Randomised trial of effect of amiodarone on mortality in patientswith left-ventricular dysfunction after recent myocardialinfarction: EMIAT
D G Julian, A J Camm, G Frangin, M J Janse, A Munoz, P J Schwartz, P Simon, for the European Myocardial Infarct
Background Ventricular arrhythmias are a major cause of
Ventricular arrhythmias are one of the main causes of
death after myocardial infarction, especially in patients
death in survivors of acute myocardial infarction. The
w it h poor left -v ent ric ular func t ion. P rev ious at t empt s
European Myocardial Infarct Amiodarone Trial (EMIAT)
t o i d e n t i f y a n d su p p r e ss a r r h y t h m i a s w i t h v a r i o u s
was conceived at the end of 1988, after many studies of
antiarrhythmic drugs failed to reduce or actually increase
antiarrhythmic drugs that block the sodium-channel (classI) showed no efficacy in the suppression of arrhythmias
mortality. A miodarone is a pow erful antiarrhythmic drug
and prevention of death in survivors of myocardial
with several potentially beneficial actions, and has shown
infarction.1–3 Later, the Cardiac Arrhythmia Suppression
benefit in several small-scale studies. We postulated that
Trials proved that in survivors of myocardial infarction
this drug might reduce mortality in patients at high risk of
at low risk of death, the suppression of ventricular
deat h aft er my oc ardial infarc t ion bec ause of impaired
extrasystoles with the sodium-channel-blocking drugs
v ent ric ular func t ion, irrespec t ive of w het her t hey had
flecainide and encainide, and possibly moricizine, was
associated with excess mortality.4,5 Apart from ␤-
Methods The European Myocardial Infarct Amiodarone Trial
adrenergic blocking agents, the only antiarrhythmic drug
(EMIAT) was a randomised double-blind placebo-controlled
that showed any promise in the late 1980s was
t rial t o assess w het her amiodarone reduc ed all-c ause
amiodarone, although the number of patients studied was
mort alit y ( primary endpoint ) and c ardiac mort alit y and
too small to allow definite conclusions about the drug’s
arrhy t hmic deat h ( sec ondary endpoint s) in survivors of
effect on all-cause mortality.6–8 We, therefore, decided to
my oc ardial infarc t ion w it h a left -v ent ric ular ej ec t ion
conduct a large trial of amiodarone in survivors of
fraction ( LVEF) of 40% or less. Intention-to-treat and on-
myocardial infarction at increased risk of death.9,10
Although the presence of frequent or complex
ventricular arrhythmias is an independent risk factor for
Findings E M I A T enr ol l ed 1486 pat i ent s ( 743 i n t he
mortality after myocardial infarction,11–13 and most trials
amiodarone g roup, 743 in t he plac ebo g roup) . M edian
of antiarrhythmic drugs have used such ventricular
follow-up was 21 months. All-cause mortality (103 deaths
arrhythmias as an entry criterion, there is no evidence
in the amiodarone group, 102 in the placebo group) and
that suppression of ventricular arrhythmias leads to a
cardiac mortality did not differ betw een the tw o groups.
reduction in overall mortality. Moreover, the single
However, in the amiodarone group, there was a 35% risk
most powerful independent predictor of mortality,
reduction (95% CI 0–58, p=0·05) in arrhythmic deaths.
including sudden death, is left-ventricular dysfunction.13,14
Interpretation Our findings do not support the systematic
Thus, we decided to adopt a different approach from the
pr ophy l ac t i c use of ami odar one i n al l pat i ent s w i t h
other trials by taking no account of the presence of
depr essed l eft - v ent r i c ul ar func t i on aft er my oc ar di al
symptomless arrhythmias and by choosing depressed left-
infarc t ion. How ever, t he lac k of proarrhy t hmia and t he
ventricular ejection fraction (LVEF) as the main entry
r e d u c t i o n i n a r r h y t h m i c d e a t h su p p o r t t h e u se o f
criterion. We expected that most patients with left-
amiodarone in patients for whom antiarrhythmic therapy is
ventricular dysfunction would subsequently suffer from
lethal arrhythmias, even if ventricular arrhythmias werenot initially present. In addition to its antiarrhythmic
properties, amiodarone has anti-ischaemic actions and
does not aggravate heart failure, properties that may havean additional beneficial effect on survivors of myocardial
*Investigators listed at end of paper
Netherhall Gardens, London (Prof D G Julian MD); Cardiological
The aim of this randomised placebo-controlled double-
Sciences, St George’s Hospital Medical School,
blind trial was to assess the effect of amiodarone on all-
London SW17 0RE, UK (Prof A J Camm MD); Sanofi Recherche,
cause mortality, cardiac mortality, and arrhythmic death
Montpellier, France (G Frangin MD, A Munoz MD, Prof P Simon MD);
in survivors of myocardial infarction with depressed left-
Department of Clinical and Experimental Cardiology,Academic Medical Centre, Amsterdam, Netherlands
(Prof M J Janse MD); and Department of Cardiology, University of
EMIAT was a joint venture between the Working
Group on Arrhythmias of the European Society of
centre. The lower limit of the normal range was 40% or more forall centres.
We recruited patients for this randomised double-blind placebo-
Patients were then randomly assigned amiodarone or placebo.
controlled trial from the coronary-care units of participating
The daily regimen was 800 mg for 14 days, 400 mg for 14 weeks,
European hospitals between Nov 30, 1990, and Oct 30, 1995. A
and then 200 mg until the end of the study follow-up. The
log-book was kept of all patients with documented myocardial
maximum follow-up was 2 years and the minimum follow-up
infarction surviving 5 days. Eligible patients were those aged
1 year, according to the time of enrolment.
18–75 years who had a LVEF on multiple-gated nuclear
Patients were assessed at baseline, 2 weeks, 2 months, and
angiography (MUGA) of 40% or less. MUGA was done 5–21
then every 4 months up to 2 years. 24 h ambulatory
days after admission to the coronary-care unit.
electrocardiographic (Holter) monitoring was done at baseline,
We excluded women of childbearing age who were not using
2 weeks, and 4 months. Chest radiography was done at baseline,
reliable contraception and individuals who had: treatment with
2 months, and annually. Laboratory data, including
amiodarone in the previous 6 months; documented bradycardia
measurement of serum concentrations of potassium creatine
(under 50 beats per min); second-degree or third-degree
kinase, aspartate aminotransferase, alanine aminotansferase, and
atrioventricular block; sinus pauses of more than 2·5 s unless
thyrotropin, were collected at baseline, 2 weeks, and at months 2,
controlled by a pacemaker; clinically significant hepatic disease; a
history of documented thyroid dysfunction, long QT syndrome,
The investigators notified the Coordinating Centre about any
severe angina or congestive heart failure refractory to
deaths and major adverse side-effects by fax within 24 h of the
conventional therapy; a need for antiarrhythmic therapy other
event. The Coordinating Centre forwarded these data to the
than ␤-blockers or digoxin; a likelihood of imminent cardiac
surgery; and other contraindications for amiodarone.
The Validation Committee reviewed deaths under masked
Computer-generated randomisation was done in balanced
conditions. We used the following criteria to classify the cause of
blocks of four patients, and lists were prepared and kept at the
Independent Statistical Centre (Clinical Pharmacology Unit,
Cardiac deaths were grouped into three categories: “sudden”,
Claude Bernard University, Lyon). This centre was also
“non-sudden”, or “unwitnessed” presumed cardiac. Sudden
responsible for conducting the interim analyses of efficacy and
cardiac deaths occurred within 1 h of new symptoms, or in a
safety, and for packaging the study treatments. Stratification was
patient with no symptoms or stable symptoms, and with no left-
based on clinical centre and ejection fraction. Treatment
ventricular failure. Non-sudden cardiac deaths occurred more
allocation was assigned under masked conditions by the EMIAT
than 1 h after the onset of symptoms. Unwitnessed, presumed
Coordinating Centre (Sanofi, Montpellier), and sent by fax to
cardiac deaths were unexpected and not witnessed, within 24 h
the investigators. The Coordinating Centre had no access to the
of the patient’s being known to be well and with no symptoms or
Before randomisation, patients from each centre were
Sudden and non-sudden cardiac deaths were grouped into
stratified according to their ejection fraction—31–40% and 30%
documented, non-documented arrhythmic deaths, or non-
or less. The accuracy of the MUGA equipment in each centre
arrhythmic deaths, such as rupture and electromechanical
was validated by the Ejection Fraction Committee. The normal
dissociation. Unwitnessed deaths were presumed to be cardiac if
range for ejection fraction was determined individually for each
there was no alternative diagnosis. We defined sudden deathfrom myocardial infarction as a cardiac death, but not as an
The Validation Committee also reviewed, under masked
conditions, all arrhythmic events, suspected cases of pulmonary
toxicity, biochemical alterations of thyroid hormones and liverenzymes, as well as evidence of clinical dysthyroidism andhepatic disorders.
The ambulatory electrocardiographic records were analysed
centrally at the Holter Reading Centre.
The study protocol was approved by an ethics committee in
each country or centre, and all patients had to give informedconsent to take part in the trial.
Study medication was stopped in patients with sustained
ventricular tachycardia (30 s), symptomatic unsustainedventricular tachycardia, pulmonary infiltrate with no specificclinical cause, or intolerable side-effects. However, all patients
were followed up and included in the intention-to-treat analysis. Compliance was assessed by follow-up visits and tablet counts. Data on mortality was sought for all patients at the end of theplanned follow-up.
The sample size calculations were based on a 15% 2-year
mortality rate in the placebo group and a 35% reduction of riskin the amiodarone group, with type I and type II erorr rates of
0·05 and 0·20, respectively. These calulations indicated that1500 patients should be enrolled.
The primary analyses were by intention to treat, but we also
did an on-treatment (efficacy) analysis of outcome events in
eligible patients while on study medicaton or within 3 months ofearly permanent discontinuation. The primary endpoint was all-cause mortality and the secondary endpoints were cardiac
mortality, arrhythmic death, and arrhythmic death plus
At baseline (day of enrolment) continuous and categorical
variables were compared by the t test and 2 test, respectively. Survival curves of the proportion of patients who remained
event-free, according to treatment group and ejection-fraction
Mean (SD) creatine kinase activity (IU/L)
Mean (SD) systolic blood pressure (mm Hg)
VPB=ventricular premature beats. ACE=angiotensin-converting enzyme.
Table 1: Baseline characteristics of patients
stratum, were calculated by the Kaplan-Meier method15 and
compared by the log-rank test stratified by ejection fraction. We
calculated the risk ratio and 95% CI values by Cox’sproportional hazards model,16 stratified by ejection fraction.
Cox’s proportional hazards model was also used to adjust for
baseline imbalances in prognostic variables between groups.
Rules for censoring of data were defined before the
Figure 2: Kaplan-Meier estimates of all-cause mortality by
randomisation code was broken. For the intention-to-treat
analysis, the follow-up was censored on the date of the last
treatment within the previous 6 months (142), essential
planned visit for complete follow-up, the end of the trial, or thedate of death. For the on-treatment analysis, data were censored
antiarrhythmic treatment (142), and other contra-
at 3 months after permanent discontinuation of study
indications (406). Thus, 1486 patients were considered
medication, or at the end of the follow-up as defined above.
eligible, gave informed consent to participate, and were
Formal interim analyses were planned by the Safety
Committee and done after the occurrence of every 60 deaths,with total mortality as the endpoint. The stopping guidelines for
mortality were based on asymmetrical boundaries for benefit andharm.17 We followed the Peto18 approach with statistical
guidelines for efficacy fixed at p<0·001. For safety monitoring, a
report was produced every 4 months, and a fixed type-I error rate
of 0·01 (one sided) was chosen as a statistical guideline. The
committee also reviewed other adverse side-effects.
The Safety Committee did three interim analyses and
recommended continuation of the study on each occasion.
However, one analysis showed higher mortality among patients
in the amiodarone group who were receiving digoxin or digitoxinthan among those who were not; this difference in mortality was
Intention-to-treat analysis by EFS and history of MINumber in analysis
not significant. The Steering Committee was informed and
advised investigators to adhere more strictly to the protocol,
which stated that the dosage of digoxin or digitoxin should be
halved on enrolment. This adverse trend did not continue.
We recruited 26 493 patients from 75 centres in 15
European countries. 7565 of the patients underwent a
MUGA scan, and in 3255 (43%) the LVEF was 40% or
less. We excluded 1769 patients with ejection fractions in
the acceptable range because of lack of consent (409),
EFS=ejection-fraction stratum; MI=myocardial infarction.
imminent cardiac surgery (286), other serious illness
*Total patient-year ex posure: placebo=1206; amiodarone=1065.
(205), congestive cardiac failure (179), amiodarone
LVF=left-venticular failure; CS=cardiogenic shock; EMD=electromechanical
Table 3: Non-arrhythmic cardiac deaths and non-cardiac deaths
p=0·67). By contrast, there was a 35% risk reduction(95% CI 0–58, p=0·05) in arrhythmic deaths among
amiodarone-treated patients. A similar risk reduction was
observed after the addition of resuscitated cardiac arrest(table 2, figure 3). There were more deaths from non-
arrhythmic cardiac and non-cardiac causes in amiodarone-
group patients than in placebo-group patients (table 3).
A history of myocardial infarction significantly increased
the risk of death (24% vs 10%, p<0·0001). We, therefore,did a further analysis to adjust for this variable and the
baseline imbalances in prognostic variables. The adjusted
differences in all-cause and cardiac mortality between the
treatment groups were not significant (table 2).
Baseline ambulatory electrocardiograms were evaluable
in 1367 patients, of whom 191 died. The mortality rate
was higher in those with frequent or complex arrhythmias
than in those without arrhythmias (112/548 [20%] vs
79/819 [10%]. However, the group without arrhythmias
was larger and comprised 60% of the trial population. Of the 191 deaths, 79 (41%) occurred in the group
Figure 3: Kaplan-Meier estimates of arrhythmic deaths and
without baseline arrhythmias, and among these deaths 36
resuscitated cardiac arrest by group and ejection fraction
(45%) were classified as arrhythmic. In intention-to-treatanalysis of the subgroup of 548 patients with arrhythmias
enrolled in the trial, with a mean time of 15 days (Ϯ3·9)
at baseline (40% of the trial population), the difference
after the index myocardial infarction. There were 743
between the amiodarone and placebo groups in
patients in the placebo group and 743 patients in the
arrhythmic deaths did not achieve significance, but there
amiodarone group. Subsequently, we found that four
was a significant reduction in the combined endpoint
patients (two in the amiodarone group, two in the placebo
of arrhythmic deaths and resuscitated cardiac deaths
group) did not fulfil all the eligibility criteria, but they
(p=0·048) in amiodarone-treated patients compared with
were retained in the trial. The trial profile shows overall
patient numbers during the study (figure 1).
During the trial, 284 (38·5%) amiodarone-group
The baseline characteristics of patients in the two
patients discontinued their study medication compared
groups were broadly similar, but there were some
with 158 (21·4%) placebo-group patients (figure 4). The
between-group differences that may have affected
main causes of discontinuation of study medication are
outcome—eg, ejection fraction, New York Heart
shown in table 5. There was no significant difference in
Association functional class, peak concentrations ofcreatine kinase, and a history of myocardial infarction
(table 1). Of especial interest was the difference in
previous myocardial infarction: 191 (26%) placebo-group
patients versus 233 (32%) amiodarone-treated patients.
Two patients were lost to follow-up and 1484 patients
were followed up to the end of the trial. 205 patients died.
The Validation Committee classified 31 of these deaths as
non-cardiac and 174 as cardiac. Of the 174 cardiac
deaths, 83 (48%) were arrhythmic (figure 1). In addition,
there were 20 survivors of resuscitated cardiac arrest.
All-cause mortality did not differ between the treatment
groups (risk ratio 0·99, p=0·96), even after patients were
grouped by ejection fraction (figure 2). Similarly, there
was no difference in total cardiac mortality (risk ratio 0·94,
Table 4: Causes of early discontinuation of study medication
thyroid disorders were common, but in the amiodaronegroup, clinical hypothyroidism and hyperthyroidism were
observed in only 11 (1·5%) and 12 (1·6%) patients,
respectively. No torsade de pointes was documented in
As expected in survivors of myocardial infarction,
several other drugs were used concomitantly throughout
the trial (table 1). After the reports of the benefits of
inhibitors of angiotensin-converting enzyme (ACE) were
published, in 1992, the proportion of patients receiving
these agents increased. Concomitant treatment with ACE
inhibitors, diuretics, calcium antagonists, digoxin, ordigitoxin had no effect on differential treatment group
mortalities. However, we found a strong tendency towards
favourable interaction between use of ␤-blockers and
cardiac mortality (figure 5), independently of left-
Figure 4: Percentage of patients continuing study medication
EMIAT showed that amiodarone was associated witha significant reduction in arrhythmic deaths and
resuscitated cardiac arrests among patients discharged
from hospital with depressed left-ventricular function afterrecent myocardial infarction. However, neither asignificant nor corresponding reduction in all-cause or
total cardiac mortality was seen because of increasednon-cardiac, or cardiac but not arrhythmic, mortality.
Thus, EMIAT does not support the systemic prophylacticuse of this regimen of amiodarone in survivors of
myocardial infarction. Nevertheless, amiodarone therapydoes lead to a reduction in arrhythmic death with no
proarrhythmic effect, only a few minor side-effects, and aneutral effect on total mortality.
Our findings warrant a discussion of the rationale and
background of this trial, the factors that might explain its
0·2 0·4 0·6 0·8 1·0 1·2 1·4 1·6 1·8 2·0 2·2
outcome, and our interpretation of the findings.
Patients recovering from myocardial infarction are at
substantial risk of sudden arrhythmic death and
Figure 5: Intention-to-treat analysis of 2-year total cardiac
symptomatic arrhythmic events.1,2 There is a sound
mortality by concomitant medication at baseline
rationale for the use of antiarrhythmic drugs in suchpatients. However, the results of trials of antiarrhythmic
the major risk factors for mortality and arrhythmic events
drugs have mostly been adverse:4,5,19,20 antiarrhythmic drug
between those patients who continued amiodarone
therapy was associated with increased deaths, in most
cases sudden and presumably arrhythmic, that were
The on-treatment analysis showed a more striking risk
attributed to a proarrhythmic effect. Yusuf and colleagues’
reduction than was shown in the intention-to-treat
meta-analysis21 of trials of antiarrhythmic drugs after
analysis in arrhythmic deaths among the patients taking
myocardial infarction showed that sodium-channel
amiodarone compared with those taking placebo (table 2).
blockers were harmful and that calcium-channel
Adverse side-effects are shown in table 7. There were
antagonists had no beneficial effects; by contrast, a
three deaths in the amiodarone group from pulmonary
favourable effect was observed with the use of amiodarone
fibrosis, which was confirmed on necropsy; however, two
in a small population. Similarly, Teo and colleagues’
of these patients had pre-existing pulmonary disease and
meta-analysis22 of placebo-controlled trials of amiodarone
should not have been included in the trial. Biochemical
after myocardial infarction or in patients with heart failuresuggested that survival substantially improved with the use
of amiodarone. The lack of efficacy of sodium-channel
blockers has been variously ascribed to proarrhythmic and
negative inotropic effects.23 An interaction with ischaemia
has also been proposed.24 On the other hand, although
amiodarone is predominantly thought of as an
antiarrhythmic drug, it is also an anti-ischaemic agent.25,26
Importantly, amiodarone seems to have little or no
proarrhythmic potential. Thus, we decided that it was
appropriate to design and conduct a mortality trial with
amiodarone among survivors of myocardial infarction at
*Cases assessed by Validation Committee. †Central and peripheral.
Unlike the antiarrhythmic drugs used in previous trials,
mainly sodium-channel and potassium-channel blockers,
endpoint analyses, a conventional two-sided test of
amiodarone has several actions.27 Amiodarone has a
significance was judged to be essential, since previous
blocking effect on sodium, calcium, and potassium
studies of antiarrhythmic drugs had clearly shown the
channels, and also has antiadrenergic and anti-ischaemic
possibility of an adverse outcome associated with the
effects. Lethal arrhythmias in postinfarction patients are
active agent. Thus, we calculated that 1500 patients were
likely to be the result of re-entry.28 In our study, the
needed. In fact, we recruited 1486 patients, and our
prolongation of the ventricular refractory period combined
a-priori assumptions proved to be correct: mortality in the
with adrenergic blockade might have contributed to the
placebo group was 14% (102 of 743), and the arrhythmic
reduction in arrhythmic deaths in the amiodarone-treated
deaths constituted half (49%) of the placebo-group
mortality (50 of 102). Thus, EMIAT had sufficient power
EMIAT showed that amiodarone therapy had no effect
to detect a 35% reduction in mortality in the amiodarone
on the primary endpoint of 2-year all-cause mortality,
group. However, a much larger trial would have been
irrespective of the stratum of left-ventricular function.
necessary to detect the small apparent reduction in all-
However, amiodarone was associated with a reduction in
arrhythmic deaths among high-risk patients after
We assessed left-ventricular function accurately by
myocardial infarction. Our finding of a significant
MUGA scanning, which was well validated and approved
reduction in arrhythmic deaths and resuscitated cardiac
centrally. Unlike some previous studies of antiarrhythmic
arrests in the amiodarone group suggests that any possible
drugs such as the Cardiac Arrhythmias Suppression Trial,4
proarrhythmic effects of amiodarone do not outweigh its
in which mortality among placebo-group patients was
antiarrhythmic benefit. But this reduction in arrhythmic
lower than expected, the selection process in EMIAT
deaths was balanced by an excess of five non-cardiac
resulted in the anticipated number of trial endpoints.
deaths and 13 cardiac but non-arrhythmic deaths. This
EMIAT showed that many arrhythmic deaths occurred in
excess may have resulted from chance or from the
patients who did not have frequent or complex ventricular
imbalance in significant prognostic factors between the
ectopic activity at baseline. Thus, if we had used such
groups, but may also have been partly attributable to the
ectopic activity as the only entry criterion, the power of the
adverse side-effects of amiodarone, such as, the three
trial would have been reduced. Other stratification
deaths caused by pulmonary fibrosis.29 In our study, fatal
criteria, such as heart rate variability31,32 or baroreceptor
reinfarction contributed to the non-arrhythmic mortality
sensitivity,33 which are more likely to predict the
associated with amiodarone treatment. This finding is,
occurrence of arrhythmic events could have been used but
at first, difficult to explain because amiodarone is an
were not fully appreciated when EMIAT was planned.
But since there was a significant reduction in arrhythmic
reinfarctions (11 of 13) occurred in patients who had a
events with amiodarone, it seems unnecessary to speculate
history of myocardial infarction before the index
whether better selection for arrhythmic risk would have
infarction, a group that was over-represented in the
changed the results of the trial. Nonetheless, it might have
amiodarone group. Another explanation for the failure of
been appropriate to exclude patients at risk of early death
amiodarone to reduce non-arrhythmic death is that, by
because of pump failure (very low ejection fraction) or
preventing arrhythmic death, amiodarone treatment
reinfarction (adverse coronary anatomy). Such exclusion
did not preclude death from other non-arrhythmic
criteria may have reduced the cardiac but non-arrhythmic
mechanisms. The results of EMIAT underline the
mortality associated with amiodarone therapy.
importance of using all-cause mortality, rather than an
In some previous trials of postinfarction antiarrhythmic
antiarrhythmic effect, as the primary endpoint in a survival
drugs, patients were recruited up to 1 year after the index
infarction. However, most patients die within the first few
After adjustment for the imbalance in covariates
months after infarction. Thus, in EMIAT patients were
between the groups, there was some evidence of a
enrolled and randomly allocated treatment before
reduction in all-cause and total cardiac mortalities. Is this
discharge from hospital. Furthermore, the loading dose of
reduction real and would it have been statistically
study medication was administered quickly to ensure drug
significant if the trial had been larger? The original power
activity before discharge. Early recruitment plus high
calculations were based on the pilot study of the Canadian
dosing may have accounted for the increased early
Arrhythmia Myocardial Infarction Amiodarone Trial,8 the
mortality in patients with very poor ventricular function
Basel Antiarrhythmia Study of Infarct Survival,6 and the
(figure 2), which contrasted with the antiarrhythmic
partially completed Polish Amiodarone Trial7 which
advantages associated with amiodarone.
showed that an all-cause mortality reduction of 35–50%
Many of the adverse side-effects associated with
might be achieved with amiodarone. We also knew, at the
amiodarone are sufficiently troublesome to force
time when EMIAT was planned, that poor ventricular
discontinuation of the drug.34 In our study clinically
function was a more powerful predictor of mortality than
relevant pulmonary,29 hepatic,35 and thyroid disorders36,37
was the frequency of ventricular ectopic activity. Two
occurred in a small proportion of amiodarone-treated
studies showed that patients with an LVEF below 40%
patients, and were much as we had anticipated, for
would have a 2-year mortality of 15%, of which the
example, there was no torsade de pointes.38 However,
arrhythmic component would be about half.13,14 However,
many biochemical abnormalities and minor side-effects
there were only limited and inconsistent data on the effect
occurred, which accounted for the high rate of
of amiodarone on the mortality of patients with poor left-
discontinuation of study medication.
ventricular function. We, therefore, elected to recruit
This high rate of discontinuation suggests that although
patients on the basis of poor ventricular function alone
the primary analysis should be by intention to treat, a
and, assuming a 2-year mortality of 15% in the placebo
second on-treatment (efficacy) analysis should also be
group, designed the study to detect a 35% reduction in
done. The effect of amiodarone continues after
all-cause mortality in the amiodarone group. For the
discontinuation of the drug. We, therefore, decided to
conduct the on-treatment analysis by censoring patients
(B Brechenmacher, J B Caillard), Tours (B Charbonnier, G Pacouret),
from further analysis for 3 months after permanent early
Vandoeuvre (B Brembilla Perrot, N Danchin); Germany: Berlin(D Andresen), Cottbus (U Kreutzer, J Kolditz), Düren (H Simon,
discontinution of study medication. But these analyses
B Zentgraf), Hamburg (T Meinertz, M Volkmer, K Kuck, M Antz),
may be of limited value because such a large proportion of
Hannover (H Klein, G Claus), Heidelberg (J Brachmann, K Freigang),
patients (amiodarone 38·5%, placebo 21·4%)
Lübeck (J Potratz, G Taubert), Magdeburg (H Klein, S Grund), Mainz(N Treese, A Liebrich), München (L Goedel-Meinen, M Hofmann,
discontinued study medication. In any event, both
F G Nowak, K Coppenrath), Münster (M Borggrefe, J Brunn);
Greece: Athens (S Moulopoulos, A Antoniou), Pireus (D V Cikkinos,
Analyses of the interaction between use of concomitant
G Athanassopoulos); Italy: Bari (P Rizzon, M Di Biase), Bologna(D Bracchetti, F Naccarella, B Magnani, A Capucci), Mestre (E Piccolo,
medications at baseline and the mortality endpoints
A Raviele), Milano (A Lotto, A Finzi), Varese (G Binaghi, A M Giorgio);
suggested that there was an important interaction with
Netherlands: Amsterdam (A Vermeulen), Capelle Aan Den Yssel
␤-blockers. There were fewer cardiac deaths among
(W M Muijs van der Moer, S L Nio), Deventer (D J A Lok,M J C Osinga-Meek), Dordrecht (J B L ten Kate), Dordrecht (P Breuls),
amiodarone-treated patients who were receiving
Heerlen (J A Kragten, J H M de Warrimont-Henquet), Rotterdam
concomitant ␤-blockers than among those who were not,
(D C A Von Hoogenhuyze), Tilburg (N J Holwerda); Norway: Bergen
probably because ␤-blockers were not prescribed to
(O Johm, K Breivik); Poland: Warsaw (L Ceremuzynski, T Chamiec);
patients at high risk of cardiac death. However, of those
Portugal: Coimbra ( LProvidência, G Alves), Lisboa (C Ribeiro,A Bordalo); Spain: Barcelona (J Cinca, A Moya, A Bayes de Luna,
patients who receive ␤-blockers, there was a substantial
X Viñolas), Madrid (C Moro Serrano, A Hernandez-Madrid), Valencia
reduction in cardiac and arrhythmic mortality among
(J Cosin, M J Sancho-Tello, A Salvador, M Amela); Sweden: Lund
those who were also given amiodarone. This finding
(B Olsson, E Pantev); Switzerland: Basel (F Burkart, M Pfisterer), Bern(M Gertsch, T Chatterjee); Geneva (W Rütishauser, M Zimmermann),
indicates that ␤-blockers confer additional benefit to the
Lausanne (L Kappenberger, M Framer), Sion (P Vogt, D Fellay), Zurich
efficacy of amiodarone, a finding also reported for the
(F W Amann, E Schuiki); UK: Derby (M W Millar-Craig, A V Joy),
treatment of sustained ventricular tachyarrhythmias.39 We
Glasgow (R J Northcote), Gwent (J Davies, W Covell), Kent (B Gould,A Brian), Isleworth (T W Greenwood, E Ramhamadany), Leicester
did not anticipate this interaction before the start of the
(J Skehan, Y Haider), London (A J Camm, H A Staunton, J Cleland,
trial, and its implications must be interpreted with
Steering Committee—D G Julian (chairman), A J Camm, G Frangin,
Our findings do not support the systematic prophylactic
M J Janse, A Munoz, P J Schwartz, P Simon.
use of amiodarone in patients with poor left-ventricular
Safety Committee—L Wilhelmsen (chairman), J P Boissel, S Pocock,
function after myocardial infarction, irrespective of the
presence of symptomless ventricular ectopic activity.
Validation Committee—F Burkart (chairman), R W Campbell
However, the 35% risk reduction of fatal and resuscitated
(co-chairman), S Levy, N Rehnqvist, G Frangin (secretary non-voting).
arrhythmic events associated with amiodarone shows that
this drug affords protection from arrhythmic death in this
Ex ecutive Committee—H Klein (chairman), G Frangin (co-chairman),
high-risk population. In addition, no proarrhythmia was
associated with amiodarone treatment. Thus, it may be
Electrocardiography Monitoring Committee—G Breithardt, R Brion,
possible to identify groups of patients at high risk of
arrhythmia for whom amiodarone will offer a substantial
Ejection Fraction Committee—M M Pfisterer (chairman), D Dymond.
survival benefit. Since we found such a substantial
Coordinating Centre—Sanofi Recherche, Montpellier, France: G Frangin
antiarrhythmic effect, balanced by non-arrhythmic
(study director), E Bienvenu (administration), C Dechamp (international
mortality, it may also be possible to achieve a significant
CRA), A d’Halliun-Sulzer (event monitoring), J Galleyrand (statistician),R Mercier (data manager), A Pelizza (international CRA).
reduction in all-cause mortality among infarct survivors athigh risk of sudden arrhythmic death with a different drug
Independent Statistical Centre—J P Boissel, A Leizorovicz, F Boutitie.
regimen, such as a lower loading dose, and a shorter
Holter Reading Centre—R Brion, G Gonnot, M Brion.
Decoding Centre—J Descotes, J M Sapori.
Our findings contrast with the disappointing results of
Signal Average Electrocardiography Analysis—G Breithardt, T Fetsch.
most previous trials of antiarrhythmic drugs aftermyocardial infarction. Thus, the clinician may be
Heart Rate Variability Analysis—A J Camm, M Malik.
encouraged to consider amiodarone for patients with
Country Monitor Group—Clinical Research Unit Sanofi: M Brunet, I
symptomatic or sustained and potentially dangerous
Carbarns, K Handelberg, P Montanari, A Müller, G Munoz, N Nikkitas,L Orö, E Ribiero, I Tjong-A-Hung, F Wider; Clinical Research
arrhythmias, even after myocardial infarction,
Although subgroup analysis suggests that amiodarone
Drug Supply—Sanofi Recherche, Montpellier, France.
may be useful in some patients, definitive indications willrequire further clinical trials. Our results must be
interpreted together with those of other placebo-
We thank the study patients for their cooperation; Genevieve Fauris,Heather Switzman, and the study personnel for their assistance;
controlled amiodarone survival studies. Review of the
Lars Wilhelmsen and Florent Boutitie for their expert help with the
pooled data may suggest new or prophylactic indications
preparation of the manuscript; Jacques Galleyrand for the statistical
for the use of amiodarone, but such conclusions must
analysis; and Caroline Smeeden for secretarial assistnce.
This study was supported by a grant from Sanofi Recherche,
await formal meta-analysis or further clinical trials.
European Myocardial Infarct Amiodarone Trial Study Participants—
Austria: Innsbruck (F Dienstl, A Dienstl), Wien (K Steinbach,M Nünberg, H Weber, T Binder); Belgium: Antwerpen (R Ranquin,
Impact Research Group. International mexiletine and placebo
C Convens), Brugges (Y Vandekerckhove, E Van der Stichele), Bruxelles
antiarrhythmic coronary trial: I report on arrhythmic and other
(I Blankoff, S Carlier, A Depaepe, C Gassavolis), Gent (L Jordaens,
findings. J Am Coll Cardiol 1984; 4: 1148–63.
R Tavernier), Liège (J Boland, J L Waucquez); Finland: Espoo
Furburg CD. Effect of antiarrhythmic drugs on mortality after
(U Idänpâän-Heikkilä, R Sipilä); France: Bayonne (J Kayanakis), Clamart
myocardial infarction. Am J Cardiol 1983; 52: 32C–36C.
(G Motte, C Sebag), Lille (J Lekieffre, L Guedon), Lyon (P Touboul,
Gottlieb SH, Achuff SC, Mellits ED, et al. Prophylactic
B de Breyne, M Gallet, M Lopez), Montpellier (R Grolleau-Raoux,
antiarrhythmic theapy of high-risk survivors of myocardial infarction:
F Leclercq), Nancy (E Aliot, N Sadoul), Nantes (H Le Marec,
lower mortality at 1 month but not at 1 year. Circulation 1987; 75:
J C Chevallier), Rennes (J C Daubert, J M Baisset), Schiltigheim
The Cardiac Arrhythmias Suppression Trial (CAST) Investigators.
23 Greenberg HM, Dwyer EM Jr, Hochman JS, Steinberg JS, Echt DS,
Preliminary report: effect of encanide and flecainide on mortality in a
Peters RW. Interaction of ischaemia and encainide/flecainide
randomized trial of arrhythmia suppression after myocardial infarction.
treatment: a proposed mechanism for the increased mortality in CAST
N Engl J Med 1989; 321: 406–12.
I. Heart 1995; 74: 631–35.
The Cardiac Arrhythmia Suppression Trial II Investigators. Effect of
24 Task Force of the Working Group on Arrhythmias of the European
the antiarrhythmic agent moricizine on survival after myocardial
Society of Cardiology. CAST and beyond: implications of the Cardiac
infarction. N Engl J Med 1992; 327: 227–35.
Arrhythmia Suppression Trial. Eur Heart J 1990; 11: 194–99 and
Burkart F, Pfisterer M, Kiowski W, Follath F, Burckhardt D. Effect of
Circulation 1990; 81: 1123–27.
antiarrhythmic therapy on mortality in survivors of myocardial
25 Ellenbogen KA, O’Callaghan WG, Colavita PG, Smith MS, German
infarction with asymptomatic complex ventricular arrhythmias: Basel
LD. Cardiac function in patients on chronic amiodarone therapy. Am
Antiarrhythmic Study of Infarct Survival (BASIS). J Am Coll CardiolHeart J 1985; 110: 376–81.
1990; 16: 1711–18.
26 Gottlieb SS, Riggio DW, Lauria S, et al. High dose oral
Ceremuzynski L, Kleczar E, Kreminska-Pakula M, et al. Effect of
amiodarone loading exerts important hemodynamic actions in
amiodarone on mortality after myocardial infarction: a double blind,
patients with congestive heart failure. J Am Coll Cardiol 1994; 23:
placebo-controlled, pilot study. J Am Coll Cardiol 1992; 20: 1051–62.
Cairns JA, Conolly SJ, Gent M, Roberts R. Post-myocardial infarction
27 Task Force of the Working Group on Arrhythmias of the European
mortality in patients with ventricular premature depolarizations: the
Society of Cardiology. The Sicilian gambit: a new approach to the
Canadian amiodarone myocardial infarction arrhythmia trial.
classification of antiarrhythmic drugs based on their actions on
Circulation 1991; 4: 550–57.
arrhythmogenic mechanisms. Eur Heart J 1994; 15: 721–38 and
Schwartz PJ, Camm AJ, Frangin G, Janse MJ, Julian DG, Simon P. Circulation 1994; 89: 2892–907.
Does amiodiarone reduce sudden death and cardiac mortality after
28 Janse MJ, Wit AL. Electrophysiological mechanisms of ventricular
myocardial infarction? The European Myocardial Infarct Amiodarone
arrhythmias resulting from myocardial ischemia and infarction. Physiol
Trial (EMIAT). Eur Heart J 1994; 15: 620–24. Rev 1989; 69: 1049–169.
10 Camm AJ, Julian DG, Janse MJ, et al. The European Myocardial
29 Dusman RE, Stanton MS, Miles WM, et al. Clinical features of
Infarct Amiodarone Trial (EMIAT). Am J Cardiol 1993; 72: 95F–98F.
amiodarone-induced pulmonary toxicity. Circulation 1990; 82: 51–59.
11 Ruberman W, Weinblatt E, Goldberg JD, et al. Ventricular premature
30 Meyer BJ, Amann FW. Additional antianginal efficacy of amiodarone
beats and mortality after myocradial infarction. N Engl J Med 1977;
in patients with limiting angina pectoris. Am Heart J 1993; 125: 297: 750–57.
12 Moss AJ, Davis HT, DeCamilla J, Bayer LW. Ventricular ectopic beats
31 Task Force of the European Society of Cardiology and the North
and their relation to sudden and nonsudden cardiac death after
American Society of Pacing and Electrophysiology. Heart rate
myocardial infarction. Circulation 1978; 60: 998–1003.
variablity: standards of measurement, physiological interpretation and clinical use. Circulation 1996; 93: 1043–65 and Eur Heart J 1996; 17:
13 Bigger JT Jr, Flaiss JL, Kleiger J, et al. The relationship among
ventricular arrhythmias, left ventricular dysfunction, and mortality in
32 Farrell TG, Bashir Y, Cripps T, et al. Risk stratification for arrhythmic
two years after myocardial infarction. Circulation 1984; 69: 250–58.
events in postinfarction patients based on heart rate variability,
14 Pitt B. Sudden cardiac death: role of left ventricular dysfunction. Ann
ambulatory electrocardiographic variables and the signal-averaged
N Y Acad Sci 1982; 382: 218–28.
electrocardiogram. J Am Coll Cardiol 1991; 18: 687–97.
15 Kaplan E, Meier P. Nonparametric estimation from incomplete
33 Schwartz PJ, La Rovere MT, Vanoli E. Autonomic nervous system and
observations. J Am Stat Assoc 1958; 53: 457–81.
sudden cardiac death: experimental basis and clinical observations for
16 Cox DR. Regression models and life-tables. J R Stat Soc Ser B 1972;
post myocardial infarction risk stratification. Circulation 1992; 85 34: 1987–202.
17 Task Force of the Working Group on Arrhythmias of the European
34 Wilson JS, Podrid PJ. Side effects from amiodarone. Am Heart J 1991;
Society of Cardiology. The early termination of clinical trials: causes,
consequences, and control, with special reference to trials in the field
35 Harris L, McKenna WJ, Rowland E, et al. Side effects of long-term
of arrhythmias and sudden death. Circulation 1994; 89: 2892–907 and
amiodarone therapy. Circulation 1983; 67: 45–51. Eur Heart J 1994; 15: 721–38.
36 Newnham HH, Topliss DJ, Le Grand BA, Chosich N, Harper RW,
18 Peto R, Pike MC, Armitage P, et al. Design and analysis of
Stockigt JR. Amiodarone-induced hyperthyroidism: assessment of the
randomized clinical trials requiring prolonged observation of each
predictive value of biochemical testing and response to combined
patient: I introduction and design. Br J Cancer 1976; 34: 585–612.
therapy using propylthiouracil and potassium perchlorate. Aust N Z J
19 Waldo AL, Camm AJ, DeRuyter H, et al. The SWORD trial. Survival
Med 1988; 18: 37–44.
with oral d-sotalol in patients with left ventricular dysfunction after
37 Martino E, Aghini-Lombardi F, Mariotti S, et al. Amiodarone iodine-
myocardial infarction: rationale, design and methods. Am J Cardiol
induced hypothyroidism: risk factors and follow up in 28 cases. Clin
1995; 75: 1023–27. Endocrin 1987; 26: 227–37.
20 Waldo AL, Camm AJ, de Ruyter H, et al. Effect of d-sotalol on
38 Honhloser SH, Klingenheben T, Singh BN. Amiodarone-associated
mortality in patients with left ventricular dysfunction after recent and
proarrhythmic effects: a review with special reference to torsade de
remote myocardial infarction. Lancet 1996; 348: 7–12.
pointes tachycardia. Ann Intern Med 1994; 121: 529–35.
21 Yusuf S, Venkatesh G, Teo KK. Critical review of the approaches to
39 Bashir Y, Sneddon JF, Paul VE, et al. A prospective study of the
the prevention of sudden death. Am J Cardiol 1993; 72: 51F–58F.
efficacy and safety of adjuvant metropolol and xamoterol in
22 Teo K, Yusuf S, Furberg C. Effect of antiarrhythmic drug therapy in
combination with amiodarone for resistant ventricular tachycardia
acute myocardial infarction: an overview of results from randomised
associated with impaired left ventricular function. Am Heart J 1992;
controlled trials. JAMA 1993; 270: 1589–95. 124: 1233–44.
VNG Instructions As requested by your doctor, you have been scheduled for a test called Videonystagmography (VNG) which is a test of inner ear and central motor function. The test will take approximately 1 ½ to 2 hours and is not painful. Please follow the instructions below to ensure your test results will be valid. The appointment time reserved for your test is critical. All testing
Orefa® Metribuzin Wirkstoff: 700 g/kg Metribuzin Wasserdispergierbares Granulat GP 024504-00/028 WIRKUNGSWEISE Blatt- und Bodenherbizid zur Bekämpfung von Unkräutern und Ungräsern in der Kartoffel (Vor- und Nachauflauf) Wirkungsmechanismus (HRAC-Gruppe): C1 WIRKUNGSSPEKTRUM Gut bekämpfbar : Einjähriges Bingelkraut, Kleine Brennnessel, Ehrenpreis, Erdrauch, Franzosenkraut,