To discuss the different techniques used to assess tissue oxygenation in critically ill patients.While historically the analysis of oxygen consumption (VO2)/oxygen delivery (DO2) relationships has provided important information, methodological limitations prevent its use at bedside. PO2 measurements, while attractive, are unfortunately of limited value in the presence of microvascular blood flow heterogeneity which is observed in many critically ill conditions including sepsis. Surrogates of tissue oxygenation are hence used. Elevated lactate levels may suggest inadequate tissue oxygenation, but other sources than tissue hypoxia can also contribute to hyperlactatemia so that lactate measurements should be used in combination with other measurements of tissue oxygenation. Venous O2 saturation can be used to evaluate the adequacy of DO2 in respect to VO2, but it can be misleading normal or even high in sepsis. Measurements of Pv-aCO2 and computation of Pv-aCO2/CavO2 are very promising as physiologically sound, easy to measure, rapidly respond to therapy, and are associated with outcome. An elevated Pv-aCO2 reflects an impaired tissue perfusion while an increased Pv-aCO2/CavO2 ratio reflects tissue dysoxia.Recent studies have highlighted the interest of surrogate measurements of tissue oxygenation and in particular PCO2 gradients.