Although ischemic preconditioning protects myocardium from infarction in isolated hearts and in anesthetized open-chest animals, its effects have not been examined in unanesthetized animals. Furthermore, it is unknown whether animals become tolerant to multiple episodes of ischemic preconditioning. Rabbits were chronically instrumented with a balloon occluder around a major branch of the left coronary artery for reversible coronary occlusion, a left atrial catheter for radioactive microsphere injections, ECG electrodes for monitoring of myocardial ischemia, and, in some cases, a carotid artery catheter for pressure measurements and timed withdrawal of reference arterial blood samples. Eight control rabbits underwent a 30-minute coronary occlusion and then 180 minutes of reperfusion. Five of the eight rabbits developed ventricular tachycardia or fibrillation during ischemia, and infarct size averaged 37.7+/-2.6% of the risk area, Eight rabbits experienced a 5-minute coronary occlusion and 10 minutes of reperfusion before the 30-minute occlusion. In these preconditioned animals, potentially fatal arrhythmias during ischemia were significantly reduced (one of eight, P<.05), and infarct size was much smaller (5.6+/-1.1%, P<.0001). The difference could not be explained by hemodynamics or collateral blood flow, which were nearly identical in the two groups. But when the 30-minute coronary occlusion was preceded by 40 to 65 five-minute occlusions during a 3- to 4-day period in seven animals, protection was markedly attenuated. Potentially lethal arrhythmias were very common, and infarct size averaged 26.5+/-2.9%, substantially larger than in rabbits with only one preconditioning occlusion (P<.0001). Finally, if an interval of 2.5 to 3 days of no coronary occlusions was interposed between this period of multiple occlusions and the terminal preconditioning protocol of 5-minute occlusion and 10-minute reperfusion preceding the long 30-minute occlusion and 180-minute reperfusion, protection was again evident, and infarct size in seven rabbits averaged 10.9+/-1.5% of the risk zone (P<.0001 versus control and P<.001 versus multiple occlusions). It can be concluded-that ischemic preconditioning significantly diminishes the incidence of ischemia-induced ventricular arrhythmias and the extent of infarction after a 30-minute coronary occlusion in unanesthetized rabbits. Unfortunately, this protection wanes after multiple 5-minute coronary occlusions have occurred but does reappear after an ischemia-free period. It is currently being hoped that ischemic preconditioning will have clinical importance, but tolerance may limit its utility in patients with recurrent angina pectoris.