Cost-Effectiveness of ICD or Amiodarone after MI
Abstract & Commentary
Synopsis: Amiodarone is cost effective as secondary prevention therapy among patients who have ejection fractions less than 0.4 after myocardial infarction.
Source: Sanders GD, et al. Ann Intern Med. 2001;135: 870-883.
Sanders and colleagues describe in this paper a Markov model cost utility analysis on the relevant benefits of amiodarone therapy, implantable cardioverter defibrillator (ICD) insertion, and placebo for the reduction of total mortality and sudden deaths in survivors of acute myocardial infarction (MI). Sanders et al created a decision model for patients who had MI that was not complicated by any symptomatic, sustained ventricular arrhythmias. For the model, patients were stratified into 3 groups according to ejection fraction: less than or equal to 0.3, 0.31 to 0.4, and greater than 0.4. Mortality and inpatient cost data for patients were estimated from the Myocardial Infarction Triage and Intervention (MITI) patient registry. The MITI registry included 40,000 patients who were admitted to cardiac care units in 19 hospitals in Seattle, Wash, between 1988-1994. For example, patients with an ejection fraction of 0.3 or less had a mean age of 66.5 ± 11.9 years with 25% reporting heart failure, 63% receiving a cardiac catheterization, and 27% undergoing revascularization during their initial hospitalization. In this group of patients, the sudden cardiac death mortality was estimated to be 7.8% at 1 year, 10.3% at 2 years, and 13.1% at 3 years. The total mortality rate was estimated to be 21.5%, 30.5%, and 34.7% at 1, 2, and 3 years follow-up, respectively. The cohorts with ejection fractions between 0.31 and 0.4 and greater than 0.4 were younger and had substantially lower total and sudden death mortality rates. For ICD therapy, it was assumed that the ICD did not affect the frequency of arrhythmic events but increased the chance of surviving if one occurred. A perioperative of mortality rate of 0.9% after ICD insertion was assumed. It was estimated that patients with ICDs would develop infections or lead failures at a combined rate of approximately 5% per year. The useful life span of the generator battery before elective replacement was estimated to be 7 years. The model allowed the effectiveness of the ICD to be evaluated over a range of efficacies for sudden death prevention from 30% to 90%. The cost of the device was evaluated over a range from $10,000 to $25,000 per unit.
Two meta-analyses of trials involving amiodarone were used to estimate the reduction of mortality with that drug. Amiodarone was estimated to result in a mortality reduction of between 11% and 19% compared to placebo.
The model predicted that ICD use would be superior to amiodarone therapy in terms of prolonging quality adjusted life, but that it resulted in much higher expenditures. Amiodarone resulted in a modest increment in quality-adjusted life span and intermediate cost. Ejection fraction had a powerful effect on the cost efficacy of ICD implant. Prophylactic therapy with the ICD was most cost effective among patients with the lowest ejection fractions. Assuming a higher efficacy rate for either the amiodarone or the defibrillator also markedly improved cost effectiveness. For patients with an ejection fraction below 0.3, the ICD had to prevent 70% of sudden cardiac deaths to reach a cost effectiveness threshold of $50,000 per quality life year gained. If it prevented only 35% of sudden deaths in this group, a cost effectiveness threshold of $100,000 per life year was achieved. In the cohorts with an ejection fraction of 0.31-0.4, the ICD had to prevent 70% of sudden deaths to reach even the $100,000 per qualify life year gained threshold. Among patients with higher ejection fractions, the ICD did not reach these cost effectiveness thresholds unless the ICD was assumed to have a high efficacy and amiodarone was assumed to have a low efficacy.
Sensitivity analysis also showed that the cost of the ICD itself was a major determinant of cost effectiveness. If the cost of the ICD was reduced from $25,000 to $10,000, the ICD became much more cost effective even in patients with moderately reduced ejection fractions or if lower rates of ICD effectiveness were assumed.
Sanders et al conclude that amiodarone is cost effective as secondary prevention therapy among patients who have ejection fractions less than 0.4 after MI. Due to the relatively low cost of amiodarone, even relatively small reductions in mortality might be clinically important and cost effective. They also conclude that the ICD, although more effective than amiodarone, met cost effectiveness thresholds only in patients with ejection fractions below 0.3.
Comment by John P. DiMarco, MD, PhD
A number of studies both in survivors of sustained ventricular arrhythmias and in high-risk CAD patients have now shown benefit from ICD therapy when it is compared to antiarrhythmic drug therapy. Since insertion of an ICD now can be accomplished with almost no mortality and relatively low morbidity, the major obstacle to more widespread use of the ICD is going to be cost. Analyses like that reported here by Sanders et al will be quite helpful in allowing physicians and society to decide whether such an expenditure is justified. The high cost of the ICD may make it the first example of a highly effective and relatively nontoxic therapy that will be restricted primarily on the basis of cost.
Unfortunately, even the cost estimates used here are fairly low since hospital mark-ups over the manufacturer’s price for the devices are not apparently included. Many in the field have hoped that the manufacturers could be persuaded to produce and market a lower-cost device for prophylactic use. This "generic" type ICD might be much cheaper and, therefore, could be used in a larger number of patients, not just the highest risk subgroups. Unfortunately, the market trend has been development and marketing of more and more sophisticated "rhythm management" devices that have been used to justify a continued high baseline price level.
One other major limitation to this paper should be mentioned. Sanders et al used data from the MITI registry. These patients were enrolled between 1988 and 1994. Since that time, we have seen dramatic advances in the management of ischemic heart disease and congestive heart failure. One would not be surprised if the mortality estimates given in this paper are considerable overestimates over what might be observed in similar populations now. Minor reductions in the baseline mortality estimates would make ICD therapy even less cost-effective than the estimates given here.
You have reached your article limit for the month. Subscribe now to access this article plus other member-only content.
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content