Immediate placement of a Hemopump (HP) ventricular assist device was undertaken in nine patients (seven men, two women) after other attempts at weaning from cardiopulmonary bypass (CPB) after coronary bypass surgery had failed. All nine patients (100%) were successfully weaned from CPB, and six (63.3%) gradually improved enough to permit removal of HP support. Five (83.3%) of the six who were weaned from CPB survived beyond hospital discharge. HP support was evaluated in terms of vital organ function, incidence of complications, and clinical outcome. In both survivors (S) and nonsurvivors (NS), serial hemodynamic measurements were taken. Although there were few differences in hemodynamic parameters between groups at 4 hr, by 24 hr the S group had markedly improved cardiac index, Glasgow Coma Scale, urinary output, and pulse pressure and required far less inotropic support than did the NS group. All four patients who required high dose inotropic agents to maintain acceptable end-organ perfusion on HP support died; three were unable to tolerate weaning from the HP, and all died within 72 hr of surgery because of ineffective myocardial recovery. None of the survivors required additional early inotropic augmentation. Renal perfusion appeared to be well maintained, even with mean arterial pressures below 60 mmHg during HP support. There were no episodes of hemolysis, infection, or limb ischemia in either group. Thus, the Hemopump provided excellent circulatory support for those patients who could not be weaned from CPB by traditional methods.