Regeneration of infarcted myocardium by injecting stem cells has been proposed to prevent heart failure. We studied the i.c. administration of human umbilical cord blood stem cells (USSC) in a porcine model of myocardial infarction (MI) and reperfusion. In 15 swine, MI was induced by balloon-occlusion of the left circumflex coronary artery (LCX) for 2 h followed by reperfusion. Five swine served as healthy controls. One week later, magnetic resonance imaging (MRI) was performed to assess left ventricular (LV) function and infarct size. Then, under immune suppression, 6 of the 12 surviving MI swine received intracoronary injection of approximately 10(8) human USSC in the LCX while the other MI-swine received medium. Four weeks later all swine underwent follow-up MRI, and were sacrificed for histology. One week after MI, end-diastolic volume (92+/-3 mL) and LV mass (75+/-2 g) were larger, while ejection fraction (42+/-2%) was smaller than in healthy control (68+/-3 mL, 66+/-3 g and 55+/-3%, all P<0.05). Regional wall thickening (-7+/-2%) in the LCX area became akinetic. No difference in global and regional LV function at 5 weeks was observed between MI animals receiving USSC or medium. Infarct size after USSC treatment was significantly larger (20+/-3 g vs. 8+/-2 g, P<0.05). USSC survived only in the infarct border zone at 5 weeks and did not express cardiomyocyte or endothelial markers. Histology showed that intracoronary injection of USSC caused micro infarctions by obstructing blood vessels. In swine with a 1 week old MI, injection of USSC via the intracoronary route does not improve LV function 4 weeks later.