Calibration of a Closed-Loop Model of Porcine Aortic Hemodynamics During Hemorrhage
Please login to view abstract download link
The integration of experimental data with computational models is essential for developing physiologically interpretable and clinically relevant biomechanical simulations, particularly in settings where direct experimentation is limited. Hemorrhagic shock represents one such challenge where ethical and practical constraints restrict systematic investigation in humans. As a result, in-silico models designed and calibrated using large-animal experimental data are increasingly employed to better understand the underlying physiology and to support the development and evaluation of treatment strategies. In this work, we present a closed-loop, zero-dimensional (0D) cardiovascular modeling framework calibrated using experimental data from 43 porcine subjects undergoing controlled hemorrhage (10\%, 20\%, and 30\% total blood volume loss). Continuous measurements of left ventricular pressure–volume dynamics, arterial pressures, and branch-specific flows were used as calibration targets through a structured data-processing pipeline to generate representative hemodynamic targets at discrete time points throughout hemorrhage. The 0D model consisted of a lumped parameter heart model with time-varying elastance function, a compartmental aortic model, Windkessel representations of peripheral vascular beds, and a compliant venous reservoir. Hemorrhage was explicitly modeled as progressive blood-volume loss from the venous compartment. Model calibration was then performed by optimizing a restricted set of vascular, cardiac, and venous parameters within physiological bounds to achieve agreement between simulated outputs and experimentally derived targets, including global metrics (mean arterial pressure and cardiac output) and regional flow distributions. The calibrated model reproduced experimental trends across hemorrhage severity and time, capturing not only mean values but also waveform morphology and pressure–volume loop dynamics. The inferred parameter trajectories exhibited physiologically consistent adaptations, including increased regional resistance, reduced arterial compliance, and declining venous unstressed volume. This study demonstrates how data-processing and inverse calibration strategies can bridge experimental measurements and reduced-order biomechanical models, providing a mechanistically interpretable platform for studying cardiovascular adaptation during hemorrhage.
