To identify factors responsible for energy consumption, a retrospective investigation of the in-vivo performance of the 100 ml electric motor-driven left ventricular assist device (ELVAD), and the e-motor 100 ml total artificial heart, was undertaken. Multivariate regression analysis of the device parameters demonstrated that device flow, and estimated outlet pressure, were the most significant independent variables for predicting changes in motor power. Weighted least-square curvefit, using the product of these two variables, showed that changes in energy consumption can be well predicted for the ventricular assist device (r2 = 0.732). However, by applying the same model to the total artificial heart (ETAH), less favorable results were achieved (r2= 0.422). In this model, device flow seemed to be more important in predicting energy consumption for the ETAH compared to the ELVAD. Therefore, changing the model by using flow to the third order significantly improved the fit (r2 = 0.6706) for the ETAH, and could compensate in part for the greater variability of the values and increased number of outliers in this group.
All Science Journal Classification (ASJC) codes