Prenatal maternal blood triglyceride and fatty acid levels in relation to exposure to di(2-ethylhexyl)phthalate: a cross-sectional study

Abstract

Objectives

The hypolipidemic effects of di(2-ethylhexyl)phthalate (DEHP) exposure in humans have not been investigated. And the influences of maternal prenatal DEHP exposure on birth outcomes are not well-known. We aimed to estimate prenatal DEHP exposure in maternal blood, and evaluate its relationships to maternal blood triglyceride (TG) and fatty acid (FA) levels and to birth outcomes.

Methods

We studied 318 mother–newborn pairs residing in Sapporo, Japan. Blood was taken one time during pregnancy for each mother. Maternal and infant characteristics were obtained from medical records and questionnaire survey. We measured DEHP metabolite, mono(2-ethylhexyl) phthalate (MEHP), along with TG and 9 FAs using maternal blood, and analyzed associations of MEHP level with maternal blood TG/FA levels and infant birth dimensions.

Results

Maternal blood TG and palmitoleic/oleic acid levels were higher, but stearic/docosahexaenoic acids and MEHP were lower during late pregnancy. Maternal blood MEHP levels inversely correlated with TG and palmitic/palmitoleic/oleic/linoleic/α-linolenic acids. After adjustment for confounders, we found that a tenfold increase in blood MEHP levels correlated with a decrease in TG of 25.1 mg/dl [95 % confidence interval (CI) 4.8–45.3 mg/dl], and similar relations in palmitic (β = −581.8; 95 % CI −906.5, −257.0), oleic (β = −304.2; 95 % CI −518.0, −90.5), linoleic (β = −348.6; 95 % CI −510.6, −186.6), and α-linolenic (β = −6.3; 95 % CI −9.5, −3.0) acids. However, we observed no correlations between maternal blood MEHP levels and infant birth weight, length, chest circumference, or head circumference.

Conclusions

Ambient DEHP exposure during pregnancy inversely correlated with maternal blood TG and 4 FA levels, but not birth outcomes.

Introduction

Di(2-ethylhexyl)phthalate (DEHP) is the most widely used plasticizer of polyvinylchloride in the manufacture of a wide variety of consumer goods, such as food packaging, building products, clothing, car products, medical devices and children’s products (but not in toys intended for mouthing) [1, 2], although the use of DEHP in the latter two goods has been decreasing recently due to several government restrictions [3, 4]. DEHP is not chemically bound to polyvinylchloride and leaks from polyvinylchloride items with time and use. Consequently, it is a ubiquitous environmental contaminant [1, 2]. Ambient exposure to DEHP in the general adult population may be in the range of 3–30 μg/kg body weight/day [5], while that in Japanese pregnant women has been estimated to be 3.45–41.6 μg/kg/day [6]: no difference in the exposure levels between the former and the latter. Fujimaki et al. [6] identified that the maximum estimated intake level per body weight reached the old Tolerable Daily Intake level of 40–140 μg/kg/day set by Japanese Ministry of Health, Labour and Welfare. Importantly, the intake level was more than the latest Tolerable Daily Intake level of 30 μg/kg/day set by Food Safety Commission of Japan [7]. Therefore, ambient DEHP exposure and potential adverse effects in Japanese warrants close concerns. After entering the human body, DEHP is first metabolized to the monoester, mono(2-ethylhexyl) phthalate (MEHP), which can be oxidized further to oxidative metabolites [8].

DEHP is a known reproductive and developmental toxicant in animals [1, 5, 8]. The growing relevant toxicity reports from experimental animals, together with widespread human exposure, raise serious concerns over the potential risks from human exposure to DEHP. There have been emerging studies conducted of human health outcomes in relation to prenatal DEHP exposure in recent years. Prenatal DEHP exposure has been associated with shorter gestational age at birth [9, 10], lower mental and psychomotor development indices [11], and lower birth weight [12]. Moreover, prenatal DEHP exposure was reported to correlate with shorter anogenital distance, reduced penile size and incomplete testicular descent [13, 14], more non-optimal reflexes [15] and reduced masculine play behavior [16] in male infants, suggesting that possible sex difference exists in DEHP toxicity. Additionally, our group recently reported that maternal exposure to DEHP decreased plasma levels of triglycerides (TG) [1] and four fatty acid (FA) components [17] in prepartum mice, including palmitic acid, oleic acid, linoleic acid (LA), and α-linolenic acid (ALA), which was suspected to correlate with adverse effects of DEHP. However, no information regarding humans is available.

To date, the potential health hazards from exposure to DEHP and/or its main metabolite MEHP in humans at risk, such as pregnant women and infants, have not been well-documented and warrant extensive investigation. Thus, this study aimed to estimate MEHP levels in the blood of pregnant women as a biomarker of ambient DEHP exposure, and to evaluate potential associations with blood levels of TG and FAs in pregnant women and term birth outcomes in infants.

Materials and methods

Study population

This study was part of the “Hokkaido Study on Environment and Children’s Health” [18–20]. Briefly, from July 2002 to October 2005, we approached pregnant women who were at 23–35 weeks of gestation and had no serious illnesses and medical complications to register with a hospital-based prospective cohort study at the Sapporo Toho Hospital in Sapporo, Hokkaido, Japan. The following were the exclusion criteria for study subjects: women with incomplete partner’s information, women who had decided to enroll in the Japanese cord blood bank, or women who had decided to deliver the baby at another hospital. Some of the women we approached did not express interest in our study, and some were unable or unwilling to participate in the study. Ultimately, 514 pregnant women (30 % of those approached) were enrolled in this study by providing written informed consent. Maternal and infant medical information were obtained from medical records of antenatal and perinatal examinations at the hospital. A self-administered questionnaire survey was completed after the second trimester to collect potential confounders, as described in detail elsewhere [18–20]. This study was approved by the Institutional Ethical Board for Epidemiologic studies of Hokkaido University Graduate School of Medicine, and Ethics Review Committee of Nagoya University Graduate School of Medicine.

Blood sampling

A blood sample of approximately 40 ml was taken from the maternal peripheral vein at the time of the next prenatal hospital examination after recruitment. If the blood could not be taken during pregnancy due to maternal anemia, it was obtained during hospitalization within a week after delivery. As blood was obtained one time for each woman, the analyses using blood samples were cross-sectional in nature. Consistent with published reports of “Hokkaido Study on Environment and Children’s Health” [19], the blood sampling period was categorized into four groups: 23–31 weeks of gestation, 32–34 weeks of gestation, 35–41 weeks of gestation, within a week after delivery. All samples were stored at −80 °C until analysis.

MEHP level in maternal blood

Blood samples (30 μl) were mixed with 120 μl 1 N HCl, 350 μl saturated saline solution and 50 μl of 10 μM MEHP-d as an internal standard. MEHP was then extracted twice with 500 μl ethyl acetate after shaking for 15 min. The ethyl acetate layer was evaporated, and the sediments were dissolved into 40 μl ethyl acetate. After adding 20 μl N-methyl-N-(tert-butyldimethylsilyl) trifluoroacetamide, the tube was left at room temperature for 60 min, and the MEHP tert-butyldimethylsilyl derivative concentration formed was measured by gas chromatography–mass spectrometry (GC/MS) (6890N, 5973N; Agilent Technologies, CA, USA) under the analytical conditions mentioned previously by Ito et al. [21]. For each sample, duplicate analysis was performed. Ultimately, MEHP levels were available from 493 maternal blood samples. The detection limit of MEHP was 1 pmol/ml. Coefficient of variation (CV) of MEHP measurements within a day was 2.0–7.8 % for 6 days, and CV of day to day for 6 days was 6.2 % at 5 pmol/ml of concentration.

TG concentration in maternal blood

The TG level in blood was measured using TG-IE kits (Wako, Osaka, Japan) after extracting lipids as described by Folch et al. [22].

FA profiles in maternal blood

FA levels in maternal blood were determined duplicately by GC–MS as described in detail in our earlier study [17] after extracting lipids according to the method of Folch et al. [22]. Nine FA species targeted for measurement included palmitic and stearic acids of saturated FAs, palmitoleic and oleic acids of monounsaturated FAs, LA and arachidonic acid (AA) of the n-6 family, and ALA, eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) of the n-3 family. Under the experimental conditions, the detection limits were 2.4 μg/ml for palmitic acid, 1.3 μg/ml for stearic acid, 0.69 μg/ml for palmitoleic acid, 3.6 μg/ml for oleic acid, and 2.0 μg/ml for the others.

Data analysis

Ten registered women were excluded due to miscarriage, stillbirth, relocation, or voluntary withdrawal from the study before follow-up. The following subjects were excluded from analysis: those with maternal pregnancy-induced hypertension (n = 11), diabetes mellitus (n = 1), fetal heart failure (n = 1), or multiple births (n = 7). We also excluded premature births (n = 23)—defined as birth at less than 37 weeks of gestation—from the data analysis to keep the focus on fetal growth [19, 23], and excluded postnatal blood samplings (n = 134) from the analysis to focus on prenatal subjects, resulting in a sample size of 327. Ultimately, the sample size available for analysis from the 327 was 318 subjects who completed the measurements of MEHP, TG and FAs.

Correlations between MEHP concentration or TG and FA levels in maternal blood and characteristics of subjects were analyzed by Spearman’s rank correlation test, Mann–Whitney U test and Kruskal–Wallis test. Finally, unadjusted and multivariable-adjusted linear regression analyses were performed to evaluate the association between MEHP levels and concentrations of TG and FAs in maternal blood. TG and FA components were dependent variables; MEHP level was independent variable. In the multivariable-adjusted models, maternal age [24], smoking and alcohol intake during pregnancy [25], inshore fish and deep-sea fish intake during pregnancy [24], and the blood sampling period (based on the correlation analyses) were included as potential confounders. We also performed unadjusted and multivariable-adjusted linear regression analyses to evaluate the association between maternal blood MEHP levels and birth outcome measures. Birth weight, birth length, chest circumference or head circumference were dependent variables; maternal blood MEHP was independent variable. In the multivariable-adjusted models, maternal age, height and pre-pregnancy weight, parity, smoking and alcohol intake during pregnancy, socio-economic status (annual household income), gestational age, infant sex [19, 20, 23], and the blood sampling period were included as potential confounders. For head circumference, the adjusted model also included delivery type [19]. The linear regression analyses were also stratified by infant gender to clarify the interaction with infant gender. Because of the skewed distribution, we treated maternal blood MEHP level as a continuous variable on a log₁₀ scale. All statistical analyses were performed using SPSS software. Results were statistically significant if p < 0.05.

Results

Maternal and infant characteristics and their association with maternal blood MEHP level

We included 318 mother-infant pairs in the study (Table 1). The women aged less than 30 years accounted for 48.1 %. Approximately, 51.9 % of the women were primiparous, and 37.4 % of them did not conceive before. Mothers had more than 13 years of education (55.3 %), and 3–5 million yen of annual household income (48.1 %). Mothers who smoked during pregnancy were 21.4, and 33.3 % had alcohol intake during pregnancy. The number of women who ate inshore fish and deep-sea fish at least 1–2 times/week was 144 (45.3 %) and 163 (51.3 %), respectively. Blood sampling was conducted during 23–31 gestational weeks for 42.5 % of women. Two-hundred seventy-three women had vaginal births (85.8 %). One hundred and fifty-one infants (47.5 %) were male, and the gestational age of 53.8 % pregnancy was less than 40 weeks. The number of infants with birth weight, birth length, chest circumference and head circumference less than 3098.0 g, 48.3, 31.5 and 33.2 cm, respectively, was 157 (49.4 %), 156 (49.1 %), 121 (38.1 %), and 157 (49.4 %), respectively. Additionally, we observed statistically significant differences in medians of MEHP levels by blood sampling period (p = 0.002), but did not find significant correlations with other characteristics. Thus, we adjusted the blood sampling period in the multivariate regression models.

Table 1 Maternal and infant characteristics, and their association with maternal blood MEHP level (n = 318)

MEHP level in blood of pregnant women

The mean (± SD) MEHP level was 0.049 ± 0.040 nmol/ml, ranging from 0.007 to 0.316 nmol/ml with a median of 0.039 nmol/ml.

TG and FA levels in maternal blood

Mean (± SD) TG was 94.1 ± 53.7 mg/dl (Table 2). Mean palmitic and stearic acids were 2052.4 (±853.4) and 569.1 (±206.3) μg/ml, respectively. Mean palmitoleic and oleic acids were 122.9 (±75.8) and 1215.5 (±562.7) μg/ml, respectively. Mean LA and AA were 735.6 (±426.6) and 71.4 (±42.6) μg/ml, respectively. Mean ALA, EPA, and DHA were 11.0 (±8.5), 10.0 (±8.4), and 30.2 (±21.8) μg/ml, respectively. Altogether, the mean (± SD) total FAs was 4818.2 ± 1982.0 μg/ml. Highly abundant and common FAs from the total FA pool were palmitic acid (42.8 %), oleic acid (25.0 %), LA (14.8 %), and stearic acid (12.4 %).

Table 2 Concentrations of TG and FA components in maternal blood (n = 318)

TG and FA relationships to subject characteristics

The relationships between TG and FA levels and subject characteristics, which were potential confounders of TG and FAs, were checked (data not shown). We found significant differences in the ALA level by maternal age (p = 0.044), in the EPA level by inshore fish intake (p = 0.001), and in TG (p < 0.001), stearic acid (p = 0.002), palmitoleic acid (p = 0.012), oleic acid (p = 0.003), and DHA (p = 0.020) levels by the blood sampling period. Specially, except for stearic acid and DHA, the levels of TG, palmitoleic and oleic acids increased in the late gestational age when blood sampling.

Relationships between MEHP levels and TG/FA concentrations in maternal blood

Table 3 showed univariate and multivariate regression model results for maternal blood TG and FAs on log₁₀-transformed MEHP concentration. Multivariate models of TG and FAs were adjusted for confounders correlated with TG and FAs at p values <0.05, factors known to be related to TG and FAs from previous reports [24, 25] and the blood sampling period. In the crude model, we found significant negative correlations of MEHP exposure with TG, palmitic acid, palmitoleic acid, oleic acid, LA and ALA in maternal blood. After full adjustment for potential confounders, we found that a tenfold increase in MEHP levels correlated with a decrease in TG of 25.1 mg/dl [95 % confidence interval (CI), 4.8–45.3 mg/dl], and similar relations in palmitic acid (β = −581.8; 95 % CI −906.5, −257.0), oleic acid (β = −304.2; 95 % CI −518.0, −90.5), LA (β = −348.6; 95 % CI −510.6, −186.6), and ALA (β = −6.3; 95 % CI −9.5, −3.0). We also evaluated the relationships of MEHP with TG and FA levels in maternal blood by blood sampling periods (Table 4). Maternal blood MEHP level significantly and negatively associated with palmitic acid, oleic acid, LA and ALA levels at 23–31 weeks of gestation, and TG, palmitic acid, palmitoleic acid, oleic acid, LA and ALA levels at 32–34 weeks in univariate and multivariate regression models. MEHP level in maternal blood negatively associated with LA in univariate regression model at 35–41 weeks, which was insignificant after adjusting for confounders.

Table 3 Regression coefficients (95 % confidence interval, CI) between log₁₀-transformed MEHP level (nmol/ml) and TG/FA in maternal blood

Table 4 Regression coefficients (95 % confidence interval, CI) between log₁₀-transformed MEHP level (nmol/ml) and TG/FA in maternal blood by blood sampling periods

Relationships between maternal blood MEHP levels and infant birth outcomes

Table 5 showed the results of univariate and multivariate regression analyses for the association between maternal blood MEHP levels and birth weight or birth size of infants. Multivariate models were adjusted for known risk factors correlated with birth outcomes from previous reports [19, 20, 23] and the blood sampling period. No significant associations with birth weight, birth length, chest circumference, or head circumference were found for prenatal MEHP level in maternal blood in all infants, males or females.

Table 5 Regression coefficients (95 % confidence interval, CI) between maternal blood log₁₀-transformed MEHP concentration (nmol/ml) and infant birth outcomes

Discussion

To our knowledge, this study was the first to focus on the potential effects of DEHP exposure on blood levels of TG and FA components in pregnant women, and to find significant inverse associations between DEHP principal metabolite MEHP levels, and TG, palmitic acid, oleic acid, LA and ALA levels in maternal blood after adjustment for confounders, using a Japanese pregnancy cohort. TG is an ester derived from glycerol and three FAs. Since palmitic acid, oleic acid and LA were predominant FA components, accounting for 82.6 % of the total FA pool, it was conceivable that these three FAs also similarly declined upon DEHP exposure, as TG decreased. These findings were quite similar to the effects of DEHP exposure on TG and FA levels in the plasma of prepartum mice in our earlier study [1, 17]. We also observed significant differences in maternal blood MEHP, TG, stearic/palmitoleic/oleic acid and DHA levels by blood sampling period (gestational weeks): MEHP, stearic acid and DHA were lower, whereas TG and palmitoleic/oleic acids were higher, in late gestation (35–41 weeks). Additionally, maternal plasma free FAs are an important source of essential FAs to the developing fetus [26]. Although MEHP exposure negatively correlated with TG and several FAs in maternal blood, no significant relationships of MEHP with birth weight and birth size of newborns were detected in the present study.

Maternal hypertriglyceridemia appears late in normal pregnancy [27]. DHA status steadily declines after a temporary increase until 18 weeks of gestation in maternal blood throughout normal pregnancy [28]. Here, maternal blood TG and palmitoleic/oleic acid levels were indeed significantly higher, whereas DHA was lower in women during late gestation (35–41 weeks) as the blood sampling period proceeded. Maternal physiological changes that occur normally in early pregnancy but are most pronounced in the third trimester have the potential to alter xenobiotic distribution and elimination [29]. The lower maternal blood MEHP level in late gestation most likely corresponded with increased renal blood flow and glomerular filtration rate, and unchanged renal tubular resorption, which enhanced the clearance of MEHP through the kidney. Additionally, the activity of hepatic DEHP metabolizing enzymes may have changed during pregnancy, which warrants further study. Taking the various alterations in maternal blood TG/FA concentrations during pregnancy into consideration, we did not consider that great influences by increased plasma volume and total body water would be likely.

The hypolipidemic effects of DEHP were first described in rats and mice by Reddy et al. [30]. Other studies confirmed the hypotriglyceridemic effects of DEHP in rats [31–33] and pregnant mice [1]. Most importantly, and in keeping with our finding of hypotriglyceridemic effects after DEHP exposure in pregnant mice [1], we further found that dietary DEHP exposure significantly reduced plasma levels of palmitic acid, oleic acid, LA and ALA in these pregnant mice with the same experimental protocol [17]. In humans, the maternal FA/lipid homeostasis environment may have dramatic changes upon exposure to DEHP during pregnancy, e.g. the composition and distribution of maternal FAs in the blood. However, no information is available, even though exposure to DEHP is common in the general population. In the present study, it was noteworthy that after adjustment for confounders including the blood sampling period, MEHP levels inversely correlated with TG, palmitic acid, oleic acid, LA and ALA levels in maternal blood. Moreover, maternal blood MEHP level inversely associated with palmitic acid, oleic acid, LA and ALA at 23–31 weeks of gestation, as well as TG, palmitic acid, palmitoleic acid, oleic acid, LA and ALA at 32–34 weeks in multivariate models. Short chain FAs with less than 20 carbon atoms may be susceptible to DEHP exposure in pregnant women. Among them, ALA may be more sensitive to DEHP, even though it accounted for only 0.2 % of total FAs. These findings were quite similar to those from pregnant mice [1, 17]. On the other hand, at 35–41 weeks of gestation, no such relationships between maternal blood MEHP and TG/some FAs were observed. This may be due to significant low concentrations of MEHP in blood compared to those of 23–34 weeks of gestation, which mitigates the effects of DEHP on TG or FAs. Of course, the exact reason must be warranted. Therefore, we could not conclude whether effects of DEHP exposure on maternal blood TG/some FA levels are related to the physiological status of pregnancy or not. To answer this question, similar epidemiological study using non-pregnant women is also required.

The effects of DEHP on TG and FA levels were observed at lower exposure level than that of birth defects: DEHP exposure decreased ALA concentration in blood of wild-type mice by 50 % at a dose of 0.01 % [17], however, birth defects were observed at 0.05 % DEHP [1]. Plasma MEHP concentrations of mice at 0.01 % DEHP exposure were 0.98 ± 0.36 nmol/ml (unpublished data). The concentrations of MEHP in blood of pregnant women were 0.049 ± 0.040 nmol/ml (0.039 nmol/ml of median, ranging from 0.007 to 0.316 nmol/ml) in the present study. Lipase activity for DEHP was fivefold higher in mice than that of non-pregnant women [34]. Body burden in the pregnant women with highest MEHP level may be roughly estimated to be similar to that of mice exposed to 0.01 % DEHP. Therefore, it may be plausible that DEHP at the exposure levels influenced some FA levels in blood of pregnant women of current study. Of course, further study is warranted to clarify the exact exposure level of pregnant women. Ait Bamai et al. [35] compared DEHP level in floor dust among several countries and reported that the levels in Sapporo, Japan, where the current study was conducted, were higher than those in USA, Germany and Denmark. Although it is questionable that whether DEHP levels in the floor dust reflects the exact body burden in humans, the subjects in the current study may be exposed to relatively high level of DEHP. The present study not only for the first time observed the hypolipidemic effects of DEHP exposure in pregnant women, but got such findings at a mean level of MEHP exposure (0.049 ± 0.040 nmol/ml, that is 0.014 ± 0.011 μg/ml), lower than a mean MEHP concentration of 0.68 ± 0.85 μg/ml reported in the human maternal plasma at term in healthy subjects [36].

We further considered the possible mechanisms by which maternal MEHP level negatively correlated with blood levels of TG and four individual FAs, although the present study did not conduct mechanism study. It is well-understood that administration of DEHP to rodents in vivo and in vitro produces pleiotropic response in the liver, which is responsible for hypolipidemic effects [1, 37, 38], especially, a hallmark response of peroxisome proliferation to DEHP in the liver [30], the activation of peroxisome proliferator-activated receptor α (PPARα)-dependent hepatic FA catabolism by DEHP exposure [39], and a decrease in microsomal triglyceride transfer protein-mediated TG transport from liver to the blood after DEHP exposure [1]. Conversely, limited DEHP-specific human data are available. Previous study further found that distinct from wild-type mice expressing PPARα in several organs including liver, DEHP exposure did not influence the plasma levels of palmitic acid, oleic acid, LA and ALA in pregnant humanized PPARα and PPARα knockout mice [17]. Humanized PPARα mice over-expressed human PPARα only in liver, suggesting that PPARα in other organs but not in liver or species difference of PPARα function may be involved in the influence of DEHP on blood TG and FAs in the pregnant women. It may be difficult to infer the mechanism of DEHP influences on TG or FAs in pregnant women from our previous animal studies [1, 17]. Thus, studies aimed at elucidating the mechanism by which DEHP exposure inversely correlated to TG and FA levels in blood of pregnant women are required.

The growth and development of the fetus and its organs depend on a sufficient supply of nutrients including FAs and lipids crossing the placenta, and fetal growth determines the birth outcomes of newborns [40]. Given the significant negative correlations of DEHP exposure with maternal blood TG and four FA levels in this Japanese pregnancy cohort, we further evaluated the effects of maternal DEHP exposure on term birth outcomes of newborns. However, we did not find any significant relationships to birth weight, birth length, chest circumference, or head circumference in univariate and multivariate linear regression analyses. No abnormal birth outcomes were noted in newborns whose mothers had a relatively greater exposure to MEHP (mean ± SD, 0.68 ± 0.85 μg/ml in maternal plasma) during the prenatal period in Italy [36]. However, under a higher maternal MEHP exposure (e.g. median, 2.9 mg/L in maternal blood with low body weight infants), the cord blood MEHP level was associated with low birth weight in a nested case–control study of Chinese newborns [12], where a higher median MEHP (2.5 mg/l) level was found in cord blood with low birth weight infants. These varied findings do not necessarily support the consideration of race difference. The insignificant association in the present study was primarily related to lower MEHP exposure level, which further lowered during late pregnancy. MEHP effects on fetal growth and infant birth outcomes may occur at a much higher exposure level. On the other hand, maternal plasma levels of AA, EPA and DHA during pregnancy were reported to be associated with birth weight, birth length or head circumference [23, 41], while maternal blood MEHP level had no influences on maternal blood levels of these long-chain polyunsaturated FAs in the present study, as a result, MEHP did not influence birth outcomes.

There are some limitations in this study. First, although this is a part of a prospective cohort study, the analyses of maternal blood MEHP, TG and FA levels are cross-sectional in nature. Second, confounders under consideration in multivariate linear regression analyses may not have been completed, e.g. second-hand smoke exposure was not assessed. Third, this study has a small sample size. Fourth, selection bias may have occurred because this cohort was based in one area hospital, which treated pregnant women in Sapporo and the surrounding areas. Fifth, this study measured MEHP alone which has a shorter half-life of elimination and tends to be influenced during sample collection and processing [8], but no oxidative metabolites of DEHP. Thus, the DEHP body burden in pregnant women might have been underestimated. In addition, pregnant women have several physiological changes: sex hormones may enhance TG/FA transport to muscles for oxidation; circulating TG is diverted to uptake by the mammary gland and by the placenta [27, 42]. Humans are exposed to a variety of environmental pollutants and chemicals which may elicit additive biological effects. These factors may cause an overestimated correlation between MEHP exposure and maternal blood TG/FA levels.

In conclusion, no adverse effects of maternal prenatal DEHP exposure on infant birth weight and birth size were observed. The hypotriglyceridemic effects of DEHP exposure in pregnant women were documented for the first time in this study, although at a lower exposure level. Furthermore, four individual FA levels in maternal blood inversely correlated with DEHP exposure. All these results raise concerns over the maternal blood FA/lipid homeostasis environment under ambient DEHP exposure during pregnancy, which warrants urgent investigation by epidemiological studies.