Control strategy for maintaining physiological perfusion with rotary blood pumps.

We present arguments and simulation results in favor of a novel strategy for control of rotary blood pumps. We suggest that physiological perfusion is achieved when the blood pump is controlled to maintain an average reference differential pressure. In the case of rotary left ventricular assist devices, our simulations show that maintaining a constant average pressure difference between the left ventricle and aorta results in physiological perfusion over a wide range of physical activities and clinical cardiac conditions. We simulated rest, light, and strenuous exercise conditions, corresponding to cardiac demands of 4.92, 7.98, and 14.62 L/min, respectively. For different exercise levels, the clinical conditions ranged from normal to failing to asystolic heart. By maintaining a constant pressure difference of 75 mm Hg between the left ventricle and aorta, with either an axial or a centrifugal blood pump, a total cardiac output close to the physiological cardiac demand was achieved, irrespective of the heart condition. The simulations of the transitions between different levels of exercise indicate that with the same reference differential pressure, the proposed approach leads to rapid adaptation of the total cardiac output to physiological levels, while avoiding suction. Comparison with the traditional control strategy of maintaining a reference rotational speed (rpm) of the pump indicates that though the traditional approach has some degree of adaptability, it is only adequate over a narrow range of cardiac demand and clinical conditions of the patient. Our results indicate that the proposed approach is superior to the alternatives in providing an adequate and autonomous adaptation of the total cardiac output over a broad range of exercise conditions (expected when an assist device is used as a destination therapy) and clinical statuses of the native heart (such as further deterioration or recovery of cardiac function), while having the potential to improve the quality of life of patients by reducing the need for monitoring and frequent human intervention. The proposed approach can be clinically implemented using simple controllers, and requires the implantation of two pressure sensors, or estimation of the pressure difference based on other available measurements.