We asked whether oxygen consumption is dependent on oxygen delivery in 17 patients who had severe adult respiratory distress syndrome (ARDS), 10 of whom had increased concentrations of plasma lactate. We determined oxygen consumption using analysis of respiratory gases while increasing oxygen delivery using blood transfusion. Oxygen consumption did not change after transfusion (from 227 +/- 83 to 225 +/- 82 ml/min, p less than or equal to 0.38). Oxygen delivery increased from 1,043 +/- 468 ml/min (24%, p less than or equal to 0.001). Even in the 10 patients who had increased concentration of plasma lactate and metabolic acidosis, oxygen consumption remained constant after increasing oxygen delivery (pretransfusion, 224 +/- 101 ml/min; post-transfusion, 225 +/- 99 ml/min; p less than or equal to 0.83). These data have more than 99% power of detecting a change in oxygen consumption of 20 ml/min after transfusion. Therefore, we conclude that directly measured oxygen consumption remains constant and independent of increases in oxygen delivery in our patients with severe ARDS. Because simultaneously determined oxygen consumption calculated from variables shared with the calculation of oxygen delivery yielded a dependent relationship, we speculate that finding dependence of calculated oxygen consumption on oxygen delivery may be the result of methodologic error.