Measuring time-varying respiratory mechanics during anesthesia.

We introduce a simple measurement technique which can track sudden and/or transient changes in respiratory mechanics even in unsteady respiration. Respiratory signals are segmented into single-breath signals. Breaths contaminated with noise produced by unsteadiness are discarded manually. A linear single-compartment model is fit to the data of "noise-free" single breaths, estimating its model parameters. respiratory mechanics is interpreted on the basis of breath-to-breath changes in the parameter estimates. The technique was tested in anesthetized subjects with unstable respiratory conditions. It was shown that the technique was noise insensitive and that the estimated model parameters well reflected the dynamic changes in respiratory mechanics. Although our method provides limited information compared with more sophisticated measurements, it may be useful when respiration is unstable, as frequently seen during light anesthesia or respiratory care.