A global hypoxia resulting in an oxygen debt is assumed to be present in patients who suffer from the different stages and degrees of sepsis including septic shock and ARDS. As a consequence, the therapeutic concept of optimal values for cardiac output and oxygen delivery for these patients was proposed. This article reviews the literature with the objective of determining whether investigations dealing with oxygen delivery and consumption and with the plasma lactate concentration support the idea of the global hypoxia in septic patients. The finding of a pathologic oxygen supply dependency and an increase in plasma lactate concentration were taken as evidence for a global hypoxia. Between 1983 and 1991, oxygen supply dependency in septic patients was reported in an increasing number of publications. The increase in plasma lactate concentration was interpreted as lactic acidosis without presentation of plasma pH values and taken as evidence of global hypoxia and oxygen debt. From 1989 on, the number of publications that failed to show oxygen supply dependency even in the presence of an increased plasma lactate concentration increased. The problem in the method of determination of oxygen supply dependency became evident. Deducing both oxygen consumption and oxygen delivery from cardiac output from a common shared variable subject to measurement error may produce errors in the calculation of the regression between oxygen delivery and consumption. Oxygen supply dependency was not demonstrated in most investigations in which oxygen delivery and consumption were measured independently of each other. No decrease in mortality could be shown in prospective randomized studies for patients with sepsis and septic shock who were treated according to the concept of the optimal values. The lactate plasma concentration was below 5 mmol/l in most studies, which represents the borderline value for a clinically significant lactic acidosis. The term acidosis is not justified without a decrease in plasma pH or a decrease in the bicarbonate plasma concentration. An increased lactate plasma concentration can be merely the result of a hypermetabolism which is often found in septic patients. There is no proven evidence for global tissue hypoxia in septic patients from the investigations of oxygen delivery and consumption. This is also true for patients in septic shock after plasma volume expansion. The dogmatic proposal to increase cardiac output and oxygen delivery to certain levels cannot be sustained. However, regional hypoperfusion (e.g., of the splanchnic vascular bed) cannot be excluded. New approaches like gastric mucosal tonometry, measurement of splanchnic blood flow, and determination of regional metabolism are currently under investigation.