To define the importance of hemodynamic performance and global tissue oxygenation in determining weaning outcome, we recorded mixed venous oxygen saturation (SvO2) continuously in eight ventilator-supported patients who failed a trial of spontaneous breathing and 11 patients who tolerated a trial and were successfully extubated. Immediately before the weaning trial, SvO2 was not statistically different in the two groups (p = 0.28). On discontinuation of the ventilator, SvO2 fell progressively in the failure group (p < 0.01), whereas it did not change in the success group. During the trial of spontaneous breathing, O2 demand was similar in the two groups, but it differed in the manner with which it was met. The success group demonstrated an increase in cardiac index (p < 0.05) and O2 transport (p < 0.02). The failure group did not increase O2 transport, partly because of elevations in right- and left-ventricular afterload, but, instead, increased O2 extraction ratio (p < 0.02) with a consequent fall in SvO2. In turn, the low SvO2 combined with greater venous admixture (p < 0.0006) led to rapid arterial desaturation (p < 0.006) and a relative decrease in O2 being supplied to the tissues. In conclusion, ventilator-supported patients who failed a trial of spontaneous breathing developed a progressive decrease in SvO2 caused by the combination of a relative decrease in convective O2 transport and an increase in O2 extraction by the tissues.