Initial favorable reports in which coronary venous retroperfusion was begun after acute coronary artery occlusion have demonstrated a reversal of ischemic injury and improved left ventricular function. However, little information has been generated to document the extent to which retroperfusion may decrease ultimate histologically determined infarct size. The objective of the present study was to evaluate the effectiveness of retroperfusion in reducing infarct size by using an accurate quantitative method in which infarct size was related to the size of the anatomic perfusion bed of the occluded artery (region at risk for infarction). In an experimental group of 5 baboons, the left anterior descending coronary artery was occluded and coronary venous retroperfusion started 1 hour after occlusion. After a 4-hour period of occlusion, retroperfusion was discontinued and anterograde perfusion was simultaneously restored. A control group of 5 baboons underwent an identical procedure without retroperfusion. Twenty-four hours after occlusion, hearts were excised and the previously occluded left anterior descending coronary artery as well as the adjacent arteries were infected with microvascular dye to delineate the perfusion bed of the occluded artery. Planimetry of serial corss-sections of the left ventricle enabled the size of the perfusion bed of the occluded artery and size of the infarct to be determined. The mean percentage of the perfusion bed infarcted in the control group was 94.1 +/- 0.9 (mean +/- standard error) and in the retroperfused group was 57.4 +/- 3.5 (p less than 0.001). Hence, the results demonstrated that when retroperfusion was initiated after 1 hour of coronary occlusion, the mean percentage of the perfusion bed salvaged was increased by 36.7%.