Patients who have experienced a myocardial infarction (MI) are at increased risk of sudden cardiac death (SCD). With the advent of implantable cardioverter-defibrillators (ICDs), accurate risk stratification has become very relevant. Numerous investigations have proven that a reduced left ventricular ejection fraction (LVEF) significantly increases the SCD risk. Furthermore, ICD implantation in patients with reduced LVEF confers significant survival benefit. As a result, LVEF is the cornerstone of current decision making for prophylactic ICD implantation after MI. However, LVEF as standalone risk stratifier has major limitations: (i) the majority of SCD cases occur in patients with preserved or moderately reduced LVEF, (ii) only relatively few patients with reduced LVEF will benefit from an ICD (most will never experience a threatening arrhythmic event, others have a high risk for non-sudden death), (iii) a reduced LVEF is a risk factor for both sudden and non-sudden death. Several other non-invasive and invasive risk stratifiers, such as ventricular ectopy, QRS duration, signal-averaged electrocardiogram, microvolt T-wave alternans, markers of autonomic tone as well as programmed ventricular stimulation, have been evaluated. However, none of these techniques has unequivocally demonstrated the efficacy when applied alone or in combination with LVEF. Apart from their limited sensitivity, most of them are risk factors for both sudden and non-sudden death. Considering the multiple mechanisms involved in SCD, it seems unlikely that a single test will prove adequate for all patients. A combination of clinical characteristics with selected stratification tools may significantly improve risk stratification in the future.