The cumulative effects of repeated, brief episodes of regional ischemia on myocardial function and dimension were examined in 14 open-chest dogs. The left anterior descending coronary artery was occluded for 5 minutes, followed by 10 minutes of reflow, repeated 16 times, and then 1 hour recovery. Systolic function decreased progressively in segments made repetitively ischemic and remained depressed even after 1 hour of recovery. Average systolic shortening decreased 20% from baseline after recovery from the first occlusion, 82% after the 8th, 91% after the 16th, and 104% after the 1 hour recovery (p less than 0.015, analysis of variance). End-diastolic segment length progressively increased in regions made repetitively ischemic, lengthening 4% after the first occlusion, 10% after the third occlusion, 19% after the sixteenth occlusion, and 16% after 1 hour of recovery (p less than 0.02). Nonischemic end-diastolic segment length also showed a smaller but parallel increase, while non-ischemic systolic function showed compensatory improvement. After the dogs were killed, myocardial staining with triphenyl tetrazolium chloride revealed no necrosis. Electron microscopy, performed in 5 dogs, showed scattered mitochondrial swelling in both postischemic and nonischemic regions, but no evidence of irreversible injury. The ratio of myocardial blood flow in the region made repetitively ischemic to nonischemic flow, as measured with microspheres, was 1.00 +/- 0.02 before the occlusions and 0.90 +/- 0.03 just before death (difference not significant). Thus, in the dog progressively abnormal regional systolic function and regional and global diastolic dilatation can be produced by repetitive, brief, coronary occlusions, which are not associated with histochemical or ultrastructural evidence of myocardial necrosis.