Controlling for confounding in the presence of measurement error in hierarchical models: a Bayesian approach

Accounting for measurement error to assess the effect of air pollution on omic signals

Studies on the effects of air pollution and more generally environmental exposures on health require measurements of pollutants, which are affected by measurement error. This is a cause of bias in the estimation of parameters relevant to the study and can lead to inaccurate conclusions when evaluating associations among pollutants, disease risk and biomarkers. Although the presence of measurement error in such studies has been recognized as a potential problem, it is rarely considered in applications and practical solutions are still lacking. In this work, we formulate Bayesian measurement error models and apply them to study the link between air pollution and omic signals. The data we use stem from the "Oxford Street II Study", a randomized crossover trial in which 60 volunteers walked for two hours in a traffic-free area (Hyde Park) and in a busy shopping street (Oxford Street) of London. Metabolomic measurements were made in each individual as well as air pollution measurements, in order to investigate the association between short-term exposure to traffic related air pollution and perturbation of metabolic pathways. We implemented error-corrected models in a classical framework and used the flexibility of Bayesian hierarchical models to account for dependencies among omic signals, as well as among different pollutants. Models were implemented using traditional Markov Chain Monte Carlo (MCMC) simulative methods as well as integrated Laplace approximation. The inclusion of a classical measurement error term resulted in variable estimates of the association between omic signals and traffic related air pollution measurements, where the direction of the bias was not predictable a priori. The models were successful in including and accounting for different correlation structures, both among omic signals and among different pollutant exposures. In general, more associations were identified when the correlation among omics and among pollutants were modeled, and their number increased when a measurement error term was additionally included in the multivariate models (particularly for the associations between metabolomics and NO2).

Ambient air pollution during pregnancy and risk of gestational diabetes in New York City

BACKGROUND

Emerging evidence suggests a potential association between ambient air pollution and risk of gestational diabetes mellitus (GDM), but results have been inconsistent. Accordingly, we assessed the associations between ambient fine particulate matter (PM_2.5) and nitrogen dioxide (NO₂) levels with risk of GDM.

METHODS

Using linked data from birth certificates, hospital discharge diagnoses, and air pollution estimates informed by the New York City Community Air Survey, we fit conditional logistic regression models to evaluate the association between residential levels of PM_2.5 and NO₂ with risk of GDM among 256,372 singleton live births of non-smoking mothers in New York City born 2008-2010, adjusting for sociodemographic factors and stratified on zip code of maternal address.

RESULTS

GDM was identified in 17,065 women, yielding a risk of GDM in the study sample of 67 per 1000 deliveries. In single pollutant models, 1st and 2nd trimester PM_2.5 was associated with a lower and higher risk of GDM, respectively. In models mutually adjusting for PM_2.5 levels in both trimesters, GDM was associated with PM_2.5 levels in the 2nd trimester (OR: 1.06, 95% CI: 1.02, 1.10 per interquartile range increase in PM_2.5), but not the 1st trimester (OR: 0.99, 95% CI: 0.96, 1.02). Conversely, GDM was associated with NO₂ during the 1st trimester (OR: 1.05, 95% CI: 1.01, 1.09), but not the 2nd trimester (OR: 1.02, 95% CI: 0.98, 1.06). The positive associations between pollutants and GDM were robust to different model specifications. PM_2.5 in the 2nd trimester was more strongly associated with GDM among mothers who were aged <35 years and not Medicaid recipients. NO₂ in the 1st trimester was more strongly associated with GDM among overweight and parous women.

CONCLUSIONS

In this large cohort of singleton births in New York City, NO₂ in the 1st trimester and PM_2.5 in the 2nd trimester were associated with higher odds of GDM, while 1st trimester PM_2.5 was weakly and inconsistently associated with lower odds of GDM.