High-frequency percussive ventilation: pneumotachograph validation and tidal volume analysis

Abstract

Introduction: High-frequency percussive ventilation (HFPV) is an increasingly used mode of mechanical ventilation, for which there is no proven real-time means of measuring delivered tidal volume (V(T)).

Objective: To validate a pneumotachograph for HFPV and then exploit flow-sensor data to describe the behavior of both low-frequency and high-frequency breaths.

Methods: Sensor performance was gauged during changes in high-frequency (4-12 Hz) and low-frequency rate and ratio, mean airway pressure, oxygen concentration, heated or heated-humidified gas flow, and endotracheal tube diameter. Glass bottle (adiabatic V(T)) and test lung (adiabatically derived low-frequency V(T)) based adiabatic conditions provided both an initial source for analog-signal calibration and an accepted standard comparator to flow-sensor measurement of high-frequency and low-frequency (flow-sensor-derived) V(T)), respectively.

Results: Pneumotachography proved accurate and precise over an array of tested settings and conditions when analyzing both high-frequency (difference between mean +/- SD high-frequency V(T) and adiabatic V(T) was -0.2 +/- 1.8%, 95% confidence interval -0.5 to 0.9%) and low-frequency breaths (mean +/- SD difference between flow-sensor-derived low-frequency V(T) and adiabatically derived low-frequency V(T) was 0.6 +/- 2.4%, 95% confidence interval 0.1-1.1%). High-frequency V(T) and frequency exhibited an exponential relationship. During HFPV, flow-sensor-derived low-frequency V(T) had a mean +/- SD of 1,337 +/- 700 mL, 95% confidence interval 1,175-1,499 mL.

Conclusions: Readily available pneumotachography provided accurate measurements of low-frequency and high-frequency V(T) during HFPV. In the setting of acute lung injury, typical HFPV settings may deliver injurious V(T).