A novel analysis of fetal breathing

We have analyzed a variety of approaches in assessing fetal breathing parameters (VT, TI, Ttot, VT/TI, VI) in eight fetal sheep during a control period and during stimulation with 6 and 9% CO2. By using conventional analysis of data blocks varying from 100 to 1500 breaths, several different conclusions could be reached regarding the respiratory response to hypercapnia: stimulation, depression, or no change in all parameters studied. A new analysis based on piecewise linear regression used as a data grouping technique indicated that a simple mean +/- SD of the individual parameters was an inappropriate description of normal or stimulated fetal breathing. Based on tests for homogeneity of regressions of VT on TI for a completely random design, it is concluded that an estimate of fetal respiratory drive is only described adequately by two to four regression regimes. These regimes, estimated from the regression technique, could be combined to give a weighted mean value based on the proportion of time they were present. Using this new approach and an analysis of variance, we found (i) that frequency and VI were similar between animals during control and hypercapnia, (ii) that breathing frequency decreased during hypercapnia, and (iii) a positive relationship between VT and TI.