Download Arterial stiffness and central hemodynamics in

130 Original article
Arterial stiffness and central hemodynamics in treated
hypertensive subjects according to brachial blood
pressure classification
Michel E. Safara, Jacques Blachera, Athanase Protogeroub and
Apostolos Achimastosb
Background International recommendations have
classified brachial blood pressure (BP) in subgroups
enabling better cardiovascular risk stratification. Central BP
is an independent predictor of cardiovascular risk, differing
from brachial BP through the predominant influence of
arterial stiffness and wave reflections. Central BP has never
been studied in relation to international guidelines for
brachial BP classification.
brachial BP classification, depending on differences in
aortic stiffness and HR. Whether aortic PWV might predict
the brachial BP classification and/or the presence of
effective BP control, as suggested in this study, needs
further confirmation. J Hypertens 26:130–137 Q 2008
Wolters Kluwer Health | Lippincott Williams & Wilkins.
Journal of Hypertension 2008, 26:130–137
Methods In 580 chronically treated hypertensive subjects
we measured: carotid–femoral pulse wave velocity (PWV),
carotid artery augmentation index (AI) and carotid blood
pressures, using applanation tonometry and pulse wave
analysis, and using brachial BP for carotid pressure wave
calibration.
Results For each given brachial value, carotid systolic
blood pressure (SBP) and PP were significantly lower
than the corresponding brachial SBP and PP. This
pressure amplification was significantly lower in the
‘optimal’ and ‘normal’ BP ranges (6.8–7.4 mmHg) than
in the higher BP ranges (10.1–11.3 mmHg), mainly
depending on heart rate (HR) and PWV levels. PWV gradually
increased as a function of brachial BP classification and was
a significant predictor of this classification independently of
age, drug treatment, atherosclerotic lesions and even mean
BP. Finally, PWV was a highly sensitive marker of the
effective BP control throughout all decades of age.
Conclusion Under chronic antihypertensive therapy,
central BP does not strictly parallel the corresponding
Introduction
Most guidelines in the literature classify the population
into blood pressure (BP) subgroups, according to both
systolic and diastolic BP, in order to better represent the
cardiovascular (CV) risk derived from such mechanical
factors. These classifications refer to peripheral (brachial
artery) BP measurements, but never involve central
(thoracic aorta or carotid artery) BP. However, large-scale
clinical trials have shown that: the reduction of cardiovascular morbidity and mortality by antihypertensive treatment cannot be fully explained exclusively from the
lowering of peripheral BP, implying largely changes in
vascular mechanics at the level of the micro- or macrocirculation and central hemodynamics; the monitoring of
Keywords: antihypertensive therapy, arterial stiffness, blood pressure
classification, central blood pressure, pressure amplification, pressure
wave reflections
Abbreviations: ACE, angiotensin converting enzyme; AI, augmentation
index; ATII, angiotensin II receptor antagonist; BMI, body mass index;
BP, blood pressure; BPC, blood pressure classification; CV, cardiovascular
risk; DBP, diastolic blood pressure; ESH, European Society of Hypertension;
GLM, general linear model; HDL, high density lipoprotein; HR, heart rate;
LDL, low density lipoprotein; LVET, left ventricular ejection time; MBP, mean
blood pressure; PP, pulse pressure; PWV, pulse wave velocity; ROC, receiver
operator curve analysis; RWTT, reflected wave time transit; SBP, systolic
blood pressure; SEM, standard error of the mean
a
AP-HP Diagnosis Center, Hôtel-Dieu Hospital, Faculty of Medicine,
Paris-Descartes University, Paris, France and bHypertension Center, Third
Department of Internal Medicine, Sotiria Hospital, University of Athens,
Greece
Correspondence and requests for reprints to Professor Michel Safar, Diagnosis
Center, Hôpital Hôtel-Dieu, 1, place du Parvis Notre-Dame, 75181 Paris Cedex
04, France
Tel: +33 1 42 34 80 25; fax: +33 1 42 34 86 32;
e-mail: [email protected]
Received 27 February 2007 Revised 16 July 2007
Accepted 22 August 2007
See editorial commentary on page 16
peripheral BP is not sufficient to describe the actual
response to drug treatment, which takes into account
central hemodynamic parameters; and that the assessment
of central BP, together with aortic stiffness and pressure
wave reflections, may give new perspectives in CV risk
assessment and reduction strategies, related to hypertension [1–8].
Systolic BP and pulse pressure (PP) are higher in
peripheral than in central arteries, for a quite similar
mean BP [9–11]. This phenomenon is called systolic
and pulse pressure amplification and results from the
propagation of pressure waves along the vascular bed,
the progressive narrowing of arterial and arteriolar
0263-6352 ß 2008 Wolters Kluwer Health | Lippincott Williams & Wilkins
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Brachial BP classification v. central hemodynamics Safar et al. 131
vessels with subsequent increase of arterial stiffness,
and mostly the summation of wave reflections [12].
Classically, SBP and PP amplifications are modulated
by two factors: age and heart rate (HR), but also are
influenced by arterial stiffness, pressure wave reflections and CV risk factors [13].
The purpose of this study was, in a cohort of 580 chronically treated hypertensive subjects, to investigate the
differences in central hemodynamics (aortic stiffness,
pressure wave reflections, carotid blood pressure and
pressure amplification) between the six subgroups of
the international brachial BP classification (optimal to
grade III hypertension), which was proposed by the
European Society of Hypertension–European Society
of Cardiology 2003 (ESH2003) guidelines [14]. This
classification was based exclusively on peripheral BP
measurement.
Methods
Study cohort
The study population consisted of a cohort of subjects
referred to our center for a CV check-up ordered by their
physician because of the presence of CV risk factors and/or
family history of CV disease. All subjects provided
informed consent for the investigation, which was
approved by our institutional review board. Patients with
all forms of secondary hypertension, with cancer, or with
severe renal insufficiency (plasma creatinine >300 mmol/l)
were excluded from the study. All subjects with secondary
hypertension were excluded on the basis of classical
laboratory tests [1]. All treated hypertensives entered
the study independently from their BP level. At inclusion
all patients had a thorough review of their medical history
for the detection of clinical events and/or signs related to
the presence of CV risk factors. Atherosclerotic lesions
(AL) were defined according to the 9th revision of the
International Classification of Diseases for coronary heart
disease, cerebrovascular disease, peripheral vascular
disease and abdominal aneurysm, as described previously
[1]. Venous blood samples were obtained from all patients
after an overnight fast for routine biochemical investigations, including plasma total cholesterol, triglycerides,
low-density lipoprotein (LDL) cholesterol and highdensity lipoprotein (HDL) cholesterol, glucose and
creatinine levels, all determined by standard methods
[1]. From the 580 hypertensive subjects, 123 (21%) subjects had diabetes mellitus, 232 (40%) had metabolic
syndrome, 82 (14%) were smokers and 124 (21%) had
AL at at least one site (either coronary, carotid, aorta or
lower limb artery disease). Two hundred and twenty-four
subjects (39%) were under treatment with diuretics, 20
(3%) with alpha-blockers (including both pre- and postsynaptic), 219 (38%) with beta-blockers, 346 (60%) with
calcium-channel blockers, 180 (31%) with angiotensinconverting enzyme inhibitors, 44 (8%) with an angiotensin
II receptor antagonist, 122 (21%) with hypolipidemic
drugs, 44 (8%) with antidiabetic drugs and 108 (19%)
with aspirin.
Hemodynamic measurements
All measurements were performed in the morning and at
stable room temperature, after an overnight fast. Brachial
BP measurements were performed by traditional mercury
sphygmomanometer in the supine position after a 15 min
rest in the laboratory, using the first and the fifth
Korotkoff sounds for SBP and DBP, respectively. The
average of the last two out of three consecutive BP
measurements was used for data analysis. Radial artery
and carotid artery applanation tonometry by a high-fidelity
Millar strain gauge transducer (SPT-301; Millar Instruments, Houston, Texas, USA) were performed as
described previously [15]. Briefly, the derived pressure
waveforms were recorded on a Gould 8188 recorder (Gould
Electronic, Ballainvilliers, France) at a paper speed of
100 mm/s. Radial pressure waveform calibrated from
brachial SBP and DBP was used for determination of
peripheral (brachial) mean BP via application of an integration method. A recent study by Verbeke et al. [16] has
reported the presence of a small pressure amplification
between the brachial and radial artery. Although this result
was observed in healthy young subjects and may not be
extrapolated in our population, it is possible that this
mode of calibration may introduce a systematic error.
We believe that this may not have significantly modified
our results. On the other hand, we have previously shown
that the use of transfer function may amplify the initial
error of calibration [17]. Since diastolic and mean BP
differences throughout the arterial tree are minor (ascending aorta to radial artery differences do not exceed 2–
3 mmHg) [9], the obtained carotid pressure wave was
calibrated using brachial diastolic and mean BP. Mean
BP of the carotid pressure waveform, computed from the
area method, was assumed to be equal to peripheral mean
BP in order to calculate the amplitude of the carotid
pressure waveform as well as carotid PP and systolic BP.
Carotid PP is considered as a close surrogate of aortic PP,
and it has been previously validated by comparison with
invasive measurements, as well as by the use of mathematical transformation [9,18–21]. Augmentation index
(AI), left ventricular ejection time (LVET) and the
reflected wave time transit (RWTT, i.e. the time needed
for the pressure wave to travel forward and then return
backward at the level of the central arteries) were
calculated according to standard definitions [22]. PP amplification between brachial PP and carotid PP was calculated
as the absolute difference of brachial PP – carotid PP
(mmHg). Carotid–femoral (aortic) PWV was measured
automatically using the Complior apparatus (Colson, Paris,
France) [23]. Its determination is based on the simultaneous recording of pulse wave in the common carotid
and femoral artery by two transducers, and calculated as
the distance separating the two transducers divided by the
time delay between the onset (foot) of the two recorded
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
132 Journal of Hypertension
2008, Vol 26 No 1
waves. The reproducibility of all these measurements has
been published previously [15,20,23].
Statistical analysis
All quantitative variables were expressed as mean standard error of the mean (SEM). Presence or absence of a trait
(e.g. drugs, metabolic syndrome) was expressed as 1 or 0,
respectively; gender as male: 1 and female: 2. Subjects
were divided into six subgroups according to their brachial
systolic and diastolic BP (mmHg) level as suggested from
the classification of the ESH2003 report (ESH2003):
optimal (SBP, <120 mmHg; DBP, <80 mmHg), normal
(SBP, 120–129 mmHg; DBP, 80–84 mmHg), high
normal (SBP, 130–139 mmHg; DBP, 85–89 mmHg),
grade I hypertension (SBP, 140–159 mmHg; DBP, 90–
99 mmHg), grade II hypertension (SBP, 160–179 mmHg;
DBP, 100–109 mmHg), grade III hypertension (SBP,
180 mmHg; DBP, 110 mmHg). In order to classify a
subject in one of the above groups, this individual should
have either both SBP and DBP in the same class, or one of
the two (SBP or DBP) in this class and the other one in a
lower class. For the comparison of qualitative data between
the groups, the x2-test was used. Comparison of mean
values of quantitative variables between the six subgroups
before and after appropriate adjustments was carried out
by ANOVA/ANCOVA provided by general linear
model (GLM) and the use of the Bonferroni post-hoc test.
Univariate and multivariate regression analysis was
applied in order to find the independent predictors of
ESH2003-BP classification (BPC). Receiver operator
curve (ROC) analysis was performed in order to evaluate
the performance of PWV on the classification of hypertensive subjects as effectively treated (ESH2003-BPC from
optimal to high normal) or ineffectively treated (ESH2003BPC from grade I to III). This type of analysis was
performed in the totality of the population and then
in subgroups according to decades of age. Statistical
analysis was performed using SPSS version 11.5 (SPSS
Inc., Chicago, Illinois, USA). P 0.05 was considered to be
the level of statistical significance.
Table 1
Results
In Table 1 the distribution of CV risk factors in the six BP
subgroups is shown. Except for a significant difference in
the prevalence of diabetes mellitus (P ¼ 0.013), as well as
a significant decrease of plasma potassium from optimal
to grade III hypertension (P ¼ 0.044), all other CV risk
factors (including age, gender, smoking habits and dyslipidemia) were evenly distributed between the six groups.
Atherosclerotic lesions tended to be more frequent in
the subgroups with better brachial blood pressure control
(P ¼ 0.063).
In Table 2 the distribution of drugs in the six subgroups is
presented. A significant increase in the mean number of
antihypertensive drugs used as well as in the duration of
the antihypertensive treatment was observed (P < 0.001
and P ¼ 0.003, respectively). The distribution of alphablockers and of calcium-channel inhibitors gradually and
significantly increased (P < 0.001) but the distribution of
beta-blockers decreased (P ¼ 0.037) from the group of
optimal BP to the group of grade III hypertension. All
other antihypertensive, antidiabetic and antiplatelet
drugs were evenly distributed between the six groups.
The hemodynamic data as a function of brachial BP
classification are described in Table 3. Carotid SBP
and PP were significantly lower than brachial SBP and
PP, within all six subgroups of BP classification. SBP,
DBP, PP and mean blood pressure (MBP) at the level of
both carotid and brachial arteries increased gradually and
significantly from the group of optimal BP to the group
with grade III hypertension (P < 0.001). Similarly, HR
gradually and significantly increased (P ¼ 0.042) and the
RWTT/LVET ratio gradually decreased (P < 0.001).
Carotid AI gradually and significantly increased from
the subgroup with the lower blood pressure to the subgroup with higher blood pressure, even after adjustment
for age, gender, heart rate (or RWTT/LVET), glucose,
height and PWV. This increase disappeared only
after additional adjustment for mean BP. This increase
Cardiovascular risk factors as a function of ESH2003-brachial blood pressure classification (ESH2003-BPC)
ESH2003-BPC
N
Age (years) (SE)
Gender (men) (%)
BMI (kg/m2) (SE)
Height (cm) (SE)
Smoking (yes) (%)
Diabetes mellitus (yes) (%)
Total cholesterol (mmol/l) (SE)
LDL cholesterol (mmol/l) (SE)
HDL cholesterol (mmol/l) (SE)
Triglycerides (mmol/l) (SE)
Plasma creatinine (mEq/l) (SE)
Plasma potassium (mEq/l) (SE)
Metabolic syndrome (yes) (%)
Presence of atherosclerosis (yes) (%)
Optimal
Normal
High normal
Grade I
Grade II
Grade III
P value
48
59.2 (1.7)
29 (60)
25.7 (0.6)
169.0 (0.01)
11 (22)
7 (15)
5.2 (0.2)
3.4 (0.1)
1.2 (0.1)
1.37 (0.1)
91.6 (6.3)
4.22 (0.01)
17 (35)
18 (37)
83
60.6 (1.3)
43 (52)
27.5 (0.5)
167.1 (0.01)
14 (17)
11 (14)
5.6 (0.1)
3.6 (0.1)
1.3 (0.1)
1.28 (0.1)
93.3 (4.8)
4.21 (0.01)
30 (36)
19 (23)
107
58.6 (1.2)
65 (61)
26.8 (0.4)
169.7 (0.01)
17 (16)
16 (18)
5.2 (0.1)
3.4 (0.1)
1.3 (0.1)
1.31 (0.1)
92.7 (4.2)
4.27 (0.01)
41 (38)
21 (17)
187
62.1 (0.9)
111 (59)
27.5 (0.3)
168.3 (0.01)
15 (8)
42 (22)
5.4 (0.1)
3.5 (0.1)
1.3 (0.1)
1.31 (0.1)
96.0 (3.1)
4.13 (0.01)
79 (42)
35 (19)
109
58.1 (1.2)
69 (63)
27.2 (0.4)
167.4 (0.01)
18 (16)
33 (31)
5.6 (0.1)
3.8 (0.1)
1.3 (0.1)
1.34 (0.1)
99.3 (4.1)
4.10 (0.01)
45 (41)
25 (23)
46
59.9 (1.8)
26 (56)
27.4 (0.6)
168.9 (0.01)
7 (15)
14 (30)
5.6 (0.2)
3.6 (0.1)
1.2 (0.1)
1.69 (0.1)
106.8 (6.4)
4.15 (0.01)
20 (43)
6 (13)
0.277
0.714
0.243
0.351
0.075
0.013
0.096
0.083
0.588
0.066
0.433
0.044
0.638
0.063
BMI, body mass index; ESH2003, European Society of Hypertension–European Society of Cardiology 2003 guidelines; HDL, high-density lipoprotein; LDL, low-density
lipoprotein; SE, standard error. Significant changes are indicated in bold type.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Brachial BP classification v. central hemodynamics Safar et al. 133
Table 2
Drug treatment as a function of ESH2003-brachial blood pressure classification (ESH2003-BPC)
ESH2003-BPC
Optimal
Number of antihypertensive drugs (SE)
Duration of drug treatment (months) (SE)
Diuretics (yes) (%)
Alpha-blockers (yes) (%)
Beta-blockers (yes) (%)
Calcium-channel inhibitors (yes) (%)
ACE inhibitors (yes) (%)
AT II (yes) (%)
Hypolipidemic drugs (yes) (%)
Antidiabetic drugs (yes) (%)
Aspirin (yes) (%)
2.06
8.4
18
3
23
18
18
3
10
1
16
(0.2)
(1.2)
(37)
(6)
(48)
(37)
(37)
(6)
(21)
(2)
(33)
Normal
2.55
7.9
42
15
37
41
27
4
20
4
16
(0.1)
(1.0)
(51)
(18)
(45)
(49)
(32)
(5)
(24)
(5)
(19)
High normal
2.17
9.1
32
13
41
59
39
5
23
6
18
(0.1)
(0.9)
(30)
(20)
(38)
(55)
(36)
(5)
(21)
(6)
(17)
Grade I
2.58
10.7
75
47
58
123
51
16
37
14
31
(0.1)
(0.6)
(40)
(25)
(31)
(66)
(27)
(9)
(20)
(7)
(17)
Grade II
2.61
12.7
39
26
48
71
32
9
19
12
21
(0.1)
(0.8)
(36)
(24)
(44)
(65)
(29)
(8)
(17)
(11)
(19)
Grade III
3.20
11.1
18
30
12
34
13
7
13
7
6
(0.2)
(1.3)
(39)
(65)
(26)
(74)
(28)
(15)
(28)
(15)
(13)
P value
<0.001
0.003
0.108
<0.001
0.037
<0.001
0.544
0.263
0.707
0.101
0.126
ACE, angiotensin-converting enzyme; AT II, angiotensin II receptor antagonists; ESH2003, European Society of Hypertension–European Society of Cardiology 2003
guidelines; SE, standard error. Significant changes are indicated in bold type.
disappeared only after additional adjustment for mean
BP. PWV increased significantly from the BP subgroup
with optimal control to that with the higher brachial BP
level, and this result remained significant even after
adjustment for age, HR, gender, glucose, mean BP and
beta-blockers. The PP amplification and SBP amplification also significantly increased as a function of higher BP
subgroup. As shown in Fig. 1 this gradual increase in PP
and SBP amplification was abolished after adjustment for
PWV and HR (or only PWV in the case of SBP amplification, data not shown) respectively.
Table 4 shows that increased PWV and AI are the
strongest independent predictors of brachial BP classification and are followed by the presence of atherosclerotic
lesions, age, HR, the number of antihypertensive drugs as
well as the duration of treatment. In order to adjust for
BP, mean BP was forced to enter in the previous model
either as a continuous variable (mmHg) or as a categorical
variable (MBP level 1, <104 mmHg; 2, 104 mmHg); in
both models PWV remained a major predictor of BP
classification (Table 5).
In Fig. 2 the ability of PWV to evaluate the effectiveness
of BP control (ESH2003-BPC 1–3 versus ESH2003-BPC
Table 3
4–6) in the totality of the population is shown by application of ROC curve analysis (area under the curve 0.69,
P < 0.001). In Table 6 the sensitivity of PWV in predicting effective BP control is further analyzed, according to
strata of 10 years of age. It should be mentioned that the
ability of MBP to predict uncontrolled BP was higher
than that of PWV (data not shown).
Discussion
The present study, in subjects chronically treated for
hypertension, explores the relation of aortic PWV and
carotid AI with the brachial BP classification widely used
in international guidelines [14] and compares the brachial
values to those of noninvasive carotid (central) BP
measurements. Two results emerge from this observational study. First, for each given value of mean BP, carotid
SBP and PP are significantly lower than brachial SBP and
PP; however, this difference (e.g. in PP amplification) is
significantly lower in the ‘optimal’ and ‘normal’ BP ranges
(6.8–7.4 mmHg) than in the higher BP ranges (10.1–
11.3 mmHg), depending mainly on the levels of HR and
PWV. Second, there is a constant influence of aortic PWV
on brachial BPC in treated hypertensive subjects. This
effect on ESH2003-BPC is independent of age, drug
treatment, the presence of atherosclerotic lesions and even
Central and peripheral hemodynamic parameters as a function of ESH2003-brachial blood pressure classification (ESH2003-BPC)
ESH2003-BPC
Brachial SBP (mmHg) (SE)y,M
Brachial DBP (mmHg) (SE)M
Brachial PP (mmHg) (SE)y,M
Brachial PP >60 mmHg (yes) (%)M
Brachial MBP (mmHg) (SE)M
Carotid SBP (mmHg) (SE)M
Carotid PP (mmHg) (SE)M
Heart rate (bpm) (SE)
RWTT/LVET (%) (SE)
Augmentation index (%) (SE)MM
Augmentation index (%) (SE)MMM
Pulse wave velocity (m/s) (SE)MMMM
PP amplification (mmHg) (SE)
SBP amplification (mmHg) (SE)
Optimal
112.5
68.9
43.5
0
83.4
105.8
36.7
63.5
33.7
116.1
124.9
11.0
6.8
6.6
(1.0)
(1.3)
(1.9)
(0)
(0.9)
(1.4)
(2.0)
(1.6)
(0.8)
(3.3)
(3.1)
(0.4)
(1.0)
(1.0)
Normal
123.4
73.9
49.4
6
90.4
114.7
42.0
64.9
33.8
116.1
120.9
11.2
7.4
8.6
(0.8)
(1.1)
(1.4)
(7)
(0.7)
(1.1)
(1.6)
(1.2)
(0.8)
(2.5)
(2.1)
(0.2)
(0.8)
(0.8)
High normal
133.1
77.2
55.8
37
95.8
123.4
45.8
66.4
32.9
116.9
122.1
12.2
10.1
9.7
(0.7)
(0.9)
(1.3)
(35)
(0.6)
(0.9)
(1.4)
(1.1)
(0.6)
(2.2)
(2.5)
(0.2)
(0.7)
(0.7)
Grade I
145.9
82.4
63.5
119
103.5
135.2
52.9
66.8
31.3
124.7
126.1
12.6
10.5
10.7
(0.5)
(0.7)
(0.9)
(64)
(0.5)
(0.7)
(1.1)
(0.8)
(0.5)
(1.7)
(1.9)
(0.2)
(0.5)
(0.5)
Grade II
164.1
89.4
74.7
97
114.2
152.6
63.6
66.1
31.0
129.3
125.1
13.3
11.0
11.4
(0.7)
(0.9)
(1.3)
(89)
(0.6)
(0.9)
(1.4)
(1.1)
(0.4)
(2.2)
(2.1)
(0.2)
(0.7)
(0.7)
Grade III
182.9
98.1
84.7
36
126.3
171.5
73.3
70.6
31.5
125.8
119.2
13.1
11.3
11.3
(1.1)
(1.4)
(1.9)
(78)
(0.9)
(1.5)
(2.1)
(1.6)
(0.6)
(3.4)
(3.1)
(0.4)
(1.1)
(1.1)
P value
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
0.042
<0.001
<0.001
0.152
0.001
0.002
0.003
DBP, diastolic blood pressure; ESH2003, European Society of Hypertension–European Society of Cardiology 2003 guidelines; LVET, left ventricular ejection time; MBP,
mean blood pressure; PP, pulse pressure; RWTT, reflected wave time transit; SBP, systolic blood pressure. Significant changes are indicated in bold type. y P < 0.05 for
paired-group comparison of brachial BPs versus corresponding carotid BPs. M After adjustment for age, gender, heart rate. MM After adjustment for age, gender, heart rate,
height, PWV. MMM After adjustment for age, gender, heart rate and MBP. MMMM After adjustment for age, gender, heart rate, MBP, glucose (or for the presence of metabolic
syndrome) and beta-blockers.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
134 Journal of Hypertension
2008, Vol 26 No 1
Multivariate model predicting ESH2004-brachial blood
pressure classification (ESH2003-BPC), including mean blood
pressure (MBP) in the model
Fig. 1
Table 5
Independent predictors of
ESH2003-BPC
MBP level
PWV (m/sec)
Number of drugs (n)
Presence of atherosclerosis (yes ¼ 1)
AI (%)
mean BP. Finally, PWV is a highly sensitive marker of the
effectively controlled BP throughout all age decades.
Considerations regarding the population and the
methodology
It is worth noting that, in the population studied herein,
the distribution of atherosclerotic lesions throughout the
ESH2003-BPC decreased but did not differ significantly;
however, the frequency of atherosclerotic lesions may
have been underestimated, since noninvasive and invasive techniques for the investigations of peripheral and
coronary arteries were not performed consistently.
Furthermore, the distribution of beta-blockers seems
to coincide with that of atherosclerotic lesions. This
observation may be explained by the fact that the
subjects with acknowledged higher CV risk were those
who were either ‘optimally’ treated (as well as in need of a
beta-blocking agent) regarding blood pressure control, by
their attending physicians, or more compliant to his
instructions. Finally, a gradual increase in the distribution
of calcium-channel inhibitors and mainly of alpha-blockers was observed. Although different antihypertensive
Multivariate model predicting ESH2003-brachial blood
pressure classification (ESH2003-BPC)
Table 4
Independent predictors
of ESH2003-BPC
PWV (m/s)
AI (%)
Presence of atherosclerosis (yes ¼ 1)
Age (years)
Number of drugs
Heart rate (bpm)
Duration of treatment (years)
Beta
P
0.434
0.262
0.175
0.214
0.137
0.144
0.092
<0.001
<0.001
<0.001
<0.001
<0.001
<0.001
0.028
95% Confidence
Interval for B
0.175
0.011
0.875
0.034
0.067
0.008
0.002
0.285
0.020
0.335
0.014
0.235
0.028
0.027
AI, augmentation index; ESH2003, European Society of Hypertension–European
Society of Cardiology 2003 guidelines; PWV, pulse wave velocity. Gender,
diabetes mellitus, body mass index, plasma creatinine, plasma lipids and smoking
were entered in the analysis but excluded from the final model. Beta, standardized
coefficients; P, level of statistical significance <0.05.
P
0.633
0.297
0.099
0.099
0.078
<0.001
<0.001
0.001
0.001
0.012
95% Confidence
interval for B
1.597
0.100
0.043
0.551
0.317
1.941
0.174
0.174
0.103
0.008
ESH2003, European Society of Hypertension–European Society of Cardiology
2003 guidelines. MBP level (1, MBP < 104 mmHg; and 2, MBP 104 mmHg)
was forced to enter in the previous model of Table 4. Similar results were found
when MBP (mmHg), instead of MBP level, was entered in the model (data not
shown). Beta, standardized coefficients; P, level of statistical significance <0.05.
drug treatments may have different effects on pressure
amplification, no independent specific drug effect on BP
was found in the present study (data not shown).
In the present study we used measurements of PWV and
pulse wave analysis in the overall population. We reported
central hemodynamic parameters assessed exclusively at
the carotid artery (Fittchet–Kelly method); that is,
measured directly without the application of the generalized transfer functions. In the past, we validated both
methods separately, using, for each of them, simultaneous
determinations of intra-arterial BP measurements
[21,22,24] and we have reported that in the present population similar results regarding central BP were found
independently of the methodology [13]. Finally, the
noninvasively acquired carotid BP was considered as a
close surrogate of intra-arterial aortic BP. In fact, the main
problem of such methodologies is that a noninvasive
calibration of radial and carotid artery BP curves requires
Fig. 2
1.0
0.8
Sensitivity
P values of the difference in pulse pressure (PP) amplification (mmHg)
between European Society of Hypertension–European Society of
Cardiology 2003 guidelines (ESH2003)-brachial blood pressure
classification groups, after adjustment for different models
(model 1, none; model 2, age and gender; model 3, age, gender and
heart rate (HR); model 4, age, gender and pulse wave velocity (PWV);
model 5, age, gender, HR and PWV) (assessed by ANOVA).
Beta
0.6
0.4
0.2
0.0
0.0
0.2
0.4
0.6
1 - Specificity
0.8
1.0
The ability of pulse wave velocity (PWV, m/s) to evaluate the
classification of hypertensives as effectively treated [European Society
of Hypertension–European Society of Cardiology 2003 guidelines
blood pressure classification (ESH2003-BPC) 1–3] or ineffectively
treated (ESH2003-BPC 4–6) was assessed by receiver operator curve
(ROC) analysis in the totality of the population. Area under the curve
0.69, P < 0.001.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Brachial BP classification v. central hemodynamics Safar et al. 135
Table 6 The sensitivity of different pulse wave velocity (PWV) cut-off
levels (increment of 1 m/s) for the evaluation of the classification of
hypertensives as ineffectively treated (ESH2003-BPC grade I to III),
as assessed by ROC curve analysis, according to decades of age
Age
Area under the curve
Significance
<40
n ¼ 35
0.81
41–50 51–60 61–70 71–80
>80
n ¼ 100 n ¼ 162 n ¼ 156 n ¼ 102 n ¼ 35
0.71
0.65
<0.001 0.002
0.004
PWV (m/s) >
10
11
12
13
14
15
16
17
18
0.70
0.69
0.81
<0.001 0.004
0.015
Sensitivity (%)
72
99
–
–
–
–
–
–
–
40
80
93
–
–
–
–
–
–
31
49
68
80
90
–
–
–
–
19
37
54
78
86
94
–
–
–
15
27
39
48
63
81
88
93
–
10
15
21
55
66
77
89
89
90
ESH2003-BPC, European Society of Hypertension–European Society of Cardiology 2003 guidelines-blood pressure classification.
a constantly adequate concomitant measurement of the
brachial artery BP. It is widely accepted that when
noninvasive brachial BP transcutaneous measurements
are used, the determination of brachial SBP is accurate.
In contrast, accuracy is not the case for DBP [25], but
because DBP (but not SBP) is nearly the same in all parts of
the arterial tree, these errors may be minimized. Finally,
we adjusted all the results of this investigation for age and
sex, due to the well-established relationships of these two
variables with arterial properties.
Another concern regarding BP classification in the present study involves the role of arterial stiffness and total
peripheral resistance in the genesis of the final pattern
(especially of SBP) in the present population. Since the
mean age of subjects studied was around 60 years, we
speculate that SBP was highly influenced by increased
arterial stiffness. The ESH2003 BP classification does not
distinguish between high SBP due to arterial stiffness or
peripheral resistance, thus the present results may not be
extrapolated to younger, healthier subjects with a less
stiff aorta.
Considerations regarding the findings
The results clearly show that the SBP and PP amplifications are lower in the subjects who are effectively treated
and mostly at the ‘optimal’ and ‘normal’ BP ranges than in
the higher BP ranges. This latter condition suggests the
presence of ‘masked’ increased left ventricular afterload,
since the decreased PP amplification in the ‘optimal’
subgroup implies that central PP is too high relative to
the ‘optimal’ PP. This may potentially explain results
observed in treated hypertensive subjects, in which a
consistent residual CV risk is observed even in the
presence of SBP <140 mmHg [26]. On the other hand,
the most important difficulty with the present data is the
cross-sectional nature of the investigation, and therefore
the lack of pretreatment BP values. Although previous
longitudinal studies support the present results [5], further
prospective studies are required to verify these data.
In accordance with previous results [13], we showed that
PWV and HR were two cardinal modulators of BP amplification. Because PWV depends not only on the distending
BP but also on the structure of the arterial wall, and
because HR is modulated by the autonomic nervous
system, these findings might suggest that altered baroreflex mechanisms are present, and that arterial stiffening
may contribute to modify heart rate control in hypertensive
subjects. Several previous examples in the literature
support this association [27,28], which remains, however,
to be explored specifically in the particular case of patients
with ‘optimal’ or ‘normal’ BP.
It should be also noted that another classical parameter that
modulates BP amplification [13], namely the pressure
wave reflections assessed by carotid AI, was increased
throughout the six subgroups of the ESH2003-BPC.
Although heart rate increased as a function of higher
brachial BP subgroup, an early shift in the timing
(RWTT/LVET) of wave reflections at the level of the
central arteries was observed and could be explained
potentially by the increased PWV. Early arrival of wave
reflections is a classical cause of high AI, yet our
data suggest that mean BP, and thus a rise in peripheral
resistance and reflection coefficient causing reflection
points to ‘come closer to the heart’, rather than high
PWV, was the main reason for the gradual increase in
AI. It is important to note that although pressure wave
reflections at the level of the carotid artery increased from
‘optimal’ to ‘grade III’ group, BP amplification did not
decrease but increased. This phenomenon, although
strange at first glance, may be explained potentially on
the basis of increased carotid stiffness and thus altered
baroreflexes, as well as being due to the previously
described differences of drug treatment on small and large
arteries [5]. These results imply the presence of higher
augmentation of the radial rather than the central pressure
wave from the reflected wave. The phenomenon of peripheral (radial or brachial) pressure wave augmentation is
poorly studied and understood. It should be clarified that
AI estimated at the level of the aorta (as in the present
study) may have a good linear correlation with AI at
the level of the radial artery, but they are not identical.
Moreover, it is logical to assume that the reflected wave at
the level of the arterioles of the hand largely enhances
the radial pressure wave, since the reflection site and the
measurement sites are very close, leading to systolic
timing.
An important result of this study was that, independently
of age, mean BP, drug treatment and atherosclerotic
lesions, aortic PWV predicted the results of ESH2003BPC in subjects under drug treatment. Increased PWV is
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
136 Journal of Hypertension
2008, Vol 26 No 1
well accepted as a significant independent predictor of
cardiovascular risk in hypertensive subjects [1]. In the past,
increased collagen cross-links of large vessels have been
proposed to explain increased aortic stiffness independently of mechanical factors in hypertensive subjects
[29]. More recently, we have shown that the progression
of aortic PWV with age is accelerated in treated hypertensive subjects, and that this increased progression
was specifically enhanced in subjects with metabolic syndrome [30]. The present results imply that mechanisms
such as endothelial dysfunction, with resulting oxidative
stress, might play a role in the observed changes in PWV.
Finally, it is worth noting that the independent predictive
value of PWV on ESH2003-BPC was reinforced by the
mean BP adjustment. This is very important, if we consider that mean BP is almost constant throughout the
arterial system and thus independent of the site of BP
measurement (brachial or carotid artery). On the other
hand, the inverse correlation of atherosclerotic lesions with
BP classification most probably reflects the more efficient
management of those subjects at higher CV risk, as
discussed previously.
Although limited by the cross-sectional design of the
study, our results imply that aortic stiffness might be a
marker of the degree of response or effectiveness of drug
treatment. At intervals of 10 years of age, cut-off levels
for PWV were shown to provide >90% sensitivity for
recognizing unsuccessfully treated subjects. The lack of
baseline (before treatment) values for both PWV and BP
values as well as the fact that the pressure-dependent
component of arterial stiffness (due to passive arterial
distention) may partly explain our results and impose the
need for a cautious interpretation of the present data. Of
course, as expected, MBP had a higher predictive value
for BP classification and effective BP control than PWV
but, as previously discussed, their effects were independent of each other. Further investigation of the predictive
role of PWV, regarding both peripheral and central BP
classification as well as the response to drug treatment, is
warranted, especially from data on 24 h ambulatory BP
monitoring and prospective studies.
In conclusion, arterial stiffness and the autonomic
nervous system, independently of age and MBP, seem
to explain the observed decrease of BP amplification in
successfully treated hypertensive subjects. This study
implies that aortic PWV might be a useful marker
for predicting the response or effectiveness of drug
treatment. This hypothesis remains to be verified in
longitudinal studies.
References
1
2
Blacher J, Asmar R, Djane S, London G, Safar M. Aortic pulse wave velocity
as a marker of cardiovascular risk in hypertensive patients. Hypertension
1999; 33:1111–1117.
Sleight P, Yusuf S, Pogue J, Tsuyuki R, Diaz R, Probstfield J. Bloodpressure reduction and cardiovascular risk in HOPE study. Lancet 2001;
358:2130–2131.
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Poulter NR, Wedel H, Dahlof B, Sever PS, Beevers DG, Caulfield M, et al.
Role of blood pressure and other variables in the differential cardiovascular
event rates noted in the Anglo-Scandinavian Cardiac Outcomes
Trial-Blood Pressure Lowering Arm (ASCOT-BPLA). Lancet 2005;
366:907–913.
Hirata K, Vlachopoulos C, Adji A, O’Rourke MF. Benefits from angiotensinconverting enzyme inhibitor ‘beyond blood pressure lowering’: beyond blood
pressure or beyond the brachial artery? J Hypertens 2005; 23:551–556.
London GM, Asmar RG, O’Rourke MF, Safar ME. Mechanism(s) of
selective systolic blood pressure reduction after a low-dose combination of
perindopril/indapamide in hypertensive subjects: comparison with atenolol.
J Am Coll Cardiol 2004; 43:92–99.
Williams B, Lacy PS, Thom SM, Cruickshank K, Stanton A, Collier D,
et al. Differential 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.
Safar ME, Blacher J, Pannier B, Guerin AP, Marchais SJ, Guynovarc’h JP,
London G. Central pulse pressure and mortality in end-stage renal disease.
Hypertension 2002; 39:735–738.
Protogerou AD, Papaioannou TG, Blacher J, Papamichael Ch, Lekakis J,
Safar M. Central blood pressures: do we need them in the management of
cardiovascular disease? Is it a feasible therapeutic target? J Hypertens
2007; 25:165–272.
Pauca AL, Wallenhaupt SL, Kon ND, Tucker WY. Does radial artery
pressure accurately reflect aortic pressure? Chest 1992; 102:1193–
1198.
Kroeker EJ, Wood EH. Comparison of simultaneously recorded central and
peripheral arterial pressure pulses during rest, exercise and tilted position in
man. Circ Res 1955; 3:623–632.
Rowell LB, Brengelmann GL, Blackmon JR, Bruce RA, Murray JA.
Disparities between aortic and peripheral pulse pressures induced by
upright exercise and vasomotor changes in man. Circulation 1968;
37:954–964.
Latham RD, Westerhof N, Sipkema P, Rubal BJ, Reuderink P, Murgo J.
Regional wave travel and reflections along the human aorta: a study with six
simultaneous micromanometric pressures. Circulation 1985; 72:1257–
1269.
Protogerou AD, Blacher J, Mavrikakis M, Lekakis J, Safar ME. Increased
pulse pressure amplification in treated hypertensive subjects with
metabolic syndrome. Am J Hypertens 2007; 20:127–133.
European Society of Hypertension–European Society of Cardiology. 2003
European Society of Hypertension–European Society of Cardiology
Guidelines for the management of arterial hypertension. J Hypertens 2003;
21:1011–1053.
Lokkndon GM, Guerin AP, Marchais SJ, Pannier B, Safar ME, Day M,
Metivier F. Cardiac and arterial interactions in end-stage renal disease.
Kidney Int 1996; 50:600–608.
Verbeke F, Segers P, Heireman S, Vanholder R, Verdonck P,
Van Bortel LM. Noninvasive assessment of local pulse pressure:
importance of brachial-to-radial pressure amplification. Hypertension
2005; 46:244–248.
Papaioannou TG, Lekakis JP, Karatzis EN, Papamichael CM,
Stamatelopoulos KS, Protogerou AD, et al. Transmission of
calibration errors (input) by generalized transfer functions to the aortic
pressures (output) at different hemodynamic states. Int J Cardiol 2006;
110:46–52.
Chen CH, Ting CT, Nussbacher A, Nevo E, Kass DA, Pak P, et al.
Validation of carotid artery tonometry as a means of estimating
augmentation index of ascending aortic pressure. Hypertension 1996;
27:168–175.
Chen CH, Nevo E, Fetics B, Pak PH, Yin FCP, Maughan L,
Kass DE. Estimation of central aortic pressure waveform by mathematical
transformation of radial tonometry pressure: validation of generalized
transfer function. Circulation 1997; 95:1827–1836.
Wilkinson IB, Fuchs SA, Jansen IM, Spratt JC, Murray GD,
Cockroft JR, Webb DJ. Reproducibility of pulse wave velocity and
augmentation index measured by pulse wave analysis. J Hypertens 1998;
16:2079–2084.
Smulyan H, Siddiqui DS, Carlson RJ, Lokkndon GM, Safar ME.
Clinical utility of aortic pulses and pressures calculated from applanated
radial-artery pulses. Hypertension 2003; 42:150–155.
Murgo JP, Westerhof N, Giolma JP, Altobelli SA. Aortic input impedence in
normal men: relationship to pressure waveform. Circulation 1980;
62:105–116.
Asmar R, Benetos A, Topouchian J, Laurent P, Pannier B, Brisac AM, et al.
Assessment of arterial distensibility by automatic pulse wave velocity
measurement: validation and clinical application studies. Hypertension
1995; 26:485–490.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Brachial BP classification v. central hemodynamics Safar et al. 137
24
25
26
27
28
29
30
Benetos A, Tsoucris-Kupfer D, Favereau X, Corcos T, Safar M.
Carotid artery tonometry: an accurate noninvasive method for central
aortic pulse pressure evaluation. J Hypertens 1991; 9 (Suppl 6):S144–
S145.
Safar ME, Smulyan H. Systolic versus diastolic blood pressure. In: Bulpitt
CJ, editor. Handbook of hypertension. Amsterdam: Elsevier Science BV;
2000. pp. 73–85.
Blacher J, Evans A, Arveiller D, Amouyel P, Ferrieres J, Bingham A,
et al. Residual coronary risk in men aged 50–59 years treated for
hypertension and hyperlipidaemia in the population: the PRIME study.
J Hypertens 2004; 22:415–423.
Monahan KD, Dinenno FA, Seals DR, Clevenger CM, Desouza CA,
Tanaka H. Age-associated changes in cardiovagal baroreflex sensitivity are
related to central arterial compliance. Am J Physiol Heart Circ Physiol
2001; 281:H284–H289.
Monahan KD, Tanaka H, Dinenno FA, Seals DR. Central arterial compliance
is associated with age and habitual exercise-related differences in
cardiovagal baroreflex sensitivity. Circulation 2001; 104:1627–1632.
Kass DA, Shapiro EP, Kawagushi M. Improved arterial compliance by a
novel advanced glycation end-product cross-link breaker. Circulation
2001; 104:1464–1470.
Safar ME, Thomas F, Blacher J, Nzietchueng R, Bureau JM, Pannier B,
Benetos A. Metabolic syndrome and age-related progression of aortic
stiffness. J Am Coll Cardiol 2006; 47:72–75.
Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.