Current issues of ACP Journal Club are published in Annals of Internal Medicine


The cost of QALY gained was high for implantable cardioverter defibrillators compared with amiodarone for sudden cardiac death

ACP J Club. 1997 May-Jun;126:61. doi:10.7326/ACPJC-1997-126-3-061

Related Content in this Issue
• Companion Abstract and Commentary: An implanted defibrillator reduced death in patients with coronary disease at high risk for ventricular arrhythmia

Source Citation

Owens DK, Sanders GD, Harris RA, et al. Cost-effectiveness of implantable cardioverter defibrillators relative to amiodarone for prevention of sudden cardiac death. Ann Intern Med. 1997 Jan 1;126:1-12.



To compare the cost-effectiveness of implantable cardioverter defibrillators (ICDs) with amiodarone for patients at high or intermediate risk for sudden cardiac death.


Cost-effectiveness analysis using a computer decision model.


United States.


A hypothetical cohort of patients at high or intermediate risk for sudden cardiac death. Patients were at risk for ventricular tachycardia or fibrillation, nonarrhythmic cardiac death, noncardiac death, illness or death related to amiodarone, and perioperative morbidity and mortality related to ICDs.


Patients were allocated to ICD; amiodarone; or amiodarone-to-ICD, which included patients receiving amiodarone who crossed over to ICD if resuscitated from ventricular fibrillation or tachycardia, or those in which severe amiodarone toxicity occurred. Base-case analysis assumed the mean patient age was 57 years, quality of life was 0.75 on either treatment, ICD use reduced mortality at 1 year by 20% to 40% relative to amiodarone, annual perioperative mortality was 1.8% and morbidity was 2% to 3%, and the ICD generator would be replaced every 4 years.

Main cost and outcome measures

Life expectancy, quality-adjusted life-years (QALYs), and marginal cost-effectiveness (in 1995 U.S. dollars) based on direct costs estimated from a survey of California hospitals.

Main results

Assuming that an ICD reduces total mortality by 20%, high-risk patients who received an ICD lived for 4.18 QALY at a cost of $88 400 compared with 3.68 QALY at a cost of $51 000 for those who received amiodarone. For intermediate-risk patients, the corresponding figures were 6.32 QALY and $110 500 for ICD and 5.81 QALY and $71 400 for amiodarone. The marginal cost-effectiveness ratio for an ICD relative to amiodarone alone was $74 400 per QALY saved for high-risk patients and $76 800 for intermediate-risk patients. For a mortality reduction of 40% with an ICD, the marginal cost-effectiveness per QALY saved was $37300 for high-risk patients and $36 300 for intermediate-risk patients. Cost-effectiveness estimates were most sensitive to the relative reduction in total mortality because of ICD use, frequency of generator replacement, quality of life with therapy, and initial cost of ICD implantation.


Cost-effectiveness data showed that the cost of quality-adjusted life-years saved with implantable cardioverter defibrillators was substantial and greatly depended on the reduction in mortality achieved.

Sources of funding: Agency for Health Care Policy and Research, and Veterans Affairs Health Services Research and Development Service.

For article reprint: Dr. D.K. Owens, Section of General Internal Medicine (111A), Veterans Affairs Palo Alto Health Care System, 3801 Miranda Avenue, Palo Alto, CA 94304, USA. FAX 415-852-3474.


The ICD aborts sudden cardiac death in patients with symptomatic ventricular tachyarrhythmias. The role of this device in preventing sudden death in patients with asymptomatic tachyarrhythmias is less clear. The landmark MADIT study found that ICD was an effective therapy for a select group of patients with asymptomatic unsustained ventricular tachycardia at very high risk for sudden death.

Patients in MADIT had previous MIs, mean ejection fractions of 25% to 27%, and electrophysiologically inducible sustained tachyarrhythmias that could not be suppressed with procainamide. These patients had a very high risk for sudden death and may represent the group that is least likely to respond to antiarrhythmic therapy (1). Despite very low ejection fractions, only 50% to 60% were treated with angiotensin-converting enzyme inhibitors. More optimal treatment of systolic dysfunction could decrease sudden death and overall mortality and lower the absolute benefits of ICD therapy. Whether participants received other therapies (such as aspirin and lipid-lowering agents), which could also decrease mortality and lower absolute benefits of ICD therapy, was not described.

MADIT compared ICD with conventional medical management determined by individual physicians. Most patients (92%) assigned to conventional therapy received antiarrhythmic agents, including β-blockers, at 1 month. Amiodarone is among the most promising pharmaceutical agents for ventricular arrhythmias. Although 74% of participants in MADIT who were assigned to conventional therapy received amiodarone, approximately 10% received class I antiarrhythmic agents; no patient received encainide, flecainide, or moricizine. Further, 23% of the conventional-therapy group were not taking any antiarrhythmic agents at the end of the trial. Of note is that 11% of the participants assigned to conventional therapy received a defibrillator during the trial, and 5% of those assigned to ICD never received a defibrillator. These factors could lead to underestimates of the benefits of ICD.

The elegant analysis done by Owens and colleagues elucidates important points that must be taken into consideration before ICD can be accepted as cost-effective for preventing sudden death in high-risk and intermediate-risk patients with asymptomatic ventricular arrhythmias. This analysis specifically compared ICD with amiodarone alone and amiodarone followed by ICD when necessary. It did not include costs of procedures, such as electrophysiologic testing, that are used by some cardiologists to identify patients likely to benefit from prophylaxis. The analysis found that cost-effectiveness strongly depended on the magnitude of reduction in total mortality associated with ICD and the frequency of generator replacements. ICD is estimated to cost more than U.S. $50 000 per QALY gained unless it reduces mortality by 30% or more compared with amiodarone. Although MADIT showed that mortality reductions as high as 50% can be achieved with ICD compared with conventional therapy in selected very high risk patients, the relative magnitude of reductions that are achievable in high-risk and intermediate-risk patients treated with ICD compared with amiodarone alone is not yet known. Further, estimates of the magnitude of mortality benefits in this relatively small study were not very robust (i.e., the CIs were wide).

In summary, MADIT showed that ICDs were more efficacious than physician- determined therapy that consisted of a heterogeneous mixture of antiarrhythmic agents for very high risk patients with tachyarrhythmias that could not be suppressed with procainamide. Actual costs of identifying and treating this select group of patients were not clearly defined and were likely to be high. The analysis by Owens and colleagues underscored the need to wait for the results of several ongoing trials before concluding that ICD is generally preferable to amiodarone. Clinicians should wait for the results of these trials before routinely recommending ICDs for their high-risk and intermediate-risk patients with asymptomatic unsustained ventricular tachyarrhythmias.

Domenic Marini, MD
Cynthia D. Mulrow, MD, MSc
Audie L. Murphy Memorial Veterans Administration HospitalSan Antonio, Texas, USA

Domenic Marini, MD
Audie L. Murphy Memorial Veterans Administration Hospital
San Antonio, Texas, USA

Cynthia D. Mulrow, MD, MSc
Audie L. Murphy Memorial Veterans Administration Hospital
San Antonio, Texas, USA


1. Waxman HL, Buxton AE, Sadowski LM, Josephson ME. The response to procainamide during electrophysiologic study for sustained ventricular tachyarrhythmias predicts the response to other medications. Circulation. 1983;67:30-7.