Polybrominated Diphenyl Ethers (PBDEs) in Breast Milk and Neuropsychological Development in Infants

Background: There is increasing interest in the potential effects of polybrominated diphenyl ethers (PBDEs) on children’s neuropsychological development, but only a few small studies have evaluated such effects. Objectives: Our goal was to examine the association between PBDE concentrations in colostrum and infant neuropsychological development and to assess the influence of other persistent organic pollutants (POPs) on such association. Methods: We measured concentrations of PBDEs and other POPs in colostrum samples of 290 women recruited in a Spanish birth cohort. We tested children for mental and psychomotor development with the Bayley Scales of Infant Development at 12–18 months of age. We analyzed the sum of the seven most common PBDE congeners (BDEs 47, 99, 100, 153, 154, 183, 209) and each congener separately. Results: Increasing Σ7PBDEs concentrations showed an association of borderline statistical significance with decreasing mental development scores (β per log ng/g lipid = –2.25; 95% CI: –4.75, 0.26). BDE-209, the congener present in highest concentrations, appeared to be the main congener responsible for this association (β = –2.40, 95% CI: –4.79, –0.01). There was little evidence for an association with psychomotor development. After adjustment for other POPs, the BDE-209 association with mental development score became slightly weaker (β = –2.10, 95% CI: –4.66, 0.46). Conclusions: Our findings suggest an association between increasing PBDE concentrations in colostrum and a worse infant mental development, particularly for BDE-209, but require confirmation in larger studies. The association, if causal, may be due to unmeasured BDE-209 metabolites, including OH-PBDEs (hydroxylated PBDEs), which are more toxic, more stable, and more likely to cross the placenta and to easily reach the brain than BDE-209.


Exposure assessment: liquid-liquid extraction protocol to measure POPs in colostrum and description of the limitis of detection (LOD) and quantification (LOQ).
Colostrum was collected the first 48 to 96 hours' postpartum at the hospital by an experienced nurse. The sample was collected in the morning at the end of the feeding, in sterile polypropylene tubes, by mechanical expression of one breast using a breast pump. Milk was transported to the laboratory in ice boxes less than 2 hours after collection, where samples were stored at -80°C until analysis. OCs (DDE,HCB and P CB congeners 28,52,101,118,138,153,180) and PBDEs (congeners 17,28,47,66,71,85,99,100,138,153,154,183,190,209) were extracted following a liquid-liquid extraction protocol. Colostrum (1 mL), n-hexane (3 mL) and concentrated H 2 SO 4 (3 mL) were added in 10 mL centrifuge glass tubes. The mixture, spiked with 2-ethylhexyl-2,3,4,5-tetrabromobenzoate (TBB) and PCB 209 as surrogate standards, was mixed (vortex, ca. 1500 rpm, 30 s) and centrifuged (ca. 3500 rpm, 10 m). Supernatant was separated, and 2 mL of hexane were added to the resting acid fraction, mixed in vortex, centrifuged twice, and yielded together with the first n-hexane fraction. Concentrated H 2 SO 4 (3mL) was added to the final volume of 7 mL of extract, mixed with vortex (ca. 1500, 90 s) and centrifuged (ca. 3500, 20 m). Clean extract was then evaporated under a gentle N 2 stream to near dryness and dissolved with isooctane to a glass chromatographic vial. Prior to injection, the extract was evaporated again and 100 µL of PCB142 were added as an internal standard for OCs. The extract (2 µL) was injected automatically in a gas chromatograph with electron capture detection (GC/ECD model 6890, Agilent; Palo Alto, CA) in split-spitless mode, with a 60m DB-5 column (J&W/Agilent) using helium as carrier gas (1.5 mL/min). The temperature program was 90°C for 2 min, 15 °C/min to 130°C, and finally 4 °C/min to 290°C, held for 15 min.
Quantification was performed using PCB142 as an internal standard to correct for volume.

Description of the imputation procedure.
• Software used and key setting: STATA 10.1 software (Stata Corporation, College Station, Texas) -ice command (with 20 cycles). •

Statistical interactions included in imputation models: imputations were done separately by each region (Sabadell and Gipuzkoa).
Figure S1. Generalized additive models (GAM) for the associations between log Σ7PBDEs and mental and psychomotor scores.
Mental score model adjusted for: region of study, sex, region of origin of the mother, gestational age, low birth weight, maternal social class and studies, day-care attendance and parity (p for non-linearity=0.12). Psychomotor score model adjusted for: region of study, sex, region of origin of the mother, gestational age, low birth weight, maternal social class and studies and maternal pre-pregnancy BMI (p for non-linearity=0.39).