At present, patients with documented sustained VT or resuscitated cardiac arrest (CA) are treated with ICDs. The aim of this study was to retrospectively evaluate if a routine electrophysiological study should be recommended prior to ICD implantation. In 462 patients referred for ICD implantation because of supposedly documented VT (n = 223) or CA (n = 239), electrophysiological study was routinely performed. In 48% of the patients with CA, sustained VT or VF was inducible. Electrophysiological study suggested conduction abnormalities (n = 11) or supraventricular tachyarrhythmias (n = 3) in conjunction with severely impaired left ventricular function to have been the most likely cause of CA in 14 (5.9%) of 239 patients. Likewise, sustained VT was only inducible in 48% of patients with supposedly documented VT. Of these inducible VTs, nine were diagnosed as right ventricular outflow tract tachycardia or as bundle branch reentry tachycardia. Supraventricular tachyarrhythmias judged to represent the clinical event were the only inducible arrhythmia in 35 (16%) patients (AV nodal reentrant tachycardia [n = 7], AV reentry tachycardia [n = 4], atrial flutter [n = 19], and atrial tachycardia [n = 5]). Based on findings from the electrophysiological study, ICD implantation was withheld in 14 (5.9%) of 239 patients with CA and in 44 (19.7%) of 223 patients with supposedly documented VT. During electrophysiological study, VT or VF was only reproducible in about 50% of patients with supposedly documented VT or CA. Electrophysiological study revealed other, potentially curable causes for CA or supposedly documented VT in 12.6% (58/462) of all patients, indicating that ICD implantation can potentially be avoided or at least postponed in some of these patients. Based on these retrospective data, routine electrophysiological study prior to ICD implantation seems to be advisable.