Estimation of an aortic pressure waveform from radial pulse using ARX model

Hardening of the arteries, and its relation to age, is far from a new phenomenon, and is an important independent risk factors for cardiovascular disease. It calls for the development of a system which measures central aortic stiffness noninvasively. This study was performed to establish an algorithm which estimates central aortic pressure waveforms from the pulse waveform measured from radial artery. Transfer functions were established using 10th order ARX model by setting radial pulse as an input and aortic pulse as an output, and were verified for the stability through residual analysis. Order of an ARX model was decided based on the FPE(final prediction error). In addition, correspondence between two pulse waves was evaluated by using systolic pressure, pulse pressure, ejection time, and RWTT(reflected wave transit time) between measured and estimated aortic pressure waves. On the comparison between measured aortic pressure waves and estimated aortic pressure waves for each variable, similar results were observed for all the variables. Errors for systolic pressure, pulse pressure, ejection time, and RWTT between measured and estimated waves were -4.75±3.86mmHg, - 5.29±3.16mmHg, 19.3±20msec, and 2.5±11.5msec, respectively. The results could be applied for the development of a system which provides the quantitative aortic stiffness using easilymeasurable radial artery pulse waves.