Input respiratory impedance measured by head generator in preschool children.

The upper airway wall motion represents a serious problem when measuring the input impedance of the respiratory system (Zrs) by the forced oscillation technique, particularly in young children. To minimize this error, it has been proposed to vary transrespiratory pressure around the head rather than directly at the mouth, using the head generator technique (HGT). The aim of this study was to collect normative data in preschool children in whom the technique may prove most useful. Zrs was measured using HGT and 4-32-Hz pseudorandom noise input in 127 healthy children. Age ranged from 2. 8 to 7.4 years and height (H) from 0.89 to 1.29 m. The fast Fourier transforms of pressure and flow allowed us to calculate respiratory system resistance (Rrs(f)) and reactance (Xrs(f)) at each frequency (f). Resonant frequency (fn), respiratory system inertance (Irs), and compliance (Crs) were derived from the Xrs(f) data. The technique was accepted by more than 95% of the children. A coherence function </=0.95 required us to reject Zrs values in the lower frequency spectrum. These missing data did not allow reliable estimation of fn, Irs, and Crs in 44% of children. Regression equations of Zrs on growth parameters were obtained. H was the only significant predictive variable. The decrease of Rrs with growth was best described by a multiplicative model [e.g., ln (Rrs(20)) = 2. 5301 - 2.3837. ln (H)]. The slope of the regression lines of Xrs on H was dependent on f: positive between 4 and 16 Hz, negative at high frequency (e.g., 32 Hz), and nonsignificant in the intermediate range (e.g., 20 Hz). Irs and fn decreased significantly and Crs increased as H increased. The current study provides reference values for input impedance of the respiratory system in preschool children with upper airway artifacts minimized. HGT should be particularly helpful at or around 20 Hz, where the signal to noise ratio is larger than at lower frequencies but where upper airway artifacts may significantly corrupt measurements with the standard generator.