Phthalate Levels and Low Birth Weight: A Nested Case-Control Study of Chinese Newborns

Abstract

Objective

To assess maternal–fetal exposure to phthalates and investigate whether in utero phthalate exposure is associated with low birth weight (LBW).

Study design

A total of 201 newborn–mother pairs (88 LBW cases and 113 controls) residing in Shanghai were enrolled in this nested case-control study during 2005–2006. Maternal blood, cord blood, and meconium specimens were collected and analyzed for phthalates by high-performance liquid chromatography–mass spectrometry. Nonparametric tests were used to compare demographic characteristics in cases and controls. Conditional logistic regression and Spearman correlation were used to analyze the association between phthalate exposure and LBW.

Results

No significant differences in gestational age, prepregnancy body mass index, prenatal care, vitamin supplementation, or socioeconomic levels were found between the LBW and control infants. More than 70% of the bio-samples had quantifiable levels of phthalates, with higher levels in the LBW infants compared with the controls. Prenatal di-n-butyl phthalate (DBP) exposure was associated with LBW, and di-2-ethylhexyl phthalate (DEHP) was negatively associated with birth length. After adjusting for the potential confounders, DBP concentrations in the highest quartile were associated with an increased risk of LBW.

Conclusions

Newborns in China are ubiquitously exposed to phthalates; significantly higher phthalate levels were detected in LBW cases compared with controls. In utero DBP and DEHP exposures were associated with LBW in a dose-dependent manner. Prenatal phthalate exposure may be a risk factor for LBW. (J Pediatr 2009;155:500–4).

Low birth weight (LBW) is one of the leading causes of mortality in children under age 5 years and is associated with increased risk of cardiovascular and metabolic disease in adulthood.^1–3 Comparisons of studies performed over the last 2 decades show geographical differences in the rate of LBW.⁴ In the United States, the percentage of LBW infants increased from 6.7% in 1985 to 7.8% in 2002, and LBW is a problem in developing countries as well (eg, rates of 5.8% in urban areas and 11.8% in rural areas of China in 1998).^5–7

Environmental factors, especially manmade chemicals, affect the health and well being of children. The World Health Organization estimates that > 30% of the global burden of disease in children can be attributed to environmental factors.⁸ Socioeconomic factors, maternal malnutrition and smoking, environmental pollutants (eg, polybrominated biphenyls, perfluorinated chemicals, dichlorodiphenyltrichloroethane) have been linked to the development of LBW.^9–13

As plasticizers used in cosmetic products, personal care products, and some medical devices (eg, blood storage bags, intravenous medical tubing), phthalates are found in almost all categories of personal care products for infants, children, and adults, resulting in widespread nonoccupational human exposure through multiple routes.¹⁴ Although phthalates are rapidly metabolized, they are detected in the environment and in humans. The effects of phthalates on fetal growth has been investigated in human epidemiologic studies.^15–17 Latini et al¹⁵ detected di-2-ethylhexyl phthalate (DEHP) and mono-2-ethylhexyl phthalate (MEHP, a metabolite of DEHP) in the cord blood of newborns, suggesting that phthalate exposure can begin in utero. Prenatal phthalate exposure at environmental levels can adversely affect male reproductive development in humans in a manner similar to that seen in rodents.¹⁷

In rodents, reduced fetal birth weight and shortened gestational period were considered the most sensitive endpoints of the effects of phthalates. In rats exposed to di-n-butyl phthalate (DBP), the lowest observed adverse effect level for decreased birth weight in the F₂ generation was 52 mg/kg/day for males and 66 mg/kg/day for females.¹⁸ In addition, insulin resistance, neurobehavioral abnormalities, and testicular dysgenesis syndrome are more common in those who were small at birth.^3,19 No data are available regarding a possible association between fetal phthalate exposure and neonatal LBW in humans, however.

We propose the following hypotheses: (1) Phthalate levels are higher in LBW newborns than in newborns of normal birth weight, and (2) in utero phthalate exposure may account for the development of LBW in newborns. Accordingly, we designed this nested case-control study to investigate the maternal–fetal phthalate exposure in a Shanghai cohort, and to explore the possible association between phthalate exposure and LBW.

Methods

Between October 2005 and December 2006, a total of 3316 infants were born at Shanghai Medical Center for Maternal and Child Health, of which 125 (3.8%) were identified as LBW. Thus, 250 mother–newborn pairs (125 cases and 125 controls) were initially recruited and paired by maternal age and newborn sex. After excluding multiple-birth pregnancies and premature deliveries, 201 pairs qualified for the case-control study, completed the questionnaires, and provided samples. The 201 singleton subjects consisted of 88 term LBW newborns (gestational age ≥ 37 weeks, birth body weight < 2500 g), and 113 control term newborns (gestational age ≥ 37 weeks, birth body weight ≥ 2500 g).

All subjects were residents of Shanghai and provided written informed consent approved by Fudan University’s Institutional Review Board. The mothers completed questionnaires eliciting information about the newborn’s birth status and maternal information (eg, medical history, menstruation, civil status, education, work, leisure activity, tobacco, drinking, dietary habits [including vitamin supplementation], degree of physical activity). All subjects were examined and investigated before the results of chemical analysis were known. Newborns’ supine length was measured using an infantometer, and birth weight, measured on a digital scale, was obtained from hospital records.

Sample Collection and Measurement

Umbilical vein blood was obtained from 88 LBW and 113 control infants immediately after delivery using a syringe. Maternal blood samples were collected after the mothers completed the questionnaire. For each infant, meconium was collected directly from every diaper during the first 48 hours after delivery, and all of these meconium samples were pooled into a single sample. All specimens were collected with glass devices to avoid contamination by phthalates during handling and storage. Frozen samples were stored in phthalate-free containers and transferred on dry ice to the Fudan University laboratory for analysis.

Three commonly used phthalates (di-ethyl phthalate [DEP], DBP, and DEHP) and 2 of their metabolites (mono-butyl phthalate [MBP] and MEHP) were analyzed in cord blood and meconium specimens. Analyses of phthalates and their monoesters in serum and meconium samples were done as described previously.^20,21 In brief, the determination of phthalates and metabolites in serum (1 mL) and meconium (1 g) involved enzymatic deconjugation of the glucuronidated metabolites, solid-phase extraction, separation with high-performance liquid chromatography (HPLC), and detection by mass spectrometry (MS). HPLC-MS analysis was performed with a Bio-Rad Variant analyzer (Bio-Rad, Hercules, California) at Fudan University’s Key Laboratory of Ministry of Education on Public Safety, using the manufacturer’s reagents and following the manufacturer’s protocol. The limit of detection (LOD) was 1.0 ng/g for meconium and 0.2 to 1.0 µg/L for serum. Analysts at the Key Laboratory of Ministry of Education on Public Safety were blind to all information concerning our subjects. This study was conducted in accordance with protocols approved by Fudan University’s Human Studies Committee.

Data Analysis

Parametric t-tests and χ² tests were used to compare demographic characteristics in the LBW infants and controls. Phthalate concentrations were skewed to the right, and medians and interquartile ranges are presented to characterize phthalate concentrations in the descriptive analysis. The value of the LOD divided by the square root of 2 was used to estimate the value of samples below the LOD.²² Analyses of potential differences in phthalate levels in serum and meconium samples between LBW cases and controls were conducted using the Mann-Whitney U-test. Spearman correlation was used to explore the associations between individual phthalate concentrations and birth weight and length in the newborns. Log-transformed phthalate concentrations were used in Spearman correlation, to minimize the potential effect of extreme values on the correlation coefficients. In addition, the distribution of each phthalate was divided into quartiles, and an odds ratio (OR) was calculated for each quartile compared with the lowest quartile. Conditional logistic regression was used to examine the relationship between each phthalate and LBW, taking into account some potential confounders and the effect modifiers (eg, gestational age, pregnancy complications, exposure to tobacco smoke at home, socioeconomic level, prepregnancy body mass index [BMI]). The analyses were considered statistically significant when P < .05. All statistical analyses were conducted using the SPSS 17.0 statistical package (SPSS Inc, Chicago, Illinois).

Results

In this subject population, first births represented 80% of the newborns, of which 51% were male (Table I). There were no significant differences in infant sex, gestational age, or delivery mode, except for pregnancy complications (including pregnancy-induced hypertension, diabetes, infection, and intrahepatic cholestasis syndrome). The majority of the mothers had completed an undergraduate education (65%) and had received 5 prenatal care examinations (55%). A few mothers (< 9%) reported family members smoking or drinking alcohol at home during the entire pregnancy.

Table I.

Selected characteristics of LBW infants and controls

Characteristic	Controls (n = 113)	LBW (n = 88)	P *
Birth weight, g†	3366 (3130∼3637)	2309 (2157∼2333)	.000
Birth length, cm‡	48.8 ± 1.5	45.3 ± 2.4	.01
Gestational age, months‡	39.7 ± 2.1	40.4 ± 2.8	.78
Sex, male/female	58/55	45/43	.98
Uncomplicated vaginal births, yes/no	70/43	57/31	.50
Maternal age, years‡	28.1 ± 3.3	28.4 ± 3.9	.75
Prepregnancy BMI </>18.5	26/87	17/71	.53
Prenatal care (Kessner index), adequate/inadequate	60/53	49/39	.72
Pregnancy complications, yes/no§	32/56	18/95	.001
Irregular menstruation, yes/no	17/96	21/67	.12
Occupational exposure, yes/no	25/88	25/63	.31
Vitamins supplement, yes/no	48/65	38/50	.46
Maternal/paternal smoking, yes/no	5/108	7/81	.22
Location, urban/county	78/35	52/36	.36
Occupation, industrial/nonindustrial worker	12/101	19/69	.08
Education, < /> high school	39/74	32/56	.71
Income, </>$400/month	22/91	23/65	.82

^*Calculated using the t-test and x² test.

^†Data shown as median (25th∼75th percentiles).

^‡Data shown as mean ± standard deviation.

^§Pregnancy complications include pregnancy-induced hypertension, diabetes, infection, and intrahepatic cholestasis syndrome.

More than 70% of the biological samples had quantifiable levels of phthalates and monoesters, with a higher detection rate in the LBW infants compared with the controls (Table II). DEP, DBP, DEHP, and MEHP, but not MBP, were detected in maternal and cord blood samples; however, only MBP and MEHP (metabolites of DBP and DEHP, respectively) were detectable in meconium samples. MEHP, the only phthalate detected in both infant and maternal samples, was detected in 60 of 88 LBW mother–newborn pairs and in 75 of 113 control pairs. A positively skewed distribution was found for phthalate and monoester levels in the subject population. Phthalate levels were approximately 20% lower in cord blood samples than in maternal blood samples. The LBW infants had significantly higher levels of DBP and MEHP in cord blood samples and up to 2-fold higher levels of MBP and MEHP in meconium samples (Table II). Cord blood and meconium have been identified as biological samples that represent the instantaneous and accumulated exposure levels of chemicals from mothers to fetus and newborns; thus, we focused our analysis mainly on the relationship between newborn phthalate levels and birth weight and length.

Table II.

Phthalate levels [median (25th to 75th percentile)] in serum and meconium samples from LBW infants (n = 88) and controls (n = 113)

	Maternal blood, mg/L					Cord blood, mg/L					Neonatal meconium, mg/g
	Controls		LBW			Controls		LBW			Controls		LBW
Phthalate	Median	% >LOD*	Median	% >LOD*	P †	Median	% >LOD*	Median	% >LOD*	P †	Median	% >LOD*	Median	% >LOD*	P †
DEP	2.2 (1.6–2.9)	79.4	2.0 (1.6–2.7)	94.3	.65	2.0 (0.9–2.4)	76.1	1.6 (1.3–2.0)	96.6	.62	-	-	-	-	-
DBP	2.2 (1.4–3.2)	82.9	2.9 (2.5–3.4)	92.9	.02	1.8 (1.2–2.7)	74.8	2.7 (2.2–3.0)	97.7	.002	-	-	-	-	-
DEHP	0.6 (0.3–1.2)	72.3	0.7 (0.5–1.3)	84.3	.58	0.5 (0.1–0.9)	67.3	0.6 (0.3–1.0)	77.1	.35	-	-	-	-	-
MBP	-	-	-	-	-	-	-	-	-	-	1.7 (1.2–2.4)	74.3	2.2 (1.6–3.6)	95.5	.003
MEHP	1.4 (1.2–2.1)	80.1	2.9 (1.8–3.5)	88.6	.000	1.1 (0.9–1.7)	71.4	2.5 (1.6–3.4)	85.2	.000	2.9 (1.8–4.4)	76.1	5.5 (3.4–9.3)	87.5	.000

^*LOD was 0.2–1.0 µg/L for serum and 1.0 ng/g for meconium.

^†P values calculated using the Mann-Whitney U-test.

After controlling for the potential covariates (including gestational age, smoking at home, socioeconomic level, and prepregnancy BMI), we used Spearman correlation to analyze the relationship between phthalate exposure and LBW. DBP exposure in utero (DBP in cord blood and MBP in meconium) was significantly associated with LBW, and DEHP exposure was associated with shorter birth length (Table III). The potential dose-response relationships were evaluated by factoring the phthalate levels into quartiles and using conditional logistic regression analyses. After adjusting for the potential confounders, DBP and DEHP concentrations in the highest quartile were found to be associated with an increased risk of LBW (Table IV). A dose-response relationship was found between MEHP level in meconium and LBW. For DBP, a dose-response relationship was found between LBW and both DBP level in cord blood and MBP level in meconium. Such differences between LBW infants and controls were not observed for DEP and DEHP.

Table III.

Spearman correlation between phthalate levels in cord serum and meconium and birth weight or size in newborns

	Cord blood				Meconium
Characteristic	DEP	DBP	DEHP	MEHP	MBP	MEHP
Birth weight
Correlation coefficient	0.03	− 0.23	−0.01	−0.18	−0.56	−0.21
P*	.76	.01	.96	.046	.000	.03
Birth length
Correlation coefficient	0.04	−0.09	−0.22	−0.25	−0.11	−0.47
P*	.60	.23	.05	.001	.16	.000

^*Test made on log-transformed concentrations.

Table IV.

Adjusted ORs (95% confidence intervals) between birth weight and phthalates in cord serum and meconium^*

	Quartile†
Groups	1 (ref)	2	3	4	P
Cord serum, mg/L
DEP	1.0	1.28 (0.92–1.65)	0.69 (0.30–1.57)	1.11 (0.43–2.28)	.38
DBP	1.0	0.54 (0.45–1.47)	2. 69 (1.30–4.74)	3.54 (1.54–6.15)	.008
DEHP	1.0	0.52 (0.23–1.16)	1.27 (0.57–2.84)	1.19 (0.79–2.05)	.11
MEHP	1.0	0.53 (0.31–1.61)	1.22 (0.90–2.52)	2.05 (1.17–3.70)	.05
Meconium, mg/g
MBP	1.0	1.58 (1.08–2.46)	2.84 (1.19–4.82)	4.68 (2.14–6.85)	.000
MEHP	1.0	1.12 (0.89–2.03)	2.89 (1.19–5.02)	3.23 (1.31–5.94)	.04

^*Adjustment made for gestational age, smoking at home, socioeconomic level, prepregnancy BMI, and other phthalate variable (in quartiles), using the backward-elimination method.

^†Quartile 1, √LOD∼24th percentile; quartile 2, 25th∼49th percentiles; quartile 3, 50th∼74th percentiles; quartile 4, 75th percentile∼maximum value.

Discussion

LBW is an important risk factor in development. The world-wide data for LBW indicate that infants with birth weight < 2500 g are at greater risk of dying in infancy or experiencing long-term disabilities. Along with malnutrition and socio-economic factors, environmental pollutants, including phthalates, may possibly play a role in the development of LBW.^2,23 Phthalates are widely used in industrial consumer products, and humans can be exposed to phthalates through ingestion, inhalation, dermal contact, and medical treatments. Numerous studies have found that the effects of phthalates are more significant in children than in adults. Prenatal exposure to DBP, benzyl butyl phthalate, di-isononyl phthalate, and DEHP could lead to underdevelopment and agenesis of androgen-dependent tissues in a dose-additive manner.^24,25

The fetus might have increased susceptibility to the potential adverse effects of phthalates.^26–30 Three common plasticizers—DEP, DBP, and DEHP—have been in commercial use for more than 50 years and account for about 70% of the total phthalate production in China. In the present study, we examined the relationship between maternal–fetal phthalate exposure and LBW by conducting a case-control study in Shanghai, China. To avoid multiple birth and preterm confounders on birth weight, 201 full-term and singleton newborns were selected as our subject population. The percentage of smoking and drinking in family members was < 9%, and there were no significant differences between the LBW infants and controls in terms of socioeconomic factors, including prenatal care, occupation, education, monthly income, and family location. More than 60% of the deliveries were uncomplicated vaginal births, and there was difference in delivery modes between the LBW infants and controls. Thus, it is unlikely that there was a significant difference between the 2 groups in exposure to DEHP resulting from medical interventions during delivery.

Meconium is an appropriate biological sample for assessing fetal exposure to chemicals.^21,31,32 Unlike other biological samples, meconium begins to accumulate in the human fetus at week 16 of gestation and is not excreted until after delivery. The accumulation of chemical pollutants in meconium, the ease of meconium collection, and the large amount of meconium that can be collected are factors contributing to the increased sensitivity of this material.³² Meconium is considered a more valid dosimeter of prenatal exposure to phthalates than cord blood, because of the short half-life of phthalates in humans. In this study, MBP and MEHP levels in meconium and cord blood were higher in the LBW infants compared with the controls. When phthalate exposure levels were calculated as a function of body weight, newborn infants (who are exposed to phthalates solely through their mothers), had 5- to 10-fold higher levels of DBP (50th percentile, 1170 µg/kg /day) and DEHP (260 µg/kg /day) compared with the reference doses established by the US Environmental Protection Agency (100 µg/kg/day for DBP and 22 µg/kg/day for DEHP).

In 2003, Latini et al¹⁵ investigated a possible relationship between DEHP and MEHP levels in cord blood and gestational age and birth weight through a cross-sectional study of 84 newborns in Italy. Although delivery mode and maternal smoking were considered, their study population included 11 preterm newborns, and no information on birth size was provided. Thus, no differences in birth weight were found between DEHP- or MEHP-positive and DEHP- or MEHP-negative newborns. In this study, DBP/MBP levels were significantly associated with LBW, and DEHP/MEHP levels were negatively associated with birth length. Furthermore, dose-response relationships were found between DBP/MBP or MEHP exposure and LBW, especially in the highest quartile (OR up to 4.67). This indicates that newborns, who are immature both developmentally and physiologically, may be at greatest risk for LBW after in utero exposure to DBP and DEHP.

The present study has some limitations. First, data on daily activity and use of personal care products were not available for most of the mothers, and thus the sources of phthalate exposure are not known. Second, smoking and drinking are risk factors for LBW. Although the percentage of mothers reporting smoking by family members in the home was low, exposure to second-hand smoke outside of the home was not assessed. Second-hand smoke in the workplace or other environments might have contributed to LBW and led to overestimation of the phthalate contribution. In addition, the analytical method was neither selective nor sensitive enough to measure all oxidative metabolites of phthalates; thus, the body burden of phthalates in neonates might have been underestimated. Consequently, although we found an association between prenatal exposure to some phthalates and LBW, measurements of the distribution of primary and secondary metabolites in placenta, cord blood, and amniotic fluid, as well as comparisons of the levels of metabolites in maternal serum and urine, would lead to better estimation of the contribution of phthalates to LBW. Continued surveillance and additional research are needed to evaluate the complex potential health risks from high exposure to phthalates.

Glossary

BMI

Body mass index

DBP

Di-n-butyl phthalate

DEHP

Di-2-ethylhexyl phthalate

DEP

Di-ethyl phthalate

HPLC

High-performance liquid chromatography

LBW

Low birth weight

LOD

Limit of detection

MBP

Monobutyl phthalate

MEHP

Mono-2-ethylhexyl phthalate

MS

Mass spectrometry

OR

Odds ratio