Letter by Dardas and Aaronson Regarding Article, “Impact of Implantable Cardioverter-Defibrillator, Amiodarone, and Placebo on the Mode of Death in Stable Patients With Heart Failure: Analysis From the Sudden Cardiac Death in Heart Failure Trial”

Abstract

HomeCirculationVol. 122, No. 15Letter by Dardas and Aaronson Regarding Article, “Impact of Implantable Cardioverter-Defibrillator, Amiodarone, and Placebo on the Mode of Death in Stable Patients With Heart Failure: Analysis From the Sudden Cardiac Death in Heart Failure Trial” Free AccessLetterPDF/EPUBAboutView PDFView EPUBSections ToolsAdd to favoritesDownload citationsTrack citationsPermissions ShareShare onFacebookTwitterLinked InMendeleyReddit Jump toFree AccessLetterPDF/EPUBLetter by Dardas and Aaronson Regarding Article, “Impact of Implantable Cardioverter-Defibrillator, Amiodarone, and Placebo on the Mode of Death in Stable Patients With Heart Failure: Analysis From the Sudden Cardiac Death in Heart Failure Trial” Todd F. Dardas, MD, MS and Keith D. Aaronson, MD, MS Todd F. DardasTodd F. Dardas University of Michigan Health Systems Division of Cardiovascular Medicine Ann Arbor, Mich (Dardas, Aaronson). Search for more papers by this author and Keith D. AaronsonKeith D. Aaronson University of Michigan Health Systems Division of Cardiovascular Medicine Ann Arbor, Mich (Dardas, Aaronson). Search for more papers by this author Originally published12 Oct 2010https://doi.org/10.1161/CIRCULATIONAHA.110.941302Circulation. 2010;122:e486To the Editor:The Sudden Cardiac Death in Heart FailureTrial (SCD-HeFT) investigators provide an important analysis of the mode of death that occurred in the placebo, amiodarone, and defibrillator arms of the SCD-HeFT.1 The authors conclude that survival from sudden cardiac death is increased among those randomized to an implantable cardioverter-defibrillator (ICD). In addition, the probability of survival from heart failure death was not reduced among those receiving ICD therapy.Another approach to measuring the effect of ICD therapy on the type of death is to assess the odds of the mode of death given that a death occurred. We used the data supplied in the SCD-HeFT authors' table and multinomial logistic regression to assess the odds of selected types of cardiac death relative to the odds of heart failure death. When comparing the odds of tachyarrythmic death to heart failure death, the odds of a tachyarrythmic death were higher in the placebo [odds ratio 2.8, 95% confidence interval 1.7–4.6] and amiodarone (odds ratio 2.2, 95% confidence interval 1.3–3.6) groups compared with the ICD group. The modeled probability of a cardiac death being a heart failure death over the course of the study was 59% in the ICD group versus 41% in the amiodarone group and 40% in the placebo groups. The probability of tachyarrythmic deaths was 30% for the ICD group, 46% for the amiodarone group, and 57% for the placebo group. From the standpoint of the type of death, given that a death occurred, we can conclude that heart failure deaths are more frequent in the ICD group.From the SCD-HeFT authors' analysis, we may conclude that ICD therapy does not alter survival from heart failure deaths, in contrast to observations that ICD placement may reduce overall mortality while increasing heart failure deaths. ICDs certainly do not hasten death from heart failure, but ICDs will allow more patients to be exposed to the baseline risk of heart failure death. This effect is well visualized in the mode of death curves reported in the Defibrillator in Acute Myocardial Infarction Trial (DINAMIT), which shows that when arrhythmic deaths are censored, the likelihood of nonarrhythmic deaths in the ICD arm increases relative to the control arm.2 Censoring sudden death has the effect of overestimating mortality from heart failure death. Simultaneously, more patients in the ICD arm are exposed to pump failure death as the result of sudden death reduction. Because ICDs are effective in reducing sudden death, the effect of increasing the number of subjects at risk for pump failure death in the ICD arm is greater than the overestimate of mortality from censoring sudden death in the control arm. As such, we would expect that pump failure deaths in the ICD arm would exceed those in the control arm. ICDs introduce no excess risk of pump failure, although among those who die, heart failure deaths are relatively more likely among those with an ICD.Todd F. Dardas, MD, MSKeith D. Aaronson, MD, MS University of Michigan Health Systems Division of Cardiovascular Medicine Ann Arbor, Mich.DisclosuresNone.