADimplantation.Results Of the 23 patients (53±9 years, 73 % male, LVEF 19±9 %) included, ICD telemetry was lost in four patients post-LVAD implantation (Saint-Jude-Medical Atlas V-193, V-240,V-243, and Sorin CRT-8750), prompting either use of a metalshield (n =1), a change in position of the programmer head(n =1) or ICD replacement (n =2). LVAD implantation wasassociated with a decrease in both RV signal amplitude (p =0.04) and RV impedance (p <0.01), and a trend towards anincreased RV pacing threshold (p =0.08), without affectingclinical outcome. Eleven patients (47.8 %) experienced VAsafter LVAD implantation, which on the whole were welltolerated. Their occurrence was strongly linked to a historyof VAs before device implantation (p <0.01).Conclusions Electromagnetic interference between LVADsand ICD telemetry may necessitate ICD replacement. LVADplacement is associated with significant changes in RV leadparameters that have minimal clinical significance. VAs occurin approximately half of LVAD patients seen and theiroccurrence is strongly related to a history of VAs prior toLVAD implantation.

Keywords Left ventricular assist device . Implantablecardioverter-defibrillator . Lead . Electromagneticinterference . Ventricular tachycardia . Ventricular fibrillation

AbbreviationsCRT-D Cardiac resynchronization therapy defibrillatorICD Implantable cardioverter-defibrillatorLVAD Left ventricular assist device

Fanny Boudghène-Stambouli and Stéphane Boulé contributed equally tothis work.

F. Boudghène-Stambouli : S. Boulé : C. Goéminne : E. Botcherby :C. Marquié :C. Kouakam : L. Guédon-Moreau :G. Schurtz :P. de Groote :N. Lamblin :M. Fertin : F. Brigadeau :D. Klug :D. Lacroix : J. Meurice : L. Wissocque : S. KacetDepartment of Cardiovascular medicine, Lille University Hospital,Lille, France

G. Schurtz :N. Lamblin :D. Klug :D. Lacroix :A. Vincentelli :S. KacetFaculty of Medicine, University of Lille 2, Lille, France

C. Goéminne : E. RobinDepartment of Cardiovascular reanimation, Lille UniversityHospital, Lille, France

A. VincentelliDepartment of Cardiovascular surgery, Lille University Hospital,Lille, France

S. Boulé (*)Pôle de Cardiologie, Hôpital Cardiologique, CHRU,50370 Lille, Francee-mail: stephane.boule@chru-lille.fr

J Interv Card Electrophysiol (2014) 39:177–184DOI 10.1007/s10840-013-9854-y

LVEF Left ventricular ejection fractionLVEDD Left ventricular end diastolic diameterNYHA New York Heart AssociationRV Right ventricularVA Ventricular tachyarrhythmia

1 Background

Left ventricular assist device (LVAD) support is beingincreasingly used in patients with advanced heart failure,either as a bridge to transplantation or as destination therapy[1, 2]. It has also become an important treatment option inend-stage heart failure patients, in whom the use of an LVADis associated with a marked improvement of both functionalstatus and quality of life [2–4]. In practice, most recipients ofan LVAD also carry an implantable cardioverter-defibrillator(ICD); however, little is known at present about the clinicalimplications of using these devices simultaneously in the samepatient. In the present study, we seek to describe the clinicaloutcomes for patients carrying both devices, focusing on theproblems that occur as a result of electromagnetic interferencebetween them, the impact of LVAD placement on rightventricular (RV) lead parameters, as well as the ventriculartachyarrhythmias (VAs) that occur post-LVAD.

2 Methods

2.1 Study population

All patients who underwent LVAD implantation with aconcomitant ICD at our institution between June 2007 andAugust 2012 were included in this study. In all cases, thespecific LVAD used was the Heart Mate II (Thoratec Corp,Pleasanton, CA, USA). Follow-up was started from the dateboth devices were first activated and was terminated on 1December 2012, unless the patient died or underwent hearttransplantation beforehand. All patients were in possession ofeither a single-chamber, a dual-chamber or a cardiacresynchronization therapy ICD (CRT-D).

In all cases, tachycardia detection was disabled at the timeof surgery to avoid the provocation of inappropriate shocks byelectrical interference during the operation. Arrhythmiadetection and pacing functions were subsequently reactivatedpost-surgery. In all cases, devices were programmed to detectventricular fibrillation (VF), defined as any ventricular activity≥240 bpm, and in all but one patient, they were programmedto recognise fast ventricular tachycardia (VT) and slow VT.For the majority of patients (21 out of 23), fast VTwas definedas ventricular activity in the range 200–240 bpm and slow VTwas set according to whether the ICD had originally been

implanted as a primary prevention measure, in someone whohad never suffered VA, or whether it had been implanted as asecondary prevention measure in someone who had. For theprimary prevention cases, slowVTwas defined as 150–200 or160–200 bpm, and for the secondary prevention cases thelower limit was set to be 10 bpm below the slowest episodeof VA, so that, for example, a patient who had sustained anepisode of VAwith ventricular rate of 140 bpm would have anICD installed with a slow VT range of 130–200 bpm.

In all cases, long detection times were programmed forboth VF and VT, to avoid unnecessary shocks during non-sustained episodes of VA. In addition, during episodes of VT,ICDs were programmed to discriminate supraventriculartachycardias and to trigger anti-tachycardia pacing as thefirst-line response, followed by shocks if this failed toterminate the arrhythmia.

2.2 Data collection

All patient data was collected retrospectively from medicalrecords kept by our institution, which included demographicinformation as well as clinical outcome for each patient. ICD-specific parameters were collated including lead impedance,intracardiac signal amplitude and capture threshold for the RVlead as well as for the right atrial and left ventricular leads inthe cases where ICDs with more than one lead had beendeployed. For each case, data from the most recent ICDinterrogation prior to LVAD implantation was included as wellas the values measured post-implantation. All interventionsinvolving any of the leads were also noted. Medicationregimens (β-blockers, amiodarone) were also recorded priorto LVAD implantation, during subsequent episodes of VA, aswell as at the end of follow-up for each patient.

All episodes of VA occurring after LVAD implantationwere classified as either VT or VF. For each patient, we notedthe total number of episodes of VA sustained as well as howlong after LVAD implantation the first episode occurred. Wealso noted how well each episode was tolerated by the patientand how each was terminated, whether by anti-tachycardiapacing or by shock. In this data, an electrical storm wasdefined as three or more separate episodes of VA (VT/VF)within a 24-h interval, resulting in the deployment of eitheranti-tachycardia pacing or shock therapy from the ICD.

2.3 End points of the study

The following issues were addressed:

(1) Electromagnetic interference between LVAD and ICDduring telemetry.

(2) Impact of LVAD implantation on RV lead parameters.(3) Occurrence of VA post-LVAD implantation.

178 J Interv Card Electrophysiol (2014) 39:177–184

2.4 Statistical analysis

Qualitative variables were expressed in terms of theirfrequency and percentage. Continuous variables wereexpressed in terms of their median and interquartile rangevalues. Non-parametric tests were used because of the samplesize (n <30). Wilcoxon matched pairs signed rank test wasused to compare pre-LVAD versus post-LVAD rightventricular lead parameters. For the comparison of patientswith or without post-LVAD ventricular arrhythmias,continuous variables were compared with the Mann–WhitneyU test and the qualitative variables with the χ2 test or Fischerexact test. Binary logistic regression modelling was used todevelop a predictive model for the occurrence of post-LVADventricular arrhythmias. Variables included in the multivariateanalysis were those reaching a significant level of p ≤0.2 onunivariate analysis. Results of multivariate analysis werereported as odds ratio (OR) with corresponding 95 %confidence interval and p value. An OR>1 indicates that thestudied factor is associated with a likelihood of occurrence ofpost-LVAD ventricular arrhythmia. Statistical analysis wasperformed using SPSS 15.0.1 (SPSS, Chicago, IL, USA). Avalue of p <0.05 was considered significant.

3 Results

3.1 Study population

Among the 26 consecutive patients who underwent HeartMate II implantation at our institution between June 2007and August 2012, two patients did not receive ICD therapyand were thus excluded from the analysis. One patient diedfrom cardiac tamponade the day after LVAD implantation andwas also excluded from our analysis due to a lack of usefulinformation provided by the case. The study populationtherefore consisted of 23 patients (17 males, 73 %), with amedian age of 50 years (the youngest being 36 years old andthe oldest being 67 years old). Median LVEF at the time ofLVAD implantation was 15 % [5–20] and baselinecharacteristics of the study population are presented inTable 1.

As regard to the clinical indication for LVAD support, themajority of patients (n =19) underwent implantation as abridge to transplantation and the remainder (n =4) wereimplanted with an LVAD as a destination therapy. Twopatients had to undergo LVAD replacement during follow-updue to device infection and the median duration of LVADsupport was 182 days [48–478]. Regarding the ICDs, 20patients already had ICDs at the time of LVAD implantationand 3 patients were implanted with these devices afterwards.The following brands were employed: Medtronic (n =9),

Saint-Jude Medical (n =5), Guidant/Boston Scientific (n =4),ELA/Sorin (n =3) and Biotronik (n =2).

Of the 23 patients concomitant carriers of an LVAD and anICD in this study, 12 died, 5 underwent heart transplantationand 6 outlived the study endpoint with ongoing LVAD support(longest duration >1,500 days).

3.2 Electromagnetic interference between LVAD and ICDduring telemetry

In four patients, electromagnetic interference preventedtelemetry being established between the ICD and itsprogrammer immediately following LVAD implantation. Asindicated in Table 2, the following devices were associatedwith this problem: Sorin Paradym CRT 8750 and Saint JudeMedical Atlas V-193, V-240, V-243. For the Sorin ParadymCRT 8750, telemetry could be re-established by simplychanging the position of the programmer head and for theSaint JudeMedical Atlas V-243, it was possible to re-establishtelemetry by placing a metal plate on the chest of the patient toshield the ICD device from interference generated by theLVAD. Both these scenarios are discussed later. For theremaining two cases (Saint Jude Medical Atlas V-193 andV-240), neither of these techniques were effective and theICDs had to be replaced with devices from othermanufacturers communicating at a different frequency withtheir programmer.

Table 1 Baseline characteristics of the study population

Variables Values

Age, years 50 [47–60]

Male gender 17 (74)

Ischemic cardiomyopathy 13 (56.5)

Heart failure duration, years 10 [2–14]

LVAD indication (% BTT) 19 (82.6)

NYHA functional class

NYHA III 4 (17.4)

NYHA IV 19 (82.6)

LVEF, % 15 [10–25]

LVEDD, mm 74.5 [66–81]

Cardiac resynchronization therapy 5 (21.7)

Ventricular arrhythmias pre-LVAD 12 (52.2)

Supraventricular arrhythmia pre-LVAD 8 (34.8)

Time between ICD and LVAD implantationa, days 460 [221–1,063]

Data are presented as number (percentage) of patients or median[interquartile range], as appropriate

BTT bridge to cardiac transplantation, ICD implantable cardioverter-defibrillator, LVAD left ventricular assist device, LVEDD left ventricularend diastolic diameter, LVEF left ventricular assist device, NYHA NewYork Heart Associationa Time between ICD and LVAD implantation is indicated for patients whounderwent LVAD placement after ICD implantation (n =21)

J Interv Card Electrophysiol (2014) 39:177–184 179

The following ICD models did not cause any problemsduring this study: Medtronic (Virtuoso, Insync Sentry, Gem7227, Entrust, Consulta, Secura), Saint-Jude Medical(Promote RF, Atlas V-168), Guidant/Boston Scientific(Vitality II, Teligen 100), ELA/Sorin (Ovatio CRT,Paradym VR) and Biotronik (Lumax 540). It isimportant to note that there were no instances of LVADinterference triggering inappropriate therapy from any ofthe ICDs during this study and that all problems weresimply related to establishing communication betweenICD and programmer.

3.3 Impact of LVAD implantation on RV lead parameters

Changes in RV lead parameters were observed after LVADimplantation and are reported in Fig. 1. In particular, RVsignal amplitude fell significantly from 12 mV [8.0–15.9] to9 mV [8.1–12.4] (p =0.04), RV impedance decreasedsignificantly from 520 ohms [483–592] to 456 ohms [407–549] (p <0.001) and there was a non-significant trend towardsan increased RV pacing threshold from 0.75 V [0.5–1.0] to0.95 V [0.74–1.03] (p =0.08) post-LVAD implantation. Noneof these changes had any clinical consequence, however, asthere were no instances of missed VA or cases where the ICDfailed to stimulate the ventricles during this study. There wasalso no evidence of RV lead dislodgment on radiologicalassessment during follow-up. In one case out of the fivepatients implanted with a CRT-D device, however, there wasdislodgement of the LV lead positioned in the epicardialsurface. No atrial lead dislodgement was seen post-LVADimplantation.

3.4 Occurrence of VA post-LVAD implantation

Among the 23 patients included in our study, 11 patients(47.8 %) experienced one or more episodes of VA followingLVAD placement. Of these, nine had had at least one previousepisode of VA and two experienced their first episode onlyafter LVAD implantation. All 11 patients experiencedepisodes of VT, 5 had episodes of VF and 7 experiencedelectrical storms. Eight patients received appropriate shocktherapy at some point and for the remaining three patientsanti-tachycardia pacing alone had been sufficient to terminateepisodes of VT. The median time from LVAD placement tothe first occurrence of a VA was 32 days [12–74] and themedian number of VA episodes post-LVAD implantation was4.5 [2–33].

Prior to LVAD implantation, 20 out of 23 patients (87 %)were on β-blockers with a median dose of 75 % [25–100] of the maximum dose. For the 11 patients whosuffered episodes of VA post-LVAD, only 4 were on β-blocker therapy, with doses ranging from 25 to 100 %of the maximal dose.

Typically, episodes of VA were tolerated well, even whenassociated with a 12-lead electrocardiogram that might havesuggested otherwise. For instance, the ECG in Fig. 2 shows aclear example of VF but was taken from an ambulatorypatient who had not lost consciousness. There was,however, one case of congestive heart failure and one caseof RV failure that resulted in patient deaths following anelectrical storm of VAs.

A univariate analysis comparing patient characteristicswith VA occurrence post-LVAD implantation is shown inTable 3. Multivariate analysis was also carried out and

Table 2 Electromagnetic interferences between implantable cardioverter-defibrillators and left ventricular assist devices observed in the studypopulation

Patients ICD manufacturer Model LVAD Type of interference Type of intervention

1 St Jude Medical Atlas V-193 Heart Mate II Loss of telemetry ICD replacement

2 St Jude Medical Atlas V-240 Heart Mate II Loss of telemetry ICD replacement

3 St Jude Medical Atlas V-243 Heart Mate II Loss of telemetry Metal shielding

4 Sorin Group Paradym CRT 8750 Heart Mate II Loss of telemetry Modification of ICD programmer head position

ICD implantable cardioverter-defibrillator, LVAD left ventricular assist device

Fig. 1 Comparison of rightventricular lead parameters beforeand after LVAD implantation

180 J Interv Card Electrophysiol (2014) 39:177–184

identified history of VA prior to LVAD implantation as astrong independent predictor of post-LVAD ventriculararrhythmias (OR 25 [2.9–213]; p =0.003). None of the othercovariates were found to be associated with post-LVAD VAoccurrence.

3.5 Inappropriate therapies received by LVAD patients

Three patients received inappropriate shock therapy whileunder LVAD support. In each case, a supraventriculararrhythmia was responsible for triggering the ICD. Onepatient received inappropriate shocks following themisdiagnosis of atrial fibrillation (as it was a “pseudo-regular”atrial fibrillation) and subsequently required anti-arrhythmicdrug therapy to terminate the episode. The other two patientsreceived shocks after developing atrial flutter and weresubsequently treated with catheter ablation of thecavotricuspid isthmus while under LVAD support withoutfurther complication. There were also no instances ofinterference from the LVAD triggering inappropriate therapyby any of the ICDs during this study.

4 Discussion

The main findings of the present study are that (1)electromagnetic interference from the LVAD can inhibittelemetry between certain ICD models and their programmerand hence necessitate replacement of the ICD; (2) LVADimplantation is associated with significant changes in RV leadparameters, but that these have no effect on clinical outcome;and (3) episodes of VA occur in about half the patientscarrying an LVAD, and that prior history of VA episodes is astrong predictor of VA occurrence post-LVAD implantation.

Although impaired telemetry between ICD andprogrammer has been reported previously for the Saint JudeMedical Atlas V-193 [21–24], to the best of our knowledge,this is the first time problems of this nature have been reportedfor the Saint JudeMedical Atlas V-240 and V-243 models andSorin Paradym CRT 8750 ICD. Loss of telemetry occurs forICDs that communicate in the frequency range 7–8 kHz asthis coincides with the Heart Mate II pulse width modulatorfrequency of 7 kHz, which ultimately interferes with ICD

Fig. 2 Twelve-leadelectrocardiogram of aspontaneous episode ofventricular fibrillation in anasymptomatic patient underLVAD support

Table 3 Comparison of patient’s characteristics according to theoccurrence of ventricular arrhythmias post-LVAD implantation.Univariate analysis

Characteristics Post-LVADventriculararrhythmias (n=12)a

No post-LVADventriculararrhythmias (n =12)a

pvalue

Age, years 47.5 [43–64] 52.5 [50–59] 0.25

Male gender 10 (83) 8 (66) 0.64

Heart failureduration, years

11 [6–17] 7 [2–12] 0.11

Ischemiccardiomyopat-hy

8 (66) 6 (50) 0.41

LVEDD, mm 77 [70–84] 70 [62–80] 0.22

LVEF, % 11 [6–17] 7 [2–12] 0.86

Atrial lead 7 (58) 2 (16) 0.09

Cardiacresynchronizati-on therapy

2 (16) 1 (0.1) 0.59

Secondaryprevention

8 (66) 3 (25) 0.04

Ventriculararrhythmiasbefore LVAD

10 (83) 3 (25) 0.001

Data are presented as number (percentage) of patients or median[interquartile range], as appropriate

LVAD left ventricular assist device, LVEDD left ventricular end diastolicdiameter, LVEF left ventricular assist device, NYHA New York HeartAssociationa A total of 25 LVAD were implanted in 23 patients. One of the twopatients who underwent two LVAD implantations was not included in theanalysis since her ICD was removed during the second procedure ofLVAD implantation (infection)

J Interv Card Electrophysiol (2014) 39:177–184 181

communication. This is the situation for all Sorin Group ICDs,which communicate in the frequency range 7–8 kHz, and forthe older models of ICD from Saint Jude Medical. For thenewer models from Saint Jude Medical, however, this is not aproblem as they communicate at a higher frequency (64 kHz),nor is it a problemwith the followingmanufacturers:Medtronic(175 kHz), Biotronik (32 kHz), Guidant (100 kHz)) [23, 25].

In this study, it was possible to overcome loss of telemetryin two out of four cases. In the case of the Sorin Paradym CRT8750 model, telemetry was re-established by altering theposition of the programmer head, following advice from SorinGroup product support. To do this, the patient was first placedin the supine position and the ICD programmer unitpositioned as far from the LVAD as possible. Then, theprogrammer head, on an outstretched cable, was advancedtowards the shoulder of the patient containing the ICD, untilthe active area was positioned next to the ICD connectorblock, where the ICD antenna was located.

In the case of the Saint Jude Medical Atlas V-243,telemetry was re-established by using a thick aluminium plate,designed by technical support at Saint Jude Medical that wasplaced on the chest of the patient between the LVAD and ICD,and very close to the ICD. Once again, the ICD programmingunit was placed as far as possible from the patient, at the headof the bed, and the metal plate effectively provided enoughelectrical shielding from the LVAD interference to allowtelemetry to be re-established of between ICD andprogrammer.

None of these strategies were effective in the other twocases (Saint Jude V-193 and V-240), which subsequentlyrequired re-implantation of new ICDs from a differentmanufacturer. Misiri recently highlighted how important it isfor surgeons implanting LVADs to be aware of this potentialconflict between LVADs and certain models of ICD, as it hasstrong clinical implications [5] and several strategies have alsobeen proposed to re-establish communication between ICDand programmer, although none are guaranteed to work in allcases [5–7].

The second finding of this study is that LVAD placement isassociated with significant changes in RV lead parameters.Here, we noted a decrease in both RV intracardiac amplitude,RV lead impedance and a trend towards an increase in pacingthreshold. These results are consistent with previous studiesand several hypotheses have been proposed to explain thesechanges [8, 9]. It is unclear, however, whether any of thesemodifications have any clinical impact. In a recentretrospective study [9], post-LVAD RV decrease inintracardiac signal amplitude led to lead revision in 13 % ofpatients, and ICD testing in 20 % of patients. In contrast,however, the results of the present study and the study byFoo et al. [8] suggest that the decrease in the R-waveamplitude have no clinical implications as neither leadrevision nor ICD testing were required, and there were no

episodes of VA undersensing during follow-up. Furtherstudies may therefore be required to address this point morethoroughly. Whether the change of RV lead parameters hasany ramification on clinical outcome or not, it is our opinionthat all patients should undergo ICD interrogation as soon aspossible after LVAD implantation and be notified beforehandof the potential need for lead revision should problems arise.

The third finding of this study is that episodes of VA arevery common in LVAD patients, and there is a strongassociation with a history of VA before LVAD implantation.In our population, half the patients experienced episodes ofeither VT or VF following LVAD implantation. Althoughthese results come from a small sample population, they arecomparable with those from previous studies, whereincidences of VA have been reported in the range 22–52 %among LVAD patients [10–13].

In our study, 11 out of 23 patients (48 %) receivedappropriate therapy from their ICD at some point and 3patients (12.5 %) received inappropriate therapy followingthe misdiagnosis of a supraventricular arrhythmia. Theseresults are consistent with previous studies on LVAD patientsin which 12–41 % received appropriate therapy from aconcomitant ICD and 4–25 % received inappropriate therapydespite short-term follow-up [10, 14–17, 24].

In this study, the median time to first occurrence of VApost-LVAD implantation was 32 days [12–74]. Previousstudies have indicated an increased risk of VT and VF in theearly post-operative period [4]. The Heart Mate II registry alsoindicates a 10-fold increase in the risk of VA occurrence in the30 days that follow LVAD implantation and various reasonshave been put forward to explain this [10, 13, 18]. In line withthe study by Refaat et al. [13] that demonstrated a linkbetween the suspension of β-blocker therapy post-LVADimplantation and increased risk of VA episodes, it is our beliefthat withholding β-blocker therapy is one of the keycontributing factors. In our study, 20 out of 23 patients(87 %) were on high dose β-blocker therapy prior to LVADimplantation, but by the end of follow-up, only 10 patients(43 %) were still receiving β-blocker therapy. Two reasonsmight explain this reduction. First, patients often requireinotropic support following LVAD implantation, so β-blocker therapy is withdrawn in the postoperative period, atleast temporarily. Second, a significant number of our patientsdeveloped RV failure under LVAD support, so clinicians mayhave been reluctant to administer β-blocker therapy in linewith common practice. No matter the cause, however, it is ourstrong belief that β-blocker therapy should be given to LVADpatients whenever possible and should be titrated to themaximum tolerable dose. Further studies are, however,required to further assess the role of anti-arrhythmic therapyin VA occurrence after LVAD implantation.

It is of general interest, perhaps, that episodes of VA in ourstudy population were asymptomatic in the majority of cases

182 J Interv Card Electrophysiol (2014) 39:177–184

and that one patient even presented in sustained VF,apparently asymptomatic, and only required non -urgentelectrical cardioversion to terminate the arrhythmia. This pointhas strong implications for ICD programming in LVADpatients because they are far less likely to lose consciousnessduring episodes of VA and therefore require very longdetection periods and high detection thresholds to avoiddelivery of painful shocks while still in a state ofconsciousness [15, 24]. It is our opinion that a high numberof anti-tachycardia pacing attempts should also beprogrammed in an attempt to avoid the need for shock therapy.That having been said, prolonged episodes of arrhythmia maylead to RV failure, thromboembolic events and worsening ofrenal and liver functions, and one patient in this study didindeed develop RV failure as a result of electrical storms. ICDprogramming in LVAD patients therefore remains achallenging task and, for the moment, should probably beindividualized for each patient.

On multivariate analysis in our study, a history of VA priorto LVAD placement was identified as a strong independentpredictor of VA occurrence post-LVAD implantation. Theseresults are consistent with previous studies on the subject andclinicians should be aware of this associated risk [14, 15,18–20]. It is also noted that even though episodes of VA arewell tolerated, they are associated with a worse prognosis inLVAD recipients [1, 3, 14, 26].

Finally, there is still a certain amount controversy overwhether LVAD patients who do not already carry an ICD atthe time of LVAD implantation should receive one on thegrounds of primary prevention. In the present study, threepatients fell into this category and were fitted with ICDs;however, none experienced any episodes of VA duringfollow-up, and in one case lead dysfunction occurred thatled to complete deactivation of the ICD. In contrast, however,there were eight patients already fitted with primaryprevention ICDs prior to LVAD implantation, who equallyhad never had a previous episode of VA, and of these, twowent on to sustain episodes of VA post-LVAD implantation. Inother studies of this issue, conflicting data exist [15, 20]. In thestudy by Oswald et al. [15], for a cohort of 61 patients withLVAD and ICD (75 % with primary prevention ICDs), a highprevalence of VAs was noted in LVAD patients, even in thosewith primary prevention ICDs. They concluded that allpatients who do not already have an ICD at the time of LVADimplantation should be implanted with one on the grounds ofprimary prevention of sudden death, and for all those whoalready carry an ICD, this should be maintained duringfollow-up. Conversely, Garan et al. [20] drew the oppositeconclusion in their recent study on a cohort of 94 LVADpatients (77 with an ICD and 17 without ICD), where theyfound that the absence of preoperative VA conferred a verylow risk of subsequently developing VAs post-LVADimplantation (4.0 vs. 45 %; p <0.001). In addition, none of

the LVAD patients without an ICD died during follow-up(mean follow-up without an ICD being 12.7 months). Theresults of our study, therefore, support the conclusion of Garanet al., although our sample size does not allow us to draw anysignificant conclusions. These results together suggest thatpatients who do not have a previous history of preoperativeVA may not require a primary prevention ICD, mostespecially when one takes into account the additional risksof ICD implantation, such as shocks while still conscious andinfections.

4.1 Study limitations

The sample population in this study is relatively small, duemainly to the low numbers of patients receiving LVADtherapy each year. Nonetheless, all patients implanted withan LVAD who were concomitant carriers of ICD at ourinstitution were included. We cannot rule out the impact ofunmeasured variables on the development of post-implantation VA. Finally, given that all implanted LVADwerecontinuous flow pumps (Heart Mate II), our results cannot beextended to other types of assist devices.

5 Conclusions

The concomitant use of LVAD and ICD therapies is a feasible,safe and effective treatment option in end-stage heart failurepatients. LVAD implantation causes loss of telemetry withspecific models of ICD, which may necessitate replacementwith another model of ICD. Systematic post-operative ICDinterrogation should be performed, as LVAD implantation isassociated with substantial changes in RV lead parameters.VAs are very common following LVAD implantation,occurring in about half of our patients; however, they aregenerally well tolerated, and their occurrence is stronglypredicted by a history of VA prior to LVAD implantation. Itis still not clear whether absence of an ICD at the time ofLVAD implantation is reason enough not to implant an ICDdevice as a primary prevention measure. The increasedtoleration of LVAD patients to episodes of VA means thatICD programming must be carried out with care to ensurepatients are not subjected to unnecessary shocks while still in astate of consciousness and, in general, ICD programming inthis group of patients still remains a challenging task.

Acknowledgments The authors thank Nicolas Iscolo (Sorin Group)and Xavier Le Groumellec (Saint Jude Medical), for their help withtechnical information.

Conflict of interest None

J Interv Card Electrophysiol (2014) 39:177–184 183

References

1. Rose, E. A., Gelijns, A. C., Moskowitz, A. J., Heitjan, D. F.,Stevenson, L. W., Dembitsky, W., et al. (2001). Long-term use of aleft ventricular assist device for end-stage heart failure. The NewEngland Journal of Medicine, 345(20), 1435–1443.

2. Slaughter, M. S., Rogers, J. G., Milano, C. A., Russell, S. D., Conte,J. V., Feldman, D., et al. (2009). Advanced heart failure treated withcontinuous-flow left ventricular assist device. The New EnglandJournal of Medicine, 361(23), 2241–2251.

3. Kirklin, J. K., Naftel, D. C., Kormos, R. L., Stevenson, L. W., Pagani,F. D., Miller, M. A., et al. (2010). Second INTERMACS annual report:more than 1,000 primary left ventricular assist device implants. TheJournal of Heart and Lung Transplantation, 29(1), 1–10.

4. Pagani, F. D., Miller, L.W., Russell, S. D., Aaronson, K. D., John, R.,Boyle, A. J., et al. (2009). Extended mechanical circulatory supportwith a continuous-flow rotary left ventricular assist device. Journal ofthe American College of Cardiology, 54(4), 312–321.

5. Biviano, A., Mancini, D., Naka, Y., Arellano, J., & Garan, H. (2009).Overcoming electromagnetic interference by LVADs on ICDfunction by shielding the ICD programmer wand and extensioncable. Pacing and Clinical Electrophysiology, 32(7), 945–948.

6. Baran, D. A., Martin, T., Blicharz, D., Carr, C., Zucker, M. J.,McBride, L., et al. (2009). Simple method to eliminate interferencebetween ventricular assist devices and cardiac rhythm devices.Pacing and Clinical Electrophysiology, 32(9), 1142–1145.

7. Jacob, S., Cherian, P. K., Ghumman, W. S., & Das, M. K. (2010).“Pseudo” Faraday cage: a solution for telemetry link interactionbetween a left ventricular assist device and an implantablecardioverter defibrillator. Journal of Interventional CardiacElectrophysiology, 28(3), 221–225.

8. Foo, D., Walker, B. D., Kuchar, D. L., Thorburn, C. W., Tay, A.,Hayward, C. S., et al. (2009). Left ventricular mechanical assistdevices and cardiac device interactions: an observational case series.Pacing and Clinical Electrophysiology, 32(7), 879–887.

9. Ambardekar, A. V., Lowery, C. M., Allen, L. A., Cannon, A. P.,Cleveland, J. C., Jr., Lindenfeld, J., et al. (2010). Effect of left ventricularassist device placement on preexisting implantable cardioverter-defibrillator leads. Journal of Cardiac Failure, 16(4), 327–331.

10. Ziv, O., Dizon, J., Thosani, A., Naka, Y., Magnano, A. R., & Garan,H. (2005). Effects of left ventricular assist device therapy onventricular arrhythmias. Journal of the American College ofCardiology, 45(9), 1428–1434.

11. Andersen, M., Videbaek, R., Boesgaard, S., Sander, K., Hansen, P.B., & Gustafsson, F. (2009). Incidence of ventricular arrhythmias inpatients on long-term support with a continuous-flow assist device(HeartMate II). The Journal of Heart and Lung Transplantation,28(7), 733–735.

12. Cesario, D. A., Saxon, L. A., Cao, M. K., Bowdish, M., &Cunningham, M. (2011). Ventricular tachycardia in the era ofventricular assist devices. Journal of CardiovascularElectrophysiology, 22(3), 359–363.

13. Refaat, M., Chemaly, E., Lebeche, D., Gwathmey, J. K., & Hajjar, R.J. (2008). Ventricular arrhythmias after left ventricular assist deviceimplantation. Pacing and Clinical Electrophysiology, 31(10), 1246–1252.

14. Brenyo, A., Rao, M., Koneru, S., Hallinan, W., Shah, S., Massey, H.T., et al. (2012). Risk of mortality for ventricular arrhythmia inambulatory LVAD patients. Journal of CardiovascularElectrophysiology, 23(5), 515–520.

15. Oswald, H., Schultz-Wildelau, C., Gardiwal, A., Lusebrink, U.,Konig, T., Meyer, A., et al. (2010). Implantable defibrillator therapyfor ventricular tachyarrhythmia in left ventricular assist devicepatients. European Journal of Heart Failure, 12(6), 593–599.

16. Ambardekar, A. V., Allen, L. A., Lindenfeld, J., Lowery, C. M.,Cannon, A. P., Cleveland, J. C., Jr., et al. (2010). Implantablecardioverter-defibrillator shocks in patients with a left ventricularassist device. The Journal of Heart and Lung Transplantation,29(7), 771–776.

17. Refaat, M.M., Tanaka, T., Kormos, R. L., McNamara, D., Teuteberg,J., Winowich, S., et al. (2012). Survival benefit of implantablecardioverter-defibrillators in left ventricular assist device-supportedheart failure patients. Journal of Cardiac Failure, 18(2), 140–145.

18. Cantillon, D. J., Tarakji, K. G., Kumbhani, D. J., Smedira, N. G.,Starling, R. C., & Wilkoff, B. L. (2010). Improved survival amongventricular assist device recipients with a concomitant implantablecardioverter-defibrillator. Heart Rhythm, 7(4), 466–471.

19. Raasch, H., Jensen, B. C., Chang, P. P., Mounsey, J. P., Gehi, A. K.,Chung, E. H., et al. (2012). Epidemiology, management, andoutcomes of sustained ventricular arrhythmias after continuous-flow left ventricular assist device implantation. American HeartJournal, 164(3), 373–378.

20. Garan, A. R., Yuzefpolskaya, M., Colombo, P. C., Morrow, J. P., Te-Frey, R., Dano, D., et al. (2013). Ventricular arrhythmias andimplantable cardioverter-defibrillator therapy in patients withcontinuous-flow left ventricular assist devices: need for primaryprevention? Journal of the American College of Cardiology,61(25), 2542–2550.

21. Matthews, J. C., Betley, D., Morady, F., & Pelosi, F., Jr. (2007).Adverse interaction between a left ventricular assist device and animplantable cardioverter defibrillator. Journal of CardiovascularElectrophysiology, 18(10), 1107–1108.

22. Oswald, H., Klein, G., Struber, M., & Gardiwal, A. (2009).Implantable defibrillator with left ventricular assist devicecompatibility. Interactive Cardiovascular and Thoracic Surgery,8(5), 579–580.

23. Mehta, R., Doshi, A. A., Hasan, A. K., Love, C. J., Pizzuto, M., Sai-Sudhakar, C., et al. (2008). Device interactions in patients withadvanced cardiomyopathy. Journal of the American College ofCardiology, 51(16), 1613–1614.

24. Kuhne, M., Sakumura, M., Reich, S. S., Sarrazin, J. F., Wells, D.,Chalfoun, N., et al. (2010). Simultaneous use of implantablecardioverter-defibrillators and left ventricular assist devices inpatients with severe heart failure. The American Journal ofCardiology, 105(3), 378–382.

25. Netzler, P. C., Vasuki, N., Peura, J. L., & Gold, M. R. (2012).Interactions between a left ventricular assist device and implantablecardioverter-defibrillator. Pacing and Clinical Electrophysiology,35(9), e272–273.

26. Bedi, M., Kormos, R., Winowich, S., McNamara, D.M., Mathier, M.A., & Murali, S. (2007). Ventricular arrhythmias during leftventricular assist device support. The American Journal ofCardiology, 99(8), 1151–1153.

184 J Interv Card Electrophysiol (2014) 39:177–184