Sudden cardiac death (SCD) is usually due to ventricular tachycardia/fibrillation and represents one of the most important medical and socio-economical problems in western countries. It accounts for approximately 1 life/1000 subjects/year. New and effective treatments are necessary to reduce such dramatic event. During the last decade implantable cardioverter-defibrillators (ICDs) showed to be an effective tool to reduce both total and SCD mortality either when used for secondary or primary SCD prevention. At present, ICD implantation guidelines suggest to implant an ICD in all the patients on the basis of a left ventricular ejection fraction < or =30-35% only. This scarcely sensitive and specific criterion implies the necessity to implant very costly devices in a wide number of patients to save only few lives. A more accurate patient selection is desirable either from a clinical or ethical or economic point of view. Fortunately, this appears to be possible using well known and proven epidemiological, clinical and risk stratification data. On the basis of such data, more selective ICD implantation criteria might be used in older patients or in patients with significant comorbidity or in those patients identified at very low risk of SCD.