All-cause mortality and morbidity following an acute myocardial infarction (AMI) are correlated to LV systolic dysfunction. The correlation is closest with mortality and morbidity associated with congestive heart failure (CHF). Prediction of arrhythmic death in patients with AMI relies on the correlation between arrhythmic death and "sudden unexpected death" defined as death within 1 hour of onset of new symptoms. Assessment of late potentials, heart rate variability (HRV), T wave alternans, arrhythmias seen on Holter monitoring or during exercise testing, electrophysiological testing, and baroreceptor assessment have all proven to be useful in the prediction of sudden death even when LV systolic function is known. In selected populations HRV is superior to LV systolic function assessment in predicting sudden death and/or arrhythmic events, and may even predict all-cause mortality with the same precision. Comparisons of other methods with LV function assessment should be interpreted with care because most methods have been evaluated in subgroups of infarct patients with a low risk of death. Results from a large series of high risk patients with AMI (the TRAndolapril Cardiac Evaluation study) have shown that even in patients with severe depressed LV systolic function around one-third of the patients will die suddenly. The current situation is that LV function appears to be the best method of predicting death whereas other methods appear very promising for detecting arrhythmic death in more selected populations. The optimal method for selecting patients at high risk of arrhythmic death has not yet been developed, but a combination of LV function and another method, i.e., HRV, appears promising. This may ensure that the enrolled patients have an increased risk of death and that this risk will be due to arrhythmic events. Patients with LVEF of 10% or less can be excluded as they will most likely not die suddenly.