While some investigators recognize "reperfusion-induced injury" as an important component of the overall injury that occurs during ischemia and reperfusion, others question its existence. Resolution of this controversy is of considerable importance, particularly in an era of thrombolysis, since reperfusion-induced injury might be amenable to treatment. Although reperfusion is an absolute prerequisite for the recovery of ischemic tissue, it undoubtedly has some unfavorable effects. The identification of four (possibly sequential) components of reperfusion-induced injury helps to clarify the situation: a) Reperfusion after brief periods of ischemia can trigger arrhythmias in tissue that is potentially salvable; there is abundant experimental and clinical evidence for this form of reperfusion injury. b) Reperfusion may also be associated with "myocardial stunning"; however, given sufficient time, this prolonged postischemic contractile and metabolic dysfunction will recover. There is good experimental evidence and some clinical evidence for the existence of this type of reperfusion-induced injury. c) Reperfusion is commonly thought to cause lethal injury in cells that, until the time of reperfusion, were potentially salvable. However, conclusive evidence that reperfusion can kill cells does not yet exist. d) Reperfusion may alter the nature of necrotic processes in tissue that has already sustained lethal injury, while not altering the number of cells that die this may change the manner in which they die; this form of reperfusion injury could lead to differences in scar formation and vulnerability to aneurysm. There is considerable evidence for the existence of this form of reperfusion-induced injury. Many candidate mechanisms have been proposed for each form of reperfusion injury. Ionic disturbances (particularly for calcium) are often cited and, most recently, free radical-induced induced injury (oxidant stress) has been suggested as important. Considerable evidence exists that oxidant stress is involved in stunning and in reperfusion-induced arrhythmias, and characterization of underlying mechanisms might lead to novel therapeutic principles, such as antioxidant therapy. However, much remains to be learned.