Methods for mediation analysis with high-dimensional DNA methylation data: Possible choices and comparisons.

Epigenetic researchers often evaluate DNA methylation as a potential mediator of the effect of social/environmental exposures on a health outcome. Modern statistical methods for jointly evaluating many mediators have not been widely adopted. We compare seven methods for high-dimensional mediation analysis with continuous outcomes through both diverse simulations and analysis of DNAm data from a large multi-ethnic cohort in the United States, while providing an R package for their seamless implementation and adoption. Among the considered choices, the best-performing methods for detecting active mediators in simulations are the Bayesian sparse linear mixed model (BSLMM) and high-dimensional mediation analysis (HDMA); while the preferred methods for estimating the global mediation effect are high-dimensional linear mediation analysis (HILMA) and principal component mediation analysis (PCMA). We provide guidelines for epigenetic researchers on choosing the best method in practice and offer suggestions for future methodological development.

Policies and Indicators to Operationalize Structural Racism for Quantitative Health Research.

The field of structural racism is developing rapidly, with researchers moving from commentaries and review papers toward empirical research that aims to quantify the impact of exposure to structural racism on racial and ethnic inequities in health. Despite the increasing recognition of the impact of structural racism on health outcomes, its measurement in the empirical literature has focused on a limited set of area-based measures. This paper proposes two ways to expand the measurement of structural racism in public health research. First, by situating indicators of structural racism within the broader context of discriminatory policies that gave rise to racial and ethnic inequities, and second, by broadening the range of area-level indicators of structural racism beyond what has been used in the extant literature. The paper provides practical tools, including an extensive list of discriminatory federal policies, a list of common and novel indicators of structural racism across multiple domains, and an applied example showing how to connect policies and indicators of structural racism. These tools will aid researchers in developing a robust empirical evidence base to examine the role of structural racism in long-standing health inequities and advance quantitative research on the impact of structural racism on health inequities.

Could differential under-reporting of loneliness between men and women bias the gender-specific association between loneliness duration and rate of memory decline? A probabilistic bias analysis of effect modification.

Gender is an observed effect modifier of the association between loneliness and memory aging. However, this effect modification may be a result of information bias due to differential loneliness under-reporting by gender. We applied probabilistic bias analyses to examine whether effect modification of the loneliness-memory decline relationship by gender is retained under three simulation scenarios with various magnitudes of differential loneliness under-reporting between men and women. Data were from biennial interviews with adults aged 50+ in the US Health and Retirement Study from 1996-2016 (5,646 women and 3,386 men). Loneliness status (yes vs. no) was measured from 1996-2004 using the CES-D loneliness item and memory was measured from 2004-2016. Simulated sensitivity and specificity of the loneliness measure were informed by a validation study using the UCLA Loneliness Scale as a gold standard. The likelihood of observing effect modification by gender was higher than 90% in all simulations, although the likelihood reduced with an increasing difference in magnitude of the loneliness under-reporting between men and women. The gender difference in loneliness under-reporting did not meaningfully affect the observed effect modification by gender in our simulations. Our simulation approach may be promising to quantify potential information bias in effect modification analyses.

Exposure to structural racism-related state laws is associated with worse cardiovascular health among US adults, BRFSS 2011 and 2013.

The objective of this study was to determine whether exposure to structural racism-related state laws is associated with cardiovascular health among a racially and ethnically diverse sample of US adults. Data were from the Database of Structural Racism-Related State Laws and the Behavioral Risk Factor Surveillance System (BRFSS). The sample included 958,019 BRFSS 2011 and 2013 respondents aged 18+ from all 50 US states. The exposure was a summary index of 22 state laws related to the criminal legal system, economics and labor, education, healthcare, housing, immigration, and political participation. The outcome was the American Heart Association's Life's Simple 7 (LS7), a summary index of seven cardiovascular health indicators. Linear regression models included fixed effects for year and state to control for time trends and unmeasured time-invariant state-level contextual factors. In the full sample, a one standard deviation increase in the structural racism state legal index was associated with a 0.06-unit decrease in the LS7 (b=-0.06; 95% CI:-0.09, 0.02; p=0.001), controlling for individual- and state-level covariates. Contrary to expectations, stratified models revealed no statistically significant differences by race and ethnicity in the association between the structural racism state legal index and the LS7.

A data harmonization project of three large prospective cardiovascular health-focused cohorts.

Few prospective studies examine multilevel resilience resources and psychosocial factors in relation to cardiovascular health and disease. Recent research indicates that resilience resources are associated with a reduction in the incidence of cardiovascular disease-related events, but few studies have examined this relationship across different racial/ethnic populations or in large cohorts. Harmonization may address these limitations because it allows data from several cohorts to be analyzed together, potentially increasing sample size and in turn power overall and in minority populations. This paper describes the process involved in combining three cardiovascular health-focused cohorts: Jackson Heart Study, Multi-Ethnic Study of Atherosclerosis, and Mediators of Atherosclerosis in South Asians Living in America Study. Using a systematic process, we identified appropriate data harmonization techniques to use in harmonizing variables across cohorts. Variables included exposures (e.g., resilience resources), outcomes (e.g., American Heart Association's Life's Simple 7), and covariates (e.g., race and ethnicity). Post harmonization examinations included psychometric analyses of the harmonized variables. A total of 13,284 participants were included in the final harmonized dataset. This project provides opportunities for future research in resilience resources and informs future studies that need to harmonize data. Results based on the harmonized dataset could inform interventions and policies.

Examining the Relationship Between Multilevel Resilience Resources and Cardiovascular Disease Incidence, Overall and by Psychosocial Risks, Among Participants in the Jackson Heart Study, the Multi-Ethnic Study of Atherosclerosis, and the Mediators of Atherosclerosis in South Asians Living in America (MASALA) Study.

We examined relationships between resilience resources (optimism, social support, and neighborhood social cohesion) and cardiovascular disease (CVD) incidence and assessed potential effect-measure modification by psychosocial risk factors (e.g., stress, depression) among adults without CVD in 3 cohort studies (2000-2018): the Jackson Heart Study, the Multi-Ethnic Study of Atherosclerosis, and the Mediators of Atherosclerosis in South Asians Living in America (MASALA) Study. We fitted adjusted Cox models accounting for within-neighborhood clustering while censoring at dropout or non-CVD death. We assessed for effect-measure modification by psychosocial risks. In secondary analyses, we estimated standardized risk ratios using inverse-probability-weighted Aalen-Johansen estimators to account for confounding, dropout, and competing risks (non-CVD deaths) and obtained 95% confidence intervals (CIs) using cluster bootstrapping. For high and medium (versus low) optimism (n = 6,243), adjusted hazard ratios (HRs) for incident CVD were 0.94 (95% CI: 0.78, 1.13) and 0.90 (95% CI: 0.75, 1.07), respectively. Corresponding HRs were 0.88 (95% CI: 0.74, 1.04) and 0.92 (95% CI: 0.79, 1.06) for social support (n = 7,729) and 1.10 (95% CI: 0.94, 1.29) and 0.99 (95% CI: 0.85, 1.16) for social cohesion (n = 7,557), respectively. Some psychosocial risks modified CVD HRs. Secondary analyses yielded similar findings. For optimism and social support, an inverse relationship was frequently most compatible with the data, but a positive relationship was also compatible. For neighborhood social cohesion, positive and null relationships were most compatible. Thus, specific resilience resources may be potential intervention targets, especially among certain subgroups.

Gestational diabetes mellitus, epigenetic age and offspring metabolism.

AIMS: No reports examine the relationship between in-utero exposure to gestational diabetes mellitus (GDM), offspring epigenetic age acceleration (EAA), and offspring insulin sensitivity. METHODS: Using data from a cohort study, we examined associations between GDM in-utero exposure and offspring EAA at approximately 10 years of age, using separate regression models adjusting for offspring chronological age and sex. We also examined associations between EAA with updated homeostasis model assessment of insulin sensitivity and secretion (HOMA2-S and HOMA2-ß) measured at approximately 10 and 16 years of age, using mixed linear regression models accounting for repeated measures after adjustment for offspring chronological age and sex. RESULTS: Compared to unexposed offspring (n = 91), offspring exposed to GDM (n = 88) had greater EAA or older extrinsic age compared to chronological age (ß-coefficient 2.00, 95% confidence interval [0.71, 3.28], p = 0.0025), but not greater intrinsic EAA (ß-coefficient -0.07, 95% CI [-0.71, 0.57], p = 0.93). Extrinsic EAA was associated with lower insulin sensitivity (ß-coefficient -0.018, 95% CI [-0.035, -0.002], p = 0.03) and greater insulin secretion (ß-coefficient 0.018, 95% CI [0.006, 0.03], p = 0.003), and these associations persisted after further adjustment for measures of maternal and child adiposity. No associations were observed between intrinsic EAA and insulin sensitivity and secretion, before or after adjustment for measures of maternal and child adiposity. CONCLUSIONS: In this study, children exposed to GDM experience greater extrinsic EAA, which is associated with lower insulin sensitivity and greater insulin secretion. Further studies are needed to determine the directionality of these associations.

Caregiver-reported household food insecurity and child-reported food insecurity in relation to eating disorder risk factors and symptoms among preadolescent children.

OBJECTIVE: To examine the relation of caregiver-reported household food insecurity (FI) and child-reported FI with eating disorder (ED) risk factors and symptoms, including effect modification by gender, in preadolescent children. METHOD: Data were from the Family Food Study, a cross-sectional study of households with incomes ≤200% of the federal poverty line in southeastern Michigan. Children aged 8-10 years (n = 194) and their female primary caregivers reported separately on FI status. Children reported ED risk factors/symptoms via the 24-item Children's Eating Attitudes Test (ChEAT-24), with higher scores indicating more ED risk factors/symptoms. Linear mixed models were used to examine associations between FI measures with the ChEAT-24 total score, plus subscale scores for dieting, food preoccupation, weight preoccupation, vomiting, and social pressure to eat/gain weight. Models were adjusted for child age, child gender, caregiver race/ethnicity, caregiver education, and household income. RESULTS: Among all children, child-reported FI, but not caregiver-reported household FI, was associated with more ED risk factors/symptoms. Child-reported FI (vs. no FI) was associated with higher average ChEAT-24 total score (ß = 2.41, 95% CI: 0.57, 4.25). Child-reported FI was also associated with more food preoccupation, more weight preoccupation, and more social pressure to eat. Caregiver-reported household FI was marginally associated with less dieting in girls, and child-reported FI was associated with more dieting in boys. DISCUSSION: Child-reported FI may be more salient than caregiver-reported household FI as a risk factor for ED-related outcomes in preadolescent children. Gender may modify the association between FI and dieting behavior. PUBLIC SIGNIFICANCE STATEMENT: More child-reported food insecurity, but not parent-reported household food insecurity, was associated with more eating disorder risk factors and symptoms among preadolescent boys and girls. These findings emphasize the need for future studies that investigate the role of food insecurity in the development of eating disorders, especially studies that measure child-reported experience of food insecurity.

Perceptions of tap water associated with low-income Michigan mothers' and young children's beverage intake.

OBJECTIVE: To quantify perceptions of tap water among low-income mothers with young children residing in Michigan and examine associations between perceptions of tap water, mothers' and young children's beverage intake, and mothers' infant feeding practices. DESIGN: Cross-sectional study. SETTING: Online survey. PARTICIPANTS: Medicaid-insured individuals who had given birth at a large Midwestern US hospital between fall 2016 and fall 2020 were invited by email to complete a survey in winter 2020 (N 3881); 15·6 % (N 606) completed eligibility screening, 550 (90·8 %) were eligible to participate, and 500 (90·9 %) provided valid survey data regarding perceptions of tap water, self and child beverage intake, and infant feeding practices. RESULTS: Two-thirds (66·2 %) of mothers reported that their home tap water was safe to drink without a filter, while 21·6 % were unsure about the safety of their home tap water. Mothers' perceptions of their home tap water were associated with their own tap and bottled water intake and their young children's tap water and bottled water intake. Mothers with more negative perceptions of tap water in general, independent of their perceptions about their home tap water, consumed more bottled water and sugar-sweetened beverages, and their young children drank bottled water and fruit drinks more frequently. Few associations were observed between mothers' perceptions of tap water and infant feeding practices. CONCLUSIONS: Uncertainty about tap water safety and negative perceptions of tap water are common among low-income Michigan mothers. These beliefs may contribute to less healthful and more costly beverage intake among mothers and their young children.

Examining relationships between perceived neighborhood social cohesion and ideal cardiovascular health and whether psychosocial stressors modify observed relationships among JHS, MESA, and MASALA participants.

BACKGROUND: Psychosocial stressors increase the risks for cardiovascular disease across diverse populations. However, neighborhood level resilience resources may protect against poor cardiovascular health (CVH). This study used data from three CVH cohorts to examine longitudinally the associations of a resilience resource, perceived neighborhood social cohesion (hereafter referred to as neighborhood social cohesion), with the American Heart Association's Life's Simple 7 (LS7), and whether psychosocial stressors modify observed relationships. METHODS: We examined neighborhood social cohesion (measured in tertiles) and LS7 in the Jackson Heart Study, Multi-Ethnic Study of Atherosclerosis, and Mediators of Atherosclerosis in South Asians Living in America study. We used repeated-measures, modified Poisson regression models to estimate the relationship between neighborhood social cohesion and LS7 (primary analysis, n = 6,086) and four biological metrics (body mass index, blood pressure, cholesterol, blood glucose; secondary analysis, n = 7,291). We assessed effect measure modification by each psychosocial stressor (e.g., low educational attainment, discrimination). RESULTS: In primary analyses, adjusted prevalence ratios (aPR) and 95% confidence intervals (CIs) for ideal/intermediate versus poor CVH among high or medium (versus low) neighborhood social cohesion were 1.01 (0.97-1.05) and 1.02 (0.98-1.06), respectively. The psychosocial stressors, low education and discrimination, functioned as effect modifiers. Secondary analyses showed similar findings. Also, in the secondary analyses, there was evidence for effect modification by income. CONCLUSION: We did not find much support for an association between neighborhood social cohesion and LS7, but did find evidence of effect modification. Some of the effect modification results operated in unexpected directions. Future studies should examine neighborhood social cohesion more comprehensively and assess for effect modification by psychosocial stressors.

Heavy metal blood concentrations in association with sociocultural characteristics, anthropometry and anemia among Kenyan adolescents.

OBJECTIVES: To measure heavy metal concentrations among Kenyan youth and quantify associations with sociocultural, demographic, and health factors as well as anthropometry. METHODS: Using data from a study of semi-nomadic pastoralists in Samburu County, Kenya, we measured blood concentrations of lead (Pb), mercury (Hg), and cadmium (Cd) in 161 adolescents. We identified sociocultural, demographic and health characteristics associated with each metal and quantified the association between metals and adolescent anthropometry. RESULTS: Median blood concentrations of Pb, Cd, and Hg were 1.82 µg/dL, 0.24 µg/L and 0.16 µg/L, respectively. Place of residence (highlands vs lowlands) was a determinant of metal concentrations. Hg was inversely related to anemia, and metals were not associated with anthropometry. CONCLUSIONS: In this population of Samburu adolescents, median Pb and Cd blood concentrations were higher than other North American or European biomonitoring studies. These findings motivate further investigation into the environmental sources of metals in this community.

A mixed-methods study on the dietary practices of early postpartum women during the summer rainy season in Belgaum, Karnataka, India.

The current mixed-method study uses Food Frequency Questionnaires and 24-hour dietary recalls (n = 41) to assess the food/nutrient intake; and qualitative interviews to identify local perceptions of food among 41 early postpartum women in Belgaum, India. The results show that total energy, protein, and most micronutrient intake were significantly lower than the Recommended Dietary Allowance of India (p < .05 individually); ninety percent of mothers restricted the consumption of some specific fruits, vegetables, and other foods during postpartum due to their perceptions of foods, folk medicines, and health beliefs. Culturally sensitive programs relevant to postpartum diet practices for women should be implemented.

Bayesian hierarchical models for high-dimensional mediation analysis with coordinated selection of correlated mediators.

We consider Bayesian high-dimensional mediation analysis to identify among a large set of correlated potential mediators the active ones that mediate the effect from an exposure variable to an outcome of interest. Correlations among mediators are commonly observed in modern data analysis; examples include the activated voxels within connected regions in brain image data, regulatory signals driven by gene networks in genome data, and correlated exposure data from the same source. When correlations are present among active mediators, mediation analysis that fails to account for such correlation can be suboptimal and may lead to a loss of power in identifying active mediators. Building upon a recent high-dimensional mediation analysis framework, we propose two Bayesian hierarchical models, one with a Gaussian mixture prior that enables correlated mediator selection and the other with a Potts mixture prior that accounts for the correlation among active mediators in mediation analysis. We develop efficient sampling algorithms for both methods. Various simulations demonstrate that our methods enable effective identification of correlated active mediators, which could be missed by using existing methods that assume prior independence among active mediators. The proposed methods are applied to the LIFECODES birth cohort and the Multi-Ethnic Study of Atherosclerosis (MESA) and identified new active mediators with important biological implications.

Lifetime stress and war exposure timing may predict methylation changes at NR3C1 based on a pilot study in a warrior cohort in a small-scale society in Kenya.

OBJECTIVES: Candidate gene methylation studies of NR3C1 have identified associations with psychosocial adversity, including war trauma. This pilot study (sample sizes from 22 to 45 for primary analyses) examined NR3C1 methylation in a group of Kenyan pastoralist young men in relation to culturally relevant traumatic experiences, including participation in coalitional lethal gun violence. METHODS: Adolescent and young adult Samburu men ("warriors") were recruited for participation. DNA was obtained from whole saliva and methylation analyses performed using mass spectrometry. We performed a data reduction of variables from a standardized instrument of lifetime stress using a factor analysis and we assessed the association between the extracted factors with culturally relevant and cross-culturally comparative experiences. RESULTS: Cumulative lifetime trauma exposure and forms of violence to which warriors are particularly susceptible were associated with DNA methylation changes in the NR3C1 1F promoter region but not in the NR3C1 1D promoter region. However, sensitivity analyses revealed significant associations between individual CpG sites in both regions and cumulative stress exposures, war exposure timing, and war fatalities. CONCLUSIONS: This study supports the importance of NR3C1 methylation changes in response to challenging life circumstances, including in a global south cultural context that contrasts in notable ways from global north contexts and from the starkly tragic examples of the Rwandan genocide and war-associated rape explored in recent studies. Timing of traumatic exposure and culturally salient means to measure enduring symptoms of trauma remain important considerations for DNA methylation studies.

Evaluating gender bias in an eating disorder risk assessment questionnaire for athletes.

Due to stigma, eating disorders are under-researched, underdiagnosed and undertreated among men. This is particularly pertinent among athletes, as athletic goals are a major risk factor for disordered eating in men. This gender stereotype may be reinforced by eating disorder risk assessment tools that better reflect female symptoms. We examine an eating disorder risk assessment questionnaire in a population of athletes to assess both (1) gender bias in individual items and (2) gender differences on a cognitive and behavioral subscale, identified through a factor analysis. Controlling for eating disorder risk, we found that female gender significantly predicted high risk scores on four items; male gender significantly predicted high risk scores on four other items. We also found that women were more likely to score above the median on a cognitive subscale than men with the same level of eating disorder risk, while men were more likely to score above the median on a behavioral subscale. These results may be applied practically to allow eating disorder risk assessment tools to better capture eating disorder risk independent of gender. These methods may be applied to other questionnaires and other social identities, to expand the scope of eating disorder research and treatment.

Methods to Account for Uncertainty in Latent Class Assignments When Using Latent Classes as Predictors in Regression Models, with Application to Acculturation Strategy Measures.

Latent class models have become a popular means of summarizing survey questionnaires and other large sets of categorical variables. Often these classes are of primary interest to better understand complex patterns in data. Increasingly, these latent classes are reified into predictors of other outcomes of interests, treating the most likely class as the true class to which an individual belongs even though there is uncertainty in class membership. This uncertainty can be viewed as a form of measurement error in predictors, leading to bias in the estimates of the regression parameters associated with the latent classes. Despite this fact, there is very limited literature treating latent class predictors as measurement error models. Most applications ignore this issue and fit a two-stage model that treats the modal class prediction as truth. Here, we develop two approaches-one likelihood-based, the other Bayesian-to implement a joint model for latent class analysis and outcome prediction. We apply these methods to an analysis of how acculturation behaviors predict depression in South Asian immigrants to the United States. A simulation study gives guidance for when a two-stage model can be safely implemented and when the joint model may be required.

Association of Alcohol Consumption and Ideal Cardiovascular Health Among South Asians: The Mediators of Atherosclerosis in South Asians Living in America (MASALA) Study.

BACKGROUND: Observational studies have shown that alcohol consumption above the recommended limit is associated with increased cardiovascular disease (CVD), although its association in South Asians is unclear. Less is known regarding the association between alcohol consumption and cardiovascular health (CVH), assessed by the American Heart Association's Life's Simple 7 (LS7) health metrics among those with South Asian ancestry. METHODS: This analysis included 701 participants without CVD from the Mediators of Atherosclerosis in South Asians Living in America (MASALA) cohort (2015 to 2018). Based on a personal history questionnaire, participants were divided into never, former, and current drinkers. The current drinking category was further classified into 1 to 3 drinks/wk, 4 to 7 drinks/wk, and >7 drinks/wk. The consumption of 5 or more drinks on 1 occasion in the past month was defined as binge drinking. Each LS7 component was given a point score of 0, 1, or 2. The total score was categorized into 0 to 6, 7 to 10, and 11 to 14 to represent poor, intermediate, and ideal CVH, respectively. We use multinomial logistic regression to examine the association between alcohol consumption and CVH. RESULTS: In the MASALA cohort (mean age = 59 years, 43% female), participants consuming >7 drinks/wk had the lowest mean CVH score. Compared with never drinkers, male participants consuming >7 drinks/wk were less likely to have intermediate CVH (0.44 [0.08, 0.91]) and ideal CVH (0.23 [0.03, 0.96]). Binge drinking was associated with significantly lower odds of ideal CVH compared with never drinkers. CONCLUSION: We found evidence of an inverse association of moderate to heavy alcohol consumption and ideal CVH in South Asian men. These findings further underscore the important relationship between alcohol consumption and CVH in this unique population of South Asians.

Bayesian shrinkage estimation of high dimensional causal mediation effects in omics studies.

Causal mediation analysis aims to examine the role of a mediator or a group of mediators that lie in the pathway between an exposure and an outcome. Recent biomedical studies often involve a large number of potential mediators based on high-throughput technologies. Most of the current analytic methods focus on settings with one or a moderate number of potential mediators. With the expanding growth of -omics data, joint analysis of molecular-level genomics data with epidemiological data through mediation analysis is becoming more common. However, such joint analysis requires methods that can simultaneously accommodate high-dimensional mediators and that are currently lacking. To address this problem, we develop a Bayesian inference method using continuous shrinkage priors to extend previous causal mediation analysis techniques to a high-dimensional setting. Simulations demonstrate that our method improves the power of global mediation analysis compared to simpler alternatives and has decent performance to identify true nonnull contributions to the mediation effects of the pathway. The Bayesian method also helps us to understand the structure of the composite null cases for inactive mediators in the pathway. We applied our method to Multi-Ethnic Study of Atherosclerosis and identified DNA methylation regions that may actively mediate the effect of socioeconomic status on cardiometabolic outcomes.

Pathogen burden and leukocyte telomere length in the United States.

BACKGROUND: Prior studies in humans have suggested that telomere shortening may be accelerated by infection, but research on multiple pathogens and use of large population-based study samples has been limited. We estimated cross-sectional associations between seropositivity to five persistent pathogens (Herpes Simplex Virus Type-1 (HSV-1), Herpes Simplex Virus Type-2 (HSV-2), cytomegalovirus (CMV), Helicobacter pylori (H.pylori), and Hepatitis B) as well as total pathogen burden and leukocyte telomere length. Data were derived from the National Health and Nutrition Examination Survey (1999-2000) for individuals 20-49 years of age, N = 1708. We analyzed the influence of each pathogen separately, a pathogen count score and a latent class model of pathogen burden on log telomere length using linear regression models, adjusted for covariates. RESULTS: Individuals in a latent pathogen burden class characterized by high probabilities of infection with HSV-1, CMV, and H. pylori, had significantly decreased log telomere length (- 0.30 [95% CI: - 0.36, - 0.24]) compared to those in a latent class characterized by low probabilities of all five infections. There were limited significant associations using other pathogen measures. CONCLUSIONS: These results suggest that infection with specific combinations of pathogens may be one mechanism contributing to accelerated cellular senescence with possible origins early in the life course.

Child and Adult Socioeconomic Status and the Cortisol Response to Acute Stress: Evidence From the Multi-Ethnic Study of Atherosclerosis.

OBJECTIVE: A long-hypothesized pathway through which low socioeconomic status (SES) harms health is through dysregulation of the physiologic stress response systems. No previous studies have tested this hypothesis by investigating cortisol reactivity and recovery to acute stress in relation to SES at different times in the life course in adults. Alteration of the cortisol response to an acute stressor could signal dysregulation of the hypothalamic-pituitary-adrenal axis and has been associated with chronic illness. METHODS: We used data on 997 adults 54 years or older from a multiethnic, multisite United States study to examine associations between life course SES and cortisol response to a laboratory stress challenge. Informed by life course theory, we hypothesized that lower child and adult SES would be associated with lower reactivity (i.e., smaller increase in cortisol) and a slower recovery rate (i.e., slower rate of decline in cortisol after the challenge). RESULTS: In demographics-adjusted multilevel piecewise linear regression models, low child and adult SES were associated with a 19% (95% CI = 4%-50%) and 27% (7%-55%) slower recovery rate compared with high child and adult SES, respectively. Compared with participants with stable high SES, those with stable low SES had a 48% (16%-70%) slower recovery rate. Differences in reactivity by SES were small. CONCLUSIONS: Our results support the hypothesis that low SES throughout life affects the hypothalamic-pituitary-adrenal axis and in turn the ability to recover from exposure to acute stressors. This mechanism can help explain how socioeconomic disparities contribute to disparities in chronic disease.