Sudden cardiac death is believed to account for a substantial proportion of deaths in patients with evidence of cardiovascular disease. The exact proportion of cardiac deaths that are classified as being sudden varies depending on the population, the underlying disease, and the definition of sudden death. Various estimates suggest that about 500,000 sudden cardiac deaths occur annually in the United States alone. The majority of sudden cardiac deaths are thought to be due to ventricular fibrillation (VF) or tachycardia (VT).
The commonest approach to preventing sudden cardiac death has been by the use of drugs that suppress ventricular ectopy. The rational for this approach is based upon an association between the presence or frequency of ventricular arrhythmia and subsequent mortality in several studies. None of the randomized controlled studies of numerous 'classical' antiarrhythmic agents (other than beta-blockers, which have only a modest effect on arrhythmia suppression) have demonstrated a reduction in sudden or non-sudden cardiac mortality. Indeed, in the recent Cardiac Arrhythmia Suppression Trial (CAST), two class Ic anti-arrhythmic agents demonstrated a 2.5 fold increase in the risk of sudden and non-sudden cardiac deaths despite excellent suppression of ventricular arrhythmia.
Given the disappointing results of most pharmacologic approaches to preventing sudden death, many investigators have turned to non-pharmacologic approaches such as surgery (endocardial resection, stellate ganglionectomy) or the implantation of devices that recognize VT or VF and deliver a shock. The greatest interest has been generated by work on the implantable cardiac defibrillator.
The study was reviewed by an ad hoc working group, the Clinical Applications and Prevention Advisory Committee, and several members of the Cardiology Advisory Committee prior to review and approval by the National Heart, Lung, and Blood Advisory Council in September 1991. The Request for Proposals was released in February 1992.
At approximately 28 clinical sites, patients with ventricular tachycardia or ventricular fibrillation were screened. Those with ventricular fibrillation or serious ventricular tachycardia were entered into a registry for long-term mortality follow-up using the National Death Index. Patients with the prospect of long-term benefit from an ICD and/or antiarrhythmic drug therapy and without exclusions to an ICD or to amiodarone and without a transient or correctible cause of the index event were entered into the trial.
Patients meeting the criteria were randomized to treatment with an ICD or treatment with antiarrhythmic drug therapy. Allocation was stratified by clinical site and index arrhythmia, either ventricular fibrillation or ventricular tachycardia. Patients assigned to the antiarrhythmic drug therapy and without contraindications to sotalol underwent subrandomization to either empiric amiodarone or sotalol, the latter treatment guided by either ambulatory monitoring or electrophysiologic testing. Patients who, after subrandomization, had low levels (less than 30 beats per hour) of ventricular ectopic beats and no inducible ventricular arrhythmias at electrophysiologic study were not treated with sotalol and instead received empiric amiodarone. The AVID protocol allowed usual clinical practice but restricted interventions to state-of-the art ICD devices and first-line antiarrhythmic agents to amiodarone and sotalol. Patients who could not take amiodarone were not included in the trial. The protocol encouraged the use of concurrent drugs such as angiotensin-converting enzyme inhibitors, aspirin, and beta-blockers when appropriate, administered before randomization and maintained throughout the study. The primary endpoint was total mortality. Secondary endpoints were cost of health care and quality of life. Nonlethal events such as ICD shock, sustained arrhythmia, or syncope were tabulated.
Patients were followed every three months for assessment of secondary endpoints, to record therapies delivered by the ICD and potential adverse effects of the ICD, and to assess compliance and potential adverse symptoms in patients treated with antiarrhythmic drugs. A 12-lead electrocardiogram was obtained every six months, and appropriate laboratory and pulmonary tests were performed at six and eighteen months on patients receiving amiodarone. The average follow-up was expected to be 2.6 years. Analysis was done by intention-to-treat. The outcome of primary interest in the subrandomization between sotalol and amiodarone was the time to withdrawal from assigned therapy.
After a review of the data by the Data and Safety Monitoring Board, the AVID study was stopped early on April 7, 1997 because of the findings that after one year, patients in the defibrillator group experienced a nearly 38 percent reduction in deaths compared to the group of patients taking an antiarrhythmic drug. The defibrillator group had about a 25 percent reduction in deaths in years two and three.