Reperfusion achieved by streptokinase infusion early after myocardial infarction (MI) is now being performed in patients, but the effect on electrical instability of increasing or decreasing perfusion in the region at risk for MI is unknown. Accordingly, 34 dogs were randomized to control (13 dogs), reperfusion (11 dogs) and retrograde bleeding (10 dogs) groups. All dogs underwent coronary artery occlusion (23 of the left anterior descending and 11 of the circumflex artery). In the control dogs, occlusion was permanent. In the reperfused dogs, the occlusion was released at 2 hours. In the retrograde bleeding dogs, retrograde flow bleeding distal to the occlusive tie was continued for 2 hours after coronary occlusion. Four days later, all dogs underwent a standard right ventricular pacing protocol. Induced arrhythmias were scored; ventricular fibrillation was assigned the highest score, followed by sustained ventricular tachycardia, nonsustained ventricular tachycardia and repetitive ventricular response. Arrhythmias provocable later in diastole were assigned higher scores than those provocable early in diastole. Infarct size was not different in the three groups (35%, 28% and 39% of the area at risk in control, reperfusion and retrograde bleeding groups, respectively). However, the electrical instability index was lower in the reperfusion group than in the other two groups (e.g., electrical instability index A at 200 beats/min: p less than 0.005 for reperfusion vs control; p less than 0.01 for reperfusion vs retrograde bleeding). Retrograde bleeding did not alter the electrical instability index from the control state. These results suggest that despite no significant reduction in infarct size, reperfusion after infarction may reduce electrical instability.