Device Use Patterns and Clinical Outcome ofImpiantabie Cardioverter Defibrillator Patientswith Moderate and Severe Impairment of LeftVentricular Function

DAVENDRA MEHTA, SANJEEV SAKSENA, RYSZARD B. KROL,THAMPI JOHN, AMARKANTH SAXENA. RAMESH RAJU, RAJ KAUSHIK, andRAVINDRA KARANAM

From the Arrythmia and Pacemaker Service, Eastern Heart Institute, Passaic, and UniversityHospital, UMDNJ, Newark, New Jersey

MEHTA, D., ET AL.: Device Use Patterns and Clinical Outcome of Impiantabie Cardioverter DefibrillatorPatients with Moderate and Severe Impairment of Left Ventricular Function. The beneficial effects ofimplanted cardioverter de/ibriJiator (ICD) therapy in patients with malignant ventricular tachyarrhyth-mias and variable degrees 0/ left ventricular (LV) dys/uncfion are debated. ICD use and patient survivalwere examined in 128 patients with malignant ventricular arrhythmias and moderate or severe LV dys-function. Group r included 64 patients with moderate LV dys/unction (LV ejection fraction of > 30%) andgroup U, 64 patients with severe LV dys/unction fLV ejection/racfion o/< 30%). Foliow-up period rangedfrom ] to 78 months. The two groups were similar in age, incidence 0/ coronary artery disease andpresenting arrhythmia. The mean LV ejection fraction in group I was 44% ± 8% and group II was 22%± 5% (P < 0.0001}. Ai 4 years of follow-up, 66% of patients from group I and 62% from group II (P = NS)had ICD activation for presumed ventricular tachyarrhythmia. Survival was calculated using actuarialanalysis. Arrhythmic or sudden death mortality at 4 years o//ol/ow-up was 4% in group I and 7% ingroup II (P = NSj. Cardiac mortality was/or group I, 7% (P < 0.05}, 12% fP < 0,01 j , 15% (P < 0.01}, and15% (P < 0.01} for follow-up years 1, 2, 3, and 4, respectively. For group II, cardiac mortality was 27%.36%, 41%, and 41%/or/ollow-up years/or 1, 2, 3, and 4, respectively. The majority 0/cardiac deaths inboth groups ivas observed in the first 2 years of follow-up. However, in both groups, cardiac mortalitywas comparable in patients who did (users) and did not (nonusersj experience appropriate ICD shocks.Thus, the incidence of long-term ICD use is comparable in patients wifh moderate and severe LV dys/unc-tion. Cardiac mortality is higher in patients with severe LV dys/unction in the first 4 years of follow-upirrespective 0/ device use. The comparable long-term clinical outcome of ICD users and nonusers inpatients with moderate or severe LV dys/unction can be related to elimination 0/ arrhythmic mortality.Long-term patient survival in ICD recipients with severe LV dys/unction remains substantial even at 4years of follow-up. (PACE, Vol. 16. /anuary. Part II 1993)

defibriiiation. sudden death, left ventricular dys/unction

Introduction left ventricular (LV) function.^"^ Different studies

Sudden cardiac death presumably due to ma- [" l̂̂ ^*^ ^]^^* implantation of a cardioverter defi-lignant ventricular arrhythmias contributes signlf- bnllator device (ICD) in these patients is associ-

icantly to total mortality in patients with impaired ^^^^ ̂ ^^^ ̂ ^°^ '"'^'^^" ^^^^^ '^^^ ^^^'"8 f°l^°^-up." *• The extent of clinical benefits of long-termsurvival is unknown. It is also unclear whether

. . , ^ „ . . , utilization of ICD therapy varies with the degreeAddress for reprints: Davendra Mehta. M.D., Ph.D., Div. of Car- t t w j r • \~. n i i- -idiology, Mt. Sinai Medical Center. 1 Gustav Levy Pt., New ot LV dysfunction present. Controlled studies WithYork, NY 10029. Fax: [212) 369-3269. ICDs have been difficult to develop due to the ethi-

PACE, Vol. 16 January, Part II 1993 179

MEHTA, ET AL.

cal or logistic considerations for placebo controlsand a lack of agreement on satisfactory alternativetherapy for comparative control groups. Therefore,quantitation of benefit with ICD therapy is una-vailable. Thus, relative comparisons with patientstreated with alternative therapies or those whohave no further sudden cardiac events are beingexamined.

Mortality components potentially favorablyaffected by the ICD include arrhythmic and sud-den death. We, therefore, hypothesized that afterimplantation of ICD clinical outcome of patientswith malignant ventricular tachyarrhythmias (sus-tained ventricular tachycardia or fibrillation) andimpaired LV function should be comparable in pa-tients who experience and those who do not expe-rience recurrent sustained ventricular arrhyth-mias. Arrhythmic and cardiac mortality were com-pared in patients who experienced ICD discharges(users) for presumed recurrent ventricular tachy-cardia and fibrillation and compared to those whodid not (nonusers). Patient outcome was also strat-ified based upon the extent of LV systolic dysfunc-tion as judged by LV ejection fraction mea-surement.

Patients and Methods

All patients included in this analysis hadsymptomatic drug-refractory ventricular tachycar-dia or ventricular fibrillation. Standard indica-tions were used for ICD implantation.^ One-hundred-twenty-eight consecutive patients whoreceived ICDs for recurrent ventricular tachycar-dia or ventricular fibrillation and were followed atour institutions were evaluated. LV function wasassessed prior to implantation of the device by car-diac catheterization and angiography or gated ra-dionuclide angiography. Anatomical and electro-physiological diagnosis was established preopera-tively in all patients by cardiac catheterization andelectrophysiological studies. ICDs were implantedusing previously described standard operative im-plantation techniques following intraoperativetesting for defibriliation thresholds.^ All but 13 pa-tients underwent a thoracotomy (median ster-notomy, left lateral, subxyphoid, or subcostal) forimplantation of epicardiai patches. In 13 patientsa nonthoracotomy lead system was implanted,^

The patient follow-up period extended fromthe date of implantation to March 1992. Implanta-tion patients were followed at 1 month and subse-quently every 2 months. In addition, they wereasked to report if they had experienced a shock.After replacement of the generator, the follow-uproutine was repeated. A clinical examination wasperformed and the generator interrogated at eachvisit. A log of symptomatic arrhythmia recur-rences and number of discharges counted by theunit was maintained. A shock was presumed to bein response to a sustained ventricular tachyar-rhythmia if there was: (1) electrocardiographicdocumentation of sustained ventricular tachycar-dia or ventricular fibrillation; (2) sudden onset ofsevere presyncope, syncope, or breathlessnessprior to shock delivery that was terminated by theshock; or (3) in asymptomatic patients, when clini-cal or electrocardiographic evidence of a supra-ventricular tachyarrhythmia was not present.

Mortality was calculated from the date of im-plantation and 30-day operative mortality was in-cluded for all actuarial survival analyses. Deathfrom any cause and explanation of the device wereconsidered as study end points. In patients whodied during follow-up, an attempt was made toestablish the cause of death from events surround-ing terminal illness, hospital records, the attend-ing physician, and findings from postmortem ex-amination. ICDs were interrogated after deathwhenever possible for shock counts and device ac-tivation. All deaths were evaluated for mechanismand classified as sudden arrhythmic, cardiac non-sudden, or noncardiac. Sudden death was definedas death within 1 hour of onset or abrupt changeof symptoms or unwitnessed death. Cardiac deathincluded sudden death, tachyarrhythmic nonsud-den death with tachyarrhythmia persisting formore than 1 hour before death, and death fromother causes of cardiac origin as determined by theelectrophysiologist.

Statistical Methods

All described data are presented as mean ±standard deviation. Presentation of the mortalitydata is actuarial. Survival analysis was performedusing the Kaplan-Meier method.^" Differences in

180 January, Part II 1993 PACE, Vol. 16

IMPLANTABLE CARDIOVERTER DEFIBRILLATORS AND LEFT VENTRICULAR DYSFUNCTION

continuous variables were computed using un-paired t-test and groups were compared using Ghi-square test. Differences were considered statisti-cally significant at P < 0.05.

Results

Median follow-up period in these 128 patientswas 38 months (range 1-79). Their mean age was62 years (range 24-78) and 103 were males. Un-derlying cardiac disease was coronary artery dis-ease in 92 patients, cardiomyopatby in 25 patients,hypertensive heart disease in 5 patients, valvulardisease in 4 patients, and 2 patients did not haveany clinical evidence of an underlying cardiac ab-normality. Presenting ventricular arrhythmia wasventricular tachycardia in 101 patients and ven-tricular fibrillation or sudden cardiac death in 27patients. Patients were categorized into twogroups. Group I included 64 patients witb LV ejec-tion fractions of > 30% and group II, 64 patientswitb LV ejection fraction of < 30%. The clinicalcharacteristics of these two groups are delineatedin Table I.

Over the follow-up period of 4 years, 42 pa-tients (65%) from group I and 40 patients (62%)from group II (P = NS) experienced ICD shock

therapy for presumed ventricular tachycardia orventricular fibrillation. In botb groups age, gender,incidence of coronary artery disease, and sponta-neous and induced arrhythmias were similar inusers and nonusers (Table II).

Arrhythmic and sudden deatb mortality at 4years of follow-up was 4% in group I (two patients;one death was bradycardia-related and the otherdue to premature battery depletion) and 7% ingroup II, P = NS (three patients; one death wasrelated to IGD system failure, another due torecorded failure to defibrillate, and the third dueto unknown causes) (Fig. lA) Gardiac mortalitywas significantly higher in group II (Fig. IB). Ingroups I and II at 1 year of follow-up, it was 7%and 27% (P < 0.05), at 2 years 12% and 36% (P <0.01), at 3 years of follow-up it was 15% and 41%(P < 0.01) and at 4 years 15% and 41% (P < 0.01),respectively,

Gardiac mortality in users as compared tononusers, in botb groups of patients, is shown inFigures 2A and 2B, respectively. There was no sig-nificant difference in cardiac mortality in eachgroup between the users and nonusers of the IGD(group I, users = 12% vs nonusers = 13%, P =NS; group II, users = 44% vs nonusers = 45%, P= NS) at 48 months of follow-up. Wben mortality

Comparison

Age (years)Gender (male:temale)LVEF (percent)Cardiac disease

CAD (n)CardiomyopathyOthers

Clinical ArrhythmiaVTVF/SCD

Induced arrhythmiaVTVF

Table 1.

of Demographic Data

Group 1

(n = 64)62 ± 11

47:1744 ± 8

45 (70%)98

51 (79%)13 (21%)

46 (72%)12 (21%)

and Clinical Variables

Group II

(n = 64)63 ± 10

56:821.2 ± 6

47 (73%)16

1

50 (78%)14 (22%)

47 (73%)15 (23%)

P

NSNS

< 0.0001

NSNS

NSNS

NSNS

CAD = coronary artery disease; LVEF = left ventricular ejection fraction; SCD = sudden cardiac death;VT = ventricular tachycardia; VF = ventricular fibrillation.

PACE, Vol. 16 lanuary, Part II 1993 181

MEHTA, ET AL.

Demographic Data

Users (% of the total Gp.)Age (years)Gender (male:female)LVEF (percent)Cardiac disease

CAD (n)CardiomyopathyOthers

Clinical ArrhythmiaVTVF/SCD

Induced arrhythmiaVTVF

NonusersAge (years)Gender (maie:female)LVEF (percent)Cardiac disease

CADCardiomyopathyOthers

Clinical ArrhythmiaVTVF/SCD

Induced arrhythmiaVTVF

Table II.

and Clinical Variables in

Group 1

n = 42 (65%)62 ± 1133:9

45 ± 8

32 (78%)55

34 (81%)8 (19%)

31 (74%)9 (21%)

n - 2263 ± 11

14:846 ± 9

1443

17 (77%)4 (23%)

15 (68%)4 (18%)

Users and Nonusers of ICD

Group II

n = 40 (62%)63 ± 12

37:322 ± 6

30 (75%)91

32 (80%)8 (20%)

29 (72%)10 (25%)

n = 2461 ± 9

19:520 ± 5

177

—

18 (75%)6 (25%)

18 (75%)5 (21%)

P

NSNS

< 0.0001

NSNS

NSNS

NSNS

NS

< 0.0001

NS

NSNS

NSNS

Abbreviations as in Table I.

in device users and non-users was compared ingroup I and II, cardiac mortality at 48 months offollow-up was significantly higher in group II bothin users (P < 0.01) and nonusers (P = 0.05] of theICD.

Discussion

The benefits of ICD therapy in patients withsustained ventricular tachyarrhythmias and im-paired LV function are still debated.^''^^ Previ-ously published data on survival of patients withventricular tachyarrhythmia and out-of-hospitalcardiac arrest treated with antiarrhythmic drugs

or surgical ablation suggest that LV function is anindependent predictor of recurrent events and sur-vival.̂ •^^•''* In patients with malignant ventriculararrhythmia and severely impaired LV function 3-year arrhythmia recurrence rates have been re-ported to be approximately 50%.^ In comparison,patients with ICDs have been reported to have verylow sudden death rates and better actuarial totalsurvival.^^ This may be accounted for by incompa-rable patient population or benefit from ICD ther-apy. The overall impact of improved sudden deathrates on total and cardiac mortality has yet to befully quantitated. Fogoros et al.," used device dis-charge rates to suggest a dismal prognosis for pa-

182 lanuary, Part 11 1993 PACE, Vol. 16

IMPLANTABLE CARDIOVERTER DEFIBRILLATORS AND LEFT VENTRICULAR DYSFUNCTION

64

64% Survival

da44

34

22

23

13

(LVEF '30%)

(LVEF -30%)

100%

&0%

60%

40%

20%

0%

100%

80%

6 0 % •

LVEF .30% (Oroup I)

LVEF -30% IQroup II)

20 30Follow-up (months)

so

20% -

B10 20 30

Follow-up (months)

Figure 1. Life table analysis using the Kaplan-Meier method showing actuarial survival fromarrhythmic (A] and all cardiac mortality (B) in patients wilh left ventricular ejection fractionsof > 30% (top trace) and < 30% (bottom (race). Arrhythmic mortality in the two groups was lowand not significantly different. However, all cardiac mortality was signi/icontJy higher in patientswith left ventricular ejection fractions af < 30% alJ along the foUow-up period. The panel abovethe graph shoivs the totaJ number of patients in each group at various stages of foJIoiv-up.

80%

80%

40%

20%

0%

4240

% Survival

•

3029

2717

p-0.001

LVEF

LVEF

S1 (LVEF *30%)

10 (LVEF i30%>

•30% «3roup 1)

<30% (Qroup II)

n •

100%

80%

80%

40%

2224% Survival

1315

\

"̂ N̂ ^- • - - -

8e

• \

- - \ -

^ LVEF

— 1 - LVEF

^ (LVEF

* (LVEF

0-0.05

• 30% (Oroup

•30% (Qroup

•30%)l30%)

1)

2)

10 20 30Follow-up (months)

30

B Follow up (months)

Figure 2. Life table analysis showing actuariaJ survival from all cardiac mortality in patientswho experienced presumahly appropriate fCD shock therapy (A) as compared to those who didnot (Bj. Format of the figures similar to Figure I. Survival from cardiac mortality at 3 and 4 yearsof follow-up was significantly higher in patients with Jeft ventricular ejection fractions of > 30%as compared to those £ 30% in users and nonusers of ICDs. However, the survival of deviceusers and nonusers was comparable in patients with either severe or moderate left ventriculardysfunction.

PACE, VoL 16 January, Part II 1993 183

MEHTA, ET AL.

tients with untreated recurrent sustained ventricu-lar arrhythmias and severe LV dysfunction (6%survival at 3 years). In the absence of a control/untreated group absolute, ICD benefit in terms ofsudden and total mortality will be unavailable. Al-ternatives such as drug treated patients, ICD im-plantees with and without ventricular tachycar-dia, and ventricular fihrillation recurrences haveto be used as comparative population.

The incidence of appropriate shocks over a2-5 year follow-up period has been reported to bein the range of 33%-7O%^*-'^^ It has also beensuggested that tbe period from device implant tofirst shock is shorter in patients witb lower LVejection fractions.^^ At 5-year follow-up the inci-dence of device use in patients with depressed LVfunction (LV ejection fraction of < 30%) was re-ported to be 45% as compared to 32% in patientswith better LV function [LV ejection fraction of >40%]. In comparison, incidence of ICD use washigher in the two groups reported by us (66% ingroup I and 62% in group II at 4 years of follow-up). This could be related to patient selection orthe frequency of concurrent use of antiarrhythmicdrugs after ICD implantation. A prior analysis ofICD activation and correlation with device logs inpatients with third-generation ICDs, wbich havestorage capabilities for treated arrhythmic eventsat our center, indicates that ICD shock therapy, asdefined in this study, is an acceptable indicator ofrecurrent ventricular arrhythmia for nonprogram-mable ICDs.̂ ^ In fact, use of prior symptoms mayactually severely underestimate recurrent ventric-ular tachycardia and ventricular fibrillation eventsin patients with programmable ICDs. From our ex-perience, the frequency of presumed ventriculartachyarrhythmia recurrences in the two groupswith varying degrees of LV dysfunction appearscomparable. Thus, benefits from sudden deathprevention by the ICD can be inferred to be likelyin both groups.

Sudden arrhythmic mortality in hoth groupswas low (five patients) and in at least two patientswas related to device malfunction. Observed car-diac mortality in the present series of patients withICDs and impaired LV function is in the lower halfof the range seen in previous reports. Initial re-ports showed a survival rate of 67%-87% at 2years of follow-up in a similar group of patientswith LV ejection fractions of < 30% as compared

to 96% in patients with ejection fractions of >30%." Recent reports with a longer follow-up pe-riod also indicate poorer prognosis in patientswith impaired LV function. A more recent studyon survival of patients with LV ejection fractionsof < 30% reported a 5-year survival rate of 96% forarrhythmic mortality and 60% from cardiac mor-tality.'' In all these reports, however, survival ratehas been reported in patients who did and thosewho did not have recurrent ventricular arrhyth-mias (delivery of shocks by ICDs taken as an indi-cator of recurrence of ventricular arrhythmia). Ax-tell et al.^ reported that 45% of their patients withsevere LV dysfunction had appropriate ICD shocksbut survival in users and nonusers of the ICD wasnot compared. In our report, 4-year survival fromcardiac mortality was 85% in patients with moder-ate LV dysfunction and 59% in patients with se-vere LV dysfunction. We ohserved a similar sur-vival in users and nonusers of the ICD. This indi-cates that ICD therapy in patients experiencingrecurrent sustained ventricular tachycardia wasassociated with a survival comparable to that ofpatients who did not have any subsequent sponta-neous recurrent ventricular tachycardia or fibrilla-tion.^^ This would suggest elimination of the ar-rhythmic death component of the total mortalityin these patients. Since device use was similar inpatients with severe LV dysfunction as comparedto those with lesser degrees of LV dysfunction,users in both groups appear to have derived a ben-efit with respect to reducing hospitalization anddeath from symptomatic recurrent ventriculartachycardia and/or cardiac arrest. However, mor-tality was higher in patients with poor LV functionirrespective of ICD shock therapy experience andit was similar in patients who did and did not haveICD shocks. The influence of this clinical variableon patient survival has been well documented inother cardiac populations such as following coro-nary artery hypass surgery and in patients withcongestive heart failure.^°

On the basis of our experience, we would sug-gest that patients with moderate as well as severeLV dysfunction derive benefit from ICD therapy asjudged by device use patterns, that patients withand without ventricular tachycardia/fibrillationrecurrences after ICD implant have comparablesurvivals.

184 lanuary. Part II 1993 PACE, Vol. 16

IMPLANTABLE CARDIOVERTER DEFIBRILLATORS AND LEFT VENTRICULAR DYSFUNCTION

References

Franciosa JA, Wilen M. Ziesche S, et al. Survivalin men with severe chronic left ventricular failuredue to either coronary heart disease or idiopathicdilated cardiomyopathy. Am J Cardiol 1983; 51:831-83B.Swerdlow CD, Winkle RA, Mason JW. Determi-nants of survival in patients with ventriculartachyarrhythmias. N Engl J Med 1983; 308:1436-1442.Bigger JT Jr, Fliess JL, Kleiger R, et al. The relation-ships among ventricular arrhythmias, left ventric-ular dysfunction and the mortality in the 2 yearsafter myocardial infarction. Circulation 1984; 69:250-258.Fogoros RN, Elson JJ, Bonnet CA, et al. Efficacy ofthe automatic implantable cardioverter-defibrilla-tor in prolonging survival in patients with severeunderlying heart disease. J Am Coll Cardiol 1990;16:381-386.Tchou PJ, Kadri N, Anderson J, et al. Automaticimplantahte cardioverter defihrillators and sur-vival in patients with left ventricular dysfunctionand malignant ventricular arrhythmias. Ann InternMed 1988; 109:529-534.Axtell K, Tchou PJ, Akhtar M. Survival in patientswith depressed left ventricular function treatedwith implantable defihrillator. PACE 1991; 14:291-296.Lehmann MH, Saksena S, and the NASPE WritingCommittee. NASPE Policy Statement: Implantablecardioverter defibrillators in cardiovascular prac-tice: Report of the policy conference of the NorthAmerican Society of Pacing and Electrophysiol-ogy. PACE 1991; 14:969-979.Thurer RJ, Luceri RM, Bolook H. Automatic Im-plantable Cardioverter Defibrillator: Technique ofimplantation and results. Ann Thorac Surgery1986; 42:143-147.Saksena S, Tullow NG, Krol RB, et at. Initial clini-cal experience with endocardial defibrillationusing an implantahle cardioverter/defihrillatorwith a triple-electrode system. Arch Intern Med1989; 149:2333-2339.

10. Kaplan BL, Meier P. Nonparametric estimation ofincomplete observation. J Am Stat Assoc 1958; 53:457-462.

11. The Cardiac Arrhythmia Suppression Trial Inves-tigators. Preliminary report: Effect of encainideand flecainide on mortality in randomized trial ofarrhythmia suppression and myocardial infarc-tion. N Engl J Med 1989; 321:406-412.

12. Furman S. AICD benefit (editorial). PACE 1989; 12:399-400.

13. Wilher DJ, Garan H, Finkelstein D, et al. Out-of-hospital cardiac arrest: Use of electrophysiologictesting in the prediction of long-term outcome. NEngl J Med 1988; 318:19-24.

14. Lawrie GM, Pacifico A. Kaushik R, et al. Factorspredictive of results of direct ablative operationsfor drug-refractory ventricular tachycardia: Analy-sis of 80 patients. J Thorac Cardiovasc Surg 1991;73:1239-1247.

15. Winkle RA, Mead RH, Ruder MA, et al. Ten yearexperience with implantahle defibrillators. Circu-lation 1991; 84(Suppl II]:II-426.

16. Myerburg RJ, Luceri RM, Thurer R, et al. Time tofirst shock and clinical outcome in patients receiv-ing an automatic implantable cardioverter-defi-hrillator. J Am Coll Cardiol 1989; 14:508-514.

17. Kim SG, Fisher JD, Choue CW, et al. Influence ofleft ventricular function on outcome of patientstreated with implantahle defibrillators. Circulation1992; 85:1304-1310.

18. Mehta D, Krol RB, Saksena S. Diagnostic use ofsignificant symptoms underestimates appropriateventricular tachycardia reversions by programma-ble pacemaker defibrillator. (abstract) PACE 1992;15:531.

19. Mehta D, Saksena S, Krol RB, et al. Comparisonof clinical benefits and outcome in patients withprogrammable and non-programmable implanta-hle cardioverter-defibrillators. PACE 1992 (Inpress).

20. Passamani E, Davis KB, Gillespie MJ, et al. A ran-domized trial of coronary artery bypass surgery.Survival of patients with a tow ejection fraction. NEngl J Med 1985; 312:1665-1671.

PAGE, Vol. 16 January, Part II 1993 185