Concentrations of per- and polyfluoroalkyl substances (PFAS) in human placental tissues and associations with birth outcomes

Abstract

Per- and polyfluoroalkyl substances (PFAS) are ubiquitous environmental contaminants commonly detected in human serum. Previous studies have observed associations between maternal serum PFAS and adverse pregnancy and birth outcomes such as lower birth weight or pre-eclampsia; however, few studies have explored these associations with birth outcomes and placental tissue PFAS concentration. The placenta is a vital contributor to a healthy pregnancy and may be involved in the mechanism of PFAS reproductive toxicity. Our goal was to measure placental PFAS concentrations and examine associations with birth outcomes (e.g., birth weight, gestational duration). Placenta samples (n=120) were collected during delivery from women enrolled in the Healthy Pregnancy, Healthy Baby cohort (HPHB) in Durham, North Carolina. All placenta samples contained detectable PFAS, with perfluorooctane sulfonic acid (PFOS), perfluorooctanoic acid (PFOA), perfluorononanoic acid (PFNA), and perfluorodecanoic acid (PFDA) being the most abundant and most frequently detected (all >96% detection frequency). While placental PFAS concentrations did not differ by infant sex, higher PFAS levels were observed in placenta from nulliparous women, suggesting that parity influences the accumulation of PFAS in the placenta. We used linear regression models to examine associations between placental PFAS and birth outcomes. After adjustment for parity, tobacco use, maternal age, and maternal race, we found that placental PFOS was associated with lower birth weight for gestational age in male infants and higher birth weight for gestational age in female infants. Similar findings were seen for PFNA for birth weight for gestational age. These differences in birth outcomes based on infant sex highlight a need to explore mechanistic differences in PFAS toxicity during gestation for male and female infants.

Graphical Abstract

1. Introduction

PFAS, or per- and polyfluoroalkyl substances, are a large group of chemical contaminants that can be found ubiquitously in the environment and in human bodies. Their occurrence is so frequent in part because PFAS are widely used for their water- and grease-repellant properties in a range of products such as non-stick cookware, textiles, carpeting, food packaging, fire-fighting foam, and consumer products (Wang et al., 2017). In addition to being used in many products, many PFAS chemicals do not break down easily due to the strength of their multiple carbon-fluorine bonds (Lau et al., 2007).

PFAS are frequently detected in human serum and urine, and over 95% of the U.S. population has PFAS in their blood according to data collected by the National Health and Nutrition Examination Survey (NHANES) within the Centers for Disease Control and Prevention (CDC) (Calafat et al., 2007a; Calafat et al., 2007b). PFAS can be persistent in the human body due to the long elimination half-lives of some PFAS, particularly longer-chain PFAS (Olsen et al., 2007; Bartell et al., 2010; Zhang et al., 2013a; Li et al., 2018). PFAS have been found in human serum (Calafat et al., 2006; Calafat et al., 2019), breast milk (Macheka-Tendenguwo et al., 2018; Jin et al., 2020), and umbilical cord blood (Chen et al., 2017; Mamsen et al., 2017; Wang et al., 2020). These findings are particularly concerning as they suggest the potential for placental and lactational transfer of PFAS to developing fetuses and infants; some PFAS are known to have developmental and reproductive toxicity (Washino et al., 2009; Bach et al., 2015; Chang et al., 2018; Blake and Fenton, 2020).

These contaminants are multi-system toxicants with effects on the liver, kidney, immune system, thyroid, cholesterol, and infant birthweight (ATSDR, 2021). Epidemiological studies on PFAS have observed significant associations with adverse health outcomes, including cancer, immune impairment, and thyroid disease, among others (Lau et al., 2007; Grandjean et al., 2012; Barry et al., 2013; NTP, 2016; Rappazzo et al., 2017; Pelch et al., 2019; Sunderland et al., 2019; Fenton et al., 2021). These observations in humans have been further reinforced by data from animal studies (Lau et al., 2007; Lilienthal et al., 2017; ATSDR, 2021; Fenton et al., 2021).

While the mechanisms of PFAS toxicity are still being elucidated, the placenta has come forward as a potential target organ of PFAS toxicity (Blake and Fenton, 2020). Placental effects have been observed after exposure to PFOA and GenX in mice (Blake et al., 2020), and transplacental transfer of PFAS has been found experimentally in mice (Fenton et al., 2009) and observationally in humans (Yang et al., 2016a; Yang et al., 2016b; Chen et al., 2017; Mamsen et al., 2017; Mamsen et al., 2019). The placenta is a vital contributor to a healthy pregnancy as it mediates the exchange of nutrients, hormones, and other factors between mother and fetus. PFAS effects on the placenta may play a role in the mechanism of PFAS reproductive toxicity (Gao et al., 2019; Bangma et al., 2020a; Blake et al., 2020; Szilagyi et al., 2020a; Szilagyi et al., 2020b; Lu et al., 2021). For example, the placenta is a very important source of reproductive hormones, and prenatal exposure to short-chain PFAS such as PFBA and PFHpA has been associated with perturbations in fetal reproductive hormones (Nian et al., 2020). A few prior studies have reported PFAS concentrations in placenta (Zhang et al., 2013b; Martin et al., 2016; Chen et al., 2017; Mamsen et al., 2017; Mamsen et al., 2019; Bangma et al., 2020b; Lu et al., 2021; Vela-Soria et al., 2021). However, the potential health implications of PFAS accumulation in the placenta remain largely unknown. To our knowledge, only one study (Bangma et al., 2020b) has examined associations between placental PFAS with health or birth outcomes in a high-risk cohort. In this study, researchers investigated whether PFAS placental levels were associated with gestational age at delivery, fetal growth, or hypertensive disorders of pregnancy; they did not find evidence for any associations.

In addition to being a potential target of toxicity, the placenta may also be a better measure of fetal PFAS exposure than maternal measures. While maternal serum PFAS is frequently used in studies focusing on birth outcomes, maternal serum PFAS concentrations during pregnancy can be impacted by other changes that occur during pregnancy. For example as pregnancy progresses, increases in blood volume, plasma volume, and glomerular filtration rate (GFR) can occur (Peck and Arias, 1979; Odutayo and Hladunewich, 2012). These changes may impact serum PFAS concentrations. The extent of change in blood volume can be highly variable, and normal pregnancies can show increases in blood volume of 20–100% of nonpregnant blood volume (Pritchard, 1965). It is unclear how these changes in blood volume and GFR affect the partitioning of contaminants between maternal serum and placental tissues.

The goals of the present study were to evaluate the concentration of several PFAS in placenta collected from women living in central North Carolina in 2010–2011 and to further explore the associations between placental PFAS and birth outcomes.

2. Materials and Methods

2.1. Study population and sample collection

Placental tissues were obtained from participants in the Healthy Pregnancy, Healthy Baby (HPHB) Study. This prospective cohort study aimed to explore racial disparities in pregnancy outcomes due to social, environmental, and maternal factors (Miranda et al., 2010; Miranda et al., 2011; Stapleton et al., 2011; Swamy et al., 2011; Maxson et al., 2012; Buttke et al., 2013; Johnston et al., 2014; Sanders et al., 2014; Edwards et al., 2015; Gona et al., 2015; Miranda et al., 2015; Leonetti et al., 2016a; Maxson et al., 2016; Grotegut et al., 2017). This study was carried out in accordance with a human subjects research protocol approved by the Duke University Medical Center Institutional Review Board (IRB), and all women provided informed consent prior to participation. In brief, pregnant women were enrolled from the Duke University Medical Center (DUMC) Obstetrics Clinic and the Durham County Health Department Prenatal Clinic. Women were excluded from participation if they were younger than 18 years old, were not English-literate, were less than 18 or greater than 28 weeks gestation at time of enrollment, lived outside of Durham County, North Carolina, had a multi-fetal gestation, had a known fetal genetic or congenital abnormality, or were not planning to deliver at DUMC (Miranda et al., 2010). This population has previously been studied while investigating associations between brominated flame retardants and thyroid hormone levels in the placenta (Leonetti et al., 2016a; Leonetti et al., 2016b).

Our analyses include a subset of women from the HPHB study who delivered between March 2010 and December 2011 and from whom sufficient placental tissue was available (n = 120). Placenta tissue subsamples (approximately 5–20 g) were taken at time of delivery at the Duke University Medical Center, and tissues were stored in Nalgene™ polypropylene screw-top cryo-vials at −80 °C until analysis.

2.2. Sample extraction and purification

Extraction of PFAS from placenta was adapted from several methods previously described in the literature (Taniyasu et al., 2005; Liu et al., 2015; Martin et al., 2016). Whole-thickness placenta tissue aliquots (~1.0 g wet weight) were thawed, lightly rinsed with ultra-pure water to remove blood, and minced with dissecting scissors in 50-mL polypropylene centrifuge tubes. Tissue was spiked with ¹³C-mass-labeled internal standards (Wellington Laboratories, Guelph, Ontario, Canada), acidified with formic acid, and extracted with acetonitrile. Extracts (30 mL acetonitrile) were concentrated under nitrogen in an N-EVAP on low heat (Organomation) to ~1.0 mL. Extracts were then purified using solid-phase extraction with weak-anion-exchange cartridges (Waters Oasis WAX) and eluted in 10 mL of methanol with 0.1% ammonium hydroxide in 15-mL polypropylene centrifuge tubes. Extracts were concentrated under nitrogen again to ~300 μL, spiked with 60 μL of 1 mM sodium hydroxide, and stored at −20 °C. Prior to LC-MS/MS analysis, 500 μL of 2 mM ammonium acetate in water was added to each extract, and samples were transferred to 1-mL polypropylene push-top LC vials (Agilent).

2.3. HPLC-MS/MS analysis

Eleven PFAS (Table S1) were quantified in each extract using an Agilent 1260 Infinity II high-performance liquid chromatograph (HPLC) instrument coupled to an Agilent 6460A triple quadrupole mass spectrometer. The mass spectrometer was operated in negative electrospray ionization mode (HPLC-ESI-MS/MS). Separation of analytes by LC was performed using a 4.6 mm (I.D.) × 50 mm Agilent ZORBAX Eclipse XDB-C18 reversed-phase HPLC column (1.8 μm particle size) preceded by a 4.6 mm × 5 mm XDB-C18 guard cartridge.

Mobile phases were 2 mM ammonium acetate in water (mobile phase A) and 2 mM ammonium acetate in methanol (mobile phase B) using a flow rate of 0.4 mL/min. Gradient conditions for chromatographic separation were as follows: initial condition (30% B) was increased to 60% B over 1.5 minutes; then increased to 95% B over 2 minutes and held for 5.5 minutes. The gradient was then increased to 100% B over 3 minutes, before finally returning to initial conditions (30% B) over 0.5 minutes, and holding for 5.5 minutes. The column temperature was set at 45°C and the injection volume was 20 μL. Data were acquired under multiple reaction monitoring (MRM) transitions using optimized parameters. Additional methods information, including transitions, is included in Tables S2, S3, and S4.

2.4. Quality control and quality assurance (QA/QC)

Laboratory processing blanks were included in each batch and extracted alongside placenta tissue samples. A Lake Michigan fish tissue Standard Reference Material (SRM 1947) (NIST, 2017) of approximately 2.0 g was analyzed for PFAS alongside samples to examine accuracy. In total we included 7 lab processing blanks, 7 SRM extracts, and 12 duplicate samples (to assess precision) for our QA/QC assessment. Isotopically-labeled internal standards (¹³C-mass-labeled, listed in Table S2) were used in all samples. Recovery of ¹³C-labeled PFAS was calculated to assess the recovery efficiency of the extraction and clean-up methods; recovery averaged 133% for M3-GenX and ranged between 56–82% for all other mass-labeled standards as listed in Table S5. Our SRM values were extremely similar to the reference values for SRM 1947 as reported in Table S6.

2.5. Statistical analysis

Method detection limits (MDLs) for each analyte were calculated as three times the standard deviation of the blank values. Concentrations of each PFAS in samples were blank-corrected using the average laboratory blank levels. Only PFAS that were detected in more than 50% of placenta samples (PFOS, PFOA, PFDA, PFNA) were included in subsequent statistical analyses. Values less than the MDL were imputed with MDL divided by 2. Statistical analyses were performed using JMP Pro (version 14.0.0) and GraphPad Prism (version 9.0.2). The following data were abstracted from medical records: gestational age at delivery, birth weight, infant sex, maternal age, parity, maternal tobacco use during pregnancy, and maternal race. Birth weight was normalized for gestational age using the INTERGROWTH-21^st standards and is reported as a percentile (Stirnemann et al., 2017). Preliminary analyses using the Shapiro-Wilk test indicated the PFAS concentration data were not normally distributed, so non-parametric tests were used for analyses. Spearman rank correlation coefficients were calculated to evaluate the associations between individual placental PFAS concentrations. To evaluate whether placental PFAS differed by demographic characteristics, the two-tailed Mann-Whitney test was performed.

Linear regression models were conducted to determine if placental PFAS and other factors were associated with continuous measures of birth outcomes, including birth weight, gestational age, and birth weight for gestational age. For these models, placental PFAS concentrations were categorized into tertiles. Each tertile included n=40 pregnancies. Pregnancies were grouped into tertile solely based on placental PFAS concentration, without regard for infant sex; previous analyses in Figure S1 showed no difference in placental PFAS by infant sex and the distribution of placental PFAS concentration data by infant sex was fairly even. Tertiles were approximately 55% female and 45% male, mirroring the sex distribution in the total study population. Gestational age in days, birth weight in grams, and birth weight for gestational age in percentile were treated as continuous variables for regression analyses.

The potential for demographic and behavioral factors to confound associations between PFAS and birth outcomes was also considered. We identified possible confounding variables based on a literature review and selected variables for inclusion in adjusted analysis if they were associated with both placental PFAS and birth outcomes.

The final adjustment set included maternal race (non-Hispanic black vs. other); parity: (nulliparous (first birth) vs. multiparous (second or later birth)); maternal smoking status (any tobacco use during pregnancy vs. no use during pregnancy) and maternal age at gestation (dichotomized at the median of 27 years for analysis). Maternal education was considered for inclusion in fully adjusted models, but it was not associated with any of the outcomes in this population and did not appreciably change models when included.

Previous work suggests sex differences in associations with PFAS and birth outcomes. Thus, analyses were stratified by infant sex to explore potential sex differences in birth outcomes. The threshold for statistical significance was p<0.05 for all analyses. Main analyses were performed on our full study population (n=120), including preterm births. However, to evaluate whether results were driven by the inclusion of premature babies, analyses were repeated excluding babies born before 37 complete weeks gestation (n=12); results were very similar in direction and magnitude (Supplemental Excel File); our conclusions and interpretations are unchanged by the inclusion or exclusion of preterm births.

3. Results and Discussion

3.1. Demographic information

Placenta samples from 120 participants were included in this study. Table 1 summarizes the demographic characteristics of the women and children in this study cohort.

Table 1:

Demographic characteristics of cohort (n = 120)

Characteristic	N	%
Total	120	100.0
Maternal age at gestation
18–19	11	9.2
20–26	44	36.7
27–34	46	38.3
35–39	14	11.7
40–46	5	4.2
Maternal race
Non-Hispanic black	72	60.0
Non-Hispanic white	26	21.7
Hispanic	12	10.0
Other	10	8.3
Male infant	54	45.0
Gestational age <37 weeks	12	10.0
Tobacco use during pregnancy	26	21.7
Parity
Nulliparous	38	31.7
Multiparous	82	68.3

Of the 120 participants, 60% were non-Hispanic black, 22% were non-Hispanic white, 10% were Hispanic, and 8% were either non-Hispanic Asian, other, or multi-racial. Approximately 54% of the women were 27 years of age or older, and approximately 78% of women reported no tobacco use during pregnancy.

3.2. PFAS in placenta

At least one PFAS analyte was detected in all placental samples (n=120) as reported in Table 2. The most frequently detected PFAS were PFOS, PFOA, PFNA, and PFDA, which were detected in more than 95% of placental samples. As displayed in Figure 1, PFOS and PFOA were the most abundant analytes with median concentrations of 0.95 and 0.27 ng/g, respectively. The concentrations of individual PFAS in placenta were correlated with the concentrations of other frequently-detected PFAS, with Spearman correlations ranging from 0.35 to 0.70, p<0.05 for all (Figure 2).

Table 2:

Detection frequency, method detection limit (MDL), minimum, median, 95^th percentile, and maximum concentrations of PFAS in ng/g wet weight in placental tissues (n = 120).

Analyte	Detection Frequency	MDL (ng/g)	Minimum (ng/g)	Median (ng/g)	95th percentile (ng/g)	Maximum (ng/g)
PFOS	99%	0.01	<MDL	0.95	2.5	7.2
PFOA	98%	0.02	<MDL	0.27	0.7	1.6
PFNA	100%	0.01	0.03	0.11	0.3	0.6
PFDA	96%	0.01	<MDL	0.06	0.2	0.3
PFBA	12%	2.36	<MDL	<MDL	19.2	29.4
PFPeA	0%	1.06	<MDL	<MDL	<MDL	<MDL
PFHxA	7%	0.03	<MDL	<MDL	0.1	0.1
PFHpA	0%	1.50	<MDL	<MDL	<MDL	<MDL
PFBS	4%	8.10	<MDL	<MDL	<MDL	10.3
PFHxS	19%	0.17	<MDL	<MDL	0.5	0.5
GenX	0%	0.06	<MDL	<MDL	<MDL	<MDL

Figure 1:

PFAS concentrations (ng/g) in placenta (n = 120) for analytes detected in more than 95% of samples. Violin plots show the distribution of concentration data with solid lines and dotted lines demarcating median and quartiles, respectively. Concentrations are plotted on a logarithmic scale.

Figure 2:

Spearman correlation coefficients for analytes detected in more than 95% of placental samples (n = 120). Color indicates strength and directionality of correlation. *indicates p<0.05

Placental PFAS concentrations for PFOS, PFOA, PFNA, and PFDA were analyzed to determine if concentrations differed based on infant sex, but no statistically significant differences were found. Figure S1 displays the concentrations of these four PFAS stratified by infant sex. However, placental PFOA, PFNA, and PFDA were significantly higher in placenta from nulliparous women compared to multiparous women; nulliparous women had 20–40% higher median concentrations (Figure 3). This is not a surprising result given the potential for transplacental and lactational transfer of these chemicals to reduce the PFAS body burden in the mother. Previous studies have found that PFAS concentration in the placenta may decrease with parity as lactational and placental transfer offloads the body burden of PFAS from the mother to the infant (Fei et al., 2007; Kim et al., 2011; Lee et al., 2013). For example, higher PFOA concentrations have been reported in nulliparous women compared to multiparous women (Lee et al., 2013), and higher parity was associated with lower serum levels for PFOA and seven other PFAS (Shu et al., 2018).

Figure 3:

Parity and placental PFAS concentrations. Placenta samples from nulliparous (n = 38) and multiparous (n = 82) women were compared. PFOA, PFNA, and PFDA placental concentrations were significantly higher in nulliparous women. Violin plots show the distribution of concentration data with thick solid lines and thin solid lines demarcating median and quartiles, respectively. *indicates p<0.05, and “ns” indicates not significant.

We also examined differences in placental PFAS concentrations by maternal race. Previous studies have observed differences in PFAS serum concentrations based on race, with non-Hispanic whites having higher serum concentrations of PFOS, PFOA, and PFHxS than non-Hispanic blacks and Mexican Americans (Calafat et al., 2006). In our study, we found no significant differences in placental PFAS by maternal race (Figure S2). However, it is important to note that the racial and ethnic breakdown of our cohort necessitated combining non-Hispanic white, Hispanic, and other races or ethnicities into a single category for comparison with non-Hispanic black women. It is likely this limited our ability to detect differences between groups.

We also analyzed a small group of anonymous placenta samples (n=10) collected more recently in 2018 for PFAS. Methods for the collection, storage, and processing of these tissues from 2018 are included in Supporting Information. Previous research in our lab has observed higher concentrations of brominated flame retardants in the fetal side of the placenta compared to the maternal side (Ruis et al., 2019). To investigate whether PFAS would also partition preferentially to one side of the placenta, we sectioned these ten placenta samples into the maternal and fetal sides, using the same method reported in Ruis et al. (2019), and analyzed them for PFAS. No significant differences were observed between the maternal and fetal placenta concentrations of PFAS. These ten placenta samples were observed to have slightly lower PFAS concentrations than the 120 placenta samples collected in 2010–2011, though this comparison is limited due to the small sample size (Table S7).

3.3. Placental PFAS concentrations and birth outcomes

Placental PFAS concentrations were analyzed for associations with birth outcomes using linear regression models and adjusted for maternal race, maternal age at gestation, parity, and tobacco use. For infant males, the highest exposure to PFOS was associated with lower birth weight for gestational age; the highest exposure exhibited a 13% decrease (95% CI: −23, −1.6) in birth weight percentile in comparison to the reference (lowest exposure) tertile (Figure 4). To put that difference into context, for a baby born at 40 weeks gestation, a 13% decrease in percentile is equivalent to approximately a 130-gram decrease in birth weight. Conversely, for infant females, the highest exposure of PFOS was associated with higher birth weight for gestational age; the highest exposure group exhibited an 11% increase (95% CI: 2.8, 19) in birth weight percentile in comparison to the reference tertile. Placental PFNA had a complex association with birth weight for gestational age; while modeling results for female infants were null and showed no difference, male infants with the highest exposure had lower weight percentiles (14% percentile decrease; 95% CI: −24, −3.9) and male infants within the middle tertile of exposure had higher weight percentiles (11% percentile increase; 95% CI: 0.2, 21) in comparison to the reference tertile.

Figure 4:

Results for regression models of birth weight for gestational age stratified by male and female infants. Analyses were performed using tertiles of placenta PFAS exposure with the lowest tertile as the reference group and were adjusted for maternal tobacco use, race, age, and parity. Horizontal bars reflect the 95% confidence interval (CI), *indicates p<0.05.

For models of gestational age, a significant association was observed in male infants exposed to the highest placental PFDA (Figure S3). Placental PFDA in males was associated with a shorter gestational age; male babies with the highest levels of exposure were born approximately ten days earlier on average. All other models of gestational age were null, suggesting no association between placental PFAS and the timing of parturition. Analyses of birth weight in grams, not normalized to gestational age, are included in (Figure S4), and a significant association was observed only in female infants exposed to the highest placental PFOS.

A sensitivity analysis was performed to examine the influence of preterm births on the model outcomes. This restricted model excluded the samples from preterm births (gestational age less than 37 complete weeks, n=108). Analyses from the restricted model were very similar in direction and magnitude as results from the full model and conclusions were not altered; additional details are included in Supplemental Information.

Many studies have explored associations between certain PFAS in maternal blood and birth weight or gestational age. Low birth weight has been associated with maternal serum PFOA and PFOS in previous studies (Fei et al., 2007; Darrow et al., 2013; Johnson et al., 2014), particularly for male infants (Marks et al., 2019). Manzano-Salgado et al. (2017) reported weak associations between maternal plasma PFOA, PFHxS, and PFNA and reduced birth weight in a Spanish birth cohort with samples collected from 2003–2008. Of note, they also report that higher PFOS in first-trimester maternal plasma (n=1,202) was associated with low birth weight in boys, similar to the results reported here (Manzano-Salgado et al., 2017).

In a study comparing placental PFAS with maternal serum and fetal tissue PFAS, Mamsen et al. (2019) noted that the placenta:maternal serum ratio was higher in pregnancies with male fetuses compared to those with female fetuses. While we did not see differences in placental concentrations between male and female pregnancies, we also did not have maternal serum available to compare with this study or to determine the placenta:maternal serum ratio. Additionally, there are several reported sex differences in placental epigenetics that may also affect infant birth weight (Martin et al., 2017; Clark et al., 2021). This sexually-dimorphic epigenetic placental signature may explain the differential response to gestational PFAS exposure.

Impacts on birth weight and gestational age are not the only potential deleterious effects of PFAS exposure. Maternal PFAS exposure has been associated with adverse reproductive outcomes such as miscarriage (Wikström et al., 2021) and earlier menopause (Ding et al., 2020), though there is evidence of reverse causation with regards to PFAS and menopause (Dhingra et al., 2017). Maternal PFAS exposure could also lead to disruptions in inflammatory pathways that impact pregnancy outcomes and reproductive health (Liu et al., 2020).

3.4. Comparisons with other placenta studies

Results from this study were compared with placenta PFAS concentrations reported in previous studies and are summarized in Table S8 and Figure S5. Overall, our PFAS concentration data are very similar to previously reported data for placental concentrations. Mamsen et al. (2017) reports PFAS concentrations in placental and fetal tissues from fetuses terminated in the first trimester of pregnancy. A second study by this group, reported in Mamsen et al. (2019), expands on their research by including tissues from the second and third trimester of pregnancy from fetuses that died in utero. Placental PFAS concentrations in their study were similar to fetal organ concentrations, and the placenta:maternal serum ratio for PFOS, PFOA, and PFNA increased in later trimesters, suggesting placental bioaccumulation or hemodilution due to plasma volume increase (Mamsen et al., 2019). Given the similarity in PFAS concentrations between placenta and fetal organs seen in Mamsen et al. (2019), it is possible that our placenta data would be similar to the tissue concentration in the infants. However, Chen et al. (2017) measured PFAS in maternal serum, placenta, and cord serum from maternal-fetal pairs, and they found that PFAS concentrations in maternal serum and cord serum were higher than in placenta.

Zhang et al. (2013b) measured PFAS in maternal blood, placenta, cord blood, and amniotic fluid. They noted that maternal transfer efficiency (moving from maternal blood to cord blood) decreased with increasing PFCA chain length. Zhang et al. (2013b) found that shorter-carbon chain PFAS (i.e., shorter than PFDA) partitioned more in the cord blood and maternal blood than in the placenta, possibly due to greater water solubility. In our study, longer-chain PFAS were abundantly detected in placenta, but shorter-chain PFAS were detected much less frequently. However, this absence of shorter-chain PFAS may also be reflective of a lack of exposure to shorter-chain PFAS in our study. The placenta may also not be the best matrix for measuring certain PFAS. While the longer-chain PFAS were frequently detected in our placenta, shorter-chain PFAS were not detected or detected very rarely. As these placenta samples were collected in 2010–2011, this may be due to a lack of exposure to the shorter-chain PFAS. It may also be an indication that these other PFAS do not accumulate in the placenta as efficiently. We were unable to compare our placenta PFAS to matched maternal serum PFAS. Having both the placenta and maternal serum, or fetal or cord serum, would have been valuable in determining whether the absence of shorter-chain PFAS was due to a lack of exposure or due to a limitation of the placenta matrix.

Bangma et al. (2020b) measured PFAS in placenta from a geographically similar cohort as our study, with placental tissues collected more recently (2015–2018 compared to 2010–2011). While their study did not find any significant associations between placental PFAS concentrations and fetal growth, gestational age, or hypertensive disorders of pregnancy, this may be explained by the fact that their study cohort was comprised of women who were at increased risk for spontaneous preterm birth and included predominantly white women.

As seen in Table S8 and Figure S5, PFOS and PFOA are the most abundant PFAS measured in placental samples across the world. Over time, it appears that placental concentrations are slowly declining for PFOS, PFOA, PFNA, and PFDA. However, geographic location plays an important role in placenta PFAS concentration in addition to year of sample collection. Although both the present study and Zhang et al. (2013b) analyzed placenta samples collected in 2010–2011, our median and maximum concentrations were considerably lower. As Zhang et al. (2013b) collected placenta from healthy women at a hospital in Tianjin, China, the differences in placental PFAS exposure may be a result of differential use of consumer products and furnishings in each region (e.g., furniture, carpeting, stain-repellent).

3.5. Limitations

Limitations to our study include the fact that our study cohort consisted exclusively of women living in central North Carolina and may not be representative of other regions. Demographic differences in our cohort may also be a limitation; our cohort is unique as it consists primarily of non-Hispanic black women, an under-represented population in contaminant exposure studies. However, while this may limit the generalizability of our findings to other populations, we do not expect the demographic characteristics of the study population to limit the internal validity of our findings; the homogeneity of our cohort may have helped to reduce unmeasured confounding. Additionally, exposures to PFAS have been changing over time as legacy PFAS such as PFOA and PFOS are phased out of use and replaced with newer, emerging PFAS. Our study examined associations with 120 placenta samples collected in 2010–2011. The placenta concentrations observed in these samples may not be consistent with current PFAS exposure as can be evidenced by the changing placental concentrations observed in more recent studies (Table S8) and in our limited observation of ten placenta samples collected in 2018 (Table S7). In addition, our results indicate that evaluating sex-specific associations between PFAS in placenta and birth outcomes is critical. However, our relatively small sample size (n=120) may have limited our power to detect more subtle associations.

4. Conclusions

Our present study shows that several PFAS are frequently detected in placenta, and our observed associations with birth outcomes indicate the potential concern for adverse health effects on infants exposed to the highest tertiles of PFAS exposure. Our concentrations are similar in magnitude to the few other studies on placenta PFAS conducted across the world, highlighting the widespread reach of PFAS exposure. In our study PFAS placental exposure was associated with sex-specific birth outcomes, similar to other studies that analyzed maternal serum PFAS. We found that the highest exposure of placental PFOS was associated with lower birth weight for gestational age in infant males and higher birth weight in female infants.

Highlights.

• Several PFAS were detected in human placental tissues collected from 2010–2011 in the U.S.

• PFOS, PFOA, PFNA, and PFDA were most abundant in placenta

• Sex-specific differences were observed in birth outcome associations with PFAS

• Placental PFAS was higher in nulliparous pregnancies