Causes of Death in the AVID Trial
abstract & Commentary
Synopsis: The ICD is more effective than an antiarrhythmic drug in reducing arrhythmic cardiac deaths while producing little or no effect on the frequency of nonarrhythmic cardiac deaths.
Source: The AVID Investigators. J Am Coll Cardiol 1999;34:1552-1559.
The antiarrhythmics versus implantable defibrillators (AVID) Trial was a randomized comparison of antiarrhythmic drug therapy and implantable cardioverter defibrillator (ICD) therapy in patients who had survived cardiac arrest or sustained hypotensive ventricular tachycardia. Patients randomized to the antiarrhythmic drug arm could receive either amiodarone or sotalol but the majority received amiodarone. This study analyzes the causes of death among the 202 patients who died during the trial.
Death was defined as the time when respiration and blood circulation ceased without recovery. Information about each death event was collected from the principal investigator. The clinical information was then edited by the coordinating center to remove information on the type of antiarrhythmic therapy the patient was receiving. The edited information was then reviewed by an events committee and classified independently of the site principal investigator’s decision. Deaths were classified as cardiac or noncardiac. Cardiac deaths were then subclassified as either arrhythmic or nonarrhythmic. The diagnosis of arrhythmic cardiac death required the absence of severe congestive heart failure or shock preceding a terminal arrhythmia. The location of death was also recorded and classified as either in-hospital (including in the emergency room) or out-of-hospital. Deaths due to pulmonary causes were reviewed after unblinding to see if amiodarone pulmonary toxicity might be responsible.
There were 122 deaths among patients assigned to an antiarrhythmic drug compared to 80 deaths among patients assigned to an ICD. A majority (78%) of these deaths were classified as cardiac. There were equal numbers of arrhythmic and nonarrhythmic cardiac deaths—79 vs. 78, respectively. The onset of terminal symptoms occurred nearly equally in-hospital vs. out-of-hospital (74 patients vs 83 patients). Most arrhythmic deaths occurred out-of-hospital vs. in-hospital (66 vs 13 patients). Most nonarrhythmic deaths occurred in-hospital vs. out-of-hospital (61 vs 17 patients). Enzymatic or electrocardiographic evidence of ischemia or myocardial infarction (MI) at the time of death was uncommon, with only 12 patients having such findings.
Arrhythmic deaths were more common in the antiarrhythmic drug therapy group. Nonarrhythmic cardiac deaths occurred with equal frequency in the ICD and drug therapy groups (39 vs 39). Noncardiac deaths were more common in the antiarrhythmic drug therapy group, with an excess of pulmonary and renal deaths accounting for most of the difference. There were 24 apparent arrhythmic deaths among patients with an ICD. Unfortunately, only seven of these patients had an ICD interrogation after death. In three of these seven, the ICD detected the arrhythmia and delivered shocks appropriately but could not terminate the arrhythmia. In four of the seven, no tachyarrhythmia was detected and either bradycardia or pulseless electrical activity was likely to have been the final rhythm. Autopsies were performed in only 14 patients and could not be used for classifying the cause of death.
Arrhythmic death was more common among patients whose index arrhythmia had been ventricular fibrillation (VF) than among patients who initially presented with ventricular tachycardia (VT). The ICD produced apparently greater benefit among the patients who presented with VF than it did among those who presented with VT.
Finally, the paper compares the classification by the principal investigators and the Events Committee. There was better than 90% agreement between the unblinded principal investigator and the blinded Events Committee for event classification.
The investigators conclude that the ICD is more effective than an antiarrhythmic drug in reducing arrhythmic cardiac deaths while producing little or no effect on the frequency of nonarrhythmic cardiac deaths.
Comment by John P. DiMarco, MD, PhD
The AVID Trial was a large, multicenter, randomized trial that established the superiority of ICD therapy vs. antiarrhythmic drug therapy in patients who had survived an episode of a life-threatening ventricular arrhythmia. This paper confirms the hypothesis that the ICD produces its greatest benefit by affecting the frequency of arrhythmic death. Importantly, there was no excess in nonarrhythmic cardiac deaths, indicating that the ICD did not merely briefly postpone deaths among patients with an irreversible deterioration in cardiac function.
The classification process used in AVID has now become standard in large-scale arrhythmia mortality trials. The most difficult decision is usually whether a terminal arrhythmia occurred in the setting of irreversible heart failure or shock. This requires extensive review of hospital records and requires a qualitative judgment by an experienced clinician. This paper is interesting since both the principal investigators and Events Committee agreed in a large majority of cases.
One disappointing feature in this report is the lack of information from ICD interrogation about the mechanism of deaths. During the period of AVID, event data storage was a relatively new feature of ICDs. Therefore, many clinicians, including many cardiologists, were not aware that important information about cause of death could be obtained by device interrogation. Programmers capable of interrogating the ICD at the time of death are still not universally available. As more and more physicians become knowledgeable about the capacity of current generation ICDs, we can hope that they will try to obtain interrogations even after out-of-hospital deaths so that we can learn more about the mechanism of death in these individuals.
In AVID, approximately 50% of the deaths occurred out-of-hospital. By contrast, another large trial that investigated the use of ICD therapy, the CABG Patch Trial, observed that 79% of their deaths occurred in-hospital. The CABG Patch Trial was unable to show a difference between the ICD and an untreated control group. This report from AVID demonstrates that the population selected for ICD therapy should have a high frequency of out-of-hospital arrhythmic events.
The slight increase in pulmonary deaths in the drug-treated group is of concern. There were nine pulmonary deaths among the antiarrhythmic drug-treated group vs. only two in the ICD group. Three of these nine deaths were classified as due to amiodarone toxicity. However, amiodarone does produce some slight progressive deterioration in pulmonary function, even in patients without full manifestations of amiodarone toxicity, and one must be concerned that there may be a subtle tendency toward pulmonary problems among amiodarone patients.
The antiarrhythmics vs. implantable defibrillators (AVID) trial showed that:
a. defibrillators reduced in-hospital deaths.
b. defibrillators reduced out-of-hospital deaths.
c. defibrillators reduced noncardiac deaths.
d. defibrillators reduced MI deaths.
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