4.5 Cardiac Output Estimation from Beat-to-Beat Radial Pressure and Pulse Wave Velocity: A Model-Based Study

Background Cardiac output (CO) monitoring remains a salient challenge. The state-of-the-art is based on generalized transfer functions and parameter estimations from pooled clinical data, which do not necessarily reflect the state of the cardiovascular system in a patient-specific way. Here, we introduce a patient-specific approach to estimate CO from sequential radial pressure measurements and carotid-to-femoral pulse wave velocity (cf-PWV). We do so by effectively tuning a generalized mathematical model of the cardiovascular system (1). Methods Initially, the method uses the measured cf-PWV to estimate arterial compliance. We consequently determine aortic flow from beat-to-beat radial pressure measurements based on the assumption of a fairly constant total peripheral resistance (TPR) over several heartbeats (2). Concretely, we developed an algorithm which, starting from an initial flow, employs a gradient-based optimization process (3) to calculate TPR at each beat. This TPR value is subsequently used as input for a new flow approximation. The process is repeated until convergence is reached. To assess the accuracy of the method, we implemented the algorithm on in vivo anonymized data from n=15 subjects (4) and compared the method-derived CO to the measured ones. Results Our results demonstrated that precise estimates of CO were yielded, with a RMSE of 0.38 L/min ( Fig. 1 ). Small variance in arterial compliance tuning did not show to significantly undermine the accuracy of the CO predictions. Conclusions: The in vivo validation allows us to conclude that our novel method accurately estimates CO in a patient-specific way. Therefore, the technique may potentially be employed for noninvasive CO monitoring in the clinical setting. Fig. 1 Scattergram of model-derived CO estimates vs. in vivo CO data