Which is better over the long term: Amiodarone or the implantable cardioverter defibrillator?

Cardiology Review® Online, November 2005, Volume 22, Issue 11

Implantable cardioverter defibrillator (ICD) implantation has been shown

to be superior to amiodarone (Cordarone, Pacerone) in reducing the risk

of mortality in patients who survived ventricular tachycardia/ventricular fibrillation (VT/VF) in the Cardiac Arrest Study Hamburg (CASH),1 Canadian Implantable Defibrillator Study (CIDS),2 and Antiarrhythmics Versus Implantable Defibrillators (AVID)3 trials. The improvement was not statistical-

ly significant in the CIDS study, however. All patients in the CASH, CIDS,

and AVID studies who had an ICD showed a marked decrease in all-cause

mortality in a meta-analysis, but the benefits were greater for some patients than for others.4 ICD therapy was superior to amiodarone in older patients

with poor left ventricular function and poor New York Heart Association (NYHA) functional class, for example, but it was not significantly better in other subgroups.5

Following the publication of the CIDS data and the meta-analysis of the three ICD trials, ICD implantation became the preferred treatment for survivors of VT/VF.6 However, the ICD benefit may decrease with time and may not extend to all subgroups. Significant differences in the benefits of amiodarone for preventing sudden cardiac death and in the tolerance of the drug have been shown in several studies.7-14 The difference between ICD and amiodarone therapy over the long-term is not known.

We hypothesized that the benefits of ICD implantation over amiodarone are time dependent and, as a result, changing our practice seemed unwarranted once the CIDS trial ended. We decided that, unless patients had major side effects or a recurrence of arrhythmia, we would not cross amiodarone patients over to ICD therapy at our hospital. We compared the long-term effects of using amiodarone as first-line monotherapy with ICD implantation over a period of 11 years.


Of the 659 patients enrolled in the CIDS trial, 120 were enrolled at our institution. The CIDS trial included patients who survived VT/VF or who had unmonitored syncope with inducible VT and a left ventricular ejection fraction below 35%. Patients with ventricular arrhythmia due to reversible causes are excluded from ICD trials, and those with frequent ventricular arrhythmia not due to reversible causes are candidates for ICD only after ventricular arrhythmia has been stabilized and appropriately treated. At the end of the trial, patients continued receiving the same treatment they received during the trial. A change to ICD therapy was not made unless a patient had serious side effects from amiodarone necessitating cessation of treatment or unless the patient had cardiac arrest, unmonitored syncope, or recurrent sustained VT.

Mortality from all causes was the primary end point of the study. Recurrence of ventricular arrhythmia, side effects associated with amiodarone use, discontinuation of amiodarone, and cause-specific death were the secondary end points, in addition to the composite of discontinuation of amiodarone because of side effects or ventricular arrhythmia recurrence or total mortality.


The mean follow-up period was 5.6 ± 2.6 years, from October 1990 until April 2002. The ICD and amiodarone groups were similar in terms of baseline characteristics. Beta blocking agents were used in 38% of patients in the ICD group and in 35% of patients in the amiodarone group. The mean age was 64 ± 9.2 years versus 64 ± 8.7 years in the ICD and amiodarone groups, respectively. The mean ejection fraction was 33.9 ± 12.5% in the ICD group and 32.1 ± 11.1% in the amiodarone group.

The majority of patients were in NYHA functional classes I and II. In the ICD group, 27% of patients (n = 16) died, compared with 47% of patients (n = 28) in the amiodarone group (P = .021). The ICD group had a 43% relative risk reduction in all-cause mortality compared with the amiodarone group. Annual mortality was significantly lower in the ICD group (2.8% per year) compared with 5.5% per year in the amiodarone group (hazard ratio, 2.011; 95% confidence interval, 1.087—3.721; P = .026). The mortality benefit of ICD was assumed to be mainly the result of fewer deaths from arrhythmia compared with amiodarone treatment (2 versus 12, respectively; P = .049). At 11 years, 40 patients (67% of all patients) failed to respond to amiodarone or died, among which 12 patients (20%) had nonfatal symptomatic arrhythmia recurrences. By actuarial analysis, all amiodarone-treated patients are predicted to have arrhythmia recurrences, experience adverse effects, or die within 11 years of follow-up (Table). The majority of patients taking amiodarone (82%; n = 49) experienced adverse effects. A reduction in dosage or discontinuation of the drug because of adverse effects was needed for 50% of patients (n = 30).


The results of our study show that, because of the occurrence of serious adverse effects, recurrence of arrhythmia, and high risk of sudden cardiac death over the long term, amiodarone should not be used as a first choice

of treatment for long-term secondary

prevention of sudden cardiac death. This analysis shows that ICD therapy is superior to amiodarone for decreasing the incidence of death for up to 11 years.

This previously unreported analy-sis shows the efficacy of ICD versus

amiodarone beyond 5 years and un-derlines the importance of using ICD therapy as a first choice for secondary prevention of sudden cardiac death. In addition, the study shows that all amiodarone-treated patients have recurrence of arrhythmia, experience adverse effects, or die by 11 years of follow-up.


Although several randomized clinical trials have shown that ICD therapy is superior to amiodarone in patients with resuscitated VT/VF,1-4 this is the first study to show that the effectiveness of ICD over amiodarone in reducing overall mortality increases with time.

The risk of sudden cardiac death is high among patients receiving amiodarone for secondary prevention of sudden cardiac death compared with patients treated with ICD implantation. Amiodarone as first-line monotherapy for long-term secondary prevention of sudden cardiac death, therefore, is not warranted.