Drug therapy versus implantation of a cardiac defibrillator for the treatment of malignant arrhythmias in left ventricular dysfunction

Drug therapy versus implantation of a cardiac defibrillator for the treatment of malignant arrhythmias in left ventricular dysfunction Michael M. Mannino, MD, Davendra Mehta, MD, Noelle M. Langan, MD, and J. Anthony Gomes, MD New York, N.Y.

Heart disease is the leading cause of death in the United States. Probably the most feared complication of this disease is sudden cardiac death, which accounts for 350,000 to 400,000 deaths annually. Despite the large number of out-of-hospital cardiac arrests, only a small proportion of such patients will survive long enough to receive medical care and, of those, only a percentage will survive to hospital discharge. 1 Patients at particular risk of ventricular tachyarrhythmias and therefore sudden cardiac death (SCD) are those with congestive heart failure (CHF). An analysis of the Framingham data revealed an age-adjusted rate of SCD of 31.4 per 1000 for men with heart failure versus 5.7 per 1000 for men without heart failure (9.9/1000 and 1.8/1000, respectively, for women).2 Although revascularization, ablation, surgery, or transplantation are options for certain patients, most do not qualify for such procedures. Therefore essentially only two widespread options exist: arrhythmia suppression by using available drug therapy or implantation of a cardiac defibrillator (ICD). Medical therapy. Large studies such as CONSENSUS and V-HEFT have demonstrated the usefulness ofvasodilators, particularly the angiotensin-converting enzyme inhibitors, in reducing mortality in patients with poor ventricular function, a, 4 In addition, past studies have suggested a possible role for beta blockers in reducing cardiac mortality in these patients.5, 6 However, neither of these agents has gained acceptance for sustained ventricular arrhythmias. As for the available class 1 agents, several From the Division of Cardiology, Mount Sinai Hospital and Medical Center. Received for publication June 23, 1995; accepted Aug. 1, 1995. Reprint requests: Davendra Mehta, MD, Division of Cardiology, Box 1030, Mount Sinai Hospital, One Gustave L. Levy Place, New York, NY 10029. Am Heart J 1996;131:1251-9 Copyright © 1996 by Mosby-Year Book, Inc. 0002-8703/96/$5.00 + 0 4/1/70085

studies involving their use in patients with poor left ventricular function as well as the results of the CAST (Cardiac Arrhythmia Suppression Trial) and ESVEM (Electrophysiologic Study Versus Electrocardiographic Monitoring) studies have shed light on the ineffectiveness and even potential harm of many of these agents. 7-9 Sotalol, a class III agent with beta-blocking activity, has demonstrated effective atrhythmia suppression in trials involving patients with sustained ventricular arrhythmias. The ESVEM trial in particular demonstrated the agent's superior efficacy over most other available antiarrhythmic agents. 9 Unfortunately, this trial did not specifically address the use of sotalol in patients with poor left ventricular (LV) function. In addition, a current trial of d-sotalol (sotalol devoid of its beta-blocking activity) in patients with prior myocardial infarction (SWORD) was recently halted because of excess mortality in the active treatment arm (3.9% vs 2%). The latter preliminary findings raise the interesting question regarding the role of the agent's betablocking activity in arrhythmia suppression. Amiodarone is also currently under much consideration as a possible alternative to ICD therapy in preventing sudden cardiac death. Although principally a class 3 antiarrhythmic, amiodarone also possesses properties of beta and alpha blockade as well as calcium-channel blockade and, to a small degree, fast sodium channel (class I) activity. Past studies, performed mainly in Europe, regarding the use of amiodarone (mostly in patients with nonsustained tachycardia) have been small or inadequately designed and have thus not produced data implying any significant mortality benefit of the drug. However, several large-scale studies (discussed below) are currently underway to assess the use of amiodarone in preventing sudden cardiac death in patients with malignant ventricular arrhythmias and reduced ejection fraction. In 1987 Neri et al.lo monitored 65 patients with 1251

1252

June1996

American HeartJournal

Mannino et al.

A

1"00k

0.°01 e.so

o.,o. o.o.

\,=._

0.500.45-

Celttml

0.40

9'0

1~0 2~0 3~0 4~0 5~0 Days from randomisation

1.00~ B

8~0 720'

C

Log-ranktest p=0.16 ~Log-rank test p=O.16

o.95] ~

1\ l O.701

Control

Daysfromrandomisation

Fig. 1. Percent

survival in patients treated with amiodarone versus control (GESICA trial) in terms of total mortality (A), death from progressive heart failure (B), or sudden death (C).

dilated cardiomyopathy and complex ventricular arrhythmias by ambulatory electrocardiographic monitoring, 41 of whom received amiodarone (200 to 400 mg daily). During a mean follow-up of 3 years, there were 18 deaths: 10 patients in the group who received amiodarone (all from progressive CHF) and 8 patients who did not receive the agent (4 from sud-

den death and 4 from progressive CHF). Side effects were common (23 of 41 patients), but only four patients required amiodarone discontinuation. A similar small study by H a m e r et al. n tracked 34 patients on low-dose amiodarone or placebo for an average of 2 years. Their findings were not unlike Neri et al. in that no episodes of sudden cardiac death were noted in the patients receiving amiodarone (6 deaths, 2 noncardiac and 4 pump failure), whereas 5 of 7 deaths in the placebo group were attributed to sudden arrhythmic deaths. Cleland and Dargie 12 in 1988 reported their study involving 132 patients receiving conventional therapy with or without lowdose amiodarone (200 mg) in a nonrandomized fashion. After 21 months of follow-up, total mortality in the study population was 41% (63 patients) of which 47 (75% of deaths) were deemed sudden, and only 5 patients were believed to have died from progressive heart failure. Estimated survival curves of the data revealed that amiodarone reduced the incidence of sudden death from 45% to 15%. Nicklas et al.13 studied 101 patients with ejection fractions <30% and nonsustained ventricular tachycardia on ambulatory monitoring receiving either low-dose amiodarone (200 rag/day) or placebo for 1 year. While the frequency of complex arrhythmias on ambulatory monitoring was significantly diminished in those patients receiving active therapy, unlike the previous studies, there was no improvement in mortality or the incidence of sudden cardiac death (26% total mortality on amiodarone vs 19% on placebo). Twenty-one patients died suddenly during the 1-year follow-up period, of which 12 were receiving amiodarone and 9 placebo. The GESICA (Grup0 de Estudio de la Sobrevida en la Insuficiencia Cardiaca en Argentina) study recently published its 2-year experience with 516 patients with CHF receiving either amiodarone (300 rag/day) and standard CHF therapy or standard therapy alone. 14 An intention-totreat analysis revealed an overall mortality reduction of 28% (33.5% vs 41.4%) in patients treated with amiodarone compared with standard therapy alone (p = 0.024). A subanalysis of the data revealed a 27% relative reduction in SCD and a 23% relative reduction in deaths from CHF (both nonsignificant trends, p = 0.16) in patients treated with amiodarone compared with conventional therapy (Fig. 1). Side effects and drug discontinuation were uncommon (<10%). It is important to note that arrhythmias were not included as an entry criteria for these patients; in fact, sustained ventricular tachycardia or ventricular fibrillation were reasons for exclusion. The findings of the CHF-STAT Veterans trial were announced at the 1994 American Heart Association

Volume 131, Number 6

Mannino et al.

American Heart Journal

1253

100

75P e r

c

50

%

e 11

t P-,007 by Log Rank Statistic

25

~Amiodarone, 0

No113

--

Conventional,

N-116

I

I

I

I

I

I

1

2

3

4

5

Amiodarono

91%

82%

76%

66%

63%

6 53%

Conventional

77%

69%

56%

52%

46%

40%

Years

0

Fig. 2. Total cardiac events in patients receiving amiodarone versus conventional therapy (CASCADE trial). (From the CASCADE investigators. Am J Cardiol 1993;72:280-7.)

meetings. This trial involved the use of amiodarone or placebo in 674 patients with poor ventricular function and asymptomatic nonsustained ventricular arrhythmias. Their findings in several hundred patients revealed that despite significant arrhythmia suppression by Holter monitoring, there was no advantage to the amiodarone group regarding either sudden death or total cardiac mortality (death rate for both groups: 17%, 28%, and 42% at 1, 2, and 3 years respectively). All these studies have demonstrated conflicting evidence for a benefit of amiodarone in reducing the incidence of sudden death in patients with complex, u n s u s t a i n e d ventricular arrhythmias. Most appear, however, to lack either sufficient power or design to properly evaluate for a significant effect of the medication, especially in those patients with s u s t a i n e d malignant arrhythmias. Other studies such as the CAMIAT (Canadian Amiodarone Myocardial Infarction Arrhythmia Trial), EMIAT (European), and SSD (Spanish Study on Sudden Death) studies are underway to assess the effect of amiodarone on mortality; however, these involve mainly post-myocardial infarction patients with nonsustained arrhythmias. The CASCADE (Cardiac Arrest in Seattle: Conventional vs. Amiodarone Drug Evaluation) study did address the use of amiodarone in patients with malignant arrhythmias (ventricular fibrillation), i5 This study randomized 228 patients (45% with CHF) to either treatment with amiodarone or conventional antiarrhythmic medication (including quinidine,

procainamide, mexilitine, combination therapy, or flecainide). After 6 years of follow-up, while the overall event rates (cardiac death, ventricular fibrillation arrest, sustained arrhythmias, ICD shocks [up to 50% of patients had ICDs] and syncope with an ICD shock) were high (57% at 4 years and 70% at 6 years), the patients treated with amiodarone were significantly less likely to have an event than those receiving conventional therapy (22% vs 48% at 2 years, 48% vs 64% at 4 years, and 59% vs 80% at 6 years for amiodarone and conventional therapy, respectively) (Fig. 2). However, side effects were common and 29% of patients taking amiodarone discontinued therapy, mainly because of pulmonary and thyroid toxicity. The implantable defibrillator. Regarding SCD, survival in patients with poor left ventricular function receiving ICDs is comparable to those with good ventricular function, usually exceeding 90% at 1 year.i6, i7 However, in the majority of patients, much of the survival benefit as a result of the ICD is seen in the first few years after implantation, when abortion of SCD results in an extended life. It is after this time that other causes of death emerge as the natural history of the patients' disease is altered. These deaths are reflected in the total mortality data (Fig. 3). Winkle et aI.,is in a study of 270 patients, reported a sudden death rate of 4% over a 3-year period in which total mortality for the same period was 20%. Kim et al., 19in a recent study of 377 patients, showed an actuarial survival rate from sudden death of 99% at both 1 and 3 years for patients with ejection frac-

June 1996

1254

M a n n i n o et al.

American Heart Journal

% survival 80%-

........ ~

•

60%

....... : ..............

40%

........................

~---~

:- -~-->~=-

:

~'~..,.

:

---:'?---~"~~~"?~

.................................. Total

20% .

.

.

.

.

.

0% 03

.

.

.

.

.

.

.

.

.

,

,

6

9

.

.

.

.

.

.

,

.

.

.

.

.

.

.

.

,

.

.

.

.

.

,

.

.

.

,

.

.

.

.

.

-:~-Arrhythmic

.

,

,

,

r

~

C a r ~ iac,

,

1 2 1 5 1 8 21 2 4 2 7 3 0 3 3 3 6 3 9 4 2 4 5 4 8

Follow up (months) Fig. 3. Actuarial survival from arrhythmic, cardiac, and total mortality of patients with malignant ventricular arrhythmias, LV dysfunction, and implanted cardioverter-defibrillators.

Table I. Benefits of ICD therapy on arrhythmic and cardiac mortality in patients with ventricular arrhythmias and left ventricular dysfunction (LVEF <30%) Follow-up Period 1 1 3 3 3 3 4

yr yr yrs yrs yrs yrs yrs

Arrhythmic mortality

Cardiac mortality

Reference

9% 2% 11% 4% 8% 7%

16% 33% 38% 20% 28% 40%

M y e r b e r g et al. 4° Kim et al. 31 Fogoros et al. 17 G r i m m e t a ] . 2° Winkle et al. is Kim et al. 19 M e h t a et al. 3s

LVEF, Left ventricular ejection

fraction.

tions >30% and 98% 1-year and 92% 3-year survival for those with an ejection fraction <30%. Regarding total mortality, the survival rates were 94% at 1 year and 85% at 3 years for patients with an ejection fraction >30%, with corresponding values of 84% and 72% for those with ejection fractions <30%. Another study by Grimm et al, 2° involving 241 patients demonstrated a total mortality of 38% at 3 years with only an 11% sudden death rate (Table I). While the dramatic reduction of sudden cardiac death by ICDs has become accepted, the total cardiac mortality rates remain disappointingly high. Therefore the relative survival benefit of the ICD is likely to be early after implantation in patients with poor ventricular function and may not be sustained. The overall survival rate is thus improved in the short run but will dissipate in the long run as the dominant mode of death becomes non-sudden. This phenome-

non of altering the rate of SCD but having total mortality remained unchanged has been termed "conversion" of the mode of death. Newman et al., 21 in their analysis of 60 recipients ofamiodarone and ICD and 120 historical controls, found that although actuarial survival curves showed a significant difference between the patients with ICD and controls (89% vs 72% at 1 year and 65% vs 49% at 3 years), the survival advantage afforded by the ICD appeared to dissipate over time, with the curves converging at 40 months. Several explanations exist for the disappointing reduction in total cardiac mortality compared with the impressive reduction in sudden deaths in patients receiving ICDs. First, sudden death comprises only a fraction of the total cardiac deaths in patients with cardiac arrhythmias, accounting for a larger proportion of deaths in patients with better New York Heart Association functional classes. Patients with poorer functional classes and lower ejection fractions tend to suffer more deaths from progressive heart failure than the above group. In fact, despite studies analyzing multiple parameters in patients with poor ventricular function, ejection fraction remains the most powerful predictor of cardiac mortality. 22 Perhaps an explanation for the large proportion ofnonarrhythmic deaths manifested by these patients with ICDs is that their arrhythmias are not necessarily markers for impending SCD but are merely a manifestation of poor left ventricular function and impending terminal left ventricular failure. 22 Unpublished data from our own institution supports these theories (Table II). In a group of

Volume 131, Number 6

AmericanHeartJournal

Mannino et al.

1255

Table II, Relation between left ventricular ejection fraction

Table III. Problems associated with ICD implantation in

and CHF symptoms (NYHA class) on mortality

patients with LV dysfunction

Index

No. of patients

Cardiac death

Sudden death

EF -> 30% EF -< 30% No CHF CHF I and II CHF III and IV

208 161 184 117 48

11.5% 26.5%* 10% 20% 48%*

5.3% 9.3% 4% 6% 23%

Relation between ejection fraction (EF) and heart failure (CHF) symptoms in 369 patients. Incidence of both total cardiac mortality and sudden death appear to depend more on symptoms of heart failure than absolute ejection fraction. NYHA, New York Heart Association; EF, ejection fraction. *p < 0.05.

Relatively high defibrillation threshold Underlying cardiac disease--ischemia or fibrosis Increased left ventricular mass Decreased tolerance to induced ventricular fibrillation during testing (more likely to have higher threshold with longer time to shock) With standard epicardial patches relatively less myocardinm is shocked Higher perioperative mortality (with epicardial patches) Prolonged anesthesia required for thoracotomy* Perioperative heart failure in hemodynamically compromised patient* Postoperative infection Higher incidence of postoperative atrial and ventricular arrhythmias* *Less of a problem with endocardial/nonthoracotomy lead system

almost 400 patients monitored for up to 8 years (receiving either medical therapy, ICD, or surgery), the incidence of both sudden death and total cardiac mortality appears to be more a reflection of the underlying ventricular function (manifested by the presence or absence and degree of congestive failure) as opposed to the absolute ejection fraction. A second reason for the discrepancy between the sudden death and total mortality death rates is that the implantation of defibrillators, especially through a thoracotomy approach, is associated with a small but significant surgical mortality, especially in patients with poor ventricular function (Table III). 23, 24 In addition, there is literature to suggest that implantation of defibrillators through a thoracotomy approach may actually worsen or provoke more arrhythmias through pericardial or myocardial irritation. 25 Finally, many patients with heart failure are predisposed to other causes of death, which have been shown to contribute to the overall cardiac mortality to a significant degree. Luu et al. 26 reviewed the terminal rhythms of 21 stable patients with class III CHF awaiting discharge from the hospital. Although all cardiac arrests can be classified as sudden, only 38% were attributed to ventricular tachycardia or fibrillation (all of which occurred in those with ischemic cardiomyopathy). The most common cause of death, responsible for 62%, was bradycardia or electromechanical dissociation. The latter was responsible for all the deaths in those patients with idiopathic cardiomyopathy and 35% of patients with ischemia. Two patients died from coronary thrombus, one from a pulmonary embolus, and two from hyperkalemia. Other centers have similarly reported cases ofbradyarrhythmias, coronary thrombosis, and aortic dissection as other causes of "sudden" death in patients with cardiomyopathy. 27 Thus if one were to reason-

ably assume that 50% of deaths in patients with left ventricular dysfunction are sudden in nature and that approximately 80% of those are tachyarrhythmic in nature, a 90% reduction in such deaths would calculate into a 36% reduction in total mortality. When adding a 3% to 5% surgical mortality for implantation of the ICD, the overall reduction is reduced to 31%. 2s "Not so sudden" deaths and surgical mortality. A l -

though progressive heart failure is understood to be an important cause of nonarrhythmic death in patients with ICDs and poor ventricular function, much discussion has been generated regarding non-SCD related to arrhythmias, such as the surgical mortality associated with implantation and the "not so sudden" arrhythmic deaths described by Guarnieri et al. 29 The latter term is used to describe patients with internal defibrillators who were initially successfully defibrillated but died shortly after from incessant malignant arrhythmias. This phenomenon, it appears, m a y be more a reflection of the severity of the individual patients' underlying ventricular disease. 22 In support of this theory is the fact that the increased presence of complex ventricular arrhythmias on ambulatory monitoring in patients with poor ventricular function has been shown to predict not those dying from sudden death, but total cardiac mortality. 12,22 In addition, as heart failure progresses, the prevalence of nonsustained ventricular tachycardia increases to a significantly greater degree than does the incidence of sudden death. 22, ao Finally, most studies assessing the use of antiarrhythmic agents have found that despite the suppression of arrhythmic activity on ambulatory monitoring, the incidence of sudden death remains high in this population, s

June 1996

1256

Mannino et al.

In addition to the non-sudden arrhythmic deaths, surgical mortality is an important contributor to mortality in patients receiving ICDs. Kim et al.31 examined the influence of left ventricular function on the outcome of patients with implanted deflbrillators. Their study of 68 patients, implanted through a thoracotomy approach, found that the incidence of SCD was not significantly different between those with ejection fractions >30% and those <30% at 1, 2, and 3 years of follow-up. However, surgical mortality related to implantation of the defibrillator was 11% for those with poor left ventricular function compared with 0% for those with an ejection fraction >30%. Consequently, the arrhythmia-related death rate, or those deaths caused by sudden death, surgical mortality-related arrhythmias, and non-sudden death (not prevented by amiodarone and ICD), was higher for those with low ejection fractions than for those without (30% vs 3% at 3 years). Concomitant surgery did not influence the surgical mortality; however, that m a y just be a reflection of the small number of patients undergoing concomitant bypass surgery in this study. A subsequent larger study by Kim et al. 19 reporting the Cleveland Clinic Foundation and Monteflore Medical Center experience also looked at the influence of left ventricular function on survival and mode of death in patients receiving an ICD. This study of 377 patients compared rates of sudden death, surgical mortality, total arrhythmia-related death, and total cardiac death in patients with an implanted device and ejection fractions of >30% to those with a device and ejection fractions <30%. The incidence of sudden death was extremely low in all patients receiving the ICD, survival being 98% for both groups at i year and 95% and 90% at 5 years for patients with ejection fractions >30% and <30%, respectively. Surgical mortality for the groups was improved from their previous study, probably representing improved surgical technique and a larger patient population adding to the statistical power. The surgical mortality was 3.9% in all: 1.8% in patients with ejection fraction >30% and 7% in those with ejection fraction <30%. Also, unlike their previous report, concomitant bypass surgery did influence the surgical mortality in an unfavorable m a n n e r (9% with vs 2.2% without for all patients). Cardiac mortality, as expected, differed depending on ejection fraction (19% for patients with ejection fraction >30% at 3 years, 28% in those with an ejection fraction <30%), with arrhythmia-related deaths constituting 63% of all deaths in the latter group. These findings of surgical mortality in the 7% to !1% range for patients with low ejection fractions

American Heart Journal

appear excessively high compared with other studies that have reported rates of 1.5% to 5.4%, 32, 33 although Gohn et al.34 found similar elevated mortality rates for patients with low (<30%) ejection fractions compared with those with ejection fractions >30%. A recent study however, comparing the newer transvenous (nonthoracotomy) leads to the traditional thoracotomy approach (epicardial leads), demonstrated an improved survival by the transvenous method. 35 The PCD Investigator Group examined the safety of the implantation procedure and clinical outcome of 1221 patients receiving either a thoracotomy approach with epicardial leads (616 patients) or a nonthoracotomy approach with endocardial leads (605 patients). The nonthoracotomy approach was associated with a significant perioperative survival advantage--0% versus 3.4% for patients with ejection fraction >30% and 2.1% versus 5.2% for those with ejection fractions <30%. When perioperarive events were excluded in their analysis, both cardiac and noncardiac mortality were comparable 3 years after follow=up. Interestingly, when one looks at the cause of death immediately after implantation of the transthoracic devices, it appears that arrhythmias account for a significant proportion, although infection and worsening heart failure also contribute. Veltri et a l . 36 r e ported a 50% incidence of death from "incessant v e n t r i c u l a r arrhythmias" during the postoperative period. Similarly, Gohn et al.34 reported that 58% of postoperative deaths were from arrhythmic complications. In the study by Kim et al., 25 two thirds of the early deaths were related to intractable arrhythmias. Unfortunately, the PCD group did not elaborate on the early perioperative causes of death. Thus from these data, the reduction in mortality by the transvenous method appears likely from, at least in part, the fact that epicardial electrodes are no longer needed and therefore spare the patient from the risks of a thoracotomy as well as pericardial and myocardial irritation and their potential for arrhythmia provocation. Although this exacerbation, when it occurs, is transient in m a n y patients, in some it may have long-term effects on the natural course of arrhythmia occurrence and may cause more frequent recurrences during the follow-up period than without ICD. 25 Crandall et al., 37 in their study of 194 consecutive patients treated with or without an ICD (thoracotomy approach) for an out-of-hospital sudden death and noninducibility by programmed electrical stimulation, suggest such a possibility. Although overall survival rates for the two groups were similar, the arrhythmia event rate of fatal and nonfatal recurrences was significantly higher (30% vs

Volume 131, Number 6

Mannino et al.

American Heart Journal

1257

% survival 1 oo%~

...............................................

40%

..................................................

20%

........................................

I ~LVEF>30% USERS

I ~/.,,:s~.;.3g,."O,,.OS,/RS " " ~LVEF<30% USERS O LVEF<30% NON-USERS

0% 0

~ 5

, 10

I 15

, 20

I 25

Follow

up

(months)

/~ 30

I 35

, 40

Fig. 4. Comparison of cardiac mortality rate in patients with good (>30%) and poor (<30%) left ventricular ejection fraction (LVEF) with reference to "appropriate" use of device. Users are patients who had appropriate ICD shocks; non-users did not receive any ICD shocks. No significant difference in cardiac mortality between users and non-users in patients with good and those with poor left ventricular function. However, beyond 32 months, cardiac mortality was significantly higher in patients with impaired left ventricular function both in users and non-users of ICD. (From Mehta et al. Am Heart J 1992;124:1608-14.)

12% at 2 years) in those receiving ICDs than those treated without. Shock occurrence a n d s u r v i v a l b e n e f i t . Regarding ICD discharge, Mehta et al., 3s in a study of 112 patients with drug refractory ventricular tachycardia or fibrillation, found that the rate of amiodarone and ICD discharge at 4 years was similar for patients with both low (<30%) and high (>30%) left ventricular ejection fractions. The 5-year cardiac mortality in that study, however, was higher in patients with lower ejection fractions (40% vs 12%), with no difference in mortality between those who received "appropriate" shocks and those who did not (Fig. 4). Grimm et al., 2° in reviewing the rate of discharge of 241 patients with ICDs, found the frequency of administered shocks to increase significantly with time--occurring in 15% of patients at 1 year and up to 76% at 5 years. This observation is in keeping with other authors and m a y suggest that the true benefit of these devices m a y not be apparent for several years. Unfortunately for those patients with poor ventricular function, this time frame may exceed their life expectancy. In addition, Grimm et al. also found that the survival among patients receiving spontaneous ICD shocks was not significantly different from those without, the only predictor of total mortality being a low ejection fraction. Forgoros et al. 39 analyzed the impact of the ICD on prolonging life measured from first shock. When comparing re-

sults in terms of life ventricular function, a significant difference between the groups was observed, with the <30% ejection fraction group surviving a mean of 11 ___14 months after initial shock compared with a mean of 42 +_ 9 months for those >30%. Myerberg eta]., 4° tracking 60 patients over a 25-month period, found in their study that the mean time to death was 14.3 _+ 13.1 months for patients receiving any prior shocks and 16.7 _+ 11.5 for those receiving no shocks. The mean time to first shock for those patients who subsequently died was 2.5 _+ 2 months, thus making the mean time to death after the first shock 11.8 months. These observations are mainly made on the assumption that a patient with a defibrillator would have died at the time of the first appropriate shock if the patient was without the ICD. This method of survival analysis, however, has important limitations because a significant proPortion of shocks are not related to potentially fatal arrhythmia. On the basis of available information, is it possible to determine the appropriateness of delivered ICD therapies? Studies with the aid of the newer third-generation devices exist to suggest that many of the arrhythmias that trigger the ICDs to fire may spontaneously terminate before ICD therapy. 41 In addition, the possibility exists that the newly available pacing-capable devices may convert stable, non-lifethreatening ventricular or even supraventricular

1258

June 1996 American Heart Journal

Mannino et al.

arrhythmias into unstable arrhythmias, requiring a discharge from the ICD. 42 The next step. Controlled studies are currently underway to possibly answer the questions of appropriateness and benefit of ICD therapy when delivered. The Multicenter Automatic Defibrillator Implantation Trial (MADIT) aims at comparing conventional therapy with ICD implantation in patients with impaired left ventricular function, coronary artery disease with nonsustained ventricular tachycardia, and inducible sustained ventricular tachycardia. The Coronary Artery Bypass Graft (CABG) Patch trial is investigating the "prophylactic" use of defibrillators in patients with a low ejection fraction and positive signal-averaged electrocardiogram who are scheduled for bypass surgery. Unlike previous studies, these trials have included only patients without a history of sustained tachyarrhythmias. Several large trials are currently underway to answer the question of whether amiodarone truly alters the incidence of SCD in this population. The Antiarrhythmics Versus Implantable Defibrillators (AVID) study is a National Institutes of Health multicenter prospective study comparing conventional antiarrhythmic drug therapy with ICDs in patients surviving cardiac arrest or sustained ventricular tachycardia. Drug therapy will consist of sotalol (through electrophysiologic and Holter guidance) or amiodarone. The ICDs will be of the third generation through a transvenous approach when possible. The Cardiac Arrest Study in Hamburg (CASH) is also investigating the use of drug therapy versus ICD implantation. This trial compares amiodarone, metoprolol, or propafenone (latter arm discontinued for excessive mortality) to an ICD through a transvenous or thoracotomy approach. To date, no differences in mortality rates have been seen between the ICD and other drug groups or between amiodarone and metoprolol. Sudden death, however, is significantly reduced with the ICD compared with medical therapy. 43 The Canadian Implantable Defibrillator (CIDS) study also compares amiodarone with ICD implantation. This trial has continued to proceed for more than 2 years without premature termination, thus suggesting that any difference between the two therapies, if any, is unlikely to be dramatic. The primary end points for all the above studies are cardiac mortality and/or arrhythmic death. It should be noted that much discussion and literature has recently been devoted to standardizing mortality analysis and definitions of the causes of death in this patient population in hopes of better defining sudden deaths from other modes of death (a major criticism in many of the previous studies). 44

The treatment of ventricular arrhythmias in patients with cardiomyopathy continues to be a difficult problem. Although advances in medical therapy for patients with CHF have made long strides, arrhythmia management remains in its infancy. While ICD therapy appears to offer hope, it is not without its limitations. Thus the medical community must continue to search for, and debate over, the best treatment for this growing population of patients. REFERENCES

1. Lombardi G, Gallagher J, Gennis P. Outcome of out-of-hospital cardiac arrest in New York City. JAMA 1994;271:678-83. 2. Kannel WB, Plehn JF, Cuppies LA. Cardiac failure and sudden death in the Framingham study. AM HEARTJ 1988;115:869-75. 3. The CONSENSUS Trial Study Group. Effects of enalapril on mortality in severe congestive heart failure: results of the Cooperative North Scandinavian Enalapril Survival Study (CONSENSUS). N Engl J Med 1987;316:1429-35. 4. Cohn JV, Archibald DG, Zeische S, Fraciosca JA, Hartsen WE, Tristani FE, Dunkman WB, Jacobs W, Francis GS, Flohr KH, Goldman S, Cobb FR, Shah PM, Saunders R, Fletcher RD, Loeb HS, Hughes VC, Baker B. Effect ofvasodilator therapy on mortality in chronic congestive heart failure: results of a Veterans Administration cooperative study. N Engl J Med 1986;314:1547-52. 5. Pitt B. The role of beta-adrenergic blocking agents in preventing sudden cardiac death. Circulation 1992;85(suppl I)I-107-1L 6. Beta-blocker heart attack trial research group: a randomized trial of propranelol in patients with acute myocardial infarction. JAMA 1982; 247:1707-14. 7. Hallstrom AP, Cobb LA, Yu BH, Weaver WD, Fahrenbruch CE. An antiarrhythmic drug experience in 941 patients resuscitated from an initial cardiac arrest between 1970 and 1985. Am J Cardio11991;68:102531. 8. The Cardiac Arrhythmia Suppression Trial (CAST) Investigators. Preliminary report: effect of encainide and flecainide on mortality in a randomized trial of arrhythmia suppression after myocardial infarction. N Engl J Med 1989;321:406-12. 9. Mason JW, The ESVEM Investigators. A comparison of seven antiarrhythmic drugs in patients with ventriculartachyarrhythmias.N Engl J Med 1993;329:452-8. 10. Neri R, Mestroni L, Salvi A, Pandullo C, Camerini F. Ventricular arrhythmias in dilated cardiomyopathy: efficacy of amiodarone. AM HEARTJ 1987;113:707-12. 11. Hamer AW, Arkles LB, Johns JA. Beneficial effects of low-dose amiodarone in patients with congestive cardiac failure: a placebo-controlled trial. J Am Coll Cardiol 1989;14:1768-74. 12. Cleland JG, Dargie HJ, Ford I. Mortality in heart failure: clinical variables of prognostic value. Br Heart J 1987;58:572-82. 13. Nicklas JM, McKenna WJ, Stewart RA, Mickelson JK, Das SK, Schork A, Krikler SJ, Quain LA, Morady F, Pitt B. Prospective, double-blind, placebo-controlled trial of low-dose amiodarone in patients with severe heart failure and asymptomatic frequent ventricularectopy. AM HEART J 1991;122:1016-2L 1 4 . Doval HC, Nul DR, and the Grupo de Estudio de Sobrevida en Insuficiecia Cardiaca en Argentina (GESICA): Randomised trial of low-dose amiodarone in severe congestive heart failure. Lancet 1994;344:493-8. 15. CASCADE Investigators. Randomized antiarrhythmicdrug therapy in survivors of cardiac arrest. Am J Cardiol 1993;72:280-7. 16. Mirowskl M Reid PR Winkle RA Mower MM Watkins L Stinson EB Griffith LSC, Kallman CH, Weisfeldt ML. Mortality in patients with implanted automatic defibrillators. Ann Intern Med 1983;98:585-88. 17. Fogoros RN, Elson J J, Bonnet CA, Feidler SB, BurkholderJA. Efficacy of the automatic implantable cardioverter-defibrillatorin prolonging survival in patients with severe underlying cardiac disease. J Am Coll Cardiol 1990;16:381-6. 18. Winkle RA, Mead RH, Ruder MA, Guadani VA, Smith NA, Buch WS, Schmidt P, Shipman T. Long term outcome with the automatic im-

Volume 131, Number 6 American Heart Journal

19.

20.

21.

22.

23.

24.

25.

26.

27.

28. 29.

30.

31.

32.

plantable cardioverter-defibrillator. J Am Coll Cardiol 1989;16:135361. Kim SG, Maloney JD, Pinski SL, Choue CW, Ferrick KJ, Roth JA, Gross J, Brodmann R, Furman S, Fisher JD. Influence of left ventricular function on survival and mode of death after implantable defibrillator therapy (Cleveland Clinic Foundation and Montefiore Medical Center Experience). Am J Cardiol 1993;72:1263-7. Grimm W, Flores BT, Marchlinski FE. Shock occurrence and survival in 241 patients with implantable cardioverter-defibrillator therapy. Circulation 1993;87:1880-8. Newman D, Suave MJ, tterre J, Langberg JJ, Lee MA, Titus C, Franklin J, Scheinman MM, Griffin JC. Survival after implantation of the cardioverter-defibrillator. Am J Cardiol 1992;69:889-903. Packer M. Lack of relation between ventricular arrhythmias and sudden death in patients with chronic heart failure. Circulation 1992; 85(suppl)I50-6. Gartman DM, Bardy GH, Allen M, Misbach GA, Ivey TD. Short term morbidity and mortality of implantation of the automatic implantable cardioverter-defibrillator. J Thorac Cardiovasc Surg 1990;100:353-9. Mosteller RD, Lehmann MH, Thomas AC, Jackson K. Operative mortality with implantation of the automatic cardioverter-defibrillator. Am J Cardiol 1992;68:130-45. Kim SG, Fisher JD, Furman S, Gross J, Zilo P, Roth JA, Ferrick KJ, Brodman R. Exacerbation of ventricular arrhythmias during the postoperative period after implantation of an automatic defibrillator. J Am Cell Cardiol 1992;18:1200-6. Luu M, Stephenson WG, Stephenson LW, Baron K, Walden J. Diverse mechanisms of unexpected cardiac arrest in advanced heart failure. Circulation 1990;80:1675-80. Radhakrishnan S, Kaul U, Bahl VI~ Talwar KK, Bhatia ML. Sudden bradyarrhythmic death in dilated cardiomyopathy: a case report. PACE 1988;11:1369-72. Sweeney MO, Ruskin JN: Mortality benefits and the implantable cardioverter defibrillator. Circulation 1994;89:1851-8. Guarnieri T, Levine JH, Griffith LC, Veltri E. When "sudden cardiac death" is not so sudden: lessons learned from the automatic cardioverter-defibrillator. AM HEARTJ 1988;115:205-7. CalifRM, McKinnis RA, Burke J, Lee KL, Harrel FE, Behar VS, Pryor DB, Wagner GS, Rosati RA. Prognostic implications of ventricular arrhythmias during 24 hour ambulatory monitoring in patients undergoing cardiac catheterization for coronary artery disease. Am J Cardiol 1982;50:23-31. Kim SG, Fisher JD, Choue CW, Gross J, Roth JA, Ferrick KJ, Brodmann R, Furman S. The influence of left ventricular function on outcome of patients treated with implantable defibrillators. Circulation 1992;85:1304-10. Gartman DM, Bardy GH, Allen M, Misbach GA, Ivey TD. Short term morbidity and mortality of implantation of the automatic implantable cardioverter-defibriUator. J Thorac Cardiovasc Surg 1990;100:353-9.

MGnTtino et al.

1259

33. Mosteller RD, Lehmann MH, Thomas AC, Jackson K. Operative mortality with implantation of the automatic cardioverter-defibrillator. Am J Cardiol 1992;68:130-45. 34. Gotm D, Edel T, Pollard C, Firstenberg M, Moore S, Simmons T, Trohman R, Maloney J. Determinants of operative mortality in implantable cardioverter-defibrillator [abstract]. J Am Coll Cardiol 1991; 17(suppl):86A. 35. The PCD Investigator Group. Clinical outcome of patients with malignant ventricular tachyarrhythmias and a multiprogrammable implantable cardioverter-defibrillator implanted with or without thoracotomy: an international multicenter study. J Am Coll Cardiol 1994; 23:1521-30. 36. Veltri E, Mower MM, Guarnieri T, Mirowski M, Levine JH, Fisher A, Juanteguy J, Watkins L, Reid PR, Griffith LSC. Clinical efficacy of the automatic implantable cardioverter-defibrillator: six year cumulative experience [abstract]. Circulation 1986;74(suppl):II-109. 37. Crandall BG, Morris CD, Cutler JE, Kudenchuk PJ, Peterson JL, Liem LB, Broudy DR, Greene HL, Halperin BG, McAnulty JH, Kron J. Implantable cardioverter-defibrillator therapy in survivors of out-of-hospital sudden cardiac death without inducible arrhythmias. J Am Coll Cardiol 1993;21:1186-92. 38. Mehta D, Saksena S, Krol RB. Survival of implantable cardioverterdefibrillator recipients: role of left ventricular function and relationship to device use. AM HEARTJ 1992;124:1608-14. 39. Fogoros RN, Elson JJ, Bonnet CA, Feidler SB, Burkholder JA. Efficacy of the automatic implantable cardioverter-defibrillator in prolonging survival in patients with severe underlying cardiac disease. J Amer Coll Cardiol 1990;16:381-6. 40. Myerberg RJ, Luceri RM, Thurer R, Cooper DK, Zaman L, Interian A, Fernandez J, Cox M, Glicksman F, Castellanos A. Time to first shock and clinical outcome in patients receiving an automatic implantable cardioverter-defibrillator. J Am Coll Cardiol 1989;14:508-14. 41. Ellenbogen K, Luceri R, Dorian P, Brachman J, Uthr J, Beckman K, Singer I, Zipes D, Saksena S, Estes M, Franklin J and Phase-1 Clinical Investigators. Clinical evaluation ofimplantable cardioverter-defibriliator (ICD) function: importance of data logging to assess device efficacy. Circulation 1991;84:II-426. 42. Hook BG, Callans DJ, Kleiman RB, Flores BF, Marchilinski FE: Implantable cardioverter-defibrillator therapy in the absence of significant symptoms: rhythm diagnosis and management aided by stored electrogram analysis. Circulation 1993;87:1897-1906. 43. Siebels J, Kuck KH. Implantable cardioverter-defibrillator compared with antiarrhythmic drug treatment in cardiac arrest survivors (The CASH Study). AM HEARTJ 1994;127:1139-44. 44. Kim SG, Fogoros RN, Furman S, Connolly, SJ, Kuck KH, Moss AJ for the participants of the Policy Conference. Standardized reporting of ICD patient outcome: the report of a North American Society of Pacing and Electrophysiology Policy Conference, February 9-10. PACE 1993; 16:1-5.