INTRODUCTION: Heart failure is a major and growing public health concern. Although heart failure has been considered an inexorable and progressive disorder, emerging evidence suggests that some patients may have reversible left ventricular dysfunction. Indeed, recent reports have documented the potential for myocardial recovery in humans in response to prolonged mechanical circulatory support with a left ventricular assist device (LVAD). However, myocardial recovery remains uncommon, and a strategy of unloading the failing left ventricle with a continuous-flow (non-pulsatile) LVAD has not been specifically developed to promote favorable myocardial remodeling. As a preliminary investigation, we developed a bovine model of chronic, ischemic heart failure and quantified the effects of different levels of support with a continuous-flow LVAD on myocardial mechanoenergetics. METHODS: Normal cows (n=8) and cows with chronic, ischemic heart failure (n=9) were studied. To induce heart failure, 90 µm micro spheres were percutaneously injected into the left main coronary artery. Heart failure developed over 60 days. In an acute surgery, a continuous-flow LVAD was implanted and operated at Low Partial Support (~1.5 L/min support, aortic valve opening every beat), High Partial Support (~3 L/min support, aortic valve opening every beat) and Full Support (~5 L/min, aortic valve maintained closed, left ventricle maximally unloaded). Cardiac and systemic arterial hemodynamics were measured with flow probes and pressure catheters. Myocardial blood flow was mapped with 15 µm fluorescent-labeled microspheres. After termination, molecular and histological markers of heart failure were quantified. RESUL TS: In normal animals, increasing levels of non-pulsatile support deranged the profile of cardiac and arterial hemodynamics. As cardiac workload decreased, myocardial vascular resistance increased, and myocardial blood flow decreased. The ratio between blood supply and demand did not change and indicated appropriate coronary autoregulation and the presence of an intact coronary reserve. Animals with chronic, ischemic heart failure exhibited hallmark signs of severe left ventricular systolic dysfunction that included a 50% reduction in ejection fraction, left ventricular dilatation, decreased cardiac output and arterial pressures, decreased end-organ blood flow, severe myocardial fibrosis, myocyte hypertrophy, and increased myocardial apoptosis. In animals with chronic heart failure, increasing levels of non-pulsatile support similarly deranged the profile of cardiac and arterial hemodynamics. As cardiac workload decreased, myocardial vascular resistance increased. However myocardial blood flow did not change and indicated a lack of a coronary reserve. Importantly, during full but not partial support, the ratio between blood supply and demand improved significantly to levels observed in normal control animals. CONCLUSIONS: After the implantation of an LVAD, full but not partial support of the failing left ventricle with an LVAD normalizes the myocardial blood supply/demand relationship. In the immediate postoperative period, the left ventricle should be completely unloaded. Chronic studies are necessary to determine whether a transition to partial support may prevent myocardial atrophy and fibrosis that is seen with prolonged full support. Our bovine model of chronic, ischemic heart failure is appropriate for such a study.