P.077 In Vivo Validation of a Non-Invasively Measured Local Pulse Wave Velocity: Improved Performance Using Aortic Valve Closure in Stead of Aortic Valve Opening

Abstract

Introduction

High resolution multiple M-line ultrasound can be used to obtain local pulse wave velocity, by tracking the propagation of characteristic time-points of the distension waveform over an arterial segment. Local PWV provides a direct estimate of the local arterial stiffness.

Method

The left and right CCA of 12 young subjects was measured with multiple M-line ultrasound, resulting in 14 diameter waveforms spaced over 16.4mm. The second derivative of the distension waveforms was calculated to identify aortic valve opening and closure (AVO and AVC). By performing linear regression on M-line positions versus time-points, the local PWV was estimated. Beats were accepted for further analysis if the regression coefficient exceeded 0.9. The DC was measured using brachial-PP and carotid-distension, giving a reference estimate of the PWV using the Bramwell-Hill equation: $PW{V_{DC}} = {1 \mathord{\left/ {\vphantom {1 {\sqrt {\rho DC} }}} \right. \kern-\nulldelimiterspace} {\sqrt {\rho DC} }}$

Results

PWV_AVC (mean = 4.91m/s) showed a better intra-subject precision (0.46m/s) than PWV_AVO (mean = 5.21m/s, precision = 1.66m/s). The range of PWV_DC (2.8–7.9m/s) is comparable to PWV_AVC (3.5–7.3m/s) but not to PWV_AVO (2.4–13.2m/s). The correlation between PWV_DC and PWV_AVC was weak but significant (r = 0.69, p<0.001). No correlation between PWV_DC and PWV_AVO exists.

Discussion

In contrast to PWV_AVO, PWV_AVC has an acceptable precision (coefficient of variation less than 10%) and correlates with PWV_DC. The poor correlation between PWV_AVO and PWV_DC estimates may be caused by interference of early arterial wave-reflections leading to a biased estimate of PWV_AVO. The correlation between PWV_DC and PWV_AVC is weak, although carotid-PP may improve DC calculation.