The diagnostic accuracy of programmed electrical stimulation was prospectively assessed in 111 patients with myocardial infarction (MI) with or without a history of spontaneous ventricular arrhythmias. In 29 patients neither ventricular tachycardia (VT) nor episodes of 10 premature ventricular depolarizations per hour was documented. Fifty patients had documented nonsustained VT and 32 had sustained monomorphic VT. One and 2 extrastimuli (twice diastolic threshold, 2 ms in duration) were given during sinus rhythm and ventricular pacing at 100, 120 and 140 beats/min in the right ventricular apex (part I). When this protocol failed to induce a sustained monomorphic VT, a third extrastimulus was introduced (part II). Repetitive ventricular responses were induced in all patients, and in 15 (14%) polymorphic ventricular arrhythmias requiring DC shock were induced. Incidence of initiation of sustained monomorphic VT and polymorphic ventricular arrhythmias requiring DC shock was related to the clinical arrhythmia and the stimulation protocol. In patients with documented sustained monomorphic VT, a third extrastimulus only increased the incidence of sustained monomorphic VT (68% to 94%), whereas in patients with documented nonsustained VT and without VT the incidence of both polymorphic and monomorphic arrhythmias increased by 7 to 12%. Sustained monomorphic VTs induced in patients without such a history were faster (p less than 0.01), depended on site of MI (p less than 0.05) and were more often preceded by nonsustained polymorphic VT (p less than 0.01) than in patients with documented sustained monomorphic VT.(ABSTRACT TRUNCATED AT 250 WORDS)