Abstract
Unhealthy behaviors may modify relationships between chronic stress and depression among diverse older adults. We analyzed nationally representative cross-sectional data from participants aged 40–79 years of the 2005–2012 National Health and Nutrition Examination Survey. Unhealthy behaviors included current smoking, excessive/binge drinking, insufficient physical activity, and fair/poor diet. Allostatic load was defined by 10 biomarkers indicating the cumulative physiologic burden of stress. Depressive disorder was assessed using the Patient Health Questionnaire. Multivariable logistic regression examined whether current smoking, excessive/binge drinking, insufficient physical activitiy, and fair/poor diet modified relationships between allostatic load and depressive disorder. Mean age of 12,272 participants was 55.6 years (standard error = 0.19), 51.9% were women, and most had at least a high school education (81.8%). Latinos (11.3%) and African Americans (10.4%) were more likely than Whites (7.1%; p<0.001) to meet depressive disorder criteria. Allostatic load was not associated independently with depressive disorder in any racial/ethnic group and this lack of a relationship did not differ by the extent of unhealthy behaviors. Although Latinos and African Americans report higher levels of depression than Whites, physiological markers of stress do not appear to explain these differences.
Keywords: Stress, Allostatic load, Health behavior, Alcohol use, Cigarette smoking, Diet, Physical activity, Depression, African Americans, Hispanics/Latinos
INTRODUCTION
One of the most significant factors that has a longitudinal effect on depression is chronic stress during adulthood.1,2 This effect may be exacerbated among racial/ethnic minority adults, as their mental health may suffer due to stress experienced from inequalities in social and economic opportunities as well as environmental conditions.
The prevalence of major depression differs by race/ethnicity. While 4.2–5.6% of African Americans and 3.9–5.3% of Whites report experiencing major depression in the past 12 months, Latinos are affected by major depression at a higher rate (7.9–8.6%).3,4 Understanding of the effects of social, economic, and environmental stressors on racial/ethnic health disparities is growing, but remains limited.
The leading behavioral causes of preventable death and disability differ by race/ethnicity.5 African Americans and Whites are most likely to smoke.6 Obesity is more common among African Americans and Latinos than Whites.7 African Americans and Latinos are more likely to drink heavily or in a binge pattern, although alcohol use is most prevalent among Whites.8
These unhealthy behaviors may influence the relationship between chronic stress and depression.9,10 Relationships between substance use or consumption of unhealthy foods and chronic stress have been explained partially by physiological mechanisms.11–14 Also, chronic stress has been shown to negatively affect levels of physical activity.15
Jackson and colleagues observed that as chronic stress increased, African Americans who engaged in unhealthy behaviors partially avoided a future episode of major depression.9 Other research has shown that among Latino older adults, as chronic stress increased those who engaged in more unhealthy behaviors were more likely to experience significant depressive symptoms in the future.10 A theoretical framework, named the Environmental Affordances model, has been proposed as an explanation of how chronic stress and risk behaviors may interact to affect health.16 This framework postulates that engaging in unhealthy behaviors attenuates the effects of stress on depression among African Americans. However, variability in the effects of unhealthy behaviors by race/ethnicity warrants further investigation to understand the biological mechanisms that link chronic stress, unhealthy behaviors, and depression.
Allostatic load has been proposed as a measure of chronic stress using biological indicators and has been found to be associated with mental and physical well-being and all-cause mortality.17,18 Although allostatic load may vary by factors such as age, abdominal fat, and medication use, allostatic load stabilizes in later life and abdominal fat, as well as medication use, is accounted for in the final score.19,20 In the present study, we used data from the National Health and Nutrition Examination Survey (NHANES) to assess whether unhealthy behaviors moderated the relationship between allostatic load and being at-risk for depressive disorder by racial/ethnic group. We hypothesized that engaging in a greater number of unhealthy behaviors (a) attenuated the relationship between allostatic load and being at-risk for depressive disorder for African Americans compared with Whites and (b) amplified the relationship between allostatic load and being at-risk for depressive disorder for Latinos compared with Whites.
METHODS
National Health and Nutrition Examination Survey
Data on adult participants aged 40–79 years from four cross-sectional waves of NHANES, 2005–2012, were analyzed in 2016.21 Older adults were assessed in order to capture the effects of chronic stress over the life course on physiologic markers. NHANES collects data from a nationally representative sample employing a stratified, multistage probability design to assess the health and nutritional status of the U.S. population. Interview data were collected in participants’ homes while supplemental health data from physical examinations and blood samples for laboratory testing were collected in mobile examination centers. The number of persons surveyed each year from 2005 to 2012 ranged between 2,564 and 3,529. Because only publicly available data from NHANES were analyzed and are not considered to involve human subjects, our investigation did not necessitate review by an Institutional Review Board.
Demographic Characteristics
Age, gender, highest level of education (less than high school, high school graduate or equivalent, some college or an associate degree, or college graduate or higher), and race/ethnicity were assessed by self-report in all survey years. Racial/ethnic groups included participants who identified as African American or Black, Latino or Hispanic, and White.
Depressive Disorder
Depressive disorder was assessed using the Patient Health Questionnaire (PHQ-9).22 Items asked about the following problems experienced in the last two weeks: little interest or pleasure in doing things; feeling down, depressed, or hopeless; trouble falling or staying asleep – or sleeping too much; feeling tired or having little energy; poor appetite or overeating; feeling bad about yourself – or that you are a failure or have let yourself or your family down; trouble concentrating on things; moving or speaking so slowly that other people could have noticed – or the opposite, being so fidgety or restless that you have been moving around a lot more than usual; and thoughts that you would be better off dead or of hurting yourself in some way. Each PHQ-9 item was scored via ordered response options corresponding to how often the participant had been bothered by each problem during the prior two weeks: not at all (0), several days (1), more than half the days (2), and nearly every day (3). Scores were summed and an indicator variable representing being at-risk for depressive disorder was created with a cut point score of ≥ 10.23
Biomarkers of Stress and Allostatic Load
To determine allostatic load, we used ten biological indicators which represent cardiometabolic risk, glucose metabolism, cardiopulmonary functioning, parasympathetic functioning, and inflammation and have been used in past research using NHANES data.24 These biomarkers included systolic blood pressure, diastolic blood pressure, body mass index (BMI), glycohemoglobin, total cholesterol, high-density lipoprotein (HDL) cholesterol, total/HDL cholesterol ratio, C-reactive protein, albumin, and creatinine clearance.
Allostatic load was calculated as an index of physiologic dysregulation for each participant with each indicator defined as high-, moderate-, or low-risk categories. The following clinically-relevant cut-points were used, respectively: systolic blood pressure: ≥150 mmHg, 120 to <150 mmHg, and <120 mmHg; diastolic blood pressure: ≥90 mmHg, 80 to <90 mmHg, and<80 mmHg; body mass index (BMI):≥30 kg/m3, 25 to <30 kg/m3, and 18 to <25 kg/m3; glycohemoglobin:≥6.5%, 5.7% to <6.5%, and <5.7%; total cholesterol:≥240 mg/dL, 200 to <240 mg/dL, and <200 mg/dL; HDL cholesterol: <40 mg/dL, 40 to <60 mg/dL, and ≥60 mg/dL; total/HDL cholesterol ratio: ≥6, 5 to <6, and <5; C-reactive protein: ≥3 mg/L, 1 to <3 mg/L, and <1 mg/L; albumin: <3.0 μg/mL, 3.0 to <3.8 μg/mL, and ≥3.8 μg/mL; and creatinine clearance: <30 mL/min/1.73 m2, 30 to <60 mL/min/1.73 m2, and ≥60 mL/min/1.73 m2. The highest risk categories for HDL cholesterol, albumin, and creatinine clearance were the categories with the lowest values for each biomarker.
Allostatic load was calculated by assigning one point for the high-risk category, a half point for moderate-risk, and zero points for low-risk. Following previous research, a half point was added to the allostatic load score of participants who reported taking medication for hypertension, diabetes, and/or cholesterol and who had a low-risk value for blood pressure, glycohemoglobin, or lipids.20 The maximum possible allostatic load score was 10 points.
Unhealthy Behaviors
We assessed the following four unhealthy behaviors: (a) current smoking, (b) excessive and/or binge drinking, (c) not meeting recommendations for moderate or vigorous physical activity per week, and (d) fair or poor overall diet. Current smoking was identified by assessing serum cotinine, a major metabolite of nicotine. Methods for measuring serum cotinine, processing and editing of cotinine data, and laboratory quality assurance and monitoring are documented on the NHANES website.21 The following cut points for serum cotinine were used to determine current smoking: ≥5.92 ng/mL for African Americans, ≥0.84 ng/mL for Latinos, and ≥4.85 ng/mL for Whites.25
Participants were asked if they had at least 12 drinks of any type of alcoholic beverage, such as liquor, beer, wine, wine coolers, and others, in any one year or in their lifetime. Among those who responded ‘yes’, excessive drinking was determined using established criteria: (a) for men <65 years old: >2 drinks per day on average and (b) for all women as well as men ≥65 years old: >1 drink per day on average 26. Binge drinking was determined for participants who reported drinking ≥4 drinks on at least one day in the past 12 months.
Physical activity was assessed by self-reported minutes per week of moderate or vigorous activity. Participants who either reported <150 minutes of moderate or <75 minutes of vigorous physical activity were determined to not meet national recommendations.27 Overall diet quality was assessed by the question, In general, how healthy is your overall diet, and included the following response options: excellent, very good, good, fair, or poor.
Four dichotomous unhealthy behavior indicators were created: current smoking, excessive or binge drinking, not meeting physical activity recommendations, and ‘fair or poor’ diet (vs. ‘excellent/very good/good’). These indicators were summed to create an unhealthy behavior index (range: 0–4).
Statistical Analyses
Means and standard errors were estimated for allostatic load score and unhealthy behavior index, while frequencies were calculated for all other variables. Chi-square and t tests tested differences in sample characteristics by race/ethnicity. To test our hypothesis, regression models included interaction terms between mean allostatic load score and specific values of unhealthy behavior index using dummy variables.
Logistic and linear regression, stratified by race/ethnicity, modeled the relationships between allostatic load, unhealthy behaviors, and depressive disorder. Bivariate models were fit prior to the construction of multivariable models. Linear regression was only used to model the relationship between allostatic load and unhealthy behavior index. Multivariable logistic regression estimated the associations of allostatic load, unhealthy behaviors, and the interactions of allostatic load score and unhealthy behavior index with depressive disorder. Models were adjusted for age, gender, and education level and included participants with non-missing data for all variables. Survey data analysis procedures for means, percentages, and regression modeling were used to account for the complex stratified sampling and weighting procedures of NHANES.28 To increase the precision of estimates, all years of data (i.e., 2005–2012) were combined. Analyses were conducted using Stata, Version 14.0 (StataCorp LP, College Station, Texas).
RESULTS
Descriptive Characteristics
From 2005–2012, 12,272 participants were between the ages of 40 and 79 years (Table 1). Twenty-five percent of the sample self-reported as African American, 27% Latino, and 48% White. With a mean age of 52.6 years, Latino participants tended to be younger than African Americans (54.6 years) and Whites (56.2 years; p<0.001). A greater proportion of African Americans were women (55.6%) compared to the proportions among Latinos (51.4%) or Whites (51.5%; p=0.001). Having less than a high school education was significantly more common among Latinos (51.4%) than African Americans (26.7%) or Whites (12.3%) while a college degree or above was significantly more common among Whites (32.6%) than African Americans (17.4%) or Latinos (11.6%; p<0.001).
Table 1.
Demographics, Unhealthy Behaviors, and Depressive Disorder by Race/Ethnicity: 2005–2012 National Health and Nutrition Examination Survey, (n = 12,272)
| Characteristics | African American n = 3,081 |
Latino n = 3,288 |
White n = 5,903 |
p |
|---|---|---|---|---|
| Demographics | ||||
| Age, mean ± SE | 54.6 (0.2) | 52.6 (0.3) | 56.2 (0.2) | <0.001 |
| Age (categorical), %a | <0.001 | |||
| 40–59 years | 70.6 | 75.3 | 63.5 | |
| 60–79 years | 29.4 | 24.7 | 36.5 | |
| Women, %a | 55.6 | 51.4 | 51.5 | 0.001 |
| Education, %a | <0.001 | |||
| Less than high school | 26.7 | 51.4 | 12.3 | |
| High school graduate or equivalent | 25.2 | 16.9 | 25.3 | |
| Some college or associate degree | 30.7 | 20.1 | 29.8 | |
| College degree or higher | 17.4 | 11.6 | 32.6 | |
| Unhealthy behaviors | ||||
| Current smoking, %a,b | 32.8 | 20.3 | 24.7 | <0.001 |
| Excessivec or binged drinking in past 12 months, %a | 36.3 | 48.1 | 37.9 | <0.001 |
| Excessive drinking in past 12 monthsc | 34.1 | 45.6 | 35.5 | <0.001 |
| Binge drinking in past 12 monthsd | 4.3 | 4.6 | 4.5 | 0.963 |
| Did not meet either recommendation for moderate or vigorous physical activity per week, %a | 67.4 | 65.9 | 57.4 | <0.001 |
| Minutes moderate physical activity per week, mean ± SE | 233 (14.7) | 255 (16.0) | 268 (12.6) | 0.273 |
| Did not meet recommendation for 150 minutes of moderate physical activity per week, %a | 73.4 | 72.5 | 64.9 | <0.001 |
| Minutes vigorous physical activity per week, mean ± SE | 130 (10.6) | 155 (13.4) | 133 (9.2) | 0.344 |
| Did not meet recommendation for 75 minutes of vigorous physical activity per week, %a | 82.8 | 81.4 | 77.5 | 0.002 |
| Fair/poor overall diet, %a | 30.7 | 35.9 | 21.1 | <0.001 |
| Unhealthy behavior indexe, mean ± SE | 1.7 (0.03) | 1.7 (0.03) | 1.4 (0.02) | <0.001 |
| 0 unhealthy behaviors | 13.5 | 11.5 | 21.0 | <0.001 |
| 1 unhealthy behavior | 36.5 | 35.3 | 37.3 | |
| 2 unhealthy behaviors | 24.3 | 29.2 | 25.3 | |
| 3 unhealthy behaviors | 19.7 | 19.5 | 13.4 | |
| 4 unhealthy behaviors | 6.0 | 4.5 | 3.0 | |
| Depressive disorderf | ||||
| PHQ-9 Score, mean ± SE | 3.4 (0.14) | 3.5 (0.14) | 2.9 (0.09) | 0.001 |
| PHQ-9 Score ≥ 10, %a | 10.4 | 11.3 | 7.1 | <0.001 |
Note. SE = standard error.
Percentages based on non-missing values.
Based on serum cotinine level ≥ 5.92 ng/mL for African Americans, ≥ 0.84 ng/mL for Latinos, and ≥ 4.85 ng/mL for Whites.
For men <65 years old: >2 drinks per day on average; for women and men ≥ 65 years old: >1 drink per day on average.
Based on ≥ 4 drinks per day on at least one day in the past 12 months.
Respondents were assigned 1 point for current smoking, excessive/binge drinking, not meeting either physical activity recommendation, and fair/poor overall diet; range: 0–4.
Based on a score of 10 or greater on the 9-item Patient Health Questionnaire (PHQ-9).
Unhealthy Behaviors
Among the unhealthy behaviors, the prevalence of current smoking was highest among African Americans (32.8%) and lowest among Latinos (20.3%; p<0.001). Greater proportions of Latinos (48.1%) reported excessive or binge drinking compared to African Americans (36.3%) and Whites (37.9%; p<0.001). Compared to Whites, both African Americans and Latinos reported higher proportions of not meeting recommendations for moderate or vigorous physical activity per week (57.4% compared to 67.4% and 65.9%, respectively; p<0.001) and engaging in zero minutes of moderate or vigorous physical activity per week (46.3% compared to 57.0% and 56.8%, respectively; p<0.001). Larger proportions of African Americans (30.7%) and Latinos (35.9%), compared to Whites (21.1%; p<0.001), reported having a fair or poor overall diet. With respect to the unhealthy behavior index, higher proportions of African Americans (25.7%) and Latinos (24.0%) engaged in 3 or 4 unhealthy behaviors than Whites (16.4%; p<0.001). More Whites reported no unhealthy behaviors (21%) compared to African Americans (13.5%) and Latinos (11.5%).
Depressive Disorder
Being at-risk for depressive disorder, as determined by a PHQ-9 score of 10 or greater, differed by race/ethnicity. Accordingly, 10.4% of African Americans and 11.3% of Latinos were at-risk for depressive disorder compared to 7.1% of Whites (p<0.001).
Biomarkers of Stress and Allostatic Load
Compared to Latinos and Whites, African Americans had the largest proportion in the high-risk group on six of ten biomarkers of stress (Table 2): systolic blood pressure (p<0.001), diastolic blood pressure (p<0.001), BMI (p<0.001), glycohemoglobin (p<0.001), C-reactive protein (p<0.001), and creatinine clearance (p<0.001). Latinos had the largest proportion in the highest risk group for HDL cholesterol (p<0.001) and total/HDL cholesterol ratio (p<0.001). Greater proportions of both Latinos and Whites compared to African Americans were in the highest risk group for total cholesterol (p<0.001) and albumin (15.0% and 21.6% compared to 9.5%, respectively; p<0.001). African American and Latinos had higher mean allostatic load scores compared to Whites (3.3 and 3.3 compared to 3.1, respectively; p<0.001).
Table 2.
Biomarkers of Stress and Allostatic Load Score by Race/Ethnicity: 2005–2012 National Health and Nutrition Examination Survey, (n = 12,272)
| African American n = 3,081 |
Latino n = 3,288 |
White n = 5,903 |
p | |
|---|---|---|---|---|
| Characteristics | ||||
| Systolic blood pressure (mmHg), %a | <0.001 | |||
| Lowest risk (<120) | 21.9 | 38.0 | 33.3 | |
| Moderate risk (120–149) | 62.6 | 53.0 | 59.4 | |
| Highest risk (≥150) | 15.5 | 9.1 | 7.3 | |
| Diastolic blood pressure (mmHg), %a | <0.001 | |||
| Lowest risk (<80) | 67.7 | 75.5 | 76.3 | |
| Moderate risk (80–89) | 21.3 | 19.3 | 18.3 | |
| Highest risk (≥90) | 11.0 | 5.2 | 5.4 | |
| Body mass index (kg/m2), %a | <0.001 | |||
| Lowest risk (<25) | 20.9 | 17.0 | 27.1 | |
| Moderate risk (25–29) | 30.3 | 40.6 | 35.5 | |
| Highest risk (≥30) | 48.8 | 42.4 | 37.4 | |
| Glycohemoglobin (HbA1c), %a | <0.001 | |||
| Lowest risk (≤5.6) | 42.3 | 51.0 | 64.6 | |
| Moderate risk (5.7–6.4) | 39.4 | 31.7 | 27.3 | |
| Highest risk (≥6.5) | 18.3 | 17.3 | 8.1 | |
| C-reactive protein (mg/L), %a | <0.001 | |||
| Lowest risk (<1) | 22.1 | 23.5 | 29.9 | |
| Moderate risk (1–3) | 30.5 | 37.3 | 35.5 | |
| Highest risk (>3) | 47.4 | 39.1 | 34.6 | |
| Creatinine clearance (mL/min), %a | <0.001 | |||
| Lowest risk (≥60) | 90.8 | 95.7 | 92.7 | |
| Moderate risk (30–59) | 7.6 | 3.8 | 7.0 | |
| Highest risk (<30) | 1.5 | 0.4 | 0.3 | |
| HDL cholesterol (mg/dL), %a | <0.001 | |||
| Lowest risk (≥60) | 36.5 | 22.2 | 31.7 | |
| Moderate risk (40–59) | 51.3 | 54.1 | 49.8 | |
| Highest risk (<40) | 12.2 | 23.7 | 18.5 | |
| Total/HDL cholesterol ratio, %a | <0.001 | |||
| Lowest risk (<5) | 86.4 | 71.3 | 78.6 | |
| Moderate risk (5 to <6) | 8.9 | 16.2 | 11.6 | |
| Highest risk (≥6) | 4.7 | 12.5 | 9.8 | |
| Total cholesterol (mg/dL), %a | <0.001 | |||
| Lowest risk (<200) | 39.7 | 34.6 | 31.6 | |
| Moderate risk (200–239) | 46.6 | 46.6 | 50.1 | |
| Highest risk (≥240) | 13.7 | 18.8 | 18.3 | |
| Albumin (mg/L), %a | <0.001 | |||
| Lowest risk (≥3.8) | 85.3 | 78.8 | 71.4 | |
| Moderate risk (3 –<3.8) | 5.2 | 6.2 | 7.0 | |
| Highest risk (<3) | 9.5 | 15.0 | 21.6 | |
| Allostatic load scoreb, mean ± SE | 3.3 (0.03) | 3.3 (0.05) | 3.1 (0.03) | <0.001 |
Note. HDL = high-density lipoprotein; SE = standard error.
Percentages based on non-missing value.
Respondents were assigned 0.5 points for each biomarker categorized as moderate risk and 1 point for each biomarker categorized as high risk; range: 1–10.
Associations between Allostatic Load and Unhealthy Behaviors
Bivariate analyses revealed that allostatic load was associated with unhealthy behaviors, but relationships differed by the individual behavior and/or race/ethnicity (Table 3). Among African Americans, greater allostatic load was significantly associated with decreased odds of current smoking (adjusted odds ratio [aOR] = 0.89, 95% confidence interval [CI] = [0.83,0.96]), decreased odds of excessive or binge drinking in the past 12 months (aOR = 0.88, 95% CI = [0.82, 0.95]), and increased odds of fair or poor overall diet (aOR = 1.13 95% CI = [1.06, 1.18]).
Table 3.
Associations Between Allostatic Loada and Unhealthy Behaviors by Race/Ethnicity: 2005–2012 National Health and Nutrition Examination Survey
| Racial/Ethnic Group | Current smokingb aOR (95% CI)d | Excessive or binge drinking in past 12 months aOR (95% CI)d | Did not meet either recommendation for moderate or vigorous physical activity per week aOR (95% CI)d | Fair/poor overall diet aOR (95% CI)d | Unhealthy behavior indexc Estimate (95% CI)de |
|---|---|---|---|---|---|
| African American | 0.89(0.83, 0.96) | 0.88(0.82, 0.95) | 1.04 (0.97, 1.12) | 1.13(1.06, 1.18) | 0.007(−0.02, 0.04) |
| Latino | 0.97(0.89, 1.06) | 0.92(0.42, 0.99) | 1.12 (1.05, 1.19) | 1.17(1.11, 1.24) | 0.05(0.02, 0.08) |
| White | 0.96(0.92, 1.01) | 0.91(0.86, 0.95) | 1.13 (1.07, 1.18) | 1.21(1.15, 1.26) | 0.04(0.02, 0.06) |
Note. aOR = adjusted odds ratio; CI = confidence interval.
Allostatic load included systolic and diastolic blood pressure, body mass index, glycohemoglobin, total and high-density lipoprotein cholesterol (HDL), total/HDL cholesterol ratio, C-reactive protein, albumin, and creatinine clearance.
Based on serum cotinine level ≥ 5.92 ng/mL for African Americans, ≥ 0.84 ng/mL for Latinos, and ≥ 4.85 ng/mL for Whites.
Unhealthy behavior index included current smoking, excessive or binge drinking in the past 12 months, did not meet either recommendation for moderate or vigorous physical activity per week, and fair/poor overall diet. Scores were summed such that one point was given for the presence of each unhealthy behavior.
Adjusted for age, gender, and education.
Modeled using linear regression.
Among Latinos and Whites, greater allostatic load was significantly associated with decreased odds of excessive or binge drinking in the past 12 months (aOR = 0.92, 95% CI = [0.42, 0.99] and aOR = 0.91, 95% CI = [0.86, 0.95], respectively), increased odds of not meeting recommendations for moderate or vigorous physical activity per week (aOR = 1.12, 95% CI = [1.05, 1.19] and aOR = 1.13, 95% CI = [1.07, 1.18], respectively), increased odds of fair or poor overall diet (aOR = 1.17, 95% CI = [1.11, 1.24] and aOR = 1.21, 95% CI = [1.15, 1.26], respectively), and a higher unhealthy behavior index (aOR = 0.05, 95% CI = [0.02, 0.08] and aOR = 0.04, 95% CI = [0.02, 0.06], respectively).
Associations with Depressive Disorder
Multivariable logistic regression estimated the associations between 1) allostatic load and being at-risk for depressive disorder (Table 4), 2) unhealthy behavior index and being at-risk for depressive disorder (Table 4), and 3) the interactive relationships between allostatic load score and unhealthy behavior index on being at-risk for depressive disorder (not presented). Among models which included interaction terms, no statistically significant interactions were observed between allostatic load and any value of the unhealthy behavior index for any of the three racial/ethnic groups.
Table 4.
Multivariable Logistic Regression Results for Depressive Disordera by Race/Ethnicity: 2005–2012 National Health and Nutrition Examination Survey
| Independent Variables | African American aOR (95% CI)b |
Latino aOR (95% CI)b |
White aOR (95% CI)b |
|---|---|---|---|
| Main variables | |||
| Allostatic loadc | 1.35 (0.91, 2.02) | 0.97 (0.55, 1.72) | 1.10 (0.80, 1.51) |
| Unhealthy behavior indexd | |||
| 1 unhealthy behavior | 1.38 (0.21, 9.07) | 1.20 (0.15, 9.58) | 1.28 (0.37, 4.37) |
| 2 unhealthy behaviors | 5.76 (1.05, 31.64) | 2.03 (0.27, 15.36) | 1.43 (0.51, 4.00) |
| 3 unhealthy behaviors | 11.56 (1.85, 72.08) | 5.31 (0.57, 49.72) | 4.03 (1.48, 10.96) |
| 4 unhealthy behaviors | 17.39 (1.85, 135.68) | 3.47 (0.22, 54.30) | 13.21 (3.12, 56.00) |
| Covariates | |||
| Age (continuous) | 0.98 (0.97, 1.00) | 1.00 (0.98, 1.01) | 0.98 (0.97, 0.99) |
| Gender | |||
| Men | Reference | Reference | Reference |
| Women | 1.99 (1.53, 2.60) | 2.60 (2.03, 3.32) | 1.78 (1.42, 2.25) |
| Education | |||
| Less than high school | Reference | Reference | Reference |
| High school graduate or equivalent | 0.64 (0.45, 0.90) | 0.94 (0.63, 1.39) | 0.66 (0.47, 0.93) |
| Some college or associate degree | 0.63 (0.42, 0.94) | 0.85 (0.56, 1.30) | 0.55 (0.37, 0.81) |
| College degree or higher | 0.60 (0.37, 0.99) | 0.48 (0.26, 0.88) | 0.30 (0.19, 0.48) |
Note. aOR = adjusted odds ratio; CI = confidence interval.
Based on a score of 10 or greater on the 9-item Patient Health Questionnaire (PHQ-9).
Adjusted for age, gender, and education.
Allostatic load included systolic and diastolic blood pressure, body mass index, glycohemoglobin, total and high-density lipoprotein cholesterol (HDL), total/HDL cholesterol ratio, C-reactive protein, albumin, and creatinine clearance.
Unhealthy behavior index included current smoking, excessive or binge drinking in the past 12 months, did not meet either recommendation for moderate or vigorous physical activity per week, and fair/poor overall diet. Scores were summed such that one point was given for the presence of each unhealthy behavior.
Greater allostatic load was not associated with being at-risk for depressive disorder in any of the three racial/ethnic groups, controlling for covariates. Among African Americans, engaging in two, three, or four unhealthy behaviors was associated with increased odds of being at-risk for depressive disorder (aOR = 5.76, 95% CI = [1.05, 31.64], aOR = 11.56, 95% CI = [1.85, 72.08], and aOR = 17.39, 95% CI = [1.85, 135.68], respectively). A similar finding was observed among Whites who engaged in three or four unhealthy behaviors (aOR = 4.03, 95% CI = [1.48, 10.96] and aOR = 13.21, 95% CI = [3.12, 56.00], respectively). Older and more educated participants tended to have a lower odds of being at-risk for depressive disorder but this association was not observed among Latinos. Women had a significantly higher odds of being at-risk for depressive disorder in all three race/ethnic groups (Table 4).
DISCUSSION
We found that African Americans and Latinos were more likely than Whites to report being at-risk for depressive disorder as well as higher levels of allostatic load and unhealthy behaviors. However, there was no significant association between allostatic load and being at-risk for depressive disorder in any of the three racial/ethnic groups. Engaging in unhealthy behaviors did not modify these results for any of the racial/ethnic groups.
Our finding that chronic stress, as measured by allostatic load, was not significantly and independently associated with being at-risk for depressive disorder differs from other literature that links chronic stress by self-report and depression or depressive symptoms.29 Our use of a physiologic index of chronic stress rather than self-report may explain these conflicting results. It could be that the subjective experience of chronic stress affects depression or depressive symptoms differently than the physiological impact.
In our study, even though engaging in unhealthy behaviors did not modify associations between allostatic load and being at-risk for depressive disorder, engaging in a higher number of unhealthy behaviors was strongly associated with being at-risk for depressive disorder in both African Americans and Whites.
There is conflicting evidence as to whether African Americans engage in unhealthy behaviors to cope with chronic stress.30,31 The Environmental Affordances model combines neurobiological research on stress and health behavior with sociological theory to predict how individual- and societal-level factors interact to influence health.16 Evidence suggests that unhealthy behaviors may blunt the effects of chronic stress along the pathway to some mental disorders.11–13 One study observed that at higher levels of self-reported chronic stress, African Americans who engaged in unhealthy behaviors were less likely, compared to those who did not engage in those behaviors, to experience a future episode of depression.9 Among the proposed mechanisms that may explain unhealthy coping behavior include the impact of stress on increased sensitivity to substances, the reinforcement of excessive ethanol consumption in response to stress, the inhibition of a metabolic feedback signal by stress and subsequent increased desire for sugary foods, and the effect of stress on increasing physical inactivity and possibly sedentary behavior.11–13,15
In our study, the Environmental Affordances model was not supported for any of the racial/ethnic groups analyzed. Our results also differed from prior studies that found unhealthy behaviors interacted with chronic stress to increase depressive symptoms among Latinos overall, specifically foreign-born Mexicans.10,32 Other studies have failed to find signficant modification effects for unhealthy behaviors on the relationship between chronic stress and depression.33,34 Inconsistency in findings could be due to differences in study designs (multiple cross-sectional samples compared to a longitudinal design) or measures of chronic stress (subjective versus physiologic). Additionally, depression and allostatic load are known to be associated with early life experiences and cardiovascular biomarkers, as well as socioeconomic, psychosocial, and environmental factors.19 These relationships reflect the complex web of factors related to depression.
Allostatic load was inversely related to excessive or binge drinking among all racial/ethnic groups and with current smoking among African Americans only. Compared to previous research which has found that heavy alcohol use was associated with higher allostatic load, our findings with respect to excessive or binge drinking are not intuitive and may need further study.35 Research between allostatic load and current smoking is sparse. By contrast, having a fair or poor diet and lacking regular physical activity were generally associated with a higher allostatic load. The expected associations with fair or poor diet and lacking regular physical activity but unexpected associations with excessive or binge drinking and current smoking may be related to the close relationships between fair or poor diet and lacking regular physical activity and a few of the allostatic load biomarkers. Additionally, a gradient was observed whereby having at least a high school degree or equivalent was protective of depressive disorder for all groups. This finding supports previous research that higher educational attainment is more protective for disadvantaged groups and supports the promotion of more education as a population-level intervention to improve health.36
Study Limitations
NHANES is a cross-sectional survey that does not follow participants over time, therefore we could not determine the temporal order of the factors or evaluate causality. Although allostatic load and current smoking were assessed with physiologic measures of chronic stress, the remaining predictors in NHANES were assessed by self-report, which may affect their accuracy. Future research distinguishing between Latinos that are foreign and U.S. born as well as taking account country of ancestry is warranted. Finally, allostatic load was designed as a physiologic measure of chronic stress, but may not fully capture the experience and full impact of chronic stress related to health.
CONCLUSIONS
In conclusion, findings from the present study did not support previous research that has suggested that African Americans may engage in unhealthy behaviors to cope with chronic stress and reduce their chances of depression. Instead, they challenge our current knowledge of these relationships among racial/ethnic minorities. To build upon existing knowledge, we investigated these relationships using 1) an objective measure of the cumulative physiologic toll of stress on the body with clinically-relevant cut-points for each biomarker, 2) four types of unhealthy behaviors, and 3) a clinically-relevant measure of depression. Although Latinos and African Americans in this study had worse levels of allostatic load, more unhealthy behaviors, and higher levels of being at-risk for depressive disorder, we did not find an independent relationships betweem allostatic load and being at-risk for depressive disorder in any racial/ethnic group. Our findings suggest that factors other than allostatic load may help explain varying levels of depression in racial/ethnic groups and that unhealthy behaviors may or may not modify these relationships depending on the types of chronic stress measures used and the timing of exposures. Results point to the complexity of the relationships between chronic stress, unhealthy behaviors, and mental health among diverse racial/ethnic groups.
Highlights.
Racial/ethnic minorities were more likely than Whites to report depressive disorder
Allostatic load was higher among African Americans and Latinos compared to Whites
At least a high school degree or equivalent was protective of depressive disorder
Allostatic load was not associated with depressive disorder
The lack of relationship was not modified by unhealthy behaviors
Acknowledgments
We express gratitude for the advice we received from faculty and scholars of the Center for Aging in Diverse Communities. This study began at the University of California, San Francisco and was supported in part by grants from the Resource Centers for Minority Aging Research of the National Institute on Aging, National Institutes of Health [Center for Aging in Diverse Communities, grant number P30 AG15272]; the Clinical and Translational Science Institute at the University of California, San Francisco [Strategic Opportunities Support, grant number K19683L]; and the National Latino Cancer Research Network of the National Cancer Institute, National Institutes of Health [Redes En Acción, grant numbers U01 CA86117, U54 CA153511]. This study was also supported by the Divisions of Intramural Research of the National Heart, Lung, and Blood Institute and the National Institute of Minority Health and Health Disparities, National Institutes of Health [grant number not available]. None of the funders were involved in the study design; collection, analysis, or interpretation of data; writing of the manuscript; or decision to submit the manuscript for publication.
Footnotes
Conflict of Interest: The authors declare there is no conflict of interest.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
REFERENCES CITED
- 1.Hammen C. Stress and depression. Annu Rev Clin Psychol. 2005;1:293–319. doi: 10.1146/annurev.clinpsy.1.102803.143938. [DOI] [PubMed] [Google Scholar]
- 2.Paykel ES. Life events and affective disorders. Acta Psychiatr Scand Suppl. 2003;(418):61–66. doi: 10.1034/j.1600-0447.108.s418.13.x. [DOI] [PubMed] [Google Scholar]
- 3.González HM, Tarraf W, Whitfield KE, Vega WA. The epidemiology of major depression and ethnicity in the United States. Journal of psychiatric research. 2010;44(15):1043–1051. doi: 10.1016/j.jpsychires.2010.03.017. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Jimenez DE, Alegria M, Chen CN, Chan D, Laderman M. Prevalence of psychiatric illnesses in older ethnic minority adults. J Am Geriatr Soc. 2010;58(2):256–264. doi: 10.1111/j.1532-5415.2009.02685.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Mokdad AH, Marks JS, Stroup DF, Gerberding JL. Actual causes of death in the United States, 2000. JAMA. 2004;291(10):1238–1245. doi: 10.1001/jama.291.10.1238291/10/1238. [DOI] [PubMed] [Google Scholar]
- 6.Jamal A, King BA, Neff LJ, Whitmill J, Babb SD, Graffunder CM. Current Cigarette Smoking Among Adults - United States, 2005–2015. MMWR Morb Mortal Wkly Rep. 2016;65(44):1205–1211. doi: 10.15585/mmwr.mm6544a2. [DOI] [PubMed] [Google Scholar]
- 7.Pan L, Galuska DA, Sherry B, et al. Differences in prevalence of obesity among black, white, and Hispanic adults - United States, 2006–2008. MMWR Morb Mortal Wkly Rep. 2009;58(27):740–744. [PubMed] [Google Scholar]
- 8.Chartier K, Caetano R. Ethnicity and health disparities in alcohol research. Alcohol Res Health. 2010;33(1–2):152–160. [PMC free article] [PubMed] [Google Scholar]
- 9.Jackson JS, Knight KM, Rafferty JA. Race and unhealthy behaviors: chronic stress, the HPA axis, and physical and mental health disparities over the life course. Am J Public Health. 2010;100(5):933–939. doi: 10.2105/AJPH.2008.143446. AJPH.2008.143446. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Rodriquez EJ, Gregorich SE, Livaudais-Toman J, Perez-Stable EJ. Coping With Chronic Stress by Unhealthy Behaviors: A Re-Evaluation Among Older Adults by Race/Ethnicity. J Aging Health. 2016 doi: 10.1177/0898264316645548. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Piazza PV, Le Moal M. The role of stress in drug self-administration. Trends in pharmacological sciences. 1998;19(2):67–74. doi: 10.1016/s0165-6147(97)01115-2. [DOI] [PubMed] [Google Scholar]
- 12.Koob GF, Roberts AJ, Schulteis G, et al. Neurocircuitry targets in ethanol reward and dependence. Alcohol Clin Exp Res. 1998;22(1):3–9. [PubMed] [Google Scholar]
- 13.Dallman MF, Akana SF, Laugero KD, et al. A spoonful of sugar: feedback signals of energy stores and corticosterone regulate responses to chronic stress. Physiol Behav. 2003;79(1):3–12. doi: 10.1016/s0031-9384(03)00100-8. [DOI] [PubMed] [Google Scholar]
- 14.Wadsworth ME. Development of Maladaptive Coping: A Functional Adaptation to Chronic, Uncontrollable Stress. Child Dev Perspect. 2015;9(2):96–100. doi: 10.1111/cdep.12112. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Lutz RS, Lochbaum MR, Lanning B, Stinson LG, Brewer R. Cross-lagged relationships among leisure-time exercise and perceived stress in blue-collar workers. J Sport Exerc Psychol. 2007;29(6):687–705. doi: 10.1123/jsep.29.6.687. [DOI] [PubMed] [Google Scholar]
- 16.Mezuk B, Abdou CM, Hudson D, et al. “White Box” epidemiology and the social neuroscience of health behaviors: The Environment Affordances Model. Soc Ment Health. 2013;3(2) doi: 10.1177/2156869313480892. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Seeman TE, Singer BH, Rowe JW, Horwitz RI, McEwen BS. Price of adaptation--allostatic load and its health consequences. MacArthur studies of successful aging. Arch Intern Med. 1997;157(19):2259–2268. [PubMed] [Google Scholar]
- 18.Schulz AJ, Mentz G, Lachance L, Johnson J, Gaines C, Israel BA. Associations between socioeconomic status and allostatic load: effects of neighborhood poverty and tests of mediating pathways. Am J Public Health. 2012;102(9):1706–1714. doi: 10.2105/AJPH.2011.300412. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.McEwen BS. Allostasis and allostatic load: implications for neuropsychopharmacology. Neuropsychopharmacology: official publication of the American College of Neuropsychopharmacology. 2000;22(2):108–124. doi: 10.1016/S0893-133X(99)00129-3. [DOI] [PubMed] [Google Scholar]
- 20.Geronimus AT, Hicken M, Keene D, Bound J. “Weathering” and age patterns of allostatic load scores among blacks and whites in the United States. Am J Public Health. 2006;96(5):826–833. doi: 10.2105/AJPH.2004.060749. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.[dataset] Centers for Disease Control and Prevention (CDC). National Center for Health Statistics (NCHS) [Accessed August 28, 2015];National Health and Nutrition Examination Survey Data. http://www.cdc.gov/nchs/nhanes/nhanes_questionnaires.htm.
- 22.Kroenke K, Spitzer RL, Williams JB. The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med. 2001;16(9):606–613. doi: 10.1046/j.1525-1497.2001.016009606.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Manea L, Gilbody S, McMillan D. Optimal cut-off score for diagnosing depression with the Patient Health Questionnaire (PHQ-9): a meta-analysis. CMAJ. 2012;184(3):E191–196. doi: 10.1503/cmaj.110829. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Beckie TM. A systematic review of allostatic load, health, and health disparities. Biol Res Nurs. 2012;14(4):311–346. doi: 10.1177/1099800412455688. [DOI] [PubMed] [Google Scholar]
- 25.Benowitz NL, Bernert JT, Caraballo RS, Holiday DB, Wang J. Optimal serum cotinine levels for distinguishing cigarette smokers and nonsmokers within different racial/ethnic groups in the United States between 1999 and 2004. Am J Epidemiol. 2009;169(2):236–248. doi: 10.1093/aje/kwn301. [DOI] [PubMed] [Google Scholar]
- 26.National Institute on Alcohol Abuse and Alcoholism (NIAAA) [Accessed December 21, 2015];How To Screen For Heavy Drinking. http://pubs.niaaa.nih.gov/publications/Practitioner/PocketGuide/pocket_guide5.htm.
- 27.U.S. Department of Health and Human Services. 2008 Physical Activity Guidelines for Americans. Washington, DC: Secretary of Health and Human Services; Oct, 2008. [Google Scholar]
- 28.Division of Health and Nutrition Examination Surveys. National Center for Health Statistics (NCHS) National Health and Nutrition Examination Survey: Analytic guidelines, 2011–2012. Hyattsville, MD: 2013. [Google Scholar]
- 29.Tennant C. Life events, stress and depression: a review of recent findings. Aust N Z J Psychiatry. 2002;36(2):173–182. doi: 10.1046/j.1440-1614.2002.01007.x. [DOI] [PubMed] [Google Scholar]
- 30.Bates LM, Barnes DM, Keyes KM. Re: “Reconsidering the role of social disadvantage in physical and mental health: stressful life events, health behaviors, race, and depression”. Am J Epidemiol. 2011;173(11):1343–1351. doi: 10.1093/aje/kwr110. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Boardman JD, Alexander KB. Stress trajectories, health behaviors, and the mental health of black and white young adults. Soc Sci Med. 2011;72:1659–1666. doi: 10.1016/j.socscimed.2011.03.024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Lipton R. The relationship betweel alcohol, stress, and depression in Mexican Americans and Non-Hispanic Whites. Behav Med. 1997;23(3):101–111. doi: 10.1080/08964289709596366. [DOI] [PubMed] [Google Scholar]
- 33.Walsh JL, Senn TE, Carey MP. Longitudinal associations between health behaviors and mental health in low-income adults. Transl Behav Med. 2013;3:104–113. doi: 10.1007/s13142-012-0189-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Keyes KM, Barnes DM, Bates LM. Stress, coping, and depression: Testing a new hypothesis in a prospectively studied general population sample of U.S.-born Whites and Blacks. Soc Sci Med. 2011;72:650–659. doi: 10.1016/j.socscimed.2010.12.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Petrovic D, Pivin E, Ponte B, et al. Sociodemographic, behavioral and genetic determinants of allostatic load in a Swiss population-based study. Psychoneuroendocrinology. 2016;67:76–85. doi: 10.1016/j.psyneuen.2016.02.003. [DOI] [PubMed] [Google Scholar]
- 36.Bauldry S. Variation in the Protective Effect of Higher Education Against Depression. Soc Ment Health. 2015;5(2):145–161. doi: 10.1177/2156869314564399. [DOI] [PMC free article] [PubMed] [Google Scholar]