The currently available automatic implantable cardioverter-defibrillator has proven highly successful for termination of ventricular tachycardia and fibrillation. Newer devices, however, permit lower energy shocks to be delivered initially and longer episodes of arrhythmia to occur before shocks are delivered. These changes may result in longer durations of arrhythmia before successful termination. Little is known about the effects of the duration of ventricular fibrillation on the efficacy of defibrillating shocks. In this study, we examined the efficacy of defibrillating shocks in 22 patients undergoing automatic implantable cardioverter-defibrillator implantation or generator change. Defibrillating shocks ranging from 300 to 600 V (5.9-24.2 J) were delivered in matched pairs after 5 and 15 seconds of ventricular fibrillation. For the 300-V shocks (5.9 J), defibrillation was accomplished in 82% of patients when the shocks were given after 5 seconds of ventricular fibrillation and in only 45% of patients when the shocks were delivered after 15 seconds (p less than 0.01). At higher energies, there was no difference in the efficacy of defibrillation shocks delivered after 5 compared with 15 seconds of ventricular fibrillation. The postshock aortic, systolic, and diastolic blood pressures were significantly lower when the shocks were given after 15 seconds of ventricular fibrillation than after only 5 seconds. We conclude that the duration of ventricular fibrillation affects defibrillation efficacy especially at energies that are relatively low compared with maximal device outputs and that longer episodes of ventricular fibrillation cause more postshock hemodynamic depression. These observations have implications for defibrillation threshold testing at the time of device implantation and for the design and programming of future automatic implantable antitachycardia devices.