Modeling Partial and Total Support of Left Ventricular Assist Device for Discrete Hemodynamic Control Framework.

Abstract

This study develops a model-based control framework integrating a left ventricular assist device (LVAD) with drug therapy to regulate multidimensional hemodynamics (mean arterial pressure, cardiac output, and left and right atrial pressures) while minimizing myocardial oxygen consumption ($MVO\_{2}$) in acute heart afailure (AHF). By modeling the hemodynamic and $MVO\_{2}$ effects of partial and total LVAD support, the framework enables combined drugs-LVAD therapy for severe AHF cases where drug therapy alone is not effective. A previously developed hemodynamic control framework was extended to incorporate LVAD support. Using circulatory equilibrium models under LVAD support, an analytical $MVO\_{2}$ model encompassing both LVAD modes was derived from the relationship between $MVO\_{2}$ and the ventricular pressure-volume area. An optimal controller explicitly accounts for transitions between partial and total LVAD support. In our cardiovascular simulator, we validated the $MVO_{2}$ model and evaluated the framework's multidimensional hemodynamic regulation and $MVO\_{2}$ minimization capabilities. The $MVO\_{2}$ model accurately predicted $MVO\_{2}$ across the two LVAD modes. In a severe AHF scenario, drug therapy or LVAD support alone failed, whereas the combined drugs-LVAD controller succeeded in achieving both hemodynamic regulation and $MVO\_{2}$ minimization. Compared with drug therapy alone, drugs-LVAD control using the proposed $MVO\_{2}$ model halved $MVO\_{2}$ while achieving multidimensional regulation. Our drugs-LVAD framework may enable multidimensional hemodynamic regulation while reducing $MVO\_{2}$ in severe AHF. Modeling both partial and total LVAD support may be essential for hemodynamic management. This work provides a theoretical basis for integrated drugs-LVAD control to achieve target hemodynamics while reducing cardiac workload.