Gestational exposure to organophosphate ester flame retardants and risk of childhood obesity in the environmental influences on child health outcomes consortium

Abstract

Introduction: Organophosphate esters (OPEs) are increasing in use as flame retardants and plasticizers and concerns have been raised given their endocrine-disrupting activities and possible obesogenic consequences. However, longitudinal studies on gestational OPE exposure and childhood obesity are scarce. This study examined whether OPE levels in maternal urine during pregnancy were associated with the risk of childhood obesity.

Methods: OPEs were analyzed in pregnancy urine samples of 5,087 individuals from 14 studies contributing to the Environmental influences on Child Health Outcomes (ECHO) Cohort. BDCPP, DBUP/DIBP, and DPHP, detected in > 80 % of the samples, were modeled continuously and by tertiles; whereas BCPP, BBOEP, and BCETP, detected in 50-80 % of samples, were modeled categorically (not-detected, low, and high). Childhood obesity was defined by BMI z-score ≥ 95th percentile according to WHO (<2 years) and the CDC (≥2 years) metrics. Adjusted modified Poisson regression models assessed childhood obesity risk and the mixture effect was assessed using Bayesian kernel machine regression (BKMR).

Results: BMI measurements were available for 3,827 children in infancy (0.5-1.9 years), 3,921 children in early childhood (2.0-4.9 years), and 2,541 children in mid-childhood (5.0-10.0 years). Obesity was present in 16-21 % of children across age groups. In mid-childhood DBUP/DIBP second and third versus first tertiles were associated with increased obesity risk (RR 1.14; 95 % CI: 1.02, 1.28; and RR 1.11; 95 % CI: 0.97, 1.27; respectively); whereas BDCPP second and third versus first tertiles reflected an inverse association with obesity risk (RR 0.85; 95 % CI: 0.80, 0.91 and RR 0.91; 95 % CI: 0.77, 1.07; respectively). No association with obesity risk was observed for DPHP, BCPP, BBOEP, and BCETP. Directions observed were consistent with those seen in BKMR models.

Conclusions: This study identified mixed associations between gestational OPE exposure and childhood obesity. Further investigation across a comprehensive range of OPE exposures is warranted.

Keywords: Childhood obesity; Flame retardants; OPE; Pregnancy.

Fig. 1.

Adjusted associations of gestational BDCPP, DBUP/DIBP, and DPHP tertiles with childhood obesity in infancy, early childhood and middle childhood. Note: Regression models reflect multiple imputation by chained equations (MICE) for covariates and were adjusted for urinary dilution, ECHO site, maternal age at birth, maternal race/ethnicity, maternal educational attainment, pre-pregnancy BMI, maternal smoking in pregnancy, parity, and child’s year of birth. Infancy 0.5–1.9 years, Early Childhood 2.0–4.9 years, Middle Childhood 5.0–10.0 years. Obesity defined ≥ 95th percentile. DPHP, diphenyl phosphate; DBUP/DIBP, composite of dibutyl phosphate and di-isobutyl phosphate; BDCPP, bis(1,3-dichloro-2-propyl) phosphate.

Fig. 2.

Adjusted associations of gestational BBOEP, BCETP, and BCPP with childhood obesity in infancy, early childhood and middle childhood. Note: Regression models reflect multiple imputation by chained equation (MICE) for covariates and were adjusted for urinary dilution, ECHO site, maternal age at birth, maternal race/ethnicity, maternal educational attainment, pre-pregnancy BMI, maternal smoking during pregnancy, parity, and child’s year of birth. Infancy 0.5–1.9 years, Early Childhood 2.0–4.9 years, Middle Childhood 5.0–10.0 years. Obesity defined ≥ 95th percentile. BCETP, bis(2-chloroethyl) phosphate; BBOEP, bis(butoxyethyl) phosphate; BCPP, bis(1-chloro-2-propyl) phosphate; LOD Limit of Detection.

Fig. 3.

Exposure response function for each organophosphate esters (OPE) from Bayesian kernel machine regression models in middle childhood.

Fig. 4.

Adjusted associations of gestational BDCPP, DBUP/DIBP, and DPHP tertiles with childhood obesity in infancy, early childhood and middle childhood by child sex. Note: Regression models reflect multiple imputation by chained equations (MICE) for covariates and were adjusted for urinary dilution, ECHO site, maternal age at birth, maternal race/ethnicity, maternal educational attainment, pre-pregnancy BMI, maternal smoking in pregnancy, parity, and child’s year of birth. Infancy 0.5–1.9 years, Early Childhood 2.0–4.9 years, Middle Childhood 5.0–10.0 years. Obesity defined ≥ 95th percentile. DPHP, diphenyl phosphate; DBUP/DIBP, composite of dibutyl phosphate and di-isobutyl phosphate; BDCPP, bis(1,3-dichloro-2-propyl) phosphate. T = Tertile.