Arterial path length estimation on brachial-ankle pulse wave velocity

Abstract

Brachial-ankle pulse wave velocity (baPWV) is one of few indices of arterial stiffness that has been incorporated into a routine clinical setting. baPWV is associated with markers of central arterial stiffness, but the absolute values are substantially higher than other PWV measures. We tested the hypothesis that the height-based formulas underestimate actual arterial path lengths overestimating PWV values. Arterial path lengths were measured in 220 adults by the three-dimensional MRI arterial tracing. Actual path lengths from the heart to the ankle (Lha) and from the heart to the brachium (L(hb)) were significantly longer than those estimated by the height-based formula (1464 ± 85 vs. 1439 ± 71 mm and 476 ± 41 vs. 335 ± 19 mm, both P < 0.0001). Consequently, the MRI-based effective path length for baPWV (L(ha) - L(hb)) was shorter than the height-based path length (988 ± 66 vs. 1105 ± 52 mm, P < 0.0001). baPWV re-calculated by using the MRI-based path length was 11% lower than that derived from the height-based path length but was still 45% greater than carotid-femoral PWV (cfPWV). Both height-based and MRI-based baPWV measures were significantly associated with cfPWV (r = 0.84 and 0.87, both P < 0.0001). We concluded that the height-based formulas to estimate arterial path length grossly overestimate the actual path length resulting in the overestimation of PWV values and that baPWV was still significantly higher than other PWV measures even when the actual path length is used to calculate. These results suggest that other methodological considerations apart from the arterial path length estimation contribute to higher absolute baPWV values.