Towards a More Widespread Clinical Use of Cardio-Ankle Vascular Index (CAVI) and CAVI0: Defining Reference Values in Healthy Russians

An ever-growing body of evidence supports the predictive value of arterial stiffness, often measured as pulse wave velocity (PWV), for future cardiovascular events and all-cause mortality.1 Indeed, stiffer arteries are responsible for increasing the pressure the heart must generate with each beat to propel blood forward into the vasculature, as well as for enhancing the transmission of pressure pulsatility to peripheral organs, which may damage their microvasculature. In recent years, first important efforts have been undertaken to prove the utility of targeting PWV for cardiovascular prevention through large randomized controlled trials.2 This work has demonstrated that the progression of arterial stiffening with age can be pharmacologically controlled if PWV measurements are performed in routine clinical examinations and used to define the treatment strategy. The definition of reference values of arterial stiffness in the healthy population is a crucial milestone towards the widespread use of arterial stiffness measurements in clinical practice. Carotid-femoral PWV (cfPWV), which estimates the propagation speed of the pressure wave in the aorta, iliac artery, and proximal femoral artery, is the gold standard metric of arterial stiffness in vivo, and the most widely used measuring technique, at least in western countries. For this reason, large population studies have provided values and age trajectories in healthy individuals to serve as a reference for the clinical use of cfPWV as a marker of vascular health.3 Nonetheless, other methods to assess arterial stiffness have been proposed over the years, each aiming to provide a different angle on vascular health or to simplify arterial stiffness measurement in the clinic. Because of inherent differences in the sampled district of the arterial tree and/or in the measuring principle/technique, arterial stiffness values derived from different methods cannot be directly compared, and each requires the establishment of dedicated reference values. The cardio-ankle vascular index (CAVI) is one such metric introduced to address two limitations of cfPWV.4 First, because of its measuring principle, cfPWV does not reflect the stiffness of the proximal aorta, which plays a vital role in the compliant function of the arterial system. Conversely, CAVI uses the sound produced by the aortic valve closure to identify the timing of the cardiac ejection, thereby allowing for characterizing the wave propagation in the proximal aorta. Second, because arteries behave non-linearly (approximately exponentially), cfPWV acutely depends on blood pressure at the time of measurement. CAVI, being based on a parameter that takes into account such exponential behavior (i.e., the stiffness index β), aimed to assess stiffness independently from the confounding effect of blood pressure. Concerning this second aspect, however, it was later shown that β and, consequently, CAVI still retain some residual pressure dependency, which is inherent to their mathematical definition. This finding led to our group introducing a corrected metric, CAVI0.4,5 To date, CAVI has found widespread use predominantly in East-Asian countries, particularly Japan, its country of development. Reference values of CAVI in Japanese and Chinese populations have long been established.6–8 However, because of potential differences in ethnic composition, socio-economic status and other factors, reference values identified for any specific population may not apply elsewhere. This has prompted researchers of non-East-Asian countries (e.g., Czech Republic9 and Spain10) to identify reference values for their target populations. In line with this, the present issue of the American Journal of Hypertension features a study by Safronova et al.,11 which provides the first reference values of CAVI and CAVI0 in a cohort of 742 presumed healthy Russians. In this study, Safronova and colleagues show that CAVI and CAVI0 display a clear increasing trend with age, comparable to that observed in other populations (Figure 1).6–10 Wang et al.6 found that healthy Chinese individuals presented a significantly lower CAVI than age-matched Japanese individuals. While CAVI in their Japanese cohort was, on average, higher than that found in other (larger) Japanese studies of Namekata et al.7 and Shirai et al.,8 collectively these works suggest that CAVI may, indeed, be consistently higher in Japanese individuals. Interestingly, age group-averages in Safronova’s Russian cohort closely matched those found in the Chinese population (Figure 1). Comparison to values reported in the Czech9 and Spanish10 studies is more difficult, since these both describe the general population rather than healthy individuals. This important difference may explain why these two cohorts (especially the Spanish) present an accelerated increase of CAVI with age in older adults.8 Cardio-ankle vascular index (CAVI) and CAVI0 reference values studies in different countries.a: CAVI reference values; b: CAVI0 reference values. Note that Japanese,6–8 Chinese,6 and Russian11 cohorts consist of apparently healthy individuals (closed symbols). Conversely, the Spanish10 and Czech9 cohorts represent the general population (open symbols). CAVI and CAVI0 correspond to the mean of age group-averages in males and females. These, in turns, are determined as means of right and left CAVI and CAVI0. Note that groups at the extremes of the age range may not exactly correspond their specific age group (e.g., Safronova et al. clustered all individuals above 60 years of age in a single group, here assigned to the 61–70 group). Reference values of CAVI and CAVI0 are typically stratified for sex. This is because large studies in the Japanese population reported slightly but consistently and significantly higher values of CAVI in men compared to women.7 This observation is consistent throughout the literature, albeit not statistically proven in all studies, likely due to their relatively small sample size.6–10 Despite reporting no significant difference between sexes in the entire cohort, overall, the study of Safronova et al.11 is in line with the literature: in all age groups except 30–40, CAVI/CAVI0 values were on average higher in men than in women (significant only for CAVI0 at 20–30 years). Curiously, however, CAVI in women aged 30–40 years was significantly higher than that of age-matched men. While the relatively small sample size of this healthy Russian cohort may have played a role in this surprising finding and constitutes the main limitation of Safronova’s work, future studies are warranted to ascertain whether sex-specific trajectories of arterial stiffness with age may indeed differ between the Russian population and other cohorts. Safronova’s work also provides some interesting insight into what variables affect CAVI and CAVI0, which is also helpful in underpinning differences between the two metrics. First, while both CAVI and CAVI0 significantly correlated with age, heart rate, and plasma glucose, only CAVI showed significant positive and negative correlations with diastolic (DBP) and systolic blood pressure (SBP), respectively. As mentioned previously, CAVI0 was introduced in 2017 by our group to correct for the residual pressure dependency of CAVI,5 sparking a debate on the merits and issues of the two metrics summarized in our review.4 At first glance, the findings of Safronova’s group appear to support the view that CAVI0 does indeed resolve the residual pressure dependency of CAVI. However, we have previously argued that cross-sectional correlations between CAVI or CAVI0 and pressure are not incontrovertible proof of the pressure (in)dependency of one or the other metric.4 This is because long-term exposure to elevated blood pressure has been shown to cause arterial remodeling, increasing arterial stiffness independent of blood pressure at the time of measurement.7,12 Second, the authors performed a multivariable linear regression analysis of the predictors of CAVI and CAVI0 within individual age groups (stratified by decade of age, see Figure 1), which led to a useful formula to calculate age group-dependent normal ranges of CAVI. They found that both CAVI and CAVI0 show negative associations with body mass index and heart rate, but only CAVI was also negatively associated with pulse pressure (PP) = SBP − DBP. This observation is not surprising when considering that CAVI and CAVI0 normalize the heart-to-ankle PWV (haPWV) with respect to mid pressure (Pm, i.e., the average of SBP and DBP) and DBP, respectively. where ρ is the blood mass density and Pref is a reference pressure (please refer to our review4 for a detailed explanation of the differences between the two formulas). CAVIunscaled corresponds to CAVI before being scaled to a PWV-like quantity through the scaling factors a and b.4 As shown in equations (1) and (2), the pressure normalization is performed by dividing the square of the measured haPWV by the normalizing pressure (i.e., either Pm or DBP). It follows that for any given DBP, a larger PP leads to a relatively stronger pressure normalization in CAVI than CAVI0, which may explain the additional negative association of CAVI with PP. It is worth noting that, because SBP and DBP follow distinct trajectories with age (the first steadily increasing and the second slightly increasing until mid-life and then dropping in the elderly) which also differ between sexes,13 this methodological difference may result in CAVI and CAVI0 following slightly different age trajectories. This may explain why Safronova et al. observed inter-sex CAVI0 differences in individuals aged 21–30, and inter-sex CAVI differences in those in the 31–40 years group. Similar CAVI vs. CAVI0 incongruences were also observed in the comparison between normotensive and hypertensive Japanese people.8 In summary, by providing reference values of CAVI and CAVI0 in healthy Russians, the study of Safronova and colleagues constitutes an important milestone towards a more widespread use of these two metrics in Russia. While larger studies are warranted in the future, their timely effort will hopefully play a pivotal role in improving vascular health awareness in their region. None.