Prenatal Exposure to Endocrine Disrupting Compounds and ...
Eva Govarts, Flemish Institute for Technological Research (VITO)
Keywords: Endocrine disrupting compounds, Flemish environment and health survey, Left censored data, Marginal model, Principal component analysis.
Background: The incidence of childhood obesity has reached ‘epidemic’ proportions globally and there is an urgent need to increase our understanding of the impact of food contaminants on obesity development. There is still much uncertainty related to the causes and underlying physiological mechanisms of obesity.
Objective: In this study, we examine the hypothesis that prenatal exposure to endocrine disrupting compounds (EDCs) in food plays a role in the development of obesity later in life. We are interested to study dose-response relationships between EDCs and both health outcomes and effect-biomarkers such as thyroid hormones.
Method: The Flemish environment and health survey (FLEHS) for human biomonitoring was introduced by Flemish Institute for Technological Research (VITO) and it was conducted in Flanders, Belgium for five-years (2002-2006). Relevant data on environmental pollution and related health indicators were also collected. In this study, we only used the data of mother-child cohort with a sample size of n = 1196. Mothers and their newborns, i.e., mother-child cohort, were selected by geographically stratified sampling resulting in 8 geographical areas covering 22% of the surface area of Flanders and 20% of the population. In our analysis, these geographical areas are treated as clusters. In addition to markers for exposures (EDCs), the other profiles considered are effect biomarkers, health outcomes, and some effect modifiers. Principal component analysis is used to cumulate mixture mechanism of EDCs on target body. For studying the dose-response relationships, a univariate and joint marginal models are fitted for both health outcomes and thyroid hormones. In environmental research, it is common to come across with measurements of exposure concentrations that are non-detectable or near limit of detection (LOD): left-censored data. Some imputation techniques are employed: single imputation, tobit regression and multiple imputation, are implemented, for sensitivity analysis reason.
Results and Conclusion: We found significant dose-response relationships between EDCs and both birth weight and head circumference from health outcomes, and between EDCs and both thyroid stimulating hormone (TSH) and free thyroxine (FT4) from thyroid hormones outcomes.
Important Dates & Deadlines
- October 9 - 11, 2013