|Home | About | Journals | Submit | Contact Us | Français|
Ethnic disparities in cardiovascular disease (CVD) may partially reflect differences in chronic stress burden that vary by social class and exposure to ethnic discrimination. Stress is associated with increased endothelin-1 (ET-1). This study examined the relationship of ET-1 to socioeconomic status (SES) and to perceived ethnic discrimination among black (n = 51) and white (n = 65) adults (mean age 36.5).
The Perceived Discrimination subscale of the Scale of Ethnic Experience measured exposure to discrimination and the Hollingshead Two-Factor Index of Social Position assessed SES. Plasma ET-1 was sampled upon awakening after an overnight admission.
SES and ET-1 levels were similar across ethnic groups, but mean discrimination scores were higher among blacks than whites (P < 0.001). Multiple regressions found that the SES × ethnicity interaction was associated with ET-1 (P < 0.05), after adjustment for gender, resting mean arterial pressure (MAP), body mass index (BMI), and exercise frequency. Regressions stratified by ethnicity revealed that lower SES correlated with higher ET-1 in whites (P < 0.001), but not blacks, and accounted for 21% of the variance. Another series of regressions revealed an interaction effect of ethnicity by discrimination on ET-1 (P < 0.05). Increased discrimination correlated with increased ET-1 among blacks (P < 0.05), but not whites, and explained 11% of the variance after adjustment for SES, gender, exercise frequency, and socially desirable response bias.
Thus, ET-1 levels increased in association with different psychosocial burdens in blacks and whites. Plasma ET-1 was higher among whites with lower SES and among blacks with higher levels of perceived ethnic discrimination, regardless of SES.
Blacks in the United States have disproportionately higher rates of cardiovascular disease (CVD) than whites, including hypertension rates that are among the highest in the world. However, lower prevalence of hypertension is found among people of African descent in the Caribbean and West Africa.1 Thus, nongenetic influences likely contribute to black–white differences in CVD. Although blacks show higher levels of certain cardiovascular risk factors (e.g., obesity), their rates of other risk factors (e.g., hypercholesterolemia) are often similar or lower than whites.2 Ethnic differences in exposure to chronically stressful conditions, including lower socioeconomic status (SES), have been proposed as underlying contributors to disparities in cardiovascular health.3 Obesity and health behaviors partially explain the links established between lower SES and CVD,4 but stress-related physiological activation may also be a contributing factor.5,6
Blacks often report greater stressors than whites, even when both groups have similar SES. Prominently embedded within the larger context of SES are ethnic differences in certain stressors, such as discrimination.3 Perceived discrimination could increase cardiovascular risk by adding to the overall burden of chronic stress.7–10 Although higher SES often confers health advantages, its influence on the physical impact of discrimination-related stress is unclear. Perceived discrimination has been associated with adverse levels of several cardiovascular health indicators,11–14 particularly increased blood pressure (BP) reactivity.15,16 Assessments of BP partly reflect underlying systemic vascular resistance, which is elevated in hypertension and is largely influenced by vasoconstriction. However, few investigations have examined the effects of chronic stress on the biochemical regulation of vasoconstriction.
Endothelin-1 (ET-1) is a potent vasoconstrictor that elicits long-lasting constriction of smooth muscle cells and induces proliferation of smooth vascular cells and cardiac myocytes.17,18 Plasma ET-1 largely reflects spillover from the vascular wall into the blood.18 Elevated ET-1 may contribute to vascular remodeling, renal dysfunction, atherosclerotic plaques, left ventricular hypertrophy, and other changes related to hypertension and heart failure.17,18 Higher ET-1 is linked to coronary heart disease19 and independently predicts mortality in patients with congestive heart failure and myocardial infarction.20 ET-1 is posited to modulate stress responses due to its distribution throughout the central nervous system, including the hypothalamus and the amygdala, which are thought to play an integrative role in physiological responses to stressful stimuli.21 However, the relationship between ET-1 in the central nervous system and circulating plasma ET-1 is uncertain.
Blacks have shown higher resting ET-1 than whites,22–25 but not consistently.26,27 Given that ET-1 has a variety of possible influences,28 the mixed findings regarding ethnic differences may be partly due to confounding by uncontrolled or unknown variables. Despite the potential role of this vasoconstrictor in cardiopathogenesis, possibly via stress-induced increases,24,25,29 literature addressing its relationship to psychosocial characteristics is limited.26,30 No studies were found that examined ET-1 for ethnic differences in the effects of discrimination. However, research revealed an association between lower SES and higher ET-1, but no ethnic group comparisons were conducted.26 It is unknown whether SES shows similar influence across ethnic groups or whether discrimination makes a unique contribution to ET-1 beyond SES. Thus, the current study examined how resting plasma levels of ET-1 are related to SES and to ethnic discrimination among blacks and whites.
The sample included 116 employed young-to-middle-aged black (n = 51; 60% women) and white (n = 65; 43% women) adults, who participated in a larger cardiovascular study between 2000 and 2005. The larger study found that psychosocial adversity (e.g., lower SES, chronic stress, anxiety, discrimination) was associated with a variety of outcomes, including greater inflammation,26,31 decreased β-adrenergic receptor sensitivity,32 and greater vascular reactivity.33
After 263 of the 385 individuals recruited via advertisements and referrals were excluded due to ineligibility or lack of interest in enrolling (usually related to time demands), the parent study sample was n = 122. Data on the current study variables were available for 116 black and white participants, who had lower to upper SES.
The study’s physicians screened for exclusions based on physical exams, laboratory tests, and self-reported medical and psychiatric histories using uniform criteria. Though 19% of participants had untreated, primary hypertension (systolic BPs: 140–168 mm Hg; diastolic BPs: 90–103 mm Hg), the sample was medication-free with no other major medical or psychiatric conditions (e.g., schizophrenia). Exclusions were diagnoses of diabetes (or a fasting glucose >120 mg/dl) and secondary hypertension (screened for renal artery stenosis (bruit upon examination), glomerular disease (proteinuria, hematuria), and renal compromise (creatinine levels >1.4 mg/dl)) (sample’s mean estimated glomerular filtration rate (eGFR) from Cockcroft-Gault calculation of creatinine clearance = 111.8 ml/min). Other exclusions were morbid obesity (body mass index (BMI) ≥40), current drug or alcohol abuse, and known sleep disorders (e.g., narcolepsy). Women were ineligible if they were pregnant, postmenopausal, taking oral contraceptives, or diagnosed with premenopausal syndrome. Ethnicity was self-reported. Participants provided informed consent in accordance with the Institutional Review Board of the University of California, San Diego (UCSD).
To determine eligibility, volunteers underwent testing conducted by a study physician (e.g., physical examination, medical history, lab work) and by medical personnel, who assessed BMI and collected three measures of seated BP at the UCSD General Clinic Research Center. Adults accepted into the study were administered demographic and psychosocial questionnaires by research staff. Participants later returned for a two-night admission, which included insertion of an intravenous catheter on the second night. At awakening the following morning, this catheter was used to obtain a blood sample for testing of ET-1.
The Hollingshead Two-Factor Index of Social Position is a widely used measure of SES that is computed by summing weighted values (1–7) that are applied to occupational and educational levels.34 Social position index scores can range from 11 to 77, with lower scores representing higher SES.
The Scale of Ethnic Experience, Perceived Discrimination subscale,35 assesses perceptions about ethnic discrimination experienced by the respondent and members of his/her ethnic group in society. Respondents endorse items on a Likert scale from “1” (strongly disagree) to “5” (strongly agree), which are summed (after reverse scoring as appropriate). Higher scores on its nine items (e.g., “In my life, I have experienced prejudice because of my ethnicity”) indicate greater perceptions of ethnic discrimination.
After catheter insertion at ~5 PM the previous evening, a blood sample was obtained upon awakening the next day at 6 AM while all participants were still in bed, to limit possible effects of movement and circadian variations on ET-1 levels.36 Whole blood was preserved with ethylenediamine-tetraacetic acid, then spun in a refrigerated centrifuge, and stored at −80 °C until assay. ET-1 levels were determined by commercial enzyme-linked immunosorbent assay (R&D Systems; Minneapolis, MN). For the ET-1 assay, sensitivity was 0.1 pg/ml and intra- and interassay coefficients of variation were 5.5 and 7.5%, respectively.
After 5 min of seated rest, three measures of mean arterial pressure (MAP) were obtained with a Dinamap 1846x monitor (Critikon, Tampa, FL) and averaged to compute resting MAP.
Weight and height measured on a calibrated scale were used to derive BMI to the nearest 0.1 kg and 0.1 cm, respectively.
The Marlowe Crowne Social Desirability Scale contains 33 true/false items representing socially desirable behaviors unlikely to occur in the general population.37 Higher scores suggest a tendency to overreport desirable behaviors and underreport negative behaviors. We have found ethnic differences in the influence of social desirability on self-reported psychosocial distress;38 thus, social desirability is included as a covariate to control for this form of response bias.
The Leisure Time Exercise Questionnaire39 includes a one item “sweat score,” which is a 3-point assessment of how frequently (“often = 1”, “sometimes = 2”, or “never/rarely = 3”) the respondent engages in activity “long enough to work up a sweat” (or make heart rate beat more rapidly) each week. This score correlates with cardiovascular fitness as measured by maximal oxygen uptake and treadmill test.40 Lower scores indicate a higher frequency of intense physical activity.
Primary analyses involved two sets of hierarchical multiple regressions to examine ET-1 for the effects of: (i) ethnicity, SES, and the SES × ethnicity interaction; and (ii) ethnicity, discrimination, and the discrimination × ethnicity interaction. For the initial regression of SES × ethnicity on ET-1, covariates were entered on Step 1, ethnicity on Step 2, SES on Step 3, and the SES × ethnicity interaction on the last step. The regression of discrimination × ethnicity on ET-1 included covariates on Step 1, ethnicity on Step 2, discrimination on Step 3, and discrimination × ethnicity on the final step. Significant interaction effects of SES × ethnicity on ET-1 and of discrimination × ethnicity on ET-1 were subsequently examined with two sets of ethnicity-stratified regressions. Whereas these follow-up models in blacks and whites all included covariates on the first step, SES was on the last step in one set and discrimination was on the last step in the other set.
Considering the power allowed by the sample size, four covariates were included in the aforementioned regression models based on prior literature22–32 and their correlations with the variables of interest or group differences. For models of SES × ethnicity on ET-1, covariates were gender, BMI, resting MAP, and exercise. Models of discrimination × ethnicity on ET-1 included the covariates gender, exercise, and socially desirable response bias, along with SES (to examine the contribution of discrimination beyond social status).
Continuous scores represented the predictors of SES (i.e., social index) and perceived discrimination, as well as the dependent variable of ET-1 (square root transformed to reduce skew). Ethnicity and gender were both dummy coded (black = 1, white = 0; women = 1, men = 0). Ethnicity correlated highly with discrimination, but no multicollinearity was detected; thus, both predictors were included in the discrimination × ethnicity models.
Table 1 shows that blacks and whites did not exhibit significantly different mean levels of ET-1 or SES (i.e., social index). Compared to whites, blacks had higher levels of perceived discrimination (P < 0.001), socially desirable responding (P < 0.01), BMI (P < 0.001), and resting MAP (P < 0.05). Whites had less hypertension (P < 0.05) and reported more frequent exercise (P < 0.01) than blacks.
Table 2 details how ET-1, SES, and discrimination correlate (r) with other study variables in blacks and whites. Among blacks, lower SES showed simple correlations with lower discrimination (P < 0.01), lower exercise, and higher smoking (Ps < 0.05). Neither ET-1 nor discrimination was correlated with other study variables in blacks. Among whites, higher ET-1 was associated with male gender (P < 0.01), lower SES (P < 0.001), and lower socially desirable responding (P < 0.05). Higher perceived discrimination in whites correlated with lower SES (P < 0.01) and higher BMI (P < 0.05). Lower SES in whites also correlated with higher MAP and higher smoking (Ps < 0.05).
The initial regression of SES × ethnicity on ET-1 was significant (P < 0.05), after controlling for gender, resting MAP, BMI, and exercise. The interaction of SES × ethnicity on ET-1 was further examined in blacks and whites (detailed in Table 3). Among whites, regression analyses indicate that plasma ET-1 increased as SES decreased (P < 0.001), with SES contributing 21% of the variance beyond that explained by gender, MAP, BMI, and exercise. The full model for SES on ET-1 with covariates was significant among whites (F(5,59) = 6.21; P < 0.001, R2 = 0.35). Among blacks, SES did not reach significance for ET-1 and the full model was not significant (F(4,45) = 0.85; P > 0.05, R2 = 0.09). Figure 1 depicts the relationship between ET-1 and SES (i.e., social index scores) for each ethnic group.
Regression results revealed that the discrimination × ethnicity interaction was significant for ET-1 (P < 0.05), after adjustment for gender, exercise, socially desirable response bias, and SES. This interaction effect was examined with regressions conducted on each ethnic group (detailed in Table 4). Among blacks, increased perceived discrimination was associated with increased ET-1 (P < 0.05), explaining 11% of the variance. Moreover, the final adjusted model for ET-1 in blacks was significant (F(5,46) = 2.77, P < 0.05, R2 = 0.23). Among whites, discrimination was not a significant contributor to ET-1 (P > 0.05). However, the full model for whites with covariates was significant for ET-1 (F(5,59) = 7.45, P < 0.001, R2 = 0.39). Figure 2 depicts the relationship between ET-1 levels and discrimination for each ethnic group.
Regressions (data not shown) were conducted for each ethnic group to explore the impact of additional covariates on the previously described models. Three separate regressions of SES on ET-1 were conducted to each include one exploratory covariate of smoking, discrimination, or eGFR, along with the original covariates (i.e., gender, BMI, MAP, and exercise). The discrimination on ET-1 regression model was altered to add MAP, BMI, or eGFR to the original covariates (i.e., gender, exercise, social desirability, and SES). None of these exploratory covariates significantly changed the aforementioned regression results for either SES or discrimination on ET-1. The only exploratory covariate showing significance in these regressions was discrimination in the SES on ET-1 model for blacks, though SES remained nonsignificant (P = 0.09).
Chronic stressors, such as discrimination and lower SES, may have deleterious effects on cardiovascular health.3,5–16 Few studies have investigated biochemical regulation of vaso-constriction in the context of these adverse life circumstances, particularly across ethnicity. To that end, we examined how SES and perceived discrimination each interacted with ethnicity to influence ET-1. We found that ethnicity interacted with both psychosocial factors, such that each relationship to ET-1 varied by ethnic group.
Consistent with a previous study,26 SES was related to ET-1, but only among whites. Lower SES showed simple correlations with higher ET-1 among whites. In fact, regression analyses indicated that SES explained more variance (22%) in ET-1 among whites than the variance accounted for by the covariates gender, resting MAP, BMI, and exercise (14%). Although lower SES in whites was correlated with higher perceived discrimination and higher smoking, exploratory regressions suggest that neither of these factors (nor eGFR) account for this link between lower SES and ET-1. Figure 1 illustrates how whites with lower SES had higher ET-1 than whites with higher SES, but levels of ET-1 were similar among blacks regardless of SES.
Individuals with lower SES may experience a variety of chronic stressors, including discrimination. Higher discrimination showed a simple correlation with lower SES among whites, but it surprisingly was associated with higher SES among blacks. Regression results indicate that exposure to ethnic discrimination may surpass SES in influence on ET-1 levels among blacks. As illustrated in Figure 2, ET-1 increased as discrimination increased among blacks, but not whites. Perceived discrimination in black participants explained 11% of the variance in ET-1 beyond the 13% contributed by significant covariates SES, exercise, and socially desirable response bias, along with gender. Inclusion of adjustments for BMI, resting MAP, and eGFR in exploratory analyses did not diminish this effect in blacks. Taken together, our findings on blacks suggest that higher SES neither shields them from exposure to ethnic discrimination nor protects them from the impact of discrimination on ET-1 levels.
Given the study’s controlled setting and methods to limit acute confounding effects on ET-1, the observed associations may reflect the chronic influence of psychosocial stress burden on basal ET-1 regulation of vascular tone. However, these findings need to be replicated and further research is needed to elucidate the mechanisms linking lower SES and higher perceived discrimination to higher resting ET-1. Related literature suggests that chronic stressors are associated with prolonged autonomic arousal and delayed recovery after stress exposure.5,6,9 Therefore, it may be productive for the field to build on earlier work that assessed ET-1 in association with reactivity to and recovery from laboratory stressors.24,25
Although blacks and whites exhibited different psychosocial vulnerabilities for increased ET-1, the mean levels of ET-1 did not vary by ethnicity. This is consistent with some studies.26,27 However, it is inconsistent with several reports suggesting that blacks may have higher ET-1 than whites.22,25 The dissimilar results between these reports and the present study may be partly due to different sample characteristics (e.g., age range, health status) and to methodology. For example, we collected blood samples upon awakening at least 12 h after catheter insertion, which likely reduced acute influences of venipuncture and daily stressors on ET-1. Furthermore, standard early morning sampling controlled for circadian variations in ET-1.36 It is unknown whether measures taken at a different time of day would reveal ethnic group differences in ET-1.
To our knowledge, this is the first study to examine resting ET-1 for associations with perceived ethnic discrimination or SES by ethnic group. Analyses were strengthened by adjustments for several potential confounders, including socially desirable response bias. However, the study could not account for all possible influences on ET-1 (e.g., sodium and alcohol intake, genetics, endothelium properties, endothelin receptor activity). This research should be replicated with a larger sample and adjustments for additional confounders. Although major medical (e.g., secondary hypertension) and psychiatric conditions are unlikely in this sample, the screenings our physicians conducted cannot conclusively rule out the presence of such conditions. It may be beneficial for future work on ET-1 to consider aspects of SES and discrimination not measured in the current study (e.g., income, subtle vs. overt discrimination). The current data cannot explain why perceptions of ethnic discrimination were stronger as social status increased in blacks and decreased in whites. Further research is needed to examine how psychological factors (e.g., vigilance, self-esteem, control, coping) might influence these SES variations in perceived discrimination among blacks and whites. In addition, assessment of vasodilation factors, such as nitric oxide, could enhance future investigations. Finally, no causal or temporal determinations can be derived from this cross-sectional, correlational study.
Overall, results suggest that the particular psychosocial stressor “getting under the skin” to affect biochemical regulation of vasoconstriction varied by ethnic group. Whites with higher SES perceived less ethnic discrimination than their lower SES counterparts and had considerably lower ET-1 levels. On the other hand, higher SES did not protect blacks from discrimination, nor did it significantly benefit them with regard to ET-1. As perceived discrimination increased among blacks, ET-1 levels also increased, controlling for SES and other factors. Given that elevated ET-1 has been linked to greater cardiovascular morbidity and mortality, the positive association between discrimination and ET-1 among blacks might contribute to their higher rates of CVD.
This study was supported by National Institutes of Health grants HL36005, RR00827, and P60 MD00220.
Disclosure: The authors declared no conflict of interest.