Serial testing of antiarrhythmic drugs by programmed electrical stimulation can be costly in time, expense and risk. The purpose of this study was to evaluate the results of serial electropharmacologic tests for similarities that might obviate the need for protracted drug testing. Serial electropharmacologic testing was performed in 283 patients with coronary artery disease and clinical sustained ventricular tachycardia (VT) or fibrillation (VF). Drug tests were defined as concordant if sustained VT or VF could be consistently induced, or failed to be consistently induced during all such trials in a given patient. The following drugs were included for testing: procainamide, quinidine and disopyramide (class IA); phenytoin, mexiletine and tocainide (class IB); and flecainide and encainide (class IC). All patients were serially tested with > or = 2 (mean and median, 3) antiarrhythmic agents regardless of results from drug-free testing or initial acute drug testing. Overall, the results of serial drug trials directed by programmed stimulation were concordant in more than two thirds of patients. Concordance was comparably high whether patients were serially tested with drugs within the same antiarrhythmic class, or with drugs from differing classes, and was not related to patients' clinical or electrophysiologic characteristics. Protracted serial electropharmacologic testing does not appear necessary for predicting successful or unsuccessful antiarrhythmic drug therapy in survivors of clinical VT or VF. Single drug testing can identify most patients whose arrhythmia will or will not respond to medications.