The defibrillation safety margin (difference between maximum energy output and lowest energy yielding 100% defibrillation success) for implantable cardioverter-defibrillators (ICD) is required to be =< 10 joules (J) at primary implant. There is no general agreement on the need to perform safety margin verification at the time of ICD pulse generator elective replacement (PGER). We tested the defibrillation efficacy of shocks at 10J below maximum ICD output in 91 consecutive pts at PGER using the identical lead configuration (epicardial patch or transvenous lead/subcutaneous patch) and shock waveform (monophasic or biphasic) as primary implantation. All pts had =< 10 J defibrillation safety margins at primary implant (mean implant defibrillation threshold [DFT] 14.4 +/- 3.6 J). Minimum 10 J safety margin verification shocks at PGER failed in 23 (25.2%) pts and were successful in 68 (74.8%). The mean difference between lowest successful shock energy at primary implant and highest unsuccessful shock energy in those pts who failed 10J safety margin verification at PGER was 10.2 +/- 5.8 J. Age, gender, index arrhythmia, underlying heart disease, ejection fraction, heart failure class, implant duration, lead configuration and shock waveform did not differ between groups. Amiodarone use at the time of PGER (odds ratio 4.3, p = 0.03) and higher primary implant DFT (odds ratio 1.2 per joule increment in implant DFT, p = 0.04) were the only independent multivariate predictors of failed 10J safety margin shocks at PGER.Conclusions: One-quarter of the study population failed 10J defibrillation safety margin shocks at ICD PGER despite adequate safety margins at primary implant. Defibrillation safety margin verification should be routinely performed at ICD PGER, particularly in those pts receiving amiodarone or with higher primary DFTs.