Abstract
Experimental studies suggest that prenatal exposure to endocrine disrupting chemicals interferes with developmental processes in the fetal brain. Yet, epidemiological evidence is inconclusive.
In a birth cohort (2008-2010, upstate New York), we quantified concentrations of perfluorooctane sulfonic acid (PFOS), perfluorooctanoic acid (PFOA), and bisphenol A (BPA) in stored newborn dried blood spots using liquid chromatography/tandem mass spectrometry. Mothers reported on children’s behavior using the Strengths and Difficulties Questionnaire at age 7 (650 singletons and 138 twins). Difficulties in total behavior (i.e., emotional, conduct, hyperactivity, and peer problems) and prosocial behavior were classified using validated cutoffs. We used logistic regression with generalized estimating equations to estimate the odds of having difficulties per exposure category.
In total, 111 children (12.1%) had total behavioral difficulties and 60 (6.5%) had difficulties in prosocial behavior. The median (interquartile range) of PFOS, PFOA, and BPA were 1.74 ng/ml (1.33), 1.12 ng/ml (0.96), and 7.93 ng/ml (10.79), respectively. Higher PFOS levels were associated with increased odds of having behavioral difficulties (OR per SD of log PFOS=1.30, 95%CI: 1.03-1.65). We observed associations between PFOS in the highest relative to the lowest quartile and behavioral difficulties (OR for PFOS1.14-1.74=1.65, 95%CI: 0.84-3.34; PFOS1.75-2.47=1.73, 95%CI: 0.87-3.43; and PFOS>2.47=2.47, 95%CI: 1.29-4.72 compared to PFOS<1.41). The associations between higher concentrations of PFOS and behavioral difficulties at age 7 years were driven by problems in conduct and emotional symptoms. Higher PFOA levels were associated with difficulties in prosocial behavior (OR=1.35, 95%CI: 1.03-1.75). There was an inverse association between BPA concentrations and difficulties in prosocial behavior but only in the 2nd and 4th quartiles. We found no interactions between sex and chemical concentrations.
Increasing prenatal exposure to PFOS and PFOA, as reflected in neonatal concentrations, may pose risk for child behavioral difficulties.
Keywords: endocrine disrupting chemicals, perfluoroalkyl substances, neonates, behavior, dried blood spots
Graphical Abstract
Capsule
In children born between 2008 and 2010, we found that higher neonatal concentration of perfluorooctane sulfonate (PFOS) were associated with behavioral difficulties at age 7 years.
Introduction
An increasing number of children is diagnosed annually with developmental disabilities and behavioral problems, while our understanding of the etiology is still incomplete. Recently, attention has been directed toward the toxic effects of environmental chemicals.1 Despite emerging concerns, data on the neurologic influences of chemicals such as bisphenol A (BPA) and perfluoroalkyl substances (PFASs) are inconclusive. While the U.S. Food and Drug Administration regulates BPA usage in baby products, the exposure to BPA is still considered ‘safe’ at the levels occurring in adults, including pregnant women. Moreover, since 2002, major US companies have been phasing out two PFASs, perfluorooctanoic acid (PFOA) and perfluorooctane sulfonate (PFOS). Nonetheless, children born after reduction in manufacturing use remained exposed to these chemicals.2,3
PFASs are widely used in textiles, furniture, and cookware. The neurotoxic effects of PFASs occur through modulating the dopaminergic system in brain regions as well as interference with thyroid function and inducing oxidative stress.4,5 Despite a detrimental effect on fetal growth and the endocrine system,6–8 most of the epidemiological studies did not show associations between maternal serum levels of PFOA and PFOS and children’s motor and mental development,9,10 behavioral and coordination problems,11,12 or attention-deficit/hyperactivity disorder (ADHD) and autism spectrum disorder (ASD).13,14 In contrast, the Taiwan Birth Panel Study (n= 239) found that cord blood PFOS was associated with poor developmental outcomes at age 2 years.15 Other researchers showed small to moderate associations between maternal PFOS and PFOA and children’s hyperactive behavior and executive function (n= 256).16,17
BPA is used in the production of aluminum cans, plastics, and thermal paper receipts. BPA is a weak estrogenic substance18 but also interferes with thyroid function.19,20 In 240 children from the Health Outcomes and Measures of the Environment (HOME) Study, Braun et al. showed a positive association between maternal urinary levels of BPA and externalizing behaviors21 and anxious-depressed behavior and hyperactivity (predominantly in girls).22 In another cohort, there was no association between urinary levels of BPA and children’s behavioral problems; however there was an indication for sex-specific associations.23–25 In INfancia y Medio Ambiente (INMA)-Sabadell cohort, maternal urinary levels of BPA was associated with increased risk of ADHD symptoms, in particular in boys, but this association did not last until a later age.26
While prospective studies with early measurement of chemical exposure can further clarify the programming effect on brain development, exploring new methodologies may further scientific progress. Most of the previous studies measured PFASs in maternal serum and BPA in maternal urine during pregnancy. It is less clear to what extent maternal levels reflect child exposure during fetal brain development throughout pregnancy. We previously reported on quantification of BPA, PFOS, and PFOA in stored newborn dried blood spots (DBS) using high-performance liquid chromatography/tandem mass spectrometry (HPLC–MS/MS).27 Here, we examined the prospective associations between neonatal levels of PFOA, PFOS, and BPA and children’s behavioral difficulties at the age of 7 years.
Materials and Methods
Participants
We used data from the Upstate KIDS Study, a matched exposure cohort in upstate New York (2008-2010).28 Upstate KIDS was originally designed to evaluate the long-term impact of infertility treatment on child development and showed that infertility treatment was not associated with children’s development after accounting for plurality.29 Recruitment was based on indications of infertility treatment and plurality on birth certificates. Briefly, all children conceived by infertility treatment and all multiples–regardless of their mode of conception-were recruited. In addition, singletons who were not conceived by infertility treatment were recruited at a 3:1 ratio to those conceived by treatment, while frequency matching on region of birth.
When children were eight months old, parents gave consent to use archived newborn DBS for chemical analysis (2071 singletons, 1040 unrelated twins, and 64 higher order multiples). In these analyses, we included singletons and twins with consent on the use of stored newborn DBS and information on child behavior at age 7 years (650 singletons and 134 twin pairs). Higher order multiples were excluded because of the small number.
The New York State Department of Health (NYSDOH) and the University at Albany Institutional Review Board (IRB) approved the study and served as the IRB designated by the National Institutes of Health for this study under a reliance agreement. Parents provided written informed consent.
Endocrine disrupting chemicals in dried blood spots
In New York State, DBSs are obtained as part of the Newborn Screening Program 1-2 days after birth. DBS were stored at 4°C after collection and later retrieved for blood spot analyses.
Punches of the residual spots were extracted in a manner as described previously.30 We measured PFOA, PFOS, and BPA in archived DBS (16-mm diameter discs) using a highly sensitive liquid-liquid extraction and HPLC–MS/MS.27 Determination of concentrations of environmental chemicals in DBS requires assays with great sensitivity. Thus, with HPLC-MS/MS, PFOS and PFOA were detected in 100% of the specimens analyzed (with the limit of detection as low as 0.03 and 0.05 ng/mL). Background contamination was taken into account by taking unspotted sections of the filter paper as field blanks to account for any contamination that arise from hospital settings and subtracting these background levels from measured concentrations. The concentrations of PFOS and PFOA measured in newborn DBS were similar to those reported earlier in the whole blood samples of newborns31 and DBS.32,33 In the Upstate KIDS sample, BPA was detected in 86% of the specimen.
Child Behavior at age 7 years
Mothers rated their children’s behavior using the Strengths and Difficulties Questionnaire (SDQ) at approximately 7 years of age.34 The SDQ is a validated and reliable questionnaire that provides a quantitative measure of a child’s behavior in 5 domains of emotional symptoms (5 items), conduct problems (5 items), hyperactivity/inattention (5 items), peer relationship problems (5 items), and prosocial behaviors (5 items). Items were scored as 0 (‘not true’), 1 (‘somewhat true’) or 2 (‘certainly true’) by parents. For each child, a total SDQ score (‘total difficulties’) as well as five scale scores were obtained by summing up the responses. The total difficulties score was generated by summing scores from all scales except the prosocial behavior scale. The SDQ’s emphasis on strengths (items such as ‘considerate of other people’s feelings’) as well as difficulties (items such as ‘restless, overactive, cannot stay still for long’) makes it particularly acceptable to community samples. The SDQ has been widely used in epidemiological studies examining the genetic and environmental determinants of children’s behavior.35,36 In line with other studies in the U.S., we used the recommended cut-offs to identify children with behavioral problems within the borderline/abnormal and abnormal range (e.g., for total difficulties score: scores 0-13=‘Normal’; scores 14-16=‘Borderline/Abnormal’; and scores 17-40=‘Abnormal’).37,38
Covariates
At four months postpartum, mothers reported on their race/ethnicity, highest acquired education level, insurance, marital/cohabitation status, weight and height, and history of smoking and alcohol drinking during pregnancy. We obtained information on parity, plurality, mother’s weight and height prior to pregnancy, infant’s sex, gestational age, birth weight, intensive care unit (NICU) admission, and breastfeeding in neonates from birth certificate. Maternal body mass index (BMI) was calculated using pre-pregnancy weight and height obtained from birth certificates (and maternal questionnaire if missing). Children with developmental disabilities (i.e., ASD, ADHD, language, learning and speech disorders, cognitive deficits, cerebral palsy, and/or sensory impairments) were identified at age 4 years using a multi-step approach, described in details previously.39
Statistical analysis
In descriptive statistics, child characteristics are shown for all singletons and a randomly selected twin of each sibling pair. We used independent sample t tests and chi-square statistics to determine differences in maternal and child characteristics between participants in this analysis and children excluded because of missing behavioral data.
All singletons and twin pairs were included in the association analysis (918 children from 788 families). We included chemical concentrations both continuously (log-transformed for normality) and categorically (in quartiles). We performed negative binomial regressions to estimate the coefficients and 95% confidence intervals of the associations between each chemical and the total behavioral difficulties scores and subdomain scores. Analysis with prosocial score did not have a good model fit and, therefore, the results with continuous score for this subscale were not reported. Analyses were conducted using generalized estimating equations (GEE) with robust standard error to account for the clustering nature of data due to the large number of twins (268 twins, 29%). Since categorical analysis of SDQ provides valuable insight about children at risk of psychopathology, we used borderline/abnormal cut-offs to define children with behavioral problems and ran generalized linear models with logit function using GEE to examine the associations between chemicals and behavioral difficulties. Models were rerun in a sample excluding children with developmental disabilities (n=22). Effect sizes were reported per standard deviation (SD) increment in the exposure level (in log scale). Sex-specific effects have been reported for prenatal exposure to PFASs and BPA.23,24,40 Therefore, we examined interactions between chemical exposures and a child’s sex in relation to behavioral problems. Models were also run stratifying on infertility treatment and according to plurality (singletons and twins). For sensitivity analysis, we imputed missing values of behavioral problems in all children with chemical data (n=2604). Fifty copies of the original dataset were generated, with missing values replaced by values randomly generated from the posterior distribution based on a fully conditional specification (FCS) method for imputation which uses a multivariate imputation by chained equations method to impute values for a data set with an arbitrary missing pattern, assuming a joint distribution exists for the data. The joint distribution is based on variables with missing values in this dataset (maternal BMI (n=6), parity (n=29), marital status (n=107), and children’s behavioral problems and age at assessment (n=2193)) and other variables (i.e., maternal age, race/ethnicity, education, and history of smoking and drinking in pregnancy, and a child’s sex, birth weight, and chemical concentrations). We then ran the models in the imputed dataset.
Models were adjusted for a child’s age (continuous) and sex (boy/girl), maternal age (continuous), pre-pregnancy body mass index (continuous), race/ethnicity (Non-Hispanic White/Non White or other), educational levels (high school or less/some college/college graduate/graduate and higher), marital status (married/living with partner), history of smoking in pregnancy (yes/no), having private insurance (yes/no), and parity (nulliparous/para 1/para 2 and higher), as these factors are associated with chemical exposures and child behavior.41–45 Models were additionally adjusted for preterm delivery in BPA models (because of BPA short half-life DBS levels were probably a reflection of infant exposure rather than maternal exposure). All models were adjusted for infertility treatment to account for the study’s design of oversampling infants conceived with infertility treatment.
Statistical analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC).
Results
We observed that children included in the analysis were more likely to be singletons (82% vs 78%, p=0.01) and first born (50% vs 44%, p<0.001) and to have larger birth weights (mean difference=93 grams, p=0.001) as compared to children excluded on the basis of missing SDQ data. Mothers of children with SDQ data were slightly older (mean difference=1.1 years, p<0.001), were more likely to be married (93% vs 90%, p=0.01) and Non-Hispanic White (89% vs 81%, p<0.001), smoked less in pregnancy (7% vs 13%, p<0.001), and had a lower BMI (mean difference=−0.6, p=0.03) as compared to mothers who did not complete the behavioral rating for their children. Factors such as a child’s sex, NICU admission, and maternal history of drinking during pregnancy did not differ between the two groups (data not shown). There was no difference in PFOS and BPA concentrations between children included in the analysis and those excluded (mean difference in PFOS=0.03 ng/mL, p=0.42 and mean difference in BPA 0.04 ng/mL, p=0.37, respectively). The mean difference in PFOA between two groups was 0.08 ng/mL, p=0.04.
Table 1 and Supplementary Table 1 summarize participants’ characteristics. Univariate analysis showed that children with neonatal PFOS concentrations in the highest quartile had a smaller birth weight compared to children with concentrations in the lowest quartiles. Parity, maternal race/ethnicity, and BMI were also associated with PFOS concentrations. Using borderline/abnormal cut-off for SDQ,34 111 (85 singletons and 26 twins, 12.1%) children had behavioral problems. Using abnormal cut-off for SDQ, this number was instead 60 (43 singletons and 17 twins, 6.5%). Using borderline/abnormal cut-off, 6.0% of children had difficulties in prosocial behavior, and with abnormal cut-off this number was 2.4%. The mean total behavioral difficulties score was 6.6 (SD=5.4, range 0–28). The median for neonatal concentration of PFOS in DBS was 1.74 ng/ml (interquartile range (IQR) = 1.33). The medians (IQR) of PFOA and BPA concentrations were 1.12 ng/ml (0.96) and 7.93 ng/ml (10.79), respectively. There was a modest correlation between PFOS and PFOA concentrations (r= 0.58, p <0.001), but no correlation was observed with BPA (r= −0.004 with PFOS and 0.005 with PFOA).
Table 1.
Participants’ characteristics. Upstate KIDS Study.
| Participants included in this analysis | Participants with consent to use DBS | |
|---|---|---|
| N=788 | N=2604 | |
| Maternal characteristics | ||
| Age, years | 31.8 (5.6) | 31.1 |
| Being married or living with partners | 721 (92.8) | 2287 (90.6) |
| Parity, nulliparous, % | 392 (50.3) | 1186 (45.6) |
| Race/ethnicity, % | ||
| Non-Hispanic White | 704 (89.3) | 2180(83.7) |
| Not White or Other | 84 (10.7) | 424 (16.3) |
| Educational levels, % | ||
| High school equivalent or less | 75 (9.5) | 367 (14.1) |
| Some college | 176 (22.3) | 743 (28.5) |
| College graduate | 208 (26.4) | 665 (25.5) |
| Graduate/professional school | 329 (41.8) | 829 (31.8) |
| Alcohol consumption during pregnancy, % | 110 (14.0) | 354 (13.6) |
| Smoking in pregnancy, % | 58 (7.4) | 290 (11.1) |
| Pre-pregnancy BMI, kg/m2 | 26.5 (6.7) | 26.9 (6.7) |
| Normal weight (BMI <25) | 417 (52.9) | 1276 (49.0) |
| Overweight (BMI 25–29.9) | 201 (25.5) | 671 (25.8) |
| Obese (BMI ≥30) | 170 (21.6) | 657 (25.2) |
| Private health insurance, % | 682 (86.6) | 2081 (79.9) |
| Infertility treatment, % | 304 (38.6) | 843 (32.4) |
| Child characteristics | ||
| Age at SDQ assessment, years | 7.1 (0.6) | - |
| Gender, male, % | 400 (50.8) | 1329 (51.0) |
| Plurality, % | ||
| Singletons | 650 (82.5) | 2071 (79.5) |
| Twins | 138 (17.5) | 533 (20.5) |
| Birth weight, grams | 3279 (667) | 3214 (665) |
| Being preterm, % | 112 (14.2) | 430 (16.5) |
| Gestational age at birth, weeks | 39 (38, 40) | 39 (37–40) |
| Any breastfeeding in neonates, % | 648 (83.2) | 2573 (81.3) |
| NICU admission, yes, % | 95 (12.1) | 333 (12.8) |
Body Mass Index: BMI; Dried Blood Spots, DBS; Neonatal Intensive Care Unit: NICU; Strengths and Difficulties Questionnaire: SDQ.
Numbers are mean (SD) for continuous variables with normal distribution, median (quartile range) for continuous variables with skewed distribution and n (%) for categorical variables. For child-level data, descriptive statistics were derived from all singletons and one randomly selected twin of each pair.
Behavioral difficulties were defined as having SDQ Total Difficulties Score within the borderline/abnormal range.
Tables 2–4 represent associations between neonatal concentrations of PFASs and BPA and behavioral problems at age 7 years. Higher PFOS levels were associated with increased odds of having behavioral difficulties (OR per SD of log PFOS=1.30, 95%CI: 1.03-1.65) (Table 2). We observed associations between PFOS in the highest relative to the lowest quartile and behavioral difficulties (OR for PFOS1.14-1.74=1.67, 95%CI: 0.84-3.34; PFOS1.75-2.47=1.73 , 95%CI: 0.87-3.43; and PFOS>2.47=2.47, 95%CI: 1.29-4.72 compared to PFOS<1.41). The associations with PFOS were not significant when continuous behavioral scores were used. There was no relationship between PFOA and BPA and total behavioral difficulties (continuous and categorical analyses).
Table 2.
Neonatal concentrations of perfluoroalkyl substances and bisphenol A with behavioral problems at age 7 years. Upstate KIDS Study (n=918).
| Total Behavioral Difficulties, Total scores 1 | Total Behavioral Difficulties, Borderline problems 2 | |||
|---|---|---|---|---|
| n | β (95%CI) | n 3 | OR (95%CI) | |
| PFOS | ||||
| Unadjusted | ||||
| Per log SD | 918 | 0.03 (−0.03, 0.08) | 111/918 | 1.22 (0.99, 1.50) |
| Q1 | 222 | reference | 11/222 | reference |
| Q2 | 236 | 0.14 (−0.01, 0.30) | 30/236 | 1.68 (0.88, 3.22) |
| Q3 | 235 | 0.05 (−0.11, 0.20) | 28/235 | 1.61 (0.85, 3.03) |
| Q4 | 225 | 0.14 (−0.01, 0.27) | 35/225 | 2.09 (1.12, 3.88) |
| Adjusted | ||||
| Per log SD | 891 | 0.04 (−0.02, 0.10) | 111/891 | 1.30 (1.03, 1.65) |
| Q1 | 213 | reference | 17/213 | reference |
| Q2 | 231 | 0.14 (−0.01, 0.28) | 28/231 | 1.67 (0.84, 3.34) |
| Q3 | 230 | 0.04 (−0.11, 0.19) | 28/230 | 1.73 (0.87, 3.43) |
| Q4 | 217 | 0.17 (0.01, 0.32) | 35/217 | 2.47 (1.29, 4.72) |
| PFOA | ||||
| Unadjusted | ||||
| Per log SD | 918 | 0.01 (−0.04, 0.07) | 111/918 | 1.15 (0.95, 1.39) |
| Q1 | 232 | reference | 25/232 | reference |
| Q2 | 227 | −0.04 (−0.19, 0.11) | 25/227 | 1.00 (0.56, 1.78) |
| Q3 | 237 | 0.08 (−0.07, 0.23) | 33/237 | 1.37 (0.78, 2.40) |
| Q4 | 222 | −0.01 (−0.16, 0.14) | 28/222 | 1.22 (0.69, 2.19) |
| Adjusted | ||||
| Per log SD | 891 | −0.01 (−0.06, 0.05) | 111/891 | 1.13 (0.91, 1.40) |
| Q1 | 221 | reference | 22/221 | reference |
| Q2 | 223 | −0.05 (−0.19, 0.10) | 25/223 | 1.07 (0.56, 2.03) |
| Q3 | 230 | 0.03 (−0.12, 0.17) | 33/230 | 1.37 (0.74, 2.53) |
| Q4 | 217 | −0.05 (−0.21, 0.12) | 28/217 | 1.30 (0.63, 2.40) |
| BPA | ||||
| Unadjusted | ||||
| Per log SD | 918 | −0.004 (−0.07, 0.06) | 111/918 | 1.03 (0.80, 1.31) |
| Q1 | 227 | reference | 30/227 | reference |
| Q2 | 233 | −0.08 (−0.22, 0.06) | 21/233 | 0.66 (0.38, 1.14) |
| Q3 | 230 | 0.005 (−0.14, 0.15) | 31/230 | 1.02 (0.60, 1.71) |
| Q4 | 228 | −0.03 (−0.18, 0.12) | 29/228 | 0.91 (0.52, 1.58) |
| Adjusted | ||||
| Per log SD | 891 | −0.01 (−0.07, 0.06) | 111/891 | 1.00 (0.77, 1.29) |
| Q1 | 223 | reference | 30/223 | reference |
| Q2 | 227 | −0.09 (−0.24, 0.05) | 19/227 | 0.49 (0.26, 0.94) |
| Q3 | 219 | −0.04 (−0.18, 0.11) | 30/219 | 0.84 (0.47, 1.51) |
| Q4 | 222 | −0.07 (−0.22, 0.08) | 29/222 | 0.80 (0.44, 1.45) |
Perfluorooctanesulfonic acid: PFOS; Perfluorooctanoic acid: PFOA; Bisphenol A: BPA.
Continuous variable as outcome.
Categorical variable as outcome. Behavioral difficulties were defined as having Strengths and Difficulties Questionnaire Total Difficulties Score within the borderline/abnormal range.
Number of children with behavioral difficulties / all children within the category of exposure. Chemical concentrations were log-transformed for normality and coefficients were estimated per standard deviation change in log-scale of concentrations.
Models were adjusted for a child’s age (year) and sex (boy/girl), and maternal age (year), pre-pregnancy body mass index (kg/m2), race/ethnicity (Non-Hispanic White/Not White or others), education (less than high school, high school, some college, college, advanced), marital status (married or living with partner), history of smoking in pregnancy (yes/no), having private insurance (yes/no), parity (nulliparous/para 1/para 2 and higher), and infertility treatment (yes/no). Models were additionally adjusted for preterm delivery (yes/no) for BPA.
Table 4.
Neonatal concentrations of perfluorooctanesulfonic acid (PFOS) with hyperactivity, conduct, peer, and emotional problems at age 7 years. Upstate KIDS Study (n=918).
| Hyperactivity Problems | Conduct Problems | Peer Problems | Emotional Problems | |
|---|---|---|---|---|
| OR (95%CI) | OR (95%CI) | OR (95%CI) | OR (95%CI) | |
| PFOS | ||||
| Unadjusted | ||||
| Per log SD | 1.03 (0.86, 1.24) | 1.06 (0.86, 1.30) | 0.98 (0.81, 1.18) | 1.33 (1.10, 1.61) |
| Q1 | reference | reference | reference | reference |
| Q2 | 1.30 (0.75, 2.23) | 1.78 (0.97, 3.27) | 1.31 (0.80, 2.14) | 2.13 (1.21, 3.74) |
| Q3 | 1.10 (0.64, 1.90) | 0.86 (0.43, 1.74) | 1.23 (0.75, 2.23) | 1.01 (0.55, 1.87) |
| Q4 | 1.45 (0.85, 2.46) | 2.22 (1.18, 4.15) | 1.45 (0.85, 2.46) | 2.62 (1.51, 4.55) |
| Adjusted | ||||
| Per log SD | 0.99 (0.81, 1.22) | 1.22 (0.97, 1.52) | 1.02 (0.83, | 1.31 (1.04, 1.63) |
| Q1 | reference | reference | reference | reference |
| Q2 | 1.23 (0.69, 2.20) | 1.78 (0.97, 3.27) | 1.27 (0.74, 2.17) | 2.08 (1.13, 3.80) |
| Q3 | 0.95 (0.52, 1.74) | 0.86 (0.43, 1.74) | 1.31 (0.77, 2.24) | 0.89 (0.47, 1.68) |
| Q4 | 1.39 (0.78, 2.46) | 2.22 (1.18, 4.15) | 1.08 (0.62, 1.88) | 2.28 (1.24, 4.18) |
| PFOA | ||||
| Unadjusted | ||||
| Per log SD | 1.05 (0.88, 1.26) | 0.89 (0.73, 1.09) | 0.95 (0.79, 1.13) | 1.20 (1.02, 1.42) |
| Q1 | reference | reference | reference | reference |
| Q2 | 0.93 (0.54, 1.58) | 0.75 (0.43, 1.31) | 0.93 (0.56, 1.54) | 1.49 (0.87, 2.55) |
| Q3 | 1.22 (0.73, 2.04) | 1.12 (0.67, 1.87) | 0.96 (0.58, 1.57) | 1.47 (0.86, 2.50) |
| Q4 | 1.12 (0.66, 1.89) | 0.58 (0.32, 1.05) | 0.87 (0.52, 1.44) | 1.32 (0.76, 2.28) |
| Adjusted | ||||
| Per log SD | 1.01 (0.82, 1.23) | 0.97 (0.76, 1.23) | 0.94 (0.77, 1.14) | 1.09 (0.90, 1.32) |
| Q1 | reference | reference | reference | reference |
| Q2 | 0.94 (0.53, 1.67) | 0.92 (0.51, 1.67) | 0.98 (0.57, 1.69) | 1.29 (0.72, 2.33) |
| Q3 | 1.09 (0.63, 1.89) | 1.31 (0.73, 2.34) | 0.94 (0.55, 1.60) | 1.16 (0.65, 2.07) |
| Q4 | 1.04 (0.56, 1.90) | 0.75 (0.37, 1.51) | 0.84 (0.47, 1.47) | 0.95 (0.52, 1.77) |
| BPA | ||||
| Unadjusted | ||||
| Per log SD | 0.97 (0.77, 1.24) | 1.01 (0.80, 1.28) | 1.01 (0.83, 1.23) | 0.86 (0.68, 1.07) |
| Q1 | reference | reference | reference | reference |
| Q2 | 0.72 (0.44, 1.21) | 0.95 (0.55, 1.62) | 1.02 (0.60, 1.72) | 0.83 (0.50, 1.37) |
| Q3 | 0.93 (0.57, 1.51) | 1.07 (0.63, 1.82) | 1.40 (0.86, 2.26) | 0.69 (0.41, 1.15) |
| Q4 | 0.81 (0.49, 1.35) | 0.96 (0.55, 1.67) | 1.14 (0.69, 1.90) | 0.71 (0.41, 1.22) |
| Adjusted | ||||
| Per log SD | 0.95 (0.76, 1.20) | 0.99 (0.78, 1.26) | 0.98 (0.79, 1.23) | 1.19 (0.66, 1.06) |
| Q1 | reference | reference | reference | reference |
| Q2 | 0.66 (0.38, 1.15) | 0.90 (0.51, 1.60) | 0.95 (0.53, 1.68) | 0.75 (0.43, 1.31) |
| Q3 | 0.76 (0.44, 1.31) | 1.04 (0.59, 1.84) | 1.27 (0.76, 2.12) | 0.64 (0.37, 1.08) |
| Q4 | 0.76 (0.44, 1.31) | 0.95 (0.53, 1.70) | 1.08 (0.63, 1.87) | 0.59 (0.33, 1.06) |
Perfluorooctanesulfonic acid: PFOS; Perfluorooctanoic acid: PFOA; Bisphenol A: BPA.
Chemical concentrations were log-transformed for normality and coefficients were estimated per standard deviation change in log-scale of concentrations.
Problems were defined as having Strengths and Difficulties Questionnaire Scores within the borderline/abnormal range.
Models were adjusted for a child’s age (year) and sex (boy/girl), and maternal age (year), pre-pregnancy body mass index (kg/m2), race/ethnicity (Non-Hispanic White/Not White or others), education (less than high school, high school, some college, college, advanced), marital status (married or living with partner), history of smoking in pregnancy (yes/no), having private insurance (yes/no), parity (nulliparous/para 1/para 2 and more), and infertility treatment (yes/no). Models were additionally adjusted for preterm delivery (yes/no) for BPA.
Higher concentrations of PFOA were linearly associated with difficulties in prosocial behavior (OR=1.35, 95%CI: 1.00-1.75) (Table 3). Categorical analysis of PFOA with difficulties in prosocial behavior revealed similar results. While the continuous analysis of BPA showed an inverse association between BPA and difficulties in prosocial behavior, this association was not present when we categorized BPA in quartiles (Table 3).
Table 3.
Neonatal concentrations of perfluoroalkyl substances and bisphenol A with difficulties in prosocial behavior at age 7 years. Upstate KIDS Study (n=918).
| Difficulties in Prosocial Behavior |
||
|---|---|---|
| n1 | OR (95%CI) | |
| PFOS | ||
| Unadjusted | ||
| Per log SD | 55/918 | 1.10 (0.86, 1.42) |
| Q1 | 12/222 | reference |
| Q2 | 13/236 | 0.83 (0.34, 2.09) |
| Q3 | 14/235 | 1.12 (0.51, 2.49) |
| Q4 | 16/225 | 1.37 (0.64, 2.94) |
| Adjusted | ||
| Per log SD | 50/891 | 1.26 (0.92, 1.72) |
| Q1 | 11/213 | reference |
| Q2 | 11/231 | 0.86 (0.35, 2.15) |
| Q3 | 14/230 | 1.72 (0.65, 4.52) |
| Q4 | 14/217 | 1.87 (0.70, 4.98) |
| PFOA | ||
| Unadjusted | ||
| Per log SD | 55/918 | 1.11 (0.87, 1.41) |
| Q1 | 11/213 | reference |
| Q2 | 15/227 | 1.34 (0.57, 3.16) |
| Q3 | 17/237 | 1.59 (0.68, 3.74) |
| Q4 | 12/222 | 1.28 (0.54, 2.99) |
| Adjusted | ||
| Per log SD | 50/891 | 1.35 (1.03, 1.75) |
| Q1 | 7/221 | reference |
| Q2 | 15/223 | 2.63 (0.97, 7.14) |
| Q3 | 16/230 | 2.93 (1.03, 8.28) |
| Q4 | 12/217 | 3.23 (1.04, 10.07) |
| BPA | ||
| Unadjusted | ||
| Per log SD | 55/918 | 0.70 (0.52, 0.94) |
| Q1 | 14/227 | reference |
| Q2 | 11/233 | 0.75 (0.32, 1.78) |
| Q3 | 23/230 | 1.62 (0.77, 3.39) |
| Q4 | 7/228 | 0.46 (0.18, 1.21) |
| Adjusted | ||
| Per log SD | 50/891 | 0.69 (0.50, 0.94) |
| Q1 | 13/223 | reference |
| Q2 | 9/227 | 0.60 (0.22, 1.60) |
| Q3 | 21/219 | 1.66 (0.74, 3.70) |
| Q4 | 7/222 | 0.52 (0.19, 1.45) |
Perfluorooctanesulfonic acid: PFOS; Perfluorooctanoic acid: PFOA; Bisphenol A: BPA.
Numbers indicate children with behavioral difficulties / all children within the category of exposure. Difficulties in prosocial behavior were defined as having Strengths and Difficulties Questionnaire Prosocial Score within the borderline/abnormal range.
Chemical concentrations were log-transformed for normality and coefficients were estimated per standard deviation change in log-scale of concentrations.
Models were adjusted for a child’s age (year) and sex (boy/girl), and maternal age (year), pre-pregnancy body mass index (kg/m2), race/ethnicity (Non-Hispanic White/Not White or others), education (less than high school, high school, some college, college, advanced), marital status (married or living with partner), history of smoking in pregnancy (yes/no), having private insurance (yes/no), parity (nulliparous/para 1/para2 and higher), and infertility treatment (yes/no). Models were additionally adjusted for preterm delivery (yes/no) for BPA.
We further examined the association between neonatal chemical concentrations and four domains of total behavioral difficulties and found that the observed association with PFOS was driven by conduct and emotional problems (Table 4 and Supplementary Table 2).
There was no interaction between chemical concentrations and sex (p for interaction with PFOS= 0.11; for PFOA=0.98; for BPA=0.20) in relation to behavioral problems. Stratification by infertility treatment showed that infertility treatment did not modify the association between chemicals and child behavioral difficulties (Supplementary Table 3). The association between PFOS and total behavioral difficulties were homogeneous in singletons and twins (Supplementary Table 4). Associations between PFOS and total behavior difficulties remained unchanged when missing values in behavior were imputed (Supplemental Table 5) or if 22 children with developmental disabilities were excluded (data not shown).
Discussion
In this large sample of children participating in a US contemporary cohort, we used archived neonatal DBS to measure PFOS, PFOA, and BPA. We found that higher concentrations of PFOS in newborns were associated with behavioral difficulties at age 7 years, driven by problems in conduct and emotional symptoms. Our analysis suggested that the associations of PFOS with conduct and emotional problems were not linear, only present for exposure concentrations to PFOS in the 2nd and 4th quartiles. Higher concentrations of PFOA were associated with having difficulties in prosocial behavior. There was no relationship between neonatal BPA and behavioral problems. There was an inverse association between BPA concentrations and difficulties in prosocial behavior, but not if BPA was analyzed in quartiles. The novelty of this study is measuring concentrations of chemicals directly in newborns rather than relying on maternal blood or urine concentrations. Our sample had a substantial number of twins which allowed us to examine, for the first time, if associations differ by plurality.
Several longitudinal studies examined the association between PFOA and PFOS exposures during gestation and child neurodevelopment with the majority of them reporting null associations.46 In design and sample size our study is similar to the study by Fei and Olsen in the Danish National Birth Cohort (DNBC), with the exception that they measured PFOS and PFOA in maternal blood around 8 week gestation.11 Behavioral outcomes were measured at age 7 years using SDQ in both studies and the patterns of confounding were very similar, with race and parity being important factors associated with chemical levels. In our study, factors such as pre-pregnancy BMI and breastfeeding were not associated with PFAS concentrations. The median concentrations of PFOS and PFOA in maternal blood were higher than the median concentrations in DBS in our cohort; even though comparison should be done with caution because of the different biospecimens used (maternal serum vs. neonatal DBS). Fei and Olsen reported no association between maternal concentrations of PFOA and PFOS and children’s SDQ scores. The inconsistency can be explained by differences in baseline exposure in the population of upstate New York and Denmark along with time of birth (2008-2010 for Upstate KIDS vs. 1996-2002 for DNBC). Examination of water sources has shown that industrialized areas have, in general, had higher contamination of PFOS and PFOA concentrations than non-industrialized areas.47 Water contamination with PFOA in particular has been a concern in upstate New York, which has attracted specific attention in recent years.48 In line with this, a cross-sectional study in a group of 9 to 11 year old children in upstate New York found associations between higher blood levels of PFASs and children’s impulsivity.49 Another explanation for different findings in our study and the DNBC study might be that neonatal concentrations are a better reflection of exposure in the child at birth relative to maternal concentrations that are often used as proxy exposures for the fetus/infant. We observed that problems in domains of conduct and emotional symptoms were associated with higher exposure to PFOS, whereas higher concentrations of PFOA were associated with difficulties in prosocial behavior. Prosocial and other four subscales of SDQ assess distinct aspects of the child’s behavior; prosocial items are closely related to the temperamental trait of callus/unemotional; while emotional items represent internalizing problems and conduct problems are an indication of externalizing problems.50 The association with psychopathology is the weakest for prosocial problems among all subscales.34 In vivo examination confirms that direct neurotoxic effects of PFOA and PFOS are through different processes during neural cell differentiation.51 This can potentially translate to various neurodevelopmental outcomes (prosocial, internalizing or externalizing). One epidemiological study has shown that prenatal exposure to PFOS but not PFOA was associated with poorer behavioral regulation.17 Yet, such domain specific associations need further investigation.
We found no association between neonatal BPA concentrations and behavioral outcomes. We observed an inverse association between BPA and difficulties in prosocial behavior, but the association was not present when BPA was analyzed in quartiles. There was no sex difference in the association between neonatal BPA and behavioral problems. Further stratification by sex, and examination of a sex interaction in specific subdomains such as emotional problems and hyperactivity, hypothesized based on findings in the HOME cohort and other studies,21,22,25 revealed null results. It is possible that behavioral abnormalities associated with BPA exposure in young children does not last till older ages, as supported by findings in the INMA cohort (no association with BPA if behavior was measured at age 7).26 However, the follow-up of children in an inner-city cohort showed that the association between prenatal BPA and anxiety and depression in boys lasted until 10-12 years of age.25 It is noteworthy that these cohorts measured BPA in maternal urine samples, which should be considered when comparing their findings to ours. Because of the short half-life of BPA, exposure assessment, especially with a single measurement in urine or other biospecimen, remains a challenge. BPA concentrations as measured in newborn DBS might be influenced by postnatal exposure. Background contamination could potentially be a problem; however, we performed several quality assurance procedures, including field blanks from unspotted sections of the DBS.27 Therefore, DBS levels reported here represent true infant levels, but could be a reflection of recent exposures in the child rather than prenatal exposures.
Our findings should be interpreted in the light of a number of limitations. First, we did not have any measure of chemical exposures in the period between infancy and age 7 years. However, Perera et al. showed that additional adjustment for the postnatal BPA levels in children did not influence the effect estimates of prenatal exposures,25 suggesting that developmental programming of chemical exposures happen early during gestation. Second, we had substantial non-response at age 7 years, which was selective with regard to sociodemographic characteristics. However, non-response was not related to chemical concentrations. Moreover, the associations between PFOS and child behavior remained unchanged when we imputed behavioral outcomes using multiple imputations. Third, SDQ is not a diagnostic tool. Nonetheless, it is a sensitive and specific screening tool for diagnosis of developmental disabilities.37 Fourth, we did not repeat measures of chemicals. Due to BPA’s short half-life (<6 hours), single measurements might not characterize exposure well, or reflect exposure during early gestation. We examined whether the study’s design for oversampling on infertility treatment impacted results but found the results remained robust after exclusion of children conceived with infertility or adjustment for treatment. We previously children conceived with infertility treatment and children conceived spontaneously show few differences with respect to their developmental and behavioral outcomes.29 In addition, we did not have information on maternal IQ, which might influence the association between exposure to persistent chemicals and child behavior.
Conclusions
We showed that neonatal concentrations of two commonly measured perfluoroalkyl substances were associated with age-seven behavioral problems, suggesting that increasing prenatal exposure to PFOS and PFOA may pose risk of child behavioral difficulties. Newborn DBS is a rich and widely available source that can be used for measurement of early exposures in neonates.
Supplementary Material
Highlights.
We quantified concentrations of chemicals in newborn dried blood spots (DBS).
Higher PFOS levels were related to increased odds of behavioral difficulties.
Higher PFOA levels were associated with difficulties in prosocial behavior.
Newborn DBS can be used as a biospecimen for measurement of persistent chemicals.
Acknowledgment
The authors thank all the Upstate KIDS participants and staff for their important contributions.
Funding sources: This work was supported by the Intramural Research Program of the Eunice Kennedy Shriver National Institute of Child Health and Human Development, Bethesda, MD (NICHD; contracts #HHSN275201200005C, #HHSN267200700019C, #HHSN275201400013C, #HHSN275201300026I/27500004). The sponsor played no role in study design; in the collection, analysis, and interpretation of data; in the writing of report; and in the decision to submit the article for publication.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Declarations of interests: None.
References
- 1.Bennett D, Bellinger DC, Birnbaum LS, et al. Project TENDR: Targeting Environmental Neuro-Developmental Risks The TENDR Consensus Statement. Environ Health Perspect 2016;124(7):A118–122. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Ye X, Kato K, Wong LY, et al. Per- and polyfluoroalkyl substances in sera from children 3 to 11 years of age participating in the National Health and Nutrition Examination Survey 2013-2014. Int J Hyg Environ Health. 2018;221(1):9–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Khalil N, Ebert JR, Honda M, et al. Perfluoroalkyl substances, bone density, and cardio-metabolic risk factors in obese 8-12 year old children: A pilot study. Environ Res 2018;160:314–321. [DOI] [PubMed] [Google Scholar]
- 4.Salgado R, López-Doval S, Pereiro N, et al. Perfluorooctane sulfonate (PFOS) exposure could modify the dopaminergic system in several limbic brain regions. Toxicol Lett 2016;240(1):226–235. [DOI] [PubMed] [Google Scholar]
- 5.Mariussen E Neurotoxic effects of perfluoroalkylated compounds: mechanisms of action and environmental relevance. Arch Toxicol 2012;86(9):1349–1367. [DOI] [PubMed] [Google Scholar]
- 6.Bach CC, Bech BH, Brix N, et al. Perfluoroalkyl and polyfluoroalkyl substances and human fetal growth: a systematic review. Crit Rev Toxicol 2015;45(1):53–67. [DOI] [PubMed] [Google Scholar]
- 7.Lind DV, Priskorn L, Lassen TH, et al. Prenatal exposure to perfluoroalkyl substances and anogenital distance at 3 months of age in a Danish mother-child cohort. Reprod Toxicol 2016. [DOI] [PubMed] [Google Scholar]
- 8.Berg V, Nost TH, Pettersen RD, et al. Persistent organic pollutants and the association with maternal and infant thyroid homeostasis: a multipollutant assessment. Environ Health Perspect 2017; 71:1002–1007. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Fei C, McLaughlin JK, Lipworth L, et al. Prenatal exposure to perfluorooctanoate (PFOA) and perfluorooctanesulfonate (PFOS) and maternally reported developmental milestones in infancy. Environ Health Perspect 2008;116(10):1391–1395. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Forns J, Iszatt N, White RA, et al. Perfluoroalkyl substances measured in breast milk and child neuropsychological development in a Norwegian birth cohort study. Environ Int 2015;83:176–182. [DOI] [PubMed] [Google Scholar]
- 11.Fei C, Olsen J. Prenatal exposure to perfluorinated chemicals and behavioral or coordination problems at age 7 years. Environ Health Perspect 2011;119(4): 573–578. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Oulhote Y, Steuerwald U, Debes F, et al. Behavioral difficulties in 7-year old children in relation to developmental exposure to perfluorinated alkyl substances. Environ Int 2016;97:237–245. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Liew Z, Ritz B, von Ehrenstein OS, et al. Attention deficit/hyperactivity disorder and childhood autism in association with prenatal exposure to perfluoroalkyl substances: a nested case-control study in the Danish National Birth Cohort. Environ Health Perspect 2015;123(4):367–373. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Ode A, Kallen K, Gustafsson P, et al. Fetal exposure to perfluorinated compounds and attention deficit hyperactivity disorder in childhood. PloS one. 2014;9(4):e95891. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Chen MH, Ha EH, Liao HF, et al. Perfluorinated compound levels in cord blood and neurodevelopment at 2 years of age. Epidemiology. 2013;24(6):800–808. [DOI] [PubMed] [Google Scholar]
- 16.Hoyer BB, Ramlau-Hansen CH, Obel C, et al. Pregnancy serum concentrations of perfluorinated alkyl substances and offspring behaviour and motor development at age 5-9 years--a prospective study. Environ Health. 2015;14:2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Vuong AM, Yolton K, Webster GM, et al. Prenatal polybrominated diphenyl ether and perfluoroalkyl substance exposures and executive function in school-age children. Environ Res 2016;147:556–564. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Lewis JB, Lapp CA, Schafer TE, et al. 4-Hydroxytamoxifen-induced cytotoxicity and bisphenol A: competition for estrogen receptors in human breast cancer cell lines. In Vitro Cell Dev Biol –Animal. 2000;36(5):320–326. [DOI] [PubMed] [Google Scholar]
- 19.Zoeller RT, Bansal R, Parris C. Bisphenol-A, an environmental contaminant that acts as a thyroid hormone receptor antagonist in vitro, increases serum thyroxine, and alters RC3/neurogranin expression in the developing rat brain. Endocrinology. 2005;146(2):607–612. [DOI] [PubMed] [Google Scholar]
- 20.Moriyama K, Tagami T, Akamizu T, et al. Thyroid hormone action is disrupted by bisphenol A as an antagonist. J Clin Endocrinol Metab 2002;87(11):5185–5190. [DOI] [PubMed] [Google Scholar]
- 21.Braun JM, Yolton K, Dietrich KN, et al. Prenatal bisphenol A exposure and early childhood behavior. Environ Health Perspect 2009;117(12):1945–1952. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Braun JM, Kalkbrenner AE, Calafat AM, et al. Impact of Early-Life Bisphenol A Exposure on Behavior and Executive Function in Children. Pediatrics. 2011;128(5):873–882. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Roen EL, Wang Y, Calafat AM, et al. Bisphenol A exposure and behavioral problems among inner city children at 7-9 years of age. Environ Res 2015;142:739–745. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Perera F, Vishnevetsky J, Herbstman JB, et al. Prenatal bisphenol a exposure and child behavior in an innercity cohort. Environ Health Perspect 2012;120(8):1190–1194. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Perera F, Nolte EL, Wang Y, et al. Bisphenol A exposure and symptoms of anxiety and depression among inner city children at 10-12 years of age. Environ Res 2016;151:195–202. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Casas M, Forns J, Martinez D, et al. Exposure to bisphenol A during pregnancy and child neuropsychological development in the INMA-Sabadell cohort. Environ Res 2015;142:671–679. [DOI] [PubMed] [Google Scholar]
- 27.Ma W, Kannan K, Wu Q, et al. Analysis of polyfluoroalkyl substances and bisphenol A in dried blood spots by liquid chromatography tandem mass spectrometry. Anal Bioanal Chem 2013;405(12):4127–4138. [DOI] [PubMed] [Google Scholar]
- 28.Buck Louis GM, Hediger ML, Bell EM, et al. Methodology for establishing a population-based birth cohort focusing on couple fertility and children’s development, the Upstate KIDS Study. Paediatr Perinat Epidemiol 2014;28(3):191–202. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Yeung EH, Sundaram R, Bell EM, Druschel C, Kus C, Ghassabian A, et al. Examining Infertility Treatment and Early Childhood Development in the Upstate KIDS Study. JAMA Pediatr 2016;170(3):251–258. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Yeung EH, Louis GB, Lawrence D, Kannan K, McLain AC, Caggana M, et al. Eliciting parental support for the use of newborn blood spots for pediatric research. BMC Medical Research Methodology. 2016;16:14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Reagen WK, Ellefson ME, Kannan K, Giesy JP. Comparison of extraction and quantification methods of perfluorinated compounds in human plasma, serum, and whole blood. analytica chimica acta. 2008;628(2):214–221. [DOI] [PubMed] [Google Scholar]
- 32.Spliethoff HM, Tao L, Shaver SM, Aldous KM, Pass KA, Kannan K, et al. Use of newborn screening program blood spots for exposure assessment: declining levels of perfluorinated compounds in New York State infants. Environmental science & technology. 2008;42(14):5361–5367. [DOI] [PubMed] [Google Scholar]
- 33.Kato K, Wanigatunga AA, Needham LL, Calafat AM. Analysis of blood spots for polyfluoroalkyl chemicals. Analytica chimica acta. 2009;656(1-2):51–55. [DOI] [PubMed] [Google Scholar]
- 34.Goodman R Psychometric properties of the strengths and difficulties questionnaire. Journal of the American Academy of Child and Adolescent Psychiatry. 2001;40(11):1337–1345. [DOI] [PubMed] [Google Scholar]
- 35.Stergiakouli E, Thapar A, Davey Smith G. Association of Acetaminophen Use During Pregnancy With Behavioral Problems in Childhood: Evidence Against Confounding. JAMA pediatrics. 2016;170(10):964–970. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Stergiakouli E, Davey Smith G, Martin J, Skuse DH, Viechtbauer W, Ring SM, et al. Shared genetic influences between dimensional ASD and ADHD symptoms during child and adolescent development. 2017;8:18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.He J-P, Burstein M, Schmitz A, Merikangas KR. The Strengths and Difficulties Questionnaire (SDQ): the Factor Structure and Scale Validation in U.S. Adolescents. Journal of abnormal child psychology. 2013;41(4):583–595. [DOI] [PubMed] [Google Scholar]
- 38.Cuffe SP, Moore CG, McKeown RE. Prevalence and correlates of ADHD symptoms in the national health interview survey. Journal of attention disorders. 2005;9(2):392–401. [DOI] [PubMed] [Google Scholar]
- 39.Ghassabian A, Sundaram R, Bell E, Bello SC, Kus C, Yeung E. et al. Gross motor milestones and subsequent development. Pediatrics. 2016;138(1). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Quaak I, de Cock M, de Boer M, Lamoree M, Leonards P, van de Bor M. et al. Prenatal Exposure to Perfluoroalkyl Substances and Behavioral Development in Children. International journal of environmental research and public health. 2016;13(5). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Brantsæter AL, Whitworth KW, Ydersbond TA, Haug LS, Haugen M, Knutsen HK, et al. Determinants of plasma concentrations of perfluoroalkyl substances in pregnant Norwegian women. Environment International. 2013;54:74–84. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Cho CR, Lam NH, Cho BM, Kannan K, Cho HS. Concentration and correlations of perfluoroalkyl substances in whole blood among subjects from three different geographical areas in Korea. Science of The Total Environment. 2015;512-513:397–405. [DOI] [PubMed] [Google Scholar]
- 43.Kristensen SL, Ramlau-Hansen CH, Ernst E, Olsen SF, Bonde JP, Vested A, et al. Long-term effects of prenatal exposure to perfluoroalkyl substances on female reproduction. Human Reproduction. 2013;28(12):3337–3348. [DOI] [PubMed] [Google Scholar]
- 44.Carwile JL, Michels KB. Urinary bisphenol A and obesity: NHANES 2003–2006. Environmental research. 2011;111(6):825–830. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Yeung EH, Sundaram R, Ghassabian A, Xie Y, Buck Louis G. Parental Obesity and Early Childhood Development. Pediatrics. 2017;139(2). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Rappazzo KM, Coffman E, Hines EP. Exposure to Perfluorinated Alkyl Substances and Health Outcomes in Children: A Systematic Review of the Epidemiologic Literature. International journal of environmental research and public health. 2017;14(7). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Kunacheva C, Fujii S, Tanaka S, Seneviratne ST, Lien NP, Nozoe M, et al. Worldwide surveys of perfluorooctane sulfonate (PFOS) and perfluorooctanoic acid (PFOA) in water environment in recent years. Water science and technology : a journal of the International Association on Water Pollution Research. 2012;66(12):2764–2771. [DOI] [PubMed] [Google Scholar]
- 48.Parker JL. Lessons from the New York Waterfront: Addressing the Risks of Drinking Water Contamination from Unregulated Contaminants. Environmental Claims Journal. 2017;29(3):242–252. [Google Scholar]
- 49.Gump BB, Wu Q, Dumas AK, Kannan K. Perfluorochemical (PFC) exposure in children: associations with impaired response inhibition. Environmental science & technology. 2011;45(19):8151–8159. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50.Goodman A, Lamping DL, Ploubidis GB. When to use broader internalising and externalising subscales instead of the hypothesised five subscales on the Strengths and Difficulties Questionnaire (SDQ): data from British parents, teachers and children. Journal of abnormal child psychology. 2010;38(8):1179–1191. [DOI] [PubMed] [Google Scholar]
- 51.Slotkin TA, MacKillop EA, Melnick RL, Thayer KA, Seidler FJ. Developmental neurotoxicity of perfluorinated chemicals modeled in vitro. Environmental health perspectives. 2008;116(6):716–722. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.