Single or multiple brief periods of ischemia (preconditioning) have been shown to protect the myocardium from infarction after a subsequent more prolonged ischemic insult. To test the hypothesis that preconditioning is the result of opening ATP-sensitive potassium (KATP) channels, a selective KATP channel antagonist, glibenclamide, was administered before or immediately after preconditioning in barbital-anesthetized open-chest dogs subjected to 60 minutes of left circumflex coronary artery (LCX) occlusion followed by 5 hours of reperfusion. Preconditioning was elicited by 5 minutes of LCX occlusion followed by 10 minutes of reperfusion before the 60-minute occlusion period. Glibenclamide (0.3 mg/kg i.v.) or vehicle was given 10 minutes before the initial ischemic insult in each of four groups. In a fifth group, glibenclamide was administered immediately after preconditioning. In a final series (group 6), a selective potassium channel opener, RP 52891 (10 micrograms/kg bolus and 0.1 micrograms/mg/min i.v.) was started 10 minutes before occlusion and continued throughout reperfusion. Transmural myocardial blood flow was measured at 30 minutes of occlusion, and infarct size was determined by triphenyltetrazolium staining and expressed as a percent of the area at risk. There were no significant differences in hemodynamics, collateral blood flow, or area at risk between groups. The ratio of infarct size to area at risk in the control group (28 +/- 6%) was not different from the group pretreated with glibenclamide in the absence of preconditioning (31 +/- 6%). Preconditioning produced a marked reduction (p less than 0.002) in infarct size (28 +/- 6% to 6 +/- 2%), whereas glibenclamide administered before or immediately after preconditioning completely abolished the protective effect (28 +/- 6% and 30 +/- 8%, respectively). RP 52891 also produced a significant (p less than 0.03) reduction (28 +/- 6% to 13 +/- 3%) in infarct size. These results suggest that myocardial preconditioning in the canine heart is mediated by activation of KATP channels and that these channels may serve an endogenous myocardial protective role.