| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Department of Psychiatry, University of California, San Diego, San Diego, California.
Address correspondence and reprint requests to Suzi Hong, PhD, 9500 Gilman Dr., La Jolla, CA 92093-0804. E-mail: s1hong{at}ucsd.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Methods: Social status of the subjects was determined by using the Hollingshead Two Factor Index of Social Position scale, and plasma sICAM-1 and ET-1 levels were assessed using immunoassays.
Results: Correlation analyses revealed positive correlations among plasma sICAM-1 levels, BP, and social status. Levels of ET-1 were also significantly correlated with BP (p < .01) and social status (p < .001). When subjects were categorized into three social classes, sICAM-1 levels were significantly higher in the lowest social class as compared with the upper (p < .05) or middle (p < .01) class. The levels of ET-1 were higher in the low (p < .01) and middle (p < .05) social classes as compared with the upper class. Multiple hierarchic regression analyses revealed that even after controlling for demographic and health characteristics (gender, ethnicity, age, body mass index, and smoking) and systolic BP, social status accounted for additional variance of sICAM-1 or ET-1 levels.
Conclusion: These results suggest that low-social-status individuals may incur risk for future vascular diseases through vascular inflammation.
Key Words: blood pressure • endothelin-1 • gender • inflammation • sICAM-1 • social status
Abbreviations: AA = African American; BMI = body mass index; BP = blood pressure; CRP = C-reactive protein; IL = interleukin; sICAM-1 = soluble intracellular adhesion molecule 1; EA = European American; ET-1 = endothelin-1; SBP = systolic blood pressure; SES = socioeconomic status; ICAM-1 = intercellular adhesion molecule 1.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
The literature of the association between SES and vascular inflammatory marker levels is still limited. Systemic inflammation is well characterized by elevated inflammatory marker levels in plasma, including cytokines; C-reactive protein (CRP); and endothelial cellular adhesion markers, including soluble intercellular adhesion molecule 1 (ICAM-1) (5–12). Large epidemiological studies have shown that the levels of inflammatory molecules (e.g., CRP) are clearly associated with increased coronary heart disease (13,14). There are a few studies of inflammatory markers as associated with SES; higher levels of CRP have been observed in individuals with lower SES, even after controlling for demographic variables, including gender, age, body mass index (BMI), smoking, and alcohol consumption (15). Low childhood SES or less education is associated with higher levels of CRP (3,16) and with higher inflammatory cytokine interleukin (IL)-6 and TNF-
levels (17).
A variety of factors is considered in determining SES. Previous studies assessed education levels (3,18) and occupation (15–17,19) of either individuals or their parents as reliable indicators of SES because there are often obvious associations among education, occupation, and economic status. Regardless of the SES-defining factors, though, CRP, IL-6, and TNF- are reported to be elevated in individuals of lower SES.
Although the link between low SES and levels of CRP and some inflammatory cytokines (i.e., IL-6 and TNF-) is previously documented and replicated, studies on the association between SES and vascular inflammatory markers such as ICAM-1 or endothelin-1 (ET-1) are scarce. ICAM-1 is a membrane-bound adhesion molecule that mediates the firm attachment of leukocytes to endothelial cells, and this process is important in homing of leukocytes to the area of infection or injury. However, its elevated expression on the vascular endothelium could lead to vascular pathogenesis such as atherosclerosis. Individuals with high BP show elevated levels of circulating soluble ICAM-1, and elevation of its circulating soluble form reflects increased expression on the endothelium, implying endothelial activation (20). Elevated levels of soluble intracellular adhesion molecule 1 (sICAM-1) have been consistently shown in cardiovascular diseases (9,10,21–23).
ET-1, which is produced by endothelial and vascular smooth muscle cells, is a potent vasoconstrictor peptide. Elevated levels of ET-1 are implicated in hypertension and vascular diseases (24) and significantly correlated with elevated systolic blood pressure (SBP) (25). ET-1 is also shown to increase expression of ICAM-1, vascular cell adhesion molecule-1, and CD44 on fibroblast-like synovial cells (26) and inflammatory cytokine IL-6 secretion (18). Like sICAM-1, elevated ET-1 levels in plasma reflect increased ET-1 production by endothelial cells and underlying inflammation and dysregulation in the vascular endothelium. Currently, little is known about the association between ET-1 and SES in the literature.
In this study, we examined the association among plasma levels of sICAM-1 and ET-1, BP, and social status (i.e., education and occupation) among 121 African Americans (AA) and European Americans (EA). We hypothesized that low social status will be associated with higher levels of sICAM-1 or ET-1, even after controlling for likely confounding factors.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Procedure
After a screening procedure by physical and basic laboratory examinations, resting BP was measured using a Dinamap Compact BP monitor (Critikon, Tampa, FL) and was defined as the average of three seated BPs taken in the right arm after having rested for 5 minutes. The Hollingshead Two-Factor Index of Social Position questionnaire was administered by experienced research staff. Subjects returned to the UCSD General Clinical Research Center for an overnight admission. Blood was collected in EDTA on awakening the next morning through an indwelling line with a saline lock for sICAM-1 and ET-1 analyses. Blood draw for all subjects was at 6:30 AM in order to control for circadian variation. In addition, this sampling was taken in bed before any activity in order to avoid possible effects of movements on sICAM-1 and ET-1 levels.
Social Status Assessment
Social status can be characterized in various ways. We chose to administer the Hollingshead Two-Factor Index of Social Position (28) that is a venerable objective measure of social status based on the highest level of formal education achieved (seven levels) and occupation (seven levels). Education level and job position were assigned numeric values of 1–7, weighted (scores multiplied by 7 and 4 for occupation and education, respectively), and combined to calculate a "social index" (29). As a secondary measure, subjects were also categorized to three "social classes" based on the value range of the social index: 11 to 27 (upper), 28 to 43 (middle), and 44 to 77 (lower) in order to compare demographic characteristics and biochemistry measures among the social classes.
sICAM-1 and ET-1 assays
Plasma was stored at –80°C until sICAM-1 and ET-1 levels were assessed using a high sensitivity enzyme-linked immunosorbent assay (ELISA) and a direct chemiluminescent immunoassay (CLIA) from R&D Systems (Minneapolis, MN), respectively. The sensitivity of the sICAM-1 assay was <0.35 ng/ml. Intra- and interassay variation (CV) were 4.6% and 6.4%, respectively. The sensitivity of the ET-1 assay was <0.15 pg/ml, and intra- and interassays variation of the ET-1 assays were 2.9% and 5.7%, respectively.
Statistical Analyses
One-way ANOVA and 
2 analyses were performed to compare demographic characteristics in the social class categories. Pearson correlation analyses were performed to examine associations between BP, social class, and the levels of sICAM-1 and ET-1 after confirming the normal distribution.
To test the main hypothesis of the study, hierarchical multiple regression analyses were performed to examine if demographic variables and resting BP affected the associations between plasma sICAM-1 or ET-1 levels and social status. Demographic characteristics (gender, ethnicity, age) were entered as a first step, health variables (BMI, SBP, and smoking) were entered as a second step, and the social class variable was entered as a third step to examine if social class further accounted for variances in sICAM-1 or ET-1 levels after adjusting for demographic and health variables. Statistical analyses were done using SPSS 11.0 (SPSS, Inc., Chicago, IL), and a significance level was determined at < 0.05.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Subjects' demographic characteristics are presented by three categories of social class (Table 1). There were no statistical differences in mean age, BMI, gender, ethnicity, or hypertension distributions among the social class categories, although age and BMI were the highest in the lower class. In addition, gender and ethnicity distribution across the classes indicated higher percentage of men and AAs in the lower class. Systolic BP was significantly different among the social classes such that the lower class exhibited higher BP than the upper or middle classes (p < .05). There were 19 smokers among 121 subjects, and the number was significantly higher (p < .01) in the lower social class as compared with the upper or middle class. Measurements of lipid and cholesterol (i.e., triglyceride, high-density cholesterol, and low-density cholesterol) were not significantly different among the social classes.
|
Demographic Characteristics and Inflammatory Markers
sICAM-1 levels were compared based on different demographic characteristics (Table 2). There was a significant gender difference such that sICAM-1 levels were higher in men than women (p < .05). In addition, sICAM-1 levels were higher among smokers as compared with nonsmokers (p < .01). However, there were no significant differences in sICAM-1 levels between the hypertensive and normotensive individuals or between AA and EA in our sample.
|
ET-1 levels were also significantly higher in men than women. ET-1 levels were slightly higher among smokers as compared with nonsmokers, but this difference was not significant. Also, there were no significant differences in ET-1 levels between the hypertensive and normotensive individuals or between AA and EA.
sICAM-1 Levels and Social Status
Correlation analyses revealed that sICAM-1 levels were significantly related to average SBP (r = 0.25, p < .01). Social index was also significantly correlated with sICAM-1 levels (r = 0.34, p < .001) such that higher sICAM-1 levels were associated with higher social index values (i.e., lower social status). BMI and age were not associated with sICAM-1 levels. There was no significant correlation between sICAM-1 and ET-1 levels (r = 0.09). Plasma sICAM-1 levels were also significantly different among the three social classes (F(2,118) = 6.45, p = .002). Bonferroni post hoc tests revealed that the levels of sICAM-1 were higher in the lower class (248.41 ± 92.65 ng/ml) as compared with the upper (203.52 ± 50.35 ng/ml) or middle (198.26 ± 57.90 ng/ml) class. However, no difference in sICAM-1 levels was found between the upper and middle social classes.
Multiple hierarchical regression analyses were performed to examine the role of social index in explaining the sICAM-1 variance (Table 3). At step 1, demographic variables (gender, ethnicity, and age) were entered. At step 2, health characteristics (BMI, current smoking, and SBP) that are known to be associated with sICAM-1 levels in the literature were entered. Social index was then entered as a step 3 in order to examine if social status contributed additional information in sICAM-1 levels, even after controlling for the demographic and health influences. As a result, the demographic variables were not significant predictors and only accounted for about 2% of the sICAM-1 variance in the model 1. Fourteen percent of the variation in sICAM-1 levels was accounted for by including health characteristics (model 2; 
R2 = 0.12), with SBP and smoking entering as the significant predictors. The addition of social index explained an additional 5% of the variation (model 3). Systolic BP and social status entered as significant predictors of sICAM-1 levels in the final model.
|
ET-1 Levels and Social Status
Correlation analyses revealed that ET-1 levels were associated with average SBP (r = 0.19, p < .01). Social index was also significantly correlated with ET-1 levels (r = 0.33, p < .001) such that higher ET-1 levels were associated with higher social index values (i.e., lower social class). BMI (r = 0.19, p < .05) and age (r = 0.17, p < .05) were also associated with ET-1 levels. Plasma ET-1 levels were also significantly different when compared among the social classes (F(2,118) = 8.65, p < .001). Levels of ET-1 were lower in the upper (2.32 ± 0.80 ng/ml) class as compared with the middle (3.02 ± 0.92 ng/ml) or lower (3.11 ± 0.90 ng/ml) class. The levels were the highest in the lower class, but the difference in ET-1 levels between the lower and middle social classes was not significant.
The results of multiple hierarchical regression analyses are presented in Table 4. At step 1, demographic variables (gender, ethnicity, and age) were entered. At step 2, health variables (BMI, SBP, and current smoking) that are associated with plasma ET-1 levels in the literature (26) were entered. Social index was then entered as a step 3 in order to examine if social status contributed additional information in ET-1 levels, even after controlling for the demographic and BP influences. Ten percent of the variation in ET-1 levels was accounted for by demographic variables (model 1), with gender entering as the significant predictor. In model 2, entering BMI, SBP, and smoking and demographic and health variables together explained 14% of the ET-1 variance (R2 = 4.2). The addition of social index explained an additional 4.6% of the variation. Gender and social status were significant predictors in the final model.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
The associations between BP and inflammatory marker levels found in our study are in agreement with previous studies (5–12). However, the difference in the plasma levels of sICAM-1 or ET-1 markers was not significant between the individuals with and without hypertension, although the hypertensive individuals exhibited slightly higher levels (229 vs. 211 ng/ml for sICAM-1 and 2.98 vs. 2.84 pg/ml for ET-1). There were 24 individuals who were categorized as hypertensive (screening average SBP >140 mm Hg and/or DBP >90 mm Hg) among our 121 participants. Thus, having only a small number of individuals with hypertension may have limited our statistical power in comparing inflammatory marker levels between individuals with normal BP (average BP of 121/72 mm Hg) and with hypertension (average BP of 147/77 mm Hg). Also, those 24 individuals exhibited relatively mild hypertension.1 Our results indicate that there is, in fact, an association between vascular inflammation and BP, even among individuals with normal to mildly high BP. However, whether the SES-inflammation relationship is stronger among more severely hypertensive individuals needs to be further investigated.
In our sample of 121 EA and AA men and women, the levels of inflammatory markers were significantly different based on gender but not on the ethnicity of the individual. Both sICAM-1 and ET-1 were higher among men as compared with women, regardless of their ethnicity. In further multiple regression analyses, we found that gender was not a significant predictor of sICAM-1 levels but was for ET-1. This finding is in agreement with Treiber and colleagues (30), who reported higher ET-1 levels both at rest and in response to acute laboratory stressors in men compared with women. Little is known about the biological mechanisms of gender effects on sICAM-1 or ET-1 levels in conjunction with future vascular disease risk. Also, it is unclear if menstrual cycle affects inflammatory marker levels among women. From the findings of our study, it appears that the association between SES and inflammatory markers is independent of gender. It may be informative to investigate the effects of gender by SES interaction on sICAM-1 or ET-1 levels in a future study.
Inflammation of the vascular endothelium is a major pathophysiological feature of vascular diseases such as atherosclerosis (31). Inflammation of the vessel wall is both influenced by and leads to a cascade of biological processes marked by elevated levels of a number of molecules that are readily measurable. Plasma levels of sICAM-1 reflect the amount of expression of adhesion molecule ICAM-1 on the endothelium and are a reliable indicator of vascular activation. ET-1, a potent vasoconstrictor, is linked to essential hypertension and to inflammatory cytokine production in vitro (18), indicating its possible causal effect on vascular inflammation. Trieber and colleagues (24) emphasize the critical role of both genetic predisposition (ET-1 gene polymorphism) and environmental factors (low SES) in BP regulation in response to a laboratory stressor. ET-1 is also shown to increase ICAM-1 expression (26) and to mediate stress-induced acute endothelial dysfunction (32). Thus, given their biological roles in vascular inflammation, sICAM-1 and ET-1 levels have significant clinical implications for cardiovascular diseases.
Our results highlight the influence of social environment and socioeconomic disadvantage on cardiovascular health. It should be noted, though, that the levels of sICAM-1 or ET-1 found in low social class individuals in the literature are not indicative of current clinical (cardiovascular) conditions and that the participants in our study were asymptomatic healthy individuals as described by the inclusion and exclusion criteria, except for elevated BP. However, a longitudinal epidemiological study showed that asymptomatic individuals with sICAM-1 concentration in the highest quartile at baseline (>260 ng/ml) had higher risk for future myocardial infarction, with relative risk of 1.6, as compared with individuals in the lower quartiles (33). Also, a higher level of sICAM-1 (257 ng/ml) was shown in individuals who had three out of six identified risk factors (age >60 years, diabetes, smoking, SBP >140 mm Hg, HDL <35 mg/dl, and triglyceride >200 mg/dl) for cardiovascular diseases (34). Our finding of higher sICAM-1 levels (249 ng/ml) in lower-social-class individuals is in agreement with previous findings regarding an elevated risk for future cardiovascular diseases.
Although sICAM-1 and ET-1 have significant clinical implications in association with high BP and vascular diseases, the associations that we found between inflammatory markers and social status are cross-sectional in nature. It remains uncertain whether or to what degree this vascular inflammation-social status relationship would lead to future vascular diseases. Future studies can employ a longitudinal prospective design to investigate development of cardiovascular diseases among socioeconomically disadvantaged individuals and to evaluate a possible prognostic value of inflammatory markers such as sICAM-1 and ET-1 for future vascular diseases among individuals with elevated proven risk factors. We can only speculate about differential effects of the environment (e.g., hospital setting) on the inflammatory marker levels among individuals of different SES. It is possible that low-SES individuals may have felt less comfortable in the hospital setting, which may have contributed to elevated inflammatory marker levels, but this question cannot be answered with the data at hand.
We did not directly assess economic status (e.g., income, assets) in our measure of social status. However, Hollingshead Two-Factor Index of Social Position (28) is a validated and widely used measure of social position and an objective measure of social status based on the highest level of formal education and occupation. There is a variety of factors that are considered to be critical in determining one's SES. The current literature shows the SES-inflammation association regardless of how SES is defined and assessed. Whether a particular dimension of the SES measurements is more closely related with elevated inflammation merits further research.2
Although it is unclear what the underlying mechanisms are, the association between low SES and increased cardiovascular disease risk may be partly explained by adverse lifestyle habits such as smoking, higher alcohol consumption, and unhealthy eating habits that often lead to a higher rate of obesity (35–37). The contribution of lifestyle, bad dietary habits, and low physical activity levels in particular to elevated low-grade vascular inflammation should be examined in a future study. In addition, psychosocial stress and/or depressive mood are a possible link between low SES and vascular inflammation, given the evidence of the association between emotional stress/depression and elevated proinflammatory cytokine levels in cardiac patients (38,39). Social disadvantage also implies limited access to medical care for both diagnosis and treatment. These unfavorable environmental, psychosocial, and behavioral factors may be compiled to pose health risks to individuals of low SES.
Regardless of how SES is defined and measured, individuals of low SES exhibit elevated inflammatory marker levels. Although the levels observed do not imply current vascular pathology, a low-grade vascular inflammation associated with socioeconomic disadvantage may lead to future development of cardiovascular diseases.
We are grateful to Dr. Loki Natarajan for her helpful comments on the statistical analyses of this manuscript.
|
NOTES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
R2 = 0.045). The associations found between ET-1 levels and social status also remained significant in the normotensive subjects (R2 = 0.044). 
2For further clarification of dimensions of SES and inflammation relationships, we performed supplemental regression analyses by including occupation and education separately in the model. For sICAM-1 levels, both occupation (ß = 0.202) and education (ß = 0.253) were significant predictors in the final models. However, only occupation (ß = 0.282) entered as a significant predictor for ET-1 levels.
This work was supported by grants RO1 HL36005, RO1 HL57265-05, MO1-RR00827-25, and P60 MD00220, National Institutes of Health.
DOI:10.1097/01.psy.0000227684.81684.07
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
