The effect of ventricular fibrillation duration in humans on defibrillation efficacy as it pertains to the time of intervention of an automatic implantable defibrillator is unknown. If a difference in defibrillation efficacy exists in the early period after ventricular fibrillation onset, it may affect algorithms used by antiarrhythmic devices for arrhythmia detection and therapy. Therefore, a prospective, randomized evaluation was performed of the effect of ventricular fibrillation durations of 10 s and 20 s on defibrillation thresholds in 10 survivors of sudden cardiac arrest undergoing implantation of an automatic cardioverter defibrillator. The initial duration of ventricular fibrillation was chosen randomly. Subsequently, each patient served as his or her own control for the alternate duration of ventricular fibrillation to that chosen initially. The mean leading edge defibrillation threshold voltage was 411 +/- 114 V when ventricular fibrillation persisted for 10 s and 419 +/- 125 V when it persisted for 20 s (p = 0.73). The mean defibrillation threshold current was 11.4 +/- 2.8 A when ventricular fibrillation persisted for 10 s and 11.4 +/- 3.2 A when it persisted for 20 s (p = 0.97). The delivered energy defibrillation threshold was 11.5 +/- 5.9 J when ventricular fibrillation persisted for 10 s and 12.0 +/- 6.9 J when it persisted for 20 s (p = 0.67). These findings show that the defibrillation threshold does not change between 10 and 20 s of ventricular fibrillation in out-of-hospital survivors of cardiac arrest at the time of surgical implantation of an automatic defibrillator. The data may have influence on the programming of defibrillator detection algorithms.