Background We sought to assess the efficacy of high‐energy shocks to restore rhythm and predictors of success in patients with sustained ventricular arrhythmias and implantable cardioverter defibrillator (ICD). Methods and results Data from 162 patients included in the UMBRELLA study that experienced one or more episodes of ventricular tachycardia (VT) for which ICD shocks of at least 30 Joules were delivered (appropriate high‐energy shocks) were analyzed. In total, 456 ventricular arrhythmia episodes were registered. Forty four episodes (9.6%) from 39 patients (24%) had at least one ineffective high‐energy shock delivered. Hypertrophic cardiomyopathy was more frequent among patients with unsuccessful shocks (10.3% vs 2.4%). Patients with ineffective shocks had higher proportion of sustained monomorphic ventricular arrhythmias (86.4%; the other 13.6% were sustained polymorphic and ventricular fibrillation VF) compared with patients with all their shocks effective (62.9%, P = 0.02). No statistical differences were found between groups in time from detection to the high‐energy shock delivery, in tachycardia cycle length, or in antitachycardia pacing, but patients with ineffective high‐energy shocks had higher proportion of previously ineffective low‐energy shock (9.1% vs 0.5%, P = 0.01). Conclusion We found a substantial rate of ineffective high‐energy shocks for the treatment of VT or VF in patients with ICD. High‐energy shock efficacy seems to be reduced by hypertrophic cardiomyopathy and by the administration of previous low‐energy shocks.