mpdstscecpcIod

gsS

4

0d

Simultaneous Use of Implantable Cardioverter-Defibrillators andLeft Ventricular Assist Devices in Patients With Severe

Heart Failure

Michael Kühne, MDa,*, Michaela Sakumuraa, Stephen Scott Reich, MDa,Jean-Francois Sarrazin, MDa, Darryl Wells, MDa, Nagib Chalfoun, MDa, Thomas Crawford, MDa,

Warangkna Boonyapisit, MDa, Laura Horwood, NPa, Aman Chugh, MDa, Eric Good, DOa,Krit Jongnarangsin, MDa, Frank Bogun, MDa, Hakan Oral, MDa, Fred Morady, MDa,

Francis Pagani, MD, PhDb, and Frank Pelosi, Jr., MDa

More patients who are receiving therapy with a left ventricular assist device (LVAD) alsohave an implantable cardioverter-defibrillator (ICD). The aim of the present study was todescribe the outcomes and device interactions of simultaneous therapy with an ICD and aLVAD. We evaluated 76 patients with class IV heart failure (age 52 � 12 years, leftventricular ejection fraction 0.13 � 0.05%, 88% men, 61% nonischemic cardiomyopathy)with both an ICD and a LVAD. The median follow-up with both devices was 156 days. ALVAD with a pulsatile and continuous flow pump was used in 53 (70%) and 23 (30%)patients, respectively. Of the 76 patients, 12 (16%) received a total of 54 ICD therapies. Ofthe ICD therapies, 88% were appropriate. Of the 76 patients, 55 (72%) underwent hearttransplantation a median of 146 days after LVAD implantation. Twelve patients (16%)died during simultaneous ICD and LVAD therapy. Interactions between the LVAD andICD occurred in 2 patients (2.7%) with continuous flow pumps (HeartMate II). In bothcases, telemetry failure occurred after LVAD implantation with 2 different models of ICDsfrom the same manufacturer. No ICD therapies occurred because of device-related inter-actions. In conclusion, simultaneous ICD and LVAD therapy in patients with severecongestive heart failure is safe and clinically feasible. Interactions between the devices areuncommon and appear limited to specific models of ICDs. © 2010 Elsevier Inc. All rights

reserved. (Am J Cardiol 2010;105:378–382)

M

MvsuDtTcftL

tDtattpafivr

Nonpharmacologic advances have emerged to treat theost serious sequelae of congestive heart failure: the im-

lantable cardioverter-defibrillator (ICD) to prevent suddeneath from ventricular arrhythmias and left ventricular as-ist devices (LVADs) for advanced pump failure.1,2 Duringhe past 25 years, the ICD has become the mainstay ofudden cardiac death prevention for patients with advancedongestive heart failure.3 More recently, LVADs havemerged as an important treatment option for advancedongestive heart failure, both as a bridge to cardiac trans-lantation and as destination therapy.4 It is becoming in-reasingly common for patients with LVADs to also haveCDs. The objective of the present study was to evaluate theutcomes of patients with both LVADs and ICDs and toetermine the frequency of device-related interactions.

aDivision of Cardiovascular Medicine, and bSection of Cardiac Sur-ery, University of Michigan Health System, Ann Arbor, Michigan. Manu-cript received July 28, 2009; revised manuscript received and acceptedeptember 20, 2009.

*Corresponding author: Tel: (�41) 61-265-2525; fax: (�41) 61-265-598.

fE-mail address: kuehnem@uhbs.ch (M. Kuhne).

002-9149/10/$ – see front matter © 2010 Elsevier Inc. All rights reserved.oi:10.1016/j.amjcard.2009.09.044

ethods

The present study was conducted at the University ofichigan Medical Center (Ann Arbor, Michigan). The Uni-

ersity of Michigan institutional review board approved thetudy. The databases maintained by our institution weresed to identify eligible patients from October 1996 toecember 2007. All patients who had undergone LVAD

herapy with a concomitant ICD were analyzed (Figure 1).he data collected and analyzed included demographic andlinical variables, including New York Heart Associationunctional class, underlying heart failure etiology and dura-ion, and device-specific information on the ICDs andVADs.

The beginning of follow-up was the first day of simul-aneous use of both devices. The last day of follow-up wasecember 31, 2007, the date of death, or the date of heart

ransplantation, whichever came first. The outcome vari-bles analyzed were survival, ventricular arrhythmias, ICDherapies, and device-related complications and interac-ions. The number of hospitalizations and duration of hos-italization during simultaneous ICD-LVAD therapy werelso analyzed. Sustained ventricular arrhythmias were de-ned as those that lasted �30 seconds or required cardio-ersion or defibrillation. All ventricular arrhythmias occur-ing after surgery were included. The initial hospitalization

or LVAD placement (or ICD implantation if after LVAD

www.AJConline.org

if

d(utv�w

Is

R

wt

T21aTdt

mt(

peCvTp

uInp

tinc

Fippaa

TB

V

AMBBE

NH

LBH

TM

I

M

BBS

TM

L

P

C

379Heart Failure/Simultaneous Use of ICD and LVAD

mplantation) was included for analysis, and hospitalizationor orthotopic heart transplantation was excluded.

Continuous variables are presented as the mean � stan-ard deviation or as the median and interquartile rangeIQR) in the case of a nongaussian distribution. For contin-ous variables, comparisons were made using Student’s test or the Mann-Whitney U test, as appropriate. Discreteariables were compared using Fisher’s exact test. p Values0.05 were considered statistically significant. Calculations

igure 1. Posteroanterior chest radiograph of 33-year-old male with non-schemic dilated cardiomyopathy and class IV heart failure requiring im-lantation of a Heartmate II LVAD. Blood enters the continuous flowump (arrow) from the left ventricular apex and is pumped into thescending aorta. The dual-chamber ICD with a lead in the right ventricularpex and in the right atrium is also shown.

able 1aseline characteristics (n � 76)

ariable Value

ge (years) 52 � 12en 67 (88%)ody mass index (kg/m2) 27 � 6ody surface area (m2) 2.1 � 0.4tiology of cardiomyopathyIschemic 30 (39%)Nonischemic 46 (61%)ew York Heart Association functional class IV 76 (100%)eart failure duration (months)Mean � SD 70 � 49Range 3–210eft ventricular ejection fraction (%) 13 � 5iventricular pacemaker 25 (33%)istory of documented arrhythmiaSustained ventricular tachycardia or ventricular fibrillation 20 (26%)Nonsustained ventricular tachycardia 26 (34%)Atrial fibrillation 11 (14%)

ere made using StatView software, version 5.0.1 (SAS m

nstitute, Cary, North Carolina) and GraphPad Prism, ver-ion 5.0a (GraphPad Software, La Jolla, California).

esults

The data from 76 patients with both an ICD and LVADere analyzed. The demographic data and clinical charac-

eristics are summarized in Table 1.Follow-up data were available for all 76 patients (100%).

he median follow-up duration was 156 days (IQR 71 to92). Of the 76 patients with an ICD implanted from July989 to May 2006, 57 (75%) received a dual-chamber ICDnd 25 (33%) received an ICD with biventricular pacing.he ICD programming in the study population was notifferent from that for standard patients. The ICD manufac-urers and models are listed in Table 2.

Of the 76 patients, 71 (93%) received LVAD therapy aedian of 591 days (IQR 183 to 946) after ICD implanta-

ion, and 5 (7%) received their ICD a median of 14 daysIQR 8 to 25) after LVAD placement.

Table 3 lists the types of LVADs implanted. Of the 76atients, 53 (70%) received a pulsatile flow LVAD. Differ-nt versions of the HeartMate LVAD (Thoratec, Pleasanton,alifornia) were used: the Vented Electric, its enhancedersion, the XVE, and the Implantable Pneumatic version.hese 3, as well as the Thoratec IVAD—all pulsatile flowumps—were used.

Of the 76 patients, 23 (30%) received LVAD therapysing axial continuous flow pumps, such as the HeartMateI (Thoratec) and the Micromed DeBakey (MicroMed Tech-ology, Houston, Texas). The first HeartMate II was im-lanted in August 2001.

Of the 76 patients, 55 (72%) underwent heart transplan-ation a median of 146 days (IQR 70 to 240) after LVADmplantation. Twelve patients (16%) died during simulta-eous ICD and LVAD therapy. In these patients, circulationould be maintained, and the LVAD was turned off for

able 2odels of implantable cardioverter-defibrillators (n � 76)

CD Patients (n)

edtronic (Gem, Micro Jewel, InSync, Marquis,Entrust, Virtuoso)

28 (37%)

oston Scientific/Guidant/CPI (Contak, Ventak, Vitality) 42 (55%)iotronik (Belos, Tachos) 3 (4%)t. Jude Medical/Ventritex (Atlas, Epic, Contour MD) 3 (4%)

able 3odels of left ventricular assist devices (n � 76)

VAD Model Patients (n)

ulsatile flow left ventricular assist device 53 (70%)HeartMateVE 15 (20%)HeartMate XVE 34 (45%)HeartMate IP-1000 1 (1%)IVAD Thoratec 3 (4%)ontinuous flow left ventricular assist device 23 (30%)HeartMate II 22 (29%)Micromed DeBakey 1 (1%)

edical reasons unrelated to pump failure, ventricular fi-

brdoht

ietclp((h(afl

wp(os

bu(o

t(trbrt0(ar

pBuCnVdpcttptamLittnotul

D

bletTdf

tirpII

TNid

R

THII

LSAVG

RAFACCCHAO

4

6d(

itt

380 The American Journal of Cardiology (www.AJConline.org)

rillation, or a mechanical problem. The cause of death wasenal failure in 4, multiorgan failure in 2, anoxic brainamage after LVAD implantation in 2, intracerebral hem-rrhage in 1, ischemic stroke in 1, major gastrointestinalemorrhage in 1, and death related to toxic megacolon dueo Clostridium difficile in 1 patient.

The 76 patients in the series had a median of 2 hospital-zations during the follow-up period. No significant differ-nce was found in the number of hospitalizations betweenhe patients with pulsatile flow LVADs and those withontinuous flow LVADs (p � 0.45). The overall medianength of stay per patient was 35 days (IQR 19 to 79) inatients with pulsatile flow pumps compared to 27 daysIQR 17 to 59) in patients with continuous flow pumpsp � 0.28). The median percentage of days spent in theospital compared to the follow-up duration was 34%IQR 14% to 92%) in patients with pulsatile flow pumpsnd 16% (IQR 14% to 27%) in patients with continuousow pumps (p � 0.02).

Of the 53 patients with pulsatile flow LVADs, 29 (55%)ere rehospitalized after the initial LVAD implantation com-ared to 14 (61%) of 23 in the continuous flow LVAD groupp � 0.8). Of a total of 132 rehospitalizations, 40 (30%)ccurred because of LVAD- or ICD-related problems. Table 4

able 4umber and causes of rehospitalization in patients with simultaneous

mplantable cardioverter-defibrillator (ICD) and left ventricular assistevice (LVAD) therapy

eason for Rehospitalization Patients (n)

otal 132 (100%)eart failure exacerbations 10 (7.6%)

mplantable cardioverter-defibrillator therapies 16 (12.1%)mplantable cardioverter-defibrillator–related (without

implantable cardioverter-defibrillator therapies)*6 (4.5%)

VAD related† 18 (13.6%)yncope/falls 8 (6.1%)trial fibrillation 2 (1.5%)entricular tachycardia/fibrillation 4 (3.0%)astrointestinal symptoms (bleeding, diarrhea, abdominalpain)

11 (8.3%)

enal failure 5 (3.8%)nemia 6 (4.5%)ever of unknown origin 6 (4.5%)ltered mental status 4 (3.0%)erebrovascular accident 2 (1.5%)hest pain 2 (1.5%)ellulitis 2 (1.5%)ypoglycemia/hyperglycemia 2 (1.5%)rteriovenous fistula 3 (2.3%)ther‡ 25 (18.9%)

* ICD-related rehospitalizations were due to generator replacement (n �) and pocket hematoma (n � 2).

† LVAD-related rehospitalizations were due to driveline infection (n �), replacement for end of life (n � 5), controller malfunction (n � 2),evice malfunction (n � 1), device infection (n � 1), driveline dehiscencen � 1), device-related bowel obstruction (n � 1), LVAD alarms (n � 1).

‡ Other hospitalizations were due to infections (pneumonia, urinary tractnfections), admission for possible heart transplantation, heart catheteriza-ion, seizure, atrial fibrillation, radiofrequency ablation of ventricularachycardia, clot in peripherally inserted central catheter.

ummarizes the number and causes for rehospitalizations. l

Ventricular arrhythmias requiring cardioversion or defi-rillation occurred in 22 (29%) of the 76 patients. Ventric-lar tachycardia or ventricular fibrillation occurred in 1325%) of 53 patients with pulsatile flow pumps and 9 (39%)f 23 patients with continuous flow LVADs (p � 0.27).

Of the 76 patients, 12 (16%) received a total of 54 ICDherapies during the follow-up period. Of these therapies, 4380%) were appropriate for ventricular tachycardia or ven-ricular fibrillation in 9 patients (12%). Three patients (4%)eceived inappropriate ICD therapies because of atrial fi-rillation. No ICD therapies occurred owing to device-elated interactions. No significant difference was found inhe ICD therapies between the 2 types of LVADs (p �.49). Of the 53 patients with pulsatile flow LVADs, 713%) received 27 ICD therapies (78% were appropriate),nd 5 (22%) of 23 patients with continuous flow LVADseceived 27 ICD therapies (81% were appropriate).

Interactions between the LVAD and ICD occurred in 2atients (2.7%) with continuous flow pumps (HeartMate II).oth cases occurred with 2 different models of ICDs man-factured by St. Jude Medical (St. Jude Medical, Sunnyvale,alifornia). In the first case, telemetry communication couldot be established between a St. Jude Atlas VR model-193 ICD and the ICD programmer after LVAD implantespite several maneuvers, including adjustments in therogrammer head position and distance from the ICD. Suc-essful telemetry was established only after replacement ofhe ICD generator with a device from a different manufac-urer (Medtronic, Minneapolis, Minnesota). This case has beenreviously reported.5 In the second case of device interaction,elemetry failure occurred with a St. Jude Epic � VR V-196fter LVAD implantation. In addition to the previouslyentioned maneuvers, metal shielding between the ICD andVAD failed to establish telemetry communication. The

nteraction occurred because the pulse width modulator ofhe Heartmate II LVAD operates at a similar frequency ashe telemetry operating frequency of the ICD; thus, a tech-ical option involving bypassing the pulse width modulatorf the LVAD was considered but was not attempted becausehis had not been tested in humans. The ICD generator wasltimately replaced, and successful telemetry was estab-ished with the new device (Medtronic).

iscussion

From our data, we have concluded that ICD-LVAD com-ination therapy is safe and clinically feasible. This is theargest published report describing a patient population withnd-stage heart failure in which 2 potentially life-savingherapies (ICDs and LVADs) were used simultaneously.he patients in this population were relatively young, pre-ominantly male, and had very severe left ventricular dys-unction due to ischemic or nonischemic cardiomyopathy.

No ICD therapies occurred owing to device-related in-eractions. In a smaller study, 1 case of electromagneticnterference during charging of the LVAD battery packesulting in an inappropriate shock was reported.6 In ouropulation, 2 patients with LVAD-ICD therapy requiredCD replacement because of an inability to interrogate theCD after LVAD placement. However, this interaction was

imited to an older generation of ICDs from a single man-

uLttTrorddffpptg(thIfo(BempLLnmI

sImd(LddvfLradgtalvL(bciac

s

tatSpwofifat

L1hgoatin

hicfldLaip

tawatrwdm

381Heart Failure/Simultaneous Use of ICD and LVAD

facturer (St. Jude Medical) used with a continuous flowVAD, the HeartMate II. The interaction occurred because

hese ICD models use an 8-kHz operating frequency, near tohat of the Heartmate II pulse width modulator (7.2 kHz).he pulse width modulator of the Heartmate II is used to

educe the 12-volt battery voltage (needed for the controllerf the LVAD) to power the motor of the LVAD, whichequires less voltage. In both patients, the interaction wasiscovered immediately after LVAD placement and reme-ied without adverse sequelae. Newer generation ICDsrom St. Jude Medical operate at much greater telemetryrequencies (64 kHz), thereby significantly reducing theotential for this type of interaction. However, because therogrammer (Merlin, St. Jude Medical) uses a short 8-kHzest pulse or “handshake” to determine whether an oldereneration (8-kHz operating frequency) or newer generation64-kHz operating frequency) device is interrogated, a po-ential for interaction might still exist.7 A similar interactionas been described with the ELA (Sorin Group, Milan,taly) devices, which use a 16-kHz telemetry operatingrequency.8,9 No interactions were seen with ICDs fromther manufacturers, which operate at higher frequenciesBiotronik, Berlin, Germany, 32 kHz; Medtronic, 175 kHz;oston Scientific, Boston, Massachusetts, 100 kHz). Nev-rtheless, we recommend that all ICDs be interrogated im-ediately after LVAD implantation. Metal shielding, reim-

lanting the device at a different site (contralateral to theVAD), and bypassing the pulse width modulator of theVAD have been reported as possible remedies but mightot be feasible for reliable interrogation.8,10,11 We recom-end that the ICD generator be replaced when this LVAD-

CD interaction is suspected.Ventricular arrhythmias occurred in 29% of our patients,

imilar to the incidence in previous reports of the HeartmateI LVAD, in which 24% of patients had ventricular arrhyth-ias during follow-up, with most occurring in the first 30

ays after LVAD implantation.12 In another study, 1252%) of 23 patients had ventricular arrhythmias afterVAD implantation. However, these arrhythmias occurreduring a much longer mean duration of support (341ays).13 A possible mechanism for the early postoperativeentricular arrhythmias includes contraction-excitationeedback after acute myocardial unloading on initiation ofVAD support. It has been shown that LVAD placement

esults not only in important hemodynamic changes, butlso in significant electrophysiologic changes. Although theecrease in myocardial tension might alleviate arrhythmo-enic mechanisms in the long term, the acute changes afterhe initiation of LVAD support might transiently promoterrhythmias by a nonuniform effect on the course of repo-arization and excitability.14–17 There appeared to be moreentricular tachycardia in patients with a continuous flowVAD (39%) than in patients with a pulsatile flow LVAD

25%). This difference was not significant but might haveeen underestimated because of the sample size. It is con-eivable that pulsatile and continuous flow LVAD therapys associated with a different type of myocardial unloadingnd also in differences in the resulting electrophysiologichanges.

ICD therapies occurred in 16% of patients, with no

ignificant difference between patients with pulsatile and

hose with continuous flow LVADs. With a follow-up ofpproximately 6 months, this was significantly greater thanhat in primary prevention ICD populations such as theudden Cardiac Death in Heart Failure Trial (SCD-HeFT)opulation.18 However, our population consisted of patientsith heart failure in its most advanced stages, and about 1/4f the patients had a history of ventricular tachycardia orbrillation. Inappropriate ICD therapies were relatively in-requent and occurred in only 4% of our patients. Theyccounted for 20% of all ICD therapies delivered, similar tohat of other ICD populations.19

Of the 76 patients, 9 (12%) were alive with simultaneousVAD and ICD therapy at the end of the follow-up period,2 patients (16%) had died during follow-up, and 55 (72%)ad undergone heart transplantation. This was a muchreater survival and transplant rate compared to a 10-year-ld analysis of 10 patients receiving simultaneous LVADnd ICD therapy, in which 30% underwent heart transplan-ation and 50% died.20 This could be explained by themprovements in treating patients with heart failure, tech-ical advances in LVADs, and organ availability.

Patients with an ICD and LVAD had approximately 2ospitalizations during follow-up. No difference was foundn the number of hospitalizations between the pulsatile andontinuous flow pumps; however, patients with continuousow pumps spent a smaller percentage of the follow-upuration in the hospital. However, most continuous flowVADs were implanted during the later stages of the study,nd this difference could be explained by more experiencen treating patients with LVADs in the continuous flowump group.

The question of whether the patients truly benefited fromhe ICD in the setting of LVAD therapy could not benswered by the present analysis because no control groupas included. The comparisons between continuous flow

nd pulsatile flow LVADs should be interpreted with cau-ion, because the present study was neither prospective norandomized. Furthermore, most continuous flow LVADsere implanted during the second part of the study, and theifferences could be explained by changes in heart failureanagement over time.

1. Adams KFJ, Zannad F. Clinical definition and epidemiology of ad-vanced heart failure. Am Heart J 1998;135:S204–S215.

2. Narang R, Cleland JG, Erhardt L, Ball SG, Coats AJ, Cowley AJ,Dargie HJ, Hall AS, Hampton JR, Poole-Wilson PA. Mode of death inchronic heart failure: a request and proposition for more accurateclassification. Eur Heart J 1996;17:1390–1403.

3. Mirowski M, Reid PR, Mower MM, Watkins L, Gott VL, Schauble JF,Langer A, Heilman MS, Kolenik SA, Fischell RE, Weisfeldt ML.Termination of malignant ventricular arrhythmias with an implantedautomatic defibrillator in human beings. N Engl J Med 1980;303:322–324.

4. Goldstein DJ, Oz MC, Rose EA. Implantable left ventricular assistdevices. N Engl J Med 1998;339:1522–1533.

5. Matthews JC, Betley D, Morady F, Pelosi FJ. Adverse interactionbetween a left ventricular assist device and an implantable cardioverterdefibrillator. J Cardiovasc Electrophysiol 2007;18:1107–1108.

6. Foo D, Walker BD, Kuchar DL, Thorburn CW, Tay A, Hayward CS,Macdonald P, Keogh A, Kotlyar E, Spratt P, Jansz P. Left ventricularmechanical assist devices and cardiac device interactions: an observa-tional case series. Pacing Clin Electrophysiol 2009;32:879–887.

7. Mehta R, Doshi AA, Hasan AK, Love CJ, Pizzuto M, Sai-Sudhakar C,

Chan DP. Device interactions in patients with advanced cardiomyop-athy. J Am Coll Cardiol 2008;51:1613–1614.

1

1

1

1

1

1

1

1

1

1

2

382 The American Journal of Cardiology (www.AJConline.org)

8. Bakhtiary F, Therapidis P, Scherer M, Dzemali O, Moritz A, Kleine P.Electromagnetic interaction between an axial left ventricular assistdevice and an implantable cardioverter defibrillator. J Thorac Cardio-vasc Surg 2008;136:1380–1381.

9. Oswald H, Klein G, Struber M, Gardiwal A. Implantable defibrillatorwith left ventricular assist device compatibility. Interact CardiovascThorac Surg 2009;8:579–580.

0. Jacob S, Cherian PK, Ghumman WS, Das MK. “Pseudo” Faradaycage: a solution for telemetry link interaction between a left ventricularassist device and an implantable cardioverter defibrillator. J IntervCard Electrophysiol Epub 2009 July 28.

1. Biviano A, Mancini D, Naka Y, Arellano J, Garan H. Overcomingelectromagnetic interference by LVADs on ICD function by shieldingthe ICD programmer wand and extension cable. Pacing Clin Electro-physiol 2009;32:945–948.

2. Miller LW, Pagani FD, Russell SD, John R, Boyle AJ, Aaronson KD,Conte JV, Naka Y, Mancini D, Delgado RM, MacGillivray TE, FarrarDJ, Frazier OH. Use of a continuous-flow device in patients awaitingheart transplantation. N Engl J Med 2007;357:885–896.

3. Andersen M, Videbaek R, Boesgaard S, Sander K, Hansen PB,Gustafsson F. Incidence of ventricular arrhythmias in patients onlong-term support with a continuous-flow assist device (HeartMate II).J Heart Lung Transplant 2009;28:733–735.

4. Harding JD, Piacentino Vr, Gaughan JP, Houser SR, Margulies KB.

Electrophysiological alterations after mechanical circulatory support

in patients with advanced cardiac failure. Circulation 2001;104:1241–1247.

5. Harding JD, Piacentino Vr, Rothman S, Chambers S, Jessup M, Mar-gulies KB. Prolonged repolarization after ventricular assist devicesupport is associated with arrhythmias in humans with congestive heartfailure. J Card Fail 2005;11:227–232.

6. Lab M. Contraction-excitation feedback in myocardium: physiologicalbasis and clinical relevance. Circ Res 1982;50:757–766.

7. Taggart P, Sutton PM, Treasure T, Lab M, O’Brien W, Runnalls M,Swanton RH, Emanuel RW. Monophasic action potentials at discon-tinuation of cardiopulmonary bypass: evidence for contraction-excita-tion feedback in man. Circulation 1988;77:1266–1275.

8. Bardy GH, Lee KL, Mark DB, Poole JE, Packer DL, Boineau R,Domanski M, Troutman C, Anderson J, Johnson G, McNulty SE,Clapp-Channing N, Davidson-Ray LD, Fraulo ES, Fishbein DP, Lu-ceri RM, Ip JH. Amiodarone or an implantable cardioverter-defibril-lator for congestive heart failure. N Engl J Med 2005;352:225–237.

9. Germano JJ, Reynolds M, Essebag V, Josephson ME. Frequency andcauses of implantable cardioverter-defibrillator therapies: is devicetherapy proarrhythmic? Am J Cardiol 2006;97:1255–1261.

0. Deng MC, Tjan TD, Asfour B, Gradaus R, Bocker D, Loick HM, BabaHA, Breithardt G, Scheld HH, Borggrefe M, Hammel D. Combiningnonpharmacologic therapies for advanced heart failure: the Münsterexperience with the assist device-defibrillator combination. Am J Car-

diol 1999;83:158D–160D.