Calibration of a Closed-loop Model of Porcine Aortic Hemodynamics during Hemorrhage

Abstract

Uncontrolled hemorrhage remains a leading cause of traumatic death, driven by rapid physiological deterioration that is often difficult to detect during the compensated phase. While large-animal models provide critical insights into these dynamics, they are resource-intensive, motivating the need for efficient computational frameworks that can mechanistically interpret cardiovascular responses. We developed and calibrated a closed-loop zero-dimensional (0D) lumped-parameter model (LPM) using hemodynamic data from 43 anesthetized swine subjected to controlled hemorrhage (10%, 20%, or 30% of total blood volume). The computational framework, incorporates a dynamic heart model with a custom time-varying elastance function, a multi-compartment aorta, and distal Windkessel models representing vascular beds. The model was calibrated at discrete time snapshots throughout the 30-minute hemorrhage protocol to reproduce group-averaged experimental waveforms for aortic flow, regional organ flows, and systemic pressures. The calibrated model successfully reproduced experimental hemodynamic targets and waveform morphology across all hemorrhage severities. Analysis of the calibrated parameters revealed distinct physiological mechanisms driving hemodynamic adaptation during hemorrhage: a preferential increase in renal resistance compared to carotid resistance, indicating flow redistribution to vital organs, and a progressive mobilization of venous unstressed volume to sustain cardiac filling. Furthermore, the model captured the distinct shift toward preload limitation state for 30% hemorrhage group. This study establishes a physiologically interpretable in-silico framework capable of predicting both global and regional hemodynamic responses to acute blood loss, providing a validated foundation for future applications in trauma care and resuscitation modeling.