Received: 21 October 2020 | Revised: 13 December 2020 | Accepted: 19 December 2020

DOI: 10.1111/jch.14209

ORIG INAL PAPER

Hypertension-­mediated organ damage regression associates

with blood pressure variability improvement three years after
successful treatment initiation in essential hypertension

Helen Triantafyllidi MD, PhD | Dimitrios Benas MD | Antonios Schoinas MD |

Dionyssia Birmpa MD | Paraskevi Trivilou MD | Efthimia Varytimiadi MD |
Dimitrios Voutsinos MD | Ignatios Ikonomidis MD, PhD

2nd Department of Cardiology Medical

School, University of Athens, ATTIKON Abstract
Hospital, Athens, Greece
Blood pressure variability (BPV) has been associated with the development, progres-
Correspondence sion, and severity of cardiovascular (CV) organ damage and an increased risk of CV
Helen Triantafyllidi, 2nd Cardiology
morbidity and mortality. We aimed to explore any association between short-­term
Department, Attikon Hospital, Medical
School, University of Athens, 83, Agiou BPV reduction and hypertension-­mediated organ damage (HMOD) regression in hy-
Ioannou Theologou, Holargos 155 61,
pertensive patients 3-­year post-­treatment initiation regarding BP control. 24-­h am-
Athens, Greece.
Email: seliani@hotmail.com bulatory blood pressure monitoring (24 h ABPM) was performed at baseline in 180
newly diagnosed and never-­treated hypertensive patients. We measured 24 h average
systolic (24 h SBP) and diastolic BP (24 h DBP) as well as 24 h systolic (sBPV) and dias-
tolic BPV (dBPV). Patients were initially evaluated and 3 years later regarding arterial
stiffness (PWV), left ventricular hypertrophy (LVMI), carotid intima-­media thickness
(cIMT), 24 h microalbumin levels (MAU), and coronary flow reserve (CFR). Successful
BP treatment was defined as 24 h SBP/DBP < 130/80 mm Hg based on 2nd ABPM and
subsequently, patients were characterized as controlled (n = 119, age = 53 ± 11 years)
or non-­controlled (n = 61, age = 47 ± 11 years) regarding their BP levels. In the whole
population and the controlled group, 24 h SBP/DBP, sBPV/dBPV, LVMI, and IMT were
decreased. Additionally, LVMI improvement was related with both sBPV (p < .001) and
dBPV reduction (r = .18, p = .02 and r = .20, p = .03, respectively). In non-­controlled
hypertensives, PWV was increased. In multiple linear regression analysis, sBPV and
dBPV reduction predicted LVMI improvement in total population and controlled
group independently of initial office SBP, mean BP, and 24 h-­SBP levels. In middle-­
aged hypertensive patients, a 3-­year antihypertensive treatment within normal BP
limits, confirmed by 24-­h ABPM, leads to CV risk reduction associated with sBPV and
dBPV improvement.

This is an open access article under the terms of the Creative Commons Attribution-­NonCommercial License, which permits use, distribution and reproduction
in any medium, provided the original work is properly cited and is not used for commercial purposes.
© 2021 The Authors. The Journal of Clinical Hypertension published by Wiley Periodicals LLC.

1150 | 
wileyonlinelibrary.com/journal/jch J Clin Hypertens. 2021;23:1150–1158.
TRIANTAFYLLIDI et al. | 1151

1 | I NTRO D U C TI O N transthoracic echocardiogram (TTE) in order to evaluate LVMI as

well as CFR of left anterior descending artery (LAD), (6) carotid ultra-
Arterial hypertension (AH), is one of the most common diseases sonography for cIMT measurement, (7) carotid-­femoral pulse wave
worldwide with a 30%–­45% prevalence among several regions which velocity (PWV) to evaluate arterial stiffness and (8) microalbumin
increases with advanced age.1,2 It represents a well-­established levels measurement (MAU) in 24 h urine collection.
risk factor for cardiovascular disease (CVD) with significant mor- Informed consent was obtained during the initial visit of the
bidity and mortality and often coexists with other cardiovascular study which was approved by the ethical committee of our hospital.
(CV) risk factors (dyslipidemia, glucose intolerance, diabetes mel- Patients with secondary hypertension, congestive heart failure,
litus) which all together contribute further to increased total CV previous myocardial infarction, stroke, cardiac valve diseases, his-
risk.3,4 AH leads to a wide spectrum of subclinical organ damages tory of coronary artery by-­pass grafting, atrial fibrillation, renal in-
(hypertension-­mediated organ damage, HMOD) or overt clinical dis- sufficiency, overt proteinuria, anemia or other hematologic disorder,
eases (ie coronary artery disease, heart failure, stroke, and chronic as well as those patients on medication for cardiovascular (except
kidney disease). 2 HMOD refers to a subclinical intermedium stage of statins for hyperlipidemia treatment) or non-­cardiovascular diseases
2
the cardiovascular continuum affecting various target-­organs and or hormonal replacement for any reason were excluded from the
leading to left ventricular hypertrophy (LVH), reduced coronary flow study. Conclusively, the participants in our study neither had any
reserve (CFR), increased carotid intima-­media thickness (cIMT), reti- concomitant disorders nor received any cardio-­metabolic medi-
nopathy, microalbuminuria increased aortic stiffness and endothelial cations and subsequently they constitute a homogenous group of
5
dysfunction. As HMOD relates to increased morbidity and mortal- newly diagnosed hypertensives.
ity, blood pressure (BP) treatment targets to HMOD prevention and/
or regression besides BP control. 2,5
BP is characterized by continuous and significant changes (blood 2.2 | Diagnostic work-­up
pressure variability, BPV) beat-­to-­beat (very short-­term BPV), over
24-­h (short-­term BPV), day-­to-­day (mid-­term BPV) and from visit-­ The protocol of the study has been described in details in a previous
to-­visit (long-­term BPV). AH diagnosis is usually based on 24 h am- study by our research group.10 However, a short description follows:
bulatory blood pressure monitoring (24 h ABPM), which provides a
variety of information during a 24-­h period apart from single blood
pressure (BP) measurements, like nocturnal BP recording, dipping 2.2.1 | Office BP measurement
status and BPV, useful for AH diagnosis as well as for treatment
evaluation. 2,6,7 BPV has been associated with the development, Morning office BP was measured in the hospital outpatient clinic, ap-
progression, and severity of CV organ damage and an increased proximately at the same morning hour, by the same cardiologist using
risk of cardiovascular morbidity and mortality.8 However, the actual a mercury sphygmomanometer {first and fifth phases of Korotkoff
size of its independent contribution to CV risk remains unknown.9 sounds taken as systolic (SBP) and diastolic (DBP) blood pressure,
Accordingly, we aimed to explore any existing relationship between respectively} after the patients had rested for 5–­10 min in the sitting
short-­term BPV improvement, derived by 24 h ABPM, and HMOD position while they were advised to avoid smoking or drinking cof-
regression in recently diagnosed and never-­treated hypertensive pa- fee for at least 2 h before examination. Three measurements were
tients 3 years after medical treatment initiation. taken at 1 min intervals, and the average was used as the office SBP
and DBP. Hypertension was diagnosed as SBP ≥ 140 mm Hg and/or
DBP ≥ 90 mm Hg. 2 PP was defined as SBP-­DBP while mean BP as
2 | PATI E NT S A N D M E TH O DS DBP + PP/3.

2.1 | Study population

2.2.2 | Ambulatory BP monitoring
We studied 350 consecutive Caucasian hypertensive patients with
recently diagnosed and never-­treated stage I-­II essential hyperten- ABPM was carried out 1–­4 days after the first evaluation of each
sion according to the 2018 guidelines of the European Society of patient in the hypertension outpatient clinic on the non-­d ominant
2
Hypertension (ESH) visiting our outpatient ESH Excellence Centre. arm using validated Spacelab 90207 (Spacelab) recorders. The
All patients were subjected to the following examinations within ABPM device was set to obtain BP readings at 15 min intervals
2 weeks: (1) The average of three (3) office BP measurements taken during the day (07.00–­23.00) and at 20 min intervals during the
in the hypertension outpatient clinic was considered as office BP night (23.00–­07.00). The time of application and the type of the
(systolic and diastolic); (2) blood and urine sampling for routine device were the same in all patients. The patients were instructed
blood chemistry (lipid profile included) and urine examination; (3) to attend their usual day-­to-­day activities but to keep still at the
standard 12-­lead electrocardiogram; (4) 24 h ABPM in order to con- times of measurements. While ABP monitoring was obtained dur-
firm hypertension diagnosis based on office BP measurements; (5) ing working days (Monday–­Friday), patients were asked to go to
1152 | TRIANTAFYLLIDI et al.

bed not later than 23.00 and to stay in bed until 07.00. If this was (angiotensin converting enzyme inhibitors or sartans) alone or in
not acceptable, information from their diaries was taken in order double combination with calcium blockers or hydrochlorothiazides
to correctly obtain data from daily and night activities according or in triple combination (RAAS inhibitors plus calcium blockers
individual patient's schedule. Recordings were analyzed to obtain plus hydrochlorothiazides). Patients were followed by our ESH
24 h, daytime and nighttime average SBP, DBP, PP and heart rates. Excellence Centre every 3–­6 months during scheduled visits. At
Systolic readings >260 or <70 mm Hg and diastolic readings >150 baseline 350 hypertensive patients were recruited. However, only
or <40 mm Hg were discarded. In order to define ABPM as valid, 200 (57%) were re-­evaluated at 3 years after treatment initiation fol-
each patient had to have no fewer than 3 successful readings per lowing the same protocol as at baseline evaluation (office BP mea-
hour during daytime and 2 during night-­t ime and ≥70% of success- surements, ABPM, assessment of HMOD). The rest 150 patients
ful readings. In only six patients (3.3%), ABPM did not meet the were lost during the follow-­up period or they refused to be submit-
above definition of validity and the patient had to repeat it dur- ted in the re-­evaluation protocol. Finally, we present results from
ing the next day. Systolic and diastolic BPV (sBPV, dBPV) were 180 patients, since we found incomplete diagnostic documentation
defined as the standard deviation of 24 h average SBP and DBP. 2 at re-­evaluation in 20/200 patients.
ΔsBPV (or ΔdBPV) was defined as sBPV at baseline minus sBPV at
3-­year post-­t reatment initiation (or dBPV at baseline minus dBPV
at 3-­year post-­t reatment initiation). 2.4 | Statistical analysis

The Shapiro-­Wilk test was used to assess the normality of dis-

2.3 | Hypertension-­mediated organ damage tribution. Almost all variables were normally distributed and are
(HMOD) evaluation expressed as mean ± SD or % incidence. However, weight, BMI,
HDL-­C , 24-­h average SBP, PWV, MAU, E/Ea, LVMI, IMT, CFR (both
a. Left ventricular hypertrophy (LVH) was estimated by LVMI, at baseline and after 3 years) as well as sBPV after 3 years were not
using the Devereux formula according to the Penn Convention normally distributed and were presented as median value plus 25%–­
Protocol with a Vivid 7 system (GE Medical Systems). LV hyper- 75% interquartile range (IQR). Categorical variables are expressed as
trophy was defined as LVMI index >115 g/m2 in men and >95 g/ absolute values and percentages. Paired sample t test and Wilcoxon
2 11
m in women. ΔLVMI was defined as LVMI at baseline minus signed-­ranked test were used for comparisons regarding normally
LVMI at 3-­year post-­treatment initiation. and non-­normally distributed parameters, respectively between
b. Coronary flow reserve was estimated by coronary velocity profiles the same group of patients at baseline and 3 years after treatment.
in the left anterior descending artery obtained by color-­guided Independent sample t test and Mann-­Whitney test were used for
pulse wave Doppler from long axis apical projections after adenos- normally and non-­normally distributed parameters, respectively
ine infusion (140 μg/kg/min) for 3 min. CFRD < 2 has been consid- in order to compare differences between two different groups of
ered as abnormal, 2–­2.5 as borderline normal and >2.5 as normal. patients. Finally, chi-­squared test was used for the comparison of
c. Carotid intima-­media thickness (cIMT) was measured by ultraso- categorical variables.
nography in 3 paired segments of both carotid arteries (at the level Pearson's analysis was used to identify any existing relation-
of the common carotid artery, the carotid bulb and the internal ships between changes regarding sBPV/dBPV and HMOD (LVMI, E/
carotid artery). In each segment, 3 measurements of the maximal Ea, CFR, MAU, IMT, and PWV) at 3-­year post-­treatment initiation.
cIMT in the far wall were averaged. The average cIMT of all 6 seg- Multiple linear regression analysis, using backward method, was per-
ments was calculated; a cIMT < 0.09 cm was considered as normal. formed in order to explore any independent relationships between
d. Carotid-­femoral PWV: Aortic stiffness was estimated by an au- differences in sBPV (ΔsBPV) or dBPV (ΔdBPV) and LVMI (ΔLVMI)
tomatic carotid-­femoral PWV measurement using a Complior SP in the whole population and controlled hypertensives, separately.
(Artech Medical), a computerized device that permits automatic Age, BMI, cholesterol, BP, and smoking at baseline evaluation were
calculation of PWV. The same examiner, who was blinded to the forced in the model as independent variables. Due to collinearity be-
patient's history, performed all measurements. Patients were ad- tween office and 24-­h BP parameters, three models were examined
vised to avoid smoking or coffee at least for 2 h before examina- (Models A, B, C); in each one we used another method of baseline BP
tion. PWV < 12 m/s was considered as normal. 2 evaluation (office SBP, office mean BP, and 24-­h SBP). The level of
e. e.MAU levels in 24 h urine collection: MAU was analyzed by neph- significance was determined as two-­sided p < .05. Statistical analysis
elometry (Immunochemical assay, BN, Prospec, Dade Behring). was performed on a SPSS 23 version (SPSS Inc).
Patients were classified as normoalbuminuric (NA) when micro-
albuminuria levels were <30 mg/24 h and microalbuminuric (MA)
when microalbuminuria levels were between 30 and 300 mg/24 h.2 3 | R E S U LT S

When baseline evaluation was completed, antihyperten- Demographic and clinical characteristics of the total population (n = 180)
sive treatment was initiated. The latter included RAAS inhibitors and studied groups (controlled and non-­controlled hypertensives) at
TA B L E 1 Study population demographic and clinical characteristics at baseline and 3 years after treatment initiation

Controlled vs. non-­ Controlled vs. non-­

Total population Controlled hypertensives Non-­controlled hypertensives controlled at baseline controlled at re-­evaluation

Characteristics Baseline Re-­evaluation p Baseline Re-­evaluation p Baseline Re-­evaluation p p P

N 180 180 –­ 119 119 –­ 61 61 –­ –­ –­

TRIANTAFYLLIDI et al.

Sex (M/F) 116/64 116/64 –­ 66/53 66/53 –­ 50 (82%) 50 (82%) –­ –­ –­

Age (years) 51 ± 12 54 ± 12 –­ 53 ± 11 56 ± 11 –­ 47 ± 11 50 ± 11 –­ .001 .001
Weight (kg) 84 (74–­96) 83 (74–­95) .79 82 (73–­94 82 (73–­94) .49 86 (78–­99) 87 (76–­98) .64 .081 .09
BMI (kg/m2) 29 (27–­32) 29 (27–­32) .51 29 (27–­33) 29 (27–­32) .36 29 (26–­32) 29 (26–­32) .89 .486 .42
Current smokers 39 (22%) 29 (16%) .08 19 (16%) 13 (11%) .26 20 (33%) 16 (26%) .18 .01 .03
(%)
Cholesterol (mg/ 214 ± 34 200 ± 32 .001 214 ± 35 193 ± 32 .01 214 ± 33 205 ± 33 .02 .917 .04
dl)
LDL-­C (mg/dl) 136 ± 34 124 ± 29 .002 135 ± 33 118 ± 28 .008 137 ± 35 134 ± 29 .29 .634 .04
HDL-­C (mg/dl) 48 (42–­58) 50 (42–­62) .90 51 (44–­63) 53 (43–­6 4) .57 44 (40–­51) 46 (41–­57) .35 .001 .04
Blood pressure and blood pressure variability parameters
Office SBP 145 (135–­ 131 (125–­140) <.001 145 (135–­ 130 (124–­140) <.001 145 (135–­ 135 (125–­144) <.001 .730 .23
(mm Hg) 160) 160) 160)
Office DBP 90 (85–­ 82 (80–­90) <.001 90 (80–­95) 80 (77–­85) <.001 95 (85–­100) 85 (80–­90) <.001 .08 .003
(mm Hg) 100)
24-­h SBP 137 (131–­ 122 (117–­129) <.001 134 (130–­ 119 (114–­123) <.001 142 (136–­ 132 (127–­138) <.001 <.001 <.001
(mm Hg) 144) 140) 150)
24-­h DBP 87 ± 9 75 ± 8 <.001 86 (77–­95) 72 (68–­76) <.001 91 (82–­100) 83 (81–­87) <.001 <.001 <.001
(mm Hg)
24-­h HR 77 ± 8 73 ± 7 <.001 77 ± 8 72 ± 8 <.001 78 ± 8 75 ± 7 .005 .28 .002
(beats/
min)
sBPV (mm Hg) 15 ± 3 13 (11–­15) .002 15 ± 3 13 ± 3 <.001 14 ± 3 13 ± 3 0.01 .08 .51
dBPV (mm Hg) 13 ± 3 11 ± 2 <.001 11 ± 2 11 ± 2 <.001 13 ± 3 11 ± 3 <.001 .96 .31
Hypertension-­mediated organ damage
LVMI (g/m2) 76 (67–­92) 73 (64–­89) .01 76 (68–­92) 71 (63–­89) .01 76 (67–­93) 79 (69–­90) .36 .94 .04
E/Ea 6.7 (5.4–­ 6.7 (5.4–­8) .32 6.7 (5.5–­ 6.7 (5.4–­8.2) .68 6.7 (5.2–­ 6.7 (5.3–­7.7) .27 .58 .38
8.7) 8.7) 8.6)
PWV (m/s) 10.8 (9.3–­ 10.8 (9.4–­12.1) .35 10.8 (9.3–­ 10.7 (9.3–­12) .60 10.8 (9.3–­ 11 (10–­13) .01 .68 .09
12.5) 12.6) 11.8)
MAU 10 (7–­16) 8 (6–­12) <.001 9 (6–­14) 7 (5–­11) .002 11 (8–­20) 9 (6–­16) .09 .01 .11
|

(mg/24 h)
1153

(Continues)
1154 | TRIANTAFYLLIDI et al.

baseline and 3-­year post-­treatment initiation are listed in Table 1. Patients

pressure variability; LVMI, left ventricular mass index; E/Ea, E wave (transmitral)/to Ea wave (Tissue Doppler Imaging ratio) ; PVW, pulse wave velocity; MAU, microalbumin; CFR, coronary flow reserve; 8
controlled at re-­evaluation

Abbreviations: BMI, body mass index; LDL-­C/HDL-­C , low/high density lipoprotein cholesterol; SBP/DBP, systolic/diastolic blood pressure; 24-­h, 24 h; HR, heart rate; sBPV/dBPV, systolic/diastolic blood
in the whole population were middle-­aged (age = 51 ± 12 years), mostly
males (64%), non-­smokers (78%), over-­weighted (BMI = 29 kg/m2) with

Controlled vs. non-­

a minor prevalence of diabetes mellitus (7%).
Successful BP treatment was defined as 24 h SBP/
DBP < 130/80 mm Hg based on 2nd ABPM and subsequently, pa-
tients were characterized as controlled (n = 119, age = 53 ± 11 years,

.54

.23
P
56% males) or non-­controlled (n = 61, age = 47 ± 11 years, 82%
males) regarding their BP levels. We compared the two groups of
controlled at baseline
Controlled vs. non-­

controlled and non-­controlled hypertensives regarding their base-

line characteristics. It appears that controlled hypertensives, were
older (p = .001) with similar BMI, Cholesterol and LDL, office SBP
and DBP levels, sBPV and dBPV, PWV, LVMI, E/Ea, cIMT, CFR com-
pared to the “non-­controlled hypertensives”. However, controlled
.06

.31
p

hypertensives had lower baseline 24-­h SBP (p < .001), 24-­h DBP
(p < .001) and MAU (p = .01) and higher HDL-­C levels (p = .001)
.53

.03

compared to non-­controlled ones. 15 controlled hypertensives (6/9

p

males/females) and 6 non-­controlled ones (5/1 males/females) had

Non-­controlled hypertensives

(0.08–­0.11)
Re-­evaluation

LVH at baseline evaluation.

2.7 (2.1–­3.4)

At re-­evaluation, controlled hypertensives had similar BMI,

office SBP levels, sBPV and dBPV, PWV, LVMI, E/Ea, cIMT, CFR
0.09

compared to non-­controlled hypertensives. However, controlled

hypertensives had lower office DBP (p = .003), 24-­h SBP (p < .001),
0.1 (0.09–­
2.6 (2.3–­
Baseline

0.12)

24-­h DBP (p < .001), Cholesterol, LDL (p < .05) and LVMI (p < .05)
3.2)

and higher HDL-­C levels (p < .05) compared to non-­controlled ones.

When we studied the 3-­year post-­treatment changes in the
whole population, we found that Cholesterol (p = .001), LDL-­C
.002
.23

(p = .002), office SBP and DBP, 24-­h SBP and 24-­h DBP (p < .001),
p

MAU levels (p < .001), cIMT (p < .001), and LVMI (p = .01) were im-
(0.08–­0.10)
Re-­evaluation

proved. Additionally, sBPV (p = .002) and dBPV (p < .001) were also
Controlled hypertensives

2.6 (2–­3.2)

reduced. Similar results were found in controlled hypertensives, that is

Cholesterol (p = .01), LDL-­C (p = .008), office SBP and DBP, mean 24 h
0.09

SBP and 24-­h DBP (p < .001), sBPV and dBPV (p < .001), MAU levels
and cIMT (p = .002) and LVMI (p = .01) were decreased. However, in
0.1 (0.08–­
2.5 (2–­2.9)
Baseline

0.12)

non-­controlled hypertensives, Cholesterol (p = .03), office SBP/DBP,

24-­h SBP/DBP (p < .001), sBPV (p = .01) and dBPV (p < .001) and IMT
Note: Data are presented as mean ± SD or % or median (25–­75 IQR).

(p = .04) were decreased while PWV was increased (p = .02).

We have to mention that at re-­evaluation, white coat hyper-
<.001
.55

tension phenomenon (WCH) was found in 34/119 (29%) controlled

p

hypertensives. On the other hand, the masked hypertension phe-

(0.08–­0.11)
Re-­evaluation

nomenon was revealed in 36/61 (59%) non-­controlled hypertensives.

2.6 (2–­3.2)

In turn, we performed Pearson's correlation analysis and we

found the following relationships between:
0.09
Total population

Italics indicate statistical significance.

cIMT, carotid intima-­media thickness.

a. Differences in LVMI (ΔLVMI) and sBPV (ΔsBPV) as well as dBPV

0.1 (0.08–­
2.5 (2–­3)
Baseline

0.12)

(ΔdBPV) (r = .25, p = .001 and r = .18, p = .02, respectively) in the

whole population and
TABLE 1 (Continued)

b. ΔLVMI and ΔsBPV as well as ΔdBPV (r = .29, p = .001 and r = .20,

p = .03, respectively) in controlled hypertensives (Table 2).
Characteristics

cIMT (cm)

Finally, we performed multiple regression analysis, using the

CFR

backward method, in order to investigate any associations between

ΔLVMI and ΔsBPV or ΔdBPV in the whole population and the
TRIANTAFYLLIDI et al. | 1155

controlled hypertensives. Age, BMI, cholesterol, BP, and smoking at primary endpoint of the study is that both systolic and diastolic BPV
baseline evaluation were inserted in the model as independent vari- decrease associate with LVMI regression only in well-­controlled hy-
ables. We examined three models (Models A, B, C) using in each one pertensive patients, the latter confirmed by 24-­h ABPM.
a different method of baseline BP evaluation; office SBP in Model BPV reflects a dynamic hemodynamic parameter, which depicts
A, office mean BP in Model B and 24-­h SBP in Model C. We found marked BP fluctuations across time. These variations can be mea-
that, ΔsBPV was independently related with ΔLVMI (in all models) in sured over a period of seconds or minutes (very short-­term BPV),
the whole population and the controlled hypertensives, (Figure 1). 24 h (short-­term BPV), between days (mid-­term BPV) and between
Additionally, ΔdBPV was associated with ΔLVMI (in models A and months or years (long-­term BPV).8 Under physiological conditions,
B) in the whole population as well as the controlled hypertensives. BPV largely represents a response to environmental stimulations
Initial 24-­h SBP was also associated with ΔLVMI in the whole popu- and challenges of daily life. It aims at maintaining the so-­called BP
lation and well-­controlled hypertensive patients. “homeostasis” which in turn is necessary to guarantee adequate
organ perfusion in response to changing metabolic and physio-
logic demands (ie during physical exercise) or to changing environ-
4 | DISCUSSION mental conditions (ie during exposure to high-­altitude hypobaric
hypoxia or weather-­related temperature changes). However, sus-
In the present prospective study, we investigated the role of the tained increases in BPV may also reflect alterations in the mech-
short-­term BPV reduction regarding HMOD regression in hyper- anisms responsible for cardiovascular homeostasis or underlying
tensive patients at 3 years after initiation of medical treatment. The pathological conditions and may represent a source of damage to

TA B L E 2 Multiple linear regression analysis regarding independent associations between differences in LVMI and BPV (systolic and
diastolic)

Left ventricular mass index differences (ΔLVMI)

Independent variables Total population (n = 180) Controlled hypertensives (n = 119)

Model A (Office SBP)

ΔsBPV β = 0.20, ΔdBPV β = 0.17, p = .04 ΔsBPV β = 0.26, p = .01 ΔdBPV β = 0.22, p = .01
p = .01
Age –­ Age –­ Age –­ Age –­
Smoking –­ Smoking –­ Smoking β = 0.19, p = .06 Smoking –­
BMI –­ BMI –­ BMI –­ BMI β = −0.20, p = .05
Cholesterol –­ Cholesterol –­ Cholesterol –­ Cholesterol –­
Office SBP –­ Office SBP –­ Office SBP –­ Office SBP –­
Model B (Office mean BP)
ΔsBPV β = 0.20, ΔdBPV β = 0.17, p = .04 ΔsBPV β = 0.26, p = .01 ΔdBPV β = 0.22, p = .02
p = .01
Age –­ Age –­ Age –­ Age –­
Smoking –­ Smoking –­ Smoking β = 0.19, p = .06 Smoking –­
BMI –­ BMI –­ BMI –­ BMI β = −0.20, p = .05
Cholesterol –­ Cholesterol –­ Cholesterol –­ Cholesterol –­
Office mean BP –­ Office –­ Office –­ Office –­
mean mean mean
BP BP BP
Model C (24-­h SBP)
ΔsBPV β = 0.20, ΔdBPV –­ ΔsBPV β = 0.19, p = .04 ΔdBPV –­
p = .01
Age –­ Age –­ Age –­ Age –­
Smoking –­ Smoking –­ Smoking Smoking –­
BMI –­ BMI –­ BMI β = −0.20, BMI β = −0.20, p = .03
p = .03
Cholesterol –­ Cholesterol –­ Cholesterol –­ Cholesterol –­
24-­h SBP β = 0.29, 24-­h SBP β = 0.30, 24-­h SBP β = 0.35, 24-­h SBP β = 0.39, p < .001
p < .001 p < .001 p < .001
1156 | TRIANTAFYLLIDI et al.

F I G U R E 1 Relationship between LVMI regression and sBPV decrease in the whole population and controlled hypertensives

the cardiovascular system.12 Each type of BPV shares a differ- Hypertension-­mediated organ damage shows increased preva-
12,13
ent underlying mechanism, although not fully revealed. Very lence among patients even in the early stages of hypertension dis-
short-­term and short-­term BPV are mainly determined by increased ease. 2,21-­23 HMOD is due to BP levels as well as variable concomitant
central sympathetic drive, reduced arterial reflexes and behavioral conditions, neurohormonal alterations and life style (ie increased
and emotional factors while long-­term variability should be shaped salt consumption24) involved in structural and functional alterations
mainly by reduced arterial compliance, seasonal changes as well as of arterial bed, heart, kidneys and central nervous system.5
14
improper dosing or poor adherence to antihypertensive treatment. The clinical significance and prognostic implications of BPV have
Despite the different substrate, both short-­ and long-­term BPV are been demonstrated by a series of recent studies in which increased
associated with the development, progression and severity of car- BPV has been associated with a higher risk of CV mortality in the
diovascular and renal complications independently of mean pressure general population, 25 future CV events26 or contributed modestly
14
elevation. However, clinical trials have shown that long-­term BPV to CV risk stratification. 27 However, 24 h ambulatory BP level re-
is associated with cardiovascular events to a greater degree com- mained the most valuable CV predictor for use in clinical practice. 28
15-­17
pared to short-­term BPV. Additionally, 24-­h BPV has been recognized as a useful index of
ABPM has been long recognized as the gold standard method for HMOD in hypertensive and general population, pointing to carotid
diagnosing AH compared to office BP measurement, providing data artery wall alterations and LVH, 29-­31 the latter representing, at car-
18,19
on BP during patient's activities and uniquely during sleep. Since diac level, the main factor associated with worse CV prognosis.32
it is the only method for nocturnal BP dipping measurement, it may Likewise, increased BPV has been associated with arterial stiffness
also calculate both day and night BP fluctuations and subsequently and LV mass and dysfunction in treated and untreated hypertensive
BPV (ABPV). 20,21 Increased ABPV is associated with AH, carotid population, suggesting that BPV may be an important determinant
artery disease, progression of small vessel disease, left ventricular of HMOD.33-­35 In a group of elderly hospitalized patients, 24-­h SBPV
20
hypertrophy (LVH) and CV events. Consequently, ABPV is consid- could reflect the degree of HMOD as it was associated with IMT,
ered as an independent CV risk factor compared to 24-­h average LVMI and MAU.36 In a 7-­year follow-­up study of a small hyperten-
20
BP levels derived by ABPM. Various methods have been used for sive group (n = 73), Frattola et al reported that the BP level achieved
BPV measurement (continuous beat-­to-­beat recordings, office BP, by treatment, the degree of HMOD at baseline evaluation and the
home BP measurement, 24-­h ABPM). Moreover, there are different long-­term BPV were the most important determinants of future
indices for BPV evaluation (ie standard deviation [SD], coefficient of end-­organ damage related to hypertension throughout the years
variation, weighted 24-­h SD, average real variability [ARV]). Since of follow-­up.37 Importantly, a recent meta-­analysis showed a weak
there is no clear indication as to which method or index should be positive correlation between several 24-­h ABPM-­derived BPV mea-
preferred, the choice should be supported by the strongest outcome surements (24-­h SD, diurnal SD, weighted SD and 24-­h ARV) and
9 8
evidence. A recent meta-­analysis pointed to the use of SD, derived LVMI.38 On the other hand, Veloudi et al concluded that BPV ap-
by 24-­h ABPM, as one of the preferred indices for 24-­h BPV evalua- peared with limited clinical utility over a 12-­month period in patients
tion, which was also investigated in our study. with uncomplicated hypertension since the changes in average 24-­h
TRIANTAFYLLIDI et al. | 1157

SBP, but not BPV, were most relevant to changes in HMOD (LVMI, In conclusion, our study provides substantial evidence that in
39
PWV). middle-­aged hypertensive patients, systolic and diastolic BPV im-
In our study, we examined a population with recently diagnosed, provements, associated with cardiovascular risk reduction (left ven-
never-­treated and uncomplicated hypertension using 24-­h ABPM at tricular mass regression), occur only in the setting of BP treatment
baseline and after 3 years of treatment initiation. We pointed out within normal limits as it is confirmed by ABPM.
associations between ΔsBPV (and ΔdBPV) and ΔLVMI (independent
from initial levels of office SBP, mean BP and 24-­h SBP) as well as AU T H O R C O N T R I B U T I O N S
between 24-­h SBP at baseline and ΔLVMI in the whole population HT, DB, AS, DB, PT, EV, DV and II substantially contributed to the
and well-­controlled hypertensive patients. Our results underscore conception or design of the work; or the acquisition, analysis, or in-
the prognostic significance of initial ABPM-­derived data (BP levels terpretation of data for the work. HT, DB, AS, DB, PT, EV, DV and
and fluctuations) regarding LVMI regression. However, no other cor- II drafted the work or revising it critically for important intellectual
relation was revealed between ΔsBPV (and ΔdBPV) and the other content. HT and DB involved in final approval of the version to be
HMOD indices studied (PWV, LVMI, E/Ea, IMT, CFR). published. HT and DB agreed to be accountable for all aspects of the
In non-­controlled hypertensive patients, no relationship was work in ensuring that questions related to the accuracy or integrity
found between ΔsBPV (or ΔdBPV) and ΔLVMI or changes of any of of any part of the work are appropriately investigated and resolved.
the other HMOD indices studied (PWV, LVMI, E/Ea, IMT, CFR). On
the contrary, we noticed that PWV was increased at 3-­year post-­ ORCID
treatment even though BP was reduced from baseline levels in that Helen Triantafyllidi https://orcid.org/0000-0001-6801-1214
group of hypertensives patients. Thus we re-­confirmed that PWV
increases over time in those hypertensive patients under treatment REFERENCES
who do not achieve the optimal BP levels since vascular aging and 1. Kallikazaros I. Arterial hypertension. Hellenic J Cardiol.
life style besides BP levels are also powerful variables over time re- 2013;54(5):413-­415.
2. Williams B, Mancia G, Spiering W, et al. Guidelines for the man-
garding arterial stiffness increase.10
agement of arterial hypertension of the European Society of
This is the first study which takes into account the outcome of a Hypertension and the European Society of Cardiology: ESH/
3-­year successful antihypertensive treatment, evaluated by ABPM ESC task force for the Management of Arterial Hypertension. J
results, in order to explore the significance of BPV (systolic and di- Hypertens. 2018;36(12):2284-­2309.
3. Mancia G, Facchetti R, Bombelli M, et al. Relationship of office, home,
astolic) as a predictor index of HMOD regression. The achievement
and ambulatory blood pressure to blood glucose and lipid variables
of BP control within normal limits should be the primary goal of our in the PAMELA population. Hypertension. 2005;45:1072-­1077.
antihypertensive treatment and if this happens, then short-­term 4. Berry JD, Dyer A, Cai X, et al. Lifetime risks of cardiovascular dis-
sBPV reduction over time is able to predict the subsequent LVMI ease. New Engl J Med. 2012;366:321-­329.
5. Piskorz D. Hypertensive mediated organ damage and hypertension
regression.
management. How to assess beneficial effects of antihypertensive
treatments? High Blood Press Cardiovasc Prev. 2020;27:9-­17.
6. Greenland P. Effective use of ambulatory blood pressure monitor-
4.1 | Study limitations ing. JAMA. 2019;322(5):420-­421.
7. Lovibond K, Jowett S, Barton P, et al. Cost-­effectiveness of options
for the diagnosis of high blood pressure in primary care: a modeling
Our clinical prospective study has several limitations. Arterial hy-
study. Lancet. 2011;378:1219-­1230.
pertension has a high prevalence in population worldwide and sub- 8. Stevens SL, Wood S, Koshiaris C, et al. Blood pressure variability
sequently the moderate number of our Caucasian patients, overall and cardiovascular disease: systematic review and meta-­analysis.
and in each study group as well as the absence of co-­morbidities like BMJ. 2016;354:i4098.
9. Parati G, Stergiou GS, Dolan E, Bilo G. Blood pressure vari-
diabetes mellitus or chronic kidney disease, does not support us to
ability: clinical relevance and application. J Clin Hypertens.
generalize our results in all treated hypertensive patients. A greater 2018;20(7):1133-­1137.
number of patients are needed in future studies in order to expand 10. Triantafyllidi H, Trivilou P, Ikonomidis I, et al. Is arterial hyper-
our results. However, our group of 180 untreated hypertensive pa- tension control enough to improve aortic stiffness in untreated
patients with hypertension? A 3-­year follow-­up study. Angiology.
tients was relatively homogenous and it was re-­evaluated after an
2015;66(8):759-­765.
adequate time period of 3-­year post-­treatment. Another limitation 11. Devereux R, Reichek N. Echocardiographic assessment of left ven-
might be that our results were based on single ABPM at baseline and tricular mass in man. Circulation. 1977;55:613-­618.
3-­year post-­treatment. However, recent ESH guidelines do not sup- 12. Parati G, Torlasco C, Pengo M, Bilo G, Ochoa JE. Blood pressure
variability: its relevance for cardiovascular homeostasis and cardio-
port the need of a second ABPM application. Finally, the absence of
vascular diseases. Hypertens Res. 2020;43(7):609-­620.
severe HOMD is probably explained by the status of our patients at 13. Mancia G. Short-­and long-­term blood pressure variability: present
baseline (newly diagnosed, never-­treated, and hypertension stage I-­ and future. Hypertension. 2012;60:512-­517.
II) since a recent initiation of hypertension disease was recorded and 14. Parati G, Ochoa JE, Lombardi C, Bilo G. Assessment and management
of blood-­pressure variability. Nat Rev Cardiol. 2013;10:143-­155.
the hypertension burden was not severe.
1158 | TRIANTAFYLLIDI et al.

15. Matsui Y, Ishikawa J, Eguchi K, Shibasaki S, Shimada K, Kario K. 30. Sega R, Corrao G, Bombelli M, et al. Blood pressure variability and
Maximum value of home blood pressure: a novel indicator of target organ damage in a general population: results from the PAMELA
organ damage in hypertension. Hypertension. 2011;57:1087-­1093. study (Pressioni Arteriose Monitorate E Loro Associazioni).
16. Kikuya M, Ohkubo T, Metoki H, et al. Day-­by-­day variability of Hypertension. 2002;39:710-­714.
blood pressure and heart rate at home as a novel predictor of prog- 31. Shintani Y, Kikuya M, Hara A, et al. Ambulatory blood pressure,
nosis: the Ohasama study. Hypertension. 2008;52:1045-­1050. blood pressure variability and the prevalence of carotid artery al-
17. Johansson JK, Niiranen TJ, Puukka PJ, Jula AM. Prognostic value of teration: the Ohasama study. J Hypertens. 2007;25:1704-­1710.
the variability in home-­measured blood pressure and heart rate: the 32. Gosse P, Cremer A, Vircoulon M, et al. Prognostic value of the ex-
Finn-­Home Study. Hypertension. 2012;59:212-­218. tent of left ventricular hypertrophy and its evolution in the hyper-
18. Muntner P, Shimbo D, Carey RM, et al. Measurement of blood pres- tensive patient. J Hypertens. 2012;30:2403-­2409.
sure in humans: a scientific statement from the American Heart 33. Shin SH, Jang JH, Baek YS, et al. Relation of blood pressure variabil-
Association. Hypertension. 2019;73(5):e35-­e66. ity to left ventricular function and arterial stiffness in hypertensive
19. O'Brien E, White WB, Parati G, Dolan E. Ambulatory blood patients. Singapore Med J. 2019;60(8):427-­431.
pressure monitoring in the 21st century. J Clin Hypertens. 34. Schillaci G, Bilo G, Pucci G, et al. Relationship between short-­term
2018;20(7):1108-­1111. blood pressure variability and large-­Artery stiffness in human
20. Zawadzki MJ, Small AK, Gerin W. Ambulatory blood pressure vari- hypertension: findings from 2 large databases. Hypertension.
ability: a conceptual review. Blood Press Monit. 2017;22(2):53-­58. 2012;60:369-­377.
21. Wachtell K, Olsen MH, Dahlöf B, et al. Microalbuminuria in hyper- 35. Omboni S, Posokhov IN, Rogoza AN. Relationships between 24-­h
tensive patients with electrocardiographic left ventricular hyper- blood pressure variability and 24-­h central arterial pressure, pulse
trophy: the life study. J Hypertens. 2002;20:405-­412. wave velocity and augmentation index in hypertensive patients.
22. Devereux RB, Bella J, Boman K, et al. Echocardiographic left ventric- Hypertens Res. 2017;40(4):385-­391.
ular geometry in hypertensive patients with electrocardiographic left 36. Li CL, Liu R, Wang JR, Yang J. Relationship between blood pressure
ventricular hypertrophy: the life study. Blood Press. 2001;10:74-­82. variability and target organ damage in elderly patients. Eur Rev Med
23. Hawkins NM, Wang D, McMurray JJ, et al. Prevalence and prognostic Pharmacol Sci. 2017;21(23):5451-­5455.
implications of electrocardiographic left ventricular hypertrophy in heart 37. Frattola A, Parati G, Cuspidi C, Albini F, Mancia G. Prognostic
failure: evidence from the CHARM program. Heart. 2007;93:59-­64. value of 24-­hour blood pressure variability. J Hypertens.
24. Marketou ME, Maragkoudakis S, Anastasiou I, et al. Salt-­induced 1993;11(10):1133-­1137.
effects on microvascular function: a critical factor in hypertension 38. Madden JM, O'Flynn AM, Fitzgerald AP, Kearney PM.
mediated organ damage. J Clin Hypertens. 2019;21:749-­757. Correlation between short-­term blood pressure variability and
25. Kikuya M, Hozawa A, Ohokubo T, et al. Prognostic significance left-­ventricular mass index: a meta-­a nalysis. Hypertens Res.
of blood pressure and heart rate variabilities: the Ohasama study. 2016;39:171-­177.
Hypertension. 2000;36:901-­906. 39. Veloudi P, Blizzard CL, Head GA, Abhayaratna WP, Stowasser M,
26. Mancia G, Bombelli M, Facchetti R, et al. Long-­term prognostic Sharman JE. Blood pressure variability and prediction of target
value of blood pressure variability in the general population: results organ damage in patients with uncomplicated hypertension. Am J
of the Pressioni Arteriose Monitorate e Loro Associazioni Study. Hypertens. 2016;29(9):1046-­1054.
Hypertension. 2007;49:1265-­1270.
27. Hansen TW, Thijs L, Li Y, et al. Prognostic value of reading-­to-­
reading blood pressure variability over 24 hours in 8938 subjects
How to cite this article: Triantafyllidi H, Benas D, Schoinas A,
from 11 populations. Hypertension. 2010;55:1049-­1057.
28. Stolarz-­Skrzypek K, Thijs L, Li Y, et al. Short-­term blood pressure
et al. Hypertension-­mediated organ damage regression
variability in relation to outcome in the International Database of associates with blood pressure variability improvement three
Ambulatory blood pressure in relation to Cardiovascular Outcome years after successful treatment initiation in essential
(IDACO). Acta Cardiol. 2011;66:701-­706. hypertension. J Clin Hypertens. 2021;23:1150–­1158. https://
29. Mancia G, Parati G, Hennig M, et al. Relation between blood pres-
doi.org/10.1111/jch.14209
sure variability and carotid artery damage in hypertension: base-
line data from the European Lacidipine Study on Atherosclerosis
(ELSA). J Hypertens. 2001;11:1981-­1989.