Effect of CO2 concentrations on acoustic inferences of airway area

To determine the effect of gas composition on the accuracy of measurements of airway area and distance using an acoustic reflection technique, we employed glass-tube models to simulate pharyngeal (Phar-model), laryngeal (Lar-model), and tracheal (Trach-model) regions of upper and central airways. We made repeated measurements of area-distance functions using gas mixtures containing 0, 2, 4, 6, 8, and 10% CO2, 80% He, and balance O2. The actual area of the model was calculated from the roentgenographic data and compared favorably with an area measured by acoustic reflections using a gas mixture containing 0% CO2. With the different gas mixtures, calculated area was overestimated only at the highest levels of CO2, with Phar-model area increasing from (mean +/- SD) 4.66 +/- 0.03 cm2 measured with 0% CO2 to 4.93 +/- 0.05 cm2 (P less than 0.05) measured with CO2 concentration of 10%. To assess the effect of CO2 concentration on measurements of distance, we isolated two discrete points located in the Phar-model and Lar-model regions. When measurements were performed using 10% CO2 mixture, Phar-model point was shifted by 1.02 +/- 0.03 cm and Lar-model point was shifted by 2.16 +/- 0.09 cm away from the microphone compared with their axial position determined, using 0% CO2 mixture (P less than 0.05). Differences in area-distance calculations at the higher levels of CO2 did not exceed the within-run variability of the technique (10 +/- 4%). We conclude that CO2 absorbers are not required during measurements of airway area by acoustic reflections, provided CO2 concentration does not exceed 10%.