Survival of implantable cardioverter-defibrillator recipients. Can the iceberg remain submerged?

are now being widely proposed as definitive antiarrhythmic therapy in patients with sustained ventricular tachycardia (VT) or ventricular fibrillation (VF) and in survivors of cardiac arrest. This approach has been recommended as a first line of treatment in sudden death victims.' Like other therapeutic cardiovascular procedures that have evolved in the last 15 years, application in early clinical trials was initially justified in patients without adequate therapeutic alternatives. Reliable device reversion of VT/VF coupled with a remarkably low incidence of recurrent sudden death in ICD recipients has advanced its therapeutic role and made competition with other antiarrhythmic therapies inevitable. Now considered "established" clinical practice in gravely ill patients, ethical considerations and use patterns have made controlled trials a formidable challenge. Because of this handicap, several types of outcome analyses have been performed in device recipients. Earliest efforts used historical control data from drug-refractory patients for comparison. This was succeeded by attempts to use the patient as an internal control by analysis of ICD use for VT/VF reversion. Scientific acceptance of such data has been uneven. Temporal disparity in patient recruitment, inevitable in the former approach, could not eliminate the influence exerted by intervening therapeutic refinements. This factor alone makes noncomparable patient populations likely. Equation of device use with aborted sudden death, used in the latter approach, is undermined by available data on device discharge for events other than sustained VT/VF. Conversely, inclusion of only symptomatic discharges in