Anda di halaman 1dari 1

Structural Equation Modeling of the Inflammatory Response to Traffic Air Pollution Baja, ES Schwartz, JD Coull, BA Wellenius, GA* Vokonas,

PS+,# Suh, HHx


Department *Center

of Environmental Health, Harvard School of Public Health, Boston, MA; Department of Biostatistics, Harvard School of Public Health, Boston, MA; for Environmental Health and Technology, Brown University, Providence, RI; +VA Normative Aging Study, VA Boston Healthcare System, Boston, MA; #Department of Medicine, Boston University, Boston, MA; xEnvironmental Health Program, NORC at the University of Chicago, Boston, MA

BACKGROUND AND OBJECTIVES


Several epidemiological studies have reported conflicting results on the effect of traffic-related pollutants on markers of inflammation. In a Bayesian framework, we examined the effect of traffic pollution on inflammation using structural equation models (SEMs) and assessed effect modification by participant characteristics.

RESULTS
Study Population
Table 1. Characteristics of study population (N = 749), NAS 20002009. Characteristic (mean SD) Characteristic Age, years 74.9 6.7 Antihypertensive User Body Mass Index, kg/m2 28.1 4.2 BMI 30 kg/m2 Mean Arterial Pressure, mmHg 90.1 11.0 Ever Cigarette Smoker Cholesterol, mg/dL 185.6 38.5 Ever Diabetica Glucose Fasting, mg/dL 105.9 22.4 Two Drinks Cumulative Cigarette Packyears, yrs 20.3 24.4 Statin User
aReport

HYPOTHESIS
We hypothesize that short-term exposure to traffic pollution is associated with increased inflammation.

(%) 65.4 26.5 70.4 19.6 18.6 47.2

of doctors diagnosis of disease or FBG >126 mg/dL

Bayesian SEM: Posterior Probability Distribution

METHODS
Study Population: Normative Aging Study (NAS) 749 NAS participants with every 3-5 year clinic visits. Health Measurements: CRP, sICAM-1 and sVCAM-1 Each clinic visit: includes extensive physical examination, lab tests, blood collection, self-administered questionnaire, CRP, sICAM-1, sVCAM-1. Plasma sICAM-1 and sVCAM-1 measured by ELISA assay. Serum CRP measured by immunoturbidimetric assay. Air Pollution and Meteorology Hourly ambient BC measured on Harvard Library, 1 km from clinic Hourly ambient NO, CO, NO2 from MA-operated monitoring sites Temperature from Logan airport Exposure Periods 1-, 2-, 3-, 7-, 14-, and 30-day moving averages pre-visit.
Predictors: Age BMI Statin Usage Hypertension Usage Smoking Status Pack-years smoked Diabetic Status 2 servings of alcohol
Density 0.4 0.3 0.2 0.1 0.0 0 10 20 30 40 1-d MA 2-d MA 3-d MA 7-d MA 14-d MA 30-d MA

Figure 3. Posterior estimated percent change in inflammation associated with an IQR increase in traffic pollution exposure at different daily moving averages

Estimated % change in inflammation

Bayesian Linear Mixed Model: Posterior Probability Distribution

Figure 4. Posterior estimated percent change in sVCAM-1 associated with an IQR increase in BC exposure at different daily moving averages

0.4 0.3 Density 0.2 0.1 0.0

1-d MA 2-d MA 3-d MA 7-d MA 14-d MA 30-d MA

Statistical Analysis Bayesian SEMs that account for repeated measures Linear mixed models with random subject-specific intercepts in Bayesian and Frequentist approaches. SEM estimates scaled to a specific traffic-related pollutant (BC) and inflammation marker (sVCAM-1).

Figure 1: Path Diagram Effect of Traffic on Inflammation: Main Effect Model


Confounding Factors: Buy SmartDraw!- purchased copies print this document without a watermark . Temperature Visit www.smartdraw.com or call 1-800-768-3729. Date

10

20

30

40

Estimated % change in sVCAM-1

Modifiers: Diabetic Status Smoking Status Obesity Status


Buy SmartDraw!purchased copies print this Figure 2: Path Diagram Effect of Traffic on document without a watermark . Visit www.smartdraw.com or call 1-800-768-3729. Inflammation: Effect Modification Model

Figure 5. Posterior estimated % change in inflammation per IQR change of traffic by diabetic status (doctors diagnosis of diabetes or Fasting Blood Glucose (FBG) >126 mg/dL, FBG 126 mg/dL)

CONCLUSIONS
This study showed positive associations between traffic pollution (BC, CO, NO2 and NO) exposure and inflammation (CRP, sICAM-1 and sVCAM-1). BSEMs are a suitable, alternative method to examine effect of traffic pollution exposure on multiple health effect outcomes in repeated measures and longitudinal studies. The Bayesian approach allowed us to incorporate the mixed effect model structure within the SEM framework. This study provides further evidence that traffic pollutants via the inflammation pathway may play a critical role in cardiopulmonary toxicity.

Regression Models Bayesian SEM Trafficit = 1 Sine(2/365.24)it + 2 Cosine(2/365.24)it + 3 Residual(Atemp)it + 4 Residual2(Atemp)it [1] Inflammationit = 1 Trafficit + 2 X1it + + k+1 Xkit + k+2 Sine(2/365.24)it + k+3 Cosine(2/365.24)it + k+4 Residual(Atemp)it + k+5 Residual2(Atemp)it + bi Bayesian Linear Mixed Model sVCAM-1it = 0 + 1 BCit + 2 X1it + + k+1 Xkit + k+2 Sine(2/365.24)it + k+3 Cosine(2/365.24)it + k+4 Residual(Atemp)it + k+5 Residual2(Atemp)it + bi

[2]

ACKNOWLEDGMENTS
[3]
Supported in part by grants from the National Institute of Environmental Health Sciences (ES014663-01A2, ES-015172, ES015774, and PO1 ES09825) and from the U.S. Environmental Protection Agency (EPA R827353 and R 83479801).