Ambulatory arterial stiffness index, pulse wave velocity and augmentation index--interchangeable or mutually exclusive measures?

Abstract

The ambulatory arterial stiffness index (AASI) has been proposed as a novel measure of arterial stiffness and has been prospectively shown to predict stroke and cardiovascular death, but not cardiac death. This index has prompted considerable controversy as to whether it is a true measure of arterial stiffness.
The present study aimed to examine three different measures of arterial stiffness - pulse wave velocity (PWV; Complior), wave reflection [augmentation index (AIx)] and AASI - in a large hypertensive population, comparing their determinants and intercorrelations, both unadjusted and adjusted for confounders, and using Bland-Altman analysis to determine 95% confidence intervals for the ability of the AASI to predict PWV, the proposed gold standard of arterial stiffness.
The AASI correlated univariately with both PWV and the AIx in individuals overall (r = 0.28 for PWV and r = 0.24 for AIx; both P < 0.001) and in those with untreated or treated hypertension. Adjustment for age in the current study negated entirely the positive correlation between the AASI, PWV and AIx. Additional adjustment for confounders did not significantly alter these nonsignificant relationships. Furthermore, the 95% prediction limits for the AASI to predict PWV were +/- 4.18 m/s and for the AASI to predict AIx were +/- 25.4%, suggesting that the methods would not be interchangeable in a clinical setting. Direct comparative studies would be required to establish the relative predictive strength of each measure and whether combining measures can provide additional risk prediction. Until such data become available, we propose that the measures should not be considered interchangeable.

Full text

Page 1

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Ambulatory arterial stiffness index, pulse wave velocity and
augmentation index – interchangeable or mutually
exclusive measures?
Paula Jerrard-Dunne, Azra Mahmud and John Feely
Background The ambulatory arterial stiffness index (AASI)
has been proposed as a novel measure of arterial stiffness
and has been prospectively shown to predict stroke and
cardiovascular death, but not cardiac death. This index has
promptedconsiderablecontroversyastowhetheritisatrue
measure of arterial stiffness.
Objective and methods The present study aimed to
examine three different measures of arterial stiffness –
pulse wave velocity (PWV; Complior), wave reflection
[augmentation index (AIx)] and AASI – in a large
hypertensive population, comparing their determinants and
intercorrelations, both unadjusted and adjusted for
confounders,andusingBland–Altmananalysistodetermine
95% confidence intervals forthe ability of the AASI to predict
PWV, the proposed gold standard of arterial stiffness.
Results The AASI correlated univariately with both PWV
and the AIx in individuals overall (rU0.28 for PWV and
rU0.24 for AIx; both P<0.001) and in those with untreated
or treated hypertension. Adjustment for age in the
current study negated entirely the positive correlation
between the AASI, PWV and AIx. Additional adjustment for
confounders did not significantly alter these nonsignificant
relationships. Furthermore, the 95% prediction limits
for the AASI to predict PWV were W4.18m/s and for the
AASI to predict AIx were W25.4%, suggesting that the
methods would not be interchangeable in a clinical
setting. Direct comparative studies would be required to
establish the relative predictive strength of each
measure and whether combining measures can provide
additional risk prediction. Until such data become
available, we propose that the measures should
not be considered interchangeable. J Hypertens 26:529–
534 Q 2008 Wolters Kluwer Health | Lippincott Williams &
Wilkins.
Journal of Hypertension 2008, 26:529–534
Keywords: arterial stiffness, arterial structure and compliance, blood
pressure measurement/monitoring, pulse wave velocity, statistical
analysis
Abbreviations: AASI, ambulatory arterial stiffness index; AIx, augmentation
index; PWV, pulse wave velocity
Department of Pharmacology and Therapeutics, Trinity College Dublin, Trinity
Centre, St James’s Hospital, Dublin, Ireland
Correspondence to Dr Paula Jerrard-Dunne, Department of Pharmacology &
Therapeutics, Trinity Centre for Health Sciences, St James’s Hospital,
Dublin 8, Ireland
Tel: +353 1 4162291; fax: +353 1 4539033; e-mail: jerrardp@tcd.ie
Received 24 April 2007 Revised 18 September 2007
Accepted 15 October 2007
Introduction
The ambulatory arterial stiffness index (AASI) – defined
as 1 – (slope of diastolic on systolic pressure during 24-h
ambulatory monitoring) – has been proposed as a novel
measure of arterial stiffness [1,2] and has been shown to
predict stroke and cardiovascular death, but not cardiac
death, in a number of independent populations [3–5].
This novel index, however, has prompted considerable
controversy and debate as to whether the index is a true
measure of arterial stiffness [6–11].
A number of direct and indirect methods of measuring
arterial stiffness have previously been described [12]. Of
these, carotid–femoral pulse wave velocity (PWV) has
been proposed as the gold standard for arterial stiffness
measurement because it is a more direct measure of
stiffness, has good reproducibility and there is evidence
from a number of large prospective independent studies
to support its use as an independent predictor of vascular
outcomes [12].
In their original articles describing the method, Li et al.
[2,13] reported positive correlations between the AASI
and two established measures of arterial stiffness, PWV
and the augmentation index (AIx). This analysis was
criticized on the basis that finding a strong positive
correlation does not necessarily imply that the AASI is
a valid surrogate measure of arterial stiffness [7,8]. Sec-
ondly, it was proposed that the authors failed to adjust
fully for potential confounding factors [6,8]. It has been
suggested that the AASI may be largely influenced by
peripheral resistance vessels and by ventriculo-arterial
coupling [6,11]. Diurnal variability in blood pressure also
appears to influence the AASI [10]. Furthermore, the
comparative measurements were performed in healthy
volunteers and in predominantly normotensive, Chinese
community-dwelling individuals [2], which may not be
representative of the hypertensive populations in whom
the AASI might be expected to be clinically applied. A
recent study correlating the AASI and PWV in a hyper-
tensive population [10] found much poorer correlations
Original article529
0263-6352 ? 2008 Wolters Kluwer Health | Lippincott Williams & Wilkins

Page 2

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
than those reported in healthy individuals, and these
correlations were not independent of potential confound-
ing factors.
The correlation coefficient, r, measures the strength of
relation between two variables, not the agreement
between them, and can be misleading when assessing
agreement between two methods of measurement [14].
Toovercomethis,BlandandAltmanproposedanalterna-
tiveapproachtocomparingmethodsofmeasurementthat
estimates the 95% confidence intervals for the ability of
one method to predict another [14] (see also: http://www-
users.york.ac.uk/?mb55/meas/diffunit.htm).
The aim of the present study was to examine three
parameters (PWV, AIx and AASI), comparing their
determinants, looking at their inter-correlations, both
unadjusted and adjusted for confounders, and using
Bland–Altman analysis to determine 95% confidence
intervals for the ability of the AASI to predict PWV,
the proposed gold standard of arterial stiffness [12].
Materials and methods
Thestudypopulationconsistedofaconsecutivesampleof
individualsreferredforassessmenttoahypertensionclinic
between May 2000 and January 2007. All first-episode
ambulatory blood pressures (n¼2067) were evaluated
and those with 70% or more valid readings (n¼1714,
52% women; mean age, 51(14) years; age range, 17–
86years) were included in the analysis. Of these 1714
subjects, 824 had undergone measurement of the AIx
(426 of whom were untreated and 398 were treated with
antihypertensive therapy) and 622 also had their PWV
measured (391 untreated and 231 treated). Of those indi-
viduals on antihypertensive monotherapy, 14% were on
b-blockers, 13% on angiotensin-converting enzyme
inhibitors,6%onthiazidediuretics,7%oncalciumchannel
blockers, 5% on angiotensin-receptor blockers, 2% on
a-blockers, 1% on other drugs and 52% on were on
multiple/combination antihypertensive therapies. All
patientsgaveinformedconsent,thestudyhadinstitutional
ethicscommitteepermission,andtheproceduresfollowed
were in accordance with institutional guidelines and the
principles of the Declaration of Helsinki.
Ambulatory arterial stiffness index
Ambulatory blood pressure monitoring was performed
using validated oscillometric devices (SpaceLabs, Issa-
quah, Washington, USA). Blood pressure was recorded
every 30min between the hours of 0600h to midnight
andhourlyfrommidnightto0600h.Nomanualeditingof
the monitor readings was performed. The AASI was
calculated automatically using the DABL software sys-
tem (DABL Ltd., Dublin, Ireland).
Patients were studied in the morning after an overnight
fast, having refrained from caffeine-containing bev-
erages, alcohol, and smoking in the previous 12h, after
which the AIx and carotid–femoral PWV were measured
as previously described [15]. Brachial blood pressure and
the heart rate were measured using an automated oscillo-
metric device (Omron Model HEM 705-CP; Omron
Corporation,Tokyo,Japan)intherightarm,inthesupine
position after 15min rest by a trained observer. Three
blood pressure readings were taken at 1-min intervals,
and the mean was used for data analysis. Dipping was
defined as at least 10% nocturnal reduction in systolic
pressure and/or diastolic pressure.
Arterial pulse wave analysis
Radial applanation tonometry was used to acquire the
aortic pressure waveform noninvasively using the gener-
alized transfer function (version 8.0; Sphygmocor, Syd-
ney, Australia). The aortic pressurewaveform was used to
derive the aortic blood pressures and the AIx (difference
in height of the first and second systolic peak in the aorta
divided by the aortic pulse pressure).
Pulse wave velocity
The carotid–femoral PWV was measured with an auto-
mated system (Artech Medical, Pantin, France) using
the foot-to-foot method. The carotid and femoral wave-
forms were acquired simultaneously with two pressure-
sensitive transducers and the transit time of the pulse was
calculated by the system software. The distance between
the two arterial sites was measured on the body using a
tape measure and the PWV was calculated as the distance
divided by time (meters per second). At least 12 succes-
sive readings were used for analysis to cover a complete
respiratory cycle. The interclass correlation coefficients
between the first and second measurements were 0.94 for
the AIx and 0.84 for PWV. The coefficients of variation
for the AIx and PWV were less than 5%. The reprodu-
cibility of the measured variables was comparable with
previous studies [16,17].
Statistical analysis
The data were analysed using SPSS (version 12.0; SPSS
Inc., Chicago, Illinois, USA). The AASI, PWV and AIx
approximatednormaldistributions.Correlations between
the three proposed measures of arterial stiffness were
calculated both before and after adjustment for age and
gender, and multivariate adjustment (age, gender, mean
arterial pressure, mean heart rate and height). To com-
pare the three methods of measuring arterial stiffness, we
then applied the method proposed by Bland and Altman
for comparing methods of measurement that give results
in different units to estimate the 95% confidence inter-
vals for the AASI to predict the measured PWV and AIx
[14] (see also: http://www-users.york.ac.uk/?mb55/meas/
diffunit.htm).
Correlations and 95% prediction limits were calculated
for the population overall and additionally for untreated
530
Journal of Hypertension
2008, Vol 26 No 3

Page 3

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited. Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
and treated hypertensive individuals, to determine
whether antihypertensive therapy altered the inter-
relationship between the different measures of arterial
stiffness.
Results
Table 1 presents the characteristics of the study popu-
lation. Tables 2 and 3 present the major determinants of
the three arterial stiffness measures. Increasing age and
mean arterial pressure were independently associated
with all three measures of arterial stiffness. The AASI
and AIx were influenced by gender (higher in females)
and body height (inverse relationship), while the heart
rate was an independent predictor of PWV and the AIx
but not the AASI (Table 2). Age, gender, mean arterial
pressure, heart rate and height were therefore included in
the multivariate regression model.
As has been previously reported [10], dippers (defined as
?10% nocturnal reduction in systolic pressure and/or
diastolic pressure) had a lower AASI than nondippers
[0.34 (16.2) versus 0.47 (0.18), P<0.001], and an inverse
relationship was observed between nocturnal systolic and
diastolic pressure reduction and the AASI (Table 2). This
relationship persisted on multivariate analysis. In con-
trast,nosuchassociationwasobservedbetweennocturnal
blood pressure reduction and either PWV or the AIx
(Table 2). There was no heterogeneity in the results
when dippers and nondippers were analysed separately.
Correlations between the different arterial stiffness
measures
Onunivariate analysis, the AASI was correlated with both
PWV and the AIx in individuals overall (correlation
coefficients r¼0.28 for PWV and r¼0.24 for AIx;
P<0.001 for both) and in those with either untreated
or treated hypertension (Table 4). Adjusting for gender
did not significantly alter these relationships. All signifi-
cant correlations were lost, however, following adjust-
ment for age. Additional adjustment for the heart rate,
mean arterial pressure and height did not significantly
alter these nonsignificant relationships. The associations
between the AASI and PWV/AIx were therefore not
independent of age.
Comparison of methods using Bland–Altman analysis
We used the method proposed by Bland and Altman for
comparing methods of measurement that give results in
different units, which estimates the 95% confidence
intervals for the AASI to predict the measured PWV
and AIx. The 95% prediction limits for the AASI to
predict PWV were ?4.18m/s. The 95% prediction limits
for the AASI to predict the AIx were ?25.4%. These
limits were not significantly different for untreated indi-
vidualsversustreatedhypertensiveindividuals (Table5).
Discussion
The AASI has been proposed as a more readily available
alternative to more operator-dependent measures of
arterial stiffness requiring specialized staff and equip-
ment [18]. The basis for this proposal is that emerging
data show a relationship between the AASI and vascular
outcomes and that strong correlations have been demon-
strated between the AASI and more established arterial
stiffness measurements [18]. The correlation coefficient,
r, measures the strength of relation between two vari-
ables, not the agreement between them, and can be
misleading when assessing agreement between two
methods of measurement [14]. The Bland–Altman
approach to comparing methods of measurement esti-
mates the 95% confidence intervals for the ability of one
method to predict another [14]. Using this method, we
found that the ability of the AASI to predict PWV was
?4.18m/s. To put this in context, the standard deviation
in PWV in the study population was 2.2m/s. In clinical
terms,aPWVdifferenceof4m/shasbeenassociatedwith
a relative risk of stroke of1.72,or a 16–20% increasedrisk
of vascular events over 9years of follow-up [19,20].
Ambulatory arterial stiffness Jerrard-Dunne et al.
531
Table 1
Study population characteristics
Overall populationUntreated hypertensive individuals Individuals on treatment
n
Age (years)
Men (%)
Mean office arterial pressure (mmHg)
Office systolic blood pressure (mmHg)
Office diastolic blood pressure (mmHg)
Office heart rate (beats/min)
Office pulse pressure (mmHg)
Body height (cm)
Body weight (kg)
Body mass index (kg/m2)
24-h systolic blood pressure (mmHg)
24-h diastolic blood pressure (mmHg)
Dippers [n (%)]
Ambulatory arterial stiffness index
Augmentation index (%)
Pulse wave velocity (m/s)
824
51 (14)
49
114 (14)
155 (22)
91 (11)
68 (12)
65 (18)
168 (11)
84 (19)
29 (5)
135 (14)
81 (10)
678 (82)
0.35 (0.17)
30 (13)
10.5 (2.2)
426
46 (13)
51
114 (14)
153 (20)
92 (10)
70 (12)
61 (16)
169 (11)
83 (18)
29 (5)
135 (14)
83 (10)
366 (86)
0.33 (0.17)
29 (13)
10.2 (2.3)
398
56 (13)
46
114 (15)
158 (23)
90 (11)
67 (13)
68 (19)
167 (12)
84 (20)
29 (6)
135 (14)
79 (10)
312 (78)
0.38 (0.17)
32 (13)
10.9 (2.1)
Data presented as the mean (standard deviation) unless stated otherwise.

Page 4

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
A confidence interval greater than 4m/s therefore
suggests that the two methods would not be interchange-
able in a clinical setting.
In this relatively large population of hypertensive indi-
viduals, we did find positive correlations between the
AASI, PWV and AIx, but the correlations were consider-
ably weaker than those previously reported and were not
independent of age and gender [2]. This is in agreement
with the recently published data from Schillaci et al. [10].
In their original study, Li et al. [2] reported a correlation
between the PWV and AASI of r¼0.51, based on
measurements taken in 166 healthy normotensive volun-
teers, and a correlation between the AIx and AASI of
r¼0.48, based on a predominantly normotensive (67%)
Chinesecommunitypopulation(n¼348).Possibleexpla-
nations for this disparity include differences in study
population (normotensive versus hypertensive; Chinese
versus European) and methodology (Li et al. used a
different methodology (Sphygmocor) to measure caro-
tid–femoral PWV). Age has been shown in numerous
studies to be the most potent predictor of arterial stiff-
ness. Although Li et al. [2] reported correlations between
the AASI and arterial stiffness in individuals above and
below 40years of age, no data adjusting for age as a
continuous variable were presented. Similarly for the
AIx, adjustments were made for the heart rate and body
height but not age or mean arterial pressure. Adjustment
for age in the current study negated entirely the positive
correlation between the AASI, PWV and AIx.
The frequency of measurements used in this study (two
measurements per hour during the day and one measure-
ment per hour during the night) must be considered a
potential study limitation. The monitors were set to
record at a lesser frequency during the night to avoid
sleep disturbance. Current practice guidelines, however,
suggest an appropriate measurement frequency of 20–
30min throughout the 24-h period. While a greater
number of readings and consistent intervals for both
day and night would be desirable to improve the accuracy
of readings, we can see no mathematical reason why the
different intervals should materially influence the AASI.
As has been shown previously [2], the correlations
between the three stiffness measurements and risk fac-
tors demonstrate certain similarities between the AASI
and the AIx, both of which were significantly influenced
by gender and body habitus. These factors did not
correlate significantly with PWV, whereas the heart rate
influenced the AIx and PWV but not the AASI, and age
and mean arterial pressure were common correlates for all
three measurements.
In spite of this lack of correlation and agreement between
the various methods, it is important to recognize that
532
Journal of Hypertension
2008, Vol 26 No 3
Table 2
augmentation index and pulse wave velocity
Univariate Pearson correlations between risk factors and arterial stiffness measures: ambulatory arterial stiffness index,
Ambulatory arterial stiffness index
(n¼1714)
Augmentation index
(n¼824)
Pulse wave velocity
(n¼622)
UnivariateCorrelation coefficient
P valueCorrelation coefficient
P valueCorrelation coefficient
P value
Age
Gender
Mean arterial pressure (mmHg)
Systolic blood pressure (mmHg)
Diastolic blood pressure (mmHg)
Pulse pressure (mmHg)
Nocturnal reduction in systolic blood pressure (%)
Nocturnal reduction in diastolic blood pressure (%)
Heart rate (beats/min)
Height (cm)
Weight (kg)
Body mass index (kg/m2)
0.432
?0.075
0.170
0.317
?0.005
0.397
?0.274
?0.434
?0.037
?0.184
?0.097
?0.029
<0.001
0.019
<0.001
<0.001
0.884
<0.001
<0.001
<0.001
0.265
<0.001
0.003
0.383
0.454
?0.317
0.336
0.293
0.154
0.271
0.000
?0.035
?0.382
?0.358
?0.336
?0.180
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
0.999
0.313
<0.001
<0.001
<0.001
<0.001
0.525
0.014
0.414
0.492
0.222
0.484
?0.081
?0.06
0.131
?0.059
0.031
0.096
<0.001
0.720
<0.001
<0.001
<0.001
<0.001
0.045
0.017
<0.001
0.141
0.438
0.016
Table 3
index and pulse wave velocity
Multivariate correlations between risk factors and arterial stiffness measures: ambulatory arterial stiffness index, augmentation
Ambulatory arterial stiffness index
(n¼1714)
Augmentation index
(n¼824)
Pulse wave velocity
(n¼622)
Multivariate Correlation coefficient
P value Correlation coefficient
P value Correlation coefficient
P value
Age
Gender
Mean arterial pressure (mmHg)
Heart rate (beats/min)
Height (cm)
0.398
0.016
0.069
?0.003
?0.097
<0.001
0.689
0.029
0.921
0.014
0.262
?0.242
?0.419
0.334
?0.162
<0.001
<0.001
<0.001
<0.001
<0.001
0.491
0.016
0.269
0.167
0.058
<0.001
0.699
<0.001
<0.001
0.168

Page 5

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
there is published evidence for each of these measures as
independent predictors of vascular outcomes [3–5]. To
date, there are no direct comparative studies of outcome
prediction for the AASI versus PWV or the AIx, although
the available prospective data demonstrate important
differences with regard to risk prediction. Specifically,
the AASI appears to be an independent predictor of
stroke mortality [3,5], incident stroke [4] and overall
cardiovascular mortality [3,5], but not of cardiac death
[3,5] or incident coronary heart disease [4]. In contrast,
PWV has been shown to predict both stroke [19] and
coronary heart disease outcomes [12,20–22]. Prospective
data on the AIx and outcomes are limited to smaller
studies in specific patient groups, but show associations
withall-cause mortality and cardiovascularmortality [23–
25]. These prospective data further support the argument
that these different measures of arterial stiffness are not
freely interchangeable as vascular risk predictors.
The relationship between the AASI and vascular out-
comes is further complicated by the finding of varying
relationships in different populations. In the Irish and
Danish studies [3,4], the relationship between the AASI
and cardiovascular mortality was linear. In contrast,
Kikuya et al. [5] found a U-shaped relationship between
the AASI and cardiovascular mortality in their somewhat
older Japanese population. This complicates the clinical
interpretationoftheAASI forrisk-evaluation purposes,as
the same AASI measurement could be interpreted very
differently depending on the reference population used
to define risk.
In summary, the AASI has been proposed as a readily
available alternative to more costly and specialized
measures of arterial stiffness [18]. The present study
demonstrates that, in this hypertensive population,
there is poor agreement between the AASI, PWV and
AIx. Compared with the original reports suggesting
strong correlations between these methods in predomi-
nantly normotensive populations, it appears that these
relationships are considerably weaker in hypertensive
patients, whether treated or untreated, to the extent that
all correlations were lost following adjustment for age
and gender. The emerging outcome data also suggest
disparityintheabilityofthesemeasurestopredictfuture
vascular events. Direct comparative studies would be
required to establish the relative predictive strength of
each measure and whether combining measures can
provide additional risk prediction. Until such data
become available, we would propose that the measures
should not be considered interchangeable.
Acknowledgement
There are no conflicts of interest.
References
1 Dolan E, Li Y, Thijs L, McCormack P, Staessen JA, O’Brien E, et al.
Ambulatory arterialstiffness index:rationaleandmethodology.BloodPress
Monit 2006; 11:103–105.
2Li Y, Wang JG, Dolan E, Gao PJ, Guo HF, Nawrot T, et al. Ambulatory
arterial stiffness index derived from 24-h ambulatory blood pressure
monitoring. Hypertension 2006; 47:359–364.
3Dolan E, Thijs L, Li Y, Atkins N, McCormack P, McClory S, et al. Ambulatory
arterial stiffness index as a predictor of cardiovascular mortality in the
Dublin Outcome Study. Hypertension 2006; 47:365–370.
Ambulatory arterial stiffness Jerrard-Dunne et al.
533
Table 4
Correlations between the ambulatory arterial stiffness index, augmentation index and pulse wave velocity
Augmentation index Pulse wave velocity
Correlation coefficient
P value Correlation coefficient
P value
All individuals
Unadjusted
Age and sex
Multivariatea
Untreated hypertensive patients
Unadjusted
Age and sex
Multivariatea
Treated hypertensive patients
Unadjusted
Age and sex
Multivariatea
n¼824
n¼622
0.240
0.041
?0.006
<0.001
0.217
0.818
0.284
0.067
0.059
<0.001
0.077
0.095
n¼426
n¼391
0.182
0.005
?0.012
<0.001
0.928
0.865
0.253
0.096
0.094
<0.001
0.081
0.102
n¼398
n¼231
0.275
0.099
0.009
<0.001
0.055
0.882
0.292
0.042
0.023
<0.001
0.530
0.766
aAdjusted for age, gender, heart rate, mean arterial pressure and height.
Table 5
augmentation index and pulse wave velocity
Prediction limits of agreement between methods by Bland–Altman analysis for the ambulatory arterial stiffness index,
AASI 95% prediction limitsAll individuals Untreated hypertension patients Treated hypertension patients
Pulse wave velocity (m/s)
Augmentation index (%)
?4.18
?25.4
?4.32
?25.7
?3.86
?24.9