Implantation of implantable cardioverter defibrillators (ICD) has widely been accepted for secondary prevention of sudden cardiac death (SCD) in cardiac arrest survivors. Currently there are increasing interests in primary prevention of SCD in selected high risk patients who have not experienced cardiac arrest. Despite of extensive investigation for risk stratification， our current ability to accurately identify patients at high risk for SCD remains very poor.

    Approximately 80% of SCD are secondary to coronary artery disease. Our experience on risk stratification is primarily gained from studies in patients with CAD， particularly following myocardial infarction (MI).  In 1980s and 1990s， electrophysiological testing (EPT) was widely used for assessment of patients suspected of having ventricular arrhythmias associated with CAD. At the same time， considerable efforts were also made to investigate the role of several invasive and non-invasive techniques for risk stratification. Since MADIT II data was published in 2002， it appears that the predictive value of left ventricular ejection fraction (LVEF) may be over-emphasized and the role of EPT and other risk factors or techniques are being ignored.

    At present， risk stratification for primary prevention of SCD is mainly based on left ventricular function. A low LVEF is recommended for use to identify patients at high risk of SCD and for guidance to select patients for ICD implantation. This concept is derived from the data of the MADIT II and SCD-HeFT trials. Although a low LVEF is certainly associated with increased risk of SCD， the number of patients needed to treat in order to save a life is relatively high when selecting patients for ICD implantation using LVEF as the only risk stratifier. This leads to the concern on cost-effectiveness of ICD implantation for primary prevention. I personally agree with Dr. Alfred E. Buxton Dr. Mark E. Josephson on that other risk factors or techniques (including EPT) in addition to LVEF should be considered to improve cost-effectiveness until a better strategy becomes available.

Patients with coronary artery disease
    There are convincing data showing that ventricular tachycardia (VT) induced by programmed electrical stimulation (PES) can be used for risk stratification of SCD in MI patients with non-sustained VT (NSVT) and left ventricular dysfunction (LVEF <40%). Unfortunately its predictive efficacy is modest (2-year risk for VT or SCD is 18% in patients with inducible VT versus 12% in registry patients， i.e.， 50% more risk if inducible). Non-inducible patients are still at relatively high risk of SCD (2-year risk for VT or SCD is 12%). In other words， patients with inducible VT are at higher risk for SCD but non-inducibility does not necessarily exclude the risk of SCD. It is my experience that the three major risk factors for SCD in patients following MI are residual ischemia， depressed left ventricular function， and ventricular arrhythmias (including VTs induced by PES). A practical approach is that patients are initially evaluated using non-invasive techniques and further stratified using EPT. One question that remains to be answered is whether the predictive value of EPT established in 1980s and 1990s has been altered or not in the era of using percutaneous coronary interventions as the primary therapeutic modality in patients with acute MI.

Patients with idiopathic dilated cardiomyopathy
    The predictive value of EPT remains controversial in patients without CAD， such as idiopathic dilated cardiomyopathy (DCM). It has generally been thought that EPT has limited role for risk stratification in patients with idiopathic DCM. We reviewed 102 patients with idiopathic DCM underwent ICD implantation prior to MADIT II publications aimed at assessing the predictive role of EPT. We found that in these patients inducible VT (using a more aggressive stimulation protocol) was associated with more appropriate ICD therapies during a median follow-up period of 20.6 months (Figure). It seems that in patients with idiopathic DCM， inducible patients are at increased risk of spontaneous arrhythmic events and potentially cardiac arrest (positive predictive value 84% and specificity 91%) but less sensitive compared to that in patients with CAD for risk stratification (with sensitivity of 59%).

Other patients at increased risk for SCD
    Risk stratification for SCD in other rare clinical settings (such as hypertrophic cardiomyopathy， arrhythmogenic right ventricular dysplasia， and Bruguda syndrome) has been difficult， partially due to limited number of patients available for large clinical trials. Proposed risk factors usually include prior history of cardiac arrest or unexplained syncope， spontaneous sustained VTs (possibly including NSVT as well) and more severe pathological changes (for example， left ventricle thickness ≥30 mm in hypertrophic cardiomyopathy). Again the predictive value of EPT remains controversial in these patients. In general， inducible VTs in these patients might be considered to be associated with higher risk for SCD， especially in patients with other risk factors.

    In summary， EPT can be used for risk stratification for SCD in patients with CAD. This may improve predictive accuracy in patients stratified using other risk factors， including a low LVEF. In patients at increased risk for SCD without CAD (including idiopathic DCM)， EPT may also provide useful predictive information. EPT might be considered in selected patients for further risk stratification for SCD in order to improve cost-effectiveness of expansive ICD implantation.