Stress-Associated Neurobiological Pathway Linking Socioeconomic Disparities to Cardiovascular Disease

Abstract

Background:

Lower socioeconomic status (SES) associates with a higher risk of major adverse cardiac events (MACE), via mechanisms that are not well understood.

Objective:

Since psychosocial stress is more prevalent among those with low SES, we sought to test the hypothesis that stress-associated neurobiological pathways involving upregulated inflammation in part mediate the link between lower SES and MACE.

Methods:

509 individuals, median age 55 years (IQR: 45–66), underwent clinically indicated whole-body ¹⁸F-FDG-PET/CT imaging and met pre-defined inclusion criteria, including absence of known cardiovascular disease or active cancer. Baseline hematopoietic tissue activity, arterial inflammation, and in a subset of 289 resting amygdalar metabolism (a measure of stress-associated neural activity) were quantified using validated ¹⁸F-FDG-PET/CT methods. SES was captured by neighborhood SES factors (e.g., median household income and crime). MACE within 5 years of imaging was adjudicated.

Results:

Over a median 4.0 years, 40 individuals experienced MACE. Baseline income inversely associated with amygdalar activity (standardized β [95% CI]: −0.157 (−0.266, −0.041), p=0.007] and arterial inflammation (−0.10 [−0.18, −0.14], p=0.022). Further, income associated with subsequent MACE (standardized HR [95% CI]: 0.67 [0.47, 0.96], p=0.029) after multivariable adjustments. Mediation analysis demonstrated that the path of: ↓neighborhood income→↑amygdalar activity→↑bone marrow activity→↑arterial inflammation→↑MACE was significant (b =−0.01, [−0.06, −0.001], p<0.05).

Conclusions:

Lower SES: 1) associates with higher amygdalar activity and 2) independently predicts MACE via a serial pathway that includes higher amygdalar activity, bone marrow activity, and arterial inflammation. These findings illuminate a stress-associated neurobiological mechanism by which SES disparities may potentiate adverse health outcomes.

Condensed Abstract:

We tested the hypothesis that lower socioeconomic status (SES) leads to a higher risk of major adverse cardiac events (MACE) via a neuro-immunologic pathway. In 509 individuals who underwent ¹⁸ F-FDG-PET/CT imaging, lower SES associated with increased activity of the amygdala (a neural tissue involved in the response to stress) and arterial inflammation; both in-turn predicted MACE. Mediation analysis demonstrated that the pre-specified serial biological pathway (↓SES→↑ amygdalar activity →↑bone marrow activity →↑arterial inflammation→ ↑MACE) was significant. These observations illuminate a biological mechanism by which lower SES may precipitate cardiovascular disease and suggests new targets

Introduction

Life expectancy in the United States varies substantially by socioeconomic status (SES), which also contributes to significant disparities in mortality by geography and race (1,2). Measures of SES, such as income and neighborhood environmental factors, strongly predict disease outcomes, including cardiovascular disease (CVD) events (3). SES-related disparities in CVD outcomes are in part attributable to differences in CVD risk factors, access to quality care, and health behaviors (3). However, those factors alone do not explain the large variance in outcomes (4,5), prompting the hypothesis that lower SES drives health inequalities through yet undiscovered biological mechanisms (6). The discovery of such mechanisms may reveal new opportunities to attenuate the burden of disease in socioeconomically disadvantaged communities.

Lower SES associates with greater psychosocial stress (7–9) and systemic inflammation (10–12), both of which in turn, are associated with an increased risk of diseases that link to low SES, such as cancer and CVD (13,14). Thus, among biological mechanisms suspected of connecting lower SES to physical disease, those involving stress and inflammation loom prominently. The stress response begins in the brain’s salience network, within which the amygdala is a key component (15–20). Activation of this network increases sympathetic nervous system output (21), which triggers increased bone marrow hematopoietic stem and progenitor cell proliferation, and accelerates innate immune cell output (22). This stress-induced leukopoiesis results in greater atherosclerotic inflammation (22), a critical driver of CVD (23–25). In light of those observations, we previously hypothesized that increased amygdalar activity associates with an increased risk of CVD and that it does so through heightened leukopoiesis and arterial inflammation (20). To test those hypotheses, we used ¹⁸F-fluorodeoxyglucose positron emission tomography/computed tomography (¹⁸F-FDG-PET/CT) to objectively quantify metabolic activity in the amygdala (AmygA, a measure of stress-associated neural activity), the bone marrow (a measure of hematopoiesis), and the arterial wall (a measure of atherosclerotic inflammation) in individuals who were subsequently followed for the development of incident CVD. In that study, we observed that AmygA robustly predicts the risk for, and timing of, subsequent CVD (20). Furthermore, mediation analysis indicated that stress significantly associated with CVD events through a serial pathway of: ↑stress → ↑AmygA → ↑hematopoietic tissue activity → ↑arterial inflammation → ↑ major adverse cardiac events (MACE). Given the association between SES and stress, we hypothesized the existence of a biological pathway linking SES to CVD whereby low SES can drive the front end (↓SES →↑stress, etc.) of the aforementioned pathway.

Accordingly, we studied individuals who underwent multi-system ¹⁸F-FDG-PET/CT imaging, and measured amygdalar activity, bone marrow activity, and arterial inflammation. Residence-specific metrics were used to estimate socioeconomic status, after which we assessed for the development of subsequent incident MACE. We then tested the hypotheses that lower socioeconomic status: 1) associates with higher amygdalar activity and arterial inflammation, 2) independently associates with MACE, and 3) links to MACE via the prespecified amygdalar-hematopoietic-arterial pathway.

Materials and Methods

Overview

The study’s findings are derived from a retrospective, longitudinal, observational imaging study that evaluated the relationship between SES, amygdalar activity, atherosclerotic inflammation, and subsequent MACE. The study protocol was approved by the Partners Human Research Committee.

Population

The study population was described previously (25). Subjects were identified from a pool of 6,088 patients who underwent ¹⁸F-FDG-PET/CT imaging for clinical evaluation (primarily cancer surveillance) at the Massachusetts General Hospital between 2005–2008 (Figure 1A). Pre-defined inclusion criteria included: 1) absence of known CVD, 2) either absence of prior cancer or remission from cancer for at least one year prior to imaging and throughout the follow-up period, 3) absence of acute or chronic inflammatory or autoimmune disease at the time of imaging, and 4) age >30 years. To ensure adequate information for MACE adjudication, subjects were required to have at least three clinical encounter notes after index imaging in the medical records. All individuals for whom SES data were available and were included in the prior “CVD events study” (25) were included in the current study.

Figure 1. (A) Study Cohort.

The study cohort was derived from a database of patients who had undergone whole-body ¹⁸F-FDG-PET/CT imaging at Massachusetts General Hospital. All subjects meeting pre-defined criteria were included. Image analyses, event adjudication, and determination of SES were performed by mutually blinded investigators. (B) Measurement of Tissue Activity. Amygdalar activity (corrected for background cerebral activity) and arterial ¹⁸F-FDG uptake (corrected for background blood activity) were measured as validated measures of stress-associated neurobiological activity and arterial inflammation, respectively.

Outcome events

Event adjudication was performed by two cardiologists blinded to clinical and imaging data. MACE was defined as: cardiac death, myocardial infarction, unstable angina, cerebrovascular accident, peripheral artery disease with revascularization, or heart failure.

Estimation of socioeconomic status

Population-based, SES measures were derived from the U.S. Census Bureau’s 2015 American Community Survey 5-Year Estimates and Massachusetts Uniform Crime Reporting database by the Federal Bureau of Investigation. We utilized subjects’ residential addresses to derive residence-specific metrics at the zip code level. Although zip-code SES measures may be less precise as a proxy for individual SES, we hypothesized the larger area level measure would be particularly salient to the relationship between neighborhood crime and income and amygdalar activity.(26) Further details are provided in the Online Appendix.

¹⁸F-FDG PET/CT Imaging Protocol

¹⁸F-FDG-PET/CT imaging was performed with a Biograph 64 (or similar) scanner (Siemens Healthcare, Erlangen, Germany). Following an overnight fast, ~370 MBq of intravenous ¹⁸F-FDG was administered. Individuals sat in a quiet waiting room after radiotracer administration before imaging was performed after approximately one hour. A non-gated, non-contrast-enhanced CT (120 keV, ~50 mAs) was acquired for attenuation correction.

Measurement of regional brain ¹⁸F-FDG uptake

Image analyses was conducted by an investigator who was blinded to all clinical and SES data. Analysis of AmygA (Figure 1B) was performed using validated approaches (20). AmygA associates with anxious temperament (18,27), clinical manifestations of stress-related disorders (27–30), risk of subsequent incident diabetes(31), noncalcified coronary artery plaques (32) and risk of MACE (20). Using a dedicated workstation (Leonardo–TrueD, Siemens Solutions), circular regions of interest (ROI) were placed in the right and left amygdalae, as well as in background cerebral regions (the right and left temporal lobes). ¹⁸F-FDG accumulation was recorded as the mean standardized uptake value (SUV) for each ROI. AmygA was calculated as the average mean SUV of both amygdalae, divided by background cerebral activity (20). Further details are provided in the Online Appendix.

Measurement of arterial inflammation and hematopoietic activity

Arterial inflammation and bone marrow metabolic activity (Figure 1B) were measured according to previously validated methods (25,33). ¹⁸F-FDG, a radioactive glucose analogue, accumulates within tissues in proportion to their glycolytic rates (34). Because inflammatory cells (especially pro-inflammatory macrophages) have relatively high glycolytic rates (34), there is substantial accumulation of ¹⁸F-FDG within inflamed tissues.¹⁸F-FDG uptake within the arterial wall provides a validated measure of atherosclerotic inflammation (34), measured as the SUV in the wall of the aorta adjusted for background venous blood activity (by calculating a target-to-background ratio (TBR)). Similarly, hematopoietic tissue activity was assessed by deriving the SUV from ROIs placed within vertebrae (from T1 to L5) (33). Coronary artery calcium score was derived from the CT images (35,36).

Statistical analysis

Statistical analyses were performed using SPSS (IBM Corp, Version 25). Continuous variables are presented as mean (±standard deviation (SD)) or as median (25^th-75^th percentile (IQR)) when not normally distributed. Multivariable associations were evaluated using linear regression models. Where variables were non-normally distributed, bootstrapping was employed. To assess associations with MACE, Cox proportional hazards models were used to derive hazard ratios (HR) and 95% confidence intervals (CIs). Additionally, Kaplan-Meier estimates of event-free survival were generated; statistical significance was evaluated using log-rank tests. The models incorporated time between imaging and the date of MACE event or last follow-up.

Mediation analysis was performed with the SPSS PROCESS macro (v2.16.3), which uses an ordinary least squares or logistic regression-based path framework to estimate direct and indirect effects and produces CIs from bias-corrected bootstrap samples (37). We examined the pre-specified hypothesized multiple mediator path (using PROCESS Model 6): SES→ AmygA → bone marrow activity → arterial inflammation → MACE. Mediation analyses incorporated age, sex, and baseline coronary artery calcium score (as a measure of pre-existing pre-clinical atherosclerotic disease burden) as covariates. Statistical significance was determined as two-tailed p<0.05 for all analyses.

Results

Baseline characteristics

Five hundred-nine (509) individuals provided arterial imaging, SES, and clinical events data; brain imaging data were available in a subset of 289. Individuals who developed subsequent MACE had a higher prevalence of several atherosclerotic risk factors, a greater burden of coronary artery calcium (indicative of pre-existing, preclinical atherosclerosis), and (as hypothesized) lower SES (Table 1).

Table 1.

Baseline characteristics

Characteristics	Full cohort (N=509)	No MACE (N=469)	MACE (N=40)	p-value
Age, years, median (IQR)	55 (45, 66)	54 (44, 65)	68 (61, 78)	<0.001
Male, N (%)	213 (42)	197 (42)	16 (40)	0.805
Caucasian N (%)	461 (91)	425 (91)	36 (90)	0.898
Current smoker, N (%)	54 (11)	43 (9)	11 (28)	<0.001
Hypertension, N (%)	179 (35)	154 (33)	25 (63)	<0.001
Diabetes Mellitus, N (%)	45 (9)	37 (8)	8 (20)	0.010
Hyperlipidemia, N (%)	141 (28)	125 (27)	16 (40)	0.072
Total cholesterol (mg/dL), Mean (SD)	192 (43)	192 (43)	186 (40)	0.445
LDL cholesterol (mg/dL), Mean (SD)	110 (39)	111 (37)	104 (34)	0.698
Statin therapy, N (%)	101 (20)	84 (18)	16 (40)	0.001
FRS, median (IQR)	3.0 (1.0, 8.0)	2.0 (1.0, 6.75)	8.0 (3.0, 13.50)	<0.001
BMI (kg/m2), median (IQR)	26.4 (23.4, 30.9)	26.3 (23.3, 30.9)	27.9 (24.1, 31.3)	0.585
Coronary Artery Calcium Score 0–10 11–99 ≥ 100	327 (74) 60 (14) 57 (13)	314 (76) 51 (12) 49 (12)	13 (43) 9 (30) 8 (27)	<0.001
History of cancer, N (%)	427 (84)	403 (86)	24 (60)	<0.001
History of Clinically-Documented Depression or Anxiety* N (%)	29 (10)	27 (10)	2 (13)	0.740
Local Median Household Income ($) Mean (SD)	82,017 (27,974)	82,815 (28,419)	72,657 (20,100)	0.027
Local Total Crime Rate Per thousand, Mean (SD)	24.6 (12.4)	24.1 (12.2)	30.2 (13.6)	0.008

Abbreviations: BMI-body mass index, FRS-Framingham risk score, IQR-interquartile range, LDL-low density lipoprotein, MACE-major adverse cardiovascular event, SD-standard deviation.

^*Data on depression/anxiety are available on 288 subjects

SES Indices vs. Amygdalar Activity and Arterial Inflammation

Neighborhood median income negatively associated with stress-related neurobiological activity, measured as resting AmygA (standardized β [95% CI]: −0.156 [−0.267, −0.042], p=0.007) after adjusting for age and gender, i.e. for every standard deviation (SD) increase in neighborhood median income, there was a 0.156 SD decrease in AmygA. Similarly, income negatively associated with arterial inflammation after adjusting for CVD risk factors (−0.098 [−0.182, −0.14], p=0.022). In sensitivity analyses, several additional indices of SES associated with AmygA (Table 2).

Table 2.

Associations Between SES Indices and Amygdalar Activity

Predictors		Associations between SES Indices and Amygdalar Activity
		β (95%CI)	p
Local Socioeconomic Factors	% High School Graduates	−0.116 (−0.227, 0.002)	0.046
	Median Household Income	−0.156 (−0.264, −0.042)	0.007
	% Living Below Poverty	0.098 (−0.016, 0.211)	0.092
	Population	0.116 (0.003, 0.220)	0.045
	Total Housing Units	0.129 (0.015, 0.230)	0.026
Crime Statistics	Violent Crime Rate	0.109 (−0.021, 0.226)	0.104
	Murder Rate	0.008 (−0.125, 0.140)	0.912
	Rape Rate	0.049 (−0.078, 0.168)	0.471
	Robbery Rate	0.097 (−0.033, 0.210)	0.151
	Aggravated Assault Rate	0.118 (−0.013, 0.237)	0.079
	Property Crime Rate	0.144 (0.013, 0.265)	0.031
	Burglary Rate	0.064 (−0.064, 0.181)	0.345
	Larceny and Theft Rate	0.150 (0.018, 0.272)	0.025
	Motor Vehicle Theft	0.125 (−0.007, 0.261)	0.062
	Rate
	Arson Rate	−0.075 (−0.199, 0.055)	0.264
	Total Crime Rate	0.141 (0.009, 0.260)	0.035

All associations were adjusted for age and sex.

Unadjusted analyses yielded similarly significant associations.

Abbreviations: CI-confidence interval, SES-socioeconomic status.

When neighborhood median income was categorized into quartiles (Figures 2A&B) we observed lower AmygA (standardized β [95% CI]: −0.18 [−0.26, −0.06], p=0.002, Figure 2A) and lower arterial inflammation (−0.11 [−0.17, −0.02], p=0.012, Figure 2B) in subjects with higher income. When neighborhood crime rate was categorized into quartiles (Figures 2C&D), we observed higher AmygA (0.15 [0.01, 0.21], p=0.034, Figure 2C), and non-significantly higher arterial inflammation (0.05 [−0.05, 0.14], p=0.308, Figure 2D), in subjects living in neighborhoods with higher crime rates.

Figure 2. Socioeconomic Status versus Amygdalar Activity and Arterial Inflammation.

Individuals were categorized according to quartiles of their neighborhood median income and neighborhood crime rates. Amygdalar activity (A) and arterial inflammation (B) were lower as neighborhood median income increased. Conversely, amygdalar activity was higher (C) and arterial inflammation trended towards an increase (D) as neighborhood crime rate increased. Amygdalar activity was adjusted for age and sex, and arterial inflammation was additionally adjusted for CVD risk factors. Error bars indicate standard error of the mean.

SES vs. Subsequent Major Adverse Cardiovascular Events

During a median (IQR) follow-up period of 4.0 (3.0, 5.0) years, 40 individuals experienced one or more MACE events (Online Table 1). As previously observed in larger cohorts (38–40), multiple indices of neighborhood SES independently associated with MACE (Online Table 2). Neighborhood median income negatively associated with MACE (standardized hazard ratio (HR) [95% CI]: 0.673 [0.472, 0.959], p=0.029), whereas neighborhood crime rate positively associated with MACE (1.587 [1.135, 2.217], p=0.007). Further, we observed a nearly 4-fold higher MACE risk in individuals in the lowest vs. highest quartile of neighborhood median income (HR [95%CI]: 3.91 [1.30, 11.77], Cox regression p=0.015, log rank p=0.009, Figure 3), remaining significant after adjusting for CVD risk factors (3.59 [1.19, 10.85], p=0.024). Similar relationships between SES and MACE were seen when individuals were categorized by neighborhood crime rate (Online Figure 1).

Figure 3. Socioeconomic Status versus MACE.

Individuals in the lowest quartile of neighborhood median income experienced a nearly 4-fold higher risk of MACE compared to those in the highest quartile (HR [95%CI]: 3.91 [1.30, 11.77], Cox regression p=0.015; log rank p=0.009). Abbreviation: Q-quartile.

We further evaluated the association between SES and MACE after considering important potential confounders: adverse health behaviors/risk factors and decreased healthcare access. After adjusting for risk factors associated with lower SES (i.e., smoking and obesity), we observed a ~6-fold higher MACE risk in individuals in the lowest (vs. highest) quartile of neighborhood median income (6.31 [1.41, 28.18], p=0.016). Similarly, in stratified subgroup analyses, wherein smokers and obese individuals were excluded, we observed persistent associations between SES and MACE (Online Table 3). Additionally, when we adjusted for healthcare access (i.e., insurance status and in-state vs. out-of-state residence), we observed a ~3.6-fold higher MACE risk in individuals in the lowest (vs. highest) quartile of neighborhood median income (3.57 [1.18, 10.83], p=0.036). Furthermore, in an analysis wherein individuals with limited health care access were excluded (i.e., uninsured individuals and individuals living outside of Massachusetts, which mandates health coverage), the association between SES and MACE remained significant (Online Table 3).

Amygdalar Activity vs MACE among those with Lower SES

We observed that AmygA remained robustly predictive of MACE after adjusting for SES (HR [95%CI]: 1.44 [1.14, 1.83], p=0.003). To evaluate whether AmygA remained predictive of MACE specifically among individuals living within neighborhoods with high socioeconomic stress, we performed subset analyses within the group of individuals living in communities with: 1) lower neighborhood median household income (lowest tertile) and 2) higher neighborhood total crime rate (highest tertile). AmygA continued robustly to predict MACE in both analyses: (1.63 [1.28, 2.07], p<0.001); and (1.59 [1.20, 2.11], p=0.001), respectively (Table 3).

Table 3.

Amygdalar Activity vs MACE in Individuals from Neighborhods with Lower SES

SES Subgroup		Amygdalar Activity Association with MACE
		Unadjusted		Adjusted for CVD Risk Factors*		Adjusted for subclinical atherosclerosis**
		HR (95%CI)	p	HR (95%CI)	p	HR (95%CI)	p
All Neighborhoods	All individuals, from all neighborhoods	1.509 (1.198, 1.900)	<0.0001	1.423 (1.088, 1.860)	0.010	1.431 (1.097, 1.867)	0.008
Neighborhoods with Lower Income	Individuals, from neighborhoods with median income < 33rd %tile	1.626 (1.276, 2.072)	<0.001	1.877 (1.312, 2.684)	0.001	1.606 (1.232, 2.092)	<0.001
Neighborhoods with Higher Crime	Individuals, from neighborhoods with crime rate > 66th %tile	1.593 (1.203, 2.112)	0.001	1.688 (1.224, 2.330)	0.001	1.557 (1.131, 2.143)	0.007
Neighborhoods with Lower Income OR High Crime	Individuals, from neighborhoods with crime rate > 66th %tile OR median income < 33rd %tile	1.642 (1.280, 2.107)	<0.001	1.738 (1.286, 2.350)	0.001	1.636 (1.242, 2.154)	<0.001
Neighborhoods with Lower Income AND High Crime	Individuals, from neighborhoods with crime rate > 66th %tile AND median income < 33rd %tile	1.560 (1.136, 2.140)	0.006	1.736 (1.165, 2.585)	0.007	1.571 (1.086, 2.273)	0.016

Standardized hazard ratios are shown.

^*CVD risk factors entered into these models were age, sex, smoking, hypertension, diabetes, and hyperlipidemia..

^**Baseline preclinical atherosclerosis determined as coronary artery calcium score at index scanning.

Abbreviations: CVD-cardiovascular disease, CI-confidence interval, HR-hazard ratio, MACE-major adverse cardiovascular event, SES-socioeconomic status.

Amygdalar-Hematopoietic-Arterial Inflammatory Activity and MACE

Prior work showed that AmygA associates with hematopoietic tissue (i.e., bone marrow) activity and arterial inflammation (20), which in turn predict MACE (20). In the present analysis, after adjusting for CVD risk factors and neighborhood median income, AmygA associated with both hematopoietic tissue activity (β [95% CI]: 0.190 [0.077, 0.306], p=0.001) and arterial inflammation (0.202 [0.100, 0.346], p<0.001) (Online Table 4A). Similarly, AmygA, (HR [95% CI]: 1.423 [1.088, 1.860], p=0.010), hematopoietic tissue activity (2.196 [1.400, 3.445], p=0.001), and arterial inflammation (1.819 [1.399, 2.366], p<0.001) each predicted MACE after adjusting for CVD risk factors and income (Online Table 4B).

Mediation Analysis

We performed mediation analysis to examine the pre-specified pathway of: ↓SES → ↑AmygA → ↑hematopoietic tissue activity → ↑arterial inflammation → ↑MACE. SES was estimated here by quartiles of neighborhood median income (Figure 4, Online Table 5). We observed that this path significantly mediated the association of SES with MACE (standardized log odds ratio [95% CI]: −0.0137 [−0.0570, −0.000=3], p<0.05). Additionally, the path of: ↓SES → ↑AmygA → ↑arterial inflammation → ↑MACE was also significant (−0.0137 [−0.0546, −0.0001], p<0.05). Together, the two paths accounted for 28% of the total efffect of SES on risk of MACE.

Figure 4. A Hypothesized Pathway Linking Lower SES to MACE.

A serial 3-mediator analysis testing the hypothesized indirect path of: ↓SES → ↑amygdalar activity → ↑bone marrow activity→ ↑arterial inflammation → ↑MACE (red arrows) was significant (standardized log odds ratio [95% CI]= −0.0137 [−0.0570, −0.0003], p<0.05. Additionally, the path of: ↓SES→ ↑amygdalar activity → ↑arterial inflammation→ ↑MACE was also significant (−0.0137 [−0.0546, −0.0001], p<0.05). Exact p-values are not available for a dichotomous outcome measure, but in the employed bootstrap model, CIs that do not cross zero indicate p < 0.05. Abbreviations: MACE-major adverse cardiovascular event, SES-socioeconomic status.

Discussion

Results of the current study indicate that lower SES associates with heightened: 1) amygdalar activity, 2) arterial inflammation, and 3) risk of subsequent MACE. Further, path analysis supports the pre-specified hypothesis that lower SES links to MACE through a serial multi-tissue pathway involving the amygdala, hematopoietic tissues, and arterial inflammation. Thus, the current study provides supportive evidence for the existence of a neuro-immune pathway linking lower SES to CVD in adult humans.

A causal interpretation of the present findings is supported by prior observations in animal models. In macaques, experimentally-induced social subordination alters immune function, even in the absence of variation in resource access, and leads to alterations in immune cell proportions, cell type-specific gene expression, and response to an immune challenge (41). Increased stress is the likely stimulus for such immune changes that are triggered by experimentally altered social status. Similarly, in mice, chronic experimental stress increases sympathetically mediated leukopoietic progenitor cell proliferation and accelerates innate immune cell output and cytokine production (22,42,43). This pro-inflammatory immune activity in turn potentiates arterial inflammation (22), a critical driver of atherosclerotic events (24,44). The present results also build upon prior studies that have demonstrated a link between human childhood SES and amygdalar activity in response to threats in adulthood, the first steps in the hypothesized pathway tested in this study (45–51). By extending these observations to demonstrate that lower SES links to MACE in adults through multi-organ pathways driven by amygdalar activity, these results suggest new approaches to reducing the risk of MACE among low SES adults (Central Illustration).

Central Illustration. A Model of Lower Socioeconomic Status Leading to Major Adverse Coronary Events:

Prior data have demonstrated a link between low SES and higher rate of CVD. This study suggests that a biological pathway contributes to this link, involving, in series, higher AmygA, increased activation of the bone marrow (with release of inflammatory cells), which in turn leads to increased atherosclerotic inflammation and its atherothrombotic manifestations. Non-biological (and likely other biological) paths also exist. While the social variables involved in this pathway are notoriously difficult to change, the biological factors are potentially more modifiable.

The hypothesized biological pathways between SES and MACE presented here should not be interpreted as encompassing all possible pathological influences. For example, a reciprocal limb of the presented model likely exists, wherein heightened systemic inflammation reciprocally impacts brain function and exacerbates stress as a manifestation of a neuro-immune circuit. (43,51–53) Additionally, it is important to note that statistically significant mediation analyses, as observed here, do not prove causation. Future studies, especially ones containing interventions that target the nodes of the postulated mechanism, will be needed to confirm a causal role for the hypothesized pathway.

The proposed model linking SES to MACE provides a potentially important construct within which to consider future interventions. Although adverse socioeconomic factors (e.g., lower income and higher crime) have proven notoriously difficult to remedy, a biological pathway may be amenable to direct interventions to prevent MACE. Specifically, several components of the amygdalar-leukopoietic-arterial axis reported herein could be therapeutically targeted with the goal of lessening the burden of SES-related disease. Potential approaches include the use of drugs that reduce arterial inflammation (e.g., anti-inflammatory drugs or statins) or drugs that inhibit the release of pro-inflammatory cells from the bone marrow in response to stress (e.g., beta adrenoreceptor antagonists) (22). The brain, the most proximal component of the biological pathway, may represent the most attractive target. Amelioration of the neural response to low SES could potentially be accomplished through stress reduction techniques, exercise (which has well-described stress-relieving benefits), or novel drugs targeting the amygdala or other stress-related brain regions. Furthermore, it is particularly notable that AmygA remains significantly predictive of MACE even among individuals with the lowest levels of SES. This raises the possibility that reducing AmygA (or it’s down-stream biological consequences) may reduce MACE, particularly in socioeconomically disadvantaged individuals.

It is additionally important to note that the relationship between income and AmygA does not remain linear at higher incomes. A closer evaluation of AmygA, as it relates to neighborhood median income (in deciles, Online Figure 2), suggests no further improvements in AmygA beyond household income levels above approximately the 6^th decile. This threshold effect is consistent with prior findings by Kahneman et al., who observed no further improvement in reported perceived stress beyond household income levels above the second income tertile (54).

Limitations

Several potential limitations should be considered while evaluating this study’s findings. First, the subjects were identified from a clinical database of patients who had undergone imaging for clinical indications (mostly cancer surveillance), thus possibly limiting generalizability of the findings. However, there is little reason to believe that any of these clinical indications should have changed the relationships among SES, amygdalar activity and MACE. In fact, it is possible that the effects would be even stronger among patients not connected to ongoing care in a health care system. Second, neighborhood SES variables were based upon 2015 Census data (because it was closest to the time of follow-up of the cohort) rather than the individual subjects’ year of imaging. Although this approach may have led to some misclassification of SES measures, multiple studies have demonstrated that there is little reclassification of neighborhood SES between Census time periods.(55) Furthermore, any misclassification is likely to have been nondifferential, thereby making it harder to find an association. Third, in the absence of individual SES data, we necessarily used neighborhood SES measures as a surrogate for individual SES. Neighborhood SES measures have an effect on individual stress beyond individual SES measures, (3) and have repeatedly been shown to associate with adverse health outcomes. (2–5) However, future studies should examine whether individual-level SES measures yield further insights into the observed associations. Fourth, it is possible that the context of cancer screening, within which the ¹⁸F-FDG-PET/CT imaging was conducted, may have affected amygdalar activity in our participants, (e.g., being evaluated for cancer may increase anxiety). However, we do not necessarily see this as a disadvantage. Individuals who are more likely to have a negative emotional reaction to being screened for cancer may also be likely to have a more negative reaction to adverse socioeconomic circumstances. Overall, we believe these limitations are substantially counterbalanced by several important innovations, including the unique use of socioeconomic-, neuro-, and somatic imaging, and cardiovascular data to study hypothesized biological mechanisms linking SES disparities to CVD events.

Conclusions

The current findings identify a potentially modifiable biological pathway driving the increased burden of CVD that encumbers the socioeconomically disadvantaged. These pathways originate in the brain, which provides a bridge between external factors and extra-neural diseases. By illuminating a multi-organ system linking SES to CVD, this study sets the stage for testing new interventions that may forestall future heart attacks and strokes and reduce SES-driven health disparities.

CLINICAL PERSPECTIVES.

Competency in Medical Knowledge:

Low socioeconomic status is linked to stress and a greater risk of major adverse cardiac events (MACE) through up-regulation of neural activity, activation of the immune system, and arterial inflammation.

Translational Outlook:

Future research should target the neuro-immunological mechanisms mediating socioeconomically driven health disparities.

Abbreviations:

AmygA

metabolic activity in the amygdala

BM

Bone Marrow

CVD

cardiovascular disease (CVD)

¹⁸F-FDG-PET/CT

18F-fluorodeoxyglucose positron emission tomography/computed tomography

MACE

major adverse cardiac events (MACE)

MBq

Megabecquerels

SES

socioeconomic status