Demographic differences in chemical mixture exposure and associated determinants among Swedish adolescents

Background and aim: Few population-based studies of adolescents have investigated the relation between socio-demographic factors and the wider chemical exposome. Adolescents are likely going through a sensitive developmental window were negative exposure-induced health effects can manifest into adulthood. We aim to identify associations of social demographic factors with the chemical exposome of Swedish adolescents as well as to identify demographic groups more at risk of higher exposure to specific substances. Methods: Using the Riksmaten Adolescents 2016-17 cohort (RMU) we analysed several social-demographic factors: gender, participant/maternal (P/M) birth country income level, parental education levels, and geographic location (longitude/latitude), and their relation to blood/serum/urine concentrations of a large range of toxic substances (N=64) across multiple chemical groups (heavy metals, chlorinated pesticides, PCBs, BFRs, PFASs, phthalates, phenols, pesticides, PAHs). Association analysis was conducted using ordinal regression models due to its robustness to handle highly skewed data. Results: P/M birth country was the most frequent determinant of toxic substance concentrations, being significantly associated with concentrations of 45 substances, followed by gender (N=41), and longitude (N=37). P/M birth country also showed the largest fold-differences in concentrations, with subjects born in high-income countries having several-fold higher estimated adjusted mean (EAM) concentrations of PCBs, BFRs and PFASs than those born in low-income countries. In contrast, adolescents born in low-income countries had higher EAM levels of heavy metals, chlorinated pesticides, and phthalates. Substances belonging to the same chemical group tended to have the same determinants. Interesting north to south or west to east gradients of substance concentrations were identified in Sweden, suggesting geographical inequalities of exposure. Conclusions: P/M birth country is an important confounder and should be included in future toxicology population-based studies. Exposure associations with latitude/longitude should be further investigated. Keywords: Adolescents, RMU, ordinal regression, exposome, metals, PCB, BFR, PFAS, pesticides, phthalates, phenols, epidemiology, birth country income