Prenatal Exposure to Polybrominated Diphenyl Ether (PBDE) and Child Neurodevelopment: The Role of Breastfeeding Duration

Abstract

Background:

Prenatal and early-life exposure to polybrominated diphenyl ethers (PBDEs) is associated with detrimental and irreversible neurodevelopmental health outcomes during childhood. Breastfeeding may be a child’s largest sustained exposure to PBDE— potentially exacerbating their risk for adverse neurodevelopment outcomes. However, breastfeeding has also been associated with positive neurodevelopment. Our study investigates if breastfeeding mitigates or exacerbates the known adverse effects of prenatal exposure to PBDEs and child neurodevelopment.

Methods:

Participants included 321 mother-infant dyads from the Center for the Health Assessment of Mothers and Children of Salinas (CHAMACOS), a longitudinal birth cohort in California. PBDE concentrations were measured in maternal serum blood samples collected during pregnancy or at delivery. Using generalized estimated equations (GEE), we estimated associations of PBDE concentrations with children’s attention, executive function, and cognitive scores assessed longitudinally between 7-12 years of age, stratified by duration of exclusive and complementary breastfeeding.

Results:

We observed that higher maternal prenatal PBDE concentrations were associated with poorer executive function among children who were complementary breastfed for a shorter duration compared to children breastfed for a longer duration; preservative errors (β for 10-fold increase in complementary breastfeeding < 7 months = −6.6; 95% Confidence Interval (CI): −11.4, −1.8; β ≥ 7 months = −5.1; 95% CI: −10.2, 0.1) and global executive composition (β for 10-fold increase < 7 months = 4.3; 95% CI: 0.4, 8.2; β for 10-fold increase ≥ 7 months = 0.6; 95% CI: −2.8, 3.9).

Conclusions:

Prolonged breastfeeding does not exacerbate but may mitigate some previously observed negative associations of prenatal PBDE exposure and child neurodevelopment.

Graphical abstract

1. Introduction

Historical widespread use of Polybrominated Diphenyl Ether (PBDEs) in households and major consumer products, such as appliances and furniture, has led to ubiquitous contamination of the environment.¹ Although PBDEs are almost universally banned,² many countries including the European Union, Brazil, Canada, Japan, South Korea, and Turkey still allow recycling and re-use of PBDE contaminated goods.³ Humans are exposed to PBDEs through multiple sources, including contaminated air, dust, food, and/or soil.⁴ PBDEs are lipophilic and persistent⁵ with a half-life of 1-12 years in humans,⁶ leading to a rapid increase and pervasive body burden of PBDE’s in human populations.^2,7,8 PBDEs are found in blood and breast milk samples of pregnant and lactating persons, creating a potential exposure and hazard to the development of the fetus and infant.^7–11

Pre- and post-natal PBDE exposure has been associated with a myriad of adverse childhood neurodevelopmental outcomes,^12,13 including lower IQ,^14–16 attention deficit/hyperactivity disorder (ADHD) and related behaviors, ^14,16–18 and poorer fine motor coordination.^12,16 In previous analyses from our Center for the Health Assessment of Mothers and Children of Salinas (CHAMACOS) birth cohort, we observed associations of higher prenatal and child PBDE concentrations with poorer attention, executive function, and cognition measured at ages 5-12 years; however, breastfeeding duration was not investigated as a moderating factor for these relationships.^16,18

Human milk is the main source of PBDE exposure for breastfeeding infants and may be a child’s largest sustained exposure to PBDEs— potentially exacerbating any risk for adverse neurodevelopment outcomes from in utero exposures.¹⁹ PBDEs are easily offloaded in breastmilk, with reported concentrations from 2,000 to 11,000 pg/g lipid.^20,21 Higher serum levels of PBDE are found among children who: initiated breastfeeding ²² and had a longer duration of exclusive breastfeeding.¹⁹ Yet, research has consistently shown that a prolonged duration of both exclusive and complementary breastfeeding is associated with improved childhood neurodevelopment, including better cognitive and socio-emotional development,²³ fewer ADHD symptoms,^24,25 and better receptive language and nonverbal intelligence.²⁶

Given the potential exposure to PBDEs through breastmilk but the beneficial effect of breastmilk on neurodevelopment, it is unclear whether exclusive or complementary breastfeeding offsets or exacerbates the adverse neurodevelopmental effects of prenatal PBDE exposure. Understanding this relationship is imperative for informing new mothers, especially those with potentially high exposure to PBDEs and other liphophilic comopunds.²⁷ The goal of our analysis was to examine if previously reported adverse associations of prenatal PBDE exposure with attention, executive function, and cognition^16,18 are modified by exclusive or complementary breastfeeding duration in the CHAMACOS cohort.

2. Methods

2.1. Study population and design

The CHAMACOS study population resides in the Salinas Valley of California and includes mostly Spanish speaking, low-income agricultural families. Pregnant women were recruited in 1999-2000 if they were at least 18 years old, less than 20 weeks gestation, Spanish or English speaking, qualified for California’s low-income health insurance (Medi-Cal), and planned to deliver at the local public hospital (N=601).²⁸ Of the recruited pregnant women, 337 remained in the study until age 9. We also excluded twins (n=10) and children with autism, down syndrome, cerebral palsy, hydrocephalus, or deafness (n=7) to be consistent with our previous studies,^16,18 leaving 321 mother-infant dyads for analysis. Our analysis is limited to participants with PBDE data collected during pregnancy or at delivery (n=318) and who completed at least one neurodevelopmental assessment at ages 7 (n=312), 9 (n=311), 10.5 (n=312), or 12 years (n=315).

2.2. PBDE Exposure Assessment

A detailed description of maternal serum blood collection^29,30 and serum PBDE analysis¹⁹ was previously reported. Briefly, maternal blood samples were collected during pregnancy (~26 weeks gestation) (n=205). For participants that did not have measurements taken during pregnancy, PBDE levels were back-extrapolated from measurements at delivery (n=57) or at 9 years postpartum (n=56). We measured for 10 PBDE congeners at the Centers for Disease Control and Prevention (CDC). Additional information on the back-extrapolation methods can be found elsewhere.¹⁸ We previously reported a very strong correlation (Pearson r ≥ 0.98, p < 0.001) between the two measurements points (i.e., during pregnancy and delivery)^12,19 — suggesting that PBDE concentrations measured during pregnancy are an accurate indication of PBDE maternal levels when initiation of breastfeeding occurs (shortly after delivery; ~1 hour post-delivery). As in previous analyses,^18,19,29 we examined PBDE concentrations as the log₁₀ lipid-adjusted sum of 4 PBDE congeners (−47, −99, −100, and −153).

2.3. Neurodevelopmental assessment

Child attention, executive function, and cognitive scores were assessed longitudinally at ages 7, 9, 10.5, and 12 years, as previously described.¹⁸ We analyzed modification by breastfeeding duration for outcomes that were adversely associated (p<0.05) with PBDE exposure in our previous analyses of PBDE and neurodevelopment in CHAMACOS.^16,18 Assessments were administered in a private room free of distraction by trained bilingual (Spanish and English) psychometricians trained by a clinical neuropsychologist.

Cognition was assessed at 7 and 10.5 years using the Wechsler Intelligence Scale³¹ for Children–Fourth Edition (WISC-IV).³² We examined WISC-IV scores for the Full Scale Intelligence Quotient (FSIQ) and Verbal Comprehension Intelligence Quotient (VCIQ) subscales (mean (M)=100, standard deviation (SD)=15).

At ages 9 and 12 years, children completed the Conners’ Continuous Performance Test (CPT),³³ a computerized assessment that assesses hit rate, accuracy, and impulse control. We examined sex- and age-standardized T-scores (M=50, SD=10) for errors of omission (false negatives) and the hit rate standard error (SE) over all scales. Higher variability in hit rate indicates performance inconsistency, a symptom of ADHD.³⁴ At ages 9 and 12, parents also completed the Conners’ ADHD/DSM-IV Scales (CADS).³⁵ We examined age- and sex-standardized T-scores for the ADHD index, designed to identify children “at risk” for ADHD, and the DSM-IV-based Inattentive subscale (M=50, SD=10).

Executive function was assessed using the computerized Wisconsin Card Sort Task (WCST)³⁶ and the parent-reported Behavior Rating Inventory of Executive Function (BRIEF)³⁷ at ages 9 and 12. The WCST is a computerized assessment that measures skills around strategic planning, ability to shift cognitive strategies, and impulse control. We examined age- and sex-standardized T-scores (M=50, SD=10) for perseverative errors from WCST and for the summary Global Executive Composite Score from the BRIEF.

2.4. Breastfeeding

Mothers self-reported breastfeeding duration at four in-person interviews during the postpartum period: at the hospital directly following delivery, and at 6, 12, and 24 months postpartum. During each interview, the mother was asked if she was currently breastfeeding. If the mother reported she was no longer breastfeeding, she was asked the total duration of breastfeeding in months. PBDE exposure through breastmilk is conditional upon amount of milk received; therefore, we analyzed exclusive breastfeeding and complementary breastfeeding separately. We adhered to a strict definition of exclusive breastfeeding, defined as infants who only received breastmilk — which is consistent with the World Health Organization (WHO) and allows comparability among studies.^38,39 Additional questions were used to determine exclusivity, including age at first introduction of formula, solid foods, or liquids other than breastmilk. Complementary breastfeeding allowed receipt of breastmilk combined with other liquids, supplements, and/or foods.

The WHO recommends exclusive breastfeeding for the first six months of life.⁴⁰ In high-income countries where alternative infant nutrition is available and safe, there is less consensus on which duration of exclusive breastfeeding is important for child health and development.⁴¹ We investigated two different durations of exclusive and complementary breastfeeding to increase evidence supporting the marginal benefits of each additional month of breastfeeding on child health in a high-income country. A dichotomous variable was created for both exclusive and complementary breastfeeding using the median duration (≥1. 4 and ≥7 months, respectively). In our primary analyses, we excluded 24 participants who did not initiate breastfeeding or were missing data on exclusive breastfeeding duration in the exclusive breastfeeding analyses and 13 participants who did not initiate complementary breastfeeding in the complementary breastfeeding analyses; we included all participants, regardless of their breastfeeding initiation status, in sensitivity analyses.

2.5. Statistical Analysis

We used Generalized Estimated Equation (GEE) models to examine associations between PBDE concentrations and child attention, executive function, and cognition outcomes by exclusive and complementary breastfeeding duration. GEE models were used to examine longitudinal associations with outcomes at multiple time points.⁴² We examined breastfeeding duration as an effect modifier by using a product-interaction term between breastfeeding (< vs. ≥ median duration) and PBDE concentrations and considered there to be evidence of statistical significance if the interaction p-value was <0.2.

We included confounders used in the previous CHAMACOS analysis of prenatal PBDE exposure and neurodevelopment, which were selected using a Directed Acyclic Graph (DAG).¹⁸ Details on measurement of these covariates have been previously described,^16,18 and were obtained from structured questionnaires administered to mothers by study staff at each visit. All models controlled for child sex; parity; preschool attendance; age at neurodevelopmental assessment; maternal age, receptive language (Peabody Picture Vocabulary Test (PPVT) score) at child age 9, and depressive symptoms (Centers for Epidemiologic Studies Depression Scale (CES-D)) at child age 9; maternal prenatal smoking status; family structure (father present/absent) at time of assessment; household income; the Home Observation for the Measurement of the Environment-Short Form (HOME-SF)⁴³ score at child age 10.5 which measures household enrichment; psychometrician or study interviewer who administered the task or rating scale; time of day assessment occurred; and child video game usage at age 9.

3. Results

Demographic characteristics are displayed in Table 1. The majority of CHAMACOS mothers were born in Mexico or another country outside of the U.S. (87.5%), had a median age of 26 years at birth, did not graduate high school (78.8%), and were living at or below the poverty level when the child was 9 years old (68.6%). Nearly all mothers initiated breastfeeding (95.9%). About half of the mothers exclusively breastfed more than one month postpartum (47.5%) with 42.8% exclusively breastfeeding for at least 1 month to 6 months and about half of the mothers complementary breastfed more than 6 months (49.7%) with 27.7% complementary breastfeeding more than 12 months. The distribution of prenatal PBDE concentrations among participants is shown in Supplementary Table 1. The highest serum concentrations were for PBDE −47 (geometric mean (95% confidence interval (CI), 15.6 (14.0, 17.3)). In total, serum concentrations of all four PBDE −47, −99, −100, 153 congeners ranged from 2.6 to 1,293.7 ng/g lipid.

Table 1.

Sociodemographic characteristics of CHAMACOS study participants (n=318).

Characteristics	n (%) or median (P25, P75)
Maternal/household characteristics
Age at delivery (years)	26 (23, 30)
Country of birth
Mexico or other	280 (88.0)
U.S.	38 (12.0)
Years in the U.S. at delivery
≤5 years	156 (49.1)
>5 years, but not born in U.S.	131 (41.1)
Born in U.S.	31 (9.8)
Education at baseline
≤6th grade	141 (44.3)
7th-12th grade	110 (34.6)
≥High school graduate	67 (21.1)
Parity at child delivery a
0	97 (30.5)
1+	221 (69.5)
Marital status at baseline
Not married/not living as married	55 (17.3)
Married/living as married	263 (82.7)
Maternal intelligence (PPVT Score) at 9-year visit
≤ 74	65 (20.4)
75-99	109 (34.3)
≥100	144 (45.3)
Maternal depression at 9-year visit (≥16 CES-D score)a
No	246 (77.4)
Yes	72 (22.6)
Family poverty at age 9 a
At or below poverty level	221 (69.5)
Above poverty level	97 (30.5)
Father present in home at 9-year visit
Father not present	74 (23.3)
Father present	244 (76.7)
HOME z-score at 10.5-year assessmenta	0.2 (−0.6, 0.8)
Child characteristics
Child’s sex
Boy	149 (46.9)
Girl	169 (53.1)
Complementary breastfeeding duration
Never breastfed	13 (4.1)
≤ 1 month	41 (12.9)
> 1 - ≤6 months	106 (33.3)
> 6 - ≤12 months	70 (22.0)
> 12 months	88 (27.7)
Exclusive breastfeeding duration
Never breastfed	13 (4.1)
≥ 1 month	144 (45.3)
> 1 - ≤6 months	136 (42.8)
> 6 - ≤_12 months	15 (4.7)
> 12 months	0 (0.0)
Missing	10 (3.1)
History of preschool attendance
Did not attend	98 (30.8)
Attended preschool	220 (69.2)
Video game usage at age 9
Never played video games	80 (26.1)
Played video games	226 (73.9)
Log10 Prenatal PBDE Concentration (ng/g lipid)	1.4 (1.2, 1.6)

^aMissing data imputed for PPVT (n=3), parity (n=1), household income (n=1), HOME score (n=16)

P=percent; U.S.=United States; PPVT= Peabody Picture Vocabulary Test; CES-D= Centers for Epidemiologic Studies Depression Scale; Home= HOME z-score is the Home Observation for the Measurement of the Environment-Short Form (HOME-SF) that measures household enrichment; PBDE= Prenatal serum concentrations of ΣPolybrominated Diphenyl Ethers (−47,−99,−100,−153)

Table 2 displays the mean, median, and range of child attention, executive function, and cognitive scores among participants at ages 7, 9, 10.5, and 12 years. All scores decreased as the participant aged; except for WCST, preservative errors score, which was higher at 12 years than 9 years. The largest range of scores was among CPT II, errors of omission scores at 9 and 12 years.

Table 2.

Child attention, executive function, and cognitive scores among CHAMACOS participants at ages 7, 9, 10.5, and 12 years

	Mean (Standard Deviation)	Median	Range
Attention Domains
CPT II (T-score)
Errors of omission (9 y)	59.29 (17.94)	54.22	40.28-168.15
Errors of omission (12 y)	48.95 (10.0)	46.10	40.10-132.58
Hit Rate SE by block (9 y)	51.21 (12.12)	51.38	0.57-92.13
Hit Rate SE by block (12 y)	50.12 (9.76)	49.07	23.21-82.50
CADS (T-score)
ADHD Index (9 y)	51.07 (9.15)	49	40-90
ADHD Index (12 y)	49.18 (7.9)	47	41-88
Inattentive subscale (9 y)	51.64 (9.71)	49	40-90
Inattentive subscale (12 y)	49.32 (8.0)	47	41-84
Executive Function Domains
WCST (T-score)
Perseverative errors (9 y)	52.35 (14.42)	49	23-81
Perseverative errors (12 y)	56.30 (13.76)	54	19-81
BRIEF (T-score)
Global Executive Composite (7 y)
Global Executive Composite (9 y)	49.07 (10.15)	48	30-88
Global Executive Composite (12 y)	48.44 (9.76)	47	33-83
Cognition Domains
WISC-IV
FSIQ (7 y)	105.50 (13.52)	72	69-141
FSIQ (10.5 y)	90.47 (10.51)	90	66-120
VCIQ (7 y)	105.07 (14.51)	104	53-142
VCIQ (10.5 y)	85.05 (11.37)	87	57-116

Abbreviations: BRIEF, Behavior Rating Inventory of Executive Function CADS, Conners’ Rating Scales; CPT, Conners’ Continuous Performance Test; FSIQ, Full Scale Intelligence Quotient; VCIQ, Verbal Comprehension Intelligence Quotient; WCST, Wisconsin Card Sort Test; WISC, Weschler Intelligence Scale for Children

Table 3 displays the adjusted associations of prenatal PBDE metabolite concentrations with child attention, executive function, and cognitive scores, for all children and stratified by the median duration of exclusive and complementary breastfeeding (1.4 and 7 months, respectively). Prenatal PBDE concentrations were associated with somewhat poorer attention among those who received exclusive and complementary breastfeeding less than the median duration, with the largest difference for Hit Rate by SE by block (β per doubling of PBDE concentration for exclusive breastfeeding < 1.4 months = 5.6; 95% CI: 1.9, 9.3; β ≥ 1.4 months = 0.0; 95% CI: −4.1, 4.0; p-int=0.29). For executive function, we observed significantly stronger adverse associations of maternal prenatal PBDE concentrations and the Global Executive Composite among those complementary breastfed less than the 7 months (β per doubling of PBDE concentration =4.3; 95% CI: 0.4, 8.2) compared to those complementary breastfed equal to or greater than 7 months (β=0.6; 95% CI: −2.8, 3.9; p-int=0.06). We found significantly stronger adverse associations of maternal prenatal PBDE concentrations and the Global Executive Composite (β per doubling of PBDE concentration =3.6; 95% CI: −0.2, 7.4) and WCST preservative errors (β per doubling of PBDE concentration = −7.2; 95% CI: (−12.2, −2.2) among those exclusively breastfeeding 1.4 months or longer. Findings for all other measures did not differ by length of complementary or exclusive breastfeeding. Including all participants – regardless of their breastfeeding initiation status resulted in similar differences between breastfeeding longer or shorter than the median (Supplementary Table 2).

Table 3.

Adjusted^a associations [β (95% CI)] of prenatal serum PBDE concentrations (log₁₀-transformed) with child attention, executive function, and cognitive scores excluding those who did not initiate breastfeeding using GEE linear regression models for all children and stratified by exclusive and complementary breastfeeding duration (median)^b, CHAMACOS.

	Orientation for Negative Association	All Children		Exclusive Breastfeeding Duration			Complementary Breastfeeding
				< 1.4 Months	≥ 1.4 Months	p-int	< 7 Months	≥ 7 Months	p-int
Attention		n	β (95% CI)	β (95% CI)	β (95% CI)		β (95% CI)	β (95% CI)
CPT II (T-score)c (9, 12 y)
Errors of omission	(+)	468	2.6 (−1.0, 6.3)	3.4 (−2.5, 9.3)	1.4 (−3.1, 5.9)	0.56	3.1 (−2.9, 9.0)	1.6 (−2.8, 6.1)	0.86
Hit Rate SE by block	(+)	468	2.8 (0.1, 5.6)**	5.6 (1.9, 9.3)**	0.0 (−4.1, 4.0)	0.29	3.7 (0.1, 7.4)**	1.5 (−2.6, 5.5)	0.85
CADS (T-score) (9, 12 y)
ADHD Indexd	(+)	465	2.1 (−0.2, 4.5)*	2.0 (−0.8, 4.9)	1.9 (−1.9, 5.7)	0.86	3.6 (0.0, 7.1)**	0.7 (−2.5, 3.9)	0.30
Inattentive subscalee	(+)	464	1.9 (−0.2, 4.1)*	1.7 (−1.2, 4.5)	2.2 (−1.2, 5.6)	0.88	2.9 (−0.2, 5.9)*	0.6 (−2.6, 3.7)	0.51
Executive Function
WCST (T-score)f (9, 12 y)
Perseverative errors	(−)	466	−5.7 (−9.3, −2.1)**	−4.5 (−9.7, 0.7)*	−7.2 (−12.2, −2.2)**	0.98	−6.6 (−11.4, −1.8)**	−5.1 (−10.2, 0.1)*	0.63
BRIEF (T-score)h (7, 9, 12 y)
Global Executive Composite	(+)	465	2.7 (0.2, 5.3)**	2.0 (−1.6, 5.6)	3.6 (−0.2, 7.4)*	0.68	4.3 (0.4, 8.2)**	0.6 (−2.8, 3.9)	0.06
Cognition
WISC-IV (7, 10.5 y)
FSIQi	(−)	444	−4.8 (−8.4, −1.3)**	−2.7 (−7.9, 2.5)	−3.5 (−8.6, 1.6)	0.72	−3.8 (−9.0, 1.3)	−3.0 (−7.6, 1.6)	0.87
VCIQj	(−)	470	−4.5 (−8.1, −1.0)**	−2.1 (−6.8, 2.7)	−3.9 (−8.9, 1.2)	0.45	−4.7 (−9.8, 0.5)*	−3.2 (−7.9, 1.5)	0.99

Abbreviations: BRIEF, Behavior Rating Inventory of Executive Function CADS, Conners’ Rating Scales; CPT, Conners’ Continuous Performance Test; FSIQ, Full Scale Intelligence Quotient; VCIQ, Verbal Comprehension Intelligence Quotient; WCST, Wisconsin Card Sort Test; WISC, Weschler Intelligence Scale for Children

^aModels adjusted for child sex, parity, duration of breastfeeding, preschool attendance, and age at assessment; maternal age, education, IQ (PPVT score), and depression at child age 9; maternal prenatal smoking status; family structure (father present/absent) at time of assessment; household income; HOME score at child age 10.5; and psychometrician or study interviewer who administered the task or rating scale. CPT II models also adjusted for time of day assessment occurred and child video game usage at age 9. Stratified models not adjusted for breastfeeding duration.

^bMedian duration for exclusive breastfeeding duration = 1.4 months; Median duration for any breastfeeding = 7.0 months

^cn= 229 for exclusive breastfeeding < median; n= 221 for exclusive breastfeeding ≥ median; n= 229 for any breastfeeding < median; n= 239 for any breastfeeding ≥ median

^dn= 226 for exclusive breastfeeding < median; n= 220 for exclusive breastfeeding ≥ median; n= 227 for any breastfeeding < median; n= 238 for any breastfeeding ≥ median

^en= 226 for exclusive breastfeeding < median; n= 219 for exclusive breastfeeding ≥ median; n= 227 for any breastfeeding < median; n= 237 for any breastfeeding ≥ median

^fn= 227 for exclusive breastfeeding < median; n= 221 for exclusive breastfeeding ≥ median; n= 228 for any breastfeeding < median; n= 238 for any breastfeeding ≥ median

^hn= 225 for exclusive breastfeeding < median; n= 221 for exclusive breastfeeding ≥ median; n= 226 for any breastfeeding < median; n= 239 for any breastfeeding ≥ median

ⁱn= 216 for exclusive breastfeeding < median; n= 210 for exclusive breastfeeding ≥ median; n= 215 for any breastfeeding < median; n= 229 for any breastfeeding ≥ median

^jn= 229 for exclusive breastfeeding < median; n= 223 for exclusive breastfeeding ≥ median; n= 229 for any breastfeeding < median; n= 241 for any breastfeeding ≥ median

^*p<0.10

^**p<0.05

4. Discussion

We examined whether previously observed adverse associations of prenatal PBDE exposure and child attention, executive function, and cognition were modified by exclusive or complementary breastfeeding duration. While breastfeeding likely exposes infants to higher levels of lipophilic PBDE chemicals, data gaps exist on whether breastfeeding exacerbates or mitigates these exposure-outcome associations, given the overwhelming evidence of the positive neurodevelopmental impacts of breastfeeding.^44,45 Our results suggest prolonged breastfeeding does not exacerbate but may mitigate some previously observed negative associations of prenatal PBDE exposure and child neurodevelopment. We observed that higher maternal prenatal PBDE concentrations were associated with poorer executive function among children who were complementary breastfed for a shorter duration compared to children breastfed for a longer duration.

To our knowledge, no study has investigated potential modification of the well-known negative associations between PBDE and neurodevelopment by breastfeeding duration; therefore, we are unable to directly compare our results to other studies. Studies investigating associations of PBDE concentrations in breastmilk and neurodevelopment have reported mixed results. A study of 70 infant-mother pairs in Taiwan found that a doubling in breastmilk PBDE levels at 1 month postpartum was associated with adverse scores on measures of cognition, but with higher scores on the language (receptive and expressive communication) scale at 8-12 months postpartum, measured by the Bayley Scales of Infants and Toddlers Development, third edition (Bayley-III).⁴⁶ Research from the Pregnancy, Infection, and Nutrition (PIN) Study (N=300) reported an association between higher breastmilk PBDE levels and increased activity, impulsivity, anxiety, and internalizing withdrawal behaviors in children at 30 months postpartum,^47,48 although improved adaptive and cognitive skills were also observed. ²² PBDE breastmilk concentrations at 3-months postpartum were also associated with improved adaptive and cognitive skills among children in the PINS Babies Study which the authors attributed to the numerous benefits that breastfeeding affords.⁴⁸

Our results suggest that the length of exclusive or complementary breastfeeding is an important indicator for child neurodevelopment and prenatal PBDE exposure. The mean duration of exclusive or complementary breastfeeding resulted in stronger and more prominent protective associations compared to the median duration cutoff — which represents a shorter duration. These findings elucidate that the duration of breastfeeding, regardless of exclusivity or complementary breastfeeding, is important for childhood PBDE exposure. PBDE concentrations in breastmilk are highest immediately after birth exacerbating the potential risk for adverse neurodevelopment outcomes due to increased exposure. In an analysis of primiparous mothers from California, investigators found that PBDE concentrations in breast milk decreased 2-3% per-month during the first six months of breastfeeding, with a continued decrease of 1% per month until 34 months postpartum.⁴⁹ Therefore, an infant’s highest daily exposure to PBDE is during the beginning of lactation (birth to one month postpartum) with a slow decrease over time. After cessation of breastfeeding, ingestion and dermal contact of dust is reported to be the main route of PBDE exposure for toddlers.^21,50

The numerous child neurodevelopment benefits that breastfeeding affords is consistently seen among infants exclusively or complementary breastfed a longer duration.^23,44,51 As a result, infants breastfed a shorter duration (less than the mean or median duration) may not reap the full extent of benefits. Human milk composition is uniquely suited to promote optimal brain function and may counteract some of PBDEs’ negative associations with neurodevelopment.⁵² Certain components of human milk, such as long-chain polyunsaturated fatty acids (LC-PUFAs), phospholipids, and cholesterol, promote myelin sheath development and synthesis.⁵³ Gangliosides are critical for cell membrane structure and promotion of healthy infant brain development,⁵⁴ and are consistently found in breastmilk throughout lactation. Therefore, as children grow and require a higher caloric intake, their prolonged and increased intake of breastmilk naturally increases their consumption of gangliosides, LC-PUFAs, phospholipids, cholesterol, and other components that promote neurodevelopment.⁵⁵ This may also explain why longer complementary breastfeeding may have mitigated the adverse impact of early PBDE exposure. Unfortunately, our ability to look at the beneficial impact of exclusive breastfeeding was limited since only half of the women exclusively breastfed longer than a month.

Investigating the critical duration of exclusive breastfeeding needed to counteract PBDE offloading is required in order to identify proper medical and public health recommendations.

Our study has some limitations. Most notably, PBDE levels measured prenatally or immediately after delivery may not accurately reflect an infant’s exposure level from breastmilk; however, previous research indicates that PBDE concentration levels analyzed in maternal serum, breastmilk, and maternal cord blood at 6-8 weeks postpartum were highly correlated, and that prenatal serum samples can estimate exposure via breastmilk when breastmilk samples are not measured directly.⁵⁶ Further, maternal serum concentrations of PBDE are stated to provide a valid and accurate estimation of PBDE breastmilk concentrations and PBDE exposure in infants.⁵⁷ Previous research in a small subset of CHAMACOS women also reported a high correlation (Pearson r= 0.98–0.99, p-values <0.001) between PBDE serum levels during pregnancy and shortly after birth;¹⁹ thus, we do not expect marked differences in PBDE concentrations during pregnancy and immediately after delivery, when breastfeeding is initiated. Our sample size prohibited a more focused analysis by the WHO’s international breastfeeding recommendations to exclusively breastfeed for 6 months, followed by complementary breastfeeding until 2 years postpartum. We were unable to differentiate between intensity of complementary breastfeeding (i.e. the child received mostly breastmilk; an equal distribution between breastmilk and other liquids or foods; or majority other liquids or foods) that directly influences the amount of PBDE concentrations received from breastmilk.

Despite some limitations, we are the first study to assess whether breastfeeding duration may mitigate or exacerbate the previously observed adverse associations of prenatal and early-life PBDE exposure with child neurodevelopment. Additional studies should examine these associations with PBDEs and other lipophilic compounds in order to develop recommendations for exclusive and complementary breastfeeding among mothers with high exposure to environmental chemicals that may be passed to their offspring.

4.1. Conclusions

Our findings indicate that prolonged breastfeeding does not exacerbate adverse effects of prenatal PBDE exposure on child neurodevelopment but may diminish some of these associations. Thus, our findings suggest that receipt of any amount of breastmilk over the first year of life could reduce the negative association of prenatal PBDE exposure on a child’s neurodevelopment. Additional research is needed to examine if breastfeeding may indeed promote child neurodevelopment among children with high levels of exposure to other lipophilic environmental chemicals and to identify the critical duration of breastfeeding needed to mitigate the potential negative associations. More importantly, we should reduce those exposure to fetuses and children of chemicals that can cause neurodevelopmental harm.

Highlights.

• Longer breastfeeding mitigates associations between PBDE and child neurodevelopment

• PBDE is associated with worse executive scores among longer exclusive breastfeeding

• Study extends prior research by examining the interaction effect of breastfeeding