CE: Tripti; JH-D-19-00010; Total nos of Pages: 8;
JH-D-19-00010
Original Article
Accuracy of fully automated oscillometric central
aortic blood pressure measurement techniques
Michael Gotzmann a, Maximilian Hogeweg b, Felix S. Seibert c, Benjamin Johannes Rohn c,
Martin Bergbauer b, Nina Babel c, Frederic Bauer c, Andreas Mügge a, and Timm H. Westhoff c
Background: Central aortic blood pressure (cBP) is a
valuable predictor of cardiovascular risk. The lack of fully
automated measurement devices impeded an
implementation in daily clinical practice so far. The present
study compares two novel automated oscillometric devices
with invasively measured cBP.
Methods: From March 2017 to March 2018, we enrolled
consecutive patients undergoing elective coronary
angiography to this cross-sectional study. Noninvasive
assessment of cBP was performed by the SphygmoCor
XCEL device and the Mobil-O-Graph NG device
simultaneously to invasive measurement.
Results: Our study included 502 patients (228 women,
274 men) with a mean age of 67.9  11.6 years. The
noninvasive measurement of cBP was successful in 498
patients (99%) with SphygmoCor XCEL device and in 441
patients (88%) with Mobil-O-Graph NG device (P ¼ 0.451).
Measurements of both devices revealed a high correlation
to invasively measured systolic (SphygmoCor R2 0.864,
P < 0.001; Mobil-O-Graph R2 0.763, P < 0.001) and
diastolic (SphygmoCor R2 0.772, P < 0.001; Mobil-O-Graph
R2 0.618, P < 0.001) cBP. Both devices slightly
underestimated systolic and overestimated diastolic central
blood pressure: biases were 5.0  7.7/0.5  6.2 mmHg
with SphygmoCor XCEL and 6.0  10.4/3.6  8.3 mmHg
with Mobil-O-Graph NG device. Correlations (R2) were
higher and biases were lower with the SphygmoCor device
(P < 0.001 each).
Conclusion: The present study is the largest validation
study of noninvasive cBP measurement techniques so far
and shows that two current automated oscillometric
monitors are able to assess cBP with acceptable accuracy.
Automated oscillometric devices may facilitate the
implementation of cBP in daily clinical practice.
Keywords: central aortic blood pressure, hypertension,
measurement, oscillometric
Abbreviations: ASCOT, Anglo-Scandinavian Cardiac
Outcomes Trial; CAFÉ, Conduit Artery Functional Endpoint;
cBP, central aortic blood pressure
INTRODUCTION
H
ypertension is a cardiovascular risk factor of out-
standing importance and accounts for up to 50% of
vascular risk [1]. Blood pressure (BP) has been
Journal of Hypertension
Copyright © 2019 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
traditionally measured at peripheral arteries. However,
because of pulse pressure amplification, peripheral blood
pressure does not represent central blood pressure. In
recent years, there is increasing evidence that central aortic
blood pressure (cBP) might be superior to peripheral BP in
the prediction of cardiovascular events [2,3]. The close
association of cBP with target organ damage and cardio-
vascular risk has anatomical reasons: both coronary and
carotid arteries arise from the proximal aorta. The ASCOT
and its substudy CAFE demonstrated that the differential
effect of antihypertensive drugs on cBP and peripheral BP
may to some extent explain differences in their effects on
cardiovascular events [4,5].
In the 1990s, first noninvasive tonometric techniques
emerged that detected the peripheral arterial pressure
waveform and calculated cBP after additional measurement
of peripheral BP. Comparative analyses showed a satisfying
agreement with invasively measured cBP [6,7] For two
reasons, however, these systems never proceeded to daily
clinical practice. First, because of their high price and
second, because the procedure is time consuming and
requires well trained medical staff. Therefore, efforts have
been made to perform pulse contour analysis in an oscil-
lometric instead of a tonometric manner and thereby to
integrate it into automated cuff-based peripheral BP mea-
surement. Some first systems are now commercially avail-
able [8], but data on the accuracy of the systems compared
with the invasive measured cBP are scarce. Therefore, the
present study compares cBP measured noninvasively by
the SphygmoCor XCEL device (AtCor Medical, Sydney,
Australia) and Mobil-O-Graph NG device (I.E.M., Stolberg,
Germany) device to invasively measured aortic BP patients
undergoing coronary angiography.
Journal of Hypertension 2019, 37:000–000
a
University Hospital St Josef Hospital Bochum, Cardiology, Ruhr University Bochum,
b
Department of Internal Medicine, Marien Hospital Witten, Witten and cUniversity
Hospital Marien Hospital Herne, Medical Department 1, Ruhr University Bochum,
Herne, Germany
Correspondence to Timm H. Westhoff, MD, University Hospital Marien Hospital
Herne, Ruhr-University Bochum, Medical Department I, Hölkeskampring 40, 44625
Herne, Germany. Tel: +49 2323 499 1671; fax: +49 2323 499 3302; e-mail:
timm.westhoff@elisabethgruppe.de
Received 3 January 2019 Revised 28 June 2019 Accepted 8 August 2019
J Hypertens 37:000–000 Copyright ß 2019 Wolters Kluwer Health, Inc. All rights
reserved.
DOI:10.1097/HJH.0000000000002237
www.jhypertension.com 1
CE: Tripti; JH-D-19-00010; Total nos of Pages: 8;
JH-D-19-00010
Gotzmann et al.
METHODS
Study population
From March 2017 to March 2018, we enrolled consecutive
patients undergoing elective coronary angiography to this
cross-sectional study. Inclusion criteria were indication for
an elective coronary angiography, informed consent and
age at least 18 years. Instable angina, instable hemodynam-
ics, myocardial infarction, and a stenosis of the radial or
brachial artery were defined as exclusion criteria. Invasive
and noninvasive blood pressure measurements were per-
formed according to the recommendations of the ARTERY
Society Consensus statement on protocol standardization
[9]. Written informed consent was obtained from all partic-
ipants before inclusion in the study. The study was
approved by the local ethics committee of the Ruhr Univer-
sity Bochum (reg. number 16–5759).
Devices for noninvasive blood pressure
measurement
For this study, the following devices were used for nonin-
vasive blood pressure measurement: SphygmoCor XCEL
device (software version 1.2, AtCor Medical) and Mobil-O-
Graph NG device (software version HMS CS 5.1; I.E.M.).
The Mobil-O-Graph (software version HMS CS 5.1;
I.E.M.) is a noninvasive blood pressure recording device.
Oscillometric signals from the brachial artery are captured
using an inflatable cuff that is connected to a pressure
sensor. The device records pulse wave oscillations in order
to provide systolic, diastolic and mean blood pressure
values after a stepwise deflation process of the cuff. Cali-
bration is performed using systolic and diastolic pressure.
This system calculates central blood pressure, aortic pulse
wave velocity and additional central hemodynamic indices,
all based on the oscillometric recording of pulse waves at
the brachial artery site. In a first step, the recorded pulse
waves are checked for their plausibility and qualified
according to predefined quality criteria. Once the input
signal has qualified as acceptable, the concepts of wave
harmonics and Fourier analysis are applied for the calcula-
tion of an aortic pressure curve. In parallel, an aortic flow
curve is derived, by means of a modified Windkessel
model. Consecutively, the pressure and flow curves are
superimposed in a virtual model. The mathematical model
applied by the Mobil-O-Graph system is based on the
equations as set forth by Moens-Korteweg and subse-
quently further developed by Bramwell-Hill, dealing with
the interrelation of velocity and wall elasticity in a
closed system.
The SphygmoCor XCEL (software version 1.2, AtCor
Medical) measures brachial blood pressure using the stan-
dard cuff oscillometeric method (NIBP) to determine bra-
chial SBP and DBP, which are used for calibration. After the
cuff deflates to zero pressure, the cuff inflates again to a set
pressure below brachial DBP. The device measures the cuff
oscillations that represent the brachial volume displace-
ment waveform. The recorded brachial cuff waveform is
calibrated to brachial SBP and DBP. Then the calibrated
brachial waveform is converted to a calibrated central
pressure waveform using a transfer function that represents
the harmonics ratio between the two pulses. Pulses are
2 www.jhypertension.com
Copyright © 2019 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
detected from the central pressure waveform and an aver-
age central pressure is calculated. The maximum and
minimum of the average central pulse represent the average
central SBP and DBP, respectively. In addition, all of the
central pulse wave analysis parameters are calculated from
the average central pressure pulse.
Process of familiarization with the equipment
Before the start of the study, an introduction was given to
the devices by experienced investigators. The practical
application was tested on 20 healthy individuals and the
measurements were repeated three times. Also, five meas-
urements were taken in the cardiac catheter laboratory
prior to the start of the study to ensure correct simultaneous
measurement of invasive and noninvasive blood pressures.
The invasive measurement was performed by experienced
examiners who had performed more than 1000 cardiac
catheterizations.
Premeasurement examinations
The medical history and blood pressure medication were
registered the day before the cardiac catheter examination.
Blood was taken from all patients for routine laboratory
examination and ECG and echocardiography were per-
formed. For analysis in this study, atrial fibrillation was
diagnosed when this atrial fibrillation occurred during
invasive and noninvasive measurement of central blood
pressure, regardless of previous episodes of paroxysmal or
persistent atrial fibrillation.
Study setting
The hemodynamic study was performed in the cardiac
catheterization laboratory. Care was taken to minimize
disturbing influences during the examination. The meas-
urements were performed on supine patients. The circum-
ference of the patient’s middle upper arm was measured
with the enclosed device tape measures. The cuff sizes were
selected according to the upper arm circumference (Mobil-
O-Graph NG device: XS 14–20 cm, S 20–24 cm, M 24–
32 cm, L 32–38, XL 38–55 cm; SphygmoCor XCEL device:
adults 23–33 cm, adults large 31–40 cm, adults extra large
38–50 cm).
Invasive assessment of central aortic blood pressure
Before the measurements, the patients lay on the examina-
tion table in a quiet position for at least 10 min. A 5-French
sized sheath was inserted via femoral artery or radial artery.
In some patients (n ¼ 15), it was necessary to administer
0.04 mg nitroglycerin immediately after positioning the
radial sheath. In this case, at least 5 min passed between
administration of the vasoactive substance and data acqui-
sition. Patients were asked not to speak during the measure-
ments. Measurements were taken before contrast agent was
applied. No vasoactive substances or contrast agents were
administered during the measurement. The guiding cathe-
ter used was 5-French Cordis infiniti trulumen (Cordis
Corporation, Miami Lakes, Florida, USA; length 100 cm,
inner diameter 0.97 mm). Under fluoroscopic control, the
catheter was positioned in the aorta ascendens, about 1 cm
above the aortic valve. The catheter was connected using
three taps and connectors to the measuring system
Volume 37  Number 1  Month 2019
CE: Tripti; JH-D-19-00010; Total nos of Pages: 8;
JH-D-19-00010
Accuracy of central aortic blood pressure measurement techniques
(Schwarz GmbH & Co. KG, Kamen, Germany). The pres-
sure transducer was attached to the patient’s heart by a fixed
suspension and remained there throughout the examina-
tion. Blood was then aspirated to remove any air bubbles
from the catheter. The catheter was then flushed with 10 ml
physiological saline solution and the measuring system was
zeroed and calibrated. The natural frequency was con-
firmed to be greater than 20 Hz, and the damping coeffi-
cient was not less than 0.3. The procedure was repeated
when there was evidence of a damped curve. The mea-
surement of systolic and diastolic central blood pressure
values was digitally recorded. The data of at least 10 heart-
beats were averaged.
Noninvasive assessment of central aortic blood
pressure
The noninvasive blood pressure measurements were per-
formed under hemodynamically stable conditions by the
SphygmoCor XCEL device and the Mobil-O-Graph NG
device simultaneously and with the invasive central blood
pressure measurements. The sequence of the noninvasive
devices alternated randomly. In the noninvasive and inva-
sive examinations runs of bigeminy, trigeminy, or isolated
premature beats and the following compensatory beats
have been removed from analysis. In some cases, the
devices reported error messages so that a sufficiently accu-
rate central blood pressure measurement was not possible.
In these cases, the measurement was repeated a second
time with the device and, if necessary, a third time. If the
third measurement was also unsuccessful, no further
attempts were made. These measurements were defined
as ‘unsuccessful measurement’.
The recommendations of the ARTERY Society regarding
the sample characteristics and the statistical requirements
were fulfilled in this study [9].
Statistics
Results are presented as mean  standard deviation (SD) in
case of normal distribution. Continuous variables without
normal distribution are presented as median (first quartile,
third quartile). Pearson correlation analyses were per-
formed for noninvasive vs. invasive values. Fisher’s z
was calculated to compare the correlation coefficients.
Moreover, comparison of systolic and diastolic central
blood pressure values of the two test devices and the
simultaneous invasively assessed reference cBP was per-
formed by Tukey mean-difference plots (Bland–Altman
plots). In accordance with Krouwer [10], the x-axis presents
the results of the invasive measurement as the gold standard
method. Bias and the limits of agreement (bias
2  standard deviation) are reported. For the calculation
of the extreme outliers on Bland–Altman plot, the upper
and lower quartiles (Q1 and Q3) of the systolic and diastolic
bias were calculated. A point beyond an outer fence (lower
outer fence: Q1  3  interquartile range, upper outer
fence: Q3 þ 3  interquartile range) was considered an
extreme outlier. Extreme outliers in the Bland–Altman plot
(systolic or diastolic bias) and unsuccessful measurement
were considered as measurement failure.
Comparison of each method’s findings in patients with
and without atrial fibrillation was performed by unpaired
Journal of Hypertension
Copyright © 2019 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
TABLE 1. Clinical characteristics of study patients
All patients (n ¼ 502)
Age (years) 67.9  11.6
Female sex (n, %) 228 (45)
BMI (kg/m2) 29  5.9
SBP (mmHg) 145  22
DBP (mmHg) 80.2  14.6
Left ventricular ejection fraction (%) 54.3  12.4
Hypertension (n, %) 454 (91)
Hypercholesterolemia (n, %) 356 (71)
Active smoker (n, %) 145 (29)
Previous smoker (n, %) 103 (21)
Diabetes mellitus (n, %) 133 (26)
Coronary artery disease (n, %) 287 (57)
Atrial fibrillation (n, %) 135 (27)
two-tailed t tests (for normally distributed variables) or a
Mann–Whitney U test (for nonnormally distributed varia-
bles), whenever appropriate. Dichotomic parameters were
compared by chi-square test (Fisher’s exact test for inde-
pendent samples). P < 0.05 was regarded significant.
RESULTS
Table 1 presents the epidemiological and medical charac-
teristics of the study population. Mean age was 67.9  11.6
ranging from 34 to 92 years. Three hundred and sixty-five
patients (71%) had sinus rhythm, 135 (27%) suffered from
atrial fibrillation at the time of the measurement procedure.
Concomitant diseases constituted coronary artery disease
(n ¼ 287, 57%), peripheral occlusive disease (n ¼ 45, 16%),
and hypertension (n ¼ 454, 91%). The median number of
antihypertensive drugs was ranging from 0 to 6, and con-
tained diuretics, calcium-channel blockers, b-blockers, ACE
inhibitors, AT1 blockers, alpha blockers, mineralocorticoid-
antagonists, and hydrazinophthalazine (Supplement Table
1, http://links.lww.com/HJH/B145).
The invasive measurement of aortic blood pressure was
performed in all patients as part of the cardiac catheter
examination. The simultaneous measurement of the non-
invasive central blood pressure using the SphygmoCor
XCEL device was successful in 498 of 502 patients (99%).
Measurements using Mobil-O-Graph NG device were suc-
cessful in 441 of 502 patients (88%; P ¼ 0.451). In the
remaining patients, no values could be determined even
after the third measurement.
Measurement failure (unsuccessful measurement or
extreme outliers in the Bland–Altman plot) occurred in
10 patients examined with the SphygmoCor XCEL device
and in 94 patients examined with Mobil-O-Graph NG
device. Patients with measurement failure and successful
examination were compared in clinical characteristics. The
results are presented in Supplement Tables 2 (http://link-
s.lww.com/HJH/B145) and 3 (http://links.lww.com/HJH/
B145). In SphygmoCor XCEL device the invasive diastolic
central blood pressure was significantly lower in the group
with measurement failure than in the group of patients with
successful measurement. In Mobil-O-Graph NG device, the
SBP was significantly higher and atrial fibrillation occurred
more frequently in the group with measurement failure
compared with the patients with successful measurement.
www.jhypertension.com 3
CE: Tripti; JH-D-19-00010; Total nos of Pages: 8;
JH-D-19-00010

Gotzmann et al.

TABLE 2. Results of simultaneous measurement of invasive and noninvasive central blood pressure with the SphygmoCor XCEL and
Mobil-O-Graph NG device
SphygmoCor measurements (n ¼ 498) Mobil-O-Graph measurements (n ¼ 441)
Invasive systolic cBP (mmHg) 131  21 Invasive systolic cBP (mmHg) 132  21
Invasive diastolic cBP (mmHg) 76.4  11.9 Invasive diastolic cBP (mmHg) 77.8  11.8
Noninvasive systolic cBP (mmHg) 127  20 Noninvasive systolic cBP (mmHg) 126  20
Noninvasive diastolic cBP (mmHg) 77  13 Noninvasive diastolic cBP (mmHg) 81.4  13.8

only the data from successful measurements of noninvasive devices are presented. cBP, central blood pressure.

Table 2 provides mean invasively and noninvasively
assessed cBP values. The performance of the different
noninvasively cBP measurement techniques was analyzed
using two different approaches: correlations between inva-
sively and noninvasively measured values, assessment of
bias and limits of agreement in a Bland–Altman analysis.
Figure 1 illustrates the correlations between invasively
and noninvasively measured values. Correlations were
highly significant for both SBP and DBP with both devices
(P < 0.001 each). With each device, correlation coefficients
were higher for systolic than for diastolic cBP. Furthermore,
the correlation coefficients in systolic and diastolic cBP
were significantly higher for the SphygmoCor XCEL device
compared with the Mobil-O-Graph NG device (P < 0.001
each).
In the overall study population, the mean systolic bias of
SphygmoCor XCEL device to invasively assessed cBP was
5.0  7.7 mmHg. The systolic bias of the Mobil-O-Graph
NG device was slightly but significantly larger
(6.0  10.4 mmHg, P ¼ 0.011). The diastolic bias of Sphyg-
moCor XCEL device was significantly lower than the Mobil-
O-Graph NG device bias (0.5  6.2 vs. 3.6  8.3 mmHg;

FIGURE 1 Relationship between systolic and diastolic central blood pressure comparing measurements of SphygmoCor device and Mobil-O-Graph device and invasively
measured central blood pressure in the overall study population (patients with and without atrial fibrillation).

4 www.jhypertension.com Volume 37  Number 1  Month 2019

Copyright © 2019 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
CE: Tripti; JH-D-19-00010; Total nos of Pages: 8;
JH-D-19-00010

Accuracy of central aortic blood pressure measurement techniques

FIGURE 2 Tukey mean-difference plots (Bland–Altmann plots) of systolic and diastolic central blood pressure comparing SphygmoCor device and Mobil-O-Graph device to
the goldstandard of invasively measured central blood pressure in the overall study population (patients with and without atrial fibrillation). Dotted lines provide mean bias
and 95% limits of agreement (2 standard deviation).

P < 0.001). Figure 2 provides the corresponding Tukey
mean-difference plots including bias and limits of agree-
ment (95% confidence interval, 2 standard deviation).
In a second approach, we investigated the impact of
atrial fibrillation on the accuracy of noninvasive cBP mea-
surement. The corresponding correlation analyses are
shown in Fig. 3. Systolic and diastolic R2 values did not
significantly differ from those of the overall study popula-
tion with the SphygmoCor device (P ¼ 0.220 and P ¼ 0.750,
respectively), and with the Mobil-O-Graph device
(P ¼ 0.062 and 0.224, respectively). In analogy to the overall
study population, systolic and diastolic correlation coeffi-
cients were higher in SphygmoCor device compared with
Mobil-O-Graph device in patients with atrial fibrillation
(P ¼ 0.003 and 0.002, respectively). The mean systolic bias
of SphygmoCor XCEL device (5.7  8.9 mmHg) was
almost identical to that of the overall study population
(P ¼ 0.191), whereas it increased by 1 mmHg diastolic
(1.5  6.9 mmHg, P ¼ 0.042). In analogy, atrial fibrillation
did not signficantly change the systolic bias of the Mobil-O-
Graph NG device (6.2  12.3 mmHg, P ¼ 0.862) or the
diastolic bias (4.7  9.6 mmHg; P ¼ 0.095). Figure 4 pro-
vides the corresponding Tukey mean-difference plots.
DISCUSSION
Until recently, noninvasive cBP measurement was not
possible in an automated manner and necessitated an
observer trained in applanation tonometry. The present
study shows that the cBP values obtained by automated
oscillometric cuff-based pulse contour analysis highly cor-
relate to invasive measurements. The noninvasive devices
thereby slightly underestimate systolic and slightly overes-
timate diastolic cBP. This inaccuracy (systematic underesti-
mation of SBP but overestimation of DBP) has been
described before and may be the consequence of the
calibration method [11]. In the present study, both devices
made use of a traditional calibration based on SBP and DBP.

Journal of Hypertension www.jhypertension.com 5

Copyright © 2019 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
CE: Tripti; JH-D-19-00010; Total nos of Pages: 8;
JH-D-19-00010
Gotzmann et al.
FIGURE 3 Relationship between systolic and diastolic central blood pressure comparing measurements of SphygmoCor device and Mobil-O-Graph device and invasively
measured central blood pressure in patients with atrial fibrillation.
This method, however, may underestimate true systolic
brachial and thereby true systolic cBP. In the meantime,
an alternative calibration using mean instead of SBP has
become available for the Mobil-O-Graph. There are some
hints that this calibration may indeed lead to a further
increase of the device’s accuracy [12].
The mean differences of 5–6 mmHg in systolic and 1–
4 mmHg in diastolic cBP in our study appear acceptable
from a clinical point of view. Nevertheless, our findings
should encourage manufacturers to adapt their general
transfer functions. Mean biases in systolic and diastolic
cBP were slightly but significantly lower for SphygmoCor
XCEL compared with Mobil-O-Graph NG (P ¼ 0.011 and
P < 0.001, respectively). The separate analysis of those
patients in the upper and lower quartiles (outliers) identi-
fied a low diastolic (SphygmoCor XCEL) and a high systolic
cBP (Mobil-O-Graph) as risk factors for an inaccurate
measurement. A high SBP and a low DBP are indicators
of increased aortic stiffness and isolated systolic hyperten-
sion. We have previously demonstrated that the accuracy of
BP measurement is reduced in this population [13].
6 www.jhypertension.com
Copyright © 2019 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
With a number of 502 participants, the current study is
the largest validation study of noninvasive cBP measure-
ment techniques so far [6]. There are some validation
studies on the prior SphygmoCor system using applanation
tonometry [14–16]. To the best of our knowledge, there are
only one small invasive validation study on the current
oscillometric SphygmoCor device (36 patients) [17] and one
small study on the Mobil-O-Graph (30 individuals) [18]. In
the latter study, the device underestimated invasively mea-
sured central BP by 3 mmHg. Noteworthy, the manufacturer
modified the transfer function since the publication of this
report. The present study thereby constitutes the first vali-
dation study of the device in its current form. Noteworthy,
our study reveals still an underestimation of systolic cBP.
The SphygmoCor device showed a slightly but significantly
higher accuracy in the estimation of systolic and diastolic
cBP compared with the Mobil-O-graph device regarding
both correlation and biases.
A recent review on the accuracy of noninvasive mea-
surement techniques identified 22 eligible studies, which
validated 11 different commercial devices in 808 study
Volume 37  Number 1  Month 2019
CE: Tripti; JH-D-19-00010; Total nos of Pages: 8;
JH-D-19-00010

Accuracy of central aortic blood pressure measurement techniques

FIGURE 4 Tukey mean-difference plots (Bland–Altmann plots) of systolic and diastolic central blood pressure comparing SphygmoCor device and Mobil-O-Graph device to
the goldstandard of invasively measured central blood pressure in patients with atrial fibrillation. Dotted lines provide mean bias and 95% limits of agreement (2 standard
deviation).

participants. [6]. The number of 502 in the present study
constitutes the largest study population investigated so far.
The vast majority of validation studies in this review used
applanation tonometry for pulse wave analysis. The mean
error in the estimation of systolic cBP was 4.5 mmHg.
Thus, the diagnostic accuracy of the present two oscillo-
metric devices was almost exactly as accurate than the
traditional tonometric devices. A measurement procedure
takes about 1–2 min including both peripheral BP mea-
surement and pulse wave analysis. This time is only slightly
longer than a standard BP measurement, and may therefore
facilitate an integration in daily clinical practice. The auto-
mated procedure does not necessitate an intense training of
staff anymore and the results are far less observer-depen-
dent than previous tonometric approaches, which required
placement of the tip of a hand-held high-fidelity tonometer
on the patient’s artery.
Despite the increasing data on the prognostic value of
cBP, its measurement is not recommended by current
hypertension guidelines [19,20]. This decision is mainly
based on the fact that randomized controlled trials with
cardiovascular endpoints made use of brachial and not cBP
so far. The present oscillometric technology could be easily
implemented in cardiovascular large-scale studies. These
studies could investigate, whether cBP is indeed superior to
brachial BP in the prediction of cardiovascular endpoints.
The most crucial limitation of the present study is a
potential selection bias. Comparison of noninvasively vs.
invasively measured aortic BP was performed in patients
undergoing cardiac catheterization and not in healthy indi-
viduals. The same is true for prior validation studies and the
majority of the manufacturers’ approaches in the develop-
ment of the devices. This limitation is inevitable to avoid
aortic catheterization of healthy probands. Noteworthy, 159
patients (32%) of the present study population with pre-
served left ventricular ejection fraction and no severe heart
valve disease turned out to have an unremarkable coronary
status. The mean systolic/diastolic biases in this subgroup
were 4.1  6 mmHg/0.1  6.2 mmHg for the SphygmoCor
device and -4.7  7.6 mmHg/2.4  6.7 mmHg for the Mobil-
O-Graph device. Thus, the present selection bias may be of
subordinate clinical relevance.

Journal of Hypertension www.jhypertension.com 7

Copyright © 2019 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
CE: Tripti; JH-D-19-00010; Total nos of Pages: 8;
JH-D-19-00010
Gotzmann et al.
The present study shows that two current automated
oscillometric BP monitors are able to assess cBP as accurate
as former tonometric techniques. The automated measure-
ment procedure shortens the time demand and minimizes
observer errors. Manufacturers should feel encouraged to
adapt the transfer function to avoid the slight underestima-
tion of systolic cBP in the future. This technique is able to
substantially facilitate the use of cBP measurement in future
studies and daily clinical practice.
ACKNOWLEDGEMENTS
We would like to thank the cardiac catheter team for their
exceptional support.
Conflicts of interest
There are no conflicts of interest.
REFERENCES
1. Lawes CM, Vander Hoorn S, Rodgers A, International Society of
Hypertension. Global burden of blood-pressure-related disease,
2001. Lancet 2008; 371:1513–1518.
2. Vlachopoulos C, Aznaouridis K, O’Rourke MF, Safar ME, Baou K,
Stefanadis C. Prediction of cardiovascular events and all-cause mortal-
ity with central haemodynamics: a systematic review and meta-analy-
sis. Eur Heart J 2010; 31:1865–1871.
3. Cheng HM, Chuang SY, Sung SH, Yu WC, Pearson A, Lakatta EG, et al.
Derivation and validation of diagnostic thresholds for central blood
pressure measurements based on long-term cardiovascular risks. J Am
Coll Cardiol 2013; 62:1780–1787.
4. Dahlof B, Sever PS, Poulter NR, Wedel H, Beevers DG, Caulfield M,
et al., ASCOT Investigators. Prevention of cardiovascular events with
an antihypertensive regimen of amlodipine adding perindopril as
required versus atenolol adding bendroflumethiazide as required, in
the Anglo-Scandinavian Cardiac Outcomes Trial-Blood Pressure Low-
ering Arm (ASCOT-BPLA): a multicentre randomised controlled trial.
Lancet 2005; 366:895–906.
5. Williams B, Lacy PS, Thom SM, Cruickshank K, Stanton A, Collier D,
et al., CAFE Investigators; Anglo-Scandinavian Cardiac Outcomes Trial
Investigators; CAFE Steering Committee and Writing Committee. Dif-
ferential impact of blood pressure-lowering drugs on central aortic
pressure and clinical outcomes: principal results of the Conduit Artery
Function Evaluation (CAFE) study. Circulation 2006; 113:1213–1225.
6. Papaioannou TG, Karageorgopoulou TD, Sergentanis TN, Protogerou
AD, Psaltopoulou T, Sharman JE, et al. Accuracy of commercial devices
and methods for noninvasive estimation of aortic systolic blood pres-
sure a systematic review and meta-analysis of invasive validation
studies. J Hypertens 2016; 34:1237–1248.
8 www.jhypertension.com
Copyright © 2019 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
7. Cheng HM, Lang D, Tufanaru C, Pearson A. Measurement accuracy of
noninvasively obtained central blood pressure by applanation tonom-
etry: a systematic review and meta-analysis. Int J Cardiol 2013;
167:1867–1876.
8. Millasseau S, Agnoletti D. Noninvasive estimation of aortic blood
pressures: a close look at current devices and methods. Curr Pharm
Des 2015; 21:709–718.
9. Sharman JE, Avolio AP, Baulmann J, Benetos A, Blacher J, Blizzard CL,
et al. Validation of noninvasive central blood pressure devices:
ARTERY Society task force consensus statement on protocol standard-
ization. Eur Heart J 2017; 38:2805–2812.
10. Krouwer JS. Why Bland-Altman plots should use X, not (Y þ X)/2 when
X is a reference method. Stat Med 2008; 27:778–780.
11. Picone DS, Schultz MG, Otahal P, Aakhus S, Al-Jumaily AM, Black JA,
et al. Accuracy of cuff-measured blood pressure: systematic reviews
and meta-analyses. J Am Coll Cardiol 2017; 70:572–586.
12. Wassertheurer S, Baumann M. Assessment of systolic aortic pressure
and its association to all cause mortality critically depends on waveform
calibration. J Hypertens 2015; 33:1884–1888; discussion 9.
13. Westhoff TH, Schmidt S, Meissner R, Zidek W, van der Giet M. The
impact of pulse pressure on the accuracy of wrist blood pressure
measurement. J Hum Hypertens 2009; 23:391–395.
14. Davies JI, Band MM, Pringle S, Ogston S, Struthers AD. Peripheral
blood pressure measurement is as good as applanation tonometry
at predicting ascending aortic blood pressure. J Hypertens 2003;
21:571–576.
15. Cloud GC, Rajkumar C, Kooner J, Cooke J, Bulpitt CJ. Estimation of
central aortic pressure by SphygmoCor requires intra-arterial periph-
eral pressures. Clin Sci (Lond) 2003; 105:219–225.
16. Laugesen E, Rossen NB, Peters CD, Mæng M, Ebbehøj E, Knudsen ST,
et al. Assessment of central blood pressure in patients with type 2
diabetes: a comparison between SphygmoCor and invasively mea-
sured values. Am J Hypertens 2014; 27:169–176.
17. Shoji T, Nakagomi A, Okada S, Ohno Y, Kobayashi Y. Invasive
validation of a novel brachial cuff-based oscillometric device (Sphyg-
moCor XCEL) for measuring central blood pressure. J Hypertens 2017;
35:69–75.
18. Weber T, Wassertheurer S, Rammer M, Maurer E, Hametner B, Mayer
CC, et al. Validation of a brachial cuff-based method for estimating
central systolic blood pressure. Hypertension 2011; 58:825–832.
19. Williams B, Mancia G, Spiering W, Agabiti Rosei E, Azizi M, Burnier M,
et al., ESC Scientific Document Group. 2018 ESC/ESH Guidelines for
the management of arterial hypertension. The Task Force for the
Management of Arterial Hypertension of the European Society of
Cardiology and the European Society of Hypertension. J Hypertens
2018; 36:1953–2041.
20. Whelton PK, Carey RM, Aronow WS, Casey DE Jr, Collins KJ, Dennison
Himmelfarb C, et al. 2017 ACC/AHA/AAPA/ABC/ACPM/AGS/APhA/
ASH/ASPC/NMA/PCNA Guideline for the Prevention, Detection, Evalu-
ation, and Management of High Blood Pressure in Adults: a report of the
American College of Cardiology/American Heart Association Task Force
on Clinical Practice Guidelines. J Am Coll Cardiol 2018; 71:e127–e248.
Volume 37  Number 1  Month 2019