What Can Epidemiological Studies Tell Us about the Impact of Chemical Mixtures on Human Health?


Table 1. Description and examples of questions related to chemical mixtures and human health that epidemiological studies can address.
Question Examples and Methods Challenges
What are the health effects of individual chemicals within a mixture?
  • Quantified the association between prenatal exposure to 52 endocrine-disrupting chemicals and children’s autistic behaviors using semi-Bayesian shrinkage methods (Braun et al. 2014).
  • Used elastic net to examine the association between 16 prenatal exposures and birth weight (Lenters et al. 2015).
  • Examined the association between 188 environmental factors and serum lipid levels using an environment-wide association study (Patel et al. 2012).
  • Some approaches may not adequately address copollutant confounding.
  • Multiple comparisons.
  • Disentangling the effect of highly correlated copollutants.
What are the interactions between chemicals within a mixture?
  • Determined if the neurotoxic effects of lead were greater among children with higher manganese exposure using product interaction terms (Claus Henn et al. 2012).
  • Identified and examined interactions between multiple metal biomarkers and child mental development using Bayesian kernel machine regression (Bobb et al. 2015).
  • Difference in toxicologic and epidemiologic definitions of interaction (Howard and Webster 2013).
  • Multiple comparisons.
  • Imprecise effect estimates and reduced statistical power for detecting interactions.
What is the health effect of cumulative chemical exposure?
  • Examined the relationship between child anthropometry and exposure to dioxins using a toxic equivalency summary measure (Burns et al. 2011).
  • Estimated the association between different chemical classes and non-Hodgkin lymphoma using weighted quantile sum regression (Czarnota et al. 2015).
  • Used principal components analysis to examine the association between phthalate exposures and child anthropometry (Maresca et al. 2015).
  • Verifying the assumption of no interaction between individual components.
  • Estimating cumulative exposure metrics for specific health outcomes.
  • Availability of information to create biologically weighted summary measures.
  • Interpretation of results from more complex statistical methods.