Characterization of ventricular dynamics in dogs using computerized variables and their statistical treatment

To improve the analysis of mechanical events of the cardiac cycle and the selection of indexes of cardiac function, it is necessary to obtain the most probable time course of a given variable or of the shape of a loop. With this objective, a special program has been devised to obtain mean time tracings of left ventricular pressure (LVP) and posterior wall thickness (PWTH). The time scale can be normalized or beats with small variation in duration can be chosen. Mean loops are obtained in dogs by relating mean LVP and PWTH values. The distribution of pairs of individual values is assumed as normal bivariate with respect to its corresponding mean value, yielding confidence and tolerance ellipses. Results show that for a 95% confidence limit the loop is open in normal conditions. We postulate that this statistical treatment, by giving a more precise course of a mechanical signal allows a more accurate comparison between different experimental or clinical situations in the whole cardiac cycle, as well as the development of cardiac indexes.