Definitions of proarrhythmia, including clinical consequence, were applied to the flecainide and encainide data bases to determine risk factors for serious proarrhythmic events or deaths. Such outcomes with flecainide were far less common for patients with benign or potentially lethal ventricular arrhythmias compared to patients with predominantly lethal ventricular arrhythmias. No deaths from proarrhythmia during flecainide therapy occurred in patients without structural heart disease. Serious proarrhythmic events and deaths were more common in patients in whom therapy was initiated in hospital than in those in whom therapy was initiated out of hospital. When the flecainide dosage for patients with lethal ventricular arrhythmias was chosen using steady-state pharmacologic principles, the occurrence of all proarrhythmic events and deaths dropped from 26% and 13% to 10% and 0%, respectively. Structural heart disease, sustained ventricular tachycardia, inpatient initiation and large-dose escalation of class IC drugs are the primary risk factors for development of proarrhythmic events.