Independent association between inflammatory markers (C-reactive protein, interleukin-6, and TNF-a) and essential hypertension

Transcription

1 (2005) 19, & 2005 Nature Publishing Group All rights reserved /05 $ ORIGINAL ARTICLE Independent association between inflammatory markers (C-reactive protein, interleukin-6, and TNF-a) and essential hypertension LE Bautista 1,2, LM Vera 3, IA Arenas 4 and G Gamarra 3 1 Department of Preventive Medicine and Biometrics, Uniformed Services University of the Health Sciences, Bethesda, MD, USA; 2 Department of Population Health Sciences, University of Wisconsin, Madison, WI, USA; 3 Centro de Investigación Epidemiológica, Universidad Industrial de Santander, Bucaramanga, Colombia; 4 Department of Physiology, University of Alberta, Edmonton, Alberta, Canada High blood pressure (HBP) has been associated with elevated C-reactive protein (CRP), a marker of chronic mild inflammation. However, the association between HBP and other inflammatory markers, particularly interleukin 6 (IL-6) and tumour necrosis alpha (TNF-a), has not been evaluated in well-controlled studies. We examined the cross-sectional relationship between IL- 6, TNF-a, and CRP and HBP in a random sample of 196 healthy subjects. All markers were measured in duplicate with high-sensitivity ELISA tests. Three blood pressure (BP) measurments were averaged for the analysis, and subjects with systolic BP X140 and/or diastolic BP X90 mmhg were considered hypertensive. Log binomial regression was used to estimate multivariate-adjusted prevalence ratios (PR) of HBP. Of the subjects, 40% (79) were hypertensive (mean age: 44 years; range 30 64). After adjustment for age, sex, body mass index, family history of HBP, and the level of the other inflammatory markers, subjects in the second (PR: 3.10, P ¼ 0.003), third (PR: 2.32; P ¼ 0.031), and fourth quartiles (PR: 2.30; P ¼ 0.036) of IL-6 were more than twice as likely to be hypertensive than those in the first quartile. Corresponding PR estimates for TNF-a levels were 1.41 (P ¼ 0.014) for the second; 1.59 (P ¼ 0.001) for the third; and 1.61 (P ¼ 0.025) for the fourth quartile. The CRP HBP association was not statistically significant. Our results suggest that TNF-a and IL-6 could be independent risk factors for HBP in apparently healthy subjects. Nevertheless, the temporal relationship between elevated inflammation markers and HBP should be ascertained in prospective cohort studies. (2005) 19, doi: /sj.jhh Published online 9 September 2004 Keywords: blood pressure; C-reactive protein; TNF-a; interleukin-6; risk factors Introduction High blood pressure (HBP) is one of the most important risk factors for cardiovascular renal disease. In spite of its high health impact, primary prevention of HBP is partly hampered because of a limited knowledge of HBP risk factors. Several prospective cohort studies have shown that acute systemic inflammation is associated with an increased risk of acute cardiovascular events 1 3 and cardiovascular mortality. 4,5 Recently published data also suggest that chronic inflammation could be an Correspondence: Dr LE Bautista, University of Wisconsin Medical School, Population Health Sciences, 610 Walnut Street, 703 WARF, Madison, WI , USA. lebautista@wisc.edu Received 16 April 2004; revised and accepted 1 August 2004; published online 9 September 2004 independent risk factor for HBP. C-reactive protein has been shown to be associated to HBP in a few well-controlled studies. 6 8 On the other hand, results from studies on the association between interleukin-6 (IL-6) and HBP have been contradictory. 9,10 Similarly, some studies in human subjects have shown a positive association between tumour necrosis factor alpha (TNF-a) level and HBP, 11,12 but others have not. 13,14 However, most of these studies were not designed to test the association between inflammation markers and HBP and have failed to control for other risk factors for HBP and for other inflammatory markers. Since C- reactive protein (CRP) production by hepatocytes is regulated by cytokines, mostly IL-6 and TNF-a, 15 it is unlikely that any one inflammatory marker will fully reflect the complexity of mild chronic inflammation and capture its relationship with HBP. Moreover, the effect of one cytokine on HBP may

2 150 be confounded or interact with the effects of the other cytokines, a problem that has not been addressed in previous studies. In this article, we report the independent effects of CRP, TNF-a, and IL-6 on blood pressure (BP) in a random sample of free-living healthy subjects. Methods Inflammatory markers and hypertension We studied a group of subjects who participated in a previous cross-sectional study of cardiovascular risk factors conducted in a random sample of free living, apparently healthy subjects from Bucaramanga, Colombia. 16 Participants X30 and o65 years old were randomly selected from an original sampling frame of 2432 participants. The study was approved by the IRB of the Universidad Industrial de Santander and written informed consent was obtained from all subjects. All subjects completed an interview aimed to ascertain cardiovascular risk factors, history of coronary heart disease, and medication used. Anthropometric measurements were carried out by duplicate and the mean of the two measurements was used during the analysis. BP was measured in the seated position, in the right arm, after a 5-min rest period, following the recommendations of the American Heart Association. 17 The average of three consecutive BP measurements, taken by two independent observers, was used as the BP value in this report. Hypertension was defined as a systolic BP X140 mmhg and/or a diastolic BP X90 mmhg, or current treatment with antihypertensive medication. Stored blood samples were tested for levels of CRP, IL-6, and TNF-a. None of the participants had any febrile illness or major trauma during the 15 days before enrollment. High-sensitivity ELISA tests for CRP, IL-6, and TNF-a were conducted in duplicate using the IMMULITE Automated Analyzer and kits (Diagnostic Products Corporation, Los Angeles, CA, USA). The CRP, IL-6 and TNF-a tests have limits of detection of 0.01 mg/dl, 5, and 1.7 pg/ml, respectively. To account for their positive skewness, values of CRP, TNF-a, and IL-6 were transformed and analysed in the natural log scale. We used censored-normal regression to estimate the mean and standard deviation of TNF-a and IL-6, while taking in consideration the presence of values below the limit of detection of the tests. Log normality of nontruncated values was graphically confirmed. Exposure groups were defined using quartiles of the observed distribution of each inflammatory marker in the group of nonhypertensive subjects. The groups so-defined were compared regarding the prevalence of HBP. We used log binomial regression 18 to calculate prevalence ratios of HBP, taking the lowest quartile of each marker as the reference group. Log binomial regression allows the estimation of the prevalence ratio instead of the prevalence odds ratio, usually obtained from logistic regression models. Contrary to the odds ratio, the prevalence ratio (PR) does not overestimate the effect of the exposure for conditions with a high prevalence, such as HBP. Age, sex, body mass index, plasma cholesterol and triglyceride levels, smoking, history of hypertension in a parent or sibling, diabetes, and use of drugs that could have affected plasma levels of inflammatory makers in the last 2 weeks were evaluated as potential confounders of the association between HBP and each inflammatory marker. Only those variables significantly associated to HBP and those that qualified as confounders were retained in the final models. 19 The overall fitness of the final regression model was evaluated using Tsiatis test. 20 Results A total of 79 out of 205 hypertensive and 117 out of 2227 nonhypertensive subjects were randomly selected and included in the study. The average age was 44 years (range: 30 64) and 37% were male. As expected, hypertensives were significantly older, had significantly higher body mass index, and were more likely to have a close relative with diagnosed hypertension than nonhypertensives (Table 1). Plasma glycaemia, cholesterol, and triglycerides were also considerably higher among hypertensives. IL-6 plasma levels were below the limit of detection of the test in 29 (15.4%) of the subjects. Similarly, 88 subjects (46.6%) had TNF-a levels below the limit of detection. Plasma levels of IL-6 and TNF-a were two to four times higher in subjects with HBP (P-values: and 0.015; Table 1). However, both groups had similar levels of plasma CRP (P ¼ 0.564). A small group of 23 subjects had received drugs that could have affected the concentration of plasma cytokines (anti-inflammatory drugs, statins, and glitazones). Levels of CRP, TNFa, and IL-6 were not significantly different among these subjects (P ¼ 0.80, 0.17, and 0.71, respectively). HBP prevalence increased progressively from 30.4% in the first (lowest) to 52.8% in the fourth (highest) quartile of TNF-a. This corresponded to a PR of 1.73 (95% CI: 1.10, 5.50; Table 2). A significant increase in HBP prevalence was also observed in subjects in the second to fourth quartiles of IL-6, as compared to those in the first quartile. Subjects in the second quartile of IL-6 were 2.95 times more likely to be hypertensive than those in the first quartile (P ¼ 0.006). Subjects in the third (PR: 2.64; P ¼ 0.017) and fourth quartiles of IL-6 (PR: 2.47; P ¼ 0.027) also had a significantly higher prevalence of HBP than those in the first quartile. Nevertheless, the relative effect of IL-6 on HBP prevalence decreased beyond the second quartile. A similar pattern was observed for quartiles of CRP, with a borderline significant PR for the second quartile (PR:

3 Inflammatory markers and hypertension Table 1 Mean values of cardiovascular risk factors by hypertension status 151 Risk factor Hypertensive (n ¼ 79) mean (95% CI) a Nonhypertensive All (n ¼ 196) Mean P-value* (n ¼ 117) Mean (95% CI) a (95% CI) a Age (years) 46.6 (44.4, 48.9) 41.8 (40.3, 43.3) 43.7 (42.4, 45.0) o0.001 Male (%) 36.7 (26.1, 48.3) 36.8 (28.0, 46.2) 36.7 (30.0, 43.9) Systolic BP (137.2, 147.3) (106.6, 111.4) (119.0, 125.8) o0.001 Diastolic BP 87.1 (84.4, 89.8) 71.6 (70.1, 73.2) 77.9 (76.1, 79.6) o0.001 BMI b 27.8 (26.6, 28.9) 25.2 (24.4, 26.1) 26.3 (25.6, 27.0) o0.001 Glycaemia (mg/dl) 89.6 (85.8, 93.4) 85.8 (83.3, 88.2) 87.3 (85.2, 89.4) 0.09 Cholesterol c Total (218.9, 237.4) (205.4, 221.2) (213.3, 225.3) HDL 42.8 (40.8, 44.8) 41.5 (39.8, 43.1) 42.0 (40.7, 43.3) LDL (138.6, 156.7) (132.2, 146.5) (137.0, 148.2) Triglycerides c (176.1, 229.0) (149.6, 191.8) (167.0, 200.0) Smokers (%) 15.4 (9.4, 23.2) 16.5 (9.1, 26.5) 15.8 (11.0, 21.7) 0.84 Relative with HBP d (%) 72.2 (60.9, 81.6) 57.0 (47.4, 66.2) 63.2 (56.0, 70.0) Diabetes (%) 5.1 (1.4, 12.5) 2.5 (0.5, 7.3) 3.6 (1.4, 7.2) IL-6 (pg/ml) 10.0 (7.9, 12.6) 5.0 (3.5, 6.9) 6.8 (5.5, 8.4) TNF-a (pg/dl) 7.9 (4.2, 15.0) 1.8 (0.8, 4.0) 3.5 (2.1, 5.8) CRP e (mg/dl) 0.29 (0.23, 0.36) 0.26 (0.22, 0.32) 0.27 (0.24, 8.37) a 95% confidence interval. b Body mass index (kg/m 2 ). c mg/ml. d Parent or sibling with high blood pressure. e C-reactive protein. *Hypertensive vs non-hypertensive. Table 2 Prevalence and PR of HBP by quartiles of inflammatory marker Quartiles of inflammatory marker Prevalence cases (%) PR (95% CI) P-value IL-6 (pg/ml) p p , o12.4 p , , TNF-a (pg/dl) p p , p , , CRP (mg/dl) p p , p , , ; P ¼ 0.051) and nonsignificant and smaller increases in the third and fourth quartiles. Adjustment by age and sex changed little the pattern of associations observed in the crude analysis, with the exception of a higher PR for the second quartile of TNF-a (PR: 1.65; P ¼ 0.004; Table 2). In a multivariate analysis, age, sex, body mass index, and a history of HBP in a close relative were the only variables retained in the model as independent risk factors or as confounding variables. After adjustment for these variables, IL-6, and TNF-a plasma levels were significantly associated with increased prevalence of HBP. Subjects in the second to fourth quartiles of IL-6 were at least twice more likely to be hypertensive than those in the lowest quartile (Table 3). Similarly, participants with TNF-a level in the second to fourth quartile were roughly 1.5 times more likely to have HBP than those in the lowest quartile. On the contrary, there was no significant association between CRP level and HBP prevalence. In a final model we included all three inflammatory markers and the above-mentioned confounding factors. According to this model, the prevalence of HBP was 3.10 times higher in subjects in the second quartile of IL-6 (P ¼ 0.003), 2.32 times higher in those in the third quartile (P ¼ 0.031), and 2.30 times higher in those in the fourth quartile (P ¼ 0.036), as compared to those in the first quartile (Table 4). Similarly, HBP prevalence increased 1.41 times (P ¼ 0.014) in the second quartile of TNF-a, 1.59 times in the third quartile (P ¼ 0.001), and 1.61 times in the fourth quartile (Po0.001), as compared to the prevalence in the first quartile. On the other hand, PRs of HBP were nonsignificant in subjects in the third and fourth quartiles and only marginally significant in the second quartile of CRP (PR: 1.30; P ¼ 0.052), as compared to subjects in the first quartile. An evaluation of goodness-of-fit showed that our model was a good representation of the data, since expected case and control counts predicted by the model were not statistically different from observed counts (P ¼ 0.14). Finally, we added interaction terms to our model in order to test whether the effect of a maker (eg IL-6) on HBP prevalence depended on the level of the other markers (eg CRP). This analysis showed no

4 Inflammatory markers and hypertension 152 Table 3 PR of HBP by quartiles of inflammatory marker, adjusted by age and sex and by traditional cardiovascular risk factors Quartiles of inflammatory marker Age- and sex-adjusted Multivariate-adjusted PR a (95% CI) P-value PR b (95% CI) P-value IL-6 (pg/ml) p p (1.48, 6.49) (1.34, 5.72) p (1.19, 5.58) (1.22, 5.36) (1.14, 5.28) (1.01, 4.61) TNF-a (pg/dl) p p (1.18, 2.30) (1.02, 2.61) p (1.02, 2.45) (0.92, 2.17) (1.18, 2.30) (1.05, 2.01) CRP (mg/dl) p p (0.94, 2.70) (0.92, 1.61) p (0.65, 2.12) (0.46, 1.13) (0.71, 2.23) (0.62, 1.41) a Age- and sex-adjusted PR. b Adjusted by age, sex, body mass index and history of high blood pressure in parent or sibling. Table 4 PR of hypertension by quartiles of inflammatory marker adjusted by cardiovascular risk factors and levels of the other inflammatory markers Quartiles of inflammatory marker PR a (95% CI) P-value IL-6 (pg/ml) p p (1.47, 6.56) p (1.08, 4.99) (1.06, 4.99) TNF-a (pg/dl) p p (1.07, 1.86) p (1.21, 2.09) (1.38, 1.90) CRP (mg/dl) p p (1.00, 1.70) p (0.57, 1.27) (0.82, 1.62) a Adjusted for age, sex, body mass index, history of hypertension in parents or siblings, and levels of the other inflammatory markers. evidence of significant interaction between IL-6 and TNF-a (P ¼ 0.110), IL-6 and CRP (P ¼ 0.928), and TNF-a and CRP (P ¼ 0.769). Discussion This population-based study shows a significant association between IL-6 and TNF-a, two markers of chronic mild inflammation, and the presence of HBP among apparently healthy subjects. The prevalence of HBP increased significantly with increased level of IL-6. However, the rate of increase was higher in the lower end of the distribution of IL- 6 than in the upper end. This suggests that instead of a straight line, the HBP IL-6 relationship follows a curve that flattens out at higher values of IL-6. Nevertheless, even subjects in the upper quartile of IL-6 had a significant increase of two times the prevalence of HBP observed among those in the first quartile. Also, the prevalence of HBP increased progressively and significantly with higher levels of TNF-a, but the PR was very similar in the third and fourth quartiles of this marker. On the contrary, after adjustment for other risk factors and for IL-6 and TNF-a, CRP was not significantly associated with HBP. These results are important, as they suggest that the relationship between mild chronic inflammation and coronary heart disease 1,2 may be mediated not only through the development of the atherosclerotic plaque, but also by an increased risk of HBP. Previous studies have reported conflicting results about the relationship between IL-6 and HBP. 3,5,9,10,12,13,21,22 With one exception, 10 these studies were not designed to test the association between IL-6 and HBP, and failed to control for other cardiovascular risk factors. In fact, our results are consistent with those from the study of Chae et al, 10 the only other study in which appropriate control of confounding factors has been conducted. An association between IL-6 and BP level is biologically plausible. IL-6 stimulates the synthesis of many acute-phase reaction proteins, including CRP, 23 and counter-regulates TNF-a and interleukin- 1 b levels. 24 A common functional polymorphism of the IL-6 gene promoter region (-174 G4C) has been shown to influence levels of circulating IL-6 in vivo. 25 CC homozygous subjects seem to respond to inflammatory stimuli with a higher increase in IL Also, systolic and diastolic BP seem to be higher in healthy subjects 27 and dialysis patients 28 with the CC genotype as compared to carriers of the -174 G allele. However, Pola et al 29 failed to show and association between this polymorphism and essen-

5 Inflammatory markers and hypertension tial hypertension in a sample of 210 cases and 177 controls older than 65 years. There is also experimental evidence that IL-6 prolongs the endothelial dysfunction associated with the administration of IL-1b and TNF-a in human volunteers. 30 Therefore, IL-6 may result in elevated BP by inducing endothelial dysfunction and increasing peripheral vascular resistance. 31,32 Some reports of studies in human subjects have shown a significant association between TNF-a and elevated BP, 11,12,21 but others have failed to show such an association. 13,14 Unfortunately, none of these studies has been designed to test this hypothesis and all have failed to control for other risk factors. However, in experimental studies the degree of activation of the TNF-a system has been positively and significantly associated with systolic and diastolic BP. 33 Besides, TNF-a decreases endothelial nitric oxide synthase mrna level by shortening its half-life. 34 This may result in decreased bioavailability of nitric oxide and lead to endothelial dysfunction, chronic vasoconstriction, and elevated BP. Finally, a common polymorphism in the promoter region of the TNF-a gene has been associated to increased TNF-a and systolic BP. 35 Contrary to previous studies aimed to test this hypothesis, 6,7,36 we did not find a significant association between CRP and hypertensive status. In our analysis the effect of CRP may be absent due to the inclusion of TNF-a and IL-6 in the final model. However, CRP effects failed to be significant even when adjusting only for TNF-a, only for IL-6, or for none of them. A lack of an adequate sample size is a more likely explanation of this finding. Indeed, the confidence intervals for the PR in the third and fourth quartiles of CRP were considerably wide, reflecting lack of power. Common biases are unlikely explanations of our results. Since both cases and controls were randomly sampled from the same population, the likelihood of selection bias is very restricted. Information error was limited because the laboratory tests used to measure inflammatory markers are highly reliable, and were conducted in duplicate, without knowledge of other cardiovascular risk factors or BP status. All additional study measurements were also highly standardized and carried out by trained personnel. Under these conditions, any information error is likely to be nondifferential and would result in underestimation of the effect of the inflammatory markers. Finally, confounding factors, including the use of drugs capable of lowering plasma levels of inflammatory markers, were controlled for by means of multivariate analysis. A limitation of our study comes from the number of values of TNF-a and IL-6 values that were below the limit of detection for their corresponding tests. Although this is of no consequence for the analysis by quartiles of TNF-a and IL-6, it precludes their analysis as continuous exposures. A more relevant limitation of our study results from its crosssectional nature. Since inflammatory markers were measured at the same point in time as BP, it is not possible to know whether one precedes the other or vice versa. It is also possible that elevated TNF-a and IL-6 may both precede and follow HBP. 32 There is experimental evidence that endothelial cells exposed to the shear stress resulting from elevated BP increase their production of IL-6 and TNF-a. 37,38 Unfortunately, there are few if any human data on the effect of elevated BP on the synthesis of cytokines and CRP. We have found no changes in CRP levels in subjects with drug-controlled hypertension as compared to noncontrolled subjects, 32 which suggests that CRP levels are not a consequence of elevated BP. Since heritability seems to be an important independent determinant of the level of inflammatory markers in normal subjects, there is a real possibility that elevated inflammatory markers could precede the development of HBP, even if the inflammatory process is also a response to the elevation of the BP. Finally, high plasma levels of orosomucoid, a1-antitrypsin, haptoglobin, and ceruloplasmin, proteins that are components of the chronic inflammatory response, have been associated with future development of high BP. 39 In conclusion, our results suggest that TNF-a and IL-6 could be independent risk factors for the development of HBP in apparently healthy subjects. Nevertheless, the temporal relationship between elevated inflammation markers and elevated BP should be ascertained in prospective cohort studies. Acknowledgements This study was funded by COLCIENCIAS. References 1 Ridker PM et al. Inflammation, aspirin, and the risk of cardiovascular disease in apparently healthy men. N Engl J Med 1997; 336: Ridker PM et al. C-reactive protein and other markers of inflammation in the prediction of cardiovascular disease in women. N Engl J Med 2000; 342: Ridker PM et al. Plasma concentration of interleukin-6 and the risk of future myocardial infarction among apparently healthy men. Circulation 2000; 101: Harris TB et al. Associations of elevated interleukin-6 and C-reactive protein levels with mortality in the elderly. Am J Med 1999; 106: Volpato S et al. Cardiovascular disease, interleukin-6, and risk of mortality in older women: the women s health and aging study. Circulation 2001; 103: Bautista LE et al. Is C-reactive protein an independent risk factor for essential hypertension? J Hypertens 2001; 19: Chul SK et al. High sensitivity c-reactive protein as an independent risk factor for essential hypertension. Am J Hypertens 2003; 16:

6 154 Inflammatory markers and hypertension 8 Bautista LE et al. Association between C-reactive protein and hypertension in healthy middle-aged men and women. Coron Artery Dis 2004; 15 (in press). 9 Rifai N et al. Inflammatory markers in men with angiographically documented coronary heart disease. Clin Chem 1999; 45: Chae CU et al. Blood pressure and inflammation in apparently healthy men. Hypertension 2001; 38: Ito H et al. Association of serum tumour necrosis factor-alpha with serum low-density lipoprotein-cholesterol and blood pressure in apparently healthy Japanese women. Clin Exp Pharmacol Physiol 2001; 28: Furumoto T et al. Association of cardiovascular risk factors and endothelial dysfunction in japanese hypertensive patients: implications for early atherosclerosis. Hypertens Res 2002; 25: Mendall MA et al. Relation of serum cytokine concentrations to cardiovascular risk factors and coronary heart disease. Heart 1997; 78: Sheu WHH et al. Plasma tumor necrosis factor alpha levels and insulin sensitivity in hypertensive subjects. Clin Exp Hypertens 2000; 22: Baumann H, Gauldie J. The acute phase response. Immunol Today 1994; 15: Bautista L et al. Prevalencia de Factores de Riesgo para Enfermedades Crónicas No Transmisibles en la Población de Bucaramanga. Encuesta Basal del Programa CARMEN. UIS Bucaramanga. 17 Perloff D et al. Human blood pressure determination by sphygmomanometry. Circulation 1993; 88: Wacholder S. Binomial regression in GLIM: estimating risk ratios and risk differences. Am J Epidemiol 1986; 123: Greenland S. Modeling and variable selection in epidemiologic analysis. Am J Public Health 1989; 79: Tsiatis AA. A note on a goodness-of-fit test for the logistic regression model. Biometrika 1980; 67: Yudkin JS et al. C-reactive protein in healthy subjects: associations with obesity, insulin resistance, and endothelial dysfunction: a potential role for cytokines originating from adipose tissue? Arterioscler Thromb Vasc Biol 1999; 19: Fernandez-Real JM et al. Circulating interleukin 6 levels, blood pressure, and insulin sensitivity in apparently healthy men and women. J Clin Endocrinol Metab 2001; 86: Woods A et al. Genetics of inflammation and risk of coronary artery disease: the central role of interleukin- 6. Eur Heart J 2000; 21: Tilg H, Dinarello CA, Mier JW. IL-6 and APPs: antiinflammatory and immunosuppressive mediators. Immunol Today 1997; 18: Fishman D et al. The effect of novel polymorphisms in the interleukin-6 (IL-6) gene on IL-6 transcription and plasma IL-6 levels, and an association with systemiconset juvenile chronic arthritis. J Clin Invest 1998; 102: Brull DJ et al. The effect of the interleukin-6-174g4c promoter gene polymorphism on endothelial function in healthy volunteers. Eur J Clin Invest 2002; 32: Humphries SE et al. The interleukin G/C promoter polymorphism is associated with risk of coronary heart disease and systolic blood pressure in healthy men. Eur Heart J 2001; 22: Losito A et al. Association of interleukin-6-174g/c promoter polymorphism with hypertension and left ventricular hypertrophy in dialysis patients. Kidney Int 2003; 64: Pola R et al. The -174 G/C polymorphism of the interleukin-6 gene promoter and essential hypertension in an elderly Italian population. J Hum Hypertens 2002; 16: Bhagat K, Vallance P. Inflammatory cytokines impair endothelium-dependent dilatation in human veins in vivo. Circulation 1997; 96: Hingorani AD et al. Acute systemic inflammation impairs endothelium-dependent dilatation in humans. Circulation 2000; 102: Bautista LE. Inflammation, endothelial dysfunction, and the risk of high blood pressure: Epidemiologic and biological evidence. J Hum Hypertens 2003; 17: Fernandez-Real JM et al. Shedding of TNF-alpha receptors, blood pressure, and insulin sensitivity in type 2 diabetes mellitus. Am J Physiol Endocrinol Metab 2002; 282: Yoshizumi M et al. Tumor necrosis factor downregulates an endothelial nitric oxide synthase mrna by shortening its half-life. Circ Res 1993; 73: Dalziel B et al. Association of the TNF-alpha -308 G/A promoter polymorphism with insulin resistance in obesity. Obes Res 2002; 10: Ford ES. C-reactive protein concentration and cardiovascular disease risk factors in children: findings from the National Health and Nutrition Examination Survey Circulation 2003; 108: Nomura S et al. High-shear stress-induced activation of platelets and microparticles enhances expression of cell adhesion molecules in THP-1 and endothelial cells. Atherosclerosis 2001; 158: Sterpetti AV et al. Shear stress increases the release of interleukin-1 and interleukin-6 by aortic endothelial cells. Surgery 1993; 114: Engstrom G et al. Blood pressure increase and incidence of hypertension in relation to inflammation-sensitive plasma proteins. Arterioscler Thromb Vasc Biol 2002; 22: