A SIMPLE METHOD TO EVALUATE UNLOADING PROPERTIES OF A CARDIAC ASSIST DEVICE IN VIVO

Abstract

In vivo evaluation of a cardiac assist device (CAD) is usually performed in large animal models of heart failure (HF). These analysis are arduous, expensive and require prior induction of HF. We propose a simple method to quickly assess the ability of CAD to unload the left ventricle (LV). A left thoracotomy was performed in 3 anesthetized, ventilated calves (71.0±2.5kg) monitored with ECG, Swan-Ganz catheter, femoral and carotid arterial pressure (APfcm and APcar) and central venous pressure (CVP). After full heparinization, a CAD (AB-180 centrifugal pump) was connected to the left atrium and the descending aorta. A jugular cannula was connected to a cardiotomy reservoir used for blood transfusion/storage and CVP adjustment. Cross-clamping the descending aorta above the CAD connection separated 2 distinct perfusion territories, one depending on the LV and the other perfused by the CAD. Alternate periods (1 min/3 min) of aortic cross-clamping/unclamping allowed repeated APfem, APcar and blood flow measurements, while avoiding coronary hypoperfusion. Prior to aortic cross-clamping, APfem and APcar was 60.7±1.2 and 70.0±2.6 mmHg, and cardiac output was 3.4±0.3 and 6.1±0.4 l/min respectively for a CVP of 4 and 10 mmHg. Under cross-clamping and pump assistance (4,500 rpm), a 58.9±2.8% (p<0.001) and 22.1±10.7% (p=ns) increase in APfem and a 38.0±7.9% (p<0.01) and 6.1±8.0% (p=ns) decrease in APcar were observed with a CVP of 4 and 10 mmHg respectively. At a low speed (3,000 rpm), carotid flow was decreased to 36.5±4.3% (p<0.005) and 45.6±7.2% (p<0.01) of the basal value for a CVP of 4 and 10 mmHg respectively. Blood flow was abolished in the carotid at 4'500 rpm. This experimental model allows efficient assessment of unloading performance by new CADs. In our example, the LV contribution to systemic perfusion was significantly decreased when the AB-180 CAD was employed.