Methods

Study participants

The Pitt County Study is a community-based, representative prospective cohort study of blacks in eastern North Carolina; participants were 25–50 years old in 1988, the baseline year. The objective of the Pitt County Study was to elucidate psychosocial and behavioral risk factors for hypertension and related disorders in blacks (21). During the study's baseline year (1988), 1,773 male and female residents of Pitt County, North Carolina were interviewed. The baseline interview was conducted in participants' homes by a trained interviewer. The content of the interview included questions about demographic factors, health history, psychosocial factors, diet, and exercise practices. Physical measurements such as blood pressure, height, and weight were also taken.

A follow-up interview was sought in 2001 with all living, noninstitutionalized members of the cohort residing within a 100 mile radius of Greenville, the principal city in Pitt County. Of the 1,540 participants meeting these criteria, follow-up data were available for 1,173 (76%). During the follow-up interview, participants' height and weight were again measured. Additional details of the follow-up study design are published elsewhere (22).

Psychosocial stress

Psychosocial stress, the main exposure of interest, was measured using a shortened form of the Perceived Stress Scale (23). This scale is one of the most common measures of subjective stress in psychosocial health research; its psychometric properties have been well described (24,25). Participants were asked eight questions, including those related to whether they have been: worried about a problem, upset by something unexpected, feeling worried or stressed, able to control important aspects of life, or feeling as though things were going their way. Four response options were provided: 1 = Never, 2 = Sometimes, 3 = Most of the time, 4 = Always. The psychosocial stress score equaled the sum (possible range 8–32) of responses to the eight questions. Scores were then classified into tertiles.

Outcomes

The study outcome was percent change in BMI (from baseline) over the 13-year follow-up period. Measured height and weight were used to calculate BMI at baseline and follow-up. Percent BMI change was calculated as baseline BMI subtracted from BMI at follow-up, divided by baseline BMI, multiplied by 100.

Covariates

Covariates measured at baseline included BMI, age, smoking (current, nonsmoker), education (<high school, high school graduation, >high school) and occupation (laborer, blue collar, white collar), and self-report of financial strain (difficulty paying for basics, yes or no).

Statistical analysis

Participants' baseline characteristics were characterized using means and standard deviations for continuous variables and percents for categorical variables. Multivariable linear regression was used to model the relationship between perceived stress and weight change. Because factors affecting weight change could differ between men and women, we stratified all models by sex. We used interaction terms to assess effect modification of the perceived stress/body weight relationship by educational level. The analyses were weighted to account for the oversampling of middle-class households at baseline and nonresponse at baseline and follow-up. Each participant was assigned a sampling weight equal to the inverse of its selection probability for inclusion in the sample. The final sampling weights accounted for the unequal probability of being selected due to the oversampling of middle-class households during the baseline survey in 1988, as well as nonresponse to both the 1988 and 2001 surveys. All analyses were conducted using SAS statistical software (SAS Institute, Cary, NC) and STATA 9 statistical software (StataCorp, College Station, TX).

This study was submitted for review and approved by the Institutional Review Boards of Duke University and the University of North Carolina at Chapel Hill, and exempted by the Committee on Clinical Investigation, Beth Israel Deaconess Medical Center, Boston.

Results

There were 757 women and 416 men who had both baseline and follow-up BMI data (Table 1). The average age of participants at baseline was 36 (s.d. = 7) years; the majority (61%) were nonsmokers. Average baseline BMI was 29.2 kg/m² (s.d. = 6.8) for women and 26.2 kg/m² (s.d. = 4.5) for men. Women also had significantly higher psychosocial stress scores (17.2, 95% CI 16.9–17.5) compared to men (15.9, 95% CI 15.6–16.3). Relative BMI increased more among women (14.8%, s.d. = 18.2) than men (10.0%, s.d. = 14.4) over the 13-year follow-up period.

Table 1 - Baseline characteristics of the study sample.

Full table (28K)

Women reporting medium or high levels of stress, relative to low levels, had greater relative increases in percentage BMI over time (Table 2). The mean adjusted percentage increase in BMI, calculated from the regression model was 12.0% (95% CI 9.6–14.4) for women with low baseline stress; 16.3% (95% CI 13.7–18.9) for women with medium levels; and 15.5% (95% CI 13.1–17.8) for women with higher levels of psychosocial stress. Additionally, the effect of baseline stress on BMI change appeared to be stronger in women who were older (baseline age >35 years), compared to younger women (baseline age 35 years) in age-stratified analyses. Among younger women, adjusted percent BMI change was 16.6% (95% CI 12.5–20.7) in the lower stress group, 19.9% (95% 15.7–24.2) in the medium stress group, and 18.8% (95% CI 15.9–21.7) in those in the higher stress group. Whereas, among older women, adjusted percent increase in BMI over 13 years was 7.1% (95% CI 4.6–9.5) in women with low stress, 12.4% (95% CI 9.4–15.4) with medium levels, and 13.4% (95% CI 9.5–17.3) with higher stress; however, the interaction term of stress with age grouping was not statistically significant. For men, no association between psychosocial stress and percent BMI change over time was observed (Table 2).

Table 2 - Association of perceived stress with percent change in BMI in women and men.

Full table (29K)

No evidence of effect modification by educational status was observed for either men or women. Additionally, our measure of financial strain was not significantly associated with BMI change.

Discussion

This study was undertaken to help clarify how psychosocial stress might influence weight gain in black adults. The observed prospective relation between psychosocial stress and BMI change differed by gender. For women, higher levels of stress in 1988 were associated with greater percentage increases in BMI over the 13 years of follow-up, whereas among black men, similar associations were not found. In keeping with other studies, women in this study, reported higher levels of stress than men (26,27), and if a threshold effect was operating it is conceivable that the higher stress levels among women contributed to the observed gender effects. It is also possible that the physiological and behavioral responses to stress in blacks differs by gender, consistent with findings of gender differences in the relationship psychosocial stress to other outcomes (28,29,30).

Studies that specifically examine the potential impact of psychosocial factors on body weight and obesity in blacks are few and all have been cross-sectional in design thus prohibiting strong causal inferences. One such study found that black men with high levels of social support had higher levels of physical activity than others, but this study found no association between social support and BMI (18). Another cross-sectional study found that perceived stress was associated with haphazard meal planning and emotional eating (31). Lastly, a study of 36 black women found a positive correlation between body weight and stress level (20).

The relation between psychosocial factors, such as stress, and obesity may be mediated in a number of ways. Psychosocial stress is known to be associated with physiological effects, such as excess levels of the stress hormone cortisol that is known to potentiate visceral fat deposition (32). Growing evidence suggests that inflammation may mediate this relationship (33,34,35). One recent study, for example, estimates that inflammation mediates 25% of the association of stress with obesity (33). In addition, psychosocial stress has been linked to higher caloric intake and lower levels of physical activity, associations that could explain part of the relationship between perceived stress and body weight (16,17).

To our knowledge, we are the first to examine the prospective relationship between BMI change and psychosocial stress in blacks; however, there are a number of considerations that may limit the interpretation of our findings. First, psychosocial stress could, and probably does, vary greatly over time; therefore a baseline assessment of stress may not adequately capture recurring exposure to high levels of stress. In those participants with perceived stress measured at baseline and follow-up, the correlation between the two measures was about 33%. Additionally, psychosocial stress, as measured in this study, is a broad measure, and does not help us understand the impact of various types of stress, and the role of coping behaviors on weight change. For example, prior work has suggested that racism, racial segregation, and income inequity may have health consequences in blacks (11,12,13,36). Moreover, although we controlled for a number of factors that might have affected the relationship between psychosocial stress and the outcomes of interest, unmeasured or poorly measured covariates could have confounded the reported relationships. Third, though a community-based cohort, our sample is geographically restricted; therefore the study's findings may not generalize to other settings. Finally, as with all longitudinal studies, some participants were lost to follow-up, which can introduce bias. We attempted to minimize this source of bias by weighting the analyses for nonresponse at both baseline and follow-up.

In conclusion, higher levels of psychosocial stress were associated with greater weight gain in black women in our study. Although more longitudinal research is clearly needed, increased consideration should be given to interventions designed to reduce psychosocial stress as part of more comprehensive weight management programs targeted to black women. Future research efforts should use high-quality instruments measuring different types of psychosocial variables in longitudinal studies with ample sample size to detect associations and mediators of possible associations. Future studies will also need to examine how the relationship of psychosocial variables with weight gain and weight-related health behaviors vary across different categories of age, race, ethnicity, and gender.

Disclosure

The authors declared no conflict of interest.

References

REFERENCES

1. Allison DB, Fontaine KR, Manson JE, Stevens J, VanItallie TB. Annual deaths attributable to obesity in the United States. JAMA 1999;282:1530–1538. | Article | PubMed | ISI | ChemPort |
2. Fontaine KR, Redden DT, Wang C, Westfall AO, Allison DB. Years of life lost due to obesity. JAMA 2003;289:187–193. | Article | PubMed | ISI
3. Peeters A, Barendregt JJ, Willekens F et al. Obesity in adulthood and its consequences for life expectancy: a life-table analysis. Ann Intern Med 2003;138:24–32. | PubMed | ISI |
4. Ogden CL, Carroll MD, Curtin LR et al. Prevalence of Overweight and Obesity in the United States, 1999–2004. JAMA 2006;295:1549–1555. | Article | PubMed | ISI | ChemPort |
5. Everson-Rose SA, Lewis TT. Psychosocial factors and cardiovascular diseases. Annu Rev Public Health 2005;26:469–500. | Article | PubMed
6. Yan LL, Liu K, Matthews KA et al. Psychosocial factors and risk of hypertension: the Coronary Artery Risk Development in Young Adults (CARDIA) study. JAMA 2003;290:2138–2148. | Article | PubMed | ISI | ChemPort |
7. Hemingway H, Marmot M. Evidence based cardiology: psychosocial factors in the aetiology and prognosis of coronary heart disease. Systematic review of prospective cohort studies. BMJ 1999;318:1460–1467. | PubMed | ChemPort |
8. Gallo LC, Matthews KA. Do negative emotions mediate the association between socioeconomic status and health? Ann N Y Acad Sci 1999;896:226–245. | Article | PubMed | ChemPort |
9. Robert SA. Neighborhood socioeconomic context and adult health: the mediating role of individual health behaviors and psychosocial factors. Ann NY Acad Sci 1999;896:465–468. | Article | PubMed | ChemPort |
10. Bennett GG, Wolin KY, James SA. Lifecourse socioeconomic position and weight change among blacks: The Pitt County study. Obesity (Silver Spring) 2007;15:172–181. | Article | PubMed
11. Cooper RS, Kennelly JF, Durazo-Arvizu R et al. Relationship between premature mortality and socioeconomic factors in black and white populations of US metropolitan areas. Public Health Rep 2001;116:464–473. | PubMed | ChemPort |
12. 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;96:826–833. | Article | PubMed
13. Fang J, Madhavan S, Bosworth W, Alderman MH. Residential segregation and mortality in New York City. Soc Sci Med 1998;47:469–476. | Article | PubMed | ChemPort |
14. Kumanyika S, Adams-Campbell LL. Obesity, diet, and psychosocial factors contributing to cardiovascular disease in blacks. Cardiovasc Clin 1991;21:47–73. | PubMed | ChemPort |
15. LaVeist T, Thorpe R, Bowen-Reid T et al. Exploring health disparities in integrated communities: overview of the EHDIC study. J Urban Health 2008;85:11–21. | Article | PubMed
16. Bennett GG, Wolin KY, Avrunin JS et al. Does race/ethnicity moderate the association between job strain and leisure time physical activity? Ann Behav Med 2006;32:60–67. | Article | PubMed
17. Schulz AJ, House JS, Israel BA et al. Relational pathways between socioeconomic position and cardiovascular risk in a multiethnic urban sample: complexities and their implications for improving health in economically disadvantaged populations. J Epidemiol Community Health 2008;62:638–646. | Article | PubMed | ChemPort |
18. Allen J, Markovitz J, Jacobs DR Jr, Knox SS. Social Support and Health Behavior in Hostile black and White Men and Women in CARDIA. Psychosom Med 2001;63:609–618. | PubMed | ChemPort |
19. Strickland OL, Giger JN, Nelson MA, Davis CM. The relationships among stress, coping, social support, and weight class in premenopausal African American women at risk for coronary heart disease. J Cardiovasc Nurs 2007;22:272–278. | PubMed |
20. Walcott-McQuigg JA. Stress, women and weight control behavior. J Cult Divers 1995;2:64–71. | PubMed | ChemPort |
21. James SA, Keenan NL, Strogatz DS, Browning SR, Garrett JM. Socioeconomic status, John Henryism, and blood pressure in black adults. The Pitt County Study. Am J Epidemiol 1992;135:59–67. | PubMed | ChemPort |
22. James SA, Van Hoewyk J, Belli RF et al. Life-course socioeconomic position and hypertension in African American men: the Pitt County Study. Am J Public Health 2006;96:812–817. | Article | PubMed
23. Cohen S, Kamarck T, Mermelstein R. A global measure of perceived stress. J Health Soc Behav 1983;24:385–396. | Article | PubMed | ISI | ChemPort |
24. Cole SR. Assessment of differential item functioning in the Perceived Stress Scale-10. J Epidemiol Community Health 1999;53:319–320. | Article | PubMed | ChemPort |
25. Ng DM, Jeffery RW. Relationships between perceived stress and health behaviors in a sample of working adults. Health Psychol 2003;22:638–642. | Article | PubMed | ISI
26. Green CA, Pope CR. Gender, psychosocial factors and the use of medical services: a longitudinal analysis. Soc Sci Med 1999;48:1363–1372. | Article | PubMed | ChemPort |
27. Koopmans GT, Lamers LM. Gender and health care utilization: the role of mental distress and help-seeking propensity. Soc Sci Med 2007;64:1216–1230. | Article | PubMed
28. Kendler KS, Thornton LM, Prescott CA. Gender differences in the rates of exposure to stressful life events and sensitivity to their depressogenic effects. Am J Psychiatry 2001;158:587–593. | Article | PubMed | ISI | ChemPort |
29. Evans O, Steptoe A. The contribution of gender-role orientation, work factors and home stressors to psychological well-being and sickness absence in male- and female-dominated occupational groups. Soc Sci Med 2002;54:481–492. | Article | PubMed
30. Carmel S, Anson O, Levenson A, Bonneh DY, Maoz B. Life events, sense of coherence and health: gender differences on the kibbutz. Soc Sci Med 1991;32:1089–1096. | Article | PubMed | ChemPort |
31. Sims R, Gordon S, Garcia W et al. Perceived stress and eating behaviors in a community-based sample of African Americans. Eat Behav 2008;9:137–142. | Article | PubMed
32. McEwen BS, Seeman T. Protective and damaging effects of mediators of stress. Elaborating and testing the concepts of allostasis and allostatic load. Ann N Y Acad Sci 1999;896:30–47. | Article | PubMed | ChemPort |
33. Hamer M, Stamatakis E. Inflammation as an intermediate pathway in the association between psychosocial stress and obesity. Physiol Behav 2008;94:536–539. | Article | PubMed | ChemPort |
34. McDade TW, Hawkley LC, Cacioppo JT. Psychosocial and behavioral predictors of inflammation in middle-aged and older adults: the Chicago health, aging, and social relations study. Psychosom Med 2006;68:376–381. | Article | PubMed
35. Ranjit N, Diez-Roux AV, Shea S et al. Psychosocial factors and inflammation in the multi-ethnic study of atherosclerosis. Arch Intern Med 2007;167:174–181. | Article | PubMed
36. Din-Dzietham R, Nembhard WN, Collins R, Davis SK. Perceived stress following race-based discrimination at work is associated with hypertension in African-Americans. The metro Atlanta heart disease study, 1999-2001. Soc Sci Med 2004;58:449–461. | Article | PubMed

Acknowledgments

This work was primarily funded by an administrative supplement to National Institutes of Health (NIH) grant R01 DK071083-01A1 (Dr Fowler-Brown). Data collection for this study was supported by the NIH (grant HL 65645).