Premature ventricular complexes and nonsustained ventricular tachycardia mark a person with structural cardiac disease as a high--risk candidate for sudden cardiac death. Such ventricular arrhythmias are considered potentially lethal and should be distinguished from both those that are benign and those that cause hemodynamic consequences (i.e., lethal or malignant arrhythmias). Noninvasive Holter monitoring is the principal technique for detecting and evaluating the presence of potentially lethal ventricular arrhythmias. These arrhythmias undergo a high degree of spontaneous variability. Thus, to define a therapeutic drug effect, a reduction in the frequency of premature ventricular complexes of at least 75% and a reduction in the frequency of nonsustained ventricular tachycardia by at least 90% are required to eliminate the likelihood of spontaneous variability as the cause of this change in the frequency of arrhythmia. To define proarrhythmia, a different algorithm must be applied. When using antiarrhythmic drugs, a quantitative ventricular arrhythmia baseline for both frequency and type of arrhythmia must be established so that after therapeutic intervention repeat Holter monitoring can determine whether efficacy, inefficacy or proarrhythmia had occurred. Holter monitoring clearly reveals differential antiarrhythmic response rates among classes of antiarrhythmic drugs in patients with benign or potentially lethal arrhythmias. However, preliminary data have not clearly defined the relation between antiarrhythmic pharmacotherapy and a reduction in sudden cardiac death. The results of large-scale clinical trials that have only recently been undertaken must be assessed to determine whether sudden cardiac death can be prevented by adequately suppressing potentially lethal ventricular arrhythmias.