The efficacy and safety of autodecremental pacing (ADP) to interrupt ventricular tachycardia (VT) and atrial flutter was examined. Once tachycardia was recognized, ADP was initiated using a short train of stimuli with gradual shortening (3%) of the interstimulus interval. ADP was applied to 13 consecutive patients during 75 episodes of VT (mostly following induction by ventricular stimulation). Successful interruption of VT occurred in 88% of the episodes. In 6 episodes (8%), ADP resulted in ventricular fibrillation and in 3 episodes VT was unaffected by ADP. The only significant discriminator between the failure or success of ADP was the rate of VT. ADP was also applied to 17 consecutive patients with an atrial flutter that was resistant to conventional antiarrhythmic agents. Successful conversion of atrial flutter to sinus was seen in only 8 patients (47%). A temporary acceleration to atrial fibrillation appeared in 3 patients (18%), and in 6 patients atrial flutter was unaffected by ADP. ADP was successful in 70% (7/10) of patients with type 1 (< 300 beats/min) atrial flutter. The authors conclude that ADP is beneficial in the interruption of VT and atrial flutter in a selected group of patients, especially with a slower rate of tachyarrhythmia (atrial rate during atrial flutter < 300 beats/min and ventricular tachycardia < 180 beats/min).