Influence of thoracic aortic inflammation and calcifications on arterial stiffness and cardiac function in older subjects

Abstract

Background

Vascular aging is accompanied by gradual remodeling affecting both arterial and cardiac structure and mechanical properties. Hypertension is suggested to exert pro-inflammatory actions enhancing arterial stiffness.

Objectives

To determine the influence of thoracic aortic inflammation and calcifications on arterial stiffness and cardiac function in hypertensive and normotensive older subjects.

Design

A prospective study.

Setting

An acute geriatrics ward of the University Hospital of Nancy in France.

Subjects

Thirty individuals ≥ 65 years were examined, including 15 hypertensive subjects and 15 controls well-matched for age and sex.

Measurements

Applanation tonometry was used to measure aortic pulse wave velocity (AoPWV) and carotid/brachial pulse pressure amplification (PPA). Left ventricular parameters were measured with magnetic resonance imaging. Local thoracic aortic inflammation and calcification were measured by 18 F-fluorodeoxyglucose positron emission tomography/computed tomography imaging. Biomarkers of low-grade inflammation were also quantified.

Results

AoPWV was higher in elderly hypertensive subjects comparatively to normotensive controls (15.5±5.3 vs. 11.9±2.5, p=0.046), and hypertensives had a higher calcification volume. In the overall population, calcifications of the thoracic descending aorta and inflammation of the ascending aorta accounted for respectively 18.1% (p=0.01) and 9.6% (p=0.07) of AoPWV variation. Individuals with high levels of calcifications and/or inflammation had higher AoPWV (p=0.003). Inflammation had a negative effect on PPA explaining 13.8% of its variation (p<0.05).

Conclusions

This study highlights the importance of local ascending aortic inflammation as a potential major actor in the determination of PPA while calcifications and hypertension are more linked to AoPWV. Assessment of PPA in the very elderly could provide complementary information to improve diagnostic and therapeutic strategies targeting ascending aortic inflammation.

Introduction

Vascular wall remodeling and arterial stiffness are major determinants of morbidity, mortality and loss of autonomy in older subjects. Both aging and hypertension lead to structural modifications of the aortic wall such as thickening, development of arteriosclerosis and dilatation. These changes in turn lead to an increase in pulse pressure and arterial stiffness (1, 2). However, there are only very few available data on the impact of inflammation and calcification on cardiac and arterial function. In the present study, arterial function was assessed with two approaches: carotid-femoral pulse wave velocity (AoPWV) which is the standard method for measuring aortic stiffness (3) and pulse pressure amplification (PPA) between central (carotid) and peripheral (brachial) arteries. Low PPA reflects large and small artery stiffness and remodeling, as well as an increase in wave reflections, AoPWV and peripheral arterial resistance (4, 5).

Recent studies have advocated the use of arterial positron emission tomography (PET) imaging with 18 F fluorodeoxyglucose (FDG) as a means of measuring arterial wall inflammation in various populations. FDG uptake has been shown to be correlated with the number of cardiovascular risk factors and even the risk of future cardiovascular events (6, 7, 8, 9, 10).

Moreover, in a previous study conducted in thirty middle-aged patients, we have shown that aortic calcifications quantified by computed tomography (CT) and local signs of inflammation detected by FDG uptake in the thoracic aorta contribute to arterial stiffness and in particular to an increase in carotid-femoral pulse wave velocity (11).

In addition, prior studies using magnetic resonance imaging (MRI) have yielded evidence for a link between aortic geometry, increased stiffness indices and ventricular concentric remodeling (12, 13, 14, 15).

The main objective of the current study was to analyze the potential impact of calcifications and inflammation of the thoracic aorta on AoPWV, PPA and left ventricular remodeling in a population of hypertensive and normotensive older individuals.

Methods

Subject selection criteria

Thirty European individuals aged over 65 years (16 men and 14 women) were prospectively recruited by the Geriatric Medicine Unit at the University Hospital of Nancy. The Geriatric Medicine Unit is a 47-bed acute geriatric medicine ward in a university hospital admitting patients mostly from the Emergency Department. The study was approved by our institutional review committee (CPP agreement n° 2009-A00907-50) and released on the ClinicalTrials.gov site under the identifier: NCT 01963221. All exams were performed after the individuals gave informed written consent and were conducted for the exclusive purpose of the study.

Individuals were considered hypertensive if they had systolic blood pressure (SBP) values > 140 mmHg and/or diastolic blood pressure (DBP) values > 90mmHg or if they were previously diagnosed as hypertensive and treated with one or more antihypertensive drug (16). In this population, normotensive and hypertensive individuals were well-matched for age and sex.

Exclusion criteria were blood glucose> 200 mg/dl or known diabetes mellitus, inflammatory disease or cancer, secondary hypertension, renal, hepatic or pulmonary insufficiency, absence of cardiac sinus rhythm, and contraindication to MRI. All subjects arrived in a fasting state at the “PET/CT Centre for Cardiovascular Imaging Studies”. Blood sampling, hemodynamic measurements and arterial tonometry were performed at rest. An ensuing MRI allowed measurement of cardiac parameters. Finally, subjects were referred for combined FDG PET/CT imaging.

Aortic pulse wave velocity (AoPWV) and arterial pressure measurements

All measurements were performed in the morning at stable room temperature (20° to 22°C) after an overnight fast. After a 15-minute rest period in supine position, blood pressure was measured at the brachial artery with a semiautomatic device (Omron 705IT, Kyoto, Japan).

Central blood pressure and AoPWV measurements were performed with a validated device (PulsePen®, Dia Tecne srl, Milan, Italy) as described previously (17, 18). AoPWV was measured at a central level, between carotid and femoral sites. PPA was expressed as the percentage of increase in PP in the brachial artery relative to central pulse pressure (PPc) according to the formula: PPA=100 × (PP-PPc)/PPc (19, 20).

Inflammatory markers measurement

Immediately after acquisition of venous blood, plasma or serum was separated by centrifugation (3000g at 4°C for 15 minutes), placed in aliquots, and stored at -70°C for the measurement of inflammatory markers.

High-sensitivity C-reactive protein (hsCRP) was measured a single time by immunonephelometry (Dade Behring BNII) (assay range 0.175 to 10 mg/L) and high sensitivity interleukin-6 (hsIL-6) by specific ELISAs (R&D systems) (assay range 0.156 to 10 pg/mL). Fibrinogen was measured using the BN II nephelometer system (SIEMENS, Marburg, Germany).

Positron Emission Tomography and Computed Tomography

Combined FDG PET/CT imaging was performed using a Biograph hybrid scanner system.

For analysis purposes, the thoracic aorta was divided into three segments: the ascending aorta, the aortic arch and the descending aorta, and standard uptake value (SUV) was calculated as described in previous publications (10, 21, 22). Regions of interest were drawn around the aorta on each trans-axial slice, allowing maximal aortic SUV (SUVmax) to be determined for each slice. These values were averaged to determine the SUVmax for the ascending aorta, the aortic arch and the descending aorta (11).

Aortic calcifications volumes (VCa) were also determined for the three predetermined segments of the thoracic aorta using the dedicated Smartscore v3.5 (General Electric Medical Systems, Milwaukee, WI) software and a threshold of 130 HU (23).

Assessment of cardiac parameters by magnetic resonance imaging

Immediately after applanation tonometry, all images were acquired with subjects in the supine position in a 3-T magnet (Sigma Excite; GE Medical Systems, Milwaukee, Wisconsin, USA) equipped with an eight-element phased-array surface coil. A steady-state free precession (SSFP) cine and through plane phase contrast images of proximal aorta were acquired at the level of the ascending thoracic aorta, near the root of the aorta, 1cm distal to the sinotubular junction for each subject in order to obtain aortic diameter.

In addition to aortic data, a cine SSFP was used to assess left-ventricular function in contiguous short axis planes, each slice being recorded during a 15-s or less breath-hold period (24). Main acquisition parameters were as follows: 8mm slice thickness, 3.8–4.3 ms repetition time, 14–16 K-space lines per segment, 30 phases per cardiac cycle with view sharing, field-of-view ranging from 32 to 38 cm and a 224x224 matrix.

Left-ventricular end-diastolic volume (LV EDV), ejection fraction (EF) and end-diastolic left-ventricular mass (LVM) were determined on the contiguous short-axis slices, using dedicated software (MASS; Medis, The Netherlands) (15). LV mass/end-diastolic volume ratio was calculated as LVM divided by left ventricular end diastolic volume determining the degree of concentric remodeling in elderly individuals (12, 15).

Statistical analysis

NCSS 2000 statistical software package (Kaysville, Utah, USA) was used for statistical analysis. Continuous variables are expressed as mean ± standard deviation (SD) and compared using a paired-sample T-test or trend analysis of variance where appropriate. The categorical variables were compared with the Chi-square test. hsCRP and hsIL-6 are presented as the median ± interquartile range and compared using the Wilcoxon Rank-Sum test. Correlations between normally distributed data were performed using Pearson’s correlation, and presented as r2 values. Spearman’s correlation was also used for non-parametric data. For the purpose of the analyses, subjects were stratified according to the median values (Low vs. High) of descending aortic calcification volume score and ascending aortic inflammation (SUVmax). Accordingly, 4 groups were identified:

• –Low/Low group (n=8) included patients with both a low grade of inflammation and a low level of VCa;
• –High/High group (n=8) included patients with both a high grade of inflammation and a high level of VCa;
• –High SUV/Low VCa group (n=8) included patients with values over the median for inflammation.
• –High VCa/Low SUV group (n=6) included patients with values over the median for calcifications.

A p<0.05 was considered to be indicative of a significant difference. When multivariate analysis was performed, parameters were included in the model if their correlation had a p value under the threshold of 0.10.

Results

Comparison between normotensive and hypertensive elderly participants

A total of 30 patients were studied (mean age 75.3±5.8 years, 47% female).

Demographic, clinical and main hemodynamic characteristics, AoPWV, systemic arterial stiffness, MRI cardiac parameters, PET/CT results as well as systemic inflammation biomarkers in age- and gender-matched normotensive and hypertensive individuals are presented in Table 1.

Table 1.

General characteristics of the elderly study population and comparison between normotensive (NT) (n=15) and hypertensive (HT) (n=15) subjects matched for age and sex

	All (n=30)	NT (n=15)	HT (n=15)	P
Demographics
Age (years)	75.3 ± 5.8	74.3 ± 6.5	76.3 ± 5.1	0.07
Female gender (%)	47	47	47
Height (m)	1.63 ± 0.10	1.66 ± 0.09	1.60 ± 0.10	0.02
BMI (kg.m-2)	26.3 ± 2.8	25.3 ± 2.8	27.3 ± 2.6	0.08
IHD (%)	10	0	20	0.08
Smoking (%)	50	53	47	0.77
Dyslipidemia (%)	53	47	60	0.49
Stroke (%)	3	0	7	0.33
Heart rate (min-1)	60.4 ± 7.3	59.8 ± 7.0	60.9 ± 7.9	0.68
Beta blockers (n)	1	0	1	0.33
ARBs (n)	7	0	7	0.004
ACE inhibitors (n)	4	0	4	0.04
Thiazides (n)	8	0	8	0.002
Anti HT tt (n)	11(36.5%)	0	11 (73%)	0.0002
Statins (%)	9 (30%)	2 (13%)	7 (47%)	0.10
Vascular parameters
SBP (mmHg)	144.2±22.5	133.8 ± 15.7	154.7 ± 23.8	0.02
SBPc (mmHg)	131.1±20.3	121.4 ± 14.4	140.8 ± 21.2	0.007
DBP (mmHg)	81.9 ± 11.7	77.7 ± 8.2	86.1 ± 13.2	0.04
PP (mmHg)	62.3 ± 17.7	56.1 ± 11.0	68.5 ± 21.1	0.07
PPc (mmHg)	49.2±14.8	43.8 ± 8.9	54.7 ± 17.7	0.042
AoPWV (m.s-1)	13.7 ± 4.4	11.9 ± 2.5	15.5 ± 5.3	0.046
PPA (%)	27.7 ± 9.1	28.7 ± 8.2	26.7 ± 10.1	0.56
MRI parameters
LVEF (%)	56.5 ± 6.0	57.1 ± 5.9	55.9 ± 6.2	0.54
Indexed LVM (g.m-2)	46.0 ± 10.0	43.1 ± 10.0	48.8 ± 9.4	0.06
LVM/EDV ratio	0.72 ± 0.18	0.67 ± 0.18	0.76 ± 0.17	0.18
Aortic diameter (cm)	3.53 ± 0.34	3.51 ± 0.33	3.56 ± 0.35	0.63
TEP parameters
SUVmax Asc Ao	2.24 ± 0.27	2.13 ± 0.26	2.34 ± 0.26	0.13
SUVmax Ao Arch	2.34 ± 0.35	2.26 ± 0.42	2.42 ± 0.26	0.28
SUVmax Desc Ao	2.27 ± 0.40	2.20 ± 0.48	2.34 ± 0.31	0.29
VCa Asc Ao (mm3)	5.60 ± 19.84	3.20 ±12.39	7.93 ± 25.50	0.50
VCa Ao Arch (cm3)	1.40 ± 1.62	0.83 ± 0.86	1.96 ± 2.01	0.047
VCa Desc Ao (cm3)	0.55 ± 0.79	0.11 ± 0.19	1.00 ± 0.91	0.003
Systemic Inflammation Biomarkers
hs CRP (mg/L)	1.50 (1.10-3.15)*	3.2 ± 3.5	2.1 ± 1.2	0.19
hsIL-6 (pg/mL)	1.90 (1.90-1.90)*	3.6 ± 3.7	4.0 ± 8.0	0.86
Fibrinogen (g/L)	4.23 ± 0.72	412 ± 0.75	4.35 ± 0.71	0.24

BMI, body mass index; IHD, ischemic heart disease; ARBs, Angiotensin II receptor antagonists; ACE, Angiotensin converting enzyme; HT tt, Anti-hypertensive treatment; SBP/DBP, systolic/diastolic blood pressure, PP, pulse pressure; AoPWV, carotid-femoral pulse wave velocity; PPA, pulse pressure amplification; MRI, magnetic resonance imaging; LV, left ventricular; LVEF, LV ejection fraction; LVM/EDV, LV mass/ end diastolic volume; Asc, ascending; Desc, descending; Ao, aorta; SUV, standard uptake value; VCa, calcification volume; All values are mean ± SD or %; *median

As shown in Table 1, aside from differences related to the selection criteria (blood pressure levels, use of antihypertensive drugs, etc.), hypertensive and normotensive individuals did not show any major differences with the exception of AoPWV which was significantly higher in hypertensive subjects: 15.5±5.3 m.s-1 versus 11.9±2.5 m.s-1 (p=0.046). In addition, hypertensive participants showed higher VCa values at the level of the aortic arch (1.96±2.01cm3 versus 0.83±0.86cm3; p=0.047) and descending aorta (1.00±0.91cm3 versus 0.11±0.19 cm3; p=0.003), respectively. No difference was found between PPA of hypertensive and non-hypertensive participants.

Hypertensive subjects showed a trend for higher vascular diseases, higher BMI and indexed LVM, although all of these differences were below the threshold of statistical significance (between 0.05 < p< 0.10). There was no difference between hypertensive aortic diameter and normotensive aortic diameter, respectively 3.56 ± 0.35 cm versus 3.51 ± 0.33 cm (p=0.63) in paired-sample T-test.

Relationship between inflammation, thoracic aortic calcifications and various arterial stiffness indices

A positive correlation was observed between SUVmax at the level of the ascending aorta and AoPWV (p=0.03, r=0.38) (Figure 1a). There was also a negative correlation between SUVmax of the ascending aorta and PPA (p=0.04, r=0.36) (Figure 1b).

Figure 1.

Univariate relationship between local inflammation in the ascending aorta assessed by SUVmax and (a) AoPWV, (b) PPA

However, as expected, the strongest positive relationship was found between AoPWV and VCa, especially at the level of the descending aorta (p=0.007, r=0.48).

A significant positive relationship was found between aortic arch SUVmax and the level of systemic inflammation (hsCRP) for the entire population, while no other trends were observed with hsIL-6 and fibrinogen.

The involvement of inflammation (ascending aortic SUVmax) and calcifications (descending aortic VCa) in the determination of AoPWV and PPA was also tested by multivariate analysis, including age and gender in the multivariate model. For AoPWV, both SUVmax and VCa were included since the univariate correlations of these two parameters with AoPWV had a p value <0.10 (p=0.03 for SUVmax; p=0.009 for VCa). PPA was correlated only with SUVmax (p=0.04) and with heart rate (p=0.03), but not with VCa (p=0.36). As shown in Table 2, age and SUVmax were both significant determinants of PPA (model 1). Old age and high inflammation had an important negative effect on PPA accounting for 2.9% (p=0.08) and 13.8% (p<0.05) of its variation, respectively (Table 2, model 1). Multivariate analysis also showed that presence of calcification increased AoPWV, accounting for 18.1% of its variability (p=0.01), while inflammation showed the same trend explaining 9.6% of its variability (p=0.07). Gender and MAP (mean arterial pressure) had no independent effect on either AoPWV or PPA. HR was an independent determinant of PPA; in this case the significant level of the contribution of SUV max was slightly lower (p=0.062 as opposed to p=0.048 when HR was not included in the multivariate model) (model 2).

Table 2.

Multiple regression analysis of the involvement of both infammation (ascending aorta SUVmax) and calcifications (descending aorta calcification volume) in the determination of PPA (Model 1: including Age, Gender and SUVmax and Model 2: including Age, Gender and SUVmax and HR) and AoPWV with inclusion of age and gender in the multivariate model

	R2	Coefficient ± SEM	Prob. Level
PPA (%)
Age (year)	2.9%	-0.26 ± 0.27	0.08
Gender	-	-	0.89
SUVmax Ascending Aorta	13.8%	-12.06 ± 5.90	0.048
Model 1	15.7%
	R2	Coefficient ± SEM	Prob. Level
PPA (%)
Age (year)	1.7%	-0.21 ± 0.26	0.43
Gender	1.4%	-2.44 ± 3.4	0.47
HR	17.2%	0.55 ± 0.22	0.02
SUVmax Ascending Aorta	10.6%	-10.71 ± 5.48	0.062
Model 2	33%
AoPWV (m.s-1)
Age (y)	-	-	0.47
Gender	-	-	0.42
SUVmax Ascending Aorta	9.6%	5.13 ± 2.67	0.07
VCa Descending Aorta (cm3)	18.1%	2.44 ± 0.92	0.01
Model	32.6%

See Table 1 for abbreviations

AoPWV variations relative to the structural parietal properties of the thoracic aorta (SUVmax and VCa) were also analyzed by dividing individuals in the 4 sub-groups described in the Methods section. AoPVW values were higher in the High/High group as compared to the Low/Low group (17.2±1.4 m.s-1, vs. 11.4±1.4 m.s-1; p<0.01). The High/Low groups showed intermediate values (respectively 13.0±1.4 m.s-1 for High SUV/Low VCA and 13.2±1.6 m. s-1 for High VCa/Low SUV, both p=0.05 vs. High/High group Figure 2).

Figure 2.

Aortic stiffness (AoPWV) according to inflammation (SUVmax) at the ascending aorta and calcification volume at the descending aorta: Low=under the median, High= over the median

A positive relationship was found between hsCRP and LVEF (p=0.01, r=0.43). The relationship between hsCRP and LVEF did not persist after adjustment for BMI and other cardiac risk factors. Moreover, a positive relationship was found in univariate analysis between LEVF and PPA (p=0.02, r=0.41) which remained significant after adjustment for age, PPc and MAP (p=0.04).

In contrast, no correlation was found between cardiac parameters and aortic SUVmax at the three levels of the thoracic aorta. A positive correlation was found between calcifications in the ascending aorta and the LVM/EDV ratio measured by MRI (p=0.01, r=0.44). Lastly, there was no correlation between indexed LVM and AoPWV.

There was no univariate relationship between aortic diameter and PPc (p=0.3), and between PPA (p=0.5) and PP (p=0.3).

Discussion

The main result of the present study is that, in the elderly, inflammation of the thoracic aorta is a significant determinant of aortic stiffness enhancement measured by AoPWV and of PPA decrease, showing that these two parameters are probably not influenced by the same artery remodeling process.

Relationship between AoPWV, PPA and thoracic aortic inflammation and calcification

A key point is that we were able to demonstrate a significant enhancement of AoPWV when older individuals presented inflammation of the thoracic ascending aorta.

In a previous study (11), we had established that, in a middle-aged population referred to the Department of Nuclear Medicine in the setting of an existing oncologic indication, thoracic aortic inflammation, assessed by dual FDG-PET/CT imaging, was associated with significant variations in AoPWV when compared to the current impact of calcifications.

The present study further raises the existence of a persistent relationship between local inflammation assessed by SUVmax at the level of the ascending thoracic aorta and arterial stiffness in the elderly.

Arterial stiffness and high pulse pressure are developed during the aging process. Presence of arterial hypertension enhances this process by at least 2 mechanisms: the increase in blood pressure levels has a direct effect by reducing arterial compliance while chronic hypertension induces arterial alteration leading to arterial stiffening and a decrease in compliance (25, 26). The presence of shorter heights in hypertensive individuals also observed in the present study could contribute to early arrival of reflected waves and an increase in pulse pressure. Several studies have demonstrated a strong association between low-grade inflammation (hsIL-6, C-reactive protein) and aortic stiffness (27, 28, 29, 30), and an inverse correlation has been reported with the atheroprotective biomarkers adiponectin and plasminogen activator inhibitor (6). These data hence suggest that inflammation plays an important role in arterial remodeling and stiffening even though the present study did not show a relationship between biological inflammatory markers and various arterial stiffness indices (probably due to the low sample size); however, the study highlights the high sensitivity of PET/CT imaging to low levels of parietal ascending thoracic aortic inflammation.

Using a multivariate analysis, we tested the determinants of AoPWV which revealed that AoPWV was enhanced by both inflammation and calcification. The same analysis conducted for PPA determinants showed that inflammation was an independent determinant for lower PPA. Interestingly, calcifications did not influence PPA, indicating that AoPWV and PPA may not be determined by the same parameters. Indeed, both AoPWV and PPA reflect aortic elastic properties and several parameters influence both of these, including among others peripheral resistance, left ventricular function, aortic geometry and heart rate. However, AoPWV and PPA can provide complementary information relative to arterial aging (4). In a previous study involving very old frail individuals, AoPWV was shown to increase with hypertension and diabetes, whereas PPA was not dependent on BP levels but was decreased in the presence of cardiovascular disease (31). Moreover, PPA was shown to be strongly associated with both cardiovascular and total mortality in elderly (20). Recent studies in retrospective series of individuals have demonstrated that FDG uptake was correlated with the number of cardiovascular risk factors and possibly the risk of future cardiovascular events (6) which could explain the strong relationship with PPA lowering. In the present study, PPA did not significantly differ between the hypertensive and normotensive elderly groups (20), while AoPWV was markedly different between these two populations.

The relationship between AoPWV and calcification has already been described in other studies, in particular in populations of patients presenting high rates of vascular calcification such as diabetes or in end stage renal disease individuals (32) as well as in a middle-aged population (11). Our study shows that the presence of hypertension was closely related with descending aortic calcifications but not with inflammation.

One could therefore suggest that assessment of both AoPWV and PPA could provide additional information for the evaluation of arterial health and cardiovascular risk in the elderly.

Relationship between cardiac parameters, inflammation and calcification

In the present study, hsCRP levels positively correlated with the LVEF of elderly individuals while this relationship did not persist after adjustment for body mass index and cardiovascular risk factors. This result should be interpreted with caution, however, since the participants were not selected on the criteria of presence of coronary heart disease or heart chronic failure in their background. In fact, certain studies have suggested that a high level of inflammation, and especially elevated CRP concentration, is linked to left ventricular dysfunction (33). However, other studies involving individuals with low LVEF found no correlation between hsCRP levels and this cardiac function parameter (34).

In our small study, among elastic aortic properties indices, only PPA was significantly associated with LVEF in univariate analysis, and this association still remained significant after adjustment for age, PPc and MAP. Such findings might be explained by the population size and by the fact that our study included a majority of healthy volunteers. Several parameters influence PPA including peripheral resistance, left ventricular function, aortic geometry and heart rate. It has moreover been demonstrated by multiple cross-sectional observational studies (Anglo-Cardiff Collaborative Trial, Asklepios, PARTAGE studies) that amplification of the central to brachial PP is inversely related to large artery stiffness (assessed by pulse wave velocity), to peripheral resistance, to characteristics of the reflected waves (AI amplification index, reflection coefficient) and more recently to effective aortic diameter (35). In fact, the contribution of each of these parameters on PPA is very difficult to delineate because of their close interrelationship.

Furthermore, we found a positive relationship between VCa of the ascending thoracic aorta and the LVM/EDV ratio. No relationship was found between AoPWV and LVM. It is known that arterial calcification develops with aging in the intima and media layers of the large and medium-sized arterial walls and is generally associated with arterial stiffening which can ultimately lead to an enhancement of ventricular afterload.

This could be interpreted as a sign of coupling of left ventricular geometry and vascular remodeling due to the changes in aortic geometry. The lack of association between AoPWV and LVM in the present study can be explained by the relative small number of subjects. Redheuil et al. (14) studied 108 subjects describing a close relationship between the decrease in aortic arch curvature and an increase in LVM, although no significant association was found with AoPWV. Unfortunately, in their study, the use of MRI did not allow to investigate the role of aortic calcifications on LVM variation. Another study conducted in over 1300 individuals examining the best predictors of LVM found that AoPWV was only a minor predictor of its variation (36). These data hence show that large populations are often needed in order to ascertain the impact of arterial stiffness on LVM.

In a study conducted in a large Taiwanese population of normotensive and untreated hypertensive volunteers (with a huge age distribution), Farasat et al. showed the existence of an inverse strong association, after multivariate analysis, between PP and aortic diameter measured by echocardiography (37). This latter study comprised a large scale of PP values of subjects of different ages. The same trend was found by Torjesen et al. in the AGES-Reykjavik study, conducted with MRI measurements of aortic area, showing that higher PP in older European individuals (hypertensive or normotensive) was associated with smaller aortic lumen area (38). These studies have shown that after adjusting for age, gender, blood pressure and other confounders, a negative association was observed between aortic diameter and pulse pressure. In our study, univariate analysis revealed a lack of association between aortic diameter and central or peripheral pulse pressure. This finding coupled with the limited number of patients thus precluded conducting a multivariate analysis to further investigate this particular issue.

Perspectives

There is growing evidence that inflammation is related to cardiovascular events. The present study, using PET/CT scan imaging and cardiac MRI methods yields a better understanding of the mechanisms underlying vascular remodeling. Indeed, the data herein provide new evidence that segmental arterial stiffness and PPA yield complementary information in older patients whereby segmental arterial stiffness is linked to calcifications and inflammation while PPA is more closely related to local thoracic aortic inflammation on ascending aorta and to left ventricular ejection fraction. The role of inflammation, calcification, vascular growth and remodeling may be all the more important given their therapeutic implications. The CAFE study (39) demonstrated that blood pressure-lowering therapies can have differing effects on central aortic pressure and especially on central pulse pressure while no difference was evidenced on brachial pulse pressure. Moreover, advocated mechanisms were a consequence of a possible change in arterial stiffness and difference in the proximity of wave reflection sites. The present results of this pilot study further open the discussion regarding the possible effects of certain anti- hypertensive drugs on local aortic inflammation.

Limitations of the study

Potential limitations of the present study need to be considered such as the small sample size of the studied population and its selectivity (only elderly individuals). Another limitation includes the time difference since regional arterial stiffness and blood pressure and FDG PET/CT imaging measurements could not be performed simultaneously in the MRI. All measurements were repeated to minimize individual variability.

Conclusion

The primary finding of the present study suggests that, in older subjects, presence of inflammation of the thoracic aorta has deleterious effects by increasing aortic stiffness, measured by AoPWV, and by decreasing PPA, an indicator of wave reflection magnitude and global arterial stiffness. Aortic calcifications have a deleterious effect only in AoPWV. Taken together, the present findings indicate that, when compared with normotensive elderly subjects, hypertensive individuals of the same age show increased systemic and regional arterial stiffness, more calcifications and a higher inflammatory status at the site of their aorta. Given its small sample size, the results of this study should however be interpreted with caution. Larger studies are needed to confirm the importance of local aortic inflammation in the pace of vascular aging.

Acknowledgments : The study was conducted with the logistic support of the Centre of Clinical Investigations (INSERM CIC-P) and (CIC-IT) of the University Hospital of Nancy. We thank Giuseppe Lio and Paolo Salvi for their contributions. We also thank Mr. Pierre Pothier for language review and stimulating discussions.

Sources of funding: This work was supported by the French Society of Hypertension (2009). Supplementary financial support was provided by Nancyclotep fundings (2008). This study has also been supported by the National Institute of Health (INSERM) and the PPF (Plan Pluri-Formation 2009) of the French Ministry of Research. This study has been released on ClinicalTrials.gov under the identifier: NCT 01963221

Conflict(s) of Interest/Disclosure(s) Statement: None

Ethical Standards: Experiments realized for this study comply with the current french laws.