BACKGROUND: ICD therapy is established therapy for secondary prevention after aborted sudden death or ventricular tachycardia. Long-term data on the incidence of appropriate and inappropriate interventions are scarce. METHODS AND RESULTS: We retrospectively studied 391 patients with an ICD for secondary prophylaxis: 247 (63%) with ischaemic heart disease (IHD) and 144 without IHD (37%). Fifty-four patients were free from left ventricular structural disease. Mean follow-up was 30.8 months. Kaplan-Meier methodology was used for survival analysis. The use of beta-blockers was high and similar in both groups (85% IHD; 88% non-IHD; P = 0.36). The incidence of appropriate interventions was identical in IHD and non-IHD (42.7% and 47.8% at 4 y; HR 1.0, P = 0.99). There was a yearly rate of first intervention around 5% even in the fourth and fifth year after implantation. The incidence of inappropriate interventions was about half that of appropriate ICD interventions (21.4% at 4 y). It was higher in patients who also had received appropriate therapy (HR: 2.73 in the IHD group, 1.61 in the non-IHD group, P < 0.001 for both). Atrial fibrillation was the most common cause of inappropriate interventions in IHD, and sinus tachycardia in those without LV disease. The incidence of inappropriate interventions was not dependent on the type of ICD (VVI vs. DDD), in any group. CONCLUSIONS: Patients with an ICD for secondary prophylaxis have a high rate of appropriate interventions, and remain at risk for developing a first intervention several years after implantation. Inappropriate interventions constitute a significant burden. Taking preventive measures (AV nodal slowing drugs, device selection and programming, patient counseling regarding allowable physical activity) is required to optimize the quality-of-life adjusted life-saving potential of ICDs.