OBJECTIVE The clinical usefulness of varying inspiratory flow waveforms during mechanical ventilation has not been adequately studied. The aim of this study was to compare the effects of three different respiratory waveforms on the pulmonary mechanics, gas exchange, and respiratory metabolism of ventilated patients with COPD. METHODS A randomized and comparative trial of consecutive patients. METHODS Medical ICUs of a 2,000-bed university hospital. METHODS Fifty-four patients with COPD were enrolled. METHODS Constant, decelerating, and sine waveforms were applied to each patient in a random order. RESULTS With tidal volume, inspiratory time, and inspiratory frequency being kept constant, the decelerating waveform produced statistically significant reductions of peak inspiratory pressure, mean airway resistance, physiologic dead space ventilation (VD/VT), PaCO(2), and symptom score. There was also a significant increase in alveolar-arterial oxygen pressure difference with the decelerating flow waveform, but there were no significant changes in mean airway pressure, arterial oxygenation, heart rate, mean BP, and other hemodynamic measurements. In addition, assessment on the work of breathing (WOB) revealed that ventilator WOB values were reduced with the decelerating waveform. Oxygen consumption and carbon dioxide output were virtually not affected by changing inspiratory flow waveforms. Except for VD/VT, the effects of constant square and sine waveforms were similar to each other and could not be separated statistically. CONCLUSIONS The most favorable flow pattern for ventilated patients with COPD appeared to be the decelerating waveform. There are possibilities for the improvement of ventilation in these patients by selecting an appropriate inspiratory flow.