Inflammation and vascular endothelial activation in an Aboriginal population: relationships to coronary disease risk factors and nutritional markers

Aboriginal people in Australia suffer very high mortality from coronary heart disease (CHD).1 Nevertheless, screening in a number of Aboriginal communities has shown that the prevalence of two conventional coronary disease risk factors, hypercholesterolaemia and hypertension, is not excessive.2,3 In contrast, pro-inflammatory and pro-oxidant stimuli — such as infection, obesity, smoking4 and low intake of dietary antioxidant compounds5 — are often prevalent. These, in turn, reflect social environmental conditions conducive to ill health. Hence, inflammation may be an important mediator of elevated CHD risk in Aboriginal populations.

Atherosclerosis is a chronic inflammatory state that is initiated and/or exacerbated by dyslipidaemia, cigarette smoking, hypertension, elevated plasma homocysteine levels, infection, obesity, insulin resistance and diabetes.6 The early stages of this process involve binding of monocytes to the endothelial surface, initially through selectins expressed on the endothelium. This is followed by firmer binding through other cell adhesion molecules, penetration into the vessel intima, and perpetuation of a local inflammatory response that may lead to the initiation of atheroma. Oxidative damage to the endothelium may be important in initiation and progression of atheroscler-osis. Elevated homocysteine levels potentially promote oxidative damage and are a CHD risk factor. Conversely, protection from endothelial damage can be provided by diets high in plant foods, antioxidant compounds (eg, vitamin C, vitamin E and carotenoids), n-3 polyunsaturated fatty acids, folate and l-arginine.7

Certain biomarkers of chronic inflammation are associated with and predictive of future cardiovascular events. These include proinflammatory cyto-kines (soluble cell-adhesion molecules shed from endothelial cells, the first of which to be upregulated is soluble E-selectin [sE-selectin]), and acute-phase reactants produced in the liver, such as C-reactive protein (CRP), which is now recognised as an independent predictor of CHD.8,9,10

Our study reports on levels of two biomarkers of inflammation and endothelial-cell activation (CRP and sE-selectin) in a cross-sectional survey of Aboriginal people. It also examines relationships between these biomarkers and indicators of dietary quality and other CHD risk factors.

Smoking, diabetes and impaired glucose tolerance were common in the survey sample (Box 1). The population showed a high degree of insulin resistance, as indicated by high fasting insulin levels and HOMA insulin resistance values. Low mean levels of carotenoids and high levels of homocysteine indicated that dietary quality was poor. Fourteen participants were currently taking antihypertensive medication, eight were taking antidiabetic medication and none were taking lipid-lowering drugs. Mean levels of CRP and sE-selectin were very high, with the majority of both men and women having levels above the upper limit of the reference ranges for CRP (2.9 mg/L) and sE-selectin (81 ng/mL). After excluding participants with CRP values indicative of clinically significant inflammation (> 10 mg/L),15 the mean CRP level was 3.7 mg/L (95% CI, 3.2–4.2 mg/L) (n = 122).

Pearson correlation coefficients (r) for CRP and sE-selectin levels with other CHD risk factors are shown in Box 2. CRP and sE-selectin levels correlated significantly with each other (r = 0.202; P = 0.013). In men, CRP level correlated significantly with age, waist: hip circumference ratio, glucose concentration and insulin resistance, and showed a significant inverse correlation with HDL-cholesterol and carotenoid concentrations. Similar associations were apparent in women (except for fasting glucose and HDL-cholesterol concentrations), and CRP concentration was also significantly associated with body mass index and blood pressure in women. Excluding participants with CRP levels > 10 mg/L from the analysis gave largely similar results, except that correlations with lipid and α-tocopherol concentrations were stronger (total cholesterol, r = 0.205, P = 0.024; HDL-cholesterol, r = –0.169, P = 0.063; triglycerides, r = 0.216, P = 0.017; and α-tocopherol, r = 0.189, P = 0.042), while the inverse correlations with concentrations of carotenoids (r = –0.169, P = 0.063) and retinol (r = –0.101, P = 0.268) were weaker.

In men, sE-selectin concentration correlated significantly with body mass index, waist: hip circumference ratio, concentrations of glucose, fasting insulin and triglycerides, and insulin resistance, and correlated inversely with carotenoid concentrations (Box 2). Similar associations were apparent in women, in whom there were also significant correlations of sE-selectin concentration with blood pressure and HDL-cholesterol concentration.

Twenty-seven per cent of men and 51% of women exhibited the metabolic syndrome as defined by the criteria of an expert panel.13 After adjustment for age, sex, smoking status and carotenoid concentration, mean CRP concentration was significantly higher in the presence of the metabolic syndrome (Box 3). The effect was similar in both men and women, as indicated by the non-significance of the interaction term of sex with the metabolic syndrome (P = 0.596). Exclusion of individuals with CRP level > 10 mg/L from the analyses made no substantial difference to the results (data not shown). Inclusion of sE-selectin as a covariate made no substantial difference to the results (data not shown).

The mean concentration of sE-selectin was also significantly higher in the presence of the metabolic syndrome (Box 3). The interaction between sex and the metabolic syndrome was not significant (P = 0.311). Inclusion of CRP concentration as a covariate made no substantial difference to the results (data not shown). In these models, neither CRP concentration (P = 0.315) nor sE-selectin concentration (P = 0.755) varied by smoking status.

Data for CRP and sE-selectin concentrations were stratified into tertiles to examine the relationship with individual plasma antioxidant concentrations (Box 4). After adjustment for age, sex, smoking status, plasma cholesterol concentration and the metabolic syndrome, there were significant linear decreases in lycopene, β-carotene, cryptoxanthin and retinol levels with increasing CRP (Box 4). Excluding participants with CRP level > 10 mg/L made little difference to the results (other than to strengthen the association with lycopene; data not shown). Inclusion of sE-selectin concentration as a covariate in these models made little difference to the results (data not shown).

After adjustment for age, sex, smoking status, plasma cholesterol concentration and the metabolic syndrome, there were significant linear decreases in concentrations of lycopene, β-carotene, cryptoxanthin and lutein + zeaxanthin with increasing sE-selectin concentration (Box 4). For the relationship of cryptoxanthin and lutein + zeaxanthin with sE-selectin, there were significant interactions with sex, such that their concentrations decreased linearly with increasing sE-selectin in men but not in women (data not shown). Inclusion of CRP as a covariate in these models made no substantial difference to the results (data not shown).

Homocysteine concentration did not vary across tertiles of either CRP or sE-selectin.

Our study shows very high levels of circulating markers of vascular endothelial damage and subclinical inflammation in a remote population of Aboriginal people. Average CRP concentration (5.4 mg/L) was within the range associated with the highest risk of CHD mortality in men and women.8,9 Average sE-selectin concentration (119 ng/mL) was well above the mean for people with incident CHD or carotid atherosclerosis in the United States (40–50 ng/mL)10 and that for Europeans with cardiovascular disease (80 ng/mL).16 Average levels of both markers were higher in the presence of the metabolic syndrome, as reported in other populations.17,18 Moreover, these markers were inversely and independently associated with diet-derived antioxidants from fruit and vegetables. Our results are consistent with the very high mortality rate from cardiovascular disease in the fourth and fifth decades of life among Aboriginal people.1 The data suggest that vascular inflammation is a major factor mediating the elevated CHD risk associated with adverse social environmental circumstances and the high prevalence of the metabolic syndrome in this population.

Our study did not exclude prevalent cases of CHD, and this may have resulted in higher average CRP and sE-selectin levels than are seen in other populations. However, the fact that both markers of vascular damage were elevated, even in the absence of risk factors such as the metabolic syndrome, suggests that endothelial activation and inflammation are preceding, and mediating, overt CHD in this population. Acute infection also raises CRP levels dramatically, but exclusion of people with CRP levels > 10 mg/L from statistical analyses did not substantially alter our conclusions.

Some major studies of CRP level as a CHD risk factor have not applied an upper limit of CRP level as an exclusion criterion,8,9 and to apply such a limit assumes that elevated circulating (and tissue) CRP levels per se are less important than the underlying source of inflammation. As this assumption is yet to be definitively tested, it is difficult to interpret CRP levels unequivocally. In addition to being a marker of inflammation, CRP may have direct detrimental effects on vascular tissues,19 which may partly explain the association of insulin resistance with inflammation and endothelial dysfunction.

Cross-sectional studies in other populations have shown relationships between chronic inflammation and body mass index, insulin resistance, diabetes, blood pressure and other metabolic syndrome components, as well as an inverse relationship with physical activity.17,20,21 The association of CRP (and sE-selectin) concentration with blood pressure was stronger in women in our study, as it was in the Mexico City Diabetes Study.22 Furthermore, CRP level predicted six-year incidence of metabolic syndrome components among women in that study.22 Like CRP, sE-selectin concentration was correlated with markers of insulin resistance and other metabolic syndrome components in the Aboriginal population we studied.

Concentrations of CRP and, to an even greater extent, sE-selectin were inversely related to plasma levels of antioxidants derived from vegetables and fruit. Such evidence for dietary antioxidant protection against endothelial dysfunction and inflammation is consistent with previous studies. Endothelial dysfunction, induced in healthy subjects by instigating hyperinsulinaemia, can be ameliorated by the administration of vitamin C.23 In people with peripheral arterial disease, vitamin C levels have been found to be inversely associated with CRP levels.24 Lycopene has also been shown to reduce the expression of sE-selectin and other cell-adhesion molecules on human aortic endothelial cells in culture, and also to modulate leukocyte binding.25 Together, these results suggest that protection from oxidative stress provided by carotenoids and other dietary components of fruit and vegetables is important for the vasculature and thus prevention of CHD. However, our cross-sectional study cannot identify cause and effect relationships, and it may be that plasma carotenoid levels (related to fruit and vegetable consumption) are markers of a generally more healthy lifestyle rather than being of specific importance in themselves. Furthermore, if inflammation and its associated metabolic abnormalities are associated with greater degrees of oxidative stress, it may be that carotenoids are being removed from circulation at a greater rate when inflammation and metabolic abnormalities are present.

Poor dietary quality has been reported in other remote Aboriginal communities5 and, coupled with high oxidant stress in the form of high prevalences of diabetes and smoking, suggests that diet and food supply may be important contributors to excessive CHD risk. Endothelial dysfunction and CHD risk have also been linked to psychological stress,26 catecholamine metabolism27 and chronic conditions such as periodontal disease and Helicobacter and Chlamydia infections.28 We believe that the inverse associations of vascular inflammatory markers with dietary antioxidants, and positive relationships to metabolic syndrome components, provide further evidence of the need for broad-based interventions, including nutrition and physical activity, such as those community-directed interventions shown to be effective.29,30 Optimal nutrition and physical activity are, in turn, determined by issues such as food supply, education, infrastructure, social inequalities and psychosocial factors.

In conclusion, high levels of markers of inflammation and vascular endothelial dysfunction suggest these processes may be important mediators of CHD risk in Aboriginal people.

3: Levels of C-reactive protein and soluble E-selectin in men and women with or without the metabolic syndrome

Plasma concentrations of C-reactive protein (A) and soluble E-selectin (B) according to presence or absence of the metabolic syndrome in men and women. Data are expressed as geometric mean (95% CI), adjusted for age and smoking status. P-values are for the difference in means between strata of the metabolic syndrome.

4: Association between plasma levels of dietary antioxidants and the inflammatory markers C-reactive protein and soluble E-selectin

Plasma antioxidant concentrations were stratified by tertiles of inflammatory marker levels. Data are expressed as mean or geometric mean (95% CI), adjusted for age, sex, plasma cholesterol concentration, metabolic syndrome status and smoking. P-values indicate significance of tests for linear trend in antioxidant concentrations across tertiles.