Background: The insulin-sensitizing drug pioglitazone has been reported to be protective against myocardial infarction. However, its precise mechanism is unclear. Methods: Rabbits underwent 30 min of coronary occlusion followed by 48h of reperfusion. Rabbits were assigned randomly to 9 groups (n=10 in each): the control group (fed a normal diet), pioglitazone group (fed diets containing 1mg/kg/day pioglitazone), pioglitazone+5HD group (fed the pioglitazone diet + i.v. 5mg/kg 5-hydroxydecanoate, a mitochondrial KATP channel blocker), pioglitazone+GW9662 group (fed the pioglitazone diet + i.v. 2mg/kg GW9662, a PPAR- antagonist), GW9662 group (fed a normal diet + i.v. GW9662), and pioglitazone+wortmannin group (fed the pioglitazone diet + i.v. 0.6mg/kg wortmannin, a phosphatidylinositol 3-kinase (PI3-kinase) inhibitor), wortmannin group (fed a normal diet + i.v. wortmannin), pioglitazone+L-NAME group (fed the pioglitazone diet + i.v. 10mg/kg L-NAME, a NOS inhibitor), L-NAME group (fed a normal diet + i.v. L-NAME). All group fed the diets for 7 days. The risk area and non-risk area of the left ventricle (LV) were separated by Evans blue dye and the infarct area was determined by TTC staining. The infarct size was calculated as a percentage of the LV risk area. Western blotting was performed to assess levels of Akt and phospho-Akt, phospho-eNOS in the myocardium following reperfusion. Results: The infarct size was significantly smaller in the pioglitazone group (21±2%) than in the control group (43±3%). This effect was abolished by GW9662 (42±3%), wortmannin (40±3%) or L-NAME (42±7%) but not by 5-HD (24±5%). Western blotting showed higher levels of phospho-Akt and phospho-eNOS in the pioglitazone group. Conclusions: Pioglitazone reduces the myocardial infarct size via activation of PPAR-, PI3 kinase, Akt and eNOS pathway, but not via opening the mitochondrial KATP channel. Pioglitazone may be a novel strategy for the treatment of diabetes mellitus with coronary artery disease.