The efficacy of the transfemoral left-ventricular assist device Hemopump<sup>TM</sup> (HP; 21 Fr outer diameter) was examined in experiments with adult sheep in two different models of cardiogenic shock (tachycardia shock; ischemia shock), and during ventricular fibrillation. During tachycardia (high frequency pacing-induced; n = 14), HP assist led to a significant increase in cardiac output (from 2.2 to 2.8 liters/min), mean aortic pressure (from 47.6 to 65.6 mm Hg), and myocardial perfusion pressure (from 25.5 to 59.0 mm Hg). Simultaneously, a normalization of body oxygen-uptake (from 1.4 to 2.5 ml/min·kg), a decrease in myocardial oxygen consumption (from 6.1 to 4.8 ml/min·lOO g), and a normalization of myocardial lactate metabolism were ovserved during HP assist. During regional myocardial ischemia (PTCA balloon occlusion of the proximal LAD (3.5 min; n = 12), HP assist led to significant decrease in LV enddiastolic pressure (from 21.1 to 12.1 mm Hg), and increase in diastolic aortic pressure (from 58 to 67 mm Hg) resulting in significant increase in coronary perfusion pressure. In the early reperfusion period, myocardial release of both lactate and potassium was significantly lowered with HP assist. During ventricular fibrillation (induced by electrical stimulation; n = 9), HP flow rates decreased from 2.5 (after 10 min) to 2.1 liters/min (after 30 min). Mean aortic pressures simultaneously decreased from 64.0 to 54.6 mm Hg. Perfusion conditions were sufficient for maintenance of aerobic myocardial metabolism, but were borderline for peripheral circulation. Our hemodynamic and metabolic data demonstrate beneficial effects of cardiac assist with the Hemopump 21 Fr in both tachycardia-induced severe cardiogenic shock and during acute regional myocardial ischemia. During ventricular fibrillation, flow conditions were sufficient to maintain aerobic myocardial metabolism, but circulatory supply to the total organism was borderline.