To assess the spontaneous variability of ventricular arrhythmias after withdrawal of apparently successful antiarrhythmic therapy, we enrolled in a washout protocol 40 patients who had completed a randomized controlled 2-year study of antiarrhythmic drugs (the Antiarrhythmic Drug Evaluation Group [ADEG] study). All of them had heart disease and were first seen with high-grade ventricular arrhythmias (Lown class 4a and 4b) at enrollment. After 2 years all of them had responded to propafenone or flecainide; patients who completed the study on a regimen of amiodarone were not considered for the washout study. A total of 27 patients discontinued flecainide and 13 propafenone. Seven days after withdrawal they underwent 24- or 48-hour ECG testing and were classified as true responders (TR) if the arrhythmias were present at the same level as 2 years previously and false responders (FR) if they were below the ADEG responsiveness level. TR patients had a third 24-hour ECG 7 days later, after rechallenge with the same treatment, and FR patients had a third ECG without drugs. Adherence to the protocol was ascertained by measuring drug plasma concentrations at every 24-hour ECG recording. No differences were found in distribution of heart disease and grade of ventricular arrhythmias between patients in the washout study and the remaining group of the ADEG trial. Twenty-four of 40 patients (60%) were true responders. In 4 of the 17 patients who had a third 24 hour ECG, the responsiveness to the same drug was not confirmed.(ABSTRACT TRUNCATED AT 250 WORDS)