The accuracy of pneumotachograph measurements during mechanical ventilation.

We hypothesized that differences between the conditions under which a pneumotachograph (PT) is calibrated and those during data collection might lead to large errors in measured flow and volume during mechanical ventilation. A Fleisch No. 1 and Fleisch No. 2 and a screen PT were connected to "ideal" tubing configurations that optimized flow characteristics, and to ventilator tubing with and without a Y-connector and endotracheal (ET) tube. Each PT was also evaluated after water had accumulated in its resistive element. Air was passed through each PT configuration, using both a continuous and a pulsatile flow pattern, and collected in a water-seal spirometer. "Measured" and "true" flow and volume were determined from the PT and the spirometer, respectively. Measured flow and volume were falsely low when the PT was adapted to ventilator tubing. Addition of a Y-connector and ET tube caused measured flow and volume to increase, and, in some cases the relationship between measured and true flow became nonlinear. Water accumulation in the PT did not lead to measurement errors. We conclude that when a PT is used during mechanical ventilation, tubing geometry must be identical during calibration and data collection, and that calibration should be performed over the entire range of relevant flows.