We tested the hypothesis that ischaemic preconditioning reduces pre-ischaemic energy demand and thereby reduces the energy supply-demand mismatch imposed by coronary artery occlusion. Experiments were performed in 52 open chest pigs anaesthetised with sodium pentobarbital. One or two cycles, each of 10 min LAD occlusion followed by 30 min reperfusion, served as the preconditioning stimuli. The degree of protection was evaluated by measuring infarct size (tetrazolium stain) as percentage of area at risk (fluorescent particles) after 45 min LAD occlusion followed by 2 h reperfusion. One preconditioning cycle reduced regional myocardial oxygen consumption (MVO2) by 15+/-3% (P<0.05), whereas two cycles of preconditioning reduced MVO2 by 25+/-3% (P<0.05 v one cycle). This reduction was probably due to reduced energy demand, as both coronary blood flow and arteriovenous oxygen differences decreased, without lactate release or reduction in peak hyperaemic flow response. Energy requirements were most likely also reduced during ischaemia since repayment of flow debt after the second ischaemic period was 33+/-7% less than after the first ischaemic period (P<0.001). One preconditioning cycle reduced infarct size from 58+/-3% of area at risk to 40+/-5% (P<0.05), whereas two cycles of preconditioning reduced infarct size to 15+/-4% of area at risk (P<0.05 v one cycle). We conclude that preconditioning with ischaemia reduces energy consumption in a graded pattern. This effect may contribute to the graded protective effect of ischaemic preconditioning.