iwrvtwca

maMa

3

0h

Long-Term Follow-Up of Implantable Cardioverter-Defibrillatorfor Secondary Prevention in Chagas’ Heart Disease

Martino Martinelli, MD, PhDa,*, Sérgio Freitas de Siqueira, MSc Enga,Eduardo Back Sternick, MD, PhD, FHRSb, Anis Rassi, Jr., MD, PhDc, Roberto Costa, MD, PhDd,

José Antônio Franchini Ramires, MD, PhDe, and Roberto Kalil Filhoe

Assessing the efficacy of implantable cardioverter-defibrillators (ICD) in patients withChagas’ heart disease (ChHD) and identifying the clinical predictors of mortality and ICDshock during long-term follow-up. ChHD is associated with ventricular tachyarrhythmiasand an increased risk of sudden cardiac death. Although ChHD is a common form ofcardiomyopathy in Latin American ICD users, little is known about its efficacy in thetreatment of this population. The study cohort included 116 consecutive patients withChHD and an ICD implanted for secondary prevention. Of the 116 patients, 83 (72%) weremen; the mean age was 54 � 10.7 years. Several clinical variables were tested in amultivariate Cox model for predicting long-term mortality. The average follow-up was45 � 32 months. New York Heart Association class I-II developed in 83% of patients. Themean left ventricular ejection fraction was 42 � 16% at implantation. Of the 116 patients,58 (50%) had appropriate shocks and 13 (11%) had inappropriate therapy. A total of 31patients died (7.1% annual mortality rate). New York Heart Association class III (hazardratio [HR] 3.09, 95% confidence interval 1.37 to 6.96, p � 0.0064) was a predictor of a worseprognosis. The left ventricular ejection fraction (HR 0.972, 95% confidence interval 0.94 to0.99, p � 0.0442) and low cumulative right ventricular pacing (HR 0.23, 95% confidenceinterval 0.11 to 0.49, p � 0.0001) were predictors of better survival. The left ventriculardiastolic diameter was an independent predictor of appropriate shock (HR 1.032, 95%confidence interval 1.004 to 1.060, p � 0.025). In conclusion, in a long-term follow-up, ICDefficacy for secondary sudden cardiac death prevention in patients with ChHD was markedby a favorable annual rate of all-cause mortality (7.1%); 50% of the cohort receivedappropriate shock therapy. New York Heart Association class III and left ventricularejection fraction were independent predictors of worse prognosis, and low cumulative rightventricular pacing defined better survival. © 2012 Elsevier Inc. All rights reserved. (Am

J Cardiol 2012;xx:xxx)

rv

M

ipB�vIi

poBtiasma

Data are conflicting regarding the efficacy of implant-able cardioverter-defibrillator (ICD) in patients with Cha-gas’ heart disease (ChHD). The evidence is based on theresults of 2 manufacturer-sponsored registries1–3 and 1retrospective cohort study.4 The reported annual mortal-ty rates in these 3 studies varied from 5.5% to 16.6%,ith the upper value greater than the mortality rates

eported in studies restricted to patients with sustainedentricular tachycardia (VT) who were treated with an-iarrhythmic drugs. Thus, the aim of the present studyas to assess, during long-term follow-up, the ICD effi-

acy of a ChHD cohort from a tertiary center, consideringll-cause mortality and appropriate ICD shock therapy

aPacemaker Clinic, dDivision of Cardiac Surgery, and eDivision ofCardiology, Heart Institute (InCor), Hospital das Clínicas da Faculdade deMedicina da Universidade de São Paulo, São Paulo, SP, Brazil; bArrhyth-

ia and Electrophysiology Unit, Biocor Instituto, Nova Lima, MG, Brazil;nd cDivision of Cardiology, Anis Rassi Hospital, Goiânia, GO, Brazil.anuscript received March 22, 2012; revised manuscript received and

ccepted May 23, 2012.*Corresponding author: Tel: (�55) 11-2661-5516; fax: (�55) 11-

081-7148.

cE-mail address: martino@incor.usp.br (M. Martinelli).

002-9149/12/$ – see front matter © 2012 Elsevier Inc. All rights reserved.ttp://dx.doi.org/10.1016/j.amjcard.2012.05.040

ates and analyzing the predictive value of the clinicalariables.

ethods

The inclusion criteria were ChHD, diagnosed by pos-tive serologic tests, and an ICD implanted for secondaryrevention of sudden cardiac death, according to therazilian guidelines.5,6 The exclusion criteria were age18 years old, advanced atrioventricular block, or pre-

ious pacemaker or cardiac resynchronization device atCD implantation (n � 76 patients). The study flow charts shown in Figure 1.

The present study was a retrospective cohort study ofatients with ChHD and an ICD from the Heart Institutef the University of São Paulo Medical School (InCor) inrazil. The ethics committee of our institution approved

he study in January 2000 and the study was implementedn June 2000. The last inclusion occurred in June 2008,nd follow-up was closed in March 2009. A databaseystem was designed to include patient characteristics,edical history, ICD indication, and functional outcomes

t implantation and during follow-up. The data were

ollected on several potential risk factors, including age,

www.ajconline.org

vtdcofc

mvMGMiitopisiatiw

aRfitc

d

cao

n for th

2 The American Journal of Cardiology (www.ajconline.org)

gender, New York Heart Association functional class,syncope, left ventricular ejection fraction (LVEF), theuse of drugs (amiodarone, � blockers, angiotensin-con-

erting enzyme inhibitors), echocardiographic parame-ers (LVEF, left ventricular diastolic diameter, left atrialiameter, and mitral regurgitation), the percentage ofumulative right ventricular pacing, and the occurrencef appropriate and inappropriate shock therapy. Radio-requency catheter ablation was undertaken on a case-by-ase basis.

Before ICD implantation, all patients received optimaledical therapy. The ICD implants included 114 trans-

enous systems, manufactured by Medtronic (Minneapolis,innesota) (33%), Biotronik (Berlin, Germany) (25%),uidant/CPI (St. Paul, Minnesota) (10%), and St. Judeedical (Sylmar, California) (32%). The ICD programming

ncluded therapy for standard VT with antitachycardia pac-ng combined with low-energy shock therapy and for ven-ricular fibrillation (VF)—shock therapy with a 300-ms cut-ff cycle interval. It was considered standard VT in theresence of sustained tachycardia with a cycle interval rang-ng from 300 to 400 ms, not discriminated as supra-VT bypecific algorithms. It was considered VF when the cyclenterval was inferior to 300 ms. ICD therapy was classifieds appropriate for VT/VF if the recorded intracardiac elec-rogram for the intervention was compatible with the clin-cal manifestation, on medical discretion. The shock therapy

Figure 1. Flow-chart of patient selection. Criteria of patient selectio

as considered inappropriate when the ICD therapy was

pplied to supra-VT, noise, myopotential oversensing, or-wave double counting. Customized programs were per-

ormed according to individual needs. The decision regard-ng activation of pacing therapy for bradycardia was left tohe discretion of each physician at the outpatient pacemakerlinic.

The follow-up protocol included regular clinical andevice evaluation performed 3 times annually.

The primary outcome was death from all causes. Theircumstances of the deaths were reviewed and categorizeds from a cardiac or noncardiac cause, and the classificationf Hinkle and Thaler7 was used to evaluate the suspected

mechanism of death.The primary data analysis examined the effect of the

potential risk factors on appropriate shocks and long-termmortality. Cumulative survival was assessed using the Coxregression method adjusted for independent covariates, andthe differences were compared using Cox proportional haz-ard models. Analyses, adjusted for potential confounders,were also performed using multivariate Cox proportionalhazard models. The covariates were applied in a forwardselection according to the unadjusted association of p �0.05in the univariate models. The Cox model assumption ofproportional hazards was found to be valid from the log-minus-log curves and the Schoenfeld test. The softwareused was SPSS, version 16.0.1, for Windows (SPSS, Chi-

e study. CRT-D � cardiac resynchronization therapy-defibrillator.

cago, Illinois).

Tt(tVpc

(mampcdrac[

p0ip

cc2is

t(opsapdtwa

atsi1

D

wttc

a

ht

N

SMLLAM

3Cardiomyopathy/Implantable Cardioverter-Defibrillator in Chagas’ Heart Disease

Results

The study cohort consisted of 116 patients. Of the 116patients, 73 were men. The patient age range was 18 to 79years (mean 54 � 10.7 years). New York Heart Associationfunctional class I-II was present in 82.7%, and the meanLVEF was 42.4 � 15.7%. Of the 116 patients, 90 (77.6%)were taking a median daily dose of 400 mg (range 100 to600) amiodarone, and 38 (32.8%) were using a � blocker.

he ICD indication was resuscitated VF or VT in 21 pa-ients (18.1%) and symptomatic sustained VT in 9581.9%). Of these, 64 patients (55.2%) underwent an elec-rophysiologic study and had hemodynamic unstable VT orF induced by �3 extra stimuli. ICD implantation waserformed from June 1990 to June 2008. The baselinelinical characteristics of the cohort are listed in Table 1.

During a mean follow-up of 45 � 32 months, 31 deaths26.7%) occurred (Table 2), corresponding to an annualortality rate of 7.1%. The potential confounder variables

pplicable to the multivariate Cox proportional hazardsodels (those associated with death on univariate analysis,�0.05) were New York Heart Association functional

lass, the use of amiodarone, LVEF, left ventricular end-iastolic diameter, left atrial diameter, nonsustained VT,ate of cumulative right ventricular pacing, and number ofppropriate shocks. New York Heart Association functionallass III (hazard ratio [HR] 3.09, 95% confidence interval

Table 1Baseline clinical characteristics (n � 116)

Characteristics at implantation Value

Men 73 (62%)Mean age (years) 54.0 � 10.7

ew York Heart Association functional classI 36 (31%)II 60 (52%)III 20 (17%)yncope 73 (63%)ean left ventricular ejection fraction (%) 42.4 � 15.7

eft atrium (mm) 41.2 � 6.7eft ventricular end-diastolic diameter (mm) 63.3 � 9.2miodarone 90 (78%)edian daily dose (mg) 400

� Blocker 38 (33%)Angiotensin-converting enzyme inhibitors 95 (82%)Electrophysiologic study 79 (68.1%)

Not induced 3 (4%)Monomorphic ventricular tachycardia 1 (1%)Monomorphic sustained ventricular tachycardia 61 (77%)Polymorphic sustained ventricular tachycardia 11 (14%)Ventricular fibrillation 3 (4%)

Mean QRS width (ms) 125.6Nonsustained ventricular tachycardia 59 (66%)Sustained ventricular tachycardia 3 (3%)Implantable cardioverter-defibrillator indication

Aborted sudden cardiac death 21 (18%)Symptomatic sustained ventricular tachycardia 95 (82%)

Type of implantable cardioverter-defibrillatorAtrioventricular 67 (58%)Ventricular 49 (42%)

Median percentage of cumulative right ventricular pacing 17%

CI] 1.37 to 6.96, p � 0.0064) was a predictor of a worse

rognosis. The LVEF (HR 0.972, 95% CI 0.94 to 0.99, p �.0442) and a low rate of cumulative right ventricular pac-ng (HR 0.23, 95% CI 0.11 to 0.49, p � 0.0001) wereredictors of better survival.

The survival curves of the entire cohort, stratified ac-ording to New York Heart Association functional class andumulative right ventricular pacing, are presented in Figuresand 3. Survival curves free of appropriate shock therapy,

ncluding the survival curve for the entire cohort, are pre-ented in Figure 4.

A total of 750 VT/VF episodes in 58 patients or 50% ofhe cohort. These patients received 339 appropriate shocks5.8/patient, range 1 to 37), and 4 patients (3.4%) receivednly appropriate antitachycardia pacing therapy. Inappro-riate shocks occurred in 18 patients (3.6/patient), corre-ponding to 15.5% of the cohort. The causes of inappropri-te shocks were atrial fibrillation in 8 patients, noise in 8atients, myopotential oversensing in 1 patient, and R-waveouble counting in 1 patient. Seventeen patients with mul-iple appropriate ICD therapy events, including 11 patientsith “electrical storm,” underwent radiofrequency catheter

blation.Mitral regurgitation and left ventricular end-diastolic di-

meter were the applicable potential confounder variables inhe multivariate Cox regression models for appropriatehocks, with the left ventricular end-diastolic diameter thendependent predictor of this outcome (HR 1.032, 95% CI.004 to 1.060, p � 0.025).

iscussion

In the present study, the all-cause annual mortality rateas 7.1%; 50% of the patients received appropriate ICD

herapy. The mean age of the cohort was 54 years. Thus,hey were, on average, 1 decade younger than patients withoronary heart disease (mean age 65 years).8–10 The mean

LVEF of 42% differed notably from that in patients withischemic heart disease, in whom the mean LVEF is lower,such as was seen in the Antiarrhythmics Versus ImplantableDefibrillators (AVID)8 study (32%) and Canadian Implant-ble Defibrillator Study (CIDS)10 (34%).

According to the functional findings in our patients, mostad nonadvanced congestive heart failure (82.7%, func-ional class I or II). Mady et al,11 in a 3-year follow-up study

of 104 male patients with ChHD and congestive heart fail-ure without sustained VT reported the excellent prognosisof the functional class II subgroup (97% survival during thefollow-up period), suggesting the importance of ventriculararrhythmia as a cause of mortality. The mode of death in

Table 2Causes of death during follow-up

Cause Patients (n)

Unknown 3 (8%)Cardiac

Heart failure 14 (45%)Sudden death 0

Non cardiac 14 (45%)Total 31

patients with ChHD is sudden in nearly 60% of patients,

s

ues of

4 The American Journal of Cardiology (www.ajconline.org)

mostly from VT or VF.12 A peculiar factor in these patientsis the occurrence of sudden cardiac death in patients withnormal or near-normal LVEF, without symptoms of con-gestive heart failure.13

Figure 2. Survival curves considering functional class. Adjusted probabilityentire cohort survival at mean values of independent covariates.

Figure 3. Survival curves considering cumulative right ventricular pacingcumulative right ventricular pacing and entire cohort survival at mean val

The annual all-cause mortality rate was low in the pres-

ent study compared to the mortality in a similar series ofpatients with sustained VT receiving only antiarrhythmicmedication. Scanavacca et al14 studied 35 patients, with ahort-term follow-up period (27 � 20 months) and observed

al curves of the ICD cohort with ChHD, according to functional class and

ted probability survival curves of ICD cohort with ChHD, according toindependent covariates.

surviv

. Adjus

an annual mortality rate of 5.1%, with sudden cardiac death

cdsficpiy

i

hafsmr(b

5sN

5Cardiomyopathy/Implantable Cardioverter-Defibrillator in Chagas’ Heart Disease

rate of 40%. Only 46% recuperated from sudden cardiacdeath or syncopal-sustained VT (i.e., more than one halfwere not eligible for ICD implantation). Rassi et al12 studied34 patients with Chagas’ disease and found an annual mor-tality rate of 11.9%, with 71% reported as sudden death.Sarabanda and Marin-Neto15 recently reported a similarmortality rate (10.7% annually) in a series of 28 patientsafter a mean follow-up of 38 � 16 months; 78% of thedeaths were sudden.

In the present study, ICD efficacy seemed to have playedan important role on the lower annual mortality rate (7.1%)and the nonoccurrence of sudden cardiac death cases. Mostprobably, amiodarone also had an effect on the survival ofthe cohort, although it was not an independent predictor.

The ICD Latin America Chagas Registry,1 which in-luded 89 patients for primary and secondary sudden car-iac death prevention with a very short follow-up period,howed a 1-year mortality rate of 6.7%, similar to ourndings. Toro et al3 published the extended analyses of thatohort, identifying LVEF �30% and age �65 years asredictors of mortality. It is important to emphasize a crit-cal point in these reports—the mean follow-up period of 1ear with a minimal period of 1 month.

Cardinalli-Neto et al4 showed a much greater annualmortality rate (16.6%) in a recent report of a cohort with 90patients with Chagas’ disease and an ICD. The main reasonfor the lower mortality rate observed in our cohort wasradiofrequency catheter ablation. Our patients underwentcatheter ablation for frequently recurrent VT before ICDimplantation and in the setting of multiple appropriate ther-apies (“electrical storm”) after ICD implantation. We alsobelieve that the high rate of patients in our study (81%)

Figure 4. Appropriate shocks and survival curves. Kaplan-Meyer curves ofof cohort.

using angiotensin-converting enzyme inhibitors (not dis-

cussed by Cardinalli-Neto et al4) seemed to have stronglynfluenced the differences in mortality rate.

The occurrence of ICD shock in 50% of our cohortighlights the high risk of recurrent life-threatening tachy-rrhythmia in patients with ChHD. At the end of 5 years ofollow-up, the probability of nonoccurrence of appropriatehocks was 23.6%. It was not considered an independentortality predictor; however, the high survival probability

ate of 63.1% certainly reflected the efficacy of ICD therapyFigure 4). This becomes evident if we consider that Sara-anda and Marin-Neto et al15 showed a similar mortality

rate after a shorter follow-up period (3 years) in patientswithout ICD.

Our study also revealed 3 particular clinical or functionalfactors regarding the effect on survival outcomes. First,patients with an advanced New York Heart Associationfunctional class (class III) had a worse prognosis (near 50%at 5 years). Mady et al11 reported a 3-year survival rate of8% in a class III subgroup versus 97% in the class IIubgroup. Our data disagree with the data from Cardinalli-eto,4 in which almost all studied patients were in class I.

Obviously, this variable did not have an effect on survivaloutcomes. Second, in addition to mild LV compromise, weobserved during clinical follow-up that the lower the LVEF,the worse the patient’s prognosis. The magnitude of thiscorrelation was measured by multivariate statistical analy-sis. Each unit of LVEF reduction increased the risk of deathby 2.8%. Mady et al11 showed that patients with ChHD(receiving an only antiarrhythmic drug) with an LVEF of0.31 to 0.50 had a 3-year survival probability of 70%, whichdecreased to 16% when the LVEF was lower than 0.30.Third, our outcomes emulated the findings from the trial by

lity of survival without appropriate shocks compared to all-cause mortality

probabi

Wilkoff et al16 and have not been demonstrated before in

6 The American Journal of Cardiology (www.ajconline.org)

patients with ChHD. ICD programming without right ven-tricular pacing was a marker of a better prognosis than apacing activated marker. This could have been related to thedeleterious effect provided by a high cumulative right ven-tricular pacing on LV function or could be a consequence ofthe natural disease progression.

The present study was retrospective and uncontrolled.The ICD recipients were not compared to patients withsimilarly high-risk features but without ICD, precluding thequantification of survival benefits. Just as in most studiesassessing the predictors of ICD therapy, the clinical vari-ables were considered at the moment of implantation with-out accounting for changes during follow-up. One couldargue that the need for ventricular pacing could be a markerof more advanced disease and, consequently, an expressionof patients with a worse clinical profile and prognosis.However, we must emphasize that patients with a previouspacemaker and patients with advanced atrioventricularblock were excluded from the present study. The small sizeof the population precluded additional subgroup analysis.

Acknowledgments: We are grateful to Julia Tizue Fuku-shima, MSc for her expert statistical analysis and to RafaelGreco, MD for the collection of data.

1. Muratore CA, Batista Sa LA, Chiale PA, Eloy R, Tentori MC, Escu-dero J, Lima AM, Medina LE, Garillo R, Maloney J. Implantablecardioverter defibrillators and Chagas’ disease: results of the ICDregistry Latin America. Europace 2009;11:164–168.

2. Dubner S, Valero E, Pesce R, Zuelgaray JG, Mateos JC, Filho SG,Reyes W, Garillo R. A Latin American registry of implantable car-dioverter defibrillators: the ICD-LABOR study. Ann NoninvasiveElectrocardiol 2005;10:420–428.

3. Toro D, Muratore C, Aguinaga L, Batista L, Malan A, Greco O,Benchetrit C, Duque M, Baranchuk A, Maloney J. Predictors of all-cause 1-year Mortality in implantable cardioverter defibrillator patientswith chronic Chagas’ cardiomyopathy. Pacing Clin Electrophysiol2011;34:1063–1069.

4. Cardinalli-Neto A, Bestetti RB, Cordeiro JA, Rodrigues VC. Predic-tors of all-cause mortality for patients with chronic Chagas’ heart

disease receiving implantable cardioverter defibrillator therapy. J Car-diovasc Electrophysiol 2007;18:1236–1240.

5. Scanavacca MI, de Brito FS, Maia I, Hachul D, Gizzi J, Lorga A, RassiA Jr, Filho MM, Mateos JC, D’Avila A, Sosa E, Brasileira de Car-diologia S. Guidelines for the evaluation and treatment of patients withcardiac arrhythmias. [Portuguese]. Arq Bras Cardiol 2002;79(Suppl5):1–50.

6. Martinelli Filho M, Zimerman LI, Lorga AM, Vasconcelos JTM, RassiA Jr. Guidelines for implantable electronic cardiac devices of theBrazilian Society of Cardiology. Arq Bras Cardiol 2007;89:e210–e238.

7. Hinkle LE Jr, Thaler HT. Clinical classification of cardiac deaths.Circulation 1982;65:457–464.

8. The Antiarrhythmics Versus Implantable Defibrillators (AVID) Inves-tigators. A comparison of antiarrhythmic-drug therapy with implant-able defibrillators in patients resuscitated from near-fatal in ventriculararrhythmias. N Engl J Med 1997;337:1576–1583.

9. Kuck KH, Cappato R, Siebels J, Rüppel R; CASH Investigators.Randomized comparison of antiarrhythmic drug therapy with implant-able defibrillators in patients resuscitated from cardiac arrest: theCardiac Arrest Study Hamburg (CASH). Circulation 2000;102:748–754.

10. Connolly SJ, Gent M, Roberts RS, Dorian P, Roy D, Sheldon RS,Mitchell LB, Green MS, Klein GJ, O’Brien B; CIDS Investigators.Canadian Implantable Defibrillator Study (CIDS): a randomized trialof the implantable cardioverter defibrillator against amiodarone. Cir-culation 2000;101:1297–1302.

11. Mady C, Cardoso RH, Barretto AC, da Luz PL, Bellotti G, Pileggi F.Survival and predictors of survival in patients with congestive heartfailure due to Chagas’ cardiomyopathy. Circulation. 1994;90:3098–3102.

12. Rassi A Jr, Rassi SG, Rassi A. Sudden death in Chagas’ disease. ArqBras Cardiol 2001;76:75–96.

13. Sternick EB, Martinelli M, Sampaio R, Gerken LM, Teixeira RA,Scarpelli RA, Scanavacca M, Nishioka SD, Sosa E, Sosa E. Suddencardiac death in patients with Chagas heart disease and preserved leftventricular function. J Cardiovasc Electrophysiol 2006;17:113–116.

14. Scanavacca MI, Sosa EA, Lee JH, Bellotti G, Pileggi F. Empirictherapy with amiodarone in patients with chronic Chagas cardiomy-opathy and sustained ventricular tachycardia. Arq Bras Cardiol 1990;54:367–371.

15. Sarabanda AV, Marin-Neto JA. Predictors of mortality in patients withChagas’ cardiomyopathy and ventricular tachycardia not treated withimplantable cardioverter-defibrillators. Pacing Clin Electrophysiol2011;34:54–62.

16. Wilkoff BL, Cook JR, Epstein AE, Greene HL, Hallstrom AP, Hsia H,Kutalek SP, Sharma A; Dual Chamber and VVI Implantable Defibril-lator Trial Investigators. Dual-chamber pacing or ventricular backuppacing in patients with an implantable defibrillator: the dual chamber

and VVI implantable defibrillator (David) trial. JAMA 2002;288:3115–3123.