At this time, we find ourselves with an abundance of guidelines for management of patients with manifest ventricular tachyarrhythmias, or at risk for such arrhythmias, in patients with coronary heart disease (CHD). The guidelines are focused primarily on the "appropriate use" of the implantable cardioverter/defibrillator (ICD). Unfortunately, the bulk of the guidelines have very little basis in the underlying pathophysiology responsible for sudden cardiac death (SCD) in patients with CHD. Rather, they are based primarily on the results of randomized clinical trials that merely sought to take broad populations at elevated total mortality risk and determining whether the ICD can reduce overall mortality. The trials were not aimed at elucidating or exploiting the varying pathophysiology responsible for the ventricular arrhythmias responsible for most sudden deaths in this setting. The goal of the trials is appropriate - to improve the survival. The problem with promoting trials that solely determine whether a broad-based population (identified by one parameter such as ejection fraction that bears no direct relation to the pathogenesis of arrhythmias) derives a survival benefit from a therapy such as the ICD, is that many patients that could benefit from the ICD are missed (not covered by the guidelines), and many patients that will never benefit from the ICD are exposed to its risks and costs. How can we advance the use of potent, but expensive therapies that carry risk such as the ICD to improve survival of patients with CHD today? There are several avenues worth pursuing, both for short-term as well as long-term gain. First, there are several models shown to have the potential to identify patients currently covered by the guidelines for ICD use, that are highly unlikely to benefit, because of the existing co-morbidities. These models are likely to be valid because there is significant overlap in the parameters identified in each model, and they have been tested retrospectively in a variety of study populations. These models are not likely to be incorporated into use guidelines, until they have been tested prospectively in a randomized trial in a contemporary patient population. This can, and should be done. Use of such a model, based on noninvasive, readily available clinical markers offers the possibility of improving the efficiency with which ICDs are used to reduce the risk of SCD in CHD patients. Second, we need to recognize the fact that SCD in this population is a result of multiple potential mechanisms. And, the electrophysiologic substrates underlying these mechanisms are influenced by interactions with the autonomic nervous system and hemodynamic conditions. While most out-of-hospital cardiac arrests do not occur in persons with overt heart failure, the presence of heart failure clearly increases the risk for SCD, likely by a variety of mechanisms. There is increasing evidence that altered left ventricular geometry may not only reduce LV mechanical efficiency, but may also have direct effects on the electrophysiologic substrate. Although there is an abundance of evidence supporting the importance of autonomic interactions in the genesis of spontaneous arrhythmias, the utility of prospectively measuring autonomic indices to predict future arrhythmic events has to date not proven to be useful. Of course, that is not to discount the significant impact of beta-adrenergic blockade on survival and reducing arrhythmic events. Future works must focus more on both animal models of post-infarction arrhythmias, as well as integrating findings from such studies into human physiology, with subsequent testing in the form of randomized clinical trials.