The socioeconomic landscape of the exposome during pregnancy

While socioeconomic position (SEP) is consistently related to pregnancy and birth outcome disparities, relevant biological mechanisms are manifold, thus necessitating more comprehensive characterization of SEP-exposome associations during pregnancy.

We implemented an exposomic approach to systematically characterize the socioeconomic landscape of prenatal exposures in a setting where social segregation was less distinct in a hypotheses-generating manner.

We described the correlation structure of 134 prenatal exogenous and endogenous sources (e.g., micronutrients, hormones, immunomodulatory metabolites, environmental pollutants) collected in a diverse, population-representative, urban, high-income longitudinal mother-offspring cohort (N = 1341; 2009-2011). We examined the associations between maternal, paternal, household, and areal level SEP indicators and 134 exposures using multiple regressions adjusted for precision variables, as well as potential effect measure modification by ethnicity and nativity. Finally, we generated summary SEP indices using Multiple Correspondence Analysis to further explore possible curved relationships.

Individual and household SEP were associated with anthropometric/adiposity measures, folate, omega-3 fatty acids, insulin-like growth factor-II, fasting glucose, and neopterin, an inflammatory marker. We observed paternal education was more strongly and consistently related to maternal exposures than maternal education. This was most apparent amongst couples discordant on education. Analyses revealed additional non-linear associations between areal composite SEP and particulate matter. Environmental contaminants (e.g., per- and polyfluoroalkyl substances) and micronutrients (e.g., folate and copper) showed opposing associations by ethnicity and nativity, respectively.

SEP-exposome relationships are complex, non-linear, and context specific. Our findings reinforce the potential role of paternal contributions and context-specific modifiers of associations, such as between ethnicity and maternal diet-related exposures. Despite weak presumed areal clustering of individual exposures in our context, our approach reinforces subtle non-linearities in areal-level exposures.