Numerical assessment and comparison of pulse wave velocity methods aiming at measuring aortic stiffness

作者：Obeid Hasan^*; Soulat Gilles; Mousseaux Elie; Laurent Stephane; Stergiopulos Nikos; Boutouyrie Pierre; Segers Patrick

来源：Physiological Measurement, 2017, 38(11): 1953-1967.

DOI：10.1088/1361-6579/aa905a

• numerical arterial model
• pulse transit time
• hemodynamic signals

Pulse waveform analyses have become established components of cardiovascular research. Recently several methods have been proposed as tools to measure aortic pulse wave velocity (aPWV). The carotid-femoral pulse wave velocity (cf-PWV), the current clinical gold standard method for the noninvasive assessment of aPWV, uses the carotid-to-femoral pulse transit time difference (cf-PTT) and an estimated path length to derive cf-PWV. Objective: The heart-ankle PWV (ha-PWV), brachial-ankle PWV (ba-PWV) and finger-toe (ft-PWV) are also methods presuming to approximate aPWV based on time delays between physiological cardiovascular signals at two locations (similar to heart-ankle PTT, ha-PTT; similar to brachial-ankle PTT, ba-PTT; similar to finger-toe PTT, ft-PTT) and a path length typically derived from the subject's height. To test the validity of these methods, we used a detailed 1D arterial network model (143 arterial segments) including the foot and hand circulation. Approach: The arterial tree dimensions and properties were taken from the literature and completed with data from patient scans. We calculated PTTs with all the methods mentioned above. The calculated PTTs were compared with the aortic PTT (aPTT), which is considered as the absolute reference method in this study. Main results: The correlation between methods and aPTT was good and significant, cf-PTT (R-2 = 0.97; P < 0.001; mean difference 5 +/- 2 ms), ha-PTT (R-2 = 0.96; P < 0.001; 150 +/- 23 ms), ba-PTT (R-2 = 0.96; P < 0.001; 70 +/- 13 ms) and ft-PTT (R-2 = 0.95; P < 0.001; 14 +/- 10 ms). Consequently, good correlation was also observed for the PWV values derived with the tested methods, but absolute values differed because of the different path lengths used. Significance: In conclusion, our computer model-based analyses demonstrate that for PWV methods based on peripheral signals, pulse transit time differences closely correlate with the aortic transit time, supporting the use of these methods in clinical practice.

• 出版日期2017-11