Various techniques are used to establish defibrillation efficacy and to evaluate defibrillation safety margins in patients with an ICD. In daily practice a safety margin of 10 J is generally accepted. However, this is based on old clinical data and there are no data on safety margins using current ICD technology with unipolar, active pectoral defibrillators. Therefore, a randomized study was performed to test if the likelihood of successful defibrillation at defibrillation energy requirement (DER) + 5 J and + 10 J is equivalent. Ninety-six patients (86 men; age 61.0 +/- 10.3 years; ejection fraction 0.341 +/- 0.132; coronary artery disease [n = 65], dilated cardiomyopathy [n = 18], other [n = 13]) underwent implantation of an active pectoral ICD system with unidirectional current pathway and a truncated, fixed tilt biphasic shock waveform. The defibrillation energy requirement (DER) was determined with the use of a step-down protocol (delivered energy 15, 10, 8, 6, 4, 3, 2 J). The patients were then randomized to three inductions of ventricular fibrillation at implantation and three at predischarge testing with shock strengths programmed to DER + 5 J at implantation and + 10 J at predischarge testing or vice versa. The mean DER in the total study population was 7.88 +/- 2.96 J. The number of defibrillation attempts was 288 for + 5 J and 288 for + 10 J. The rate of successful defibrillation was 94.1% (DER + 5 J) and 98.9% (DER + 10 J; P < 0.01 for equivalence). Charge times for DER + 5 J were significantly shorter than for DER + 10 J (3.65 +/- 1.14 vs 5.45 +/- 1.47 s; P < 0.001). A defibrillation safety margin of DER + 5 J is associated with a defibrillation probability equal to the standard DER + 10 J. In patients in whom short charge times are critical for avoidance of syncope, a safety margin of DER + 5 J seems clinically safe for programming of the first shock energy.