Response to Dipping Deeper Into the Ambulatory Arterial Stiffness Index
Although we regret that Adiyaman et al1 misinterpreted some of our results, we are glad to see that they agree with the main findings of our study,2 in which we described a strong and previously undetected inverse association between ambulatory arterial stiffness index (AASI) and day-night blood pressure (BP) fall. Indeed, they were able to replicate our findings in 2 different cohorts.3,4
Adiyaman et al1 argue that the estimate of AASI may be influenced by the night:day ratio of the number of BP readings. However, to support their view, they rely on ambulatory BP recordings performed with a very low frequency of nocturnal BP readings. In fact, the night:day ratio of the number of BP readings in their study was only ≈0.30 with 26 to 30 measurements during the day and 8 to 9 measurements at night (see the figure in the letter by Adiyaman et al1) because of a longer nocturnal interval between automated measurements. In our representative patients (Figure 3 of our article),2 the number of valid BP measurements was 34 during nighttime and 57 during daytime (and not ≈25 and ≈35, as erroneously indicated by Adiyaman et al1). Such a distribution, rather than including a “disproportionally large” number of nocturnal readings, is the desirable result of having BP evenly measured every 15 minutes throughout the 24 hours. By avoiding an artificially lower BP sampling frequency at night than during the day, in our article the night:day ratio in the number of BP readings simply mirrors the physiologically different duration of the nighttime and daytime subperiods.
Our demonstration that day-night BP reduction is a major determinant of AASI provides a rational explanation for the dependency of AASI on the number of diurnal and nocturnal readings. Given that nocturnal diastolic and systolic BP reductions are strongly related to each other, the physiological nocturnal BP fall generates a number of nocturnal systolic and diastolic BP values, which are considerably lower than the corresponding daytime values. This, in turn, increases the regression coefficient of diastolic BP on systolic BP (B) and reduces AASI. If the number of nocturnal readings is artificially reduced by adopting longer between-measurement intervals at night, this reduces B and increases AASI. As a matter of fact, in our cohort, AASI calculated on the basis of daytime readings, thus excluding the nocturnal values, was much higher than 24-hour AASI (0.48±0.26 versus 0.31±0.17). In other words, we postulate that a low number of nocturnal readings artificially increases AASI, irrespective of any association with an increase in arterial stiffness.
Another point raised by Adiyaman et al1 is the high intercorrelation between daytime systolic BP and nocturnal fall in systolic BP, which should prevent these variables from being included simultaneously in a multivariate model. This observation is incorrectly made with reference to our study, however, in which we included in the same model daytime BP and percentage of nocturnal BP fall. In our population, there was indeed no significant correlation between the 2 variables (r=0.07; P=0.08), and the examined model was found to be free from collinearity using standard diagnostic techniques (tolerance: 0.89 to 0.99).
Moreover, Adiyaman et al1 argue that information on the correlation between left ventricular mass and pulse wave velocity was incomplete in our article. Indeed, it was completely absent. We did not address the association between pulse wave velocity and left ventricular mass index simply because this was not the aim of our study. Our observations focused on what was the primary target of our study, ie, the actual meaning of AASI. The important result of our article was that the previously reported5 association with a common index of organ damage, such as left ventricular mass, depends on the shared association with the degree of nocturnal BP reduction.
We agree that Bland-Altman plots may give a different and complementary view of the concordance between 2 measures, in particular when comparing 2 estimates of the same parameter. Conversely, we disagree that a Bland-Altman plot could be helpful to assess concordance between AASI and aortic pulse wave velocity. These 2 parameters measure quite different biological traits, which are distributed over different numerical ranges. Because data of different kind cannot be averaged or subtracted, absolute values of means, differences, and coefficients of variation cannot be given in such nondimensional graphs (see Figure 3, panel 3, in the article by Li et al6).
Finally, we do not agree with Adiyaman et al1 that adjustment for common determinants is inappropriate when assessing concordance between a newly proposed and an established trait. At least, the confounding effect of age should be taken into account when assessing the clinical value of any new surrogate marker of atherosclerosis. In our population, the bivariate relation between AASI and aortic pulse wave velocity (r=0.28; P<0.001) lost its significance when the effect of age was taken into account (partial r=0.05; P=0.39), not even considering the impact of mean arterial pressure and other variables. Although the actual physiological significance of AASI needs to be elucidated further, these data confirm our view that, at least in untreated hypertensive subjects, AASI may not be considered as a surrogate marker of arterial stiffness tout court.
Adiyaman A, Boggia J, Li Y, Wang J-G, O’Brien E, Richart T, Thijs L, Staessen JA. Dipping deeper into the ambulatory arterial stiffness index. Hypertension. 2007; 50: e59–e60.
Schillaci G, Parati G, Pirro M, Pucci G, Mannarino MR, Sperandini L, Mannarino E. Ambulatory arterial stiffness index is not a specific marker of reduced arterial compliance. Hypertension. 2007; 49: 986–991.
Li Y, Wang JG, Dolan E, Gao PJ, Guo HF, Nawrot T, Stanton AV, Zhu DL, O’Brien E, Staessen JA. Ambulatory arterial stiffness index derived from 24-hour ambulatory blood pressure monitoring. Hypertension. 2006; 47: 359–364.
Leoncini G, Ratto E, Viazzi F, Vaccaro V, Parodi A, Falqui V, Conti N, Tomolillo C, Deferrari G, Pontremoli R. Increased ambulatory arterial stiffness index is associated with target organ damage in primary hypertension. Hypertension. 2006; 48: 397–403.