Automatic calibration of the inlet pressure sensor for the implantable continuous-flow ventricular assist device.

Significant progress in the development of implantable ventricular assist devices using continuous-flow blood pumps has been made recently. However, a control method has not been established. The blood pressure in the inflow cannula (inlet pressure) is one of the candidates for performing an adequate control. This could also provide important information about ventricle sucking. However, no calibration method for an inlet pressure sensor exists. In this study, an automatic calibration algorithm of the inlet pressure sensor from the pressure waveform at the condition of ventricle sucking was proposed. The calibration algorithm was constructed based on the consideration that intrathoracic pressure could be substituted for atmospheric pressure because the lung is open to air. We assumed that the inlet pressure at the releasing point of the sucking would represent the intrathoracic pressure, because the atrial pressure would be low owing to the sucking condition. A special mock circulation system that can reproduce ventricle sucking was developed to validate the calibration algorithm. The calibration algorithm worked well with a maximum SD of 2.1 mmHg for 3-min measurement in the mock circulation system. While the deviation was slightly large for an elaborate calibration, it would still be useful as a primitive calibration. The influence of the respiratory change and other factors as well as the reliability of the calibration value should be investigated with an animal experiment as a next step.