Effect of pressure-flow relationship of centrifugal pump on in vivo hemodynamics: a consideration for design.

We have been developing centrifugal pumps for an implantable left ventricular assist device. We manufactured 2 prototype centrifugal pumps (PI, PII). These two have similar designs except for the PII having a volute casing and a large output port. To determine the differences in the hydraulic characteristics between the PI and PII, we carried out in vitro and in vivo experiments. In vitro study showed that the PII had a shallower H-Q curve than that of the PI, and the PII required a pump speed faster than the PI for the same flow rate and pressure head. On the other hand, in vivo study showed that the PII demonstrated a flow pulsatility greater than that of the PI at 1,900 rpm and 8 L/min although no significant change was observed at low pump speeds (< or = 1,500 rpm). This greater pulsatility consisted of a large discharge according to the small differential pressure during the systolic phase and a small discharge according to the large differential pressure during the diastolic phase. In contrast, the PI, having the steeper H-Q curve, showed a small discharge in the systolic phase and a large discharge in the diastolic phase. These results showed that pulsatility synchronized with the native heart beating depended on the slope of the H-Q curve. As a result, the slope of the H-Q curve is important to determine the component of pulsatility synchronized with native cardiac output. Regarding the slope of the H-Q curve, a pump having a volute casing and a large outlet port demonstrates a shallow slope in the H-Q curve. In conclusion, we suggest that a centrifugal pump for use in left ventricular aortic bypass should be designed considering the effect on the native heart pulsatility.