Cardio–ankle vascular index and carotid–femoral pulse wave velocity

The aorta and other large elastic arteries play a major role in the regulation of blood pressure (BP) and peripheral blood flow. Indeed, thanks to their viscoelastic properties, large arteries have the task of damping the pulsatile output of the left ventricle and of changing the rhythmic, intermittent and discontinuous activity of the cardiac pump into a continuous one. Thus, elastic arteries guarantee the propelling activity of the circulating system during the diastolic phase of the cardiac cycle and can be considered as the diastolic pump in the cardiovascular system [1]. Therefore, the evaluation of the mechanical properties of the aorta and of the other elastic arteries is a main topic in research and clinic. The measurement of pulse wave velocity (PWV) represents the simplest noninvasive way to measure the stiffness of a specific arterial segment, as it is an easy test, reproducible and supported by considerable scientific literature. The pulse wave is transmitted through the arterial vessels, and its speed is inversely related to the viscoelastic properties of the wall itself: the higher the velocity, the less elastic the wall. Carotid–femoral PWV (cf-PWV) is considered the gold-standard method for the noninvasive assessing of aortic stiffness [2]. The proximal transducer is placed on the common carotid artery and the distal transducer on the femoral artery. In this way, the viscoelastic properties of the aorta are assessed. Several epidemiological studies showed the predictive meaning of high cf-PWV values for cardiovascular diseases, over and above other traditional major risk factors [3–8]. However, cf-PWV does not exactly assess aortic distensibility (Fig. 1). First, the ascending aorta is excluded by this PWV measurement. It is well known that the ascending aorta has a crucial buffer function in absorbing left ventricular ejection and in damping pulsatile flow. The structural changes induced by ageing are particularly marked in proximal aortic wall, such as fragmentation and thinning of elastin fibres as well as loss of orderly arrangement in the elastic laminae [9]. These changes induce dilation, stiffening and elongation of the aorta. Unlike the other arterial segments, the ascending aorta length is age-related and increases 12% per decade (from about 44 mm at 20 years to 98 mm at 80 years) [10]. Second, the iliac arteries and the initial part of femoral arteries are included in the evaluation of cf-PWV. These arteries present different anatomical structure and viscoelastic properties compared with the thoracic aorta, reducing the elastic component and increasing the muscular component in the tunica media of arterial wall. Third, brachiocephalic trunk and common carotid artery are included in the cf-PWV measurement, even if in this arterial district pulse wave travels in opposite direction and at different speed. Despite these limitations, cf-PWV at present represents the most reliable noninvasive index for assessing aortic distensibility.FIGURE 1: Pressure wave pathway in carotid–femoral (left panel) and cardio–ankle (right panel) pulse wave velocity assessment.The VaSera device (Fukuda Denshi Co. Ltd, Tokyo, Japan) uses the cardio–ankle vascular index (CAVI) to assess arterial stiffness. PWV as used in CAVI is given by the ratio between heart-to-ankle distance and the pulse pressure transit time from the heart to the tibial artery. The cardio–ankle distance is automatically estimated from the patient's height. More complex is the measurement of the pulse wave transit time delay [11]. Thanks to a phonocardiograph, opening and closing of aortic valve can be recorded. As the exact time of the opening of the aortic valve is not well detectable in the phonocardiogram, the time between the closing of the aortic valve and the dicrotic notch in the brachial pulse waveform is assessed, considering this latter as analogous to the time between the opening of the aortic valve and the foot of the brachial artery. However, this is an approximate estimation that does not take into account the differences between central and peripheral ejection time due to heart rate variation [12]. The cardio–ankle transit time results from the sum of this latter estimated transit time and the time delay between the feet of brachial and tibial artery pressure waves. In summary, although the cf-PWV can be regarded as a reliable index of the viscoelastic properties of the aorta, the cardio–ankle PWV is a parameter that provides information about mechanical properties of the aorta, femoral artery and tibial artery as a whole. If the measurement of cardio–ankle PWV, compared with cf-PWV, includes also ascending aorta, on the other hand, the large arteries of the lower limbs throughout their course are also included in the measurement of cardio–ankle PWV. Lower limb arteries are predominantly muscular arteries. PWV presents different characteristics through elastic and muscular arteries [13] (Table 1). First, an increase in aortic PWV with age is a well established phenomenon. Ageing is characterized by a generalized increase in arterial wall stiffness related to structural changes. Increased elastase activity and reduced elastin synthesis cause thinning and breakage of elastin fibres, resulting in a decrease in the elastin and collagen ratio. However, the relationship between age and PWV in large elastic arteries is not linear, but increases exponentially with age [1,14]. Thus, cf-PWV is considered a reliable parameter representing vascular ageing. Moreover, the influence of sex on cf-PWV is negligible [15]. On the contrary, PWV in the upper arm muscular arteries, evaluated by carotid–radial PWV, is relatively unchanged with age and does not provide information about the vascular ageing. PWV in lower limbs, evaluated by femoral–tibial PWV, is higher in men than in women [16]; although in men, PWV does not change with age, in women, lower limbs PWV presents a weak, but significant, linear increase in PWV with ageing [16]. Second, aortic PWV is relatively insensitive to sympathetic system activation [17,18] and an increase in cf-PWV mainly reflects structural changes in the aortic wall. cf-PWV values are relatively stable between repeated measures [19] and with time, making the follow-up of patients reliable. On the contrary, PWV in muscular arteries is deeply affected by sympathetic system activity [20–22]. High values of PWV in muscular arteries are mostly related to functional changes in the arterial wall; thus, the reliability of this measurement in longitudinal studies may be questionable. Third, high cf-PWV values are closely associated with markers of atherosclerosis (coronary artery calcifications score, carotid and femoral intima–media thickness) and coronary artery disease [23], and large epidemiological studies have demonstrated the predictive value of high cf-PWV values for cardiovascular events regardless of other major risk factors known [4,5,7,8,24]. In contrast, PWV measured at the upper or lower limb is not associated with the major biomarkers of cardiovascular risk except mean arterial pressure [8,25,26] and has not shown any prognostic value for cardiovascular mortality and morbidity. Fourth, elastic and muscular arteries show different anisotropic properties and responses to pressure loading, which are mainly determined by the relative amounts and disposition of collagen and elastin fibres and of smooth muscle cells [27]. A higher pulsatile stress in carotid or in brachial arteries may lead to an increase in arterial diameter to keep tensile stress within certain limits and therefore to preserve local compliance despite wall stiffening. Conversely, in femoral arteries, a decrease in the inner arterial diameter is often observed, as arterial wall thickens for the development of atherosclerotic changes, which finally leads to arterial narrowing [28].TABLE 1: Main differences between pulse wave velocity in elastic and muscular arteriesA further important methodological difference between CAVI and cf-PWV measurements is represented by the use of β-stiffness index in the assessment of arterial distensibility by CAVI. CAVI's promoters increasingly emphasize the concept of pressure-dependence of cf-PWV, trying to insinuate the idea that PWV is closely related to the BP values (a given pressure value corresponds to a given PWV value), de-facto delegitimizing the usefulness of the measurement of the cf-PWV in clinical practice and research. First of all, it is important to understand the principle of ‘pressure-dependence’ and what it means. Several aspects in the relationship between BP and arterial distensibility need to be clarified. High BP, together with ageing, is a major factor causing structural alteration in arterial wall. Hypertension is characterized by a persistent increase in transmural pressure. A greater biosynthesis of collagen in arterial wall occurs to counterbalance the high BP values, reducing the elastin/collagen fibres ratio and consequently the arterial distensibility [1]. This process can be associated with functional changes, ‘remodeling’ of the arterial wall, hypertonia, hypertrophia and hyperplasia of smooth muscle cells and alterations in the endothelial function. Considering these structural alterations in arterial wall induced by high BP values, it is inevitable that an association between PWV and BP values is found in cross-sectional studies. Apart from the chronic effect of high BP on the vascular wall, actual wall pressures loading affects arterial distensibility. Due to the multilayered and inhomogeneous nature of the arterial wall, the relationship between stress and strain is nonlinear across a range of loading arterial pressures. That is, arterial wall is not a perfectly elastic material, and its elastic properties vary according to different distending pressures. The relationship between arterial pressure and diameter in a single artery can be described by an exponential curve, as shown by Hayashi et al.[29], although evidence from in-vitro experiments in aortas ex vivo are not all consistent with this assumption. As load-bearing components in the aortic wall shift from elastin to collagen, when aortic pressure increases, the compliance of the aortic wall follows a complex behaviour, which may be better described by an arctangent-type model in which a first increase in compliance, with increasing pressures, is followed by a decrease [30]. Nevertheless, if the exponential approximation is considered, the exponent of the pressure–diameter curve may be considered as the increase in arterial stiffness for a given increase in distending pressure. This parameter is often named as β-stiffness index and represents the stiffening of a localized arterial segment for a given increase in distending pressure. In the search for a ‘pressure independent’ parameter for arterial stiffness, numerous attempts have been made for the evaluation of the β-stiffness index. The CAVI advocates to be a measure of the β calculated by combining the simplified Bramwell–Hill equation with the equation describing an exponentially distending wall with pressure [11,31]. The procedure of estimating the β-stiffness index from heart–ankle PWV and BP could represent a not completely clarified methodological issue. Whereas β-stiffness index should be estimated in single localized artery from pressures and diameters, the CAVI uses a PWV from a long and inhomogeneous arterial path from the heart to the lower limbs. The assumptions of calculation of β-stiffness index should constrain the formulation in a localized arterial segment instead of a long arterial path. A previous study investigated the relationship between CAVI and the β-stiffness index measured in different aortic regions with pressures and diameters, and found a weak-moderate positive relationship, which was stronger in the thoracic (r = 0.49) than in the descending aorta (r = 0.30) [32]. Apart from technical and methodological aspects, the more relevant open issue is whether it is useful to consider the aortic PWV or another parameter, such as the β-stiffness index derived from PWV, for the clinical application in the cardiovascular assessment. Although the use of the long aortic path may be considered a questionable approach for the calculation of β-stiffness index (designed to be used in a single artery), a proven clinical usefulness could overcome doubts in its application. The estimation of β-stiffness index, as applied in the calculation of CAVI, is described by the following equation: where ρ is blood density, a and b are proprietary constants added for scaling and squared PWV is estimated over the heart-to-ankle vascular bed. However, as recently highlighted by Spronck et al.[33], this formula presents a residual ‘pressure-dependency’, as there are some theoretical drawbacks in the approximation of the β-stiffness index residing in the use of systolic and diastolic pressures instead of diastolic and reference pressure values. These authors proposed a straightforward correction for the calculation of an intrinsic, pressure-independent measure of β-stiffness index, from PWV, diastolic and a ‘reference’ pressure [33]. To clarify the relationship between aortic PWV and β-stiffness index, we calculated the CAVI from cf-PWV and BP values in a large population study including 2727 patients [14]. A strong quadratic relationship exists between cf-PWV and β-stiffness index (Fig. 2). As expected, the greater deviation from this correlation is for higher PWV values. The cf-PWV and the β-stiffness index as measured by CAVI are sometimes viewed as complementary evaluations of the arterial properties. But, as a strong relationship exists between these two parameters when the same PWV is introduced in the equation for the calculation of β (R2 = 0.97), the differences may not be attributed to the procedure of calculation of the β, but rather on the way, PWV is calculated and on which arterial trajectory is considered.FIGURE 2: Correlation between carotid–femoral pulse wave velocity (PWV) and the β-stiffness index as estimated from cardio–ankle vascular index equation (see text). Pulse wave velocity (m/s), SBP and DBP values (Pascal). ρ = 1050 kg/m3 were included in the equation. Proprietary scaling parameters (a and b, see text for details) were not included in the calculation. Data obtained from a general population study including 2727 patients (mean age 56.9 ± 25.5 years).The study by Topouchian et al.[34] in this issue of the Journal of Hypertension tries to throw some light in the interpretation of these different views of arterial stiffness assessment by analysing the factors that differently impact CAVI and cf-PWV in a high-risk population of patients with metabolic syndrome. As expected, cf-PWV was increased in patients with metabolic syndrome. On the contrary, no differences in CAVI values were found between the group of patients with and without metabolic syndrome. The higher cf-PWV values found in metabolic syndrome confirm what is already known in the literature, showing a cumulative contribution of the different metabolic syndrome constitutive elements in the stiffening of the aorta [35,36]. The components of metabolic syndrome were also considered separately. Whereas all the five components defining metabolic syndrome were positively associated with age and sex-adjusted cf-PWV, CAVI was positively correlated only with high values of BP and glycaemia and negatively related to waist circumference. The differences observed between CAVI and cf-PWV do not have a straightforward interpretation and are likely due to a different impact of the components of metabolic syndrome on the aorta or on the muscular arteries of lower extremities. A validation of this hypothesis comes from the results of studies carried out by Wohlfahrt et al.[37]. These authors analysed the contribution of metabolic syndrome to aortic or lower limbs PWV and found that metabolic syndrome constitutive elements have a large contribution on aortic stiffening and a low or even negative effect on lower limbs PWV. The relevance of age and high BP values in determining the stiffening of arterial vessels is confirmed for both the methods, cf-PWV and CAVI. The contribution of age to CAVI appears to be even higher than to cf-PWV. The ageing of the arterial wall implies mainly the degradation of the elastin component, which is highly represented in the proximal aorta and less in the peripheral arteries. But given the cumulative action over time of cardiovascular risk factors, both on the aortic segment and on the arteries of the lower limbs, the ageing of the vasculature may be more pronounced in the lower limb arteries of patients with multiple cardiovascular risk factors [16,38]. Another aspect discussed in the study by Topouchian et al.[34] regards the advocated pressure-independency of the CAVI from BP values, when compared with cf-PWV. This deduction was gathered from the less pronounced significance of the association of high BP with CAVI than with cf-PWV in this study population. On the contrary, the authors did not consider BP as a continuous variable in their multivariate analysis but rather as categorical. The purported BP-independency of CAVI may not be fully confirmed or denied by the data of the current study, as within patient variations of BP were not available. In a previous study, our research group showed that the role of mean arterial pressure as determinant of PWV significant changes at different ages, being negligible in the young and in the elderly (<5% in patients under 20 years and over 70 years of age). In any case, also in the adults aged from 20 to 70 years, mean arterial pressure explained less than 11.1% of the PWV variance [14]. This interaction between BP and PWV also includes the alterations in the viscoelastic properties of large arteries induced by structural changes caused by hypertension, and described above, and only in part are the result of transmural pressure at the time of examination. Certainly, a measure of the arterial properties which does not display this intrinsic dependency upon BP changes may represent an attractive perspective for the clinical application of arterial stiffness, and particularly for the investigation of the effects of interventions aimed at changing the arterial wall properties independently from a BP reduction. Beyond issues regarding the two measurements of arterial stiffness considered in the current study, CAVI and cf-PWV, there is an urgent need to clarify some more general uncertainties. The heterogeneous arterial composition and function, when going from the heart to the periphery of the arterial tree, makes it hard to indicate the best arterial path for the clinical evaluation of PWV. Currently, the carotid-to-femoral arterial trajectory is regarded as the preferable for the quantification of aortic stiffness. The hypothesis that the clinical value of arterial stiffness measurement may be increased by including lower limb muscular arteries is considered by the CAVI. In addition, the calculation of β-stiffness index from PWV and BP values is an interesting evolution that could overcome limitations and ambiguities derived from the BP dependency of the properties of the arterial wall, although the way of calculating this parameter is still debated and the strong correlation with the PWV used in the calculation might make unnecessary this procedure. The use of devices with algorithms or procedures which does not provide an open output of all the variables included in the calculation does not facilitate the task to find the most appropriate way to evaluate arterial stiffness, as often the marketing logic does not always meet the purposes of research. The metaphor presenting CAVI and cf-PWV as two pictures of the same patient from different angles, being thus complementary, may be at present misleading. Indeed, the role of cf-PWV is well defined, and its usefulness in research and in daily clinical practice is consolidated and undisputed. We know well what we are investigating through cf-PWV and the meaning of the measurement provided by this method. The significance of aortic stiffening evaluated by cf-PWV as a biomarker of subclinical vascular damage is confirmed by a large number of clinical trials. On the contrary, the meaning and interpretation of the CAVI is still debated, despite the large amount of literature published about it. Only longitudinal studies, comparing different methods and devices in large population samples and in different clinical conditions, such as the study by Topouchian et al.[34], will be able to clarify whether CAVI may be useful in clinical practice, complementary to cf-PWV. ACKNOWLEDGEMENTS Conflicts of interest P.S. is consultant for DiaTecne s.r.l., Milan, Italy.