Joint and independent neurotoxic effects of early life exposures to a chemical mixture: A multi-pollutant approach combining ensemble learning and g-computation

Impact of antenatal exposure to a mixture of persistent organic pollutants on intellectual development

OBJECTIVE

Strong experimental evidence exists that several endocrine disrupting chemicals (EDCs) have neurobehavioral toxicity. However, evidence of associations between prenatal exposure and child's cognitive development is inconsistent. Moreover, toxicants are generally analyzed one by one without considering aggregate effects. We examined here the impact of a prenatal exposure to a mixture of persistent organic pollutants (POPs) on intellectual abilities in preschool children, and compared their effects to those described in the literature.

METHODS

Sixty-two children were included in a longitudinal cohort. Four organochlorine pesticides, four polychlorinated biphenyls (PCBs) and seven perfluorinated compounds (PFCs) were measured in cord blood. Intellectual abilities were assessed at 6 years of age using the Wechsler Preschool and Primary Scale of Intelligence 4th ed. (WPPSI-IV). We examined the associations between a mixture of POPs and cognitive performances using principal components approach (PCA) and weighted quantile sum (WQS) regression taking sex difference into account.

RESULTS

No negative correlation was found when analyses were performed on boys and girls together. In sex-stratified analyses, lower scores in full scale intelligence quotient (FSIQ) and fluid reasoning index (FRI) were observed in boys most exposed to a mixture of POPs. Increase of the WQS index was also associated with lower verbal comprehension index (VCI) scores in girls only. No other negative correlation was found using both WQS and PCA models.

CONCLUSION

Our study suggests deleterious associations between antenatal exposure to a mixture of POPs and sex-specific cognitive level, clarifying some trends described in the literature.

Perinatal exposure to polychlorinated biphenyls (PCBs) and child neurodevelopment: A comprehensive systematic review of outcomes and methodological approaches

BACKGROUND

Polychlorinated biphenyls (PCBs), extensively used in various products, prompt ongoing concern despite reduced exposure since the 1970s. This systematic review explores prenatal PCB and hydroxylated metabolites (OH-PCBs) exposure's association with child neurodevelopment. Encompassing cognitive, motor development, behavior, attention, ADHD, and ASD risks, it also evaluates diverse methodological approaches in studies.

METHODS

PubMed, Embase, PsycINFO, and Web of Science databases were searched through August 23, 2023, by predefined search strings. Peer-reviewed studies published in English were included. The inclusion criteria were: (i) PCBs/OH-PCBs measured directly in maternal and cord blood, placenta or breast milk collected in the perinatal period; (ii) outcomes of cognitive development, motor development, attention, behavior, attention-deficit/hyperactivity disorder (ADHD), and autism spectrum disorder (ASD) among children≤18 years old. Quality assessment followed the National Heart, Lung, and Blood Institute's tool.

RESULTS

Overall, 87 studies were included in this review. We found evidence for the association between perinatal PCB exposure and adverse cognitive development and attention issues in middle childhood. There appeared to be no or negligible link between perinatal PCB exposure and early childhood motor development or the risk of ADHD/ASD. There was an indication of a sex-specific association with worse cognition and attention scores among boys. Some individual studies suggested a possible association between prenatal exposure to OH-PCBs and neurodevelopmental outcomes. There was significant heterogeneity between the studies in exposure markers, exposure assessment timing, outcome assessment, and statistical analysis.

CONCLUSIONS

Significant methodological, clinical and statistical heterogeneity existed in the included studies. Adverse effects on cognitive development and attention were observed in middle childhood. Little or no apparent link on both motor development and risk of ADHD/ASD was observed in early childhood. Inconclusive evidence prevailed regarding other neurodevelopmental aspects due to limited studies. Future research could further explore sex-specific associations and evaluate associations at lower exposure levels post-PCB ban in the US. It should also consider OH-PCB metabolites, co-pollutants, mixtures, and their potential interactions.

Prenatal Exposure to Perfluoroalkyl Substances and Child Behavior at Age 12: A PELAGIE Mother-Child Cohort Study

BACKGROUND

Per- and polyfluoroalkyl substances (PFAS) are chemical substances spread throughout the environment worldwide. Exposure during pregnancy represents a specific window of vulnerability for child health.

OBJECTIVE

Our objective was to assess the impact of prenatal exposure to multiple PFAS on emotional and behavioral functions in 12-y-old children.

METHOD

In the PELAGIE mother-child cohort (France), prenatal exposure to nine PFAS was measured from concentrations in cord serum samples. Behavior was assessed at age 12 y using the parent-reported Strengths and Difficulties Questionnaire (SDQ) and the self-reported Dominic Interactive for Adolescents (DIA) for 444 children. Associations were estimated using negative binomial models for each PFAS. Bayesian kernel machine regression (BKMR) models were performed to assess the exposure mixture effect on children's behavior.

RESULTS

In our study population, 73% of mothers had spent more than 12 y in education. Higher scores on SDQ externalizing subscale were observed with increasing cord-serum concentration of perfluorooctanoic acid (PFOA) and perfluorononanoic acid (PFNA) [adjusted mean ratio ( aMR ) = 1.18 , 95% confidence interval (CI): 1.03, 1.34, and aMR = 1.14 (95% CI: 1.00, 1.29) for every doubling of concentration, respectively]. Results for the hyperactivity score were similar [aMR = 1.20 (95% CI: 1.04, 1.40) and aMR = 1.18 (95% CI: 1.02, 1.36), respectively]. With regard to major depressive disorder and internalizing subscales, perfluorodecanoic acid (PFDA) was associated with higher self-reported DIA scores [aMR = 1.14 (95% CI: 1.01, 1.27) and aMR = 1.11 (95% CI: 1.02, 1.21), respectively]. In terms of the anxiety subscale, PFDA and PFNA were associated with higher scores [aMR = 1.11 (95% CI: 1.02, 1.21) and aMR = 1.10 (95% CI: 1.01, 1.19), respectively]. Concurrent increases in the PFAS concentrations included in the BKMR models showed no change in the SDQ externalizing and DIA internalizing subscales scores.

CONCLUSION

Prenatal exposure to PFNA and PFOA were associated with increasing scores for measures of externalizing behaviors, specifically hyperactivity. We also identified associations between PFNA and PFDA prenatal exposure levels and increasing scores related to internalizing behaviors (general anxiety and major depressive disorder), which adds to the as yet sparse literature examining the links between prenatal exposure to PFAS and internalizing disorders. https://doi.org/10.1289/EHP12540.

Challenges in Obtaining Valid Causal Effect Estimates with Machine Learning Algorithms

Unlike parametric regression, machine learning (ML) methods do not generally require precise knowledge of the true data generating mechanisms. As such, numerous authors have advocated for ML methods to estimate causal effects. Unfortunately, ML algorithmscan perform worse than parametric regression. We demonstrate the performance of ML-based single- and double-robust estimators. We use 100 Monte Carlo samples with sample sizes of 200, 1200, and 5000 to investigate bias and confidence interval coverage under several scenarios. In a simple confounding scenario, confounders were related to the treatment and the outcome via parametric models. In a complex confounding scenario, the simple confounders were transformed to induce complicated nonlinear relationships. In the simple scenario, when ML algorithms were used, double-robust estimators were superior to single-robust estimators. In the complex scenario, single-robust estimators with ML algorithms were at least as biased as estimators using misspecified parametric models. Double-robust estimators were less biased, but coverage was well below nominal. The use of sample splitting, inclusion of confounder interactions, reliance on a richly specified ML algorithm, and use of doubly robust estimators was the only explored approach that yielded negligible bias and nominal coverage. Our results suggest that ML based singly robust methods should be avoided.

Estimating Causal Effects of Interventions on Early-life Environmental Exposures Using Observational Data

PURPOSE OF REVIEW

We discuss how epidemiologic studies have used observational data to estimate the effects of potential interventions on early-life environmental exposures. We summarize the value of posing questions about interventions, how a group of techniques known as "g-methods" can provide advantages for estimating intervention effects, and how investigators have grappled with the strong assumptions required for causal inference.

RECENT FINDINGS

We identified nine studies that estimated health effects of hypothetical interventions on early-life environmental exposures. Of these, six examined air pollution. Interventions evaluated by these studies included setting exposure levels at a specific value, shifting exposure distributions, and limiting exposure levels to less than a threshold value. Only one study linked exposure contrasts to a specific intervention on an exposure source, however. There is growing interest in estimating intervention effects of early-life environmental exposures, in part because intervention effects are directly related to possible public health actions. Future studies can build on existing work by linking research questions to specific hypothetical interventions that could reduce exposure levels. We discuss how framing questions around interventions can help overcome some of the barriers to causal inference and how advances related to machine learning may strengthen studies by sidestepping the overly restrictive assumptions of parametric regression models. By leveraging advancements in causal inference and exposure science, an intervention framework for environmental epidemiology can guide actionable solutions to improve children's environmental health.

Maternal per- and poly-fluoroalkyl substances exposures associated with higher depressive symptom scores among immigrant women in the Chemicals in Our Bodies cohort in San Francisco

BACKGROUND

Exposure to per- and poly-fluoroalkyl substances (PFAS) remains an important public health issue due to widespread detection and persistence in environmental media, slow metabolism in humans, and influences on physiological processes such as neurological signaling. Maternal depression is highly prevalent during pregnancy and postpartum and is potentially sensitive to PFAS. The health risks associated with PFAS may be further amplified in historically marginalized communities, including immigrants.

OBJECTIVE

Evaluate maternal concentrations of PFAS in association with depression scores during pregnancy and whether effects differ between US born and immigrant women.

METHODS

Our study sample included 282 US born and 235 immigrant pregnant women enrolled in the Chemicals in Our Bodies prospective birth cohort based in San Francisco, CA. We measured 12 PFAS in serum samples collected in the second trimester and depressive symptom scores were assessed using the Center for Epidemiologic Studies Depression Scale. Associations were estimated using linear regression, adjusting for maternal age, education, pre-pregnancy body mass index, and parity. Associations with a PFAS mixture were estimated using quantile g-computation.

RESULTS

In adjusted linear regression models, a twofold increase in two PFAS was associated with higher depression scores in the overall sample, and this association persisted only among immigrant women (β [95 % confidence interval]: perfluorooctane sulfonic acid (2.7 [0.7-4.7]) and methyl-perfluorooctane sulfonamide acetic acid (2.9 [1.2-4.7]). Quantile g-computation indicated that simultaneously increasing all PFAS in the mixture by one quartile was associated with increased depressive symptoms among immigrant women (mean change per quartile increase = 1.12 [0.002, 2.3]), and associations were stronger compared to US born women (mean change per quartile increase = 0.09 [-1.0, 0.8]).

CONCLUSIONS

Findings provide new evidence that PFAS are associated with higher depression symptoms among immigrant women during pregnancy. Results can inform efforts to address environmental factors that may affect depression among US immigrants.

Air quality improvement and incident dementia: Effects of observed and hypothetical reductions in air pollutant using parametric g-computation

INTRODUCTION

No evidence exists about the impact of air pollution reduction on incidence of dementia. The aim of this study was to quantify how air quality improvement leads to dementia-incidence benefits.

METHODS

In the French Three-City cohort (12 years of follow-up), we used parametric g-computation to quantify the expected number of prevented dementia cases under different hypothetical interventions with particulate matter measuring <2.5 μm (PM_2.5 ) reductions.

RESULTS

Among 7051 participants, 789 participants developed dementia. The median PM_2.5 reduction between 1990 and 2000 was 12.2 (μg/m³ ). Such a reduction reduced the risk of all-cause dementia (hazard ratio [HR], 0.85; 95% confidence interval [CI], 0.76 to 0.95). If all study participants were enjoying a hypothetical reduction of more than 13.10 μg/m³ (median reduction observed in the city of Montpellier), the rate difference was -0.37 (95% CI, -0.57 to -0.17) and the rate ratio was 0.67 (95% CI, 0.50 to 0.84).

DISCUSSION

These findings highlight the possible substantial benefits of reducing air pollution in the prevention of dementia.

Childhood exposure to per- and polyfluoroalkyl substances and neurodevelopment in the CHARGE case-control study

BACKGROUND

Per- and polyfluoroalkyl substances (PFAS) are shown to have neurotoxic effects on animals, but epidemiological evidence for associations between childhood PFAS exposure and neurodevelopment is inconclusive. We examined if childhood PFAS concentrations are associated with a diagnosis of autism spectrum disorder (ASD), developmental delay (DD), and other early concerns (OEC) in development.

METHODS

We included 551 children 2-5 years old from the CHildhood Autism Risks from Genetics and Environment (CHARGE) case-control study. Children were clinically diagnosed and classified as having ASD, DD, OEC, and typical development (TD). Fourteen PFAS were quantified in child serum samples collected when diagnostic assessments were performed. We used multinomial logistic regression models to investigate the cross-sectional associations of individual PFAS concentrations with neurodevelopmental outcomes and weighted quantile sum (WQS) regression models with repeated holdout validation to investigate the associations with PFAS mixtures.

RESULTS

Childhood perfluorooctanoic acid (PFOA) was associated with increased odds of ASD (odds ratio [OR] per ln ng/mL increase: 1.99, 95% confidence interval [CI]: 1.20, 3.29) and DD (OR: 2.16, 95% CI: 1.21, 3.84) versus TD. Perfluoroheptanoic acid (PFHpA) was associated with increased odds of ASD (OR: 1.61, 95% CI: 1.21, 2.13). However, perfluroundecanoic acid (PFUnDA) was associated with decreased odds of ASD (OR: 0.43, 95% CI: 0.26, 0.69). From mixture analyses, the WQS index was associated with increased odds of ASD (average OR: 1.57, 5th and 95th percentile: 1.16, 2.13). Child's sex and homeownership modified associations of perfluorodecanoic acid (PFDA) with DD and ASD, respectively.

CONCLUSIONS

In this case-control study, childhood PFOA, PFHpA, and a PFAS mixture was associated with increased odds of ASD, while PFUnDA was associated with decreased odds of ASD. Because we used concurrent measurements of PFAS, our results do not imply causal relationships and thus need to be interpreted with caution.

State-of-the-art methods for exposure-health studies: Results from the exposome data challenge event

The exposome recognizes that individuals are exposed simultaneously to a multitude of different environmental factors and takes a holistic approach to the discovery of etiological factors for disease. However, challenges arise when trying to quantify the health effects of complex exposure mixtures. Analytical challenges include dealing with high dimensionality, studying the combined effects of these exposures and their interactions, integrating causal pathways, and integrating high-throughput omics layers. To tackle these challenges, the Barcelona Institute for Global Health (ISGlobal) held a data challenge event open to researchers from all over the world and from all expertises. Analysts had a chance to compete and apply state-of-the-art methods on a common partially simulated exposome dataset (based on real case data from the HELIX project) with multiple correlated exposure variables (P > 100 exposure variables) arising from general and personal environments at different time points, biological molecular data (multi-omics: DNA methylation, gene expression, proteins, metabolomics) and multiple clinical phenotypes in 1301 mother-child pairs. Most of the methods presented included feature selection or feature reduction to deal with the high dimensionality of the exposome dataset. Several approaches explicitly searched for combined effects of exposures and/or their interactions using linear index models or response surface methods, including Bayesian methods. Other methods dealt with the multi-omics dataset in mediation analyses using multiple-step approaches. Here we discuss features of the statistical models used and provide the data and codes used, so that analysts have examples of implementation and can learn how to use these methods. Overall, the exposome data challenge presented a unique opportunity for researchers from different disciplines to create and share state-of-the-art analytical methods, setting a new standard for open science in the exposome and environmental health field.

In-home environmental exposures predicted from geospatial characteristics of the built environment and electronic health records of children with asthma

PURPOSE

Children may be exposed to numerous in-home environmental exposures (IHEE) that trigger asthma exacerbations. Spatially linking social and environmental exposures to electronic health records (EHR) can aid exposure assessment, epidemiology, and clinical treatment, but EHR data on exposures are missing for many children with asthma. To address the issue, we predicted presence of indoor asthma trigger allergens, and estimated effects of their key geospatial predictors.

METHODS

Our study samples were comprised of children with asthma who provided self-reported IHEE data in EHR at a safety-net hospital in New England during 2004-2015. We used an ensemble machine learning algorithm and 86 multilevel features (e.g., individual, housing, neighborhood) to predict presence of cockroaches, rodents (mice or rats), mold, and bedroom carpeting/rugs in homes. We reduced dimensionality via elastic net regression and estimated effects by the G-computation causal inference method.

RESULTS

Our models reasonably predicted presence of cockroaches (area under receiver operating curves [AUC] = 0.65), rodents (AUC = 0.64), and bedroom carpeting/rugs (AUC = 0.64), but not mold (AUC = 0.54). In models adjusted for confounders, higher average household sizes in census tracts were associated with more reports of pests (cockroaches and rodents). Tax-exempt parcels were associated with more reports of cockroaches in homes. Living in a White-segregated neighborhood was linked with lower reported rodent presence, and mixed residential/commercial housing and newer buildings were associated with more reports of bedroom carpeting/rugs in bedrooms.

CONCLUSIONS

We innovatively applied a machine learning and causal inference mixture methodology to detail IHEE among children with asthma using EHR and geospatial data, which could have wide applicability and utility.

How to investigate human health effects related to exposure to mixtures of per- and polyfluoroalkyl substances: A systematic review of statistical methods

BACKGROUND

Humans are exposed to several per- and polyfluoroalkyl substances (PFAS) daily; however, most previous studies have focused on individual PFAS. Although attention to effects of exposure to mixtures of PFAS has grown in recent years, there is no consensus on the appropriate statistical methods that can be used to assess their combined effect on human health.

OBJECTIVES

We aim to perform a comprehensive review of the statistical methods used in the existing studies which evaluate the association between exposure to mixtures of PFAS and any adverse human health effect.

METHODS

The online databases PubMed, Embase and Scopus were searched for eligible studies, published during the last ten years (last search performed on April 08, 2021). Covidence software was used by two different reviewers to perform a title/abstract screening, followed by a full text revision of the selected papers.

RESULTS

A total of 3640 papers were identified, and after the screening process, 53 papers were included in the current review. Most of the studies were published between 2019 and 2021 and were conducted mainly in North America and Europe; more than half of the studies (28 out of 53) were conducted on mother and child pairs. WQS (Weighted Quantile Sum) Regression and BKMR (Bayesian Kernel Machine Regression) were used in 36 out of 53 papers to model mixtures' effects. Health outcomes included in the studies are immunotoxicity (n = 8), fetal development (n = 7), neurodevelopment (n = 9), reproductive hormones (n = 6), thyroid hormones (n = 7), outcomes related to metabolic pathways (n = 16).

CONCLUSION

Studies on human exposure to PFAS as complex mixtures and health consequences have substantially increased in the last few years. Based on our findings, we propose that addressing risk from PFAS mixtures will likely require combinations of approaches and implementation of constantly evolving statistical methods. Specific guidelines and tools for quality assessment and publication of mixture observational studies are warranted.

Imputation of Below Detection Limit Missing Data in Chemical Mixture Analysis with Bayesian Group Index Regression

There is growing scientific interest in identifying the multitude of chemical exposures related to human diseases through mixture analysis. In this paper, we address the issue of below detection limit (BDL) missing data in mixture analysis using Bayesian group index regression by treating both regression effects and missing BDL observations as parameters in a model estimated through a Markov chain Monte Carlo algorithm that we refer to as pseudo-Gibbs imputation. We compare this with other Bayesian imputation methods found in the literature (Multiple Imputation by Chained Equations and Sequential Full Bayes imputation) as well as with a non-Bayesian single-imputation method. To evaluate our proposed method, we conduct simulation studies with varying percentages of BDL missingness and strengths of association. We apply our method to the California Childhood Leukemia Study (CCLS) to estimate concentrations of chemicals in house dust in a mixture analysis of potential environmental risk factors for childhood leukemia. Our results indicate that pseudo-Gibbs imputation has superior power for exposure effects and sensitivity for identifying individual chemicals at high percentages of BDL missing data. In the CCLS, we found a significant positive association between concentrations of polycyclic aromatic hydrocarbons (PAHs) in homes and childhood leukemia as well as significant positive associations for polychlorinated biphenyls (PCBs) and herbicides among children from the highest quartile of household income. In conclusion, pseudo-Gibbs imputation addresses a commonly encountered problem in environmental epidemiology, providing practitioners the ability to jointly estimate the effects of multiple chemical exposures with high levels of BDL missingness.

Marine pollutant exposures and human milk extracellular vesicle-microRNAs in a mother-infant cohort from the Faroe Islands

BACKGROUND/AIMS

Early life exposures to marine contaminants can adversely impact child health but modes of action are unclear. Human milk contains extracellular vesicles (EVs) that can transport biologically relevant cargo from mother to infant, including microRNAs (miRNAs), and may partly mediate the effects of pollutants on child health. However, the role of marine pollutants on miRNA expression in milk EVs is unexplored.

METHODS

We isolated EV RNA from 333 milk samples collected between 2 and 74 days postpartum from a Faroese birth cohort born 1997-2000 and sequenced 2083 miRNAs using a targeted library preparation method. We quantified five perfluoroalkyl substances (PFAS), pesticide metabolite p,p'-dichlorodiphenyldichloroethylene (DDE), and the sum of three major polychlorinated biphenyls (ΣPCBs) in maternal serum at 34 weeks of gestation and maternal hair total mercury (Hg) at birth. We used negative binomial regressions to estimate associations between individual pollutants and 418 reliably expressed EV-miRNAs adjusted for potential confounders. We performed sparse principal components (PCs) analysis to derive the first four components of the EV-miRNA data and examined associations between pollutants and PCs using Bayesian kernel machine regression (BKMR).

RESULTS

We observed no associations between pollutants and individual EV-miRNA expression after controlling the false discovery rate at 0.1. However, BKMR suggested that Hg was positively associated with PC1 and negatively associated with PC3, while ΣPCBs was negatively associated with PC3, and two PFAS were associated with PC4. Exploration of PC loadings followed by pathway analyses suggested that miRNAs in PC1 (miR-200b-3p, miR-664a-3p, miR-6738-5p, miR-429, miR-1236-5p, miR-4464, and miR-30b-5p) may be related to Hg neurotoxicity, while remaining PCs require further research.

CONCLUSIONS

Our findings suggest that groups of milk EV-miRNAs may better serve as environmental biomarkers than individual miRNAs. Future studies are needed to elucidate the role of milk EV-miRNAs in child health following prenatal exposures.

Prenatal exposure to per- and polyfluoroalkyl substances (PFAS) and associations with attention-deficit/hyperactivity disorder and autism spectrum disorder in children

BACKGROUND

Prenatal exposure to per- and polyfluoroalkyl substances (PFAS) may be a risk factor for neurodevelopmental deficits and disorders, but evidence is inconsistent.

OBJECTIVES

We investigated whether prenatal exposure to PFAS were associated with childhood diagnosis of attention-deficit/hyperactivity disorder (ADHD) or autism spectrum disorder (ASD).

METHODS

This study was based on the Norwegian Mother, Father and Child Cohort Study and included n = 821 ADHD cases, n = 400 ASD cases and n = 980 controls. Diagnostic cases were identified by linkage with the Norwegian Patient Registry. In addition, we used data from the Medical Birth Registry of Norway. The study included the following PFAS measured in maternal plasma sampled mid-pregnancy: Perfluorooctanoic acid (PFOA), perfluorononanoic acid (PFNA), perfluorodecanoic acid (PFDA), perfluoroundecanoic acid (PFUnDA), perfluorohexane sulfonate (PFHxS), perfluoroheptanesulfonic acid (PFHpS), and perfluorooctane sulfonate (PFOS). Relationships between individual PFAS and ADHD or ASD diagnoses were examined using multivariable adjusted logistic regression models. We also tested for possible non-linear exposure-outcome associations. Further, we investigated the PFAS mixture associations with ASD and ADHD diagnoses using a quantile-based g-computation approach.

RESULTS

Odds of ASD was significantly elevated in PFOA quartile 2 [OR = 1.71 (95% CI: 1.20, 2.45)] compared to quartile 1, and PFOA appeared to have a non-linear, inverted U-shaped dose-response relationship with ASD. PFOA was also associated with increased odds of ADHD, mainly in quartile 2 [OR = 1.54 (95% CI: 1.16, 2.04)] compared to quartile 1, and displayed a non-linear relationship in the restricted cubic spline model. Several PFAS (PFUnDA, PFDA, and PFOS) were inversely associated with odds of ADHD and/or ASD. Some of the associations were modified by child sex and maternal education. The overall PFAS mixture was inversely associated with ASD [OR = 0.76 (95% CI: 0.64, 0.90)] as well as the carboxylate mixture [OR = 0.79 (95% CI: 0.68, 0.93)] and the sulfonate mixture [OR = 0.84 (95% CI: 0.73, 0.96)].

CONCLUSION

Prenatal exposure to PFOA was associated with increased risk of ASD and ADHD in children. For some PFAS, as well as their mixtures, there were inverse associations with ASD and/or ADHD. However, the inverse associations reported herein should not be interpreted as protective effects, but rather that there could be some unresolved confounding for these relationships. The epidemiologic literature linking PFAS exposures with neurodevelopmental outcomes is still inconclusive, suggesting the need for more research to elucidate the neurotoxicological potential of PFAS during early development.

Childhood exposure to per- and polyfluoroalkyl substances (PFAS) and neurobehavioral domains in children at age 8 years

BACKGROUND

Toxicological studies have raised concerns regarding the neurotoxic effects of per- and polyfluoroalkyl substances (PFAS). However, observational evidence from human studies investigating the association between childhood PFAS and neurobehavior is limited and remains unclear.

OBJECTIVES

To examine whether childhood PFAS concentrations are associated with neurobehavior in children at age 8 years and whether child sex modifies this relationship.

METHODS

We used data from 208 mother-child dyads in the Health Outcomes and Measures of the Environment (HOME) Study, a prospective pregnancy and birth cohort (Cincinnati, OH, USA). We quantified PFAS in child serum at 3 and 8 years. We assessed neurobehavioral domains using the Behavior Assessment System for Children-2 at 8 years. We used multiple informant models to estimate score changes per ln-increase in repeated PFAS concentrations.

RESULTS

Childhood PFAS were not associated with Externalizing or Internalizing Problems at 8 years. However, we noted effect measure modification by sex, with higher scores in Externalizing Problems among males per ln-unit increase in perfluorononanoate (PFNA) at 3 years (β = 4.3 points, 95% CI: 1.0, 7.7) while females had lower scores (β = -2.8 points, 95% CI: -4.7, -1.0). More Internalizing Problems were observed among males per ln-unit increase in concurrent PFNA concentrations (β = 3.7 points, 95% CI: 0.7, 6.8), but not in females (β = -1.7 points, 95% CI: -4.6, 1.2). Childhood PFNA concentrations were associated with lower scores for attention problems and activity of daily living.

CONCLUSION

While findings do not consistently support an association between childhood PFAS serum concentrations and neurobehavior, child sex may play a role in this relationship.

Prenatal exposure to per- and polyfluoroalkyl substances (PFAS) and neurobehavior in US children through 8 years of age: The HOME study

BACKGROUND

Studies of prenatal per- and polyfluoroalkyl substances (PFAS) and attention deficit hyperactivity disorder (ADHD)-related behaviors in children are inconsistent.

OBJECTIVES

To examine associations between maternal serum PFAS concentrations and child behavior in 241 mother-child dyads within the Health Outcomes and Measures of the Environment (HOME) Study.

METHODS

We quantified perfluorooctanoate (PFOA), perfluorooctane sulfonate (PFOS), perfluorohexane sulfonate (PFHxS), and perfluorononanoate (PFNA) in maternal serum collected during pregnancy or at delivery. We evaluated a total of 17 outcomes of child behavior using the Behavioral Assessment System for Children-2 (BASC-2) at 5 and 8 years (n = 240) and ADHD diagnostic symptoms and criteria with the Diagnostic Interview Schedule for Children-Young Child (DISC-YC) at 5 years (n = 190). We used linear mixed models and logistic regression with generalized estimating equations to assess associations between PFAS and continuous or dichotomous "at risk" BASC-2 scores; negative binomial regression to calculate incident rate ratios for counts of ADHD symptoms; and Poisson regression with robust standard errors to calculate relative risks of meeting ADHD diagnostic criteria.

RESULTS

Each ln-unit increase in PFOS, PFHxS, and PFNA was associated with higher BASC-2 scores and increased odds of "at-risk" scores for externalizing behaviors, including hyperactivity (PFOS: odds ratio [OR] 2.7, 95% confidence interval [CI] 1.2, 5.9; PFHxS: OR 2.5, 95% CI 1.5, 4.3; PFNA: OR 3.2, 95% CI 1.3, 8.0). PFHxS was also associated with internalizing problems (OR 2.0, 95% CI 1.1, 3.4) and somatization (OR 2.2, 95% CI 1.2, 4.0). PFOS and PFNA were significantly associated with 50-80% more DISC-YC symptoms and diagnostic criteria related to hyperactive-impulsive type ADHD. Prenatal PFNA was associated with increased risk of any-type ADHD.

CONCLUSIONS

Prenatal PFOS and PFNA were consistently associated with measures related to hyperactive-impulsive type ADHD across two validated assessment instruments. PFHxS was associated with increased problems with both externalizing and internalizing behaviors. No associations were noted between PFOA and child neurobehavior.

Machine Learning Within Studies of Early-Life Environmental Exposures and Child Health: Review of the Current Literature and Discussion of Next Steps

PURPOSE OF REVIEW

The goal of this article is to review the use of machine learning (ML) within studies of environmental exposures and children's health, identify common themes across studies, and provide recommendations to advance their use in research and practice.

RECENT FINDINGS

We identified 42 articles reporting upon the use of ML within studies of environmental exposures and children's health between 2017 and 2019. The common themes among the articles were analysis of mixture data, exposure prediction, disease prediction and forecasting, analysis of complex data, and causal inference. With the increasing complexity of environmental health data, we anticipate greater use of ML to address the challenges that cannot be handled by traditional analytics. In order for these methods to beneficially impact public health, the ML techniques we use need to be appropriate for our study questions, rigorously evaluated and reported in a way that can be critically assessed by the scientific community.

Maternal Experience of Multiple Hardships and Fetal Growth: Extending Environmental Mixtures Methodology to Social Exposures

BACKGROUND

Women can be exposed to a multitude of hardships before and during pregnancy that may affect fetal growth, but previous approaches have not analyzed them jointly as social exposure mixtures.

METHODS

We evaluated the independent, mutually adjusted, and pairwise joint associations between self-reported hardships and birthweight for gestational age z-scores in the Chemicals in Our Bodies-2 prospective birth cohort (N = 510) using G-computation. We examined financial hardship, food insecurity, job strain, poor neighborhood environment, low community standing, caregiving, high burden of stressful life events, and unplanned pregnancy collected via questionnaire administered in the second trimester of pregnancy. We used propensity scores to ensure our analyses had sufficient data support and estimated absolute differences in outcomes.

RESULTS

Food insecurity was most strongly associated with reduced birthweight for gestational age z-scores individually, with an absolute difference of -0.16, 95% confidence interval (CI) -0.45, 0.14. We observed an unexpected increase in z-scores associated with poor perceived neighborhood environment (0.18, 95% CI -0.04, 0.41). Accounting for coexposures resulted in similar findings. The pairwise joint effects were strongest for food insecurity in combination with unplanned pregnancy (-0.45, 95% CI -0.93, 0.02) and stressful life events (-0.42, 95% CI -0.90, 0.05). Poor neighborhood environment in combination with caregiving was associated with an increase in z-scores (0.47, 95% CI -0.01, 0.95).

CONCLUSIONS

Our results are consistent with the hypothesis that experiencing food insecurity during pregnancy, alone and in combination with stressful life events and unplanned pregnancy, may affect fetal growth.