Reperfusion of ischemic myocardium may accelerate necrosis of injured myocytes. To determine the role of neutrophil leukocytes in this process, we examined whether neutrophil depletion during reperfusion could modify infarct size in anesthetized dogs. The proximal circumflex coronary artery was occluded for 90 minutes and then reperfused for 2 hours via an extracorporeal circuit with either whole blood (n = 11) or with blood depleted of neutrophils by leukocyte filters (n = 11). The leukocyte filters caused near-total neutropenia in blood reperfusing the ischemic myocardium (7 +/- 7 neutrophils/microliters compared with 2,551 +/- 317/microliters in controls, mean +/- SEM; p less than 0.001. Infarct size was measured by planimetry of myocardial slices stained with triphenyltetrazolium chloride (TTC), and the accuracy of TTC for identifying necrotic myocardium was verified by electron microscopy. The size of the ischemic risk region was the same in the control (41.6 +/- 1.0%) and neutropenic (41.8 +/- 2.1%) groups. Collateral blood flow to the risk region was the same in control (0.15 +/- 0.03 ml/min/g) and neutropenic (0.13 +/- 0.03 ml/min/g) groups. Among dogs with collateral flow less than 0.2 ml/min/g, infarct size was reduced in the neutropenic group (27.7 +/- 6.7% of risk region, n = 8), compared with control dogs (52.5 +/- 5.7%; n = 7; p = 0.02). Multiple linear regression described the relation between infarct size, risk region size, and collateral flow in the control group, and the same regression relation was used to predict infarct size for the neutropenic group. Mean predicted infarct size in the neutropenic group (n = 11) was 16.8 +/- 3.4% of left ventricle, whereas mean observed infarct size was 9.6 +/- 3.1% (p less than 0.01). The extent of the no-reflow zone (absence of thioflavin-S-fluorescence) was also less in the neutropenic than the control group (2.2 +/- 0.8% vs. 8.1 +/- 2.7% of the risk region, p less than 0.05). Neutropenia limited to the reperfusion period is associated with significant reductions in the extent of the infarct and no-reflow zones after 90 minutes of ischemia. These findings support the hypothesis that reperfusion necrosis occurs after prolonged myocardial ischemia and indicate that neutrophil leukocytes are important mediators of such reperfusion injury.