During coronary angioplasty, a stair-step decrease in peak S-T segment elevation from the first to the second coronary occlusion has been assumed to indicate a preconditioning (PC) effect. This association was evaluated with myocardial electrograms in rabbits, which revealed that two sequential 5-min coronary occlusions resulted in a marked decrease in the area under the S-T segment voltage-time curve (P < 0.05) with no change during a third occlusion. Pretreatment with either 5-hydroxydecanoate, a mitochondrial ATP-sensitive potassium (K(ATP)) channel blocker, or anisomycin, an activator of stress-activated protein kinases, had no effect on the stair-step decline in the S-T segment voltage between the first two occlusions. HMR-1883, a potent closer of sarcolemmal K(ATP) channels, abolished changes in S-T segment elevation after brief coronary occlusions but had no effect on the infarct-sparing property of the two preconditioning 5-min occlusions. Interestingly, HMR-1883 blocked myocardial protection from diazoxide, raising doubt that the latter opens only mitochondrial channels. Therefore, myocardial protection and S-T segment changes during ischemia are dissociated. These data suggest that it is the mitochondrial K(ATP) channel that protects the myocardium, and it is the sarcolemmal channel that is responsible for changes in S-T elevation. Therefore, it cannot always be inferred that changes in S-T segment elevation reflect the state of myocardial protection.