Persistent organic pollutants exposure in newborn dried blood spots and infant weight status: A case-control study of low-income Hispanic mother-infant pairs

Abstract

Persistent organic pollutants (POPs) are believed to alter metabolic homeostasis during fetal development, leading to childhood obesity. However, limited studies have explored how fetal chemical exposures relate to birth and infant weight outcomes in low-income Hispanic families at the highest risk of obesity. Therefore, we sought to determine associations between neonatal POPs exposure measured in newborn dried blood spots (DBS) and prenatal diet quality, birth weight, and overweight status at 18 months old. We conducted a case-control study nested within the Starting Early Program randomized controlled trial comparing POPs concentrations in infants with healthy weight (n=46) and overweight status (n=52) at age 18 months. Three categories of POPs, organochlorine pesticides (OCPs), polybrominated diphenyl ethers (PBDEs) and perfluoroalkyl substances (PFASs) were measured in archived newborn DBS. We assessed correlations between prenatal diet quality and neonatal POPs concentrations. Multivariable regression analyses examined associations between POPs (dichotomized at the mean) and birth weight z-score and weight status at 18 months, controlling for confounders. Seven of eight chemicals had detectable levels in greater than 94% of the sample. Higher protein, sodium and refined grain intake during pregnancy were correlated with lower POPs in newborn DBS. We found that high concentrations of perfluorooctanesulfonate (unstandardized coefficient [B]: −0.62, 95% confidence interval [CI]: −0.96 to −0.29) and perfluorohexanesulfate (B: −0.65, 95% CI: −0.99 to −0.31) were related to lower birth weight z-scores compared to those with low concentrations. We did not find associations between PBDEs, OCPs, and the other PFASs with birth weight z-scores, or between any POPs and weight status at 18 months. In conclusion, two PFASs were associated with lower birth weight, an important indicator of child health and growth, although direct associations with infant overweight status were not found. Whether neonatal POPs exposures contribute to economic and ethnic disparities in early obesity remains unclear.

Summary of Main Findings

Persistent organic pollutants (POPs) were associated with prenatal diet and lower birth weight z-scores, but were not associated with overweight status at age 18 months.

Graphical Abstract

INTRODUCTION

Persistent organic pollutants (POPs) are common chemical contaminants known to degrade slowly (La Merrill and Birnbaum, 2011). In general, POPs are ubiquitous and are known to remain persistent in the body with their half-life on the order of 7 to 8 years (Olsen et al., 2007). Even though the use of many of these compounds has been discontinued, they have been replaced by similar compounds, highlighting their potential for continued adverse health impacts (Khalil et al., 2018; Ye et al., 2018). Infant exposure to POPs can take place prior to conception (via toxicity to sperm and oocytes) and in utero (via mother’s dietary intake and environmental exposure during pregnancy) (WHO, 2010). Since maternal fat stores are mobilized throughout pregnancy, toxicants and pollutants can be transferred to the embryo and the fetus through the placenta and consequently interfering with the fetal programming of endocrine-signaling pathways (Yeung et al., 2019). Because of their overall estrogenic, anti-estrogenic, anti-thyroid, anti-progestin, and anti-androgenic effects, POPs have the potential to disrupt the endocrine system in the human body, which may interfere with adipocyte differentiation, energy storage, and growth during the early stages of life (Darbre, 2017; Tang-Péronard et al., 2011; Vandenberg et al., 2012). Therefore, the endocrine-disruptive adverse effects of POPs exposure may be potentially obesogenic (WHO, 2010).

Several categories of POPs have been postulated to alter metabolic homeostasis during fetal development, potentially increasing the risk of obesity, diabetes, and metabolic syndrome (Buha Djordjevic et al., 2020; Darbre, 2017; Nelson et al., 2010), leading researchers to focus on better understanding the adverse effects of these chemical exposures during this sensitive period of the life course. Three specific categories of POPs, organochlorine pesticides (OCPs), polybrominated diphenyl ethers (PBDEs) and perfluoroalkyl substances (PFASs), have been suggested to adversely affect child health through early pathways related to growth (Goudarzi and Yamazaki, 2020; Ouidir et al., 2020). OCPs are lipophilic chlorinated hydrocarbons previously used for agriculture and mosquito control, and have been found in fish, dairy, and other fatty foods (Loganathan and Kannan, 1994). PBDEs are lipophilic chemicals, commonly used as flame-retardants that have been phased out in the United States since 2004. However, because of their long half-life and bioaccumulation, PBDEs are still present in many commercial and household products, including textiles, plastics, and electronics. Although their route of exposure through diet is less common, they are still detected in foods with high fat content, such as fatty fish. However, diet may be a common route of exposure for PFASs, which are amphiphilic POPs, that have been linked to drinking contaminated water, eating fish caught from contaminated water, eating foods that are packaged in materials containing PFASs, and using non-stick cookware, leading to contamination of the food being cooked (Kannan, 2011). Given that rapid growth and overweight status during infancy are consequential because they strongly predict later obesity (Ekelund et al., 2007; Ogden et al., 2014; Ong et al., 2006; Taveras et al., 2009; Wang et al., 2016) and associated co-morbidities (Belfort et al., 2007; Ong et al., 2004; Péneau et al., 2016), understanding whether early POPs exposure may serve as an antecedent to the development of obesity during infancy is critical.

Pregnancy represents a vulnerable period in which chemical exposures can alter fetal programming and POPs can pass through the placental barrier (Covaci et al., 2002; Jaraczewska et al., 2006). In some studies, prenatal exposure to PFASs has been shown to accelerate growth and obesity risk later in life (Halldorsson et al., 2012; Newbold et al., 2007). Increased child obesity risk has also been associated with prenatal exposure to organochlorine pesticides such as dichlorodiphenyldichloroethylene (DDE) (Valvi et al., 2012). However, some studies have either found mixed results or reported an inverse association between prenatal exposure to PFASs and child weight. Fetal exposure to perfluorooctanesulfonate (PFOS) and perfluorooctanoate (PFOA) has been associated with lower birth weight (Apelberg et al., 2007; Bach et al., 2016; Chen et al., 2012; Fei et al., 2007), and lower weight and body mass index (BMI) in early infancy (Andersen et al., 2010). Within the Danish National Birth Cohort, prenatal exposure to PFOA or PFOS was linked to lower birth weight, weight, and BMI through infancy in males, but not in females (Andersen et al., 2010). In this same cohort, although not significant, maternal PFOA and PFOS levels were inversely linked to child waist circumference, BMI, and risk of overweight at 7 years of age (Andersen et al., 2013). Within the Upstate KIDS Study of young children from upstate New York, newborn dried blood spot (DBS) PFOA and PFOS concentrations were also associated with lower BMI, and not early obesity (Yeung et al., 2019). Given these conflicting findings, it remains unclear how prenatal exposures to PFOS and PFOA, as well as other POPs, influence birth weight and growth in early childhood.

The majority of studies exploring associations between POPs exposure and child weight, have used maternal biospecimens, which reflect maternal POPs exposure rather than child exposure. While cord blood samples or blood drawn directly from infants may be the ideal method to assess child POPs exposure, it is often less feasible to collect in low-income underserved communities. This lack of feasibility may be due to disparities that exist in recruitment, retention, and trust in research studies contributing to biorepositories among families with limited literacy, low socioeconomic status, and racial and ethnic diversity (George et al., 2014). The use of archived newborn DBS to assess exposures in newborns may serve as an approach to enhance the feasibility of obtaining samples in newborn populations at greatest risk of obesity. This is particularly important given that the majority of studies exploring prenatal POPs exposure have focused on middle- to high-income families and rural populations (Bell et al., 2018). Studying prenatal chemical exposures in low-income, urban, Hispanic families, a group with the highest rates of obesity in early childhood in the United States (Ogden et al., 2014), will advance the literature by increasing our understanding of how these exposures may contribute to known socioeconomic and ethnic disparities in the rates of early child obesity beginning in infancy.

In order to fill these knowledge gaps, we conducted a case-control study comparing POPs concentrations in residual and archived newborn DBS in Hispanic infants with healthy weight (n=46) and overweight status (n=52) at the age of 18 months. The study aimed to examine: 1) associations between prenatal diet quality and concentrations of OCPs, PBDEs, and PFASs in newborns; and 2) associations between newborn POPs concentrations and both birth weight and weight status at age 18 months.

MATERIALS AND METHODS

Study Design and Participants

We conducted a case-control study nested within the Starting Early Program (StEP) trial, a randomized controlled trial (RCT) designed to help promote healthy infant feeding and growth, and prevent childhood obesity in low-income Hispanic families. The StEP RCT took place in the prenatal and pediatric clinics of a New York City municipal hospital system. Pregnant women (n=533) were enrolled during their third trimester between August 2012 and December 2014 (Gross et al., 2016). Pregnant women were eligible to participate in the StEP RCT, if they were at least 18 years old with a singleton, uncomplicated pregnancy, from Hispanic/Latina origin, English or Spanish speaking, and willing to receive their future pediatric care at the study sites. Participating women were not eligible if they were homeless, had significant medical, psychiatric or substance use disorders, or their fetus had severe anomalies on ultrasound examination. While the participants in the StEP RCT were mother-child pairs, because enrollment occurred during pregnancy, there were no additional inclusion/exclusion criteria for the infants after they were born. After enrollment, the pregnant women were randomized to receive either the StEP intervention or standard prenatal and pediatric health care. The StEP intervention consisted of individual nutrition and lactation counseling sessions with a registered dietitian during pregnancy and the post-partum period, followed by dietitian-led parenting groups consisting of 4 to 8 mother-child pairs coordinated with all well child health care visits during the first three years of life. The long-term follow-up of the StEP trial is ongoing.

For this case-control study nested within the StEP trial, we were able to reach 338 (63%) mother-child pairs at child age 3-years-old (the completion of the original trial), to obtain consent to use their child’s stored and residual newborn DBS for chemical analysis. Of the 338 pairs reached, 5 pairs refused, leaving 333 mother-child pairs who consented for these chemical analyses. A consecutive sample of 52 mother-child pairs (15.6% of those consented) with infants with overweight status at 18 months old (defined as weight for length z-scores ≥85^th percentile) and 46 pairs (13.8% of those consented) with infants with healthy weight status at 18 months old (defined as weight for length z-scores <85% percentiles) (Kuczmarski and National Center for Health Statistics, 2002), had their newborn DBS analyzed for POPs concentrations. Therefore, this subsample of 98 mother-child pairs, with complete POPs exposure data, were included in these analyses (Figure 1).

Figure 1:

Study overview

The Institutional Review Board of New York University School of Medicine, and New York City Health + Hospitals approved the trial, which is registered on clinicaltrials.gov (NCT01541761).

Measures

Chemical Concentrations

Participants in the StEP RCT were consented to obtain their newborn’s residual DBS from the NYS Newborn Screening Program to measure POPs concentrations. Newborn screening tests measure metabolites and enzyme activities from whole blood on specialized filter paper, by obtaining the blood from a neonate’s heel within 24 to 48 hours after birth (Olshan, 2007) and sending dried papers with blood spots for analysis. The NYS newborn screening program receives DBS from about 250,000 newborns each year and began archiving unused residual samples in 1997 for notifying patients and providers of identified abnormalities and for public health research (Ma et al., 2013; Spliethoff et al., 2008). We measured polybrominated diphenyl ether congeners #28 (PBDE28), #47 (PBDE47), #99 (PBDE99), dichlorodiphenyldichloroethylene (DDE), perfluorooctanoate (PFOA), perfluorooctanesulfonate (PFOS), perfluorononanoate (PFNA), and perfluorohexanesulfate (PFHxS) from DBS (16-mm diameter discs) using a highly sensitive liquid-liquid extraction and high-performance liquid chromatography and tandem mass spectrometry, using the methods described elsewhere (W. Ma et al., 2013; Ma et al., 2014; W. L. Ma et al., 2013).

Prenatal Diet Quality

Participants’ usual frequency and quantity of dietary intake in the past twelve months overlapping their pregnancy was assessed at baseline (28 to 32 weeks gestational age) using the 2005 bilingual English/Spanish version of Block Food Frequency Questionnaire (FFQ). The Block FFQ has been validated in pregnant women (Johnson et al., 2007) and in low-income families (Block et al., 1995). The FFQ included 118 foods based on the 1999–2002 National Health and Nutrition Examination Survey (NHANES), and foods commonly eaten in Latin American communities. The US Department of Agriculture (USDA) Food and Nutrient Database for Dietary Studies version 1.0 evaluated the FFQs.

StEP assessed diet quality of the participants using the Healthy Eating Index (HEI)-2015 (Krebs-Smith et al., 2018; Reedy et al., 2018). The HEI-2015 has thirteen food groups that reflect the dietary recommendations based on the 2015–2020 Dietary Guidelines for Americans. HEI includes nine adequacy (e.g., total vegetables, greens and beans, total fruits, whole fruits [total fruit excluding juice], whole grains, dairy, total protein foods, seafood and plant proteins, and fatty acids) and four moderation (e.g., sodium, added sugars, refined grains and saturated fat) components (Berube et al., 2019). Higher adequacy component scores represented higher intake, whereas higher moderation component scores represented lower intake. The overall total HEI-2015 score between zero and 100 represented a sum of these individual component scores. A higher HEI total score is an indicator of higher diet quality that aligns with 2015–2020 Dietary Guidelines for Americans. HEI-2015 Statistical Analysis Software Code for FFQ (National Cancer Institute, 2019) and MyPyramid Equivalents Database calculated the individual component and total scores of food groups from the HEI-2015 and Block FFQ.

Infant Anthropometric Data

All anthropometric data for infants including birth weight, weight, and length were retrieved by medical record review of the primary care visits and birth hospitalization. Infant birth weight, sex, and gestational age were used to generate birth weight z-scores based on Fenton growth curves (Fenton and Sauve, 2007). Weight for length z-scores (WFLz) were calculated using the World Health Organization Anthro macro, and biologically implausible z-scores were manually reviewed and kept if deemed plausible (Rifas-Shiman et al., 2005). Weight for length z-scores at 18 months were categorized as healthy weight status (defined as weight for length z-scores <85^th percentiles) and overweight status (defined as weight for length z-scores ≥85^th percentile).

Family Characteristics

Baseline family characteristics, including maternal age, parity, and education, were collected by an interviewer-administered survey. Prenatal depressive symptoms were defined as a score of 5 or greater on the Patient Health Questionnaire-9 (Kroenke et al., 2001) and household food insecurity was defined as a score of 3 or more on the US Adult Food Security Survey Module, developed by USDA (Bickel et al., 2000). Gestational age and maternal pre-pregnancy BMI (classified as healthy (BMI<25) vs. overweight or obese (BMI≥25) were obtained from the electronic medical record (Kuczmarski and National Center for Health Statistics, 2002).

Statistical Analysis

Descriptive statistics assessed the distribution and frequency of the main study variables, including chemical concentrations and prenatal diet quality, for those in the StEP trial, in the total analytic sample, and in those with healthy weight and overweight status. Chemical concentrations were log-transformed due to skewed distributions. Categorical variables for the log-transformed chemical concentrations were defined as low (below the mean of the log-transformed concentration) and high (concentrations at or above this mean). We assessed the relationships between prenatal diet quality (total HEI, adequacy and moderation component scores) and chemical concentrations using Pearson correlations. Linear regression analyses were used to assess associations between chemical concentrations (low vs. high) and birth weight z-scores, using both unadjusted and adjusted models controlling for covariates. Models using continuous chemical concentrations were also conducted. Binomial logistic regression analyses were used to assess associations between chemical concentrations (low vs. high) and weight status at 18 months (reference: healthy weight status). Models were adjusted for covariates associated with both chemical exposures and child weight status, such as maternal age, maternal education, prenatal depressive symptoms, pre-pregnancy BMI, gestational age, parity and intervention status. For analyses examining the associations of chemicals with birth weight z-scores and weight status at 18 months, we applied the Bonferroni adjustment to avoid a type I error due to multiple comparisons. The aim examining the association between prenatal diet and chemicals was exploratory and hypothesis generating. Therefore, Bonferroni adjustments were not applied when the correlations between prenatal diet with chemicals were tested. In addition, to determine sex-specific differences in chemical exposure associations with birth weight z-scores and weight status at 18 months, all analyses were done separately for males and females. Data analyses were performed using SPSS statistical software version 25.0 (SPSS Inc, Chicago, Il).

RESULTS

Study Sample Characteristics

The study sample is described in Table 1. All women self-identified as Hispanic/Latina with a mean age of 28.7 years (SD 5.6) at enrollment. Women were primarily non-US born (84%), with the majority of women originating from Mexico, Ecuador and Dominican Republic. Thirty-six percent of participants had less than a high school level of education, 73.2% had pre-pregnancy overweight or obese weight status, and many had high rates of poverty-related risks, including household food insecurity (29.6%) and depressive symptoms (28.6%).

Table 1:

Sample characteristics of the original StEP cohort, the total analytic sample with newborn dried blood spots, and subgroups based on infant weight status at 18 months old.

	StEP RCT	Newborn Dried Blood Spot (DBS) Samples Analyzed
Characteristics	Total StEP Sample (n=533)	Total DBS Sample (n=98)	Healthy Weight Sample (n=46)	Overweight Sample (n=52)
Expectant Mother
Maternal age, years [mean (SD)]	27.7 (5.9)	28.7 (5.6)	27.6 (6.0)	29.8 (5.2)
Primiparous, %	37.3	34.7	39.1	30.8
US born, %a	20.1	16.3	17.4	15.4
Married or living as married, %	71.1	73.5	69.6	76.9
Education (< high school), %	33.2	35.7	41.3	30.8
Pre-pregnancy weight status, %
Underweight	1.6	2.1	0.0	4.0
Healthy weight	34.4	24.7	32.6	17.6
Overweight	33.5	35.1	37.0	33.3
Obese	30.5	38.1	30.4	45.1
Household food insecurity, %	31.0	29.6	30.4	28.8
Prenatal depressive symptomsa, %	34.0	28.6	19.6	36.5
Birth
Male sex, %	49.0	51.0	41.3	59.6
Cesarean delivery, %	19.5	29.6	23.9	34.6
Gestational age [mean (SD)]	39.1 (1.33)	39.3 (1.32)	39.0 (1.46)	39.4 (1.17)
Prenatal Diet Qualityb
Total score [mean (SD)]	69.0 (9.4)	69.7 (9.0)	70.2 (8.4)	69.3 (9.6)
Total vegetablesc, %	34.7	42.7	45.7	40.0
Greens and beansc, %	65.5	71.1	73.9	68.6
Total fruitc, %	54.5	58.8	63.0	54.9
Whole fruitc, %	51.6	57.9	60.9	55.1
Whole grainsc, %	2.9	4.1	2.2	5.9
Dairyc, %	25.2	25.0	23.9	26.0
Total proteinc, %	62.4	63.9	58.7	68.6
Seafood and plant proteinc, %	61.5	66.3	60.0	72.0
Fatty acidsc, %	15.6	14.6	15.6	13.7
Sodiumc, %	2.5	2.1	2.2	2.0
Refined grainsc, %	20.2	22.9	26.1	20.0
Saturated fatsc, %	7.1	3.1	2.2	3.9
Added sugarsc, %	33.5	44.7	44.4	44.9
Newborn Blood Spot Chemical Concentrations (ng/ml) d,e
PBDE28	-	0.024 (0.018)	0.024 (0.017)	0.025 (0.021)
PBDE47	-	0.280 (0.191)	0.287 (0.163)	0.265 (0.239)
PBDE99	-	0.154 (0.143)	0.181 (0.203)	0.152 (0.134)
DDE	-	0.524 (0.728)	0.671 (1.482)	0.300 (0.561)
PFOA	-	0.376 (0.249)	0.346 (0.218)	0.391 (0.262)
PFOS	-	0.440 (0.364)	0.522 (0.366)	0.365 (0.339)
PFNA	-	0.151 (0.123)	0.162 (0.134)	0.138 (0.115)
PFHxS	-	0.108 (0.065)	0.109 (0.074)	0.104 (0.064)

^aPrenatal depressive symptoms were defined as a score of 5 or greater on the Patient Health Questionnaire-9.

^bOf the 541 participants who fully completed the Food Frequency Questionnaire, 519 women (95.9%) had plausible energy intakes between 500 and 5,000 kcal and were included in analyses.

^cPercentage of participants meeting HEI recommendations for each of the 13 adequacy and moderation components of the Healthy Eating Index [HEI].

^dPBDE28, polybrominated diphenyl ether 28; PBDE47, polybrominated diphenyl ether 47; PBDE99, polybrominated diphenyl ether 99; DDE, dichlorodiphenyldichloroethylene; PFOA, perfluorooctanoate; PFOS, perfluorooctanesulfonate; PFNA, perfluorononanoate; PFHxS, perfluorohexanesulfate.

^eMedian (IQR) of concentrations provided for participated with detectable levels. Sample size varied secondary to excluding undetectable levels (PBDE28 n=58; PBDE47 n=98; PBDE99 n=95; DDE n=94; PFOA n=99; PFOS n=100; PFNA n=100; PFHxS n=97).

The mean HEI-2015 total score was 69.7 (SD 9.0), with scores ranging from 45.7 to 94.8. While over 50% of women met recommendations for total protein foods (≥2.5 oz equivalent/1,000 kcal), seafood and plant proteins (≥0.8 oz equivalent/1,000 kcal), greens and beans (≥0.2 cup equivalent/1,000 kcal), total fruits (≥0.8 cup equivalent/1,000 kcal), and whole fruits (≥0.4 cup equivalent/1,000 kcal), the majority of participants did not meet recommendations for the other adequacy components. In addition, most women overconsumed saturated fats (≥16% of energy), sodium (≥2.0 grams per 1,000 kcal), refined grains (≥4.3 oz equivalent per 1,000 kcal), and added sugars (≥26% of energy), and therefore did not meet recommendations for the moderation components (National Cancer Institute, 2019).

Newborns were 51% male with mean gestational age (SD) of 39.3 (1.32). Table 1 also describes the median concentrations or interquartile range (IQR) of the POPs detected in the newborn DBS ranging from 0.108 ng/ml to 0.524 ng/ml. Seven of the eight chemicals assessed (PBDE47, PBDE99, DDE, PFOA, PFOS, PFNA, and PFHxS) had detectable levels in greater than 94% of the samples. PBDE28 (medium 0.024 ng/ml) was the only chemical with low rates of detection (58%), and therefore was excluded from the remainder of the analyses.

Correlations between Prenatal Diet Quality and Persistent Organic Pollutants

Overall, healthier components of the prenatal diet quality defined by HEI-2015, were correlated with lower concentrations of PFASs (Table 2). Higher total protein was associated with lower PFOA (r=−0.26, p=0.01) and PFHxS (r=−0.21, p=0.04) concentrations, and with higher PBDE47 concentrations (r=0.22, p=0.04). Lower sodium was associated with lower PFOS (r=−0.23, p=0.03) concentrations. Lower refined grains was associated with lower DDE (r=−0.27, p=0.01) and lower PFHxS (r=−0.40, p<0.001) concentrations. There were no significant associations between any of the chemicals and the following HEI components: total vegetables, greens and beans, total fruit, whole fruit, whole grains, dairy, seafood and plant protein, fatty acids, saturated fats and added sugars.

Table 2:

Correlations between persistent organic pollutants and prenatal diet quality using the Healthy Eating Index-2015 total scores and individual adequacy and moderation components.^a

Prenatal Diet Quality	PBDE47	PBDE99	DDE	PFOA	PFOS	PFNA	PFHxS
Total score	−0.06	−0.10	−0.06	−0.16	−0.04	−0.12	−0.19
Adequacy componentsb
Total vegetables	0.002	−0.02	−0.07	−0.04	0.19	0.09	0.01
Greens and beans	−0.05	−0.04	0.02	−0.14	0.001	−0.06	−0.03
Total fruits	−0.06	−0.14	−0.11	−0.06	−0.13	−0.18	−0.004
Whole fruits	−0.03	−0.12	−0.02	−0.04	−0.04	0.01	0.04
Whole grains	−0.10	−0.03	0.08	−0.19	−0.03	−0.10	−0.10
Dairy	0.09	0.09	−0.05	0.07	0.05	0.06	−0.10
Total protein foods	0.22*	0.10	0.15	−0.26*	−0.10	−0.08	−0.21*
Seafood and plant proteins	0.01	−0.08	0.11	−0.19	−0.03	−0.09	−0.08
Fatty acids	−0.09	−0.15	0.05	−0.14	0.01	−0.05	−0.03
Moderation componentsc
Sodium	−0.16	−0.12	−0.05	0.12	−0.23*	−0.19	0.13
Refined grains	−0.12	−0.09	−0.27*	−0.12	0.02	−0.06	−0.40**
Saturated fats	0.04	0.08	−0.05	−0.01	0.06	−0.01	0.07
Added sugars	−0.11	−0.08	0.14	−0.07	0.06	0.10	−0.02

^aCorrelation coefficients displayed.

^bHigher score indicates higher consumption for the adequacy components.

^cHigher score indicates lower consumption for the moderation components.

^*p<0.05

^**p<0.001

Associations between Persistent Organic Pollutants and Child Weight

Higher concentrations of perfluoroalkyl substances were related to lower birth weight z-scores. In adjusted analyses after Bonferroni adjustment (p < 0.00714), we found that high PFOS (B: −0.62, 95% CI: −0.96 to −0.29) and PFHxS (B: −0.65, 95% CI: −0.99 to −0.31) were significantly related to lower birth weight z-scores compared to those with low concentrations, respectively. We did not find significant associations between PBDE47, PBDE99, DDE, PFOA, and PFNA and birth weight z-scores (Table 3). We also examined associations with continuous chemical concentrations and the associations were not significant (data not shown).

Table 3:

Associations between persistent organic pollutants (POPs) and birth weight z-score

Higher POPs (> the mean level)	Unadjusted B	95% CI	Adjusted Ba	95% CI
PBDE47	−0.37	−0.79 to 0.05	−0.23	−0.59 to 0.12
PBDE99	−0.01	−0.44 to 0.41	−0.05	−0.41 to 0.32
DDE	−0.28	−0.70 to 0.15	−0.20	−0.55 to 0.16
PFOA	−0.01	−0.44 to 0.42	−0.26	−0.63 to 0.11
PFOS	−0.54	−0.95 to −0.13*	−0.62	−0.96 to −0.29**
PFNA	−0.30	−0.72 to 0.12	−0.42	−0.77 to −0.07*
PFHxS	−0.74	−1.14 to −0.34**	−0.65	−0.99 to −0.31**

^aLinear regression models adjusted for maternal age, maternal education, maternal depressive symptoms, pre-pregnancy BMI, gestational age, parity and intervention status.

^*Denoting p < 0.05

^**Denoting statistical significance p < 0.00714 (Bonferroni adjustment; 0.05/7 = 0.00714).

Higher concentrations of DDE were linked to lower odds of being overweight compared to lower concentrations (adjusted OR: 0.32, 95% CI: 0.13 to 0.81) at age 18 months. However, we did not find associations between any of the chemicals and child weight status at child age 18 months after applying Bonferroni adjustments (Table 4). Sex-specific differences were detected for associations between PFOS and lower birth weight z-scores, with associations more prominent in males than in females, and between DDE and lower child overweight status, with associations more prominent in females than in males.(Supplemental Tables 1 and 2).

Table 4:

Associations between persistent organic pollutants (POPs) and child weight status at 18 months

Higher POPs (>the mean level)	Unadjusted OR	95% CI	Adjusted ORa,b	95% CI
PBDE47	0.73	0.33 to 1.61	0.81	0.33 to 1.98
PBDE99	1.27	0.58 to 2.82	1.38	0.56 to 3.42
DDE	0.36	0.16 to 0.81*	0.32	0.13 to 0.81*
PFOA	0.98	0.44 to 2.17	0.91	0.36 to 2.29
PFOS	0.53	0.24 to 1.18	0.43	0.17 to 1.09
PFNA	0.48	0.21 to 1.08	0.42	0.17 to 1.04
PFHxS	0.72	0.33 to 1.60	0.75	0.30 to 1.85

^aLogistic regression models adjusted for maternal age, maternal education, maternal depressive symptoms, pre-pregnancy BMI, gestational age, parity and intervention status.; OR: odds ratio; C.I.: confidence interval for exp (B)

^bReference: Healthy weight status

^*Denoting p < 0.05

^**Denoting statistical significance p < 0.00714 (Bonferroni adjustment; 0.05/7 = 0.00714).

DISCUSSION

Our main study findings are that healthier prenatal diet quality, defined using the Healthy Eating Index-2015, was correlated with lower POPs concentrations, and that higher PFOS and PFHxS concentrations detected in newborn DBS were associated with lower birth weight. However, we did not detect associations between POPs concentrations and overweight status at child age 18 months. Our study adds to the literature by using archived newborn DBS to measure fetal exposure to these chemicals, and by exploring these associations in a low-income high-risk Hispanic families.

In the current study, higher total protein intake was associated with lower PFOA and PFHxS concentrations (amphiphilic POPs). Associations between prenatal diet quality and PBDEs (lipophilic POPs) were less consistent, with higher total protein foods being correlated with higher PBDE47. We also found that higher refined grains intake was correlated with higher DDE and PFHxS concentrations and higher sodium consumption was associated with higher PFOS concentrations. Animal studies have detected lower total POPs concentrations, particularly amphiphilic POPs (PFASs), in the adipose tissue of mice consuming high protein diets, subsequently protecting them developing later obesity (Myrmel et al., 2016). One potential mechanism may be that since POPs are excreted through bile flow, and different protein sources are involved in bile metabolism regulation, dietary protein level can modulate bile flow and consequently eliminate POPs, potentially explaining observed associations between high protein intake and lower POPs concentration (Liaset et al., 2009).

Despite postulations that these prenatal chemical exposures are linked to child obesity, a prior meta-analysis (Johnson et al., 2014) and several more recent studies have identified associations with maternal PFAS levels and lower birth weight (Bach et al., 2016; Meng et al., 2018; Negri et al., 2017; Spratlen et al., 2019). A hospital-based US cross-sectional study (Apelberg et al., 2007) and the Danish National Birth Cohort both found that both PFOS and PFOA measured in cord blood were negatively associated with birth weight measures (Andersen et al., 2013; Fei et al., 2007). However, the Taiwan Birth Panel Study reported an inverse association with only PFOS levels not PFOA in cord blood of newborns and birth weight (Chen et al., 2012). A recent study on the combined serum mixture effect of perfluorinated alkyl acids (PFAAs) in pregnant Danish women found that higher-serum PFAA-induced xenoestrogenic activities were associated with lower birth weight and birth length in offspring, suggesting a mechanism involving PFAAs disruption of the estrogen receptor function (Bjerregaard-Olesen et al., 2019). Findings from the Upstate KIDS study, a primarily white high-income sample using newborn DBS, did not find associations with birth weight (Bell et al., 2018). In contract, our current study using dried NBS in low-income Hispanic families found that in utero exposure to PFOA and PFHxS were related to lower birth weight. Given that birth weight is an important indicator of child health, more research is needed to comprehend how associations between PFAS exposure and lower birth weight effect future growth. A study of 1,954 singletons and 966 twins from the Upstate KIDS Study showed that PFOS and PFOA concentrations in newborn DBS were associated with lower weight-for-length and BMI at age three years old, despite not finding associations with rapid infant weight gain in the first year of life (Yeung et al., 2019). More research is needed to better understand if there is a critical period in early childhood when associations between POPs exposure and obesity risk emerge.

Strengths of our current case-control study include the utilization of newborn DBS to assess neonatal POPs concentrations, which may reflect exposures that occurred prior to or during pregnancy, leading to direct fetal exposure by these chemicals crossing the placenta, and thus being detected in the newborn (Inoue et al., 2004; Olsen et al., 2009; Spratlen et al., 2019). Of note, this is unique in that the majority of previous studies assessing correlations between POPs and child weight have used maternal biospecimens, which reflect maternal POPs levels rather than direct offspring exposure. Another strength is that very few studies have previously assessed associations between neonatal POPs concentrations and prenatal diet quality (Serrano et al., 2014). In addition, this study included newborn DBS from a primarily low-income Hispanic sample living in an urban area. A few studies have assessed POPs plasma concentrations of pregnant mothers and the majority of them did not include the Hispanic population living in an urban area (Yeung et al., 2019). Compared to the total US population, economic costs associated with exposures to environmental chemicals are significantly higher in non-Hispanic Blacks ($56.8 billion) and Mexican Americans ($50.1 billion), possibly due to their higher exposure to persistent pesticides and flame-retardants (Attina et al., 2019; Malits et al., 2018). Therefore, studying chemical exposures in high-risk populations is critical.

This study has several limitations. Given our small sample size, we may have been limited in our ability to detect small associations. Another potential limitation was using medical record reviews to obtain anthropometrics. While studies have demonstrated that clinical and research weight measures correlate well, inaccuracies are common in clinically-measured length and height (Rifas-Shiman, Sheryl L.; Rich-Edwards, Janet W.; Scanlon, Kelley S.; Kleinman, Ken P.; Gillman, 2005). Because of the potential discrepancies in measured child lengths in our study, we conducted sensitivity analyses to test whether these discrepancies were associated with chemical exposures. Our results revealed similar associations with POPs using both weight for length and weight for age z-scores. In addition, this study did not account for other body composition measures, including total body fat, abdominal and subcutaneous fat, skinfold thickness, and soft lean tissue mass. However, weight for length z-score is commonly used as a crude measure of adiposity especially in early childhood period (Roy et al., 2016). In addition, other persistent chemicals, such as polychlorinated biphenyls, and non-persisting chemicals, such as phthalates, that might influence infant weight were not measured, and should be included in future studies. The current study was nested within an RCT of an obesity prevention intervention. While direct effects of the intervention on these chemical exposures were not detected, we controlled for intervention status in our models to decrease potential bias related to sample distribution (Krauss, 2018). Finally, our current analysis is limited to children at 18 months of age, therefore whether effects may be observed in the later years remains unknown.

CONCLUSIONS

We found that POPs were detectable in newborn DBS despite public health efforts to remove them from the environment. Our study found significant associations of neonatal POPs exposure in low-income Hispanic families measured using newborn DBS and lower prenatal diet quality and lower birth weight, an important indicator of future child health. However, associations with child overweight status at 18 months were not found. Larger longitudinal studies in diverse populations are needed to better understand how birth weight associations with POPs relate to later child growth, and if they contribute to known disparities in the rates of early child obesity.

HIGHLIGHTS.

• Persistent organic pollutants (POPs) were detected in newborn dried blood spots.

• Healthier prenatal diet quality was correlated with lower POPs exposure.

• Perfluoroalkyl substances were associated with lower birth weight z-scores.

• Newborn POPs exposure was not associated with overweight status at age 18 months.