Introduction To test the hypothesis that administration of adenosine during reperfusion attenuates endothelial dysfunction and extension of infarct size by inhibiting polymorphonuclear neutrophil (PMN)-mediated events and apoptosis,Methods Anesthetized dogs were subjected to 1 h coronary artery occlusion and 6 h of reperfusion with infusion of saline (vehicle, n=8) or 140 mu g/kg per min adenosine, n=8) continuously into the left atrium starting 5 min before reperfusion and continuing for 2 h,Results There was no intergroup difference in collateral myocardial blood flow measured by using colored microspheres in the area at risk during ischemia. infusion of adenosine transiently improved segmental shortening (4.1 +/- 3.1% versus -2.5 +/- 2.3%, P< 0.05) and segmental work (41.4 +/- 22 versus 15 +/- 13 mmHg/mm, P<0.05) after 4 h of reperfusion, Infusion of adenosine reduced size of infarct (determined by staining with triphenyltetrazolium chloride) from 27 +/- 2% with vehicle to 14+/-1%(P<0.05). This was confirmed by measuring that it lowered activity of plasma creatine kinase (from 19+/-2 versus 8+/-1 IU/g protein, P<0.05). It also reduced the proportion of terminal deoxynucleotidyl transferase-mediated dUTP nick-end labeling-positive nuclei in the perinecrotic zone from 17.3 +/- 1,6 to 10.3 +/- 1.0% (P< 0.05) and reduced the appearance of DNA ladders in gel electrophoresis. In addition, it significantly decreased accumulation of PMN in the ischemic area (determined by immunohistochemistry with anti-CD18 antibody) and activity of cardiac myeloperoxidase compared with vehicle (439 +/- 52 versus 183 +/- 20 PMN/mm(2) myocardium and 1,1 +/- 0.1 versus 2.4 +/- 0.2 U/100 mg tissue, P<0.05, respectively). Furthermore, infusion of adenosine during reperfusion preserved vascular endothelial function expressed in terms of a decrease in adherence of PMN to postischemic coronary artery endothelium (63 +/- 3 versus 36+/-4 PMN/mm(2) endothelium, P<0.05, basal function) and agonist (acetylcholine)-induced endothelium-dependent relaxation (negative logarithm to base 10 of concentration (mol/l) for half-maximal effect 77 +/- 0.1 versus 7.2 +/- 0.1, P < 0.05, stimulated function). Infusion of adenosine directly inhibited generation of superoxide radical from canine PMN in vitro dose dependently from 27.8 +/- 6.3 to 5.8 +/- 2.1 nmol/l/5 x 10(6) PMN (P< 0.05).Conclusion Intra-atrial infusion of adenosine during reperfusion reduced accumulation of PMN in area at risk, preserved vascular endothelial function after ischemia reperfusion by inhibiting interaction between PMN and endothelial cells, and decreased extension of infarct, possibly by limiting apoptosis, Coronary Artery Dis 10:617-628.