OBJECTIVE The aim was to examine whether ischaemic preconditioning can be renewed by a second cycle of brief coronary occlusions in pigs subjected to two different reperfusion intervals (1 h or 4 d). METHODS Ischaemic preconditioning was induced by a cycle of two 10 min occlusions of the left anterior descending coronary artery separated by 30 min of reperfusion. Infarction was induced with a subsequent 1 h occlusion and a 2 h reperfusion period. There were four experimental groups: in group I (n = 5), a 30 min reperfusion was interposed after the preconditioning cycle prior to the sustained occlusion; in group II (n = 5), this time frame was extended to 1 h; in group III (n = 5), the preconditioning cycle was renewed 1 h after the first cycle; in group IV (n = 5), the second cycle was performed 4 d later. Control pigs (n = 5) were subjected to 1 h coronary occlusion and 2 h reperfusion without previous short occlusions. Infarct size was measured with p-nitro blue tetrazolium and was expressed as a percent of area at risk. RESULTS The percent of the risk region infarcted was 69.9 (SEM 3.8)% for controls, 22.9 (3.5)% in group I (p < 0.001 v controls), 67.3 (5.2)% in group II, 66.3 (4.2)% in group III, and 17.9 (3.9)% in group IV (p < 0.001 v controls). Regional wall function measured with ultrasonic crystals deteriorated through the reperfusion intervals, indicating different underlying mechanisms for ischaemic preconditioning and myocardial stunning. CONCLUSIONS Ischaemic preconditioning with two 10 min occlusions reduced infarct size resulting from a 60 min coronary occlusion when that was performed 30 min after the last short occlusion. This effect was lost after 1 h. Preconditioning could be renewed by a second cycle of brief coronary occlusion and reperfusion 4 d but not 1 h after the first cycle. These results suggest the release of a mediator from an exhaustible pool.