ICD Implantation and Arrhythmia-Free Survivalin Patients with Depressed LV Function FollowingSurgery for Valvular Heart DiseaseFELIX YANG, M.D.,* BINDI SHAH, M.D.,† SEI IWAI, M.D.,† STEVEN M. MARKOWITZ, M.D.,†BRUCE B. LERMAN, M.D.,† and KENNETH M. STEIN, M.D.†From the *Department of Cardiology, Beth Israel Medical Center, University Hospital of the Albert Einstein Collegeof Medicine, New York, New York; and †Greenberg Division of Cardiology, Weill Medical College of CornellUniversity, New York, New York

Background: Although prophylactic implantable cardioverter-defibrillator (ICD) implantation is benefi-cial in patients with severe ischemic cardiomyopathy, it is unclear whether patients with cardiomyopathydue to valvular heart disease have a similar benefit.

Methods: We followed 17 patients (14 men/three women, age 62 ± 13 years, left ventricular ejectionfraction [LVEF] 29 ± 10%) who had nonischemic valvular cardiomyopathy, underwent valvular heartsurgery (aortic valve replacement, mitral valve replacement, and/or mitral valve repair), and subsequentlyhad an electrophysiology study (EPS), for a median of 2.8 years. These patients were compared with 34patients with prior myocardial infarction and no significant valvular heart disease, who were matched (1:2)for age, gender, LVEF, EPS result, T-wave alternans result, and ICD placement. Occurrence of arrhythmiaswas ascertained from ICD device clinic follow-up and vital status was determined using the National DeathIndex.

Results: There were no differences between the groups in overall survival (P = 0.24) or arrhythmia-freesurvival (P = 0.38), and the 2-year arrhythmia-free survival was 82% for the valvular patients versus73% for the ischemic patients. Among patients with ICDs, there was no difference between the groups inoverall survival (P = 0.34), time to first appropriate ICD therapy (P = 0.54), and arrhythmia-free survival(P = 0.51).

Conclusion: Patients with valvular cardiomyopathy and residual left ventricular dysfunction follow-ing valvular surgery who underwent a tailored approach to ICD implantation had similar overall andarrhythmia-free survival as patients with ischemic cardiomyopathy. (PACE 2008; 31:1419–1424)

implantable cardioverter-defibrillator, valvular heart disease

IntroductionPatients with residual left ventricular dys-

function after valve surgery have an increasedfrequency of ventricular arrhythmias and in-creased mortality.1,2 Prophylactic implantablecardioverter-defibrillator (ICD) implantation isbeneficial in patients with severe ischemic car-diomyopathy3–6 and in the broad group of patientswith nonischemic cardiomyopathy.6,7 However, itis unclear whether patients with cardiomyopathymore specifically due to valvular heart disease sim-ilarly benefit. We therefore sought to investigatethe incidence of ventricular arrhythmias and mor-tality in patients with nonischemic valvular car-diomyopathy who were referred for electrophysi-ology study (EPS) and evaluated for possible ICDimplantation.

Address for reprints: Kenneth M. Stein, M.D., Greenberg Divi-sion of Cardiology, Weill Medical College of Cornell University,525 East 70th Street, Starr 4, New York, NY 10021. Fax: 212-746-6951; e-mail: kstein@med.cornell.edu

Received April 28, 2008; revised July 27, 2008; accepted August3, 2008.

MethodsStudy Population

The patients in this study were part of aprospectively acquired database that includes allpatients who underwent EPS and T-wave alter-nans (TWA) testing at the New York Presbyte-rian Hospital-–Weill Cornell Medical Center. Wesearched through our database of 1,121 patientswho had undergone EP testing from January 2001to December 2003 for patients who had valvu-lar heart surgery without significant ischemicheart disease. A vast majority of patients werereferred for nonsustained ventricular tachycardia(NSVT). Patients were included in the study pop-ulation if they had a history of aortic valve re-placement (AVR), mitral valve replacement (MVR),or repair (MVr). Patients with a history of my-ocardial infarction (MI) or coronary artery dis-ease (CAD) requiring intervention were excludedfrom this study group. As patients were selectedfor the primary prevention of sudden cardiacdeath (SCD), survivors of cardiac arrest or clin-ical sustained ventricular tachycardia were alsoexcluded.

C©2008, The Authors. Journal compilation C©2008, Blackwell Publishing, Inc.

PACE, Vol. 31 November 2008 1419

YANG, ET AL.

Patients from this nonischemic valvular car-diomyopathy group were matched 1:2 with the is-chemic control patients who had a prior MI andno significant valvular heart disease. Patients werematched for age, gender, left ventricular ejectionfraction (LVEF), TWA result, EPS result, and ICDplacement. This study was reviewed and approvedby the Institutional Review Board.

TWA Testing

All patients underwent TWA testing at thetime of EPS during 5 minutes of high right atrialpacing at a cycle length of 550 ms. TWA wasrecorded using the spectral analysis algorithm ofthe Cambridge Heart, Inc. CH2000 or HearTwave(Cambridge Heart, Bedford, MA, USA) and was in-terpreted using standard criteria.8 A positive TWAtest was defined as the presence of sustained TWAat rest or during atrial pacing. A negative test wasdefined as failure to meet criteria for positivity, andby at least 1 minute of artifact-free data without sig-nificant alternans while the heart rate was ≥105bpm. Tests that met criteria for neither positivenor negative TWA (e.g., due to excessive ectopy,noise, or AV Wenckebach) were labeled “indeter-minate.” Based on prior studies, positive and inde-terminate results were prospectively grouped to-gether as “nonnegative.”9,10 β-blockers or calciumchannel blockers were not routinely held prior tostudy.

Electrophysiological Testing

All patients underwent EPS with up to tripleventricular extrastimuli at two different drive traincycle lengths from two right ventricular siteswith and without infusion of isoproterenol. Pos-itive EPS was defined as the induction of sus-tained monomorphic ventricular tachycardia (VT)with up to triple ventricular extrastimuli, or theinduction of sustained polymorphic VT or ventric-ular fibrillation (VF) with up to double ventric-ular extrastimuli. Sustained ventricular arrhyth-mias were defined as episodes lasting at least30 seconds or associated with syncope and/orhemodynamic compromise requiring either anti-tachycardia pacing or electrical cardioversion. ICDimplantation, including device selection and pro-gramming, was at the discretion of the treatingphysician.

Data Collection and Follow-Up

All patients were followed prospectivelythrough chart review and query of the NationalDeath Index (NDI). For those patients who receivedICDs or permanent pacemakers, periodic deviceinterrogations (approximately every 3–6 months)were used to survey for arrhythmic events. De-vice electrograms were interpreted by blinded ob-

servers, and episodes for which therapy was de-livered were classified as VT, VF, or inappropriate.The primary endpoint was all-cause mortality. Thesecondary endpoints were time to first VT or VFand arrhythmia-free survival.

Statistical Analysis

Demographic and clinical variables were an-alyzed using descriptive statistics. For pairwisecomparisons, Fisher’s exact test was used for cat-egorical variables, and the t-test was used to com-pare continuous variables. Kaplan-Meier survivalcurves were constructed for overall mortality andarrhythmia-free survival among all patients, aswell as for freedom from arrhythmia among pa-tients with ICDs. Life tables associated with thesecurves were used to calculate estimated 2-yearevent rates. The log-rank test was used to test forsurvival differences between groups. All statisti-cal analyses were performed using SPSS Version12.0 for Windows (SPSS Inc., Chicago, IL, USA).For all purposes, a P value <0.05 was required toreject the null hypothesis.

ResultsClinical Characteristics

Seventeen patients (14 men and three women)met the inclusion criteria for nonischemic valvularcardiomyopathy. Six patients had severe aortic re-gurgitation, four had severe aortic stenosis, threehad severe aortic regurgitation and mitral regur-gitation, one had severe mitral regurgitation, onehad severe aortic stenosis and severe mitral steno-sis, one had severe aortic regurgitation and aor-tic stenosis, and one had severe aortic regurgita-tion and mitral regurgitation. These patients whohad either an AVR alone (n = 11), AVR plus MVR(n = 5), or MVr alone (n = 1) were followed fora median of 2.8 years. The average age of the pa-tients was 62 ± 13 years. The mean time elapsedfrom valve replacement to EPS was 817 days. Allof the patients were referred for NSVT. The aver-age LVEF at the time of EPS was 29 ± 10%. Sixof the seventeen patients (35%) had positive EPSand all underwent ICD implantation. Four patientswith negative EPS, but nonnegative TWA, under-went ICD implantation due to severe left ventric-ular dysfunction (n = 1) or as part of implantationof a biventricular pacing system (n = 3). There wasno significant difference in LVEF between patientswho received and did not receive an ICD. These17 patients were matched with 34 patients withprior MI and no significant valvular heart disease(Table I, panels A and B).

There were no significant differences betweenthe valvular and ischemic groups with regard tosex; occurrence of atrial fibrillation; or use of

1420 November 2008 PACE, Vol. 31

VALVULAR CARDIOMYOPATHY AND ICD IMPLANTATION

Table I.

Demographic Characteristics. Panel A, Unmatched Variables and Panel B, Matched Variables

Valvular (N = 17) Ischemic (N = 34) P Value

A: Unmatched Variables

Hypertension (%) 53 79 0.05Diabetes (%) 5.9 35 0.02Atrial fibrillation (%) 18 15 0.09ACEI/ARB (%) 59 59 1.00Aldactone (%) 5.9 8.8 0.71Antiarrhythmics (%) 0 0 −Aspirin (%) 29 97 <0.01β-blockers (%) 82 88 0.57Calcium channel blocker (%) 18 21 0.80Diuretics (%) 65 62 0.83Digoxin (%) 35 12 0.05Nitrates (%) 0 15 0.10Plavix (%) 0 21 0.04Statins (%) 18 59 0.01BUN (mg/dL) 20 ± 6.9 25 ± 12 0.11Creatinine (mg/dL) 1.1 ± 0.7 1.3 ± 0.59 0.35QRS duration (ms) 132 ± 27 114 ± 27 0.04QTc duration (ms) 457 ± 48 430 ± 40 0.03

B: Matched Variables

Age (years) 62 ± 13 66 ± 14 0.34Gender—male (%) 82 91 0.36Left ventricular ejection fraction (%) 29 ± 9.8 29 ± 8.4 0.89TWA result (% nonnegative) 71 71 NSEPS result (% positive) 35 35 NSICD placement (%) 59% 59% NS

β-blocker, angiotensin-converting enzyme in-hibitor, or angiotensin receptor blocker, nitrates,diuretics, or aldactone. Neither group had patientson antiarrhythmic medications. Ischemic patientshad significantly higher use of digoxin, aspirin,statins, and clopidogrel. In addition, they had a sig-nificantly higher proportion of hypertension anddiabetes. There were no significant differences be-tween the groups with respect to blood urea nitro-gen, or creatinine. However, QRS (132 ms ± 27valvular, 114 ms ± 27 ischemic; P = 0.04) andQTc (457 ms ± 48 valvular, 430 ms ± 40 ischemic;P = 0.03) duration were significantly longer in thevalvular group.

Follow-Up

Life table analysis revealed no differences be-tween the groups in overall survival (Fig. 1, P =0.24) or arrhythmia-free survival (time to first VT,first VF, or death) (Fig. 2, P = 0.38). Two-yeararrhythmia-free survival was 82% for the valvu-lar patients versus 73% for the ischemic patients.

Among the subset of patients with ICDs, there wasno difference between the groups in overall sur-vival (Fig. 1, P = 0.34), time to first appropriateICD therapy (Fig. 3, P = 0.54), and arrhythmia-freesurvival (Fig. 2, P = 0.51). Among the subset of pa-tients who did not receive ICDs, there was also nodifference between valvular and ischemic patientsin overall survival (P = 0.47).

DiscussionThe principal finding of this study is that in

patients with valvular heart disease with resid-ual left ventricular dysfunction following valvesurgery who underwent a tailored approach to ICDimplantation, overall survival and arrhythmia-freesurvival is similar to that observed in patientswith ischemic cardiomyopathy who underwentthe same EP-guided approach. This suggests thattreatment based on this strategy yields a compara-ble benefit for patients with valvular heart diseaseand patients with ischemic heart disease.

PACE, Vol. 31 November 2008 1421

YANG, ET AL.

Figure 1. Overall survival following electrophysiologicstudy (EPS). There was no significant difference in over-all survival between patients with valvular cardiomy-opathy versus patients with ischemic cardiomyopathy.There was also no significant difference in overall sur-vival in the subset of patients treated with ICDs followingEPS.

Current data support ICD placement for pri-mary prevention in patients with nonischemic car-diomyopathy.6,7 These trials, however, did not fur-ther differentiate among causes of cardiomyopathyin these patients and, as a result, include a hetero-geneous group of disorders. We sought to specif-ically examine the role of electrophysiologic riskstratification and ICDs in the prophylactic treat-ment of patients with valvular heart disease whohad undergone valve surgery. Since valvular car-diomyopathy patients have an increased risk of ar-rhythmic events, ICD placement in these patientsmay be an important consideration. Nevertheless,although this is a potentially large patient popula-tion, it likely represents only a small minority ofpatients enrolled in primary prevention ICD trials.

Valve disease is known to predispose patientsto arrhythmias. Patients with aortic and mitralvalvular heart disease have an increased incidenceof NSVT.11–13,16 These arrhythmias are not onlymarkers of underlying ventricular pathology, butalso a contributing factor in the sudden deaths ob-served in patients with valve disease. In the firstNatural History Study (NHS-1), 25 of 462 patients(5.4%) with aortic stenosis died suddenly, and oth-ers have shown that VT is a cause of sudden deathamong patients with aortic valve disease.14,15

Figure 2. Arrhythmia-free survival following electro-physiologic study (EPS). There was no significant differ-ence in arrhythmia-free survival between patients withvalvular cardiomyopathy versus patients with ischemiccardiomyopathy. There was also no significant differ-ence in arrhythmia-free survival in the subset of patientstreated with ICDs following EPS.

Figure 3. Time to first documented ventricular arrhyth-mia in the subset of patients treated with ICDs followingelectrophysiologic study (EPS). There was no significantdifference in time to a first arrhythmia in patients withvalvular cardiomyopathy versus patients with ischemiccardiomyopathy.

1422 November 2008 PACE, Vol. 31

VALVULAR CARDIOMYOPATHY AND ICD IMPLANTATION

Patients with valvular disease may be pre-disposed to ventricular arrhythmia and SCD bymultiple possible mechanisms.16 Increased my-ocardial mass, ventricular dilatation, ventricularwall stress, and subendocardial ischemia despitenormal coronary arteries may cause an increasedincidence of spontaneous complex ventriculartachyarrhythmias that may be associated with sus-tained ventricular tachyarrythmias and SCD.17–23

In addition to these effects of valve disease uponthe myocardium, ventricular arrhythmias may bea consequence of long-term myocardial damageand fibrosis from surgery. Michel et al. even re-port that ventricular arrhythmias seem to increasein frequency soon after valve surgery.24 Rates ofinducible sustained bundle branch reentry are in-creased after valve surgery.25 Overall rates of SCDin patients with prosthetic valves vary consider-ably and range from 15% to 30% with an esti-mated annual risk of 0.2% to 0.9%.26 In a largeseries of 1,533 patients who underwent aortic ormitral valve replacement, 6% of patient deathswere caused by arrhythmias.27 In a VA study, sud-den death accounted for 23% of deaths for MVRand 16% for AVR.28 Necropsy studies demonstratethat a majority of sudden deaths are caused by thenatural disease process, the most common beingventricular arrhythmias.29 Of note, ventricular ar-rhythmias are more common in aortic rather thanmitral valve replacements. Although the rate ofSCD decreases after an aortic valve replacement,it is still a common cause of death with an inci-dence of 10–40%. The incidence of ventricular ar-rhythmia is especially high in patients with car-diomegaly, residual cardiac hypertrophy, and leftventricular impairment.26

LimitationsThis study is limited due to the small sample

size. Accordingly, outcomes of specific valvular le-

sions could not be separately evaluated. Kaplan-Meier curves of overall survival, arrhythmia-freesurvival, and VT or VF did not show a differ-ence between the two study populations; however,event numbers were very small and a survival dif-ference may not have been detected. Our investiga-tion is also a case-controlled study and thereforeis subject to selection bias. Although patients inthe study population were matched for age, gen-der, LVEF, TWA result, EPS result, and ICD place-ment, there may have been additional confoundersthat may have affected our results. In addition, re-cent data suggest that ICD shocks may not be anaccurate surrogate for SCD in patients.30 Patientswith nonischemic cardiomyopathy in the ICD armof the DEFINITE trial were noted to have twice asmany appropriate shocks than SCD/cardiac arrestevents in the standard medical therapy arm. In ad-dition, episodes of VT and VF may terminate spon-taneously. Our combined endpoint of time to firstVT or first VF or death thus may be an overestimateof the actual mortality risk in these patients.

ConclusionCurrently, there is no conclusive evidence that

patients with nonischemic valvular cardiomyopa-thy derive the same benefit from ICD implan-tation as ischemic nonvalvular patients. To ourknowledge, ours is the first investigation into ICDimplantation specifically into patients with signif-icant valvular heart disease. Using a tailored ap-proach to ICD implantation, overall survival andarrhythmia-free survival in patients with nonis-chemic valve disease and residual left ventriculardysfunction following valve surgery is similar tothat observed in patients with ischemic cardiomy-opathy. Further studies will be needed to verifythe efficacy of ICD implantation in this patientpopulation and to determine optimal implantationcriteria.

References1. Michel PL, Mandagout O, Vahanian A, Cormier B, Iung B, Lux-

ereau P, Acar J. Ventricular arrhythmias in aortic valve disease be-fore and after aortic valve replacement. Acta Cardiol 1992; 47:145–156.

2. Hochreiter CA, Borer JS, Yin A, Supino PG, Herrold EM, KriegerK, Isom OW. Ventricular arrhythmias in mitral regurgitation: Fre-quency, clinical and prognostic importance, management before andafter mitral valve surgery. Adv Cardiol 2004; 41:108–111.

3. Moss AJ, Hall WJ, Cannom DS, Daubert JP, Higgins SL, Klein H,Levine JH, et al. Multicenter automatic defibrillator implantationtrial investigators. Improved survival with an implanted defibril-lator in patient with coronary disease at high risk for ventriculararrhythmia. NEJM 1996; 335:1933–1940.

4. Buxton AE, Lee KL, Fisher JD, Josephson ME, Prystowsky EN, HafleyG. Multicenter unsustained tachycardia trial investigators. A ran-domized study of the prevention of sudden death in patients withcoronary artery disease. NEJM 1999; 341:1882–1890.

5. Moss AJ, Zareba W, Hall WJ, Klein H, Wilber DJ, Cannom DS, DaubertJP, et al. Prophylactic implantation of a defibrillator in patients withmyocardial infarction and reduced ejection fraction. NEJM 2002;346:877–883.

6. Bardy GH, Lee KL, Mark DB, Poole JE, Packer DL, Boineau R,Domanski M, et al. Amiodarone or an implantable cardioverter-defibrillator for congestive heart failure. NEJM 2005; 352:225–237.

7. Kadish A, Dyer A, Daubert JP, Quigg R, Estes NA, Anderson KP,Calkins H, et al. Prophylactic defibrillator implantation in patientwith nonischemic dilated cardiomyopathy. NEJM 2004; 350:2151–2158.

8. Bloomfield DM, Hohnloser SH, Cohen RJ. Interpretation and clas-sification of microvolt T wave alternans tests. J Cardiovasc Electro-physiol 2002; 13:502–512.

9. Hohnloser SH, Ikeda T, Bloomfield DM, Dabbous OH, Cohen RJ. T-wave alternans negative coronary patients with low ejection fractionand benefit from defibrillator implantation. Lancet 2003; 362:125–126.

10. Chow T, Kereiakes D, Bartone C, Booth T, Schloss EJ, Waller T,Chung ES, et al. Prognostic utility of microvolt T-wave alternans inrisk stratification of patients with ischemic cardiomyopathy. J AmColl Cardiol 2006; 47:1820–1827.

11. Klein RC. Ventricular arrhythmias in aortic valve disease: Analysisof 102 patients. Am J Cardiol 1984; 53:1079–1083.

PACE, Vol. 31 November 2008 1423

YANG, ET AL.

12. Von Olshausen KV, Schwarz F, Apfelbach J, Rohrig N, Kubler W. De-terminants of the incidence and severity of ventricular arrhythmiasin aortic valve disease. Am J Cardiol 1983; 51:1103–1109.

13. Delahaye JP, Gare JP, Viguier E, Michel PL, Thomas D. Preoperativeventricular arrhythmias in mitral regurgitation. Acta Cardiol 1992;47:167–173.

14. Wolfe RR, Driscoll DJ, Gersony WM, Hayes CJ, Keane JF, Kidd L,O’Fallon WM, et al. Arrhythmias in patients with valvular aorticstenosis, valvular pulmonary stenosis, and ventricular septal defect:Results of 24-hour ECG monitoring. Circulation 1993; 87(Suppl I.):I-89–I-101.

15. Von Olshausen K, Witt T, Schmidt G, Meyer J. Ventricular tachycar-dia as a cause of sudden death in patients with aortic valve disease.Am J Cardiol 1987; 59:1214–1215.

16. Martinez-Rubio A, Schwammenthal Y, Schwammenthal E, Block M,Reinhardt L, Garcia-Alberola A, Sierra G, et al. Patients with valvu-lar heart disease presenting with sustained ventricular tachyarrhyth-mias or syncope: Results of programmed ventricular stimulation andlong-term follow up. Circulation 1997; 96:500–508.

17. Leclercq JF, Coumel P, Maison-Blanche P, Cauchemez B, Zimmer-mann M, Chouty F, Slama R. The mechanism of sudden death: Acooperative study of 69 cases recorded by the Holter method. ArchMal Coeur 1986; 79:1024–1033.

18. Bayes de Luna A, Coumel P, Leclercq JF. Ambulatory sudden cardiacdeath: Mechanisms of production of fatal arrhythmia on the basis ofdata from 157 cases. Am Heart J 1989; 117:151–159.

19. Von Olshausen KV, Witt T, Pop T, Treese N, Bethge KP, Meyer J.Sudden cardiac death while wearing a Holter monitor. Am J Cardiol1991; 67:381–386.

20. Peterson KL, Tsuji J, Johnson A, DiDona J, Le Winter M. Diastolic leftpressure-volume and stress-strain relations in patients with valvularaortic stenosis and left ventricular hypertrophy. Circulation 1978;58:77–89.

21. Johnson LL, Scalacca RR, Ellis K, Weiss MB, Cannon PJ. Reduced leftventricular myocardial blood flow per unit mass in aortic stenosis.Circulation 1978; 57:582–590.

22. Vincent WR, Buckberg GD, Hoffman JIE. Left ventricular subendo-cardial ischemia in severe valvular and supravalvular aortic steno-sis: A common mechanism. Circulation 1974; 49:326–333.

23. Buckberg GD, Eber L, Heiman M, Gorlin R. Ischemia in aorticstenosis: Hemodynamic prediction. Am J Cardiol 1975; 35:778–784.

24. Michel PL, Mandagout O, Vahanian A, Cormier B, Iung B, LuxereauP, Acar J. Ventricular arrhythmias in aortic valve disease before andafter surgery. J Heart Valve Dis 1992; 1:72–79.

25. Narasimhan C, Jazayeri MR, Sra J, Dhala A, Deshpande S, Biehl M,Akhtar M, et al. Ventricular tachycardia in valvular heart disease:Facilitation of sustained bundle-branch reentry by valve surgery.Circulation 1997; 96:4307–4313.

26. Groves P. Surgery of valve disease: Late results and late complica-tions. Heart 2001; 86:715–721.

27. Blackstone EH, Kirklin JW. Death- and time-related event after valvereplacement. Circulation 1985; 72:753–767.

28. Hwang MH, Burchfiel CM, Sethi GK, Oprian C, Grover FL, Hender-son WG, Hammermeister K. Comparison of the causes of late deathfollowing aortic and mitral valve replacement. VA Co-operativestudy on valvular heart disease. J Heart Valve Dis 1994; 3:17–24.

29. Burke PA, Farb A, Sessums L, Virmani R. Causes of sudden cardiacdeath in patients with replacement valves: An autopsy study. J HeartValve Dis 1994; 3:10–16.

30. Ellenbogen KA, Levine JH, Berger RD, Daubert JP, Winters SL,Greenstein E, Shalaby A, et al. Are implantable cardioverter de-fibrillator shocks a surrogate for sudden cardiac death in patientswith nonischemic cardiomyopathy? Circulation 2006; 113:776–782.

1424 November 2008 PACE, Vol. 31