Previous studies indicate impairment of coronary arterial ring relaxation and loss of coronary vasodilator reserve after coronary artery occlusion and reperfusion. These changes are mediated in part through loss of endothelium-derived relaxing factor (EDRF) and/or myocardial neutrophil accumulation. To examine if superoxide dismutase (SOD), a scavenger of superoxide radicals, would modify the diminished coronary vasodilator reserve after temporary coronary occlusion in the intact animal, open-chest mongrel dogs were subjected to 1 hour of circumflex (Cx) coronary artery occlusion followed by 1 hour of reperfusion and treated with saline or SOD. Before Cx occlusion, coronary blood flow increased, and vascular resistance decreased (both p less than 0.01) in response to EDRF-dependent vasodilator acetylcholine as well as EDRF-independent vasodilator nitroglycerin. After Cx reperfusion, resting Cx coronary blood flow and vascular resistance were similar to the preocclusion values. In the saline-treated animals, there was evidence of myocardial dysfunction, which was measured by segmental shortening (-6 +/- 2% vs. 10 +/- 2%). Furthermore, increase in Cx coronary blood flow and reduction in vascular resistance in response to both vasodilators were significantly (p less than 0.01) impaired; these occurrences suggested loss of coronary vasodilator reserve. Myocardial histology showed extensive capillary plugging by neutrophils in the Cx-supplied myocardium. Myocardial myeloperoxidase activity, an index of neutrophil infiltration, was also increased in the Cx compared with the left anterior descending coronary artery region (p less than 0.02). Treatment of dogs with SOD, started at the end of Cx occlusion and continued during reperfusion, exerted significant (p less than 0.01) protective effect against reperfusion-induced attenuation of coronary vasodilator reserve in response to both acetylcholine and nitroglycerin. Loss of myocardial function (segmental shortening 5 +/- 1% vs. 10 +/- 1%) was less than in the saline-treated animals (p less than 0.01). Cx region-myocardial neutrophil accumulation and myeloperoxidase activity were also less (p less than 0.02) in the SOD-treated than in the saline-treated dogs. These observations suggest that coronary reperfusion impairs coronary vasodilator reserve in intact dogs. This impairment can be modified by treatment of animals with SOD before reperfusion. Capillary plugging by neutrophils may contribute to the altered coronary vasodilator reserve observed in the immediate postreperfusion period, and SOD modifies this reperfusion-induced impairment.