Origins of the

1
Origins of the flame within Social and
Physical Correlates of Inflammation in U.S.
Children.

• Jennifer Beam Dowd, Hunter College,
• CUNY Institute for Demographic Research (CIDR)
• Anna Zajacova, Allison Aiello, Center for Social
  Epidemiology and Population Health, University of
  Michigan

2
Background and Motivation

• Disparities in health by SES in the U.S. begin in
  childhood (Case, Lubotsky, Paxson 2002)
• Biological pathways linking SES to health,
  especially in children, are not clear.
• Early environments can shape developing
  physiological systems (critical and sensitive
  periods)

Background Data/Methods Results Conclusions
3
Inflammation

• Integral part of the human stress and immune
  response
• Pro-inflammatory cytokines regulate the
  production of acute-phase proteins such as
  C-reactive protein (CRP) which fight infection
  and promote repair of damaged issues.
• Little is known about the predictors of low-grade
  inflammation in children

Background Data/Methods Results Conclusion
4
Contributors to inflammation

• Independent predictors of increased inflammation
  in adults
• Higher BMI (inflammatory cytokines expressed in
  adipose tissue)
• Smoking
• Poor sleep quality/short sleep
• Diet high in saturated and trans fat
• Chronic Infections

Background Data/Methods
Results Conclusion
5
Health Consequences of Inflammation

• Elevated CRP is associated with risk of
• myocardial infarction, stroke, atherosclerosis
• insulin resistance, Type II diabetes
• vascular dementia, Alzheimers disease.
• Life-long inflammatory burden may shape later
  life patterns of aging and mortality. (Crimmins
  and Finch 2006).

Background Data/Methods Results Conclusion
6
Inflamm-aging

• Chronic immune activation, including a persistent
  inflammatory status, may drive what we considered
  age related declines in functioning and immune
  response
• Thus large differences across groups
  (race/ethnicity/SES) in burden of inflammation
  could play a role in observed differences in
  aging rates and longevity.

Background Data/Methods Results Conclusion
7
SES differences in Infection Burden in U.S.
children
8
Previous Work

• Lower SES associated with higher CRP in U.S.
  adults
• Mixed results for race/ethnicity-some studies
  show highest levels for blacks, some for
  Hispanics
• European studies have not found social
  inequalities in CRP in childhood, differences
  emerge later.
• To our knowledge, no existing studies looking at
  CRP disparities in U.S. children

Background Data/Methods Results Conclusion
9
Primary Research Questions

• Are physical (infections, BMI, etc) and social
  (family income, race/ethnicity) risk factors
  associated with inflammation in U.S. children?
• Do physical risk factors mediate the relationship
  between social factors and levels of inflammation
  in U.S. children?

Background Data/Methods Results Conclusion
10
Secondary Question

• Hygiene hypothesis Mixed evidence on whether
  higher infectious burden in childhood promotes
  better or worse regulation of inflammation later
  in life
• Are chronic infections related to inflammation in
  U.S. children? Are proxy measures of pathogen
  exposure related to inflammation in U.S.
  children?

Background Data/Methods Results Conclusion
11
Data

• National Health and Nutrition Examination Survey
  (NHANES), 1999-2004
• Cross-sectional, representative sample of
  non-institutionalized U.S. population
• Face-to-face interview, medical exam, collection
  of blood and urine
• Our sample consists of children aged 3-17, N 6338

Background Data/Methods Results Conclusion
12
Measures Outcome

• High Sensitivity C-reactive Protein (CRP) mg/L
• Distribution is right-skewed, transformed to
  Ln(CRP)

Background Data/Methods Results Conclusion
13
Measures Physical Predictors

• Infections Positive Serostatus for
• Cytomegalovirus (CMV)
• Herpes Simplex Virus Type 1 (HSV-1)
• Helicobacter Pylori (H Pylori)
• Cryptosporidium
• Toxoplasmosis
• Hepatitis A Virus (HAV)
• Infectious Burden (Factor Score)

Background Data/Methods Results Conclusion
14
Measures Physical Predictors

• Body Mass Index (BMI) (kg/m2)
• Illness in the last 30 days (0/1)
• Low birth weight (0/1)
• Mother Smoked during pregnancy (0/1)
• Currently a smoker in the Household (0/1)
• Cotinine (log transformed)
• Triclosan (log transformed, N557)
• White Blood Cell Count
• Vitamin D (log transformed)

Background Data/Methods Results Conclusion
15
Measures Social Predictors

• Age (continuous)
• Sex
• Foreign Born (0/1)
• Household size (continuous)
• Race/ethnicity (White/Black/Mexican-American)
• SES
• Poverty-Income Ratio (Ratio of Family Income to
  Poverty Line)
• Years of Education of the Household Reference
  Person

Background Data/Methods Results Conclusion
16
Methods

• OLS Regressions
• Ln(CRP) a ß1(Social) ß2(Physical) e
• Infection burden score created with M-Plus,
  confirmatory factor analysis with
  full-information maximum likelihood estimation
• All analyses conducted with STATA 10.0 SVY
  commands to account for complex survey design

Background Data/Methods
Results Conclusion
17
Descriptive Statistics
18
Results
19
Results
20
Conclusions

• BMI and current/recent illness are strong
  predictors of CRP in U.S. children
• Differences in CRP levels by income largely
  accounted for by BMI and recent illness.
• Higher levels for Mexican-American race/ethnicity
  not explained by physical vars.

21
Conclusions

• Hygiene Hypothesis Still a mystery
• --Pro increased HH size associated with lower
  CRP, Being foreign-born associated with lower CRP
• --Cons infection coefficients all reflect
  positive effects on CRP, foreign-born effect
  explained by BMI, Triclosan coefficient negative

22
Conclusions/Next Steps

• Higher BMI and potentially more frequent acute
  infections contribute to greater levels of
  low-grade inflammation among U.S. children with
  lower family income.
• What explains higher levels for Mexican-American
  children?
• Potential life-course health implications aging,
  CVD, cognition and learning?

23
Acknowledgements

• Thanks to collaborators Anna Zajacova and Allison
  Aiello, Research assistance from Megan Todd
• Support from the NIH 1R21NR011181-01