Download Should everyone with an ejection fraction less than or equal to 30

CONTROVERSIES IN
CARDIOVASCULAR MEDICINE
Should everyone with an ejection fraction less
than or equal to 30% receive an implantable
cardioverter-defibrillator?
Downloaded from http://circ.ahajournals.org/ by guest on June 12, 2017
Not Everyone With an Ejection Fraction <30% Should
Receive an Implantable Cardioverter-Defibrillator
Alfred E. Buxton, MD
A
lthough the total number of deaths attributed to
cardiovascular disease has decreased during recent
decades, an increase in the percentage of cardiac
deaths occurring suddenly (from 56% in 1989 to 63% in
1998) has actually resulted in an increase in the total number
of sudden deaths in the United States.1 The problem of
sudden death spans all racial and cultural groups and is
equally important in men and women. It is now 40 years since
continuous ECG monitoring revealed that ventricular
tachycardia and ventricular fibrillation (VT/VF) are responsible for the majority of sudden deaths. Dissemination of this
information resulted in a marked change in practice. Until the
1970s, physicians treated only symptomatic VTs. Therapeutic
options were limited to a few antiarrhythmic drugs, administered empirically. With the advent of telemetric ECG
monitoring and continuous ambulatory (Holter) ECG monitoring came the recognition that frequent and multiform
ventricular ectopy often preceded episodes of VF in the acute
phase of myocardial infarction (MI). Ventricular ectopy and
nonsustained VT was documented 1 to 2 weeks after MI in
many patients. The presence of frequent ventricular premature complexes and runs of nonsustained VT in patients with
significant left ventricular dysfunction were then demonstrated to be an independent predictor of sudden cardiac death
late after MI. This information then led to the practice of
administering antiarrhythmic drugs to suppress ventricular
ectopy under the presumption that suppression of asymptomatic ectopy would prevent sudden death. This fashion lasted
almost 20 years, until the results of the Cardiac Arrhythmia
Suppression Trial (CAST) trial demonstrated the harm caused
by this approach.2 Thus, a sea change in the practice of
cardiology had occurred, from treating only symptomatic
patients to therapy aimed at primary prevention of sudden
death. The initial efforts using antiarrhythmic drugs failed to
reduce mortality.
While many physicians in the 1970s were administering
prophylactic drugs to survivors of MI, Michel Mirowski was
developing the implantable cardioverter-defibrillator (ICD),
introduced to clinical medicine in 1980.3 It took more than a
decade before randomized, controlled trials evaluating the
ability of ICDs to reduce mortality in survivors of cardiac
arrest or sustained VT appeared in print. These studies
demonstrated the superiority of ICDs over amiodarone and
other pharmacological agents for secondary prevention of
sudden death in survivors of cardiac arrest.4 –7
After publication of the secondary-prevention studies, 3
trials were published demonstrating that ICDs could reduce
mortality in patients with chronic coronary disease and
The opinions expressed in this article are not necessarily those of the editors or of the American Heart Association.
From the Cardiology Division, Brown Medical School, Providence, RI (A.E.B.), and the Cardiology Unit, Department of Medicine, University of
Rochester Medical Center, Rochester, New York (A.J.M.).
Correspondence to Alfred E. Buxton, MD, Professor of Medicine, Brown University Medical School, 2 Dudley St, Suite 360, Providence, RI 02905 (e-mail
[email protected]) or Arthur J. Moss, MD, Heart Research Follow-up Program, Box 653, University of Rochester Medical Center, Rochester, NY
14642 (e-mail [email protected]).
(Circulation. 2005;111:2537-2549.)
© 2005 American Heart Association, Inc.
Circulation is available at http://www.circulationaha.org
DOI: 10.1161/01.CIR.0000165057.88551.2C
2537
2538
Circulation
May 17, 2005
Relationship of 1-year mortality to EF after MI. The black columns represent mortality reported by the Multicenter Postinfarction Research Group in 198311; the gray columns represent
mortality documented by CIMS in 1996.12 Reprinted from Reference 12, Copyright 1996, with permission from the American
College of Cardiology Foundation.
Downloaded from http://circ.ahajournals.org/ by guest on June 12, 2017
abnormal left ventricular systolic function that had not yet
developed spontaneous sustained VT or VF.8 –10 The first 2 of
these trials, the Multicenter Automatic Defibrillator Implantation Trial (MADIT) and Multicenter Unsustained
Tachycardia Trial (MUSTT), in addition to requiring that
participants have chronic coronary artery disease and abnormal left ventricular function, stipulated that patients have
documented nonsustained VT, as well as sustained VT/VF
inducible by programmed ventricular stimulation (electrophysiological testing).8,9 The MUSTT trial required patients
to have an ejection fraction (EF) ⱕ40% because it is at an EF
of 40% that trials in both the pre- and the post-thrombolytic
era have shown the break point separating patients at relatively low, versus higher, mortality risk (Figure).11,12 The
MADIT investigators required an EF of 35% for entry into
that trial.8 These investigators subsequently published a
substudy demonstrating that MADIT patients with an EF of
26% to 35% did not have improved survival with an ICD in
comparison with those randomized to conventional therapy,
whereas patients with an EF ⬍26% did have significantly
improved survival with an ICD.13 These findings presumably
influenced the design of the MADIT-II trial, in which the
major entry requirements were occurrence of MI at least 1
month prior to entry and an EF ⱕ30%.10 It should be noted
that although none of these trials required patients to have
heart failure, approximately two thirds of patients enrolled in
each did have symptomatic heart failure, of New York Heart
Association (NYHA) functional class II or III at entry (in
MADIT-II, 5% of patients were NYHA functional class IV).
Thus, patients enrolled into these trials had advanced coronary disease, marked not only by low EF but also by a high
percentage with symptomatic heart failure.
More recently, the potential role of ICDs in patients with
symptomatic heart failure resulting from both coronary as
well as noncoronary disease has been studied.14 –16 These 3
studies all have been marked by relatively low mortality rates
in the untreated control arms, and as a result, the first 2 appear
to have been somewhat underpowered.14,15 The German
Cardiomyopathy Trial enrolled 104 patients with nonischemic dilated cardiomyopathy diagnosed within 9 months,
NYHA functional class II or II heart failure symptoms, and an
EF ⱕ30%. Patients were randomized to therapy with or
without an ICD.14 Although the average EF was 24%, 2-year
mortality was ⬍10% in both treatment groups, and no benefit
was demonstrated with ICD therapy.
The DEFibrillators In Non-Ischemic cardiomyopathy Treatment
Evaluation (DEFINITE) trial enrolled 458 patients with nonischemic dilated cardiomyopathy, EF ⱕ35%, a history of symptomatic
heart failure, and spontaneous nonsustained VT or frequent ventricular ectopy.15 Patients were randomized to therapy with or without
an ICD. The average duration of heart failure at the time of entry
into the trial was 2.8 years. The 2-year mortality in the control group
was 14.1%. Although the 2-year mortality in the ICD-treated group
was 7.9%, the survival difference was not statistically significant.
The Sudden Cardiac Death in Heart Failure Trial (SCDHeFT), published earlier this year, enrolled 2521 patients with
heart failure resulting from coronary disease or nonischemic
dilated cardiomyopathy, NYHA class II or III symptoms for at
least 3 months, and EF ⱕ35%. Patients were randomized to 1 of
3 treatments: Placebo, amiodarone, or ICD. Nearly equal numbers of patients with coronary disease and nonischemic cardiomyopathy were enrolled, and the average EF was 25%. The
average duration of heart failure symptoms was 2.5 years before
enrollment into the trial. The 2-year mortality in the control
group was ⬇14%, and mortality in patients randomized to ICD
therapy was 23% lower than that of patients randomized to
placebo over an average follow-up of almost 4 years (P⫽0.007).
Patients with heart failure from nonischemic disease derived a
similar amount of benefit from ICD therapy as did patients with
coronary disease. Treatment with amiodarone conferred no
survival benefit.
To summarize briefly, multiple studies completed within
the past decade have demonstrated that ICDs can improve
survival in patients with coronary artery disease and abnormal left ventricular systolic function, reflected by a variety of
EF cutoffs. The most marked ICD survival benefits have been
demonstrated in trials that required markers for mortality risk
beyond just a reduced EF (such as spontaneous and inducible
ventricular arrhythmias). The present mortality of patients
with symptomatic heart failure and abnormal left ventricular
function resulting from nonischemic cardiomyopathies appears to be lower than that of patients with similar degrees of
ventricular dysfunction from coronary disease. It appears that
ICD therapy does reduce mortality in the latter patients, but
the mortality benefit may be more difficult to demonstrate.
Does the demonstration of reduced mortality with ICD
therapy in the study populations mean that the entry criteria
for the clinical trials should be adapted as our practice
guidelines? Also, to address the question at hand, should we
base the decision to implant ICDs for primary prevention of
sudden death solely on a certain EF value? If we were to
implant ICDs in all patients with an EF ⱕ30%, what would
be the impact on the overall occurrence of sudden death?
If we could demonstrate a causal relationship between a
certain critical left ventricular EF and the occurrence of
Ejection Fraction and ICD Therapy
Downloaded from http://circ.ahajournals.org/ by guest on June 12, 2017
potentially lethal ventricular tachyarrhythmias, then it would
seem rational to base implantation decisions on this factor.
Unfortunately, the mechanisms responsible for sudden death
are complex. They vary with both the presence and the type
of underlying heart disease. In addition, mechanisms underlying arrhythmias may depend on the stage of disease. For
example, in patients with coronary artery disease but no
previous MI, the most common mechanism responsible for
sudden death seems to be ventricular fibrillation induced by
acute ischemia. Most of these patients do not have markedly
abnormal ventricular function. In fact, in a prospective
registry of cardiac arrest in the Netherlands, ⬎50% of the
deaths in patients with coronary artery disease occurred in
patients whose EF was ⬎30%, and one fifth occurred in
patients with an EF ⬎50%.17,18 Once a patient with coronary
disease experiences MI, multiple additional mechanisms
leading to VT/VF may enter the picture. Intramyocardial
reentry resulting from nonuniform anisotropy and other
electrophysiological abnormalities created by fibrous tissue
formed during the healing process after infarction may result
in VT. Such reentry circuits do not require ischemia to
become activated. Compensatory hypertrophy in noninfarcted
myocardium may predispose to VF.19 Furthermore, occurrence of infarction does not offer protection from recurrent
ischemia as a cause of potentially lethal arrhythmias. Progressive ventricular remodeling after infarction may lead to
the syndrome of congestive heart failure, with its attendant
neural and hormonal abnormalities. In addition, structural and
functional abnormalities resulting from remodeling of the left
ventricle may by themselves result in ventricular arrhythmias
differing in mechanism from reentry or ischemia. Thus, patients
with coronary disease are subject to a spectrum of mechanisms
that may cause sudden death. These mechanisms do not remain
constant; they evolve over the course of the disease.
These observations have several implications when trying
to develop and apply tests for risk stratification, such as the
measurement of ventricular function. That multiple mechanisms may contribute to the occurrence of VT/VF implies
that no single test is likely to accurately predict risk of sudden
death for all patients. That coronary disease is an evolutionary
process suggests that the results of individual tests will not
remain stable throughout the course of the disease, and will
require repetition. For example, a decreasing EF that does not
fall below 30% may carry equal prognostic significance as a
markedly reduced (⬍30%), but stable EF; however, I am not
aware of any data to substantiate this hypothesis.
In addition to the evolutionary nature of both coronary and
noncoronary disease, the risk of sudden death is nonlinear
over time. Previous analyses have demonstrated a clustering
of sudden death events. The period of highest risk is the first
6 to 12 months after infarction, and recurrence of cardiac
arrest is highest within the first year after an initial event.20
The current therapies of acute MI, such as thrombolytic
therapy and primary angioplasty, as well as the use of
␤-adrenergic blocking agents may be changing this natural
2539
history. A recent prospective survey in Finland showed that
total mortality and the occurrence of nonsudden deaths
remain highest within the first 18 months after infarction.21
The occurrence of sudden death was delayed, with a majority
of events occurring 20 to 40 months after infarction, however.
The average EF of patients who died suddenly was 41%,
whereas the average EF of patients who experienced nonsudden death was 37%. Importantly, the average EF of patients
who died suddenly did not differ significantly from the EF of
survivors. The Maastricht Circulatory Arrest Registry in the
Netherlands has recently documented that the average time
from first infarction to sudden death is 9 years.18
What are the implications if the timing of sudden death
after MI is changing? It seems likely that these events result
either from a reentrant VT circuit that remained quiescent for
years until activated or from anatomic change, such as
progressive ventricular remodeling. Could such changes be
discovered by periodic remeasurement of EF? Perhaps, but
this remains to be proven. Of note, the MADIT-II investigators observed an effect of time from infarction to enrollment
in the study on the observed ICD benefit. Patients who
entered the trial between 1 and 17 months after the most
recent infarction experienced similar survival, whether randomized to ICD or conventional therapy. In contrast, maximal ICD benefit was observed in patients enrolled ⬎10 years
after infarction because mortality in the control group was
highest in that time period.22 This seems counterintuitive and
suggests the possibility of selection bias for enrollment into
the trial. Furthermore, the recently reported Defibrillator IN
Acute Myocardial Infarction Trial (DINAMIT; a study of
prophylactic ICD implantation in patients sustaining an acute
MI within the previous 40 days, with EF ⬍35%) also failed
to demonstrate improved survival with ICD therapy.23,24 In
any case, these observations suggest that EF alone is not
sufficient to identify appropriate candidates for ICD therapy.
Rather, other clinical factors should influence this decision,
possibly including time interval after infarction.
If EF were an ideal risk stratification test, then both its
sensitivity and specificity would approach 100%, with its
predictive accuracy remaining stable over time. If EF warranted consideration as the sole test on which to determine
ICD placement, then it would identify patients specifically at
risk for sudden death and exclude those likely to die from
nonarrhythmic causes. This would be the case if EF directly
reflected the pathophysiology underlying sudden death. How
close does EF approach these ideals?
What is the sensitivity of EF in patients with coronary
disease for prediction of sudden death or cardiac arrest?
Several lines of evidence suggest EF lacks sufficient sensitivity to be useful by itself. The most recent report from the
Maastricht prospective registry found that only 19% of
patients with known coronary disease who experienced cardiac arrest had an EF ⱕ30% before the event.18 An earlier
report from the same registry demonstrated that the EF was
2540
Circulation
May 17, 2005
Downloaded from http://circ.ahajournals.org/ by guest on June 12, 2017
⬎30% in 56.5% of patients with a documented infarction
who subsequently experienced cardiac arrest.17
The limited sensitivity of EF to predict risk of sudden death also
is illustrated by the European Autonomic Tone and Reflexes After
Myocardial Infarction (ATRAMI) study (1284 patients enrolled
within 1 month of acute MI who underwent multiple risk stratification tests).25 In this study, the mean EF was 49%.25 After an
average follow-up of 21 months, a total of 49 deaths occurred. In
only 22 of those experiencing cardiac arrest after infarction was the
EF ⬍35% when measured early after MI.
We also can examine the EF in survivors of cardiac arrest.
The average value of EF measured in the Antiarrhythmics
Versus Implantable Defibrillators trial, the Canadian Implantable Defibrillator Study, and the Cardiac Arrest Study Hamburg (AVID, CIDS, CASH) was 32%, 34%, and 45%,
respectively.4 – 6 Thus, most patients in these trials had an EF
⬎30%. These data are even more compelling with the
knowledge that the measurements of EF were made soon
after resuscitation from arrest, when ventricular function is
depressed.26 It seems likely that the EF before the cardiac
arrest would have been higher in most cases. Thus, EF lacks
useful sensitivity as a risk-stratification tool.
How does EF perform with regard to specificity for
prediction of sudden death? This question is difficult to
answer because it requires accurate determination of mode of
death (sudden versus nonsudden, with the presumption that
sudden deaths are mediated by arrhythmia). Other authors
have noted the potential errors when attempting to assign
mechanisms to fatal events in clinical trials.27,28 It is clear that
not all sudden deaths are the result of primary arrhythmic
events. Nevertheless, the majority of instantaneous deaths in
patients with coronary disease who do not have end-stage
heart failure appear to be attributable to VTs. Were this not
the case, the percentage of reduction in sudden death in
ICD-treated patients would be expected to be similar to the
reduction in total mortality. In fact, in the primary-prevention
trials of patients with coronary disease, the percentage of
reduction in sudden death markedly exceeded the reduction in
total mortality in patients treated with ICDs.8,9,29
In an effort to assess the specificity of EF as a predictor of
sudden death, the MUSTT investigators noted no significant
difference in the percentage of deaths attributed to arrhythmia
in patients whose EF was ⬍30% versus those whose EF was
30% to 40%.30 If there were a causal relationship between EF
and occurrence of sudden death, then we would expect to find
a significantly higher percentage of patients whose EF was
⬍30% to have deaths attributed to arrhythmia. Further
evidence suggesting low specificity of EF taken in isolation
for risk prediction can be obtained from examining the
ATRAMI study.31 In this analysis, EF ⬍35% alone did not
confer significantly increased mortality risk. Only when EF
⬍35% was combined with other factors, such as nonsustained
VT or abnormal baroreflex sensitivity, were patients at
significantly increased mortality risk.
Another way to examine the utility of EF alone and its
interaction with other factors may be seen from analysis of the
MUSTT study data.30 In that analysis, total mortality was
affected independently by EF as well as inducibility of VT. The
total mortality of patients whose EF was ⱕ30%, regardless of
the presence or absence of inducible sustained VT, was higher
than that of patients whose EF was ⬎30%. Patients whose EF
was ⱕ30% without inducible sustained VT, however, had a
similar risk of arrhythmic death or cardiac arrest as patients with
EF ⬎30% and inducible sustained VT. Thus, EF alone cannot
predict the mode of death or the likelihood of responding to ICD
therapy. In addition, this observation demonstrates that subgroups of patients whose EF exceeds 30% (eg, those with
inducible sustained VT) may be at equal risk for sudden death as
many patients whose EF is ⱕ30%. Thus, by focusing only on
patients whose EF is ⱕ30%, there is the danger of missing a
significant population of patients that may benefit from ICD
therapy.
Although EF combined with spontaneous and inducible
ventricular tachyarrhythmias has been proven to identify
patients at low versus higher mortality risk, a number of other
less well-appreciated factors have been shown to modify risk
and influence event rates observed in randomized trials. For
example, studies of cardiac arrest survivors have documented
that mortality is related to the site of cardiac arrest.32 Patients
who experience in-hospital cardiac arrest have significantly
higher mortalities than do patients experiencing out-ofhospital cardiac arrest.32 It also appears that in-hospital versus
out-of-hospital location at the time of recruitment into primary prevention of sudden death trials affects mortality. In
the MUSTT trial, outpatients at the time of enrollment had
significantly better (by ⬇33%) survival than did those who
were identified as inpatients.33 A multivariate analysis of
these factors showed that EF was not associated with patient
location. In contrast, outpatients were significantly younger,
had multivessel coronary disease less often, had a history of
heart failure less often, and had less advanced NYHA
functional class. In addition, outpatients were significantly
less likely to have inducible sustained VT than inpatients at
the time of recruitment into the trial.
The realization that sudden death and total mortality risk
are influenced by numerous factors in addition to EF led the
MUSTT investigators to develop a predictive model based on
multiple parameters.34 Factors having the highest utility for
prediction of total mortality included digitalis therapy, EF
ⱕ30%, left bundle-branch block or nonspecific intraventricular conduction delay, discovery of nonsustained VT while an
inpatient, and age ⬎65 years. With this model, patients whose
only risk factor was EF ⱕ30% had a 2-year total mortality
risk of only 6.2% and a 2-year risk for arrhythmic death or
cardiac arrest of only 3.5%. The reason that the risk predicted
by this model is less than one third that observed in the
control arm of the MADIT-II study (19.8% total mortality at
the average follow-up of 20 months)10 is because the majority
of patients enrolled into the MADIT-II study had other risk
Ejection Fraction and ICD Therapy
Downloaded from http://circ.ahajournals.org/ by guest on June 12, 2017
factors: Two thirds of patients had symptomatic heart failure
(NYHA class ⱖII), almost 50% of the study patients had left
bundle-branch block or a nonspecific intraventricular conduction delay on the ECG at the time of enrollment, and ⬎50%
were receiving digitalis. Another example of the degree to
which patient selection influenced observed mortality in the
MADIT-II trial is illustrated by comparing the 2-year mortality of MADIT-II control patients (21%) with that observed
in placebo-treated patients with coronary disease in the
SCD-HeFT trial (19%). Because patients enrolled in the
SCD-HeFT trial were required to have symptomatic heart
failure their mortality would be expected to exceed that of
patients enrolled in the MADIT-II study, which did not
require the presence of heart failure. The fact that the
observed mortality in these 2 studies is similar suggests that
enrollment into MADIT-II was biased toward a group of
patients with advanced illness. Thus, not all patients whose
EF is ⱕ30% are at equal mortality risk. I suspect that
physicians’ motivation to implant prophylactic ICDs is different for patients with a predicted 2-year mortality risk of 6%
versus 20%, especially if it is taken into consideration that
arrhythmic deaths account for ⬇50% the total mortality.
Patients enrolled in the MUSTT trial shared a number of
characteristics with those enrolled in MADIT-II: 63% had
symptomatic heart failure (NYHA class II or III), one-quarter
of patients had left bundle-branch block or a nonspecific
intraventricular conduction delay,35 and ⬎50% were receiving digitalis. Three quarters of patients in the MUSTT study
were identified as inpatients.33 The point of this analysis is
that the characteristics of patients enrolled in these clinical
trials profoundly influence observed outcomes, and the results may not be generalizable to all patients meeting the
entry criteria specified in the protocols. If the conduct of a
clinical trial inadvertently results in a study population at
significantly higher risk than that of the universe of patients
meeting the entry criteria specified by the protocol, then the
results may have misleading implications for clinical practice.
Stated another way, clinical trial design, including entry and
exclusion criteria as well as protocol conduct, is influenced
by a number of considerations. These considerations include
the desirability of choosing a study population at high risk for
developing the end point that an intervention (such as ICD
therapy) seeks to modify to enhance the likelihood of obtaining a positive treatment effect. Ideally, it would be nice to be
able to apply trial results directly to clinical practice. Practice
guidelines must be based on objective evidence, including
clinical trial results; however, common sense would dictate
that practice guidelines be influenced by broader considerations, including the need to place trial results in context.
In summary, EF is far from an ideal risk-stratification test
on which to base prophylactic ICD therapy. Multiple factors
interact with EF to influence the mortality of patients with
similar degrees of left ventricular dysfunction. Thus, we need
combinations of tests based on individual patient characteristics if we are to use ICD therapy most efficiently for
2541
primary prevention of sudden death. In essence, not everyone
with an EF ⱕ30% should receive an ICD. Other riskstratification tools are needed in patients whose EF is ⬍30%,
as well as in those having an EF ⬎30% if the number of
patients likely to benefit from ICDs is to be maximized.
Disclosure
Dr Buxton was supported in part by grants No. UO1 HL45700 and
UO1 HL45726 from the National Heart, Lung, and Blood Institute,
National Institutes of Health.
References
1. Zheng Z-J, Croft JB, Giles WH, Mensah GA. Sudden cardiac death in the
United States, 1989 to 1998. Circulation. 2001;104:2158 –2163.
2. Echt DS, Liebson PR, Mitchell LB, Peters RW, Obias-Manno D, Barker
AH, Arensberg D, Baker A, Friedman L, Greene HL. Mortality and
morbidity in patients receiving encainide, flecainide, or placebo. The
Cardiac Arrhythmia Suppression Trial. N Engl J Med. 1991;324:781–788.
3. Mirowski M, Reid P, Mower M. Termination of malignant ventricular
arrhythmias with an implanted automatic defibrillator in human beings.
N Engl J Med. 1980;303:322–324.
4. The Antiarrhythmics Versus Implantable Defibrillators (AVID) Investigators. A comparison of antiarrhythmic-drug therapy with implantable
defibrillators in patients resuscitated from near-fatal ventricular arrhythmias. N Engl J Med. 1997;337:1576 –1583.
5. Connolly S, Gent M, Roberts R, Dorian P, Roy D, Sheldon R, Mitchell L,
Green M, Klein G, O’Brien B. Canadian Implantable Defibrillator Study
(CIDS): a randomized trial of the implantable cardioverter defibrillator
against amiodarone. Circulation. 2000;101:1297–1302.
6. Kuck K-H, Cappato R, Siebels J, Ruppel R. Randomized comparison of
antiarrhythmic drug therapy with implantable defibrillators in patients
resuscitated from cardiac arrest: the Cardiac Arrest Study Hamburg
(CASH). Circulation. 2000;102:748 –754.
7. Connolly SJ, Hallstrom AP, Cappato R, Schron EB, Kuck K-H, Zipes DP,
Greene HL, Boczor S, Domanski M, Follmann D, Gent M, Roberts RS.
Meta-analysis of the implantable cardioverter defibrillator secondary prevention trials. AVID, CASH and CIDS studies. Antiarrhythmics vs
Implantable Defibrillator study. Cardiac Arrest Study Hamburg .
Canadian Implantable Defibrillator Study. Eur Heart J. 2000;21:
2071–2078.
8. Moss AJ, Hall WJ, Cannom DS, Daubert JP, Higgins SL, Klein H, Levine
JH, Saksena S, Waldo AL, Wilber D, Brown MW, Heo M. Improved
survival with an implanted defibrillator in patients with coronary disease
at high risk for ventricular arrhythmia. Multicenter Automatic Defibrillator Implantation Trial Investigators. N Engl J Med. 1996;335:
1933–1940.
9. Buxton AE, Lee KL, Fisher JD, Josephson ME, Prystowsky EN, Hafley
G; the Multicenter Unsustained Tachycardia Trial Investigators. A randomized study of the prevention of sudden death in patients with coronary
artery disease. N Engl J Med. 1999;341:1882–1890.
10. Moss AJ, Zareba W, Hall WJ, Klein H, Wilber DJ, Cannom DS, Daubert
JP, Higgins SL, Brown MW, Andrews ML; the Multicenter Automatic
Defibrillator Implantation Trial II Investigators. Prophylactic implantation of a defibrillator in patients with myocardial infarction and reduced
ejection fraction. N Engl J Med. 2002;346:877– 883.
11. The Multicenter Postinfarction Research Group. Risk stratification and
survival after myocardial infarction. N Engl J Med. 1983;309:331–336.
12. Rouleau JL, Talajic M, Sussex B, Potvin L, Warnica W, Davies RF,
Gardner M, Stewart D, Plante S, Dupuis R, Lauzon C, Ferguson J, Mikes
E, Balnozan V, Savard P. Myocardial infarction patients in the
1990s—their risk factors, stratification and survival in Canada: the
Canadian Assessment of Myocardial Infarction (CAMI) Study. J Am Coll
Cardiol. 1996;27:1119 –1127.
13. Moss AJ, Fadl Y, Zareba W, Cannom DS, Hall WJ; the Multicenter
Automatic Defibrillator Implantation Trial Research Group. Survival
benefit with an implanted defibrillator in relation to mortality risk in
chronic coronary heart disease. Am J Cardiol. 2001;88:516 –520.
2542
Circulation
May 17, 2005
Downloaded from http://circ.ahajournals.org/ by guest on June 12, 2017
14. Bansch D, Antz M, Boczor S, Volkmer M, Tebbenjohanns J, Seidl K,
Block M, Gietzen F, Berger J, Kuck KH. Primary prevention of sudden
cardiac death in idiopathic dilated cardiomyopathy: the Cardiomyopathy
Trial (CAT). Circulation. 2002;105:1453–1458.
15. Kadish A, Dyer A, Daubert J, Quigg R, Estes NA, Anderson K, Calkins
H, Hoch D, Goldberger J, Shalaby A, Sanders W, Schaechter A, Levine
J; Defibrillators in Non-Ischemic Cardiomyopathy Treatment Evaluation
(DEFINITE) Investigators. Prophylactic defibrillator implantation in
patients with nonischemic dilated cardiomyopathy. N Engl J Med. 2004;
350:2151–2158.
16. Bardy GH, Lee KL, Mark DB, Poole JE, Packer DL, Boineau R,
Domanski M, Troutman C, Anderson J, Johnson G, McNulty SE, ClappChanning N, Davidson-Ray LD, Fraulo ES, Fishbein DP, Luceri RM, Ip
JH; Sudden Cardiac Death in Heart Failure Trial (SCD-HeFT) Investigators. Amiodarone or an implantable cardioverter-defibrillator for congestive heart failure. N Engl J Med. 2005;352:225–237.
17. de Vreede-Swagemakers JJ, Gorgels AP, Dubois-Arbouw WI, van Ree
JW, Daemen MJ, Houben LG, Wellens HJ. Out-of-hospital cardiac arrest
in the 1990’s: a population-based study in the Maastricht area on
incidence, characteristics and survival. J Am Coll Cardiol. 1997;30:
1500 –1505.
18. Gorgels AP, Gijsbers C, de Vreede-Swagemakers J, Lousberg A, Wellens
HJ. Out-of-hospital cardiac arrest-the relevance of heart failure. The
Maastricht Circulatory Arrest Registry. Eur Heart J. 2003;24:
1204 –1209.
19. Misier AR, Opthof T, van Hemel NM, Vermeulen JT, de Bakker JM,
Defauw JJ, van Capelle FJ, Janse MJ. Dispersion of “refractoriness” in
noninfarcted myocardium of patients with ventricular tachycardia or ventricular fibrillation after myocardial infarction. Circulation. 1995;91:2566–2572.
20. Myerburg RJ, Kessler KM, Castellanos A. Sudden cardiac death:
structure, function, and time-dependence of risk. Circulation. 1992;85:
I2–I10.
21. Huikuri HV, Tapanainen JM, Lindgren K, Raatikainen P, Makikallio TH,
Juhani Airaksinen KE, Myerburg RJ. Prediction of sudden cardiac death
after myocardial infarction in the beta-blocking era. J Am Coll Cardiol.
2003;42:652– 658.
22. Wilbur D, Zareba W, Hall W, Brown M, Lin A, Andrews M, Burke M,
Moss AJ. Time dependence of mortality risk and defibrillator benefit after
myocardial infarction. Circulation. 2004;109:1082–1084.
23. Hohnloser SH, Connolly SJ, Kuck K-H, Dorian P, Fain E, Hampton JR,
Hatala R, Pauly AC, Roberts RS, Themeles E, Gent M. The defibrillator
in acute myocardial infarction trial (DINAMIT): study protocol. Am
Heart J. 2000;140:735–739.
24. Hohnloser SH. DINAMIT-Defibrillator in acute myocardial infarction
trial. Paper presented at: Annual Meeting of the American College of
Cardiology; March 7 to 10, 2004; New Orleans, La.
25. La Rovere MT, Bigger JT Jr, Marcus FI, Mortara A, Schwartz PJ.
Baroreflex sensitivity and heart rate variability in prediction of total
cardiac mortality after myocardial infarction. ATRAMI (Autonomic Tone
and Reflexes After Myocardial Infarction) Investigators. Lancet. 1998;
351:478 – 484.
26. Laurent I, Monchi M, Chiche J-D, Joly L-M, Spaulding C, Bourgeois B,
Cariou A, Roszeberg A, Carli P, Weber S, Dhainaut J-F. Reversible
myocardial dysfunction in survivors of out-of-hospital cardiac arrest.
J Am Coll Cardiol. 2002;40:2110 –2116.
27. Greene HL, Richardson DW, Barker AH, Roden DM, Capone RJ, Echt
DS, Friedman LM, Gillespie MJ, Hallstrom AP, Verter J. Classification
of deaths after myocardial infarction as arrhythmic or nonarrhythmic (the
Cardiac Arrhythmia Pilot Study). Am J Cardiol. 1989;63:1– 6.
28. Pratt CM, Greenway PS, Schoenfeld MH, Hibben ML, Reiffel JA. Exploration of the precision of classifying sudden cardia death. Implications for
the interpretation of clinical trials. Circulation. 1996;93:519 –524.
29. Greenberg H, Case RB, Moss AJ, Brown MW, Carroll ER, Andrews ML.
Analysis of mortality events in the multicenter automatic defibrillator
implantation trial (MADIT-II). J Am Coll Cardiol. 2004;43:1459 –1465.
30. Buxton AE, Lee KL, Hafley GE, Wyse DG, Fisher JD, Lehmann MH,
Pires LA, Gold MR, Packer DL, Josephson ME, Prystowsky EN, Talajic
MR; MUSTT Investigators. Relation of ejection fraction and inducible
ventricular tachycardia to mode of death in patients with coronary artery
disease: an analysis of patients enrolled in the multicenter unsustained
tachycardia trial. Circulation. 2002;106:2466 –2472.
31. La Rovere MT, Pinna GD, Hohnloser SH, Marcus FI, Mortara A, Nohara
R, Bigger JT Jr, Camm AJ, Schwartz PJ. Baroreflex sensitivity and heart
rate variability in the identification of patients at risk for life-threatening
arrhythmias: implications for clinical trials. Circulation. 2001;103:
2072–2077.
32. Epstein AE, Powell J, Yao Q, Ocampo C, Lancaster S, Rosenberg Y,
Cannom DS, Herre JM, Greene HL. In-hospital versus out-of-hospital
presentation of life-threatening ventricular arrhythmias predicts survival.
Results from the AVID Registry. J Am Coll Cardiol. 1999;34:1111–1116.
33. Pires LA, Lehmann MH, Buxton AE, Hafley G, Lee KL; the Multicenter
Unsustained Tachycardia Trial Investigators. Differences in inducibility
and prognosis of in-hospital versus out-of-hospital identified nonsustained ventricular tachycardia in patients with coronary artery disease:
clinical and trial design implications. J Am Coll Cardiol. 2001;38:
1156 –1162.
34. Buxton AE, Lee KL, Hafley GE, Prystowsky EN, Josephson ME, Fisher
JD, Gold MR. A simple model using the MUSTT database can stratify
total mortality and sudden death risk of coronary disease patients. J Am
Coll Cardiol. 2004;43:425A.
35. Zimetbaum PJ, Buxton AE, Batsford W, Fisher JD, Hafley GE, Lee KL,
O’Toole MF, Page RL, Reynolds M, Josephson ME. Electrocardiographic predictors of arrhythmic death and total mortality in the Multicenter Unsustained Tachycardia Trial. Circulation. 2004;110:766 –769.
Everyone With an Ejection Fraction Less Than or Equal to
30% Should Receive an Implantable Cardioverter-Defibrillator
Arthur J. Moss, MD
T
he automatic implantable defibrillator was introduced
into clinical medicine by Mirowski et al1 in 1980,
when they published their report on the life-saving
benefit of the implanted defibrillator in 3 patients with
recurrent episodes of ventricular tachyarrhythmias. Thereafter, there were numerous reports of effective defibrillator
therapy in individual high-risk patients and in retrospective
analyses of series of patients who had a defibrillator im-
Ejection Fraction and ICD Therapy
Downloaded from http://circ.ahajournals.org/ by guest on June 12, 2017
planted because of documented episodes of hemodynamically
compromising ventricular tachycardia or ventricular fibrillation. Most of these studies compared survival after defibrillator implantation with either historical controls or assumed
death when the defibrillator fired. When the implantable
defibrillator was first introduced, the device provided only
shock therapy for fast ventricular tachyarrhythmias. Subsequent technological improvements have included transvenous
leads, improved rhythm-detection algorithms, enhanced data
storage, pacing capability, antitachycardia pacing features,
augmented programmability, extended battery life, and reduction in the size of the implanted device. With these
features, the device became known as the implantable
cardioverter-defibrillator (ICD).
When the preliminary 1989 report of the Cardiac Arrhythmia Suppression Trial (CAST) showed that antiarrhythmic
drugs were associated with increased mortality as compared
with placebo,2 it was clear that randomized ICD trials would
have to be carried out. Since 1990, 10 large randomized ICD
trials involving at-risk patients have been published or presented at international cardiology meetings. Seven of the
trials were for primary prevention, in that the ICD was
implanted to reduce mortality in patients with left ventricular
dysfunction from ischemic or nonischemic cardiomyopathy.
These primary prevention trials involved patients who had
not yet experienced a documented life-threatening ventricular
tachyarrhythmia. Three trials were for secondary prevention
and involved implantation of the ICD in patients who either
had survived an aborted cardiac arrest or had experienced a
documented life-threatening ventricular tachyarrhythmia.
Considerable data exist regarding the inverse relationship
between the level of the ejection fraction (EF) and subsequent
cardiac mortality.3 As the EF falls below 0.40, there is a
progressive increase in the 1-year cardiac mortality rate
(Figure 1). Approximately 50% of all cardiac mortality is the
result of sudden death according to the 1-hour definition for
sudden death. Presently, there are no reliable methods for
identifying in advance those patients who are more likely to
die from sudden than from nonsudden cardiac death. Thus, a
reduced EF has become the major eligibility criterion for
enrolling patients in primary-prevention ICD trials to reduce
mortality. The question at the heart of this controversy-incardiology discussion is whether an EF ⱕ0.30 alone is
sufficient to select patients who should have an ICD implanted. A few of the risk parameters frequently discussed are
do patients with an EF ⱕ0.30 require additional risk stratification such as increased QRS duration, inducible ventricular
tachyarrhythmias, or advanced New York Heart Association
functional class? A corollary to this question is can a low-risk
group of patients be identified within the low EF group with
such a low mortality that the ICD is unlikely to improve
survival within a reasonable time horizon?
This article reviews the 10 major randomized ICD trials involving 8002 patients in which each trial had at least 90 patients per
treatment arm and used all-cause mortality as the end point.4–13 The
2543
Figure 1. One-year mortality rate as a function of the EF measured at hospital discharge after MI. Reprinted with permission
from N Engl J Med.3 Copyright 1983, Massachusetts Medical
Society.
particular focus of my presentation will be derived from randomized
ICD trials in which the relative benefit of ICD therapy was
evaluated in patients with EFs ⱕ0.30. The Multicenter Unsustained
Tachycardia Trial (MUSTT) was not a randomized ICD trial, but
rather a randomized study to determine whether electrophysiologically guided antiarrhythmic drug therapy would reduce the risk of
sudden death among patients with coronary artery disease, EF
ⱕ0.40 (median EF 0.30), and asymptomatic unsustained ventricular
tachycardia.14 Some of the observational ICD findings in MUSTT
provide useful outcome information about ICD effectiveness in
patients with reduced EF, and findings from MUSTT are discussed.
It is my position based on available evidence from these 10
randomized ICD trials plus MUSTT that the overwhelming majority of cardiac patients with an EF ⱕ0.30 should receive a prophylactic ICD.
Randomized Trials
Primary ICD Prevention Studies
The design characteristics of 7 ICD primary-prevention trials
are presented in Table 1, with the relative hazard ratios (HR)
and the absolute ICD-related mortality reductions for ICD
versus conventional therapy for each trial shown in Figures 2
and 3, respectively. Four of the trials involved only patients
with ischemic heart disease, 1 study focused exclusively on
patients with nonischemic cardiomyopathy, and 2 studies
enrolled patients having ischemic or nonischemic cardiomyopathy. In the primary prevention trials, eligibility involved
patients with an EF ⱕ0.35 in 5 trials (n⫽3907) and ⱕ0.30 in
1 trial (n⫽1232). Although there is some variability in the
2544
Circulation
May 17, 2005
TABLE 1.
Randomized Primary-Prevention ICD Trials
Date Started; Date Published
No. of
Patients
Eligibility
Average
Follow-Up, mo
MADIT (ICD vs conventional treatment)4
12/27/1990; 12/27/1996
CABG Patch (ICD vs conventional treatment)5
08/14/1990; 11/27/1997
196
Previous MI, EF ⱕ0.35, NSVT, EPS⫹
27
900
CAD, abnormal SAECG, CABG surgery
MADIT-II (ICD vs. conventional treatment)9
32
07/11/1997; 03/21/2002
1232
Previous MI, EFⱕ0.30
20
COMPANION (CRT-D vs OPT)12*
01/20/2000; 05/10/2004
903
I & NICM, NYHA class III or
IV, EF ⱕ0.35, QRS ⱕ0.12 s
16
DEFINITE (ICD vs conventional treatment)10
07/09/1998; 05/10/2004
458
NICM, EFⱕ0.35, NSVT or ⬎10 PVC/24 h
29
SCD-HeFT (ICD vs placebo)13†
09/16/2001; 03/08/2004
1676
I & NICM, EF ⱕ0.35
48
2000; 03/08/2004
674
Within 6–40 d of acute MI,
EF ⱕ0.35, HRV (SDNN) ⱕ70 ms
30
Trial
11
DINAMIT (ICD vs conventional treatment)
Downloaded from http://circ.ahajournals.org/ by guest on June 12, 2017
CAD indicates coronary artery disease; CRT-D, cardiac-resynchronization therapy with pacemaker-defibrillator; EF,
nonischemic cardiomyopathy; HRV, heart rate variability; NSVT, nonsustained ventricular tachycardia; OPT, optimal
ventricular contraction; SAECG, signal-averaged electrocardiogram; and SDNN, standard deviation of normal heart beat
*Three-arm trial (CRT-D, CRT, and OPT), but comparison in this analysis only for CRT-D vs optimal pharmacological
†Three-arm trial (ICD, amiodarone, and placebo), but comparison in this analysis only for ICD vs placebo.
results, the overall average HR of all 7 trials is 0.72, a value
indicating a 28% reduction in the risk of death in the
ICD-treated patients as compared with the conventionally
treated patients (Figure 2). The absolute reduction in mortality between ICD- and conventionally treated patients also was
variable among the various studies, with an overall absolute
2-year mortality reduction of 3.0 percentage points, from
17.3% in the conventionally treated patients to 14.3% in the
ICD-treated patients (Figure 3).
Secondary ICD Prevention Studies
The design characteristics of 3 secondary-prevention trials are
presented in Table 2, with the relative HRs and the absolute
ICD-related mortality reductions for ICD versus conventional ther-
Figure 2. HR (ICD:conventional treatment) and 95% confidence
limits for mortality in 7 primary-prevention trials. The average HR
for the 7 trials is 0.72 (n⫽6039). COMPANION indicates Comparison of Medical Therapy, Pacing, and Defibrillation in Heart
Failure trial; DEFINITE, DEFibrillators In Non-Ischemic cardiomyopathy Treatment Evaluation trial; and SCD-HeFT, Sudden Cardiac Death in Heart Failure Trial. Other abbreviations as in text.
ejection fraction; I & NICM, ischemic and
pharmacological therapy; PVC, premature
intervals.
therapy.
apy for each trial shown in Figures 4 and 5, respectively. Patients
enrolled in these 3 trials had documented or suspect life-threatening
ventricular tachyarrhythmias, including patients who had been
resuscitated from ventricular fibrillation, those with ventricular
tachycardia associated with syncope or severe hemodynamic compromise, and those who experienced unmonitored syncope in
whom a full workup suggested a ventricular tachyarrhythmia as the
cause of the loss of consciousness. The patients in these 3 trials had
mean EFs in the range of 0.31 to 0.47, values considerably higher
than in the primary-prevention trials. The efficacy of ICD therapy in
these 3 trials was compared against antiarrhythmic drugs (mostly
amiodarone). The HR and 95% confidence intervals for all-cause
mortality in each of the 3 studies for ICD-treated patients compared
with those receiving antiarrhythmic drugs are presented in Figure 4.
The overall average HR of 0.65 indicates a 35% reduction in the
risk of death from ICD therapy. The mortality difference between
ICD- and conventionally treated patients in these secondaryprevention trials showed an overall absolute 2-year mortality reduc-
Figure 3. Two-year mortality rates in 7 primary-prevention ICD
trials. The average 2-year mortality rates in the conventional and
ICD treatment groups were 17.3% and 14.3%, respectively, with
an absolute 2-year mortality reduction of 3.0 percentage points
with ICD therapy.
TABLE 2.
Ejection Fraction and ICD Therapy
2545
Average
Follow-Up, mo
Reference
Randomized Secondary-Prevention ICD Trials
Date Started;
Date Published
No. of
Patients
AVID (ICD vs AAD)
06/01/1993; 11/27/1997
1016
Near-fatal VF, VT with syncope or
hemodynamic compromise
18
6
CIDS (ICD vs amiodarone)
10/01/1990; 03/21/2000
659
Near-fatal VF, VT with syncope or hemodynamic compromise,
unmonitored syncope thought to be result of VT
36
7
CASH (ICD vs amiodarone
vs metoprolol)
03/01/1997 08/15/2000
288
Resuscitated from cardiac arrest
secondary to sustained VT/VF
48
8
Trial
Eligibility
AAD indicates antiarrhythmic drugs; VF, ventricular fibrillation; and VT, ventricular tachycardia.
tion of 7.5 percentage points, from 23.0% in the conventionally
treated group to 15.5% in the ICD-treated patients (Figure 5).
Downloaded from http://circ.ahajournals.org/ by guest on June 12, 2017
Electrophysiologically Guided Drug
Study (MUSTT)
The primary aim of the MUSTT study was to evaluate the
efficacy of electrophysiologically guided antiarrhythmic therapy. In patients who had an unsuccessful drug test, an ICD could
be recommended.14 The patients who received nonrandomized
ICD therapy had a significantly lower risk of cardiac arrest or
death from arrhythmia than those not receiving an ICD. In a
secondary analysis involving MUSTT patients who did not
receive antiarrhythmic therapy, total mortality and arrhythmic
deaths/cardiac arrest occurred more frequently in patients with
EF ⬍0.30 than in those with an EF of 0.30 to 0.40.15
EF and ICD Efficacy
There are 7 primary prevention trials with HR data for
mortality in patients with EFs ⱕ0.30, and 6 primary prevention trials with relevant outcome data for EF ⬎0.30 (Figure
6). For those with EFs ⱕ0.30, the average hazard ratio was
0.71, indicating a 29% reduction in the risk of death with ICD
therapy in this subgroup. This finding is in contrast to patients
with EFs ⬎0.30 who had an average HR of 1.01, a value
indicating no benefit from ICD therapy relative to conventional therapy in this subgroup. In the nonrandomized ICD
portion of the MUSTT trial, the relative risk for death from all
causes was 0.40 (P⬍0.01) for those receiving an ICD as
Figure 4. HR (ICD:conventional treatment) and 95% confidence
limits for mortality in 3 secondary-prevention trials. The average
HR for these 3 trials is 0.65 (n⫽1963). CIDS indicates Canadian
Implantable Defibrillator Study; CASH, Cardiac Arrest Study
Hamburg. Other abbreviations as in text.
compared with those without an ICD, but a breakdown was
not provided for EF above and below 0.30.14
Regarding secondary-prevention trials, data from the Antiarrhythmics Versus Implantable Defibrillators (AVID)
study are relevant.16 The AVID findings indicate that patients
with an EF ⱖ0.35 have virtually the same survival with ICD
therapy as with antiarrhythmic drugs. Although there are no
reported data for patients with an EF ⱕ0.30, those with EFs
⬍0.20 and those with EFs in the range of 0.20 to 0.34
obtained considerable improvement in survival with the ICD.
Commentary
The available data indicate that ICD therapy provides a
meaningful and significant reduction in mortality in high-risk
patients with ischemic and nonischemic cardiomyopathy as
part of primary- and secondary-prevention strategies. These
populations contain considerable risk heterogeneity, with
clear survival benefit in patients with EFs ⱕ0.30, but not in
patients with EFs ⬎0.30. The meta-analysis findings indicate
that sicker patients (ie, those with lower EFs) benefit more
from ICD therapy.17 The MUSTT study indicates that those
Figure 5. Two-year mortality rates in 3 secondary-prevention
trials. The average 2-year mortality rates in the conventional and
ICD treatment groups were 23.0% and 15.5%, respectively, with
an absolute 2-year mortality reduction of 7.5 percentage points
with ICD therapy.
2546
Circulation
May 17, 2005
Downloaded from http://circ.ahajournals.org/ by guest on June 12, 2017
Figure 6. HR (ICD:conventional treatment) and 95% confidence
limits for mortality by EF. The number of patients in the DINAMIT study by EF was not available, so the DINAMIT study was
excluded from computation of the average HR. A, EF ⱕ0.30
(average HR⫽0.71, n⫽4752, excluding DINAMIT). B, EF ⬎0.30
(pooled HR⫽1.01, n⫽1485, excluding DINAMIT).
with an EF ⬍0.30 are at greater mortality risk than those with
higher EFs,15 so it is likely that the ICD benefit in MUSTT
was concentrated in those with lower EFs.
Patients with ischemic or nonischemic cardiomyopathy and
EF ⱕ0.30 should be considered for defibrillator implantation. It
would be overly simplistic to recommend that everyone with an
EF ⱕ0.30 should receive an ICD. Patients with a recent
myocardial infarction or recent coronary artery bypass graft
(CABG) surgery do not seem to benefit from ICD therapy,
according to results from the CABG Patch trial and the Defibrillator IN Acute Myocardial Infarction Trial (DINAMIT).
Such patients were excluded from eligibility in the successful
trials. Patients with major noncardiac comorbidity who were not
enrolled in the randomized ICD trials are not candidates for ICD
therapy. Within the qualifying group of patients with EF ⱕ0.30,
there are surely some low-risk patients with mortality rates so
low that ICD therapy could not possibly improve outcome. For
example, in unpublished analyses from the Multicenter Auto-
matic Defibrillator Implantation Trial (MADIT-II), routine interrogations of the ICD revealed that 35% of patients received
appropriate ICD therapy with antitachycardia pacing or shock
for potentially life-threatening ventricular tachyarrhythmias
within 3 years after device implantation. The question is how to
identify in advance patients who will and will not require ICD
therapy. To date, appropriate risk-stratification studies have not
provided validated criteria for selecting or excluding patients
with reduced EF for prophylactic ICD therapy. Our research
group has been working on this problem, and we recently carried
out a retrospective analysis of outcomes in the MADIT-II study
population.18 Preliminary findings indicate 6 independent baseline risk factors associated with mortality in the conventional
therapy group: age ⱖ72 years, EF ⱕ0.25, atrial fibrillation, New
York Heart Association class ⱖ class III, creatinine ⱖ1.4
mg/dL, and QRS ⬎0.13 seconds. The 2-year mortality rates in
those with 0, 1 to 2, and ⱖ3 of these risk factors were 3.0%,
20%, and 41%, respectively. The ICD:Conventional therapy
HRs in these 3 risk groups in the total MADIT-II population
were 3.2 (P⫽0.16) for 0 risk factors, 0.54 (P⬍0.01) for 1 to 2
risk factors, and 0.55 (P⬍0.01) for ⱖ3 risk factors. The low-risk
group with 0 of the 6 risk factors made up 20% of the population
and did not receive any benefit from ICD therapy. In fact, this
low-risk group seemed to do worse with ICD therapy, although
the difference was not statistically significant. Replication and
validation of these findings in other at-risk populations would
provide useful risk-stratification data for identifying patients
who may not benefit from an ICD.
An issue has surfaced regarding the use of QRS duration as
a major criterion for selecting at-risk patients who should or
should not receive ICD therapy. In MADIT-II, prespecified
QRS durations at ⬍0.12 seconds, 0.12 to 0.15 seconds, and
⬎0.15 seconds revealed that patients with wider QRS complexes seemed to achieve a better benefit from the ICD than
patients with narrower QRS complexes, although there were
no significant differences in ICD efficacy among these 3
subgroups.9 The US Center for Medicare & Medicaid Services carried out a retrospective, post-hoc analysis of the
MADIT-II data and reported that patients with a QRS ⬎0.12
seconds achieved a significantly greater benefit from ICD
therapy than did patients with QRS ⱕ0.12 seconds.19 A metaanalysis of relevant primary prevention trials (Figure 7) reveals
significant benefit with ICD therapy in patients with QRS
duration ⱖ0.12 seconds (HR⫽0.70, P⫽0.01, n⫽2414), and a
slightly smaller but still a significant ICD benefit in those with
QRS duration ⬍0.12 seconds (HR⫽0.82, P⬍0.03, n⫽2533).
These analyses indicate that QRS duration alone should not be
used to select patients who should receive an ICD.
Induction of ventricular tachycardia or ventricular fibrillation with programmed electrical stimulation has been used for
some time in the selection of patients for ICD therapy. In
general, inducibility is more useful in identifying patients for
risk of ventricular tachyarrhythmias in those with wellpreserved ventricular function than in those with moderately
to severely reduced ventricular function. In MADIT-II, elec-
Ejection Fraction and ICD Therapy
2547
before device replacement is required. Other analyses have
applied Markov model– based methods with nonverifiable assumptions that rely on speculative extrapolation of survival
curves and cost patterns to encompass longer time horizons.
None of the cost-effective analyses dealt exclusively with
patients having an EF ⱕ0.30. Post-trial follow-up studies need to
be performed to obtain a clearer picture of the lifetime costs and
benefits from ICD therapy.
Conclusion
Downloaded from http://circ.ahajournals.org/ by guest on June 12, 2017
The findings from relevant randomized defibrillator trials
indicate that the ICD saves lives in a spectrum of cardiac
patients with left ventricular dysfunction and an EF ⱕ0.30.
Within this large population of potential ICD recipients, some
patients should not receive an ICD because of co-morbid
conditions. Attempts to identify low-risk patients who are
unlikely to benefit from ICD therapy are in progress. At
present, I recommend that ICDs be implanted in patients with
an EF ⱕ0.30 unless there are specific contraindications
because survival is unequivocally improved with this therapy
in this patient group.
References
Figure 7. HR (ICD:conventional treatment) and 95% confidence
limits for mortality by QRS duration. The number of patients in the
DINAMIT study by QRS duration was not available, so the DINAMIT study was excluded from computation of the average HR. A,
QRS ⱖ0.12 s (average HR⫽0.70, n⫽2514, excluding DINAMIT). B,
QRS ⬍0.12 s (average HR⫽0.82, n⫽2535, excluding DINAMIT).
trophysiological testing was carried out in 593 patients who
had an ICD, and 36% of these patients were inducible into
ventricular tachycardia or ventricular fibrillation. Inducibility
into ventricular tachyarrhythmias was associated with greater
appropriate ICD therapy for ventricular tachycardia but reduced ICD therapy for ventricular fibrillation during
follow-up than patients who were not inducible. These
MADIT-II findings suggest that in patients with EF ⱕ0.30,
inducibility with electrophysiological testing is not clinically
useful in identifying patients who should receive an ICD for
the prevention of sudden cardiac death.
Cost-effectiveness data have not been used in developing my
position because existing studies are limited. Only a few costeffectiveness studies based on real data have been reported.20 –22
The primary, clinical trial analyses involve relatively short time
horizons of only 2 to 5 years after ICD implantation— considerably shorter than the assumed 5- to 7-year longevity of the ICD
1. Mirowski M, Reid PR, Mower MM, Watkins L, Gott VL, Schauble JF,
Langer A, Heilman MS, Kolenik SA, Fischell RE, Weisfeldt ML. Termination of malignant ventricular arrhythmias with an implanted automatic
defibrillator in human beings. N Engl J Med. 1980;303:322–324.
2. Preliminary report: effect of encainide and flecainide on mortality in a
randomized trial of arrhythmia suppression after myocardial infarction.
The Cardiac Arrhythmia Suppression Trial (CAST) Investigators. N Engl
J Med. 1989;321:406 – 412.
3. Risk stratification and survival after myocardial infarction. N Engl J Med.
1983;309:331–336.
4. Moss AJ, Hall WJ, Cannom DS, Daubert JP, Higgins SL, Klein H, Levine
JH, Saksena S, Waldo AL, Wilber D, Brown MW, Heo M. Improved
survival with an implanted defibrillator in patients with coronary disease at
high risk for ventricular arrhythmia. Multicenter Automatic Defibrillator
Implantation Trial Investigators. N Engl J Med. 1996;335:1933–1940.
5. Bigger JT Jr. Prophylactic use of implanted cardiac defibrillators in
patients at high risk for ventricular arrhythmias after coronary-artery
bypass graft surgery. Coronary Artery Bypass Graft (CABG) Patch Trial
Investigators. N Engl J Med. 1997;337:1569 –1575.
6. A comparison of antiarrhythmic-drug therapy with implantable defibrillators in patients resuscitated from near-fatal ventricular arrhythmias. The
Antiarrhythmics versus Implantable Defibrillators (AVID) Investigators.
N Engl J Med. 1997;337:1576 –1583.
7. Connolly SJ, Gent M, Roberts RS, Dorian P, Roy D, Sheldon RS, Mitchell
LB, Green MS, Klein GJ, O’Brien B. Canadian implantable defibrillator
study (CIDS): a randomized trial of the implantable cardioverter defibrillator
against amiodarone. Circulation. 2000;101:1297–1302.
8. Kuck KH, Cappato R, Siebels J, Ruppel R. Randomized comparison of
antiarrhythmic drug therapy with implantable defibrillators in patients
resuscitated from cardiac arrest: the Cardiac Arrest Study Hamburg
(CASH). Circulation. 2000;102:748 –754.
9. Moss AJ, Zareba W, Hall WJ, Klein H, Wilber DJ, Cannom DS, Daubert
JP, Higgins SL, Brown MW, Andrews ML. Prophylactic implantation of
a defibrillator in patients with myocardial infarction and reduced ejection
fraction. N Engl J Med. 2002;346:877– 883.
10. Kadish A, Dyer A, Daubert JP, Quigg R, Estes NA, Anderson KP,
Calkins H, Hoch D, Goldberger J, Shalaby A, Sanders WE, Schaechter A,
Levine JH. Prophylactic defibrillator implantation in patients with nonischemic dilated cardiomyopathy. N Engl J Med. 2004;350:2151–2158.
11. Connolly SJ, Hohnloser SH. DINAMIT: the defibrillator in acute myocardial infarction trial. Available at: http://www.medscape.com/
viewarticle/472003. Accessed April 22, 2004.
2548
Circulation
May 17, 2005
Downloaded from http://circ.ahajournals.org/ by guest on June 12, 2017
12. Bristow MR, Saxon LA, Boehmer J, Krueger S, Kass DA, De Marco T, Carson
P, DiCarlo L, DeMets D, White BG, DeVries DW, Feldman AM. Cardiacresynchronization therapy with or without an implantable defibrillator in
advanced chronic heart failure. N Engl J Med. 2004;350:2140–2150.
13. Bardy G. SCD-HeFT: the Sudden Cardiac Death in Heart Failure Trial.
Available at: www.sicr.org/scdheft_results_acc_lbcc.pdf. Accessed April
22, 2004.
14. Buxton AE, Lee KL, Fisher JD, Josephson ME, Prystowsky EN, Hafley
G. A randomized study of the prevention of sudden death in patients with
coronary artery disease. Multicenter Unsustained Tachycardia Trial
Investigators. N Engl J Med. 1999;341:1882–1890.
15. Buxton AE, Lee KL, Hafley GE, Wyse DG, Fisher JD, Lehmann MH,
Pires LA, Gold MR, Packer DL, Josephson ME, Prystowsky EN, Talajic
MR. Relation of ejection fraction and inducible ventricular tachycardia to
mode of death in patients with coronary artery disease: an analysis of
patients enrolled in the multicenter unsustained tachycardia trial. Circulation. 2002;106:2466 –2472.
16. Domanski MJ, Sakseena S, Epstein AE, Hallstrom AP, Brodsky MA, Kim S,
Lancaster S, Schron E. Relative effectiveness of the implantable cardioverterdefibrillator and antiarrhythmic drugs in patients with varying degrees of left
ventricular dysfunction who have survived malignant ventricular arrhythmias. AVID Investigators. Antiarrhythmics Versus Implantable Defibrillators. J Am Coll Cardiol. 1999;34:1090–1095.
17. Moss AJ. Implantable cardioverter defibrillator therapy: the sickest
patients benefit the most. Circulation. 2000;101:1638 –1640.
18. Vyas A, Moss AJ, Zareba W, Hall W, McNitt S. Risk stratification of
MADIT-II patients for predicting mortality and ICD benefit. Circulation.
2004;110:III-503.
19. Centers for Medicare & Medicaid Services. National coverage analysis:
implantable cardioverter defibrillators (#CAG-00157N). Tracking sheet.
Available at: http://www.cms.hbs.gov/ncdr/trackingsheet.asp?id⫽39.
Accessed April 22, 2004.
20. Mushlin AI, Hall WJ, Zwanziger J, Gajary E, Andrews M, Marron R, Zou
KH, Moss AJ. The cost-effectiveness of automatic implantable cardiac
defibrillators: results from MADIT. Multicenter Automatic Defibrillator
Implantation Trial. Circulation. 1998;97:2129 –2135.
21. O’Brien BJ, Connolly SJ, Goeree R, Blackhouse G, Willan A, Yee R,
Roberts RS, Gent M. Cost-effectiveness of the implantable cardioverterdefibrillator: results from the Canadian Implantable Defibrillator Study
(CIDS). Circulation. 2001;103:1416 –1421.
22. Larsen G, Hallstrom A, McAnulty J, Pinski S, Olarte A, Sullivan S, Brodsky
M, Powell J, Marchant C, Jennings C, Akiyama T. Cost-effectiveness of the
implantable cardioverter-defibrillator versus antiarrhythmic drugs in survivors of serious ventricular tachyarrhythmias: results of the Antiarrhythmics
Versus Implantable Defibrillators (AVID) economic analysis substudy. Circulation. 2002;105:2049–2057.
Response to Moss
Alfred E. Buxton, MD
D
r Moss argues that based on randomized trials, all
patients, except those with standard implantable
converter-defibrillator (ICD) contraindications,
whose ejection fraction (EF) is ⱕ30% should be treated with
an ICD. For this argument to be valid, the characteristics of
patients enrolled in these trials must be representative of
patients that we care for in practice, not simply those enrolled
in the trial. Does the evidence support this?
Except for Multicenter Automatic Defibrillator Implantation Trial-II (MADIT-II), every study cited required risk
factors other than EF: Sudden Cardiac Death in Heart Failure
Trial, Comparison of Medical Therapy, Pacing, and Defibrillation in Heart Failure, and DEFibrillators In Non-Ischemic
cardiomyopathy Treatment Evaluation (SCD-HeFT, COMPANION, and DEFINITE) required that patients have congestive heart failure. MADIT and the Multicenter Unsustained Tachycardia Trial (MUSTT) required spontaneous and
inducible ventricular tachycardia. In secondary-prevention
trials, ⬍50% of patients had an EF ⱕ30%. Thus, only
MADIT-II required an EF ⱕ30 alone.
How representative of all patients with an EF ⱕ30% is the
MADIT-II population? There is evidence of significant selection bias for recruitment into that trial. Two thirds of
patients enrolled had congestive heart failure.1 The highest
mortality in the control group occurred in patients enrolled
⬎10 years after the most recent infarction, contrary to all of
the other data (showing that highest mortality occurs within 1
year postinfarction).2 The 2-year mortality of 22% reported
by MADIT-II far exceeds that observed in population-based
postinfarct trials, such as Autonomic Tone and Reflexes After
Myocardial Infarction (ATRAMI) (⬍10%),3 and is even
higher than that observed in the control arm of the Sudden
Cardiac Death in Heart Failure Trial (SCD-HeFT) (14.4%).4
The problem is that these trials were designed to demonstrate that the ICD reduces mortality in selected populations;
they were not designed to test how best to use the ICD. That
is the task before us.
References
1. Moss AJ, Zareba W, Hall WJ, Klein H, Wilber DJ, Cannom DS, Daubert
JP, Higgins SL, Brown MW, Andrews ML; the Multicenter Automatic
Defibrillator Implantation Trial II Investigators. Prophylactic implantation of a defibrillator in patients with myocardial infarction and reduced
ejection fraction. N Engl J Med. 2002;346:877– 883.
2. Wilbur D, Zareba W, Hall W, Brown M, Lin A, Andrews M, Burke M,
Moss AJ. Time dependence of mortality risk and defibrillator benefit after
myocardial infarction. Circulation. 2004;109:1082–1084.
3. La Rovere MT, Pinna GD, Hohnloser SH, Marcus FI, Mortara A, Nohara
R, Bigger JT Jr, Camm AJ, Schwartz PJ. Baroreflex sensitivity and heart
rate variability in the identification of patients at risk for life-threatening
arrhythmias: implications for clinical trials. Circulation. 2001;103:
2072–2077.
4. Bardy GH, Lee KL, Mark DB, Poole JE, Packer DL, Boineau R,
Domanski M, Troutman C, Anderson J, Johnson G, McNulty SE, ClappChanning N, Davidson-Ray LD, Fraulo ES, Fishbein DP, Luceri RM, Ip
JH; the Sudden Cardiac Death in Heart Failure Trial (SCD-HeFT) Investigators. Amiodarone or an implantable cardioverter-defibrillator for congestive heart failure. N Engl J Med. 2005;352:225–237.
Ejection Fraction and ICD Therapy
2549
Response to Buxton
Arthur J. Moss, MD
N
Downloaded from http://circ.ahajournals.org/ by guest on June 12, 2017
ot everyone with a low ejection fraction benefits from
an implantable cardioverter-defibrillator, but considerable evidence exists that an ejection fraction ⱕ30%
identifies a meaningful proportion of patients at risk for sudden
death. Buxton argues that additional risk stratifiers beyond
ejection fraction ⱕ30% should be used, but which stratifiers?
Electrophysiological testing and increased QRS duration do not
reliably identify patients at increased risk for sudden death in this
high-risk group. Buxton suggests that an ejection fraction ⱕ30%
is an inadequate risk stratifier because it does not approach
100% sensitivity and specificity in the prediction of patients
likely to die suddenly. Medicine is replete with examples in
which good clinical practice involves overtreatment of high-risk
patients to prevent serious complications. For example, because acute appendicitis cannot be diagnosed correctly 100%
of the time, exploratory abdominal surgery is indicated when
the diagnosis is uncertain. Finding a normal appendix a
certain percentage of the time must be acceptable if the
surgeon is to prevent life-threatening problems from complicating a missed acute appendicitis. Similarly, until we have
more reliable tests to identify cardiac patients at risk for
sudden death, it is my position that in the absence of
contraindications, patients with ejection fractions ⱕ30%
secondary to ischemic or nonischemic cardiomyopathy
should receive an implantable cardioverter-defibrillator to
improve survival and prevent sudden death.
Not Everyone With an Ejection Fraction ≤30% Should Receive an ICD
Alfred E. Buxton
Circulation. 2005;111:2537-2549
doi: 10.1161/01.CIR.0000165057.88551.2C
Downloaded from http://circ.ahajournals.org/ by guest on June 12, 2017
Circulation is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 2005 American Heart Association, Inc. All rights reserved.
Print ISSN: 0009-7322. Online ISSN: 1524-4539
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://circ.ahajournals.org/content/111/19/2537
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Circulation can be obtained via RightsLink, a service of the Copyright Clearance Center, not the Editorial
Office. Once the online version of the published article for which permission is being requested is located,
click Request Permissions in the middle column of the Web page under Services. Further information about
this process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Circulation is online at:
http://circ.ahajournals.org//subscriptions/