SummaryLung injury, including pneumonia, can occur in the early postoperative period following thoracic surgery. Pulmonary oxygen consumption is thought to increase in patients with pulmonary infection. This study measured oxygen consumption in relationship to lung injury in the early postoperative period after thoracic surgery. Thirty-five patients who underwent thoracoabdominal oesophagectomy for oesophageal cancer were studied. Measured oxygen-consumption was obtained by indirect calorimetry and calculated oxygen-consumption was simultaneously determined by the reverse Fick method. The difference in oxygen consumption was attributed to pulmonary oxygen consumption. The difference in oxygen consumption increased to 23.1 ml.min .m )2 on postoperative day 2. In patients with pneumonia the difference in oxygen consumption increased significantly to 39.0 ml.min )1 .m )2 the day before clinical onset of pneumonia, and it increased further to 65.7 ml.min )1 .m )2 on the day that pneumonia became clinically apparent. These findings suggest that the difference in oxygen consumption may be useful for estimating the extent of lung injury and for predicting pulmonary complications in the postoperative period. In recent years, there has been increasing interest in the measurement of oxygen consumption (VO 2 ) in critically ill patients as VO 2 is related to the clinical outcome of this population. VO 2 can be measured either by indirect calorimetry, using analysis of respiratory gas exchange, or with the reverse Fick method, as the product of thermodilution cardiac output and arterial-venous oxygen content difference. Lung tissue and a considerable part of the bronchial tree are metabolically active and consume oxygen. Theoretically, the difference in oxygen content between arterial and venous blood does not include the effect of pulmonary tissue and bronchial tree oxygen uptake. Thus lung VO 2 (VO 2pulm : including that from part of the bronchial tree) should be reflected as a difference between VO 2 measured by gas exchange (VO 2meas ) and that derived using Fick's principle (VO 2Fick ). Despite this sound theoretical basis, methodological limitations can make this small difference undetectable under clinical conditions. Under carefully controlled experimental conditions, however, a constant difference between VO 2meas and VO 2Fick has been demonstrated. In previous studies [1][2][3][4][5], the difference was widely variable and some of the measurements indicated greater values with reverse Fick than with indirect calorimetry. Errors and variation of gas exchange, thermodilution cardiac output, haemoglobin and saturation measurements may accumulate, and the difference between VO 2meas and VO 2Fick may be less than the cumulative methodological error. Previous studies have demonstrated that the difference between VO 2meas and VO 2Fick , taken to represent VO 2pulm , increases once pneumonia develops [1,2,[5][6][7][8][9][10]. Accordingly, VO 2pulm might be one of the indices reflecting the extent of lung injury.In the ear...