Functional recovery with surgical revascularization of acutely ischemic myocardium has not been compared with its nonsurgical counterpart in experimental preparations of coronary occlusion. This study compares the functional and metabolic recovery of ischemic (1 hr coronary occlusion) segments revascularized either by restoration of coronary patency (simulating nonsurgical recanalization, e.g., angioplasty) or by surgical revascularization with multidose hypothermic potassium blood cardioplegic solution. Twenty-two anesthetized open-chest dogs were instrumented with Millar micromanometer-tip catheters to measure left ventricular and aortic pressures. Piezoelectric ultrasonic dimension gauges were implanted in the subendocardium supplied by the left anterior descending coronary artery to measure segmental contractile function. In five dogs, only biopsy samples were obtained for control measurements of ATP, creatine phosphate, and tissue water content. In the remaining 17 dogs, the left anterior descending artery and collaterals were ligated for 1 hr. The ligatures were removed in eight dogs and coronary perfusion continued for 2 hr, simulating nonsurgical reperfusion. The remaining nine dogs were placed on cardiopulmonary bypass and the hearts were arrested for 1 hr with multidose (every 20 min) blood cardioplegic solution enhanced with glutamate and aspartate, simulating surgical revascularization (coronary artery bypass grafting). The coronary ligatures were not released until the second cardioplegic infusion, simulating graft placement. One hour of coronary occlusion placed 39.4 ± 2.5% of the left ventricle at risk, and converted active systolic shortening to persistent paradoxical bulging (25.2 ± 2.2% to -5.8 ± 1.2% systolic shortening). Surgical revascularization resulted in greater recovery of postischemic systolic shortening (7.95 ± 1.90% vs -6.90 ± 2.01% systolic shortening; p < .05), lower subendocardial water content (79.36 ± 0.50% vs 81.14 ± 0.53%; p < .05), and less gross histochemically apparent (triphenyltetrazolium chloride) tissue damage (4.2 ± 4.2% vs 30.9 ± 6.8% of area at risk; p < .05) compared with hearts reperfused nonsurgically. There were no differences in regional ATP and creatine phosphate levels between surgically and nonsurgically revascularized hearts despite the marked differences in functional recovery. We conclude that (1) myocardium subjected to 1 hr of coronary occlusion is potentially recoverable depending on the modality of the early phase of reperfusion and (2) surgical revascularization avoids reperfusion injury resulting in immediate postischemic functional recovery.