Verheugt 1 argues, with strong reasons, that reperfusion therapy should start in the ambulance. However, reperfusion brings with it reperfusion injury, resulting both from reactive oxygen species (ROS) and a metabolic mechanism, a rise in cytosolic calcium with cell contracture. "Reperfusion injury" is a euphemism for cell death, which occurs at least in part through ATP-requiring apoptosis triggered by ROS and calcium. 2 That such reperfusion-induced cell death also occurs in the reperfused human acute myocardial infarction is strongly suggested by the recent study of Staat et al, 3 who reduced enzyme release by about one third in ST-elevation myocardial infarction by mechanical postconditioning. The latter process can only be applied at the time of PCI, not in the ambulance, thus leaving the persistent risk of reperfusion-induced cell death.It is time to take this understanding to the ambulance as well. The principles of antiapoptotic therapy that can be applied at reperfusion include decreased cytosolic calcium levels that can be achieved by increased provision of glucose-mediated generation of glycolytic ATP. Early in reperfusion, excessive use of fatty acids inhibits glycolysis and glucose oxidation. Conversely, insulin decreases the adverse effects of free fatty acids, 4 which are high during early reperfusion, thereby promoting glucose oxidation. Hence, it would be logical to test provision of insulin at the time of reperfusion. An even more cogent reason for giving insulin at reperfusion is that it stimulates the protective signaling paths that activate the key signaling molecule Akt to promote cell survival and powerfully to lessen apoptosis. Jonassen et al 5 have shown that insulin must be given at the time of reperfusion, or within 15 minutes thereof, to protect and decrease experimental infarct size. Testing this hypothesis in the ambulance would require a trial specifically designed to give reperfusion covered by insulin with glucose added to prevent hypoglycemia, also given in the ambulance. If reperfusion only occurred later, the same need for insulin coverage would hold. The earlier insulin is started, the better, because some spontaneous reperfusion is possible during the early stages when the thrombus might be forming with some breakdown. A major clinical trial to test such metabolic therapy at reperfusion could theoretically aim to reduce infarct size by one third, thus having a major impact on mortality and subsequent development of heart failure. It is possible that reperfusion should start in the ambulance, just after starting myocardial metabolic support.