Abstract Background Percutaneous left ventricular assist device (pLVAD) can unload the left ventricle (LV). However, stable operation of pLVAD is challenging under hemodynamically unstable conditions. We have developed an automated pLVAD system (Auto-VAD) that induces stable LV unloading without suction for severe cardiogenic shock (CS) conditions. This system automatically regulates pump speed to maintain the LV pressure at a desired LV unloading level, applying a dedicated negative feedback control mechanism. Purpose This study aimed to evaluate the control accuracy and clinical benefits of Auto-VAD through a large animal experiment using goats. Methods We used 16 goats (54.9±9.3 kg), inserted a pLVAD through the left carotid artery, and placed it in the LV. We set the LV systolic pressure (LVSP) = 40 mmHg as the target LV unloading level. In protocol 1 (n=4), we assessed the operation of Auto-VAD in goats with severe CS for 24 hours under general anesthesia. We induced CS by micro-embolization of the left coronary arteries and recorded LV pressure simultaneously. In protocol 2, we compared LVSP maintenance between Auto-VAD and expert manual pump control, selecting P1-P8, in CS goats over a one-hour period. We defined LVSP within 32 to 48 mmHg as the target range. We also validated two conditions of pLVAD: isolated pLVAD (n=6) and combination treatment with venoarterial extracorporeal membrane oxygenation (VA-ECMO) (ECPELLA: n=6). Results In both protocols, coronary micro-embolization succeeded in inducing CS with an LV ejection fraction below 35% and unstable hemodynamics. As shown in Figure 1, Auto-VAD maintained LVSP at 41.6±1.8 mmHg for 24 hours in protocol 1. The percentage of LVSP within the target range was 91.4±9.2%. Figure 2 demonstrates the hemodynamic changes during an hour in each condition for protocol 2. In the isolated pLVAD conditions, Auto-VAD achieved a higher percentage of LVSP within the target range (86.4±16.8%), similar to expert control (74.8±23.0%). Additionally, Auto-VAD significantly reduced the number of drug modifications compared to expert control (3.3±3.3 vs. 7.8±2.6 times, p=0.03). The same trends were observed in the ECPELLA conditions. Auto-VAD achieved significantly more accurate control of LVSP compared to experts (94.3±4.8 vs. 72.7±22.3 %, p=0.04). Auto-VAD also attenuated LV suction across both protocols. Conclusion Auto-VAD provided stable LV unloading with minimal management effort in severe CS goats with or without VA-ECMO support. The control accuracy of Auto-VAD was comparable to or superior to that of experts. Auto-VAD represents the future of intensive care for severe CS requiring mechanical circulatory support.
Morita et al. (2025) studied this question.