We sought to determine whether administration of a very low, nonvasodilating dose of a highly selective adenosine A(2A) receptor agonist (ATL-193 or ATL-146e) would be cardioprotective in a canine model of myocardial stunning produced by multiple episodes of transient ischemia. Twenty-four anesthetized open-chest dogs underwent either 4 (n=12) or 10 cycles (n=12) of 5-min left anterior descending coronary artery (LAD) occlusions interspersed by 5 or 10 min of reperfusion. Left ventricular thickening was measured from baseline through 180 min after the last occlusion-reperfusion cycle. Regional flow was measured with microspheres. In 12 of 24 dogs, A(2A) receptor agonist was infused intravenously beginning 2 min prior to the first occlusion and continuing throughout reperfusion at a dose below that which produces vasodilatation (0.01 microg x kg(-1) x min(-1)). Myocardial flow was similar between control and A(2A) receptor agonist-treated animals, confirming the absence of A(2) receptor agonist-induced vasodilatation. During occlusion, there was severe dyskinesis with marked LAD zone thinning in all animals. After 180 min of reperfusion following the last cycle, significantly greater recovery of LAD zone thickening was observed in A(2A) receptor agonist-treated vs. control animals in both the 4-cycle (91 +/- 7 vs. 56 +/- 12%, respectively; P<0.05) and the 10-cycle (65 +/- 9 vs. 8 +/- 16%, respectively; P<0.05) occlusion groups. The striking amount of functional recovery observed with administration of low, nonvasodilating doses of adenosine A(2A) agonist ATL-193 or ATL-146e supports their further evaluation for the attenuation of postischemic stunning in the clinical setting.