Prenatal exposure to a wide range of environmental chemicals and child behaviour between 3 and 7 years of age - An exposome-based approach in 5 European cohorts

Abstract

Background: Studies looking at associations between environmental chemicals and child behaviour usually consider only one exposure or family of exposures.

Objective: This study explores associations between prenatal exposure to a wide range of environmental chemicals and child behaviour.

Methods: We studied 708 mother-child pairs from five European cohorts recruited in 2003-2009. We assessed 47 exposure biomarkers from eight chemical exposure families in maternal blood or urine collected during pregnancy. We used the Strengths and Difficulties Questionnaire (SDQ) to evaluate child behaviour between three and seven years of age. We assessed associations of SDQ scores with exposures using an adjusted least absolute shrinkage and selection operator (LASSO) considering all exposures simultaneously and an adjusted exposome-wide association study (ExWAS) considering each exposure independently.

Results: LASSO selected only copper (Cu) as associated with externalizing behaviour. In the ExWAS, bisphenol A [BPA, incidence rate ratio (IRR): 1.06, 95% confidence interval (95%CI): 1.01;1.12] and mono-n-butyl phthalate (MnBP, IRR: 1.06, 95%CI: 1.00;1.13) were associated with greater risk of externalizing behaviour problems. Cu (IRR: 0.90, 95%CI: 0.82;0.98), perfluoroundecanoate (PFUnDA, IRR: 0.92, 95%CI: 0.84;0.99) and organochlorine compounds (OCs) were associated with lower risk of externalizing behaviour problems, however the associations with OCs were mainly seen among women with insufficient weight gain during pregnancy. Internalizing score worsen in association with exposure to diethyl thiophosphate (DETP, IRR: 1.11, 95%CI: 1.00;1.24) but the effect was driven by the smallest cohort. Internalizing score improved with increased concentration of perfluorooctane sulfonate (PFOS, IRR: 0.92, 95%CI: 0.85;1.00), however the association was driven by the two smallest cohorts with the lowest PFOS concentrations.

Discussion: This study added evidence on deleterious effects of prenatal exposure to BPA and MnBP on child behaviour. Other associations should be interpreted cautiously since they were not consistent with previous studies or they have not been studied extensively.

Keywords: Birth cohort; Child behaviour; Internal exposome; Prenatal exposure; Strengths and Difficulties Questionnaire.

Graphical abstract

Fig. 1

Sensitivity analysis (n = 708. BiB n = 46; EDEN n = 193; INMA n = 218; KANC n = 83; RHEA n = 168). Cohort-specific associations between prenatal exposures and SDQ externalizing (A) and internalizing (B) scores detected by the ExWAS (p value of association <0.05 except of diethyl thiophosphate for which the p value = 0.053). Regression models were adjusted for cohort, season of conception, child sex and age at SDQ assessment, parity, maternal: education level, work status, age, pre-pregnancy BMI and prenatal active smoking status. The “All cohorts” estimates are those obtained in the main ExWAS. IRRs are reported with 95%CIs and correspond to the change in the probability of the SDQ scores increasing by one unit for an IQR change in the log₂ of the biomarker concentration in maternal blood or urine. We relied on the following threshold for I² interpretation: I² < 0.3 low heterogeneity, 0.3 ≤ I² < 0.6 moderate heterogeneity, I² ≥ 0.6 substantial to high heterogeneity. The black squares display the IRRs (size of the square reflects the relative size of each cohort) and the horizontal lines their 95%CIs. Abbreviations: BiB = Born in Bradford. EDEN = Étude des Déterminants Pré et Postnatals du Développement et de la Santé de l'Enfant. INMA = Infancia y Medio Ambiente. KANC = Kaunas Cohort. RHEA = Mother-Child Cohort in Crete. CI = confidence interval of the IRR estimate. ExWAS = exposome-wide association study. IQR = inter-quartile range. IRR = incidence rate ratio. BMI = body mass index. SDQ = Strengths and Difficulties Questionnaire.

Fig. 2

Sensitivity analysis for exposure-SDQ externalizing score associations stratified on gestational weight gain. We stratified on three categories of gestational weight gain as defined in the revised recommendations of the US Institute of Medicine (Institute of Medicine (US) and National Research Council (US) Committee to Reexamine IOM Pregnancy Weight Guidelines 2009): adequate (in black), excessive (in green) and insufficient (in blue). For pre-pregnancy BMI < 18.5 kg/m² recommended total weight gain was 12.5–18.0 kg, for BMI = 18.5–24.9 kg/m²: 11.5–16.0 kg, for BMI = 25.0–29.9 kg/m²: 7.0–11.5 kg and for BMI ≥ 30.0 kg/m² recommended weight gain was 5.0–9.0 kg. We ran one negative binomial regression model per exposure and outcome for each sub-population. IRRs are reported with 95%CIs and correspond to the change in the probability of the SDQ scores increasing by one unit for an IQR change in the log₂ of the biomarker concentration in maternal blood. Each point represents the IRR estimate and the vertical line its 95%CI. Regression models were adjusted for cohort, season of conception, child sex and age at SDQ assessment, parity, and maternal factors: education level, work status, age and prenatal active smoking status. Abbreviations: BMI = body mass index. CI = confidence interval of the IRR estimate. IQR = inter-quartile range. IRR = incidence rate ratio. SDQ = Strengths and Difficulties Questionnaire. DDE = dichlorodiphenyldichloroethylene. DDT = dichlorodiphenyltrichloroethane. HCB = hexachlorobenzene. PCB = polychlorinated biphenyl. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)