Background:The aim of this study was to test the hypothesis that, during weaning from mechanical ventilation, when the pressure support level is reduced, oxygen consumption increases more in patients unable to sustain the decrease in ventilatory assistance (weaning failure). Methods: Patients judged eligible for weaning were enrolled. Starting from 20 cm H 2 O, pressure support was decreased in 4-cm H 2 O steps, lasting 10 min each, until 0 cm H 2 O; this level was kept for 1 h. The average oxygen consumption from the last 3 min of each step, along with other ventilatory variables, was measured by indirect calorimetry (M-CAiOVX "metabolic module," Engstrom Carestation; GE Healthcare, Madison, WI) and recorded. Patients were defined as belonging to the failure group if, at any moment, they developed signs of respiratory distress according to standard criteria, or to the success group otherwise. Results: Twenty-eight patients were studied. In most patients, the minimum oxygen consumption was not recorded at the highest pressure support applied. Sixteen patients were able to complete the weaning trial successfully, whereas 12 failed it; the success group had a minimum oxygen consumption lower than failure group (mean Ϯ SD: 174 Ϯ 44 vs. 215 Ϯ 53 ml/min, P Ͻ 0.05). Moreover, although respiratory drive (assessed by P0.1) increased more in the failure group, this group had a lower increase in oxygen consumption, contradicting our hypothesis. Conclusions: Patients failing a decremental pressure support trial, in comparison with those who succeed, had an higher baseline oxygen consumption and were not able to increase their oxygen consumption in response to an increased demand.