Carotid endothelial shear stress reduction with aging is associated with plaque development in twelve years

Atherosclerosis 251 (2016) 63e69

Contents lists available at ScienceDirect

Atherosclerosis journal homepage: www.elsevier.com/locate/atherosclerosis

Carotid endothelial shear stress reduction with aging is associated with plaque development in twelve years Claudio Carallo a, b, *, Cesare Tripolino a, Maria Serena De Franceschi a, Concetta Irace a, Xiao Yun Xu b, Agostino Gnasso a a b

Metabolic Diseases Unit, Department of Clinical and Experimental Medicine, “Magna Græcia” University, Viale Europa, 88100, Catanzaro, Italy Department of Chemical Engineering, Imperial College London, South Kensington Campus, London, SW7 2AZ, United Kingdom

a r t i c l e i n f o

a b s t r a c t

Article history: Received 12 August 2014 Received in revised form 15 May 2016 Accepted 27 May 2016 Available online 28 May 2016

Background and aims: Atherosclerosis is associated with clinical, biochemical and haemodynamic risk factors. In a group of subjects studied twelve years apart, we evaluated carotid plaque development in relation to baseline and to changes at follow-up in common carotid haemodynamic profile. Methods: Forty-eight participants were recruited to a cardiovascular disease prevention programme. Atherosclerotic plaques were evaluated and scored by echography. Endothelial shear stress, circumferential wall tension, and Peterson’s elastic modulus as an index of arterial stiffness, were computed by echo-Doppler, along with blood viscosity data. Binary logistic regression analyses were used to test the association among the development of atherosclerosis, cardiovascular risk factors and haemodynamic variations. Analyses were also performed on participants who presented at the follow-up with carotid haemodynamic variations in the left or right common carotid only. Results: Participants (69% male) were aged 64.5 ± 9.7 years at follow-up. Peak and mean endothelial shear stress was significantly lower at follow-up as previously reported; circumferential wall tension and arterial stiffness were significantly higher. Carotid plaque scores increased after 12 years (0.39 ± 0.72 vs. 0.67 ± 0.86, p < 0.01). Of the 96 common carotids analysed, shear stress reduction with aging was an independent predictor of carotid atherosclerosis (B ¼ 0.063; odds ratio ¼ 0.94; p ¼ 0.01). Out of 48 participants, 21 (44%) showed shear stress reduction with aging in only one side of the body and, on this side, the plaque score increased (0.52 ± 0.98 vs. 0.90 ± 0.94, p < 0.05), remaining unchanged in the contralateral carotid tree. Conclusions: Aging-related shear stress reduction is an independent predictor of atherosclerosis development. © 2016 Elsevier Ireland Ltd. All rights reserved.

Keywords: Carotid arteries Haemodynamics Atherosclerosis

1. Introduction Atherosclerosis is a systemic disease, with focal manifestations strongly dependent on local haemodynamic factors. Endothelial shear stress (also called wall shear stress, or shear stress), the friction exerted by blood moving on the endothelium, and circumferential wall tension, acting circumferentially due to the perpendicular component of haemodynamic forces (i.e. blood pressure), are two important haemodynamic factors involved in

* Corresponding author. Metabolic Diseases Unit, Department of Clinical and Experimental Medicine, “Mater Domini” Hospital, “Magna Græcia” University, Catanzaro, Italy. E-mail address: [email protected] (C. Carallo). http://dx.doi.org/10.1016/j.atherosclerosis.2016.05.048 0021-9150/© 2016 Elsevier Ireland Ltd. All rights reserved.

arterial structure and function [1e3]. In this field, most of the data has been derived from lab models and patient-specific flow modelling data, or from cross-sectional human studies [4e6]. Several human studies have demonstrated that acute reduction of wall shear stress causes endothelial dysfunction [7], but there is a lack of longitudinal studies investigating the chronic effect of decreasing shear stress on the development of atherosclerosis. This is probably due to the fact that reliable techniques for the measurement of wall shear stress have only been developed relatively recently [8e10]. Nonetheless, longitudinal studies, which were followed-up within months, have already demonstrated the progression of atherosclerotic lesions in native segments of the coronary arteries presenting low wall shear stress [10e12]. Furthermore, decreased wall shear stress seems to be associated with increased wall

64

C. Carallo et al. / Atherosclerosis 251 (2016) 63e69

thickness in venous bypass grafts of lower limbs, after a follow-up period of 9e12 months [13]. A single prospective study demonstrated a reduction of internal carotid wall shear stress when followed-up 2.7 years later, confirming previous cross-sectional reports. This study, however, did not evaluate the development/progression of atherosclerosis [14]. In the late 1990s, we conducted a comprehensive common carotid haemodynamic evaluation using echo-Doppler and blood viscometry as part of a cardiovascular disease prevention programme [15]. Twelve years later, the same population was recalled in order to repeat the examinations, which showed that shear stress had decreased significantly. This age-related change was independently associated with increases in local intima-media thickness, an early atherosclerosis marker [16]. The aim of the present study is to evaluate the influence on atherosclerotic plaque development of baseline wall shear stress and aging-related haemodynamic changes in common carotid wall shear stress, circumferential wall tension and stiffness. 2. Patients and methods 2.1. Subjects and study design Subjects were part of a longitudinal observational study; the baseline cross-sectional study was previously reported [17,18]. Details of follow-up analysis after 11.6 years, in terms of the studied population and applied methods have been reported elsewhere [16]. Briefly, subjects were participants who voluntarily joined a cardiovascular disease prevention programme. All participants were informed about the aim of the campaign and provided informed consent. From 1996, the evaluation of haemodynamic forces within the common carotid arteries was added and performed in 147 selected subjects. Exclusion criteria were: arrhythmia at ECG, common carotid arteries that were not straight, or that presented carotid plaques at the arterial site of the haemodynamic measurements, a high degree stenoses of the carotid arteries downstream to the common carotid, which can possibly reduce blood flow volume upstream, females not in menopause, use of diuretic and anticoagulant drugs, severe anaemia (haemoglobin <10 g/dL), polycythaemia (RBC>6  106 cells per mL), clinically significant renal, hepatic or pulmonary disease, severe hyperglycaemia (fasting plasma glucose >250 mg/dL), heart failure. In 2008, the subjects (n ¼ 147) were recalled by phone and mail: 8 subjects had died, 70 were unavailable or had moved far, and 19 refused to participate. Fifty subjects agreed to participate: after clinical examination, one subject was excluded because of severe anaemia and one because of a malignancy; 48 (15 women and 33 men) underwent the haemodynamic examination and were included in the follow-up. The study protocol was approved by the Ethics Committee. 2.2. Clinical and blood parameters All subjects included in the study underwent complete clinical examination and blood tests at baseline and follow-up. Body mass index was computed as weight (in kg) divided by height (in m) squared. Systolic (SBP) and diastolic (DBP) blood pressure were measured, on the right arm, after the participant had been resting for at least 5 min, with a standardized sphygmomanometer. The average of the second and third of three readings was computed. Mean blood pressure was computed as: DBP þ ((SBPDBP)/3) [15]. Venous blood for biochemical and viscosity analyses was collected after overnight fasting. Blood glucose and lipids were measured by routine methods. Diabetes was defined as fasting

blood glucose 126 mg/dL and/or use of antidiabetic agents. Hyperlipidaemia was defined as total cholesterol and/or triglycerides exceeding 200 mg/dL and/or use of lipid lowering drugs. Hypertension was defined as SBP/DBP 140/90 mmHg and/or use of antihypertensive agents. Blood viscosity was measured within 2 h of blood withdrawal, at shear rate 225 s1 and 37  C with a cone-plate viscometer (WellsBrookfield DV-III, U.S.A.); the same procedure was followed at both examinations. The viscometer was regularly checked during the 12 year study period and was overhauled, with several parts being replaced, before the follow-up examination in order to ensure that it was functioning perfectly. In both the examinations, a standard fluid with a known viscosity (Standard Fluid 5 CPS, WellsBrookfield, USA) was used to calibrate the instrument [17]. 2.3. Ultrasound study For clinical examination, the internal, external, bulb and common carotid were examined to evaluate the presence of plaques and to eliminate subjects who did not meet the inclusion criteria. Atherosclerotic plaque was defined as localized intima-media lesions of the upper and/or bottom walls on B-mode images with a thickness of at least 1.3 mm, no spectral broadening, or only in the deceleration phase of systole and a systolic peak velocity of less than 120 cm/s. Normal segments were scored 0, and those with plaque were scored 1. The thickness cut-off for plaque was used in the former examination in the 1990’s, and was replicated at followup; a global carotid plaque score was computed by adding the scores of all segments [19,20]. Echo-Doppler examination for arterial diameter and blood flow velocity measurements was performed in a quiet room at 22  C, as in the baseline examination, with an ECG-triggered high-resolution instrument that was equipped with a multi-frequency linear probe. The instrument used at baseline was an ATL UltraMark 9 HDI (Philips), equipped with a linear phased array multi-frequency 5e10 MHz probe. In the follow-up study we used an ATL HDI 3000 (Philips), equipped with a similar linear phased array multifrequency 5e10 MHz probe. In common carotid arteries that did not present atherosclerotic plaques at the measurement site EchoDoppler parameters were acquired 1 cm proximal to the line dividing the common carotid and the carotid bulb. Curved common carotids within 10 cm upstream of the measurement site were also excluded because of the impact on haemodynamics. Internal diameter (ID) was defined as the distance between the leading edge of the echo produced by the intima-lumen interface of the near wall and the leading edge of the echo produced by the lumen-intima interface of the far wall. ID was measured at the R and T waves of the ECG, representing the minimum (IDR) and maximum (IDT) carotid diameter, respectively. Blood flow velocity was detected with the sample volume reduced to 1 mm and placed in the centre of the vessel. The angle between the ultrasound beam and the longitudinal vessel axis (q) was kept between 44 and 56 , and was recorded. The maximum Doppler frequency shift, i.e. systolic peak velocity (VSP), and mean velocity (VM) were automatically recorded with auto-tracking as the mean of three cardiac cycles. 2.4. Calculations Peak (tP ) and mean (tM ) wall shear stress were calculated using the following equations [4], valid for a Newtonian Poiseuille parabolic flow profile (please also see the Limitations section below):






tP dynes cm2 ¼ 4hVSP =IDT

C. Carallo et al. / Atherosclerosis 251 (2016) 63e69



tM ðdynes cm2 Þ ¼ 4hVM =IDR where V is expressed in cm/s, ID in cm, and blood viscosity (h) in poise. Peak (TP) and mean (TM) circumferential wall tension [21] (an index of normal wall stress) were computed as:

TP ¼ SBPðIDT =2Þ TM ¼ MBPðIDR =2Þ where T is expressed in dynes/cm, blood pressures in dynes/cm2 and ID in cm. Peterson’s elastic modulus (the ratio between pulse pressure and carotid distensibility) [22] was used as an index of arterial stiffness, according to the following relationship:

 Pðdynes cm2 Þ ¼ ðSBP  DBPÞIDR =ðIDT  IDR Þ The brachial blood pressure was used as an index of central arterial pressure (please also see the Limitations section below) [23]. All examinations at follow-up were performed in the same way and by the same operator who was blinded with regard to the results of the first examination. Furthermore, for each patient, we used the same setting (recorded during the first examination), e.g., focal depth, transducer aperture size, beam steering, depth and size of sample volume, beam-vessel angle, and the exact distance from the carotid bulb where the diameters were measured, in order to obtain consistent and comparable data. Because both the right and left common carotid arteries were examined and analysed separately, the total sample size is 96. 2.5. Reproducibility The reproducibility of the measurements has been previously reported, including in terms of coefficients of variations [4,16]. In summary, the haemodynamic data of 36 subjects who participated in the baseline examination were compared to 36 new subjects matched for age, sex and cardiovascular risk factors: no significant differences were detected regarding blood viscosity and velocity, arterial diameter, and shear stress. Mean coefficient of variation for peak shear stress values was 4.0 ± 2.2, range 1.9e7.6; for mean shear stress was 5.0 ± 2.4, range 2.6e8.7.

65

development of carotid atherosclerosis. This last model was structured in two blocks. In order to correct for the presence of risk factors at baseline, the first block (method: enter) included baseline values of age, male gender, smoking, obesity, hypertension, diabetes, dyslipidaemia, plus each individual haemodynamic measurement. The second block (method: stepwise forward) included the development at the follow-up visit of smoking, obesity, hypertension, diabetes, dyslipidaemia, plus the percentage variation in the individual haemodynamic measurements included in the first block. In order to exclude a cluster effect derived by the analysis of both carotids together, the intraclass correlation coefficient (ICC) was calculated to estimate the variance explained by withinpatient variability, on haemodynamic variables that were associated to carotid atherosclerosis. Furthermore, mixed model analyses were applied to the sets of variables included in all the binary logistic regression analyses. Then, analyses were restricted to participants presenting, at the follow-up, a decrease both in peak and in mean wall shear stress on one common carotid only, to see if the observed relation between endothelial shear stress and atherosclerosis could be side dependent in some participants. The cut-off value chosen was a reduction by 5%, corresponding to the mean coefficient of variation of the present method for shear stress measurement [4] at baseline (coefficient of variation at follow-up was unchanged). In this group, basal and follow-up values of the carotid plaque score, haemodynamic variables and their determinants were compared within each body side (decreased or unchanged in haemodynamics). The percentage variations in the carotid diameters and velocities at the follow-up were also compared between the two sides of the body. The Student t-test for paired data and the Wilcoxon test were used as appropriate. ICC and mixed model analyses were performed using Stata 13 (StataCorp LP, College Station, TX). All other statistical analyses were performed by PASW Statistics 18.0.0 (IBM Corporation, Armonk, NY). 3. Results The clinical data of the 48 participants are reported in Table 1. All hypertensives, dyslipidaemic and/or diabetic patients were on lifestyle intervention, plus pharmacological therapy if needed, according to good clinical practices of the respective time periods. Mean follow-up was 11.6 ± 0.8 years. At follow-up, systolic blood pressure and prevalence of hypertension significantly increased. Blood viscosity increased from 4.5 ± 0.4 to 4.8 ± 0.5 cP.

2.6. Statistical analysis 3.1. Regression analyses on the whole population All continuous variables, except carotid plaque score, followed a normal distribution. Male sex, cardiovascular risk factors (smoking, obesity, hypertension, diabetes, dyslipidaemia) and carotid atherosclerosis (both presence at baseline and development at follow-up) were used as categorical variables (yes/no). Chi square and Student t-tests for paired data were used to compare baseline and follow-up data as appropriate. Pearson’s correlation coefficient was applied to evaluate the association between baseline and follow-up wall shear stress, for both peak and mean values, and between all shear stress values and arterial stiffness. Binary logistic regression analysis was used to verify if the presence of cardiovascular risk factors and haemodynamic values at baseline were associated to the presence of carotid atherosclerosis 12 years later. A further binary logistic regression analysis was used to verify if the appearance of cardiovascular risk factors and haemodynamic variations during follow-up predicted the

Wall shear stress values at baseline and after 12 years of followup were strongly related (peak values: r2 0.64, p < 0.0001; mean values: r2 0.57, p < 0.0001, Fig. 1). All haemodynamic parameters of the 96 common carotids analysed worsened with aging (Table 1). Peak and mean wall shear stress decreased by 12% and 18%, respectively; peak and mean circumferential wall tension significantly increased, as well as arterial stiffness by Peterson’s modulus. Carotid plaque score almost doubled after 12 years. Considering the two carotids of the same patient as a possible cluster, the ICC was calculated for all shear stress values. The estimated correlations between measurements ranged from 0 to 0.28, indicating that the variability of the within-patient measurements was similar to the between-patients variability. Baseline values of age and carotid wall shear stress were independent predictors of carotid plaque presence at follow-up by

66

C. Carallo et al. / Atherosclerosis 251 (2016) 63e69 Table 1 Clinical features of 48 participants and haemodynamic values of 96 common carotids at baseline and follow-up visits (male gender ¼ 69%).

Age (years) Systolic blood pressure (mmHg) Diastolic blood pressure (mmHg) Body mass index (Kg/m2) Smoking (%) Hypertension (%) Drugs for hypertension (%) Dyslipidaemia (%) Drugs for dyslipidaemia (%) Obesity (%) Diabetes mellitus (%) Drugs for diabetes (%) Peak endothelial shear stress (dynes/cm2) Mean endothelial shear stress (dynes/cm2) Peak circumferential wall tension (104 dynes/cm) Mean circumferential wall tension (104 dynes/cm) Peterson’s modulus (105 dynes/cm) Carotid plaque score

Baseline

Follow-up

52.9 ± 10.0 121.8 ± 15.5 78.6 ± 7.8 27.3 ± 3.1 7 24 22 28 7 17 20 13 20.6 ± 5.6 10.8 ± 2.7 4.9 ± 0.8 3.5 ± 0.5 9.1 ± 7.2 0.39 ± 0.72

64.5 ± 9.7 138.7 ± 21.2* 80.6 ± 9.8 28.7 ± 3.5 2 59* 51x 33 24* 28 25 22# 18.0 ± 5.5x 8.9 ± 3.1x 5.9 ± 1.2x 4.0 ± 0.7x 12.4 ± 8.0x 0.67 ± 0.86x

Baseline vs. follow-up Student paired t-test statistical significances (c square for hypertension, and for drugs use; Wilcoxon test for Echo Doppler Score).*p < 0.001; xp  0.01.

Fig. 1. Comparison between mean values of endothelial shear stress in the common carotid at baseline and after 12 years of follow-up. (A) Pearson’s correlation (n ¼ 96; r2 ¼ 0.57, p < 0.001). (B) Bland Altman plot (SD ¼ standard deviation).

binary logistic regression (data not shown), but when performing mixed model analyses with the same sets of variables in order to consider the possible cluster effect of the two carotids from the same body side, age only remained significant (Tables 1 and 2 of the Supplementary data). Other baseline values, including circumferential wall tension and Peterson’s modulus, did not enter into any of the models. Looking at the incidence of cardiovascular risk factors, and at the changes in vessel parameters during the follow-up, apart from age, only the percentage variation of wall shear stress after 12 years was significantly and independently related to plaque progression or development: subjects with more marked wall shear stress reduction also developed higher carotid plaque score (Tables 2 and 3). The percentage variation of circumferential wall tension and Peterson’s modulus did not enter into the model. Furthermore, mixed model analyses were performed with the same sets of variables of the above reported logistic regression analyses, and the results were similar (Tables 3 and 4 of the Supplementary data).

Both the baseline and follow-up values of Peterson’s modulus were weakly, but significantly, inversely related to baseline and follow-up shear stresses (r 0.39, p < 0.01 between baseline stiffness and baseline mean shear stress, r 0.24, p ¼ 0.02 between baseline stiffness and follow-up mean shear stress; other analyses gave similar results). 3.2. Carotids with asymmetric haemodynamic evolution within the same subject Aging-related wall shear stress reduction was asymmetric in 21 (44%) subjects who showed peak and mean shear stress reduction 5% at follow-up on one side of the body only (57% of them on the right common carotid). Variations in peak and mean values of haemodynamic variables showed similar results; to avoid redundancy, therefore, only mean values are presented (Fig. 2). On the side with shear stress reduction (21% for peak and-28% for mean values), the plaque score increased from 0.52 ± 0.98 to 0.90 ± 0.94

C. Carallo et al. / Atherosclerosis 251 (2016) 63e69

67

Table 2 Binary logistic regression analysis on 96 common carotids (final model, variables with significance). Dependent variable: atherosclerotic plaque development at follow-up visit (presence/absence). Variable

B

Odds ratio

95% confidence interv.

p

Age (years) Mean endothelial shear stress variations (dynes/cm2)

0.11 0.03

1.12 0.97

1.03e1.21 0.94e0.99

0.007 0.047

Independent variables were entered in two blocks: first block (method: enter) included baseline values of age, male gender, smoking, obesity, hypertension, diabetes, dyslipidaemia, common carotid mean shear stress; second block (method: stepwise forward) included the appearance at the follow-up visit of smoking, obesity, hypertension, diabetes, dyslipidaemia, plus percentage variations in common carotid mean shear stress.

Table 3 Binary logistic regression analysis on 96 common carotids (final model, variables with significance). Dependent variable: atherosclerotic plaque development at follow-up visit (presence/absence). Variable

B

Odds ratio

95% confidence interv.

p

Age (years) Peak endothelial shear stress variations (%)

0.125 0.063

1.13 0.94

1.03e1.24 0.89e0.99

0.007 0.010

Independent variables were entered in two blocks: first block (method: enter) included baseline values of age, male gender, smoking, obesity, hypertension, diabetes, dyslipidaemia, common carotid peak shear stress; second block (method: stepwise forward) included the appearance at the follow-up visit of smoking, obesity, hypertension, diabetes, dyslipidaemia, plus percentage variations in common carotid peak shear stress.

Fig. 2. Comparison between carotids on the two body sides within the same subject at baseline and after 12 years of follow-up, among participants with common carotid peak and mean endothelial shear stress decrease  5% at follow-up on one common carotid only (n ¼ 21). (A) Mean shear stress; (B) Mean circumferential wall tension; (C) Peterson’s modulus; (D) Carotid Plaque Score. Student t-test statistical significances (Wilcoxon test for Carotid Plaque Score). *p < 0.05 vs. baseline.

(p  0.05). In the opposite carotid artery, with stable shear stress, no atherosclerosis developed despite a worsening in circumferential wall tension. These variations were a consequence of enlarged carotid diameters and reduced blood flow velocities, differently into the two sides of the body. IDR increased by 10.2 ± 1.3% in the carotid with shear stress reduction vs 3.9 ± 0.7% in the opposite side (p ¼ 0.02). VSP decreased by 27.6 ± 12.7% vs. 11.6 ± 15.0% respectively (p < 0.001). Asymmetrical changes in circumferential wall tension and arterial stiffness were found in two and one subjects, respectively, indicating a lack of side dependence of these carotid haemodynamic features in the studied population.

4. Discussion The present findings demonstrate, for the first time in humans, that aging-related wall shear stress reductions are independent predictors of carotid plaque development and/or progression. Other parameters commonly used to evaluate the arterial wall, such as circumferential wall tension and wall stiffness, seem to play a minor role. These prospective data are concordant with the results of previous cross-sectional studies on the association between wall shear stress and atherosclerosis [4e6]. Furthermore, the analysis of subjects showing asymmetric shear stress reduction, clearly demonstrates that the development and/or progression of plaques

68

C. Carallo et al. / Atherosclerosis 251 (2016) 63e69

occur exclusively on the side where shear stress decreases. In a previous study, in which we used a cross-sectional approach, and compared the left and right common carotids of the same individual, we found a strong association between wall shear stress and plaques in subjects with asymmetrical carotid atherosclerosis, fully in line with the present results [19]. The present findings confirm previous cross-sectional reports, and emphasize the importance of the role of haemodynamic forces in the development of atherosclerosis. The development of plaque in arterial segments with reduced shear stress has already been demonstrated in the coronaries after a few months of follow-up [10e12]. In contrast to the previously reported data, the present results focused on aging-related wall shear stress reduction, with a follow-up of ~12 years; furthermore, the present results looked at the carotid bifurcation, thereby also making the haemodynamic evaluation more comprehensive. In addition, the present results expand our previous prospective data on the relationship between shear stress and intima-media thickness in the same population [16]; in the present data, established carotid atherosclerosis has been considered; tensile forces have been taken into account, and the asymmetric atherosclerosis model has ruled out many possible confounding factors. That previous study [16] had already demonstrated that shear stress reductions with aging rely on blood flow velocity reductions partly due to arterial enlargement; the present data confirms this conclusion in a context of the different behaviour exhibited between the two carotid arteries of the same patient. Explanations of these findings are beyond the aims of the present study. Previously it has been hypothesized that increases in diameter with aging might be due to fatigue failure of the arterial wall and/or to overcompensation due to the process of atherosclerosis [15 16]. The differences between the two sides are less easy to explain. Considering that head positions/rotations have a great effect on common carotid haemodynamics [24], it is possible that head movements and consequent shear stress variations influence common carotid geometry differently during one’s lifespan. In the present investigation we found no independent association between the incidence of carotid plaque and common carotid circumferential wall tension. We were not able to find any study in the literature investigating this specific association using a longitudinal approach, as in our study. Moreover, in the present study, common carotid stiffness showed conflicting results regarding its association with carotid plaque development. Here, the association has been partially verified in a side to side model (i.e. the right vs. left carotid comparison), but in the multiple regression analysis including the whole population this association did not reach statistical significance. Quite possibly, in the present study the limited sample size accounts for this discrepancy. Even if arterial stiffness is usually associated with the incidence of cardiovascular events, due to the wide pathophysiological implications of its alterations [25], measurement methods and the anatomic site chosen to investigate arterial distensibility and plaque presence can strongly influence the results [26]. In a recent study with similar methods and length of follow-up, however, carotid arterial stiffness was found to be associated with the incidence of ischemic stroke [27], disproving a finding presented in a previous study [28]. Notably, apart from age, no other baseline value (clinical, biochemical, or haemodynamic) demonstrated a clear association with plaque presence at follow-up. Probably, the explanation for this discrepancy should be found in the limited sample size and in the very long observation period; the latter probably emphasizes the role of the worsening risk factors for atherosclerosis during the observation period rather than their baseline values. Looking at blood viscosity increases with aging, a previous

paper analysed these findings and their possible implications [29]. 4.1. Clinical implications This study strongly supports the importance of local haemodynamic factors in the development of atherosclerosis. The evaluation of these “physical” factors, namely low wall shear stress [30], may deserve the kind of clinical attention now reserved almost exclusively for classical “chemical” cardiovascular risk factors (the exception being arterial hypertension, which also impacts on haemodynamics) [31]. 4.2. Limitations In the recent papers by Samady and Stone [11,12], more complex haemodynamic profiling has been applied during invasive procedures, based on coronary angiography, intravascular ultrasound, and computational modelling of fluid dynamics. Despite the costs, time, procedural risk, and patient discomfort, these methodologies enable the measurement of wall shear stress at the same arterial point of plaque presence, whereas the simpler methodology developed in the 1990s and applied in the present follow-up study allowed us to test the association between shear stress in an arterial point and atherosclerosis presence nearby. We think that these two widely different methodological approaches are addressed to different study designs. In the present protocol, however, it was clearly necessary to retain the same experimental setting as in the baseline study in the mid-1990s. The population is quite heterogeneous, and some subjects began treatment for the correction of cardiovascular risk factors during the observation period. The multiple regression statistical approach and the side to side comparison have probably overcome these threats. Due to the small number of recruited subjects, it was not possible to perform subgroup analyses. A methodological issue might arise due to the assumption of Hagen-Poiseuille flow in the common carotid artery, both at baseline and at follow-up examination, as discussed extensively elsewhere [16]. Here, it is possible to add that, very recently, it has been demonstrated that a Poiseuillean common carotid mean wall shear stress computation is applicable to at least two-thirds of patients [32]. Blood pressure was measured in the brachial artery but used for carotid wall stiffness and tension computations. Due to the different structures and positions of these two arteries in the arterial tree with respect to the reflected pressure waves, this assumption tends to lead to an overestimation of the measurement that might not be systematic in the present study because it depends on age [23]. Again, the side to side approach should have limited this problem. Finally, the formula for circumferential wall tension and Peterson’s modulus computations is valid assuming an isotropic diseased-free vessel wall under linear elasticity constraint [33]. 4.3. Conclusions Aging-associated wall shear stress reductions predict the development/progression of atherosclerotic plaques in the carotid arteries independently of known cardiovascular risk factors. Conflict of interest The authors declared that they do not have anything to disclose regarding conflict of interest with respect to this manuscript.

C. Carallo et al. / Atherosclerosis 251 (2016) 63e69

Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.atherosclerosis.2016.05.048. References [1] E. Cecchi, C. Giglioli, S. Valente, C. Lazzeri, G.F. Gensini, R. Abbate, L. Mannini, Role of hemodynamic shear stress in cardiovascular disease, Atherosclerosis 214 (2011) 249e256. [2] Y.S. Chatzizisis, R. Blankstein, P. Libby, Inflammation goes with the flow: implications for non-invasive identification of high-risk plaque, Atherosclerosis 234 (2014) 476e478. [3] M.Y. Lee, C.M. Wu, K.H. Yu, C.S. Chu, K.T. Lee, S.H. Sheu, W.T. Lai, Association between hemodynamics in the common carotid artery and severity of carotid atherosclerosis in patients with essential hypertension, Am. J. Hypertens. 21 (2008) 765e770. [4] A. Gnasso, C. Carallo, C. Irace, V. Spagnuolo, G. De Novara, P.L. Mattioli, A. Pujia, Association between intima media thickness and wall shear stress in common carotid arteries in healthy male subjects, Circulation 94 (1996) 3257e3262. [5] S.K. Samijo, J.M. Willigers, R. Barkhuysen, P.J.E.H.M. Kitslaar, R.S. Reneman, P.J. Brands, A.P.G. Hoeks, Wall shear stress in the human common carotid artery as function of age and gender, Cardiovasc. Res. 39 (1998) 515e522. [6] G.D. Schmidt-Trucks€ ass, A. Schmid, R. Boragk, C. Upmeier, J. Keul, M. Huonker, Structural, functional, and hemodynamic changes of the common carotid artery with age in male subjects, Arterioscler. Thromb. Vasc. Biol. 19 (1999) 1091e1097. [7] N.T. Jenkins, J. Padilla, L.J. Boyle, D.P. Credeur, M.H. Laughlin, P.J. Fadel, Disturbed blood flow acutely induces activation and apoptosis of the human vascular endothelium, Hypertension 61 (2013) 615e621. € nn, [8] M.M. Pedersen, M.J. Pihl, P. Haugaard, K.L. Hansen, T. Lange, L. Lo M.B. Nielsen, J.A. Jensen, Novel flow quantification of the carotid bulb and the common carotid artery with vector flow ultrasound, Ultrasound Med. Biol. 40 (2014) 2700e2706. [9] P. Tortoli, T. Morganti, G. Bambi, C. Palombo, K.V. Ramnarine, Noninvasive simultaneous assessment of wall shear rate and wall distension in carotid arteries, Ultrasound Med. Biol. 32 (2006) 1661e1670. [10] P.H. Stone, A.U. Coskun, S. Kinlay, M.E. Clark, M. Sonka, A. Wahle, O.J. Ilegbusi, Y. Yeghiazarians, J.J. Popma, J. Orav, R.E. Kuntz, C.L. Feldman, Effect of endothelial shear stress on the progression of coronary artery disease, vascular remodeling, and in-stent restenosis in humans: in vivo 6-month follow-up study, Circulation 108 (2003) 438e444. [11] H. Samady, P. Eshtehardi, M.C. McDaniel, J. Suo, S.S. Dhawan, C. Maynard, L.H. Timmins, A.A. Quyyumi, D.P. Giddens, Coronary artery wall shear stress is associated with progression and transformation of atherosclerotic plaque and arterial remodeling in patients with coronary artery disease, Circulation 124 (2011) 779e788. [12] P.H. Stone, S. Saito, S. Takahashi, Y. Makita, S. Nakamura, T. Kawasaki, A. Takahashi, T. Katsuki, S. Nakamura, A. Namiki, A. Hirohata, T. Matsumura, S. Yamazaki, H. Yokoi, S. Tanaka, S. Otsuji, F. Yoshimachi, J. Honye, D. Harwood, M. Reitman, A.U. Coskun, M.I. Papafaklis, C.L. Feldman, PREDICTION Investigators. Prediction of progression of coronary artery disease and clinical outcomes using vascular profiling of endothelial shear stress and arterial plaque characteristics: the PREDICTION Study, Circulation 126 (2012) 172e181. [13] M. Jackson, N.B. Wood, S. Zhao, A. Augst, J.H. Wolfe, W.M.W. Gedroyc, A.D. Hughes, S.A.M. Thom, X.Y. Xu, Low wall shear stress predicts subsequent development of wall hypertrophy in lower limb bypass grafts, Artery Res. 3 (2009) 32e38. [14] F.M. Box, J. van der Grond, A.J. de Craen, I.H. Palm-Meinders, R.J. van der Geest, J.W. Jukema, J.H. Reiber, M.A. van Buchem, G.J. Blauw, PROSPER Study Group, Pravastatin decreases wall shear stress and blood velocity in the internal carotid artery without affecting flow volume: results from the PROSPER MRI study, Stroke 38 (2007) 1374e1376.

69

[15] C. Carallo, C. Irace, A. Pujia, M.S. De Franceschi, A. Crescenzo, C. Motti, C. Cortese, P.L. Mattioli, A. Gnasso, Evaluation of common carotid hemodynamic forces. Relations with wall thickening, Hypertension 34 (1999) 217e221. [16] C. Irace, C. Carallo, M.S. De Franceschi, F. Scicchitano, M. Milano, C. Tripolino, F. Scavelli, A. Gnasso, Human common carotid wall shear stress as a function of age and gender: a 12-year follow-up study, Age (Dordr) 34 (2012) 1553e1562. [17] C. Irace, F. Scavelli, C. Carallo, R. Serra, A. Gnasso, Plasma and blood viscosity in metabolic syndrome, Nutr. Metab. Cardiovasc. Dis. 19 (2009) 476e480. [18] A. Gnasso, M.C. Calindro, C. Carallo, G. De Novara, M. Ferraro, G. Gorgone, C. Irace, P. Romeo, D. Siclari, V. Spagnuolo, R. Talarico, P.L. Mattioli, A. Pujia, Awareness, treatment and control of hyperlipidaemia, hypertension and diabetes mellitus in a selected population of southern Italy, Eur. J. Epidemiol. 3 (1999) 421e428. [19] A. Gnasso, C. Irace, C. Carallo, M.S. De Franceschi, C. Motti, P.L. Mattioli, A. Pujia, In vivo association between low wall shear stress and plaque in subjects with asymmetrical carotid atherosclerosis, Stroke 28 (1997) 993e998. [20] C. Irace, F. Scavelli, C. Carallo, R. Serra, C. Cortese, A. Gnasso, Body mass index, metabolic syndrome and carotid atherosclerosis, Coron. Artery Dis. 20 (2009) 94e99. [21] N. Masawa, S. Glagov, C.K. Zarins, Quantitative morphologic study of intimal thickening at the human carotid bifurcation, II: the compensatory enlargement response and the role of the intima in tensile support, Atherosclerosis 107 (1994) 147e155. [22] D. Liao, D.K. Arnett, H.A. Tyroler, W.A. Riley, L.E. Chambless, M. Szklo, G. Heiss, Arterial stiffness and the development of hypertension. The ARIC study, Hypertension 34 (1999) 201e206. [23] J.E. Davies, J. Baksi, D.P. Francis, N. Hadjiloizou, Z.I. Whinnett, C.H. Manisty, J. Aguado-Sierra, R.A. Foale, I.S. Malik, J.V. Tyberg, K.H. Parker, J. Mayet, A.D. Hughes, The arterial reservoir pressure increases with aging and is the major determinant of the aortic augmentation index, Am. J. Physiol. Heart Circ. Physiol. 298 (2010) H580eH586. [24] F.P. Glor, B. Ariff, A.D. Hughes, P.R. Verdonck, D.C. Barratt, A.D. Augst, S.A. Thom, X.Y. Xu, Influence of head position on carotid hemodynamics in young adults, Am. J. Physiol. Heart Circ. Physiol. 287 (2004) H1670eH1681. [25] S.J. Zieman, V. Melenovsky, D.A. Kass, Mechanisms, pathophysiology, and therapy of arterial stiffness, Arterioscler. Thromb. Vasc. Biol. 25 (2005) 932e943. [26] M.J. Blaha, M.J. Budoff, J.J. Rivera, R. Katz, D.H. O’Leary, J.F. Polak, J. Takasu, R.S. Blumenthal, K. Nasir, Relationship of carotid distensibility and thoracic aorta calcification: multi-ethnic study of atherosclerosis, Hypertension 54 (2009) 1408e1415. [27] E.Y. Yang, L. Chambless, A.R. Sharrett, S.S. Virani, X. Liu, Z. Tang, E. Boerwinkle, C.M. Ballantyne, V. Nambi, Carotid arterial wall characteristics are associated with incident ischemic stroke but not coronary heart disease in the Atherosclerosis Risk in Communities (ARIC) study, Stroke 43 (2012) 103e108. [28] F.U. Mattace-Raso, T.J. van der Cammen, A. Hofman, N.M. van Popele, M.L. Bos, M.A. Schalekamp, R. Asmar, R.S. Reneman, A.P. Hoeks, M.M. Breteler, J.C. Witteman, Arterial stiffness and risk of coronary heart disease and stroke: the Rotterdam Study, Circulation 113 (2006) 657e663. [29] C. Carallo, C. Irace, M.S. De Franceschi, F. Coppoletta, R. Tiriolo, C. Scicchitano, F. Scavelli, A. Gnasso, The effect of aging on blood and plasma viscosity. An 11.6 years follow-up study, Clin. Hemorheol. Microcirc. 47 (2011) 67e74. [30] Y. Richter, E.R. Edelman, Cardiology is flow, Circulation 113 (2006) 2679e2682. [31] J. Levenson, A. Simon, Opposite effects of hypertension and smoking on large artery wall shear conditions, J. Hypertens. Suppl. 6 (1988) S176eS178. [32] J.P. Mynard, B.A. Wasserman, D.A. Steinman, Errors in the estimation of wall shear stress by maximum Doppler velocity, Atherosclerosis 227 (2013) 259e266. [33] D.J. Patel, J.S. Janicki, T.E. Carew, Static anisotropic elastic properties of the aorta in living dogs, Circ. Res. 25 (1969) 765e779.