The associations between maternal and fetal exposure to endocrine-disrupting chemicals and asymmetric fetal growth restriction: a prospective cohort study

Recent evidence has revealed associations between endocrine-disrupting chemicals (EDCs) and placental insufficiency due to altered placental growth, syncytialization, and trophoblast invasion. However, no epidemiologic study has reported associations between exposure to EDCs and asymmetric fetal growth restriction (FGR) caused by placenta insufficiency. The aim of this study was to evaluate the association between EDC exposure and asymmetric FGR. This was a prospective cohort study including women admitted for delivery to the Maternal Fetal Center at Seoul St. Mary’s Hospital between October 2021 and October 2022. Maternal urine and cord blood samples were collected, and the levels of bisphenol-A (BPA), monoethyl phthalates, and perfluorooctanoic acid in each specimen were analyzed. We investigated linear and non-linear associations between the levels of EDCs and fetal growth parameters, including the head circumference (HC)/abdominal circumference (AC) ratio as an asymmetric parameter. The levels of EDCs were compared between fetuses with and without asymmetric FGR. Of the EDCs, only the fetal levels of BPA showed a linear association with the HC/AC ratio after adjusting for confounding variables (β = 0.003, p < 0.05). When comparing the normal growth and asymmetric FGR groups, the asymmetric FGR group showed significantly higher maternal and fetal BPA levels compared to the normal growth group (maternal urine BPA, 3.99 μg/g creatinine vs. 1.71 μg/g creatinine [p < 0.05]; cord blood BPA, 1.96 μg/L vs. −0.86 μg/L [p < 0.05]). In conclusion, fetal exposure levels of BPA show linear associations with asymmetric fetal growth patterns. High maternal and fetal exposure to BPA might be associated with asymmetric FGR.

FGR is a common manifestation and clinically important in obstetrics as a cause of perinatal morbidities and mortalities (25)(26)(27).Although the etiologies of FGR vary, placenta insufficiency is a major one.Placenta-mediated FGR is characterized by an asymmetric pattern of fetal growth, and it shares a pathological process with hypertensive disorders of pregnancy (28,29).Asymmetric growth patterns are thought to occur when chronic hypoxia and malnutrition due to placental insufficiency lead to adaptation to chronic hypoxia.This adaptation involves a brain-sparing effect, where blood flow to the brain increases, resulting in an increase in head circumference (HC), while the abdominal circumference (AC) of the fetus becomes relatively small (30 -32).
There are various mechanisms proposed for how EDCs may affect fetal growth.Including alterations in hormone homeostasis, inflammation with oxidative stress, and epigenetic changes (33).EDCs may affect not only the fetus directly but also the size and function of the placenta (34,35).Early exposure to these chemicals has been linked to altered syncytialization, trophoblast invasion, and spiral artery remodeling, which can result in placental insufficiency and subsequently lead to FGR through chronic hypoxia and malnutrition (36-39).Based on the laboratory evidence, EDCs could induce FGR through placental insufficiency, and it could result in fetal asymmetric growth.
Despite this evidence, most epidemiological studies to date have focused on the associations between different EDCs and birth weight itself.In addition, information on fetal exposures is limited, and population-based studies on the association between fetal asymmetric growth and EDC exposures are lacking.We aimed to investigate the potential effects of maternal and fetal exposure to EDCs on the asymmetric pattern of FGR.

Study design and population
This analysis considered data from a prospective cohort study designed to examine fetal-maternal exposure to EDCs during the COVID-19 pandemic and the association between EDC exposure and obstetric complications.The cohort study enrolled 323 women admitted for delivery to the Maternal Fetal Center at Seoul St. Mary's Hospital between October 2021 and October 2022.We collected anthropometric measurements, socioeconomic status, medical history, and maternal urine samples after obtaining informed consent and cord blood samples after delivery.Ultimately, 146 mother-child pairs with singleton pregnancies with adequate amounts of maternal urine and cord blood samples for analyzing EDCs and adequate data of fetal growth at admission were included in the study.We evaluated the distributions; fetal-maternal transfer of EDCs; and the association between levels of EDCs in each specimen and fetal growth parameters, including asymmetric indices.In addition, the study population was divided into groups according to the presence of fetal growth restriction or asymmetricity, and the levels of EDCs were compared between the groups.This study was conducted after informed consent was obtained from all participants and was approved by the institutional review board (IRB) of the Catholic Medical Center in South Korea (IRB no.XC21ONDI0125).

Assessment of EDC exposure
Maternal urine samples were obtained after hospitalization for delivery, and cord blood was collected directly from the cord immediately after birth.Cord blood serum was fractioned by centrifugation at 3000 rpm for 10 min.The samples were stored at −80°C until analysis.We selected a set of environmentally disruptive chemicals for assessment, focusing on bisphenol-A (BPA), monoethyl phthalates (MEPs), and perfluorooctanoic acid (PFOA); it is believed that pregnant women are commonly exposed in daily lives to these chemicals, which have been extensively studied during the COVID-19 pandemic period (3,4) but there are controversial results in the relation between these EDC exposure and fetal growth (15-24).The levels of EDCs were analyzed with ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS).In detail, BPA and MEP were analyzed with Agilent 1,260 Infinity UPLC system (Agilent Technologies, Santa Clara, CA) and MS, the Agilent Triple Quadrupole 6,460 system with a specialized type of ESI interface.PFOA was analyzed using a different system, i.e., UPLC of Thermofisher Scientific Ultimate (Thermo Fisher Scientific, Waltham, MA). and Bruker EVOQ Qube LC-Triple Quadrupole (Bruker  (40).To adjust the urine dilutions, metabolite levels were divided by creatinine (Cr) levels, then analyzed using an automatic biomedical analyzer (HITACHI 7020; Hitachi, Ltd., Tokyo, Japan).Detailed methods and calibration curves for analyzing the EDCs are described in the Supplementary materials.

Definition of fetal growth restriction
The fetal growth parameters measured using transabdominal ultrasound by skilled physicians after hospitalization for delivery.The median gestational age at the time of ultrasound was 38.3 weeks of gestation.Fetal growth parameters include biparietal diameter (BPD), head circumference (HC), abdominal circumference (AC), femur length (FL), and estimated fetal weight (g).Since these parameters vary greatly depending on the gestational age, z-scores were calculated using the Intergrowth-21 chart (41,42).Asymmetric growth parameters were assessed using the HC/AC ratio (43,44), which was calculated as the ratio of HC (mm) to AC (mm).Fetal growth restriction was defined according to the Society for Maternal-Fetal Medicine guidelines as an AC measurement below the 10th percentile or an estimated fetal weight below the 10th percentile (45).Asymmetric growth was defined as an HC/AC ratio above the 95th percentile based on the criteria published by Campbell and Thoms (43).

Covariates
We collected maternal demographic characteristics and socioeconomic status indicators, including maternal age, parity, pre-pregnancy body mass index (BMI), pre-pregnancy smoking, pre-pregnancy alcohol consumption, pre-existing hypertension, pre-existing diabetes mellitus, gestational diabetes mellitus, and pregnancy-associated hypertension.Obstetric and delivery outcomes were acquired, including gestational age at delivery, cesarean section delivery, fetal sex, birth weight, and neonatal intensive care unit admission.We selected five confounding variables according to a ruleof-thumb of multivariable linear regression analysis (46).The adjusted confounders that might influence fetal growth parameters included maternal age, pre-pregnancy BMI, pre-pregnancy smoking, gestational age at ultrasound examination, and fetal sex.

Statistical analysis
Continuous data are presented as mean ± standard deviation values and categorical variables are presented as a number (%).The analyses from urine samples were conducted using Cr-adjusted levels of EDCs.To assess the distribution of environmental hormones, the geometric means (GMs) of BPA, MEP, and PFOA levels were calculated.For further statistical analyses, the Cr-adjusted levels of EDCs were transformed by Log2-transformation because they were not normally distributed.For the analysis of maternal to fetal exposure of EDCs, the correlation between the level of EDCs in maternal urine and those in fetal cord blood was assessed using a generalized additive model and linear regression model, and Pearson's correlation coefficient was calculated.
To investigate the association between levels of EDCs and fetal growth parameters, we first used generalized additive models to assess the linearity or non-linearity of the associations.We then used multivariable linear regression analysis or generalized additive model analysis to identify any significant relationships between EDC levels and fetal growth parameters, while adjusting for confounding variables.The regression coefficients from these analyses represented the difference in growth parameters per two-fold increase in EDC levels.To examine the association between EDC levels and fetal growth restriction (FGR), we divided the study population into FGR and non-FGR groups, then compared the concentrations of EDCs using the Mann-Whitney U test.The same methods were used to compare EDC levels between the FGR with asymmetry group and the non-FGR group.We conducted these analyses using IBM SPSS Statistics version 25.0 software (IBM Corp., Armonk, NY, USA) and R version 4.2.1 1 (R Foundation for Statistical Computing, Vienna, Austria).

Characteristics of the total study population
A total of 146 women were included in this study, and their baseline characteristics, obstetric history, and delivery outcomes are presented in Table 1.The median maternal age was 35 years, and the median BMI was 21.2 kg/m 2 .The median gestational age at delivery was 38.6 weeks, and most of the study population delivered after 37 weeks of gestation.The median birth weight was 3,118 g, and 92.5% of neonates had an adequate birth weight, ranging from 2,500 g to <4,000 g.

The distribution of BPA, MEP, and PFOA levels in maternal urine and cord blood
The distribution of BPA, MEP, and PFOA levels in maternal urine and cord blood is shown in Supplementary Table S1.BPA was detected in 93.2 and 91.1% of maternal urine and cord blood samples, respectively.The geometric mean for urine BPA was 1.220 μg/g creatinine, while that for cord blood was 0.751 μg/L.MEP was detected in 91.1 and 77.4% of maternal urine and cord blood samples, respectively.The geometric mean for urine MEP was 10.523 μg/g creatinine, while that for cord blood was 0.106 μg/L.PFOA was detected in 49.3 and 100% of maternal urine and cord blood samples, respectively.The geometric mean for urine PFOA was 0.026 μg/g creatinine, while that for cord blood was 2.315 μg/L.
The correlation of BPA, MEP, and PFOA levels in maternal urine and cord blood blood BPA was 0.22, while that between maternal urine MEP and fetal cord blood MEP was 0.19, indicating a statistically significant positive correlation (both p < 0.05).However, there was no significant correlation among PFOA levels between maternal urine and fetal cord blood (correlation coefficient, 0.05; p = 0.583).A linear association between maternal urine and cord blood samples in BPA and MEP was confirmed by both non-linear and linear association analyses using a generalized additive model and linear regression model, respectively (Supplementary Figure S1).

Association between EDC exposure and asymmetric fetal growth restriction
Figure 2 shows associations between the levels of EDCs and the HC/AC ratio as an asymmetric growth parameter.Among the EDCs, fetal cord blood BPA showed a positive linear association with the HC/AC ratio (effective degree of freedom, 1; R 2 = 0.082; p < 0.05).After analysis using a multivariable linear regression model, a positive linear association was still observed between cord blood BPA and the HC/AC ratio after adjusting for confounding variables (β = 0.003, p < 0.05) (Table 2).The results of comparing the EDC levels between groups are presented in Table 3.There were no significant differences in the levels of EDCs between the FGR and non-FGR groups.However, significant differences in BPA levels were observed between the asymmetric FGR and non-FGR groups.The asymmetric FGR group showed significantly higher maternal and fetal BPA levels compared to the control group (maternal urine BPA, 3.99 μg/g of Cr vs. 1.71 μg/g of Cr, p < 0.05; cord blood BPA, 1.96 μg/L vs. 0.86 μg/L, p < 0.05).Considering baseline characteristics between asymmetric FGR and non-FGR groups, there were no significant differences in other characteristics except for estimated fetal weight and birth weight (Supplementary Table S2).

Association between EDC exposure and fetal growth parameters
When examining the relationship between each growth parameter and EDC levels using generalized additive modeling, we identified significant linear associations between the maternal BPA level and HC and between the maternal MEP level and FL, respectively, while the fetal PFOA level showed a non-linear association with AC (Supplementary Figure S2, all p < 0.05) After multivariable linear regression analysis, positive linear associations remained between the maternal MEP level and FL z-score (β = 0.067, p < 0.05) (Supplementary Table S3).Fetal PFOA also showed a significant non-linear association with AC after adjusting for confounding factors (effective degree of freedom, 6.545; R 2 = 0.167; p < 0.05).Statistical values for the generalized additive model are presented in Supplementary Table S4.

Main findings
Our study was conducted to investigate the associations between asymmetric fetal growth restriction and EDC exposure.According to our findings, there was a statistically significant positive correlation between the asymmetric growth indicator HC/AC ratio and the fetal cord blood BPA level.Furthermore, in cases of asymmetric FGR, both  Associations of the levels of EDCs with HC/AC ratio using generalized additive model.The asterisks denote statistically significant associations.maternal urine and cord blood samples showed significantly higher BPA concentrations compared to those of the normal growth group.
Based on these results, it can be inferred that elevated BPA levels in fetuses, resulting from maternal BPA exposure, are associated with asymmetric FGR.

Interpretation
While numerous epidemiologic studies have suggested a negative impact of BPA on fetal growth (47)(48)(49)(50)(51)(52)(53)(54), limited research has been conducted on its relationship with asymmetric fetal growth restriction.Most studies have examined associations between BPA and birth weight itself (54), with only a few investigating postnatal asymmetric parameters, such as the ponderal index (17, 55-58).The ponderal index is a postnatal asymmetric growth parameter used to assess the body composition of infants that provides a convenient value derived from each infant's weight and height; to date, it has been studied in relation to environmental hormones, unlike prenatal indicators such as HC/ AC or Doppler indices.Some studies have reported no association between the ponderal index and BPA exposure (17, 55, 56) while others have reported an association between higher levels of BPA and increased ponderal index values (57,58).
In our study, we used the prenatal HC/AC ratio as an asymmetric growth indicator, reflecting the brain-sparing effect in growth restricted fetuses and reflecting the prenatal asymmetric growth patterns associated with adverse outcomes (31, 32, 59).Although there is currently a lack of research on the relationship between prenatal asymmetric growth indicators and EDCs, some studies have reported associations between BPA exposure during the fetal period and smaller AC and larger HC measurements (16,53,(60)(61)(62).For instance, Zhou et al. reported that higher BPA levels were associated with decreased AC and noted a sex-specific correlation between head circumference and BPA, where increased BPA levels were associated with larger head sizes in girls (16).They explained the observed differences in the effects of BPA on males and females by highlighting the influence of certain hormones and epigenetic mechanisms.A recent study by Uldbjerg et al. also reported a negative association between BPA level and AC (60).However, there are also numerous studies that do not show an association between BPA level and a smaller AC or larger HC, indicating the need to conduct further research in this area (53,61,62).
Asymmetric FGR is believed to be attributed to placental insufficiency and the brain-sparing effect in growth restricted fetuses (63,64).Several in vitro studies have shed light on the impact of BPA on placental insufficiency.It has been documented that BPA can induce alterations in the size and morphology of the placenta and exert a negative influence on spiral artery remodeling (36-38).Moreover, BPA has been shown to have detrimental effects on trophoblast invasion, syncytiotrophoblast differentiation, and syncytiotrophoblast zone size, which are critical for proper placental function (39,65).BPA has also been found to induce oxidative stress, a known contributor to placental insufficiency (66, 67).Additionally, EDCs like BPA can interfere with the expression of crucial nutrition transporters like amino-acid transporters, thereby disrupting the exchange of nutrients (68).Recent research has highlighted an association between BPA and the antiangiogenic factor sFlt-1/PlGF, which plays a role in placental insufficiency (69,70).Taken together, it is hypothesized that BPA's potential impact on the placenta may be a contributing factor to the observed asymmetric pattern of fetal growth restriction, as demonstrated in our study.
There exist several epidemiological studies suggesting that MEP and PFOA may contribute to fetal growth impairment (56,(71)(72)(73)(74)(75)(76).However, in our study, we did not observe an association between these two EDCs and fetal growth restriction.Nevertheless, MEP showed a positive association with FL, while PFOA exhibited a non-linear association with AC.These findings differ from those of some previous studies where MEP has been reported to be associated with smaller HCs (56,71,77).However, multiple studies have also reported no significant relationship between MEP concentrations and fetal growth parameters (53,60,61,70,78,79).Similarly, the relationship between PFOA and fetal growth parameters, such as HC and AC, has been investigated in several studies whose results were inconsistent (75,80,81).The discrepancies may be due to differences in study methodologies, such as different sampling times or types and diverse population groups.
We assessed the exposure levels of BPA, MEP, and PFOA during the COVID-19 pandemic.Despite the assumption that plastic usage would increase during the COVID-19 pandemic, leading to greater exposure to BPA and phthalates, our results showed similar or even lower levels of BPA, MEP, and PFOA compared to earlier studies conducted in South Korea (58, 82-86).It is speculated that, although plastic usage or exposure to environmental hazards may have increased, many products used by mothers are now labeled as BPA-free, and there have been environmental regulations on BPA, phthalates, and PFOA, which may have contributed to the decreased levels of exposure.In addition, we observed a correlation between BPA and MEP levels in both maternal samples and cord blood samples.This finding is consistent with existing knowledge that many environmental chemicals can pass through the placenta and contribute to exposure in utero (87-90).

Strengths and limitations
Our study holds significance as the first investigation to explore the association between asymmetric fetal growth restriction and EDC exposure.It was conducted prospectively, enabling the examination of EDC correlations between maternal urine and cord blood samples and their relationship with fetal growth.Additionally, the present study was conducted during the COVID-19 pandemic, providing information about environmental hormone exposure during this period.
However, our study also has several limitations.Recently, Doppler parameters have been recognized as better predictors of FGR outcomes than the HC/AC ratio, yet our study did not include severe FGR cases exhibiting significant Doppler abnormalities.In addition, we did not provide additional information related to placental dysfunction, such as placental biopsy results or sFlt-1/ PlGF results.The number of FGR cases was limited, which prevented us from presenting adjusted results after controlling for confounding variables.Although this study was conducted prospectively, the collection of maternal urine and cord blood samples at the time of delivery and the assessment of fetal growth at that point limited our ability to establish causality between EDC exposure and fetal growth.

Conclusion
BPA can impede fetal growth through various mechanisms, with particular attention given to those associated with placental insufficiency.Our study demonstrated an association between greater maternal BPA exposure and elevated fetal BPA levels, providing preliminary evidence of a link between BPA exposure and asymmetric fetal growth restriction due to placental insufficiency.Further research using large datasets is needed to strengthen these findings, paying particular attention to investigating the relationship between Doppler abnormalities and EDC exposure.

Figure 1
Figure1shows the correlation between each sample.The Pearson correlation coefficient between maternal urine BPA and fetal cord

FIGURE 1
FIGURE 1Correlations between EDCs in maternal urine and fetal cord blood.The red squares represent the correlations between maternal urine and fetal cord blood for each EDC.The size of the blue circles indicates the strength of the positive relationship, and the size of the red circles indicates the strength of the negative relationship.The asterisks denote statistically significant associations.EDC, endocrine disrupting chemicals; BPA, bisphenol-A; MEP, monoethyl phthalate; PFOA, perfluorooctanoic acid.

TABLE 1
Baseline characteristics of total study population.

TABLE 2
Associations of the levels of EDCs with HC/AC ratio using multivariable linear regression model.
The bold font means remained statistically significance after adjusting for maternal age, body mass index, past smoker, gestational age at exam, and fetal sex.EDC, endocrine disrupting chemicals; HC, head circumference; AC, abdominal circumference; CI, confidence interval; BPA, bisphenol-A; MEP, monoethyl phthalate; PFOA, perfluorooctanoic acid.

TABLE 3
Differences in EDC concentrations between non-FGR group and FGR group, or between non-FGR group and FGR with asymmetry group.Comparison between FGR group and Non-FGR group.b Comparison between FGR with asymmetry group and Non-FGR group.FGR group included fetuses with FGR with asymmetric growth and FGR with symmetric growth.EDC, endocrine disrupting chemicals; FGR, fetal growth restriction; BPA, bisphenol-A; MEP, monoethyl phthalate; PFOA, perfluorooctanoic acid.The bold value means remained statistically significance. a