Respiratory inductance plethysmography in healthy infants: a comparison of three calibration methods

Abstract

Respiratory inductance plethysmography (RIP) measures respiration from body surface movements. Various techniques have been proposed for calibration in order that RIP may be used quantitatively. These include calculation of the proportionality constant of ribcage to abdominal volume change (K). The aims of this study were to 1) establish whether a fixed value of K could be used for calibration, and 2) compare this technique with multiple linear regression (MLR) and qualitative diagnostic calibration (QDC) in normal healthy infants. Recordings of pneumotachograph (PNT) flow and RIP were made during quiet (QS) and active sleep (AS) in 12 infants. The first 5 min in a sleep state were used to calculate calibration factors, which were applied to subsequent validation data. The absolute percentage error between RIP and PNT tidal volumes was calculated. The percentage error was similar over a wide range of K during QS. However, K became more critical when breathing was out of phase. A standard for K of 0.5 was chosen. There was good agreement between calibration methods during QS and AS. In the first minute following calibration during QS, the mean absolute errors were 3.5, 4.1 and 5.3% for MLR, QDC and fixed K respectively. The equivalent errors in AS were 11.5, 13.1 and 13.7% respectively. The simple fixed ratio method can be used to measure tidal volume with similar accuracy to multiple linear regression and qualitative diagnostic calibration in healthy unsedated sleeping infants, although it remains to be validated in other groups of infants, such as those with respiratory disease.