Key words aldosterone - pulse wave velocity - intima media thickness - hypertension - atherosclerosis - atherosclerotic plaque
Introduction
Primary aldosteronism (PA) is recognized as the most frequent cause of endocrine
hypertension and accounts for about 5–15% of hypertensives [1 ]
[2 ].
However, it is not so long since PA was considered as a rare cause of endocrine
hypertension, characterized by its classic triad of aldosterone excess, metabolic
alkalosis and hypokalemia. The higher prevalence in more recent studies is based on
the use of the aldosterone-to-renin-ratio (ARR). The ARR allows detecting the
normokalemic subtype of PA, which is much more frequent and also considered as a
milder form of PA [3 ]
[4 ].
Aldosterone itself physiologically impacts on water and sodium homeostasis to
maintain blood pressure stable. However, in PA long-term exposure to inadequate high
aldosterone levels increases cardiovascular risk beyond blood pressure effects [5 ]
[6 ]
[7 ]
[8 ]
[9 ].
Basic research studies suggest that aldosterone may directly impact on vascular
function, inflammation and fibrosis and therefore favoring arterial stiffening [10 ]. Data from clinical studies suggest that
arterial stiffness and intima media thickness are increased in patients with PA
[11 ]
[12 ]
[13 ]
[14 ]. However, the major drawbacks of these
studies often deal with difficulties concerning sample size as well as adjustment
for confounders (e. g., diabetes mellitus or smoking status) in addition to
an overrepresentation of patients harboring a severe phenotype of PA [10 ]
[12 ]
[13 ]
[14 ]
[15 ]
[16 ]
[17 ]. Furthermore, to the best of our knowledge,
data on atherosclerotic burden in patients with PA are still missing. Nowadays,
based on increased screening intensity for PA using the ARR, more and more milder
cases of PA are diagnosed [4 ]. In this context
it has been speculated that atherosclerosis and cardiovascular risk could depend on
PA phenotype and be consequently lower in patients with milder forms of PA [18 ].
The current study thus aimed on providing data on subclinical atherosclerotic burden
but also on arterial stiffness in a large collective of patients with milder
subtypes of both unilateral and bilateral PA in comparison to patients with
essential hypertension (EH).
Patients and Methods
Patients and controls
The study protocol was approved by the ethics committee of the University of
Munich. All patients were prospectively enrolled at the Munich center of the
German Conn's Registry and gave written informed consent. To select a
cohort with milder forms of PA, only PA patients with baseline aldosterone
levels of lower than 200 ng/l, corresponding to the 50th
percentile of aldosterone levels in our cohort of patients with PA published
before [19 ], were admitted to study
participation. Due to our eligibility criteria patients with a history of
carotid artery disease, or peripheral artery disease were also not included in
the current study. Based on the eligibility criteria 95 patients could be
initially included in the study. Following 1:1 matching for age, gender, BMI and
cardiovascular risk factors (markers of lipid and glucose metabolism,
body-mass-index (BMI), smoking status and arterial blood pressure) seven
participants were excluded from the analysis. The final study cohort consisted
of forty-four patients with PA and forty-four patients with EH.
All study examinations followed standard operating procedures. Diagnosis of PA
was made according to the Endocrine Society Practice Guidelines [20 ]. The screening test consisted of a
baseline plasma aldosterone-to-renin ratio (ARR; cut-off
12.0 ng/U, sitting position). If elevated, diagnosis of PA was
ruled in by an abnormal confirmatory test (e. g. salt loading test,
captopril challenge test or both). Antihypertensive medication was stopped
before testing, if possible. Otherwise it was replaced by the alpha 1-adrenergic
receptor blocker doxazosin or calcium-channel blocker verapamil. The subtype
diagnosis between unilateral and bilateral disease was based on adrenal vein
sampling in thirty of forty-four PA patients and revealed unilateral PA in seven
cases [21 ]. Baseline examination was
performed shortly after testing for PA and before initiation of specific
treatment for PA.
Ten patients could be appropriately re-evaluated after a median of two years
after initiation of treatment. Seven of these patients were treated with
mineralocorticoid receptor antagonists (MRA; spironolactone at a dose of
25–50 mg/d), and three patients with unilateral PA
underwent unilateral adrenalectomy with histologic confirmation of
diagnosis.
Non-invasive vascular measurements
Settings for non-invasive vascular measurement were standardized (examinations in
the morning, room temperature at 18–21°C, 10 min of resting). At
baseline and at follow-up visit, at least three consecutive blood pressure
readings were obtained in the sitting position after not less than 10 min of
rest using a validated automatic oscillometric device. The measurement was
attended by a study nurse or an investigator and the average of the measurements
recorded. To ensure proper cuff size for blood pressure measurement the upper
arm circumference was measured at each visit. Concerning the ankle-brachial
index (ABI) systolic blood pressure was measured bilaterally over the brachial
arteries, distal posterior tibial arteries and dorsalis pedis arteries with a
hand-held continuous wave-Doppler probe after at least 10 min of rest. The ABI
was calculated for each leg as the ratio of the maximum systolic blood pressure
of the ankle arteries and the maximum systolic arm blood pressure [22 ]. For the patient-based analysis, the
lower ABI of both legs was taken into account. Peripheral artery disease was
defined by an ABI<0.9 in at least one leg, whereas an ABI>1.3
was considered indicative of poorly compressible arteries [23 ].
The measurement of the central pulse wave velocity (cPWV, in m/s) was
performed using an oscillometric device and applying the Gesenius–Keller
method (AngE Pro 8, software version 1.18.33; SOT Medical Systems, Maria Rain,
Austria). Cuffs were placed on both forearms and ankles and ECG-triggered pulse
waves were recorded at different cuff pressures [24 ]
[25 ].
Ultrasound examinations
All patients underwent a standardized vascular ultrasound examination. Ultrasound
examinations were conducted using a GE Logic E9 ultrasound machine (General
Electric, Munich, Germany) by experienced vascular sonographers with more than
ten years in professional experience (AK, CL), blinded to clinical and
biochemical information. Except the abdominal aorta which was evaluated with a
2–6 MHz convex probe, all arterial segments were examined with
an 8 MHz linear probe. The following arterial segments were examined in
the longitudinal and the cross-sectional view: common carotid arteries (CCA),
carotid bulbs (CB), external carotid arteries (ECA), internal carotid arteries
(ICA), vertebral arteries (VA), subclavian arteries (SA), abdominal aorta (AA),
common femoral arteries (CFA), profunda femoris arteries (PFA), superficial
femoral arteries (SFA) and popliteal arteries (PoA). Each arterial segment was
assessed for the presence of atherosclerotic plaques, stenosis
(<50%,≥50%) and occlusion. Atherosclerotic
plaque was defined according to the Mannheim Intima-Media Thickness Consensus as
focal thickening of the intima-media complex of≥50% compared to
the surrounding IMT and/or focal intimal thickening protruding into the
arterial lumen≥0.5 mm [26 ]. IMT measurements of the CCA (CCA-IMT) were conducted
1–2 cm proximal to the bifurcation within a region free of
plaque. Three CCA-IMT measurements were conducted on each side. Mean CCA-IMT of
each patient was calculated as the mean value of all six measurements [27 ]. The diameter of the AA was measured by
using the leading edge method [28 ].
Absence or presence of aortic ectasia (≥2.0 cm) or aortic
aneurysm (≥3.0 cm) were documented. A modified plaque score was
calculated by adding up the number of arterial segments without atherosclerotic
changes (0 points), with the presence of non-stenotic atherosclerotic plaques (1
point), arterial stenosis<50% (2 points), arterial
stenosis≥50% (3 points) and arterial occlusion (4 points) [29 ]. Given a total number of 21 segments
analyzed, the plaque score could range from 0–84 points.
Statistical analysis
This study aimed to compare (markers of) subclinical atherosclerotic plaque
burden between patients with PA and EH. We estimated the sample size on the
basis of current literature on CCA-IMT, which also serves as a surrogate
parameter for atherosclerotic plaque burden [11 ]. Assuming a power of 0.8 and a two-sided type I error of
α=0.05 we calculated that a total sample size of 71 subjects
(sum of patients with PA and EH) would be sufficient to detect a difference of
0.1 mm±0.15 mm in CCA-IMT between both groups. To avoid
confounding both cohorts were matched for age, gender, BMI, and cardiovascular
risk factors (markers of lipid and glucose metabolism, body-mass-index (BMI),
smoking status, and arterial blood pressure) as mentioned above. For the
follow-up evaluation we calculated based on data of Matsuda and colleagues that
a sample size of 8 patients with PA would be sufficient to detect a difference
of 0.1 mm±0.1 mm in CCA-IMT between baseline examination
and follow-up, assuming a power of 0.8 and a two-sided type I error of
α=0.05 [30 ].
All numerical values are expressed as median, 25th and 75th percentile if not
mentioned otherwise. Body mass index (BMI) was calculated as weight in kilograms
divided by the square of the height in meters. Data between groups were compared
using Mann–Whitney U test or chi-square test for numerical or
categorical variable, respectively. Within-group changes from baseline to
follow-up were calculated by Wilcoxon signed-rank test. Spearman's Rank
correlation coefficient was used to perform bivariate correlation analysis.
Two-tailed probability values of<5% were considered to be
statistically significant. Statistical analysis was performed using standard
statistical software (IBM SPSS Statistics for Windows, Version 26. Armonk, NY:
IBM Corp.).
Results
Clinical and biochemical baseline characteristics of the cohorts
Clinical and biochemical characteristics of the study cohort are shown in [Table 1 ]. As expected, according to the
study design, patients with PA had higher aldosterone and lower potassium and
renin levels compared with EH ([Table
1 ]). Apart from that patients with PA and EH were well matched for age,
gender, BMI, and cardiovascular risk factors such as diabetes mellitus and
smoking status. Blood pressure and the intensity of antihypertensive treatment,
as measured by defined daily doses of antihypertensives (DDD; both: 1.0;
p=0.688), were also comparable between both subgroups.
Table 1 Baseline characteristics of the total cohort.
Patient characteristics
Patients with PA (n=44)
Patients with EH (n=44)
p
Age [years]
61 [51; 67]
61 [53; 68]
0.622
Sex [f/m]
24/20
23/21
0.831
Current cigarette smoking [n]
3
4
0.717
Diabetes mellitus [n]
6
4
0.502
Duration of hypertension [months]
124 [53; 205]
62 [12; 144]
0.044
Statin therapy [n]
8
7
0.777
Cardiovascular events [n]
6
7
0.764
BMI [kg/m²]
27.1 [24.6; 29.7]
26.7 [22.9; 30.1]
0.460
Plasma aldosterone [ng/l]
128 [97; 154]
91 [66; 110]
<0.001
Plasma renin [mU/l]
2.1 [2.0; 3.8]
7.3 [2.0; 21.2]
<0.001
SBP [mmHg]
151 [140; 169]
150 [140; 162]
0.637
DBP [mmHg]
92 [84; 100]
90 [80; 97]
0.169
DDD [n]
1.0 [0.5; 2.4]
1.0 [0.0; 2.6]
0.688
Number of antihypertensive drugs [n]
1.0 [1.0; 2.0]
1.0 [0.0; 2.0]
0.535
Serum potassium [mmol/l]
4.0 [3.7; 4.2]
4.2 [3.9; 4.3]
0.028
GFR [ml/min/1.73 m²]
90 [78; 100]
83 [71; 100]
0.246
HDL-C [mg/dl]
58 [48; 66]
56 [47; 69]
0.894
LDL-C [mg/dl]
125 [99; 150]
134 [98; 157]
0.649
Triglycerides [mg/dl]
109 [68; 163]
98 [70; 154]
0.848
Total cholesterol [mg/dl]
199 [176; 230]
206 [166; 238]
0.793
HbA1c [mmol/mol]
36 [33; 39]
38 [36; 38]
0.165
Albuminuria [mg/d]
11.6 [9.3; 15.4]+
9.6 [6.5; 14.6]+
0.088
Pulse wave velocity [m/s]
7.2 [6.0; 8.7]
7.1 [5.8; 7.8]
0.372
Intima media thickness [mm]
0.77 [0.65; 0.90]
0.75 [0.67; 0.85]
0.997
ABI
1.1 [1.1; 1.2]
1.2 [1.1; 1.2]
0.148
Plaque Score
4 [0; 11]
7 [2; 13]
0.159
Supra-aortic plaques [n]
64
84
0.135
Aortic plaques [n]
21
27
0.158
Infra-aortic plaques [n]
75
88
0.399
Aorta [cm]
1.7 [1.5; 1.8]
1.7 [1.5; 1.9]
0.667
Data are given as median, 25th and 75th percentile in square brackets.
Significance is marked in bold. Comparisons between baseline values were
performed by Mann-Whitney U test and chi-square test. DBP: Diastolic
blood pressure; DDD: Defined daily dose; GFR: Glomerular filtration
rate; HbA1c: Glycated hemoglobin; HDL-C: High-density lipoprotein
cholesterol; LDL-C: Low-density lipoprotein cholesterol; SBP: Systolic
blood pressure.+Owing to incomplete data, the calculations were
performed with a reduced number of patients (PA, n=42; EH,
n=36).
Atherosclerotic burden, ABI, CCA-IMT, and cPWV
The amount of vascular plaques was comparable in all 21 arterial segments
including the supraaortic arteries, abdominal aorta, and arteries of the lower
limbs ([Fig. 1 ]). Hence atherosclerotic
plaque burden, as expressed by the plaque score, was also similar in patients
with PA and EH (4 vs. 7; p=0.159). CCA-IMT was also comparable between
both groups with a median of 0.77 mm and 0.75 mm respectively
(p=0.997). While we could not detect any plaques in 16 patients (10
patients with PA, 6 patients with EH; p=0.269), 12 patients had relevant
arterial stenosis (5 patients with PA and 7 patients with EH; p=0.534).
Only 1 patient with PA exhibited segmental stenosis of more than 50%
showing unilateral high-grade stenosis of the SFA (middle segment), also
indicated by an ABI of 0.76. Apart from that ABI did not differ between both
groups (PA: 1.1 vs. EH: 1.2; p=0.148). In total 6 patients, 3 patients
with PA and 3 patients with EH, had an ABI higher than 1.3, thereof 2 patients
with manifest diabetes mellitus type 2. In total seven patients had aortic
ectasia, including 2 patients with PA and 5 patients with EH (p=0.237).
Otherwise the measures of the abdominal aorta were in normal range and similar
in both subgroups (both: 1.7 cm; p=0.667). cPWV did not differ
significantly in patients with PA and EH (7.2 m/s vs.
7.1 m/s; p=0.372). In the total cohort, CCA-IMT, cPWV
and plaque score (all: p<0.005) were strongly correlated with each
other. Although we could not detect univariate correlation between CCA-IMT, cPWV
and aldosterone or renin levels, we observed a significant correlation of the
plaque score with higher renin (r=0.302; p=0.005) but not with
aldosterone levels at baseline (r=–0.013; p=0.908).
Those results were in most part replicable when analyzing PA and EH patients
separately (data not shown). However, the latter group showed a strong
correlation of the plaque score with renin levels at baseline (r=0.413;
p=0.006), which was not the case in the PA subgroup (r=0.137;
p=0.376).
Fig. 1 Distribution of atherosclerotic plaques in patients with
essential hypertension and primary aldosteronism. CCA: Common carotid
artery; CB: Carotid bifurcation; ECA: External carotid artery; ICA:
Internal carotid artery; VA: Vertebral artery; SA: Subclavian artery;
AA: Abdominal aorta; CFA: Common femoral artery; PFA: Profunda femoris
artery; SFA: Superficial femoral artery; PoA: Popliteal artery.
Follow-up
After a median of 2 years following initiation of treatment 10 patients with PA
(7 with bilateral and 3 with unilateral disease) underwent reassessment of cPWV
and CCA-IMT. At that time systolic blood pressure levels (135 mmHg vs.
167 mmHg at baseline; p=0.012) and albuminuria had significantly
decreased (7.3 mg/d vs. 18.4 mg/d at baseline; p=0.008),
whereas potassium levels had increased (4.4 mmol/l vs.
3.8 mmol/l at baseline; p=0.005). CCA-IMT was unaltered
(0.66 mm vs. 0.70 mm; p=0.150), whereas cPWV was
significantly reduced from 8.2 m/s to 6.5 m/s (p=0.017;
[Fig. 2 ]).
Fig. 2 Changes of central pulse wave velocity (cPWV) from baseline
to follow-up in individual patients with primary aldosteronism.
Abbreviations: ADX: adrenalectomy; MRA: mineralocorticoid receptor
antagonist treatment
Discussion
It is well known that chronically high aldosterone levels in combination with
inadequate high dietary salt intake cause cardiovascular damage [5 ]. In this context there is accumulating
evidence that aldosterone impacts on endothelial dysfunction, namely via vascular
remodeling, inflammation, alterations in vascular tone and induction of early
atherosclerosis at least in patients suffering from a severe phenotype of PA [10 ]
[11 ]
[12 ]
[13 ]
[14 ]
[15 ]
[30 ]. Exemplary, already in 2007 Holaj et al.
reported lower CCA-IMT in patients with EH compared to patients with PA. In this
study almost 50% of PA patients had unilateral disease, with another
39% of patients being unclassified by refusing or unsuccessful adrenal
venous sampling (AVS). Mean aldosterone and potassium levels were
389 ng/l and 3.6 mmol/l respectively. Furthermore,
those patients still had elevated systolic blood pressure of 160 mmHg,
despite receiving a mean of four antihypertensive drugs, altogether illustrating
that these patients harbored a severe phenotype of PA.
Over the years we observed a shift to milder forms of PA not only in our center of
the German Conn’s Registry, presumably due to increased screening intensity
using ARR [4 ]. In the context with Murata and
colleagues suggesting lowered cardiovascular risk in patients with mild PA,
indistinguishable to patients with EH, our study aimed to compare (subclinical)
atherosclerotic burden and arterial stiffness between patients with EH and patients
with PA manifesting milder phenotype of PA [4 ]
[18 ].
The patients included in the present study showed mild phenotype of PA, illustrated
by a low percentage of hypokalemia (14%; n=6) in the cohort. Further
aldosterone levels as well as DDD were lower than reported before (128 vs.
219 ng/l; 1.0 vs. 3.0) in the presence of unchanged SBP (151 vs.
151 mmHg) [19 ]. As patients with
milder forms of PA more frequently show bilateral PA, AVS was only performed in 30
of 44 PA patients, a strategy, which is in agreement with current guidelines [4 ]
[20 ].
Consequently, we could identify only seven patients (16%) with unilateral
disease. Altogether the mild biochemical and clinical manifestation could be the
reason that the time to diagnosis in this cohort does not differ from previous
studies [31 ].
Our study results indicate that arterial stiffness as well as atherosclerotic burden
in patients with mild PA do not differ from patients with EH, neither using
non-invasive vascular measurements (CCA-IMT, cPWV, ABI) nor using plaque score
(determined by ultrasound examination of 21 different arterial regions), which in
part contrasts findings from patients with PA and a more severe phenotype [11 ]. However, in the current study markers of
atherosclerosis and arterial stiffness strongly correlated with each other but also
with cardiovascular risk factors such as presence of diabetes mellitus, high blood
pressure levels and male sex (data not shown). When comparing absolute values of
CCA-IMT and cPWV with results of other studies, in this study patients with PA are
at the lower end of the range, which would be in accordance with a limited impact of
mild PA on vascular remodeling [11 ]
[12 ]
[14 ]
[15 ]. This is further underlined
by the fact that CCA-IMT in this study was also lower compared to unpublished data
from our Munich center of the German Conn’s Registry, collected between 2008
and 2013 and showing CCA-IMT of 1.0 mm in patients with more severe
phenotype of PA. In contrast, CCA-IMT and cPWV in patients with EH were midrange
compared to recent studies [11 ].
Moreover, we found a correlation between atherosclerotic burden, measured by plaque
score and statin use (r=0.315; p=0.003) and also for CCA-IMT and
statin use (r=0.235; p=0.028). Altogether these could be the reasons
that in our study CCA-IMT and cPWV in patients with PA were indistinguishable from
patients suffering from EH, supposing a less pronounced aldosterone effect in mild
cases of PA, which would be in accordance with recent findings [18 ]. Nonetheless initiation of specific
treatment for PA resulted in a statistically significant reduction of albuminuria
and cPWV and should be therefore expected to translate into favorable for
cardiovascular outcomes ([Fig. 2 ]). Arterial
stiffening itself negatively impacts microcirculation and may be an indicator of
premature vascular aging in hypertensive subjects. Improvement of cPWV under
treatment, as documented at follow-up in a small subset of PA patients in our study,
may be a hint on the effect of aldosterone excess on vascular aging via arterial
stiffening. Although it is of note that the reduction of cPWV and albuminuria have
also been reported in patients with (essential) hypertension on MRA treatment, which
should promote a more liberal use of MRA in patients with arterial hypertension
[32 ]
[33 ].
Interestingly renin levels correlated positively with plaque score in the total
cohort. After splitting into patients with PA and EH we could observe a strong
correlation for renin and plaque score in patients with EH (r=0.413;
p=0.006), but not in patients with PA (r=0.137; p=0.376).
These findings fit well with recent data from us and others, suggesting an
association between renin levels and atherosclerotic burden in patients with EH but
not in patients with PA [34 ]
[35 ]
[36 ]
[37 ].
Our study has several strengths. This prospective study investigates, for the first
time, atherosclerotic burden in a substantial number of patients with PA and
controls particularly when compared to earlier studies dealing with CCA-IMT [11 ]. Patients with mild PA and controls were
very well matched with regard to demographics and cardiovascular risk factors. All
clinical and sonographic data as well as biomaterial were collected in a
standardized manner within the structured diagnostic workup of the German
Conn’s Registry. In a holistic approach, we studied 21 different arterial
regions by high resolution sonography to provide a phenotype of atherosclerotic
plaque distribution in patients with PA and EH. It has been shown that plaque
assessment of different arterial regions shows stronger association with coronary
artery disease than CCA-IMT alone [27 ]
[38 ]
[39 ].
We acknowledge some limitations of our study. First, mild PA is not yet well-defined.
The upper cut-off (200 ng/l) used in our study to define patients
with mild PA was chosen deliberately: it defines a collective of patients with
aldosterone levels below the 50th percentile in our center and is based on a recent
publication [19 ]. Furthermore, while CCA-IMT
is a known surrogate parameter of subclinical atherosclerosis, still no consensus
exists on the determination of plaque burden on ultrasound, and most studies are
limited to assessment of the carotid arteries [39 ]
[40 ]. Ultrasound is an
investigator-dependent diagnostic method but offers the highest spatial resolution
when it comes to detect initial changes of the arterial vessel wall. In our study,
sonographers were highly experienced and blinded with regard to the underlying cause
of arterial hypertension. Another limitation of the study is the small group of
patients in the follow-up. Here, initially 12 patients underwent follow-up. However
due to technical problems data of two patients could not be analyzed and we
therefore ended up with 10 patients.
In summary, our data show that patient with mild PA do not differ from patients with
EH with respect to subclinical atherosclerotic burden as well as arterial stiffness.
This does not contradict the previous observation of impaired vascular function in
patients with PA but rather sheds light on a second momentum – that of a
less pronounced impact of milder PA forms on vascular function and structure.
Nevertheless, treatment for PA was followed by an improvement of markers of
cardiovascular risk and therefore confirms daily practice to intensify screening for
PA and argues for a more liberal use of MRA in patients with arterial hypertension.
Further longitudinal studies are necessary to determine whether the long-term risk
of mild PA is indeed comparable with patients suffering from EH and to evaluate a
specific effect of aldosterone excess on vascular aging.
