SEARCH

SEARCH BY CITATION

Keywords:

  • atherosclerotic plaque;
  • C-reactive protein;
  • cytokines;
  • macrophage

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of interest statement
  9. References

Background

It is well established that subjects with moderately elevated plasma levels of C-reactive protein (CRP) have an increased risk of development of cardiovascular events. As atherosclerosis is a disease characterized by chronic arterial inflammation, it is possible that moderate increases in CRP level reflect the presence of plaque inflammation. To investigate this possibility, we compared plasma levels of hsCRP the day before carotid endarterectomy with the degree of inflammation in the excised plaque tissue.

Methods

Luminex immunoassays were used to determine the levels of IL-6, IL-10, monocyte chemoattractant protein-1 and tumour necrosis factor-α (TNF-α) in plasma and in homogenized plaque tissue from 160 endarterectomy specimens. Plaque sections were stained with antibodies against CD68 to determine the plaque macrophage content.

Results

Plasma high-sensitivity (hs)CRP levels were significantly correlated with plasma IL-6 and TNF-α. However, there were no significant associations between plasma hsCRP concentration and plaque cytokine levels or macrophage contents.

Conclusions

The present findings strongly argue against hsCRP as a marker of plaque inflammation. Hence, it is more likely that elevated hsCRP is a sign of a subclinical systemic inflammation and this in turn may contribute to progression of cardiovascular disease.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of interest statement
  9. References

C-reactive protein (CRP) is a highly conserved acute-phase protein that is primarily synthesized by liver hepatocytes in response to IL-6 [1]. It has an important role in innate immunity through its ability to recognize cell surface components on invading pathogens thereby activating the complement system. The development of high-sensitivity (hs) assays to detect CRP levels below those observed in acute infections has made it possible to study the association between low-grade inflammation and development of cardiovascular disease. This was an important development in view of the key role of inflammation in atherosclerosis and it seemed reasonable to assume that a chronic inflammatory process affecting parts of the arterial tree would manifest itself in elevated levels of circulating inflammatory markers. In 1997, Ridker et al. [2] first reported that baseline levels of hsCRP in the Physician's Health Study were higher in individuals who subsequently developed myocardial infarction and stroke. The association between moderately increased levels of hsCRP and increased cardiovascular disease risk has since been confirmed in a large number of epidemiological studies [3-6]. The question of whether hsCRP is an independent predictor of cardiovascular events has been much debated. To address this issue, the Emerging Risk Factor Collaboration reviewed data from 52 prospective trials involving 246 661 participants and concluded that assessing hsCRP in addition to other conventional risk factors in those at intermediate risk could help prevent one additional event over 10 years for every 400–500 individuals screened [7].

Whether CRP plays an active role in the disease process or is only a marker of risk has also been the focus of considerable debate. The findings of several in vitro studies have suggested that CRP has pro-inflammatory properties including stimulation of endothelial adhesion molecules and chemokine expression [8-10]. However, studies in experimental animals have provided conflicting results regarding the role of CRP in atherosclerosis [11-13]. The notion that CRP itself does not have a direct role in cardiovascular disease is supported by human genetic studies demonstrating that variations in the CRP gene associated with lifelong elevation of CRP levels do not confer increased risk of cardiovascular disease [14]. It is also possible that hsCRP may predict cardiovascular disease risk because it reflects the degree of atherosclerotic burden and/or disease activity. In this scenario, IL-6 and possibly other cytokines produced by inflammatory cells in atherosclerotic plaques would diffuse into the circulation and activate CRP expression in the liver. In accordance with this notion, several (but not all) studies have demonstrated an independent association between hsCRP and common carotid artery intima–media thicknesses (IMT) [15]. Plasma hsCRP has also been shown to correlate with the presence of both calcified and noncalcified coronary atherosclerotic plaques as quantified by computed tomography angiography [16].

In the present study, we directly addressed the question of whether hsCRP is a marker of atherosclerotic plaque inflammation by comparing hsCRP levels in plasma collected on the day before carotid endarterectomy with the degree of inflammation in the excised plaque tissue.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of interest statement
  9. References

Patients

Patients with ipsilateral carotid artery occlusion or restenosis after previous carotid surgery were excluded from this study. A total of 160 human carotid plaques were collected at carotid endarterectomy. The indications for surgery were plaques associated with ipsilateral symptoms [transient ischaemic attack (= 33), stroke (= 40) or amaurosis fugax (= 14)] and stenosis as measured by duplex sonography of >70%, or plaques not associated with symptoms and stenosis of >80% (= 73). All patients were evaluated preoperatively by a neurologist. One day before surgery, a venous blood sample was collected (EDTA-plasma) from each patient and stored for analysis. White blood cell (WBC) counts and plasma levels of hsCRP, triglycerides, HDL and LDL cholesterol and creatinine were determined by routine laboratory methods. The presence of diabetes and hypertension was recorded based on medical records. All patients gave written informed consent prior to the study. The study was approved by the local ethics committee.

Sample preparation

Plaques were removed by endarterectomy and immediately snap-frozen in liquid nitrogen. Portions (1 mm thick) from the most stenotic region of the frozen plaques were removed for histology. Plaques were weighed, cut into pieces whilst still frozen and homogenized in 5 mL of a homogenization buffer consisting of 50 mmol L−1 Tris-HCl (pH 7.5), 0.25 mol L−1 sucrose, 2 mmol L−1 tris (2-carboxyethyl) phosphine HCl, 50 mmol L−1 NaF, 1 mmol L−1 sodium orthovanadate, 10 mmol L−1 sodium glycerophosphate, 5 mmol L−1 sodium pyrophosphate, Complete Protease Inhibitor Cocktail (EDTA-free; Roche Diagnostics, Indianapolis, IN, USA), 1 mmol L−1 benzamidine and 10 mmol L−1 phenylmethylsulfonyl fluoride.

Histology

Cryosections (8 μm thick) were thawed, fixed with Histochoice (Amresco, Solon OH, USA), dipped in 60% isopropanol and then in 0.4% Oil Red O in isopropanol (60%; for 20 min) to stain lipids. The primary antibody used for macrophage assessment was monoclonal mouse anti-human CD68, clone KP1 (DakoCytomation, Glostrup, Denmark), diluted 1 : 100 in 10% rabbit serum, and the secondary antibody was polyclonal rabbit anti-mouse, rabbit F(ab′)2 (DakoCytomation), diluted 1 : 200 in 10% rabbit serum. The primary antibody used for staining vascular smooth muscle cells (α-actin) was monoclonal mouse anti-human smooth muscle actin clone 1A4 (DakoCytomation) diluted 1 : 50 in 10% rabbit serum, and the secondary antibody was biotinylated rabbit anti-mouse Ig (DakoCytomation) diluted 1 : 200 in 10% rabbit serum. Measurements of the area of plaque (% area) for the different components were quantified blindly using Biopix iQ 2.1.8 image analysis software (Gothenburg, Sweden) after scanning with ScanScope Console version 8.2 (LRI imaging AB, Vista, CA, USA).

Cytokine measurements

Aliquots of plaque homogenate (50 μL) were centrifuged at 13 000 g for 10 min. Next, 25 μL of the supernatant was removed and used for measuring the concentrations of interleukin (IL)-6, IL-10 monocyte chemoattractant protein (MCP-1) and tumour necrosis factor-α (TNF-α). The procedure was performed according to the manufacturer's instructions (Human Cytokine/Chemokine immunoassay, Millipore Corporation, MA, USA) and analysed with Luminex 100 IS 2.3 (Austin, TX, USA). Plaque cytokine levels were normalized against the wet weight of the plaque. Parallel measurements of the same cytokines were performed in plasma from the same patient before plaque donation.

Statistical analysis

Non-normally distributed variables were presented as median and range and normally distributed variables as mean and standard deviation (SD). Two-group comparisons were performed with Mann–Whitney test or Student's t-test (two-tailed), as appropriate. Spearman's rho was used for correlation analysis. Differences were considered statistically significant at < 0.05. ibm spss 21.0 (IBM Software, Armonk, NY, USA) was used for all statistical analyses.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of interest statement
  9. References

A total of 160 atherosclerotic plaques collected at carotid endarterectomy were included in the study. The clinical characteristics of the study group are shown in Table 1. Eighty-seven of the plaques were obtained from patients who had suffered a clinical event. Mean time between the clinical event and surgery was 15.2 ± 8.4 days. These are referred to as symptomatic plaques. The remaining 73 plaques were obtained from patients with more than 80% stenosis, but without a prior cerebrovascular event (referred to as asymptomatic plaques). Because the levels of inflammatory markers in plasma may have been influenced by the prior cerebrovascular event, we also analysed symptomatic and asymptomatic plaques separately. Patients with symptomatic plaques were older and were more likely to be diabetic; they were also less likely to be smokers, and treatment with beta-blockers was less common in this group. The median plasma hsCRP level in the study group was 4.0 mg L−1, and there was no significant difference in level between patients with symptomatic and asymptomatic plaques (Table 1). Plasma hsCRP levels were lower in men and in patients treated with statins (Table 2). In patients with diabetes (= 57), there was a significant association between HbA1c and plasma hsCRP levels (= 0.27, < 0.05).

Table 1. Baseline clinical characteristics of the study cohort
 All plaques (= 160)Asymptomatic plaques (= 73)Symptomatic plaques (= 87)
  1. Unless otherwise stated, all values are presented as mean ± standard deviation except triglycerides and high-sensitivity C-reactive protein (hsCRP) that are presented as median (range). WBC, white blood cell; BMI, body mass index. *< 0.05 and ***< 0.005; difference between asymptomatic and symptomatic plaques.

Age (years)69.2 ± 9.566.6 ± 6.771.3 ± 9.0***
Gender (% male)65.669.962.1
BMI (kg m−2)26.7 ± 3.826.8 ± 4.026.6 ± 3.7
Current smoker (%)32.541.125.3*
Diabetes (%)39.830.146.0*
Hypertension (%)75.679.572.4
Medication (%)
Statins88.891.886.2
Beta-blockers41.253.431.0***
Laboratory parameters
Triglycerides (mmol L−1)1.3 (0.4–5.0)1.3 (0.4–5.0) 1.2 (0.4–3.9)
HDL (mmol L−1)1.1 ± 0.41.2 ± 0.41.2 ± 0.5
LDL (mmol L−1)2.6 ± 1.12.5 ± 1.02.7 ± 1.2
WBC count (106 mL−1)8.0 ± 2.08.1 ± 1.98.0 ± 2.2
hsCRP (mg L−1)4.0 (0.2–86.7)3.8 (0.2–32.9)4.2 (0.5–86.7)
Table 2. Plasma hsCRP levels in relation to cardiovascular disease risk factors and therapies
 YesNo P
  1. hsCRP levels are expressed as mg L−1. ns, nonsignificant.

Male gender3.5 (0.2–86.7)4.7 (0.3–43.2)<0.05
Current smoker4.6 (0.5–48.3)3.8 (0.3–32.8)ns
Former smoker3.9 (0.2–86.7)3.8 (0.3–32.8)ns
Diabetes4.0 (0.2–48.3)3.9 (0.3–86.7)ns
Hypertension3.9 (0.3–86.7)4.0 (0.2–48.3)ns
Medication
Statins3.7 (0.2–86.7)5.0 (1.2–48.3)<0.05
Beta-blockers3.5 (0.3–43.2)4.2 (0.2–86.7)ns

We used two different strategies to assess plaque inflammation. First, immunohistochemical staining of sections from the most stenotic region of the plaque was used to assess the contents of macrophages (CD68), smooth muscle cells (α-actin) and lipids (Oil Red O). Secondly, the levels of the inflammatory cytokines IL-6, IL-10, MCP-1 and TNF-α were assessed in the plaque homogenate, which represents almost the entire plaque. The same biomarkers were also measured in plasma. Significant associations were found between plasma hsCRP and the WBC count as well as with plasma levels of IL-6 and TNF-α (Table 3). There were no significant associations between plasma hsCRP and plaque cytokines, macrophages, lipids or smooth muscle cells (Table 3), irrespective of whether all plaques were analysed together or asymptomatic and symptomatic plagues were analysed separately. Assuming that hsCRP levels in the circulation would increase in response to cytokines diffusing out of atherosclerotic lesions, it would be important not only to assess the relative degree of inflammation in the plaque, but also the absolute size of the plaque. As carotid endarterectomy involves removal of the entire plaque, measuring the weight of the excised tissue will provide a good estimate of the plaque size. There was no significant association between the plaque weight and plasma hsCRP level. Moreover, using the plaque weight to calculate the total plaque content of IL-6, IL-10, MCP-1 and TNF-α, we again found no significant associations with plasma hsCRP.

Table 3. Associations between hsCRP and other inflammatory markers in plasma and carotid plaques
 All plaques (= 160)Asymptomatic plaques (= 73)Symptomatic plaques (= 87)
  1. WBC, white blood cell; ns, nonsignificant. = 0.05, *< 0.05, **< 0.01 and ***< 0.005. Plaque staining for smooth muscle cells (α-actin), macrophages (CD68) and lipids (Oil Red O) were measured as percentage of total plaque area.

Plasma lipids and WBCs
Triglycerides (mmol L−1)nsns ns
HDL (mmol L−1)nsnsns
LDL (mmol L−1)nsnsns
WBC count (106 mL−1)0.25***ns0.30**
Plasma cytokines
IL-6 (pg mL−1)0.22**ns 0.25*
IL-10 (pg mL−1)nsnsns
TNF-α (pg mL−1)0.17*nsns
MCP-1 (pg mL−1)nsnsns
Plaque cytokines
IL-6 (pg g−1 plaque tissue)nsns ns
IL-10 (pg g−1 plaque tissue)nsnsns
TNF-α (pg g−1 plaque tissue)nsnsns
MCP-1 (pg g−1 plaque tissue)nsnsns
Plaque staining
α-actinnsnsns
CD68nsnsns
Oil Red Onsnsns

In order to determine whether similar information is provided by staining of sections from the area of maximum stenosis in the lesion and by cytokine analysis of plaque homogenates, we analysed the correlation between observations made using the two different approaches. Significant associations were found between CD68 macrophage staining and the plaque content of MCP-1 and IL-10 (= 0.21; = 0.01 and = 0.28; < 0.001, respectively) as well as between Oil Red O lipid staining and the plaque content of MCP-1 and IL-6 (= 0.30; < 0.001 and = 0.28; < 0.001, respectively). There was also an inverse association between smooth muscle cell α-actin staining and the plaque content of MCP-1 (= −21, < 0.01).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of interest statement
  9. References

A moderately increased plasma level of hsCRP is well established as a marker of cardiovascular disease risk [2-6]. However, the mechanisms responsible for the association between increased hsCRP and development of myocardial infarction and stroke remain to be fully elucidated. In view of the inflammatory nature of atherosclerosis, the main cause of ischaemic cardiovascular events, it has been reasonable to assume that elevated hsCRP reflects ongoing inflammation in atherosclerotic plaques. Accordingly, association between hsCRP and carotid artery IMT has been demonstrated in many but not all studies [15]. This notion is also in line with observations of associations between the plasma levels of hsCRP and cytokines that are produced by plaque macrophages, such as TNF-α, which are known to contribute to atherogenesis [17, 18]. It would thus be possible that IL-6 produced by macrophages in an atherosclerotic artery could enter the circulation and activate synthesis of CRP in the liver. However, we have demonstrated in the present study that although plasma hsCRP significantly correlates with the plasma levels of both IL-6 and TNF-α,, it does not correlate with the inflammatory activity in carotid atherosclerotic plaques. These findings provide strong evidence that hsCRP is not a marker of atherosclerotic disease activity.

There are several limitations of the present study that need to be considered. Compared with the epidemiological studies of association between hsCRP and cardiovascular disease risk, the present study is very small and could be of insufficient power to detect a correlation between plaque inflammation and hsCRP. However, the sample size was sufficient to detect significant associations with r values as low as almost 0.15, suggesting that significant associations that could have been identified in a larger study are unlikely to be of major biological importance. Most epidemiological studies have been conducted in populations with a low or moderate risk of cardiovascular disease, whereas the present study was performed in subjects with established disease and a high risk of future cardiovascular events. Accordingly, the present study cohort is not representative of the large majority of subjects in population-based epidemiological studies. At the same time, it is reasonable to assume that if hsCRP is a marker of plaque inflammation, then it would be easier to identify such an association in a group with more advanced disease.

Another important limitation of studies performed on plaques obtained at the time of surgery is that inflammatory markers in both plasma and the plaque could remain affected by a previous cerebrovascular event if such an event occurred only a couple of weeks before surgery. In our study, plasma levels of hsCRP tended to be higher in the group with a previous clinical event, but this difference did not reach statistical significance. Moreover, there was still no association between plasma hsCRP and plaque inflammation when plaques from patients without a previous clinical event were analysed separately. The technical challenges in assessing plaque inflammation should also not be underestimated. Changes in the plaque cytokine content make take place after the lesion has been surgically removed, but to minimize this risk, the plaques were snap-frozen in liquid nitrogen in the operating theatre. Additionally, we used two independent methods to determine plaque inflammation: immunohistochemical analysis of plaque macrophages and measurement of cytokines in extracts from homogenized plaques.

To standardize the macrophage staining, we used sections taken from the level of maximum stenosis of each lesion. However, several studies have shown that plaques in general rupture upstream to the area of maximum stenosis and that inflammation is more prominent at such sites [19-21]. Therefore, the sections analysed in the present study may not be representative of the most inflamed region of the plaque. Analysis of cytokine levels in plaque homogenates has the advantage of assessing inflammation in almost the entire plaque whilst the immunohistochemical analysis provides information from a limited section of the lesion. It should also be noted that although several of the plaque cytokines correlated with increased staining for macrophages and lipids and inversely with increased staining for smooth muscle cells, these associations were not strong. One possible explanation for these weak associations could be that sections taken from the area of maximum stenosis are not representative of the most inflamed parts of the plaque as discussed above. It is also possible that only measuring the absolute levels of cytokines in plaques provides incomplete information about the inflammatory activity as this may be influenced by the expression of the corresponding receptors, neutralizing soluble receptors and sequestration to connective tissue proteins [22]. The possibility that other factors in the homogenate may have influenced the analysis also cannot be excluded. Finally, it could be argued that the inflammatory activity in a single lesion may not be representative of plaques in the rest of the arterial system. Although it may not be feasible to fully investigate this possibility, it is well established that the severity of carotid disease assessed by carotid IMT is a predictor of risk of future coronary events [23].

Despite the limitations discussed above, our findings do not support the notion that hsCRP is a marker of plaque inflammation. Instead, elevated hsCRP is more likely to represent subclinical systemic inflammation that may subsequently contribute to the progression of cardiovascular disease.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of interest statement
  9. References

This study was supported by grants from the Swedish Research Council, the Marianne and Marcus Wallenberg Foundation, the Swedish Heart and Lung Foundation, the Swedish Medical Society, Regional Research Funds (Region Skåne), Malmö University Hospital Funds, the Ernhold Lundström Foundation and the Swedish Foundation for Strategic Research. We would like to thank Lena Sundius for technical assistance.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. Conflict of interest statement
  9. References
  • 1
    Ansar W, Ghosh S. C-reactive protein and the biology of disease. Immunol Res 2013; 56: 13142.
  • 2
    Ridker P, Cushman M, Stampfer M, Tracy R, Hennekens C. Inflammation, asprin, and the risk of cardiovascular disease in apparently healthy men. N Engl J Med 1997; 336: 9739.
  • 3
    Ridker PM, Rifai N, Pfeffer MA et al. Inflammation, pravastatin, and the risk of coronary events after myocardial infarction in patients with average cholesterol levels. Cholesterol and Recurrent Events (CARE) Investigators. Circulation 1998; 98: 83944.
  • 4
    Ridker PM, Hennekens CH, Buring JE, Rifai N. C-reactive protein and other markers of inflammation in the prediction of cardiovascular disease in women. N Engl J Med 2000; 342: 83643.
  • 5
    Koenig W, Sund M, Frohlich M et al. C-Reactive protein, a sensitive marker of inflammation, predicts future risk of coronary heart disease in initially healthy middle-aged men: results from the MONICA (Monitoring Trends and Determinants in Cardiovascular Disease) Augsburg Cohort Study, 1984 to 1992. Circulation 1999; 99: 23742.
  • 6
    Ballantyne CM, Hoogeveen RC, Bang H et al. Lipoprotein-associated phospholipase A2, high-sensitivity C-reactive protein, and risk for incident coronary heart disease in middle-aged men and women in the Atherosclerosis Risk in Communities (ARIC) study. Circulation 2004; 109: 83742.
  • 7
    Kaptoge S, Di Angelantonio E, Pennells L et al. C-reactive protein, fibrinogen, and cardiovascular disease prediction. N Engl J Med 2012; 367: 131020.
  • 8
    Pasceri V, Willerson JT, Yeh ET. Direct proinflammatory effect of C-reactive protein on human endothelial cells. Circulation 2000; 102: 21658.
  • 9
    Yaron G, Brill A, Dashevsky O et al. C-reactive protein promotes platelet adhesion to endothelial cells: a potential pathway in atherothrombosis. Br J Haematol 2006; 134: 42631.
  • 10
    Grad E, Danenberg HD. C-reactive protein and atherothrombosis: Cause or effect? Blood Rev 2013; 27: 239.
  • 11
    Tennent GA, Hutchinson WL, Kahan MC et al. Transgenic human CRP is not pro-atherogenic, pro-atherothrombotic or pro-inflammatory in apoE−/− mice. Atherosclerosis 2008; 196: 24855.
  • 12
    Kovacs A, Tornvall P, Nilsson R, Tegner J, Hamsten A, Bjorkegren J. Human C-reactive protein slows atherosclerosis development in a mouse model with human-like hypercholesterolemia. Proc Natl Acad Sci USA 2007; 104: 1376873.
  • 13
    Paul A, Ko KW, Li L et al. C-reactive protein accelerates the progression of atherosclerosis in apolipoprotein E-deficient mice. Circulation 2004; 109: 64755.
  • 14
    Nordestgaard BG. Does elevated C-reactive protein cause human atherothrombosis? Novel insights from genetics, intervention trials, and elsewhere. Curr Opin Lipidol 2009; 20: 393401.
  • 15
    Baldassarre D, De Jong A, Amato M et al. Carotid intima-media thickness and markers of inflammation, endothelial damage and hemostasis. Ann Med 2008; 40: 2144.
  • 16
    Bamberg F, Truong QA, Koenig W et al. Differential associations between blood biomarkers of inflammation, oxidation, and lipid metabolism with varying forms of coronary atherosclerotic plaque as quantified by coronary CT angiography. Int J Cardiovasc Imaging 2012; 28: 18392.
  • 17
    Branen L, Hovgaard L, Nitulescu M, Bengtsson E, Nilsson J, Jovinge S. Inhibition of tumor necrosis factor-alpha reduces atherosclerosis in apolipoprotein E knockout mice. Arterioscler Thromb Vasc Biol 2004; 24: 213742.
  • 18
    Ohta H, Wada H, Niwa T et al. Disruption of tumor necrosis factor-alpha gene diminishes the development of atherosclerosis in ApoE-deficient mice. Atherosclerosis 2005; 180: 117.
  • 19
    Dirksen MT, van der Wal AC, van den Berg FM, van der Loos CM, Becker AE. Distribution of inflammatory cells in atherosclerotic plaques relates to the direction of flow. Circulation 1998; 98: 20003.
  • 20
    Yilmaz A, Lipfert B, Cicha I et al. Accumulation of immune cells and high expression of chemokines/chemokine receptors in the upstream shoulder of atherosclerotic carotid plaques. Exp Mol Pathol 2007; 82: 24555.
  • 21
    Fagerberg B, Ryndel M, Kjelldahl J et al. Differences in lesion severity and cellular composition between in vivo assessed upstream and downstream sides of human symptomatic carotid atherosclerotic plaques. J Vasc Res 2010; 47: 22130.
  • 22
    Tedgui A, Mallat Z. Cytokines in atherosclerosis: pathogenic and regulatory pathways. Physiol Rev 2006; 86: 51581.
  • 23
    Lorenz MW, Markus HS, Bots ML, Rosvall M, Sitzer M. Prediction of clinical cardiovascular events with carotid intima-media thickness: a systematic review and meta-analysis. Circulation 2007; 115: 45967.