Cardiogenic shock (CS) is a life-threatening condition characterized by severe ventricular dysfunction, reduced cardiac output, and impaired tissue perfusion, with mortality rates of 50–80%. Its pathophysiology is complex and highly heterogeneous, posing major challenges to standardized treatment strategies and highlighting the need for individualized hemodynamic management. Computational physiological models provide a mechanistic framework to support personalized decision-making in critical care.
This study presents the development of a comprehensive mathematical framework to simulate targeted clinically relevant biosignals under CS conditions. We developed a zero-dimensional global closed-loop cardiovascular model describing heart chamber mechanics, systemic and pulmonary circulation, respiratory dynamics, gas exchange and transport, blood pH regulation, and short-term control mechanisms. Clinically relevant outputs include heart and respiratory rates, stroke volume, arterial, peripheral and venous pressures and O2-CO2 partial pressures. All 312 input parameters were investigated through local sensitivity analysis, followed by a computationally efficient global sensitivity analysis performed on parameter classes. Parameter identifiability and collinearity were then assessed on the most influential parameters to select an optimal, reduced subset. Finally, to simulate targeted CS scenarios, this subset was calibrated using the Covariance Matrix Adaptation Evolutionary Strategy.
Transient simulations from baseline parametrization to the CS ones showed progressive hemodynamic deterioration consistent with the targeted CS scenarios, i.e. a marked reduction in cardiac output, O2 venous concentration, mean arterial pressure, together with increased pulmonary capillary wedge pressure, vascular resistances and heart rate. The proposed framework can simulate targeted, clinically relevant biosignals under CS conditions and highlights the key mechanisms driving shock dynamics, supporting its potential use for in silico experimentation.
Future developments will include the integration of a pharmacokinetic–pharmacodynamic module to simulate the effects of vasoactive agents, enabling virtual testing of therapeutic interventions.