Background. Successful defibrillation by an implantable cardioverter- defibrillator (ICD) depends on its ability to deliver shocks that exceed the defibrillation threshold. This study was designed to identify clinical characteristics that may predict the finding of an elevated defibrillation threshold and to describe the outcome of patients with high defibrillation thresholds. Methods and Results. The records of 1,946 patients from 12 centers were screened to identify 90 patients (4.6%) with a defibrillation threshold ≥25 J. Excluding three patients who received ICDs that delivered >30 J, there were 81 men and six women with a mean age of 59.5±10.1 years, a mean left ventricular ejection fraction of 0.32±0.14, and a 76% prevalence of coronary artery disease. Sixty-one patients (70%) were receiving antiarrhythmic drugs, and 45 (52%) were receiving amiodarone. Seventy-one patients (82%) received an ICD. Death occurred in 27 patients-19 of the 71 (27%) with an ICD (eight arrhythmic), and eight of the 16 (50%) without an ICD (four arrhythmic). Actuarial survival for all patients at 5 years was 67%. Actuarial survival rates at 2 years for patients with and without an ICD were 81% and 36%, respectively (p=0.0024). Actuarial survival at 5 years for the ICD patients was 73%; no patient without an ICD has lived longer than 32 months. Actuarial survival free of arrhythmic death in the ICD patients at 5 years was 84%. Although the only variable to predict survival was ICD implantation (p=0.003), it is entirely possible that in this retrospective analysis, clinical selection decisions to implant or to not implant an ICD differentiated patients destined to have better or worse outcomes, respectively. Conclusions. Antiarrhythmic drug use may be causally related to the finding of an elevated defibrillation threshold. When patients with high defibrillation thresholds receive an ICD, arrhythmic death remains an important risk (42% of deaths in these patients were arrhythmia related, with 16% actuarial incidence at 5 years). Vigorous testing to optimize patch location can potentially benefit patients by enhancing the margin of safety for effective defibrillation.