Continuous cardiac output monitoring is urgently required because of the growth of elderly population and shortage of clinical staff. However, the reliable, less-invasive solution of monitoring cardiac output has not been determined. On the other hand, the measurement of peripheral arterial blood pressure waveform can be conveniently collected. In the dissertation, we proposed novel physiologic waveform analysis techniques which can estimate cardiac output from a peripheral arterial blood pressure waveform. The techniques can overcome the confounding reflection by estimating the cardiac output over multiple cardiac cycles. Validation results have indicated that our method was comparative with other traditional methods at the baseline level and became increasingly superior to others with more challenging cardiac output changes.We also introduced pulse transit time to overcome the estimation error due to varying arterial compliance. As a direct function of arterial compliance, pulse transit time can reduce the estimation variation and thus improve the accuracy. Two conventional approaches for measuring pulse transit time were compared, and both methods were sensitive to artifact. Therefore, a novel method was developed which can robustly estimate the pulse transit time, with which the accuracy of estimating cardiac output was significantly improved.A tube-load model with an adaptive transfer function was further proposed. The model can reconstruct the central blood pressure of flow waveforms from the peripheral measurements beat by beat. In this study, the reconstructed central blood pressure waveform is an earlier marker of hypovolemia than the peripheral blood pressure waveform in a simulated hemorrhage experiment.Cardiac output can also be derived from the reconstructed central blood pressure waveform. In addition, an extended model was proposed for estimating cardiac output from a pulmonary artery pressure. This extension allows to simultaneously estimating cardiac output and left atrial pressure, which are two important determinants of pulmonary artery pressure. Validation in a large critical/medical population database demonstrated the proof of concept. The techniques will ultimately be employed for continuous and non-invasive cardiac output monitoring for clinical settings. One potential future direction would be end-diastolic volume and ejection fraction estimation from a peripheral arterial blood pressure waveform.
Read
