Postconditioning is known as the phenomenon whereby brief intermittent ischemia applied at the onset of reperfusion following index ischemia limits myocardial infarct size. While there is evidence that the algorithm of the postconditioning stimulus is an important determinant of the protective efficacy, the importance of the duration of index ischemia on the outcome of the effects of postconditioning has received little attention. Pentobarbital-anesthetized Wistar rats were therefore subjected to index ischemia produced by coronary artery occlusions (CAO) of varying duration (15-120-min) followed by reperfusion, without or with postconditioning produced by 3 cycles of 30-s of reperfusion and re-occlusion (3POC30). 3POC30 limited infarct size produced by 45-min CAO (CAO45) from 45±3% to 31±5%, and CAO60 from 60±3% to 47±6% (bothP 0.05). In contrast, 3POC30 increased infarct size produced by CAO15 from 3±1% to 19±6% and CAO30 from 36±6 to 48±4% (both P0.05). This deleterious effect of 3POC30 was not stimulus-sensitive as postconditioning with 3POC5 and 3POC15 after CAO30 also increased infarct size. The cardioprotection by 3POC30 after CAO60 was accompanied by an increased stimulation of Akt-phosphorylation measured after 7 min of reperfusion, and a 36% lower superoxide production measured after 2 hours of reperfusion. Consistent with these results, cardioprotection by 3POC30 was abolished by phosphatidylinositol-3-kinase (PI3K)-inhibition, as well as NO-synthase inhibition. The deleterious effect of 3POC30 after CAO15 was accompanied by an increased superoxide production with no change in Akt phosphorylation, and was not affected by NO-synthase inhibition. In conclusion, the effect of cardiac postconditioning depends critically on the duration of the index ischemia and can be either beneficial or detrimental. These paradoxical effects of postconditioning may be related to the divergent effects on Akt phosphorylation and superoxide production.