In this thesis we have explored strategies to predict the risk of arrhythmic sudden cardiac death (SCD) in patients with ischemic cardiomyopathy after myocardial infarction (MI). We focus on the efficacy of current risk stratification in implantable cardioverter-defibrillator (ICD) patients and the interaction between T-wave alternans (TWA), beat-to-beat variability of repolarization (BVR) and post-MI ventricular remodeling. In chapter 1 we evaluated the value of classical risk factors to discriminate useful from useless ICD implantations in ischemic cardiomyopathy patients. Although low left ventricular ejection fraction (LVEF) is the key parameter to warrant prophylactic ICD implantation, it was also related to a higher number of futile ICD implantations. This finding emphasizes the problematic link between low LVEF and ICD-resistant mortality. There is a need for newer risk predictors of SCD that discriminate the arrhythmic death risk from the non-arrhythmic death risk. In chapter 2 and 3 we investigated the relationship between TWA, BVR and infarct size as a quantifier of post MI ventricular remodeling in a pig model and in patients. TWA and BVR were not interrelated and occurred at a different time after MI. BVR was present during early post-MI ventricular remodeling and TWA was detectable in case of a larger myocardial scar late after MI. Infarct size, timing after MI, exercise capacity and beta-blocker intake are all interfering with the outcome of TWA testing and should strictly be controlled in future studies on the predictive value of TWA. In chapter 4 we explored the link between local intracardiac repolarization alternans and body surface TWA through an innovative pacing paradigm. TWA was highly sensitive for picking up subtle repolarization alternans from limited zones of myocardium but did not discriminate between concordant versus discordant local ventricular alternans.This thesis concludes that repolarization variability is an extremely interesting research domain that integrates a multitude of complex pathophysiological mechanisms. This complexity however is a major obstacle for the implementation of these techniques in routine clinical practice. To decide on the real predictive value of these tests, future clinical research should closely monitor the effect of interfering variables and refine the interpretation of test results.