This prospective study compared two weaning modalities in COPD patients requiring mechanical ventilation (MV) for acute respiratory failure. Nineteen patients with COPD were studied when their precipitating illness was controlled. Although they satisfied the conventional bedside weaning criteria, they could not tolerate any reduction in the respirator rate below 10 cycles/min. At this time, patients were randomized into two groups receiving either synchronized intermittent mandatory ventilation (SIMV) with pressure support ventilation (PSV) (group 1) or SIMV alone (group 2). The volumetric support of ventilation (SIMV rate) was progressively decreased in both groups according to the patient's tolerance with a concurrent decrease in the barometric support of ventilation (PSV levels from 15 cm H2O to 6 cm H2O). At each step of SIMV rate, we found no difference between group 1 and group 2 in arterial blood gases, blood pressure, heart rate, airway occlusion pressure, maximal inspiratory pressure, and oxygen cost of breathing (OCB). At each step, however, group 1 patients showed significantly higher spontaneous tidal volume and lower spontaneous breathing frequency than did group 2 patients. We found a slight but not significant tendency to a shorter weaning period with than without PSV, but no difference in the weaning success. We concluded that (1) conventional weaning criteria might be inaccurate in COPD patients, (2) SIMV appeared very useful in weaning COPD patients from MV, (3) PSV marginally reduced the weaning period when added to SIMV, and (4) the OCB was not significantly improved with PSV.