A new electric method for non-invasive continuous monitoring of stroke volume and ventricular volume-time curves

BACKGROUND In this paper a new non-invasive, operator-free, continuous ventricular stroke volume monitoring device (Hemodynamic Cardiac Profiler, HCP) is presented, that measures the average stroke volume (SV) for each period of 20 seconds, as well as ventricular volume-time curves for each cardiac cycle, using a new electric method (Ventricular Field Recognition) with six independent electrode pairs distributed over the frontal thoracic skin. In contrast to existing non-invasive electric methods, our method does not use the algorithms of impedance or bioreactance cardiography. Instead, our method is based on specific 2D spatial patterns on the thoracic skin, representing the distribution, over the thorax, of changes in the applied current field caused by cardiac volume changes during the cardiac cycle. Since total heart volume variation during the cardiac cycle is a poor indicator for ventricular stroke volume, our HCP separates atrial filling effects from ventricular filling effects, and retrieves the volume changes of only the ventricles. METHODS ex-vivo experiments on a post-mortem human heart have been performed to measure the effects of increasing the blood volume inside the ventricles in isolation, leaving the atrial volume invariant (which can not be done in-vivo). These effects have been measured as a specific 2D pattern of voltage changes on the thoracic skin. Furthermore, a working prototype of the HCP has been developed that uses these ex-vivo results in an algorithm to decompose voltage changes, that were measured in-vivo by the HCP on the thoracic skin of a human volunteer, into an atrial component and a ventricular component, in almost real-time (with a delay of maximally 39 seconds). The HCP prototype has been tested in-vivo on 7 human volunteers, using G-suit inflation and deflation to provoke stroke volume changes, and LVot Doppler as a reference technique. RESULTS The ex-vivo measurements showed that ventricular filling caused a pattern over the thorax quite distinct from that of atrial filling. The in-vivo tests of the HCP with LVot Doppler resulted in a Pearson's correlation of R = 0.892, and Bland-Altman plotting of SV yielded a mean bias of -1.6 ml and 2SD =14.8 ml. CONCLUSIONS The results indicate that the HCP was able to track the changes in ventricular stroke volume reliably. Furthermore, the HCP produced ventricular volume-time curves that were consistent with the literature, and may be a diagnostic tool as well.