The Exposome and its Associations with Broad Mental and Physical Health Measures in Early Adolescence

Abstract

Environment is key to human development, yet the complex network structure of exposures (i.e., exposome) makes it challenging to investigate. Here, we analyzed data from the Adolescent Brain and Cognitive Development (ABCD) Study , a large, diverse sample of US adolescents (N=11,235, mean age=10.9, 52% male) with phenotyping at multiple levels of environmental exposure. Applying data-driven iterative factor analyses and bifactor modeling, we reduced dimensionality from hundreds of exposures to six exposome subfactors and a general (adverse) exposome factor. These factors revealed quantitative differences among racial and ethnic groups. Exposome factors increased variance explained in mental health by 10-fold (from <4% to >38%), over and above other commonly used sociodemographic factors. The general exposome factor was associated with psychopathology (Beta=0.27) and key health-related outcomes: obesity (OR=1.4) and advanced pubertal development (OR=1.3). Findings highlight the exposome role in adolescent health and demonstrate the critical need to study environment using the exposome framework.

Exposome and Trans-syndromal Developmental Trajectories Toward Psychosis

“…To capture the exposome in early adolescence, we recently applied an approach using dimensionality reduction and bifactor modeling of the exposome in 2 youth cohorts (the Philadelphia Neurodevelopmental Cohort [PNC] and the Adolescent Brain Cognitive Development [ABCD] Study), showing that such an approach indeed facilitates generalizability across cohorts ( 49 ). Briefly, we combed through each cohort’s data and reduced dimensionality of all environmental exposures (798 exposures in the ABCD Study and 29 exposures in PNC) to a limited set of correlated environmental factor scores that capture different levels of exposure (e.g., household, neighborhood).…”

“…The general ES receives loadings from all environmental exposures included in the model and therefore represents a weighted sum of multilevel environmental exposures for each study participant. After calculating the general exposome factor in the 2 independent datasets, we have shown that the general ES is associated with youth mental and general health measures in an almost identical magnitude (i.e., effect size), despite the fact that ABCD Study and PNC data were collected in different places and at different times with substantial differences in environmental phenotyping ( 49 ). We suggest that this general ES can potentially be useful because it captures a latent environment factor that is less likely to be skewed by a specific environmental exposure and might allow for better generalization of environment modeling across different cohorts, which is often very different.…”

Estimating the Association Between Exposome and Psychosis as Well as General Psychopathology: Results From the ABCD Study

“…Psychosis spectrum disorder (PSD) has a complex, multifaceted, interconnected environmental etiology including exposures from different levels (e.g., family-, household-, neighborhood-, and state-level) ( 1 , 2 ). Evidence highlights the role of various environmental exposures such as pre- and perinatal complications, childhood adversities (e.g., abuse, neglect, nonintentional adversities), minority status, cannabis use, and urbanicity ( 3 , 4 ).…”

“…Furthermore, not only significant life events but also day-to-day experiences play an important role ( 1 , 12 , 13 , 14 ). The environment in which a child grows up, including factors such as family, neighborhood, state, or country, can influence their development and psychopathology ( 1 , 2 , 15 , 16 ). Furthermore, environmental exposures do not occur in isolation.…”

“…The exposome entails exposures from internal (e.g., inflammation) as well as external (e.g., chemical, lifestyle, psychosocial) domains ( 26 ). Recently, approaches (such as the identification of individual correlates, mechanisms, and exposome domains) have been applied to uncover exposomic vulnerability for child development by using large numbers of variables ( 2 , 27 , 28 ). Furthermore, guided by the exposome framework, we previously applied predictive modeling to calculate environmental vulnerability for schizophrenia to estimate an aggregate weighted score (exposome score for schizophrenia) that takes into account the specific loading of individual exposures on schizophrenia, as well as their interdependencies (i.e., correlations) ( 22 ).…”