Accuracy of the respiratory inductive plethysmograph during loaded breathing

Indirect methods of measuring ventilation, such as the respiratory inductive plethysmograph (RIP), operate on the assumption that the respiratory system possesses two degrees of freedom of motion: the rib cage and abdomen. Accurate measurements have been obtained in many patients with pulmonary disease who possess additional degrees of freedom. Since calibration and validation of the RIP was carried out during quiet breathing in these patients, the amount of asynchronous or paradoxic breathing was presumably similar during the calibration and validation runs. Conversely, accuracy might be lost if following the initial calibration procedure the magnitude of chest wall distortion increased during subsequent validation runs. We calibrated the RIP during quiet breathing and examined its accuracy while subsequently breathing against resistive loads that required the generation of 20-80% of the subject's maximum inspiratory mouth pressure (Pm(max)). We compared the relative accuracy of three commonly employed calibration methods: isovolume technique, least-squares technique, and single position loop-area technique. Up to 60% of Pm(max), 89% of the RIP values with the least-squares technique were within ± 10% of simultaneous spirometric (SP) measurements and 100% were within ± 20% of SP, compared with 63 and 91%, respectively, for the loop-area technique and 19 and 54%, respectively, for the isovolume technique. At 70 and 80% of Pm(max) accuracy deteriorated. Accuracy of respiratory timing was judged in terms of fractional inspiratory time (TI/TT). Up to 60% of Pm(max), 69 and 96% of TI/TT values with the least-squares technique were within ± 10 and 20% of simultaneous SP measurements, respectively, whereas the comparable values for the loop-area technique were 69 and 89%, respectively, and 63 and 89%, respectively, for the isovolume technique. The results indicate that RIP provides reasonably accurate measurements even with high inspiratory loads.