Temporary coronary occlusion followed by reperfusion severely reduces contractile function in the involved segment. We tested whether an uncoupling exists between O2 utilization (MVO2) and systolic shortening in the ischemic-reperfused segment subjected to repetitive coronary occlusion and reperfusion. In 10 anesthetized open-chest dogs, left ventricular pressure and segment length (sonomicrometry) relations were measured in the left anterior descending (LAD, ischemic-reperfused) segment and circumflex coronary artery (nonischemic segment). Four 12-min LAD occlusions were each followed by 30 min reperfusion. MVO2 was determined in both segments by transmural blood flow (15 microns microspheres) and regional coronary arterial-venous O2 extraction after each occlusion-reperfusion period. The four occlusion-reperfusion periods did not produce necrosis by staining with triphenyltetrazolium chloride. LAD occlusion produced dyskinesis [control = 16 +/- 3.0% systolic shortening (SS) vs. -8.8 +/- 1.5%, P less than 0.0001]. The first reperfusion restored SS only to 2.3 +/- 2.0%, which progressively deteriorated to -3.9 +/- 1.1% (P less than 0.05) with subsequent occlusion-reperfusion episodes. Relative to the nonischemic segment, MVO2 in the ischemic-reperfused segment decreased by only 18% despite dyskinesis. Pressure-length analysis showed systolic stiffening during reperfusion with displacement of the passive ischemic pressure-length loop to the left. Segment work (integral of each loop) continued to be generated at 34.5% of control levels after the last occlusion-reperfusion event in contrast to the negative SS. We conclude that 1) MVO2 in the ischemic-reperfused segment without necrosis remains elevated despite severe reductions in systolic shortening, and 2) the discrepancy between systolic shortening and MVO2 is partially due to persistent development of segment work.