Environmental Exposures and Childhood Obesity: An Exposome Analysis

Abstract

BackgroundWith an increasing number of suspected developmental obesogens, there is a need to identify those most relevant for childhood obesity risk.ObjectiveTo systematically test associations between many environmental exposures and obesity and adiposity in children, using an exposome approach.MethodsThe HELIX study measured obesity outcomes in 1,301 children from 6 European countries, aged 6-11 years, including age-and-sex specific SD-scores for body mass index (zBMI), waist circumference, skinfold thickness, and body fat mass. HELIX estimated over 90 prenatal and over 110 childhood environmental exposures to outdoor, chemical and lifestyle factors. Exposome-outcome associations were tested using agnostic exposure-wide analyses (ExWAS) and with the deletion-substitution-addition (DSA) algorithm that considers all exposures simultaneously.ResultsFor prenatal exposures, maternal smoking-related variables (active and passive smoking, cotinine and cadmium concentrations) were associated with higher child zBMI in the ExWAS, although these associations did not pass the multiple testing corrected p-value (0.001). Maternal passive and active smoking during pregnancy were also selected in the DSA model. For childhood exposures, 18 exposures were associated with zBMI in the ExWAS after multiple testing correction (p<0.0006). In the DSA model, exposure to indoor particulate matter and copper were associated with increased zBMI, and HCB and PBDE-153 with decreased zBMI. Other adiposity outcomes showed similar results.ConclusionsEarly life exposure to several pollutants is associated with childhood obesity, with smoking and indoor air pollutants raising particular concern. Results for childhood exposures may suffer from reverse causation bias. The exposome’s systematic approach can be used to prioritize factors for further investigation and intervention, and avoids problems of selective reporting, co-exposure confounding, and false positive results.