Two comparable groups of four Holstein calves were implanted with different left ventricle assist devices: the Centrimed centrifugal pump (CP) (Sarns Inc.) or the UTAH 85 VAD pneumatic ventricle (University of Utah). Operative procedure, inflow and outflow cannulae, monitoring, heparinization, were identical in both groups. No transfusion was ever required. The study was terminated after three days and autopsy was performed on the calves. Left ventricular unloading provided by both devices was complete (LVEDP less than 1 mmHg) and identical. LV bypassed flow rate was higher in CP (0.045 L/min/kg) than in UTAH 85 VAD (0.035 L/min/kg) but with no statistical difference. Blood trauma was comparable in the two groups. Daily blood samples did not show any significant changes from baseline values in creatinine, hematocrit, fibrinogen. Platelet loss from initial level was 30%; serum lactate dehydrogenase rose 150% with no significant difference in the two groups; plasma free hemoglobin never reached significant values. At autopsy, thrombotic deposits on cannuale and renal infarction rate were similar. CP housing had to be changed every day, whereas no technical failure was ever observed with the UTAH 85 VAD. Clinical response and blood damage of the two pumps used as LVAD were the same. Considering the CP has to be replaced every 24 h, the cost of three CP's would be comparable to one Utah UVAD-85 if polyurethane tricusp semi-lunar valves are used in the latter. Until the UTAH-85 VAD becomes commercially available, simple valveless low-cost CP are very attractive for short-term mechanical support.