Cardiac arrest or sustained VT (ventricular tachycardia) in patients with heart disease is best treated with an ICD (implantable cardioverter defibrillator). However, the ICD alone is not appropriate therapy for patients with frequent VT episodes. In fact frequent shocks for VT may predict a poorer prognosis. Anti-arrhythmic drugs are co-administered with ICDs in up to 50% of patients to prevent VT episodes, but antiarrhythmic drugs may have harmful effects. Thus improved drugs to prevent VT without interfering with ICD function are needed. Recent data including our own suggest that clonidine may be a new therapy to prevent ICD shocks. It may act centrally on sympathetic outflow and peripherally and selectively on cardiac Purkinje, to suppress and control VT occurring in patients. Our purpose is to test the hypothesis that clonidine reduces frequent VT better than beta blocker in patients with ICDs. After informed consent patients will be randomized in a single blind fashion to either clonidine or metoprolol given three times per day. Other prescribed drugs may be adjusted to promote toleration of the study drug. ICD interrogations of episodes of VT will be the primary endpoint. Device based NIPS (non-invasive programmed stimulation) testing in a subset of these patients will allow mechanistic understanding of the clonidine effect. All of the procedural techniques are in place as performed clinically; preliminary data are given showing feasibility of the project.