Hypertension XCEL PWV validation

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
CE: Namrta; HJH/203117; Total nos of Pages: 7;
HJH 203117
Carotid-femoral pulse wave velocity assessment using
novel cuff-based techniques: comparison with
tonometric measurement
Mark Butlina
, Ahmad Qasema
, Francesca Battistab
, Erwan Bozecc
, Carmel M. McEnieryd
,
Euge´nie Millet-Amauryc
, Giacomo Puccib
, Ian B. Wilkinsond
, Giuseppe Schillacib
,
Pierre Boutouyriec
, and Alberto P. Avolioa
Background: Carotid-femoral pulse wave velocity, a
predictor of cardiovascular outcome, is conventionally
measured using a tonometer sequentially placed upon the
carotid and femoral arteries, gated using an
electrocardiogram. Leg cuff detection of the femoral pulse
removes the need for signal gating, reduces the time
required for a single measurement, but gives different
pulse wave velocity values to tonometric analysis. A novel
algorithm to correct for the transit time and distance
related to the additional femoral segment was applied to
the cuff-based approach in this study.
Method: Eighty-eight individuals were recruited across
four centres and carotid-femoral pulse wave velocity
measured in triplicate using two operators with both a
tonometer-based device and a device using an inflated
thigh cuff with and without the use of the novel
algorithm. Comparison was made by Bland–Altman and
regression analysis.
Results: The unadjusted cuff-based approach gave lower
pulse wave velocity values than the tonometer-based
approach (6.11 Æ 1.27 and 7.02 Æ 1.88 m/s, P < 0.001).
With application of the algorithm, the cuff-based device
gave similar pulse wave velocity values (7.04 Æ 1.72 m/s) as
the tonometer-based approach (P ¼ 0.86). Analysis of
covariance with age showed a difference between the
tonometer and cuff-based methods (P < 0.001), with a
dependence upon age (P ¼ 0.004). The adjusted cuff-based
method gave similar results to the tonometer-based
method (P ¼ 0.94), with no dependence upon age
(P ¼ 0.46).
Conclusion: This study provided validation of a cuff-based
assessment of carotid-femoral pulse wave velocity against
the universally accepted tonometric method. Adjusting the
cuff-based method for the additional femoral segment
measured gives results comparable to the tonometer-based
method, for which the majority of population data exist to
date.
Keywords: aortic stiffness, arterial compliance, device
validation, hemodynamics, measurement, pulse wave
velocity
Abbreviation: cfPWV, carotid-femoral pulse wave velocity
INTRODUCTION
C
arotid-femoral pulse wave velocity (cfPWV) is
increasingly used in population studies [1–3] and
is recommended by the European Society of
Hypertension for use in the management of hypertension
[4]. As a result, there is a demand for devices that accurately
and easily measure cfPWV. The majority of cfPWV popu-
lation data have been collected using sequential recording
of the carotid and femoral pulse using applanation tono-
metry [2,5]. The two waveforms are synchronized using a
simultaneously recorded ECG. This technique is widely
used but does require a certain degree of operator skill
to accurately acquire the pulses. Applanation tonometry
also takes some time to acquire the two signals sequentially
and is intrusive in that it requires palpation of the femoral
pulse near the groin.
The use of an inflated cuff around the upper leg to
acquire the femoral pulse permits hands free, minimally
intrusive, user-independent acquisition of the femoral
pulse, simultaneously with carotid tonometer applanation.
Simultaneous, in place of sequential measurement of the
pulses, along with elimination of the use of ECG leads, is
estimated to reduce the time of measurement by slightly
more than 50%. However, previous results have shown that
leg cuff based measurement of cfPWV underestimates PWV
compared with the tonometer-based technique, especially
at higher values of cfPWV [6]. This may be due to the
inclusion of a longer femoral segment of artery in the
cuff-based approach (Fig. 1), the cuff being lower down
the leg than the inguinal site of tonometric acquisition of the
Journal of Hypertension 2013, 31:000–000
a
The Australian School of Advanced Medicine, Macquarie University, Sydney, New
South Wales, Australia, b
Universita´ degli Studi di Perugia and Struttura Complessa di
Medicina Interna, Ospedale ‘S. Maria’, Terni, Italy, c
Universite´ Paris Descartes, INSERM,
UMR970, Assistance Publique-Hoˆ pitaux de Paris, Hoˆ pital Europe´en Georges Pompi-
dou, De´partent of Pharmacologie, Paris, France and d
Clinical Pharmacology Unit,
University of Cambridge, Addenbrooke’s Hospital, Cambridge, UK
Correspondence to Mark Butlin, The Australian School of Advanced Medicine,
Macquarie University, 2 Technology Drive, Sydney, NSW 2109, Australia. Tel: +61
2 9812 3500; fax: +61 2 9812 3600; e-mail: mark.butlin@mq.edu.au
Received 18 March 2013 Revised 15 May 2013 Accepted 31 May 2013
J Hypertens 31:000–000 ß 2013 Wolters Kluwer Health | Lippincott Williams &
Wilkins.
DOI:10.1097/HJH.0b013e328363c789
Journal of Hypertension www.jhypertension.com 1
Original Article

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
HJH 203117
femoral pulse. This study aimed to investigate the cuff-
based technique of cfPWV measurement and applied an
algorithm to adjust for the additional femoral segment
measured. Both techniques were compared with the
tonometer-based technique. Results of this study constitute
a validation of the novel SphygmoCor XCEL (AtCor
Medical, Sydney, New South Wales, Australia) cuff-based
device that employs such an algorithm to adjust for the
femoral segment measured.
MATERIALS AND METHODS
Participants
The method for comparison of cfPWV techniques adhered
to the ARTERY Society guidelines [7]. Individuals recruited
for the study were equally distributed across three age
ranges (18–30, 30–60 and greater than 60 years) with at
least 25 individuals in each range, each containing a
minimum of 40% men and 40% women. Participants were
excluded if they were pacemaker dependent, not in sinus
rhythm, pregnant, had a BMI greater than 30 kg/m2
or had
known carotid or femoral artery stenosis. Recruitment
and measurement was conducted across four centres:
Macquarie University, Sydney, Australia; University of Peru-
gia, Perugia, Italy; Hoˆpital Europe´en Georges Pompidou,
Universite´ Paris Descartes, Paris, France; Addenbrooke’s
Hospital, Cambridge University, Cambridge, UK. Ethics
approval was received from the ethics bodies of the
respective institutions and the study was conducted with
the understanding and the consent of each participant.
Carotid-femoral pulse wave velocity
measurement
The standard for comparison was taken as cfPWV measured
by the conventional or ‘classic’ SphygmoCor device
(software version 9; AtCor Medical), which allows sequen-
tial recording of the pulse at the carotid and femoral site by
applanation tonometry, gating these two signals using the
QRS complex of a simultaneously recorded ECG. The
SphygmoCor XCEL device (software version 1; AtCor
Medical) records the cfPWV between the diastolic foot of
the carotid waveform, recorded by tonometric applanation,
and the diastolic foot of the simultaneously recorded
femoral pulse, detected using the volume displacement
waveform measured in a cuff placed around the upper
thigh, inflated to subdiastolic pressure.
The diastolic foot of the femoral and carotid waveforms
in both the XCEL and the tonometer-based PWV assessment
devices was defined as the point of intersection of lines of fit
during late diastole and early systole, often referred to as
the intersecting tangents method. The transit time (t) of the
pulse was measured as the time between the diastolic feet
of the carotid and femoral pulse for the cuff-based device
(tC) and the tonometer-based device (tT).
Arterial path length (Fig. 1) for the cuff-based assessment
of cfPWV (PWVC) was estimated by measuring the linear
distance from the suprasternal notch to the top of the cuff
at the centre line of the leg, approximately at the location of
the femoral artery (dsfC), and subtracting the distance from
the suprasternal notch to the location wherein the carotid
pulse could be palpated (dsc). Distance for the tonometer-
only based assessment of cfPWV (PWVT) was measured as
the linear distance from the suprasternal notch to the site
where the femoral pulse could be palpated and measured
with the tonometer (dsfT), minus the linear distance from
the suprasternal notch to the site where the carotid pulse
could be palpated (dsc). cfPWV was calculated as the
measured distance divided by the transit time (Eqs. 1–2).
PWV C ¼
dsfC À dsc
tC
(1)
PWV T ¼
dsfT À dsc
tT
(2)
An algorithm, built into the SphygmoCor XCEL device,
corrects for two possible sources of difference in cfPWV
measurement from the tonometer-based technique. First,
the transit time was reduced by a constant (k1) associated
with the delay in the transmission and registration of the
volume displacement pulse from the cuff to the pressure
sensor and the recording circuitry in the device. The value
of k1 was ascertained by AtCor Medical. Second, the transit
time and distance were reduced in an attempt to subtract
the segment of artery between the femoral site wherein
a tonometer is usually used and the location of the leg cuff.
The distance was reduced by an operator-measured
distance from the site wherein the femoral pulse can be
palpated to the top of the cuff (dfTfC, Fig. 1). The transit time
was reduced by a constant factor (k2) multiplied by this
femoral distance. The constant factor, k2, was derived from
an average femoral transit time per unit distance measured
in a separate group of 15 individuals (six women, 34 Æ 10
years of age, 24–64 years old). The adjusted cfPWV
(PWVCA) was calculated according to Eq. (3).
PWV CA ¼
dsfC À dsc À dfTfC
tC À k1 À k2 Á dfTfC
(3)
C
S
d
d
fT
fC
sc
sfT
dsfC
dfTfC
FIGURE 1 Diagram showing distance measurements on the body for carotid-fem-
oral pulse wave velocity calculation and the additional segment measured when
using a cuff to acquire the femoral pulse (dfTfC). The linear distances (d) were
measured between the suprasternal notch (s), the site wherein the carotid pulse
could be palpated (c), the site wherein the femoral pulse could be palpated and
measured with a tonometer (fT), and the top of the thigh cuff approximately
above the location of the femoral artery (fC).
Butlin et al.
2 www.jhypertension.com Volume 31 Number 00 Month 2013

Copyright © Lippincott Williams Wilkins. Unauthorized reproduction of this article is prohibited.
HJH 203117
cfPWV measurements were taken in triplicate using each
device; two operators used alternating devices in a pattern
such that both operators used both devices throughout
the study [7]. The three measurements taken with each
device were averaged to give a single cfPWV reading
for each individual from each device. Blood pressure,
as measured using a brachial oscillometric cuff, and
heart rate were recorded before and after cfPWV
measurements to ensure that blood pressure and heart
rate had not significantly changed during the cfPWV
measurements.
Statistical analysis
Bland–Altman plots [8] were generated for comparison
of the tonometer-based method to both the thigh cuff-
based method, and the adjusted thigh cuff-based method.
One-way, repeated measures analysis of variance (ANOVA)
with posthoc Bonferroni adjusted pairwise t-test was
conducted to detect differences between the three methods
of measurement. The linear correlation coefficient, the
intercept within a 95% confidence interval and the root
mean square error (RMSE) of regression between the two
devices were calculated. Past studies of cuff-based tech-
niques showed a nonlinear relationship to the tonometer-
based approach [6]. Therefore, a second-order polynomial
was also fitted to the data as a means of quantifying any
nonlinearity.
Analysis of covariance (ANCOVA) was used to test
whether the adjusted cuff-based method offered a signifi-
cantly different result to the unadjusted cuff-based method,
as compared with the tonometer-based cfPWV, accounting
for any interaction of the choice of algorithm with increas-
ing values of cfPWV.
The mean difference and standard deviation (SD) of
the difference between methods was calculated. Using
the mean and SD of the difference between devices, the
cuff-based and adjusted cuff-based methods were then
graded as either ‘excellent’, ‘acceptable’ or ‘poor’, according
to the cut-off values proposed in the ARTERY Society
Guidelines [7].
The concept that cfPWV increases with age permitted
further investigation with regard to whether there was a
significant underestimation of cfPWV using the cuff-based
methods at an older age and higher cfPWV. ANCOVA was
used to test the interaction between the method of cfPWV
measurement and age.
Statistical analysis was conducted using the statistical
package, R (Version 2.15.0) [9]. All results are given as
mean Æ SD, unless otherwise stated.
RESULTS
Ninety-eight individuals were recruited across the four
centres (29 from Sydney, 41 from Perugia, 22 from Paris,
six from Cambridge). As required by the ARTERY Society
guidelines, individuals with a BMI greater than 30 kg/m2
were excluded (n ¼ 10). The remaining 88 individuals
(45 male) had a mean age of 46 Æ 20 years, were evenly
distributed amongst the three designated age groups and
had an even balance of men and women in each group
(Table 1).
The mean transit time for the cuff-based approach was
99 Æ 22 ms and for the tonometer-based approach was
67 Æ 14 ms, a mean difference of 32 Æ 12 ms (P 0.001).
The mean transit time for the adjusted cuff-based approach
was 67 Æ 13 ms, the mean difference from the tonometer-
based approach being 1 Æ 6 ms (P ¼ 0.10). The application
of the adjustment for the femoral segment to the cuff-based
approach increased the correlation to the tonometer-
based system (cuff-based approach R ¼ 0.87, cuff-based
approach-adjusted R ¼ 0.92, Fig. 2a,b, Table 2), with a
decrease in slope from 1.28 Æ 0.08 to 0.82 Æ 0.04, and
intercept from 13 Æ 5 to 11 Æ 3 ms, the difference in these
regressions being significant (ANCOVA, P 0.001). There
was also an interaction with increasing values of transit
time (P 0.001), indicating a divergence in the two
techniques with greater values of transit time.
The mean cfPWV for the cuff-based approach was
6.11 Æ 1.27 m/s and for the tonometer-based approach
was 7.02 Æ 1.88 m/s, the mean difference being 0.91 Æ
0.84 m/s (P 0.001). According to the ARTERY Society
guidelines, the accuracy of the cuff-based approach was
‘acceptable’ (mean difference less than 1.0 m/s, SD of
difference less than 1.5 m/s). After applying the adjustment
for the femoral segment to the cuff-based approach,
the mean cfPWV was 7.04 Æ 1.72 m/s, with an average
difference from the tonometer-based approach of
0.01 Æ 0.71 m/s (P ¼ 0.86). According to the ARTERY
Society guidelines, the accuracy of the adjusted cuff-based
approach was ‘excellent’ (mean difference less than
0.5 m/s, SD of difference less than 0.8 m/s). The application
of the adjustment for the femoral segment to the cuff-based
TABLE 1. Individual demographics across the three age ranges
Parameter All
18–30 30–60 60
years years years
n 88 31 31 26
Age (years) 46 Æ 20 26 Æ 2 44 Æ 10 72 Æ 7
Age (min/max, years) 22/83 22/29 30/60 61/83
Male 45 (51%) 14 (45%) 18 (58%) 13 (50%)
Height (m) 1.7 Æ 0.1 1.7 Æ 0.1 1.7 Æ 0.1 1.7 Æ 0.1
Weight (kg) 68 Æ 12 67 Æ 12 68 Æ 12 69 Æ 10
BMI (kg/m2
) 23 Æ 3 23 Æ 2 23 Æ 3 25 Æ 3
SBP (mmHg) 126 Æ 19 119 Æ 12 126 Æ 21 135 Æ 19
DBP (mmHg) 73 Æ 10 70 Æ 9 76 Æ 12 73 Æ 8
SBP and DBP were measured using an oscillometric brachial cuff device with individuals in the supine position.
Cuff-based pulse wave velocity measurement

HJH 203117
(a)
40 60 80 100 120 140 160
40
60
80
100
120
140
160
T, femoral tonometer (ms)
T,thighcuff(ms)
40 60 80 100 120 140 160
0
20
40
60
Devices mean t (ms)
ΔTbetweendevices(ms)
(c)
(d)
4
4 6 8 10 12
6
8
10
12
PWV, femoral tonometer (m/s)
PWV,thighcuff(m/s)
4 6 8 10 12
−4
−2
0
2
Devices mean PWV (m/s)
ΔPWVbetweendevices(m/s)
FIGURE 2 Comparison of cuff-based results (gray) and cuff-based results adjusted for the additional femoral segment (black) compared with the tonometer-based method
for both transit time (left panels) and cfPWV (right panels). Linear regression statistics are presented in Table 2. Fitted second-order polynomials have been applied by least
squares regression to capture the movement of the cfPWV measurement away from linearity with increasing cfPWV. PWVC ¼ À0:03 Â PWV2
T þ 1:11 Â PWVT À 0:07m=s
(R ¼ 0.94, P 0.001); PWVCA ¼ À0:04 Â PWV2
T þ 1:48 Â PWVT À 1:28m=s (R ¼ 0.93, P 0.001).
TABLE 2. Results for fitting of a linear relationship by least squares regression to each of the datasets displayed in transit time (t) and
carotid-femoral pulse wave velocity (Fig. 2) data for the cuff-based (C) and the adjusted cuff-based approach (CA), including
regression for the between-device differences (Dt, DPWV)
Slope Intercept Intercept 95% confidence interval R RMSE P
Tonometer-based transit time against:
tC 1.28 Æ 0.08 13 Æ 5 3–23 0.87 10 0.001
tCA 0.82 Æ 0.04 11 Æ 3 5–17 0.92 5 0.001
Tonometer-based cfPWV against:
PWVC 0.63 Æ 0.03 1.70 Æ 0.19 1.33–2.07 0.93 0.46 0.001
PWVCA 0.85 Æ 0.04 1.08 Æ 0.27 À0.45–0.61 0.93 0.64 0.001
Device mean transit time against:
DtC 0.40 Æ 0.05 À1 Æ 5 À11–9 0.62 9 0.001
DtCA À0.12 Æ 0.04 7 Æ 3 1–13 À0.27 6 0.01
Device mean PWV against:
DPWVC À0.40 Æ 0.04 1.72 Æ 0.27 1.19–2.25 À0.74 0.56 0.001
DPWVCA À0.09 Æ 0.04 0.66 Æ 0.30 0.07–1.25 À0.23 0.69 0.03
Data are presented as mean Æ standard error with intercept, intercept confidence intervals and RMSE values given in ms for transit time measurements, and m/s for cfPWV
measurements. The sample 95% confidence interval was calculated as the mean Æ1.96 times the standard error. cfPWV, carotid-femoral pulse wave velocity; PWV, pulse wave velocity;
RMSE, root mean square error.
Butlin et al.

HJH 203117
approach did not alter the correlation with the tonometer-
based system (R ¼ 0.93 in both cases, Fig. 2c,d, Table 2).
However, the slope of regression moved closer to linearity
(0.63 Æ 0.03–0.85 Æ 0.04) coupled with a decrease in the
intercept (1.70 Æ 0.19–1.08 Æ 0.27 m/s), the difference in
these regressions being significant (ANCOVA, P 0.001).
There was also an interaction with increasing values
of cfPWV (P 0.001), indicating a divergence in the two
techniques with greater values of cfPWV.
There was a significant effect of age on PWV (P 0.001),
a PWV method related effect (P 0.001), and an interaction
effect between age and the method of PWV analysis chosen
(P ¼ 0.01, Fig. 3). This was evident for the cuff-based
approach compared with the tonometer-based approach
for both the method (P 0.001), and the method interaction
with age (P ¼ 0.004), when compared with the tonometer-
based approach. However, this was not evident for the
adjusted cuff-based approach for either the method
(P ¼ 0.94) or the method interaction with age (P ¼ 0.46)
when compared with the tonometer-based approach.
The age/cfPWV relationship was well fitted by linear
regression (Fig. 3), similar to other studies of similar size
[6,10] but not showing the second-order relationship seen
in larger populations studies [2].
DISCUSSION
This study aimed to validate a method of adjustment of cuff-
based cfPWV assessment that accounts for the effect of
the additional femoral segment measured. Tonometer-
based cfPWV was chosen as the standard for comparison,
as the majority of population data to date [1–3] have been
collected using a tonometer-based approach, and equi-
valence to these data is a necessary requirement in order
for cuff-based cfPWV measurements to be meaningful and
useful in risk assessment. The correlation between cfPWV,
in general, and aortic PWV was not addressed in this
study, but has been quantified in purpose-designed
studies employing phase-contrast MRI [10] or invasive aortic
pressure measurement [11]. As advised by the ARTERY
Society guidelines [7], tonometer-based assessment of
cfPWV was chosen as the standard of reference, as it is
this parameter, not aortic stiffness per se, that has been used
in large population studies and in the clinic.
Using a volume displacement cuff around the thigh
for recording of the femoral pulse resulted in an under-
estimation of cfPWV compared with the use of a tonometer
to detect the femoral pulse, especially at higher values
of cfPWV (Fig. 2c,d, Table 2). This underestimation, also
found in other cuff-based cfPWV measurement systems [6],
could originate from either a difference in the algorithm that
detects the waveform foot, critical differences in waveform
shape or transmission due to the use of a cuff instead of a
tonometer, or anatomical differences due to the change in
site at which the femoral pulse is detected.
In a validation study conducted by Millasseau et al. [12]
investigating the Complior device, the foot detection
algorithm was shown to be the cause of a comparative
bias in the measurement of cfPWV. However, the current
study used the same algorithm for both the tonometer-
based and cuff-based cfPWV assessment, eliminating this as
a source of the difference in cfPWV.
The waveform shape of the pulse measured by the cuff is
significantly different from the waveform measured using
the tonometer. This is in part not only due to the measure-
ment of the pulse at a different site but also due to the
inherent damping characteristics of the volume displace-
ment method of pulse acquisition. However, the late
diastolic and early systolic components of the waveform
are not vastly different between the two techniques (Fig. 4)
and did not drastically alter the detected position of
the foot.
The adjustment technique used a constant (k2) femoral
PWV, derived from 15 individuals, of the order of existing
published femoral PWV values [13]. This adjustment tech-
nique assumes the same femoral PWV for all individuals,
independent of sex, age and blood pressure. A previous
30 40 50 60 70 80
5
6
7
8
9
10
11
Age (years)
PWV(m/s)
Tonometer-based method
Cuff-based method, adjusted
Cuff-based method
FIGURE 3 Carotid-femoral pulse wave velocity, as measured by the three methods,
plotted against age. ANCOVA confirmed that the cuff-based method was offset
from the tonometer-based method (P 0.001) with an interaction with age
(P ¼ 0.004), whereas the adjusted cuff-based method showed a similar result to
the tonometer-based method (P ¼ 0.94) with no interaction with age (P ¼ 0.46).
Regression and ANCOVA was conducted on the individual data, not age-averaged
points as plotted.
0.0 0.1 0.2 0.3 0.4
0.0
0.2
0.4
0.6
0.8
1.0
1.2
Magnitude(arbitraryunits)
c
Time (s)
b
a
FIGURE 4 The foot of the femoral pulse as recorded by volume displacement
with a thigh cuff (solid line) and by applanation tonometry (dashed line, time
shifted to align the diastolic feet of the waves), highlighting the similarity in the
shape of the waveform foot. Waveforms shown are from individuals in each
of the three nominal age groups: (a) 79-year-old man, BMI 20 kg/m2
, cfPWV
10.2 m/s; (b) a 45-year-old man, BMI 17 kg/m2
, cfPWV 7.6 m/s; (c) a 26-year-old
woman, BMI 25 kg/m2
, cfPWV 5.2 m/s.

HJH 203117
study showed that PWV in the leg changes little with age, on
average increasing 0.06 m/s per year [13], the same trend
being confirmed in other peripheral arterial beds such as
the arm [13,14]. However, it is known that there are ethnic
and disease-related differences in the arteries of the leg [15].
A population-based look-up table, or measurement of
individual femoral PWV, could improve correlation of
the adjusted cuff-based methods to tonometric results
and may be important for results in different ethnic groups
or disease subtypes. Femoral PWV is also dependent
upon the distending pressure, and this aspect too could
be individualized.
This is the first study to use the only guidelines to date for
the validation of devices that measure cfPWV. The ARTERY
Society guidelines [7] recommend the use of simultaneous
tonometric recording of carotid and femoral pulses as the
standard for comparison. However, in this study, the sec-
ondary reference cited by the guidelines, tonometry using
ECG gating, was used. The secondary reference was used,
as it enabled the study to be carried out in a multicentre
fashion, all centres holding the same cfPWV tonometer
device (SphygmoCor; AtCor Medical). Individuals with a
BMI greater than 30 kg/m2
were excluded from the analysis,
as recommended by the guidelines. The inclusion of these
10 individuals did not significantly alter any results (analysis
not shown). However, further studies are required to
ascertain whether the same results hold true within
such subpopulations.
The accuracy of measurement of arterial distances has
not been addressed in this study, as it has been in other
studies [11,16–18]. However, distances for both techniques
were measured in the same manner to avoid this confound-
ing factor. Inflation of the thigh cuff above the diastolic
pressure alters the pressure waveform during the diastolic
period by occlusion of the artery when the cuff pressure
exceeds the instantaneous arterial pressure. So as to avoid
the diastolic portion of the wave being altered in some
individuals, as would be the case if a constant cuff pressure
was applied to all individuals, the cuff pressure was indi-
vidualized below the measured brachial diastolic pressure.
This assumed that brachial and femoral diastolic pressures
were approximately equal.
Although there is no statistical difference between
cfPWV measured by tonometry and the adjusted cuff-based
method, the relationship between the two devices has a
slope less than unity. On average, the value of cuff-based
cfPWV is lower than tonometry-based assessment for
cfPWV greater than 7.2 m/s, and greater at values less than
7.2 m/s. For example, an individual with a tonometry-based
cfPWV assessment of 12 m/s will have an average-adjusted
cuff-based technique cfPWV reading of 11.28 m/s.
Although the adjusted cuff-based results are within the
ARTERY Society guideline ‘excellent’ category, the impact
of this difference in an individual measurement is depend-
ent upon treatment guidelines, and upon whether changes
to patient treatment are implemented using this value.
The effect in population-based studies is highlighted in
Fig. 3. Without any adjustment, the cuff-based approach
underestimated cfPWV compared with the conventional,
tonometer-based approach, and this effect was statistically
greater at higher cfPWV and thus also with greater age.
However, applying an adjustment correcting for the
additional femoral segment gives results similar to that
obtained using the tonometer-based approach.
cfPWV is accepted as a powerful predictor of cardio-
vascular disease and has traditionally been measured
using tonometry. However, the sequential measurement
used in tonometric measurement and the intrusive nature of
taking the femoral pulse limit the use in clinical settings.
Use of cuff-based acquisition of the femoral pulse permits
assessment of cfPWV in a shorter time with less intrusion.
However, the values are different than those when assessed
using the tonometric approach. Adjustment of the cuff-
based measured cfPWV for the additional femoral segment
measured gives results similar to those of the tonometric
approach. Therefore, the cuff-based approach, when
adjusted, provides a practical arterial stiffness assessment
tool for clinical use and screening, with cfPWV values able
to be compared with existing population data.
ACKNOWLEDGEMENTS
None.
Conflicts of interest
All sites, other than Cambridge University, received funding
from AtCor Medical, Australia, for the collection of cfPWV
data using the SphygmoCor ‘classic’ and SphygmoCor XCEL
devices. No restrictions were placed on the publication of
results. The values of the constants used in the algorithm for
adjustment of cfPWV could not be published, as they form
part of an integrated proprietary algorithm that is the subject
of a patent pending (AtCor Medical). A.Q. is an employee
of AtCor Medical, Australia, and a Visiting Fellow at
Macquarie University.
REFERENCES
1. Boutouyrie P, Tropeano AI, Asmar RG, Gautier I, Benetos A, Lacolley P,
Laurent S. Aortic stiffness is an independent predictor of primary
coronary events in hypertensive patients: a longitudinal study. Hyper-
tension 2002; 39:10–15.
2. McEniery CM, Yasmin, Hall IR, Qasem A, Wilkinson IB, Cockcroft JR.
Normal vascular aging: differential effects on wave reflection and aortic
pulse wave velocity: the Anglo-Cardiff Collaborative Trial (ACCT). J Am
Coll Cardiol 2005; 46:1753–1760.
3. Mitchell GF, Hwang SJ, Vasan RS, Larson MG, Pencina MJ, Hamburg
NM, et al. Arterial stiffness and cardiovascular events: the Framingham
Heart Study. Circulation 2010; 121:505–511.
4. Mancia G, De Backer G, Dominiczak A, Cifkova R, Fagard R,
Germano G, et al. 2007 guidelines for the management of arterial
hypertension: the task force for the management of arterial hyper-
tension of the European Society of Hypertension (ESH) and of the
European Society of Cardiology (ESC). J Hypertens 2007; 25:1105–
1187.
5. Adji A, O’Rourke MF, Namasivayam M. Arterial stiffness, its assessment,
prognostic value, and implications for treatment. Am J Hypertens 2011;
24:5–17.
6. Hickson SS, Butlin M, Broad J, Avolio AP, Wilkinson IB, McEniery CM.
Validity and repeatability of the Vicorder apparatus: a comparison
with the SphygmoCor device. Hypertens Res 2009; 32:1079–1085.
7. Wilkinson IB, McEniery CM, Schillaci G, Boutouyrie P, Segers P,
Donald A, Chowienczyk PJ. ARTERY Society guidelines for validation
of noninvasive haemodynamic measurement devices: Part 1, arterial
pulse wave velocity. Art Res 2010; 4:34–40.
8. Bland JM, Altman DG. Statistical methods for assessing agreement
between two methods of clinical measurement. Lancet 1986; 1:307–
310.
Butlin et al.

HJH 203117
9. R Core Team. R: A language and environment for statistical computing
Vienna. Austria: R Foundation for Statistical Computing; 2012.
10. Hickson SS, Butlin M, Graves M, Taviani V, Avolio AP, McEniery CM,
Wilkinson IB. The relationship of age with regional aortic stiffness and
diameter. J Am Coll Cardiol Cardiovasc Imaging 2010; 3:1247–
1255.
11. Weber T, Ammer M, Rammer M, Adji A, O’Rourke MF, Wassertheurer S,
et al. Noninvasive determination of carotid-femoral pulse wave
velocity depends critically on assessment of travel distance: a com-
parison with invasive measurement. J Hypertens 2009; 27:1624–1630.
12. Millasseau SC, Stewart AD, Patel SJ, Redwood SR. Chowienczyk PJ
Evaluation of carotid-femoral pulse wave velocity: influence of timing
algorithm and heart rate. Hypertension 2005; 45:222–226.
13. Avolio AP, Chen SG, Wang RP, Zhang CL, Li MF, O’Rourke MF. Effects
of aging on changing arterial compliance and left ventricular load in a
northern Chinese urban community. Circulation 1983; 68:50–58.
14. van der Heijden-Spek JJ, Staessen JA, Fagard RH, Hoeks AP, Boudier
HA, van Bortel LM. Effect of age on brachial artery wall properties
differs from the aorta and is gender dependent: a population study.
Hypertension 2000; 35:637–642.
15. Strain WD, Chaturvedi N, Dockery F, Shiff R, Shore AC, Bulpitt CJ,
Rajkumar C. Increased arterial stiffness in Europeans and African
Caribbeans with type 2 diabetes cannot be accounted for by conven-
tional cardiovascular risk factors. Am J Hypertens 2006; 19:889–896.
16. Girerd N, Legedz L, Paget V, Rabilloud M, Milon H, Bricca G, Lantelme
P. Outcome associations of carotid-femoral pulse wave velocity vary
with different measurement methods. Am J Hypertens 2012; 25:1264–
1270.
17. Huybrechts SaM, Devos DG, Vermeersch SJ, Mahieu D, Achten E,
de Backer TLM, et al. Carotid to femoral pulse wave velocity: a
comparison of real travelled aortic path lengths determined by MRI
and superficial measurements. J Hypertens 2011; 29:1577–1582.
18. Laurent S, Cockcroft J, Van Bortel L, Boutouyrie P, Giannattasio C,
Hayoz D, et al. Expert consensus document on arterial stiffness:
methodological issues and clinical applications. Eur Heart J 2006;
27:2588–2605.
Reviewers’ Summary Evaluations
Reviewer 1
This is a clear and straightforward methodological study
on the – successful – validation of the new version of
the Sphygmocor system for carotid-femoral pulse wave
velocity measurements (using a tonometer on the carotid
and cuff for the femoral signal). The study is conducted
in compliance with guidelines that were provided by
the Artery Society and is exemplary for this type of studies.
The only – albeit minor – limitation is the fact that not
all methodological details (values of some parameters) can
be disclosed for commercial reasons.
Reviewer 2
The strengths of this paper relate to the excellent design
and methodology to examine tonometer-based versus
thigh cuff-based methods for determining carotid-femoral
pulse wave velocity. The authors convincingly validate the
adjustment algorithm for cuff-based methods against the
universally accepted tonometric methods. There are no
apparent weaknesses.

Hypertension XCEL PWV validation

More Related Content

What's hot (20)

Similar to Hypertension XCEL PWV validation (20)

Hypertension XCEL PWV validation