Thoracocardiographic-Derived Left Ventricular Systolic Time Intervals

Abstract

Thoracocardiography noninvasively estimates left ventricular performance by recording ventricular volume curves with inductive plethysmography. We studied timing of these curves to evaluate their potential to accurately track systolic time intervals in comparison with standard methods. Thoracocardiographic left ventricular volume curves, carotid pressure pulses determined by applanation tonometry, the phonocardiogram and ECG were recorded simultaneously in ten normal subjects at various body positions achieved with a tilt table. An equation was derived to predict preejection period from onset of ejection in thoracocardiographic curves. Ventricular ejection time was calculated as total electromechanical systole obtained by phonocardiography minus preejection period. The equation was validated prospectively in 31 measurements in critically ill patients. In normal subjects, the interval ECG Q wave to ejection onset in thoracocardiographic curves correlated well with preejection period from applanation tonometry and phonocardiography (r = 0.92; standard error of estimate (SEE), 8 ms; p < 0.001). Thoracocardiographic curves showed a delay that varied with body position according to the regression equation: delay = 40 ms + 10 x sine (tilt angle) (where r = 0.62; SEE, 7 ms; p < 0.001). Application of this equation in the prospective study in patients revealed close agreement in systolic time intervals from thoracocardiography and simultaneous applanation tonometry plus phonocardiography, respectively. The mean difference +/- SD between methods in preejection periods was 3 +/- 7 ms and in the ratios of preejection period to left ventricular ejection time, 0.02 +/- 0.05. Trends of changes in systolic time intervals were identical for the two methods. We conclude that thoracocardiography combined with phonocardiography provides accurate systolic time intervals when corrected for a position-dependent delay of its waveforms.