Urinary metabolite concentrations of phthalate and plasticizers in infancy and childhood in the UNC Baby Connectome Project

Abstract

Introduction:

Existing evidence suggests that exposure to phthalates is higher among younger age groups. However, limited knowledge exists on how phthalate exposure, as well as exposure to replacement plasticizers, di(isononyl) cyclohexane-1,2-dicarboxylate (DINCH) and di-2-ethylhexyl terephthalate (DEHTP), change from infancy through early childhood.

Methods:

Urine samples were collected across the first 5 years of life from typically developing infants and young children enrolled between 2017 and 2020 in the longitudinal UNC Baby Connectome Project. From 438 urine samples among 187 participants, we quantified concentrations of monobutyl phthalate (MnBP), mono-3-carboxypropyl phthalate (MCPP), monoisobutyl phthalate (MiBP), monoethyl phthalate (MEP) monobenzyl phthalate (MBzP), and metabolites of di(2-ethylhexyl) phthalate (DEHP), diisonoyl phthalate (DiNP), DINCH and DEHTP. Specific gravity (SG) adjusted metabolite and molar sum concentrations were compared across age groups. Intraclass correlation coefficients (ICCs) were calculated among 122 participants with multiple urine specimens (373 samples).

Results:

Most phthalate metabolites showed high detection frequencies (>80% of samples). Replacement plasticizers DINCH (58–60%) and DEHTP (>97%) were also commonly found. DiNP metabolites were less frequently detected (<10%). For some metabolites, SG-adjusted concentrations were inversely associated with age, with the highest concentrations found in the first year of life. ICCs revealed low to moderate reliability in metabolite measurements (ρ = 0.10–0.48) suggesting a high degree of within-individual variation in exposure among this age group. The first 6 months (compared to remaining age groups) showed an increased ratio of carboxylated metabolites of DEHP and DEHTP, compared to other common metabolites, but no clear age trends for DINCH metabolite ratios.

Conclusion:

Metabolites of phthalates and replacements plasticizers were widely detected in infancy and early childhood, with the highest concentrations observed in the first year of life for several metabolites. Higher proportions of carboxylated metabolites of DEHP and DEHTP in younger age groups indicate potential differences in metabolism during infancy.

1. Introduction

Phthalates are a group of high production volume chemicals, commonly used as plasticizers, and found in a variety of consumer and building products. Widespread phthalate use has led to universal human exposure (Wang et al., 2019), through dermal absorption, inhalation, and ingestion, with dietary sources representing a significant pathway (“Chronic Hazard Advisory Panel on phthalates and phthalate alternatives (with appendices),” 2014). Infants and children may be more susceptible to exposure due to physiologic, metabolic, and behavioral differences, such as hand to mouth behavior (U.S. Environmental Protection Agency (EPA), 2005). Moreover, prenatal and early-life exposure to phthalate has been linked to adverse neurodevelopmental and behavioral outcomes (Balalian et al., 2019; Ejaredar et al., 2015; Li et al., 2020).

Population-based biomonitoring through the National Health and Nutritional Examination survey (NHANES) reports higher urinary metabolite concentrations of many phthalates for children aged 6–11 years as compared to older age groups (Centers for Disease Control and Prevention, 2019). In the 2015–2016 NHANES cycle, where phthalate data for children aged 3–5 years were also available, the highest urinary phthalate metabolite concentrations were seen in this age group. Although there is no systematic biomonitoring of children under 3 years, a number of small observational studies in the US have quantified phthalate metabolite concentrations in infancy and early childhood, also finding widespread exposure in infancy and early childhood (Balalian et al., 2019; Brock et al., 2002; Watkins et al., 2014). Studies performed in unique settings, such as the neonatal intensive care unit (NICU), similarly report high prevalence of phthalate exposure (Stroustrup et al., 2018; Weuve et al., 2006). However, few studies have examined age-related trends in exposure and consistency of phthalate measurements by age across infancy and early childhood. Moreover, little data is available on exposure to replacement plasticizers, including di(isononyl) cyclohexane-1,2-dicarboxylate (DINCH) and di-2-ethylhexyl terephthalate (DEHTP) in infancy and early childhood. Therefore, to address these gaps, we examined metabolite concentrations of phthalates and replacement plasticizers in urine samples collected from infants and young children in the University of North Carolina Baby Connectome Project (BCP).

2. Materials and methods:

2.1. Study population and Design

The BCP is a longitudinal study of child development in typically developing children across the first 5 years of life (Howell et al., 2019). Enrollment was conducted at the University of Minnesota and the University of North Carolina at Chapel Hill (UNC) between 2017 and 2020. The BCP used an accelerated cohort design, where enrollment was staggered across age-specific cohorts spanning different periods of infancy and childhood, in order to encompass a wider age range during a shorter follow-up (Howell et al., 2019). Participants were eligible to enroll in the BCP if they were born >37 weeks gestation, with normal birth weight (>2500 g) and if parents reported no medical or genetic conditions related to neurodevelopment. In total, 837 participants were enrolled into the BCP across both enrollment sites. The current study is limited to enrollment that occurred at the UNC site (n = 238) where there was a urine collection attempted at every participant visit. We excluded participants from this study if we were unable to collect urine at any visit (n = 31) or there was no usable developmental data at any visit (n = 20), leaving 187 participants for analysis (Figure 1).

Figure 1.

Selection of samples and participants in a study to characterize infancy and early childhood phthalate and replacement plasticizer exposure in the UNC BCP.

2.2. Urine Sample collection

Urine was collected using one of two methods: 1) For infants and young children wearing diapers, at the start of every in-person visit research assistants were instructed to provide parents with a fresh diaper (Bambo Nature Premium) and Waterwipes to remove residual lotion or creams prior to placing 4–5 hospital-grade cotton balls in the diaper. After the visit, cotton balls were collected in a sterile, phthalate-free container and stored at 4° C, and 2) For toilet trained children, a sterile phthalate-free toilet hat was placed in the toilet available to catch urine at any time during the visit. Once collected, samples were sent to the UNC BioSpecimen Processing Facility (median processing time ~15 hours). Samples compromised by fecal matter were discarded to ensure purity. The urine was carefully extracted from cotton balls using a phthalate-free syringe and stored securely, emphasizing the rigorous approach to avoid contamination and ensure reliable results.

Samples were aliquoted and stored at −80°C within 24 hours of processing. Among the participants eligible for this study (n = 187), 545 urine samples were collected. We excluded 4 urine samples with insufficient volume for metabolite analyses (< 400 μL), and an additional 100 urine samples that were collected within 7 days of another sample in the analysis set (exclusions were based on volume of samples and/or concurrent (same-day) collection of developmental data, such as the Mullen Scales of Early Learning) leaving 441 samples for measurement of phthalate and replacement plasticizer metabolites. The distribution of analyzed samples by age at collection is provided in sFigure 1.

2.3. Measurement of exposure to phthalates

We used a highly sensitive TSQ Quantis triple quadrupole mass spectrometry interfaced with a Vanquish UHPLC to perform a targeted quantification analysis for 17 metabolites of phthalate and replacement plasticizers (sTable 1), including monobutyl phthalate (MnBP), mono-3-carboxypropyl phthalate (MCPP), monoisobutyl phthalate (MiBP), monoethyl phthalate (MEP) monobenzyl phthalate (MBzP), mono-2-ethylhexyl phthalate (MEHP), mono-2-ethyl-5-hydroxyhexyl phthalate (MEHHP), mono-2-ethyl-5-oxohexyl phthalate (MEOHP), mono-2-ethyl-5-carboxypentyl phthalate (MECPP), mono-2-carboxymethylhexyl phthalate (MCMHP), monooxononyl phthalate (MONP), monoisononyl phthalate (MCOP), monocarboxyisononyl phthalate (MCNP), cyclohexane-1,2-dicarboxylic acid, monohydroxy isononyl ester (MHNCH), cyclohexane-1,2-dicarboxylic acid, monocarboxy isooctyl ester (MCOCH), mono-2-ethyl-5-hydrohexyl terephthalate (MEHHTP), and mono-2-ethyl-5-carboxypentyl terephthalate (MECPTP). Specific gravity (SG) was measured using a Reichert TS Meter Model TS400 refractometer. Previous research has established SG (compared to creatinine) as a suitable adjustment method in infancy and early childhood (Pearson et al., 2009; Wang et al., 2015). Urine samples were randomly assigned to analytic batches with all samples for a given participant analyzed in the same batch (n = 18–19 sample per batch). Batches also included 2–3 laboratory-blinded quality control (QC) pooled urine samples. The detailed urine sample preparation and mass spectrometry analysis have been described previously in Hammel et al., (2023). Metabolite could not be measured from 3 samples due to insufficient sample volume leaving 438 with measured metabolites for analysis.

We calculated overall and batch-specific coefficients of variation (CV) from pooled urine samples included in each analytic batch using the following formula: CV(%) = (mean/SD) * 100.

Metabolite values were adjusted for SG to account for urinary dilution using the following formula: ${\text{SGmet}}_{\text{i}}={{\text{met}}_{\text{i}}}^{\ast }\left[(\text{GMean}\left({\text{SG}}_{\text{i}}\right)-1)/({\text{SG}}_{\text{i}}-1)\right]$

$${\text{SGmet}}_{\text{i}}=\text{SG-adjusted metabolite concentration for sample}i$$

$${\text{met}}_i=\text{unadjusted measured metabolite concentration for sample}i$$

$$\text{GMean}=\text{overall geometric mean of metabolite concentration in analysis sample}$$

$${\text{SG}}_{\text{i}}=\text{specific gravity measurement for sample}i$$

Values below the LOD were imputed as LOD/√2. SG-adjusted values of MONP, MCOP, and MCNP were not calculated due to low detection frequencies (<10%) and these metabolites were excluded from subsequent analyses. We calculated molar sums of metabolite concentrations of di(2-ethylhexyl) phthalate (DEHP): ∑DEHPm = MEHP + MECPP + MEHHP + MEOHP; DEHTP: ∑DEHTPm = MECPTP + MEHHTP, and DINCH: ∑DINCH = MCOCH+ MHNCH, in units of μmol/L, by summing SG-adjusted metabolite concentrations (with LOD/√2 imputation) after dividing by the molecular weight. MCMHP was not included in the molar sum of DEHP to facilitate comparison to previous studies that only measured four metabolites. Moreover, inclusion of MCMHP did not meaningfully change concentrations of ∑DEHTPm.

2.4. Statistical analysis

All analyses were performed in SAS v. 9.3 (SAS Institute Inc., Cary, NC, USA). We calculated frequencies and percentages of participants by demographic characteristics. For descriptive analyses of metabolite levels, we calculated the overall detection frequencies (percent of samples with metabolites above the LOD), as well as concentration percentiles and geometric means and standard errors (SEs) by age (at sample collection) groups: 0–6 months, >6 months-1 year, >1–1.5 years, >1.5–2 years, >2–3 years, >3–4 years, >4–5 years, and >5 years. Geometric means and SEs were estimated in a linear mixed model using PROC MIXED to account for multiple samples/measurements per person. We calculated p-values from tests of trend (p_trend)_. in a linear mixed model by modeling SG-adjusted metabolite concentrations using age group categories numbered as a continuous variable as the independent variable (0–6 months = 1, >6 months-1 year = 2, etc.). We created box and whisker plots by age groups and calculated Spearman correlation of SG-adjusted metabolite concentrations.

To quantify the consistency of metabolite measurements over time, we calculated intraclass correlation coefficients (ICCs) of SG-adjusted metabolite concentrations from participants with multiple urine specimens (n = 122 participants, 373 samples), using the SAS ICC9 macro ( https://ysph.yale.edu/cmips/research/software/icc9/). ICC analyses were repeated within the following age groups: 0–1 year, >1–3 years, >3 years, to examine short-term reliability across ages of infancy and early childhood.

To explore age-related differences in DEHP, DINCH and DEHTP metabolism, we calculated ratios of SG-adjusted concentrations for common metabolites of parent compounds from individual samples and created box and whisker plots by age groups. For DEHP, which has a two-step metabolism involving initial hydrolysis and subsequent oxidation steps (sFigure 2), we compared all measured oxidative metabolites to the hydrolysis metabolite (MEHHP:MEHP, MECPP:MEHP, MEOHP:MEHP, and MCMHP:MEHP), the secondary to primary hexyl oxidative metabolite (MECPP:MEHHP, MEOHP:MEHHP), the secondary hexyl carboxylated to hexyl oxidated metabolite (MECPP:MEOHP) and secondary ethyl to hexyl carboxylated metabolite (MECPP:MCMHP). For DINCH and DEHTP we compared the ratios of carboxylated and oxidative metabolites (DINCH: MCOCH:MHNCH and DEHTP: MECPTP:MEHHTP).

3. Results:

The average of the batch-specific CVs from laboratory-blinded QC pooled urine samples for all metabolites ranged from 1.6 to 10.1% (sTable 2). Except for batch 1, batch specific CVs rarely exceeded 20%. In batch 1, we observed elevated CVs for DEHP and DINCH metabolites, however inspection of CVs from laboratory unblinded QCs found no issues for any metabolite. The remaining laboratory-blinded phthalate metabolite CVs for batch 1 were in range of all other batches. Moreover, in the analysis sample, average concentrations of metabolites from batch 1 did not differ from remaining batches. Therefore, we concluded that there was likely a problem with one QC sample in batch 1 but not a systematic problem with metabolite concentrations from analysis samples.

The 187 participants in the analytic sample comprised approximately equal numbers of boys and girls, 48% had a first sample collected in the first year of life and 62% contributed multiple samples (Table 1). Caregivers of participants had high educational attainment and income, with 76% of mothers having a 4-year degree (or higher) and 51% having a total household income at or above $100,000 per year. Characteristics of participants by age groups (at enrollment) is shown in Supplementary Table 3.

Table 1.

Characteristics of participants in a study to characterize infancy and early childhood phthalate and replacement plasticizer exposure in the UNC BCP (N = 187).

Characteristic		n	(%)
Participant age at enrollment
	0–6 m	27	(22)
	>6 m −1 y	32	(26)
	>1–1.5 y	29	(24)
	>1.5–2 y	19	(16)
	>2–3 y	15	(12)
	>3–4 y	23	(12)
	>4–5 y	27	(14)
	>5 y	15	(8.0)
Number of urine samples
	1	65	(35)
	2	61	(33)
	3	29	(15)
	4	13	(7.0)
	5+	21	(11)
Participant sex, male		92	(49)
Participant race/ethnicity
	non-Hispanic white	132	(71)
	non-Hispanic black	19	(10)
	Other	36	(19)
Birth year
	2015 or before	71	(38)
	2016	43	(23)
	2017	50	(27)
	2018 or after	23	(12)
Gestational age, weeks, mean, SD		39.6	1.1
Birthweight, grams, mean, SD		3505	420
Maternal age at childbirth, year, mean, SD		31.1	4.4
Nulliparous		32	(19)
Maternal education
	<4 year degree	43	(24)
	4 year degree	49	(27)
	Graduate degree	90	(49)
Total household income
	<$50,000	31	(18)
	$50,000–99,999	52	(31)
	≥$100,000	85	(51)

Detection frequencies for most phthalate metabolites were high (>80%), with MnBP MBzP, and MECPP universally detected (sTable 4) and MiBP, MEHHP and MEOHP nearly universally detected (>99%). Metabolites of replacement plasticizers DINCH (67–68%) and DEHTP (>97%) were also widely detected. DiNP metabolites were infrequently detected (<10%). When examining samples by age of collection, detection frequencies of MCPP, MEHP, and MCMHP were lower between 0–6 months (68%, 61%, and 68%, respectively) compared to remaining age groups (>82%). DINCH metabolites, MHNCH and MCOCH, were infrequently detected between 0–6 months (7 and 14%, respectively) but detected in most samples (>50%) in older age groups. DEHTP metabolites, MEHHTP and MECPTP, were universally detected in all age groups except MEHHTP between 0–6 months (82%). Metabolite and molar sum levels by demographic characteristics are shown in Supplementary Table 5.

Metabolites of phthalates from different parent compounds were low to moderately correlated after adjustment for SG (Table 2), ranging from ρ = 0.07 to 0.52, whereas correlations between metabolites of phthalates and replacement plasticizers were low (ρ < 0.25). Metabolites from the same parent compound (DEHP, DINCH, or DEHTP) were moderately to strongly correlated (ρ = 0.52– 0.96).

Table 2.

Spearman correlation coefficients of specific gravity adjusted phthalate and replacement plasticizer metabolite concentrations^*

Metabolite1	MnBP	MCPP	MiBP	MEP	MBzP	MEHP	MECPP	MEHHP	MEOHP	MCMHP	MCOCH	MHNCH	MECPTP	MEHHTP
MnBP	1.00	0.42	0.52	0.29	0.47	0.20	0.46	0.37	0.43	0.25	-0.04	0.00	0.17	0.07
MCPP		1.00	0.26	0.35	0.29	0.27	0.43	0.34	0.37	0.33	0.07	0.11	0.20	0.12
MiBP			1.00	0.23	0.36	0.16	0.44	0.38	0.43	0.25	0.03	0.01	0.14	0.09
MEP				1.00	0.32	0.12	0.37	0.28	0.31	0.26	0.10	0.10	0.16	0.11
MBzP					1.00	0.07	0.40	0.34	0.36	0.22	0.00	0.06	0.17	0.12
MEHP						1.00	0.49	0.48	0.49	0.43	0.09	0.06	0.18	0.14
MECPP							1.00	0.88	0.92	0.73	0.07	0.07	0.23	0.17
MEHHP								1.00	0.96	0.86	0.12	0.16	0.20	0.24
MEOHP									1.00	0.82	0.06	0.09	0.18	0.19
MCMHP										1.00	0.19	0.19	0.17	0.24
MCOCH											1.00	0.85	0.28	0.34
MHNCH												1.00	0.27	0.37
MECPTP													1.00	0.84
MEHHTP														1.00

^*Borders indicate common metabolites of a parent phthalate compound.

¹Metabolite concentrations adjusted for specific gravity with values below the LOD imputed as LOD/√2.

Levels of SG-adjusted MnBP, MCPP, and MiBP were inversely associated with age at collection (p_trend < 0.01, Figure 2; sTable 6). These metabolites, as well as MEP and MBzP had highest SG-adjusted concentrations in the first year (or first 6 months) followed by more consistent exposure levels at older ages. For example, mean SG-adjusted levels of MnBP and MiBP were nearly twice as high between 0–6 months (24.6 and 26.8 ug/L, respectively) compared to samples from children over 1 year (range: 12.0–14.6 ug/L and 10.1–12.2 ug/L). For DEHP metabolites, MECPP was inversely associated with age (p_trend = 0.05), with all other metabolites and the molar sum (∑DEHP) relatively constant across age groups (sTable 7). MCOCH and MHNCH concentrations were lower between 0 and 6 months (0.03 ng/mL and 0.16 ng/mL, respectively) and relatively consistent across remaining age groups (range: 0.13 −0.28 ng/mL and 0.61–1.10 ng/mL). MECPTP and MEHHTP concentrations were generally highest between 6 months and 3 years. We explored whether calendar year at the time of urine collection influenced metabolite concentrations in a linear mixed model with age groups and calendar year (discrete indicators for each age-group and calendar year, with 2019–2020 condensed to a single group due to small numbers) included as dependent variables of metabolite concentrations and found no association between calendar year of urine collection and metabolite concentrations (p-value > 0.3), indicating age-related changes are not due to trends in metabolite concentrations over time.

Figure 2.

Box and whisker plots of specific gravity adjusted concentrations by age group of non-DEHP phthalate metabolites (A), metabolites of DEHP and molar sum (B), metabolites of DINCH and molar sum (C), and metabolites of DEHTP and molar sum (D) from 438 urine samples collected among 187 participants in the UNC BCP

¹Metabolite concentrations adjusted for specific gravity using following formula: SGmeti = meti * (GM(SGi) – 1 / SGi – 1)

²Estimated using a linear mixed model to account for multiple samples per participants with log-transformed molar sum concentrations.

³Molar sums concentrations calculated by summing specific gravity adjusted metabolite concentrations (with values below the LOD imputed as LOD/√ 2) after dividing by the molecular weight.

∑DEHPm = molar sum of MEHP, MECPP, MEHHP, and MEOHP

∑DINCHm = molar sum of MCOCH and MHNCH

∑DEHTPm = molar sum of MECPTP and MEHHTP

ICCs for phthalate metabolites and replacement plasticizers ranged from ρ = 0.10–0.48 among participants with multiple samples (n = 122 participants, 373 samples), indicating low to moderate reliability and a high degree of within-individual variation in exposure in infancy and early childhood. Increasing age was generally associated with higher estimates of reliability, but this pattern was not seen across all metabolites (Table 3).

Table 3.

Intraclass correlation coefficients (ICCs) of specific gravity adjusted phthalate and replacement plasticizer metabolite concentrations overall and by age groups among participants with multiple samples in the UNC BCP (N = 122).

	Overall	0–1 years	>1–3 years	>3 years
	(n = 122 participants, 373 samples)	(n = 26 participants, 73 samples)	(n = 53 participants, 155 samples)	(n = 45 participants, 102 samples)
Metabolite/molar sum1,2	ICC (95% CI)3	ICC (95% CI)3	ICC (95% CI)3	ICC (95% CI)3
MnBP	0.34 (0.24, 0.46)	0.32 (0.13, 0.60)	0.23 (0.09, 0.47)	0.42 (0.23, 0.62)
MCPP	0.10 (0.03, 0.26)	0.0 (0.0, 1.00)	0.15 (0.05, 0.39)	0.37 (0.18, 0.62)
MiBP	0.41 (0.31, 0.53)	0.56 (0.35, 0.75)	0.34 (0.18, 0.54)	0.47 (0.28, 0.68)
MEP	0.46 (0.35, 0.56)	0.39 (0.17, 0.67)	0.51 (0.37, 0.66)	0.59 (0.40, 0.75)
MBzP	0.48 (0.37, 0.59)	0.46 (0.23, 0.70)	0.46 (0.30, 0.63)	0.62 (0.45, 0.77)
∑DEHPm	0.35 (0.24, 0.48)	0.37 (0.15, 0.67)	0.36 (0.21, 0.55)	0.50 (0.30, 0.69)
∑DINCHm	0.22 (0.13, 0.35)	0.01 (0.00, 1.00)	0.26 (0.12, 0.47)	0.38 (0.19, 0.60)
∑DEHTPm	0.18 (0.09, 0.33)	0.0 (0.0, 1.00)	0.21 (0.08, 0.43)	0.22 (0.06, 0.54)

¹Modeled as log-transformed metabolite concentrations adjusted for specific gravity with values below the LOD imputed as LOD/√2.

²Molar sums calculated by summing specific gravity adjusted metabolite concentrations (with values below the LOD imputed as LOD/√ 2) after dividing by the molecular weight.

³ICCs and 95% confidence intervals calculated using SAS ICC9 macro (https://www.hsph.harvard.edu/donna-spiegelman/software/icc9/).

∑DEHPm = molar sum of MEHP, MECPP, MEHHP, and MEOHP

∑DINCHm = molar sum of MCOCH and MHNCH

∑DEHTPm = molar sum of MECPTP and MEHHTP

When examining ratios of SG-adjusted metabolite concentrations of common parent compounds (Figure 3, sTable 8), we found the ratio of MECPP to other DEHP metabolites decreased with age (p_trend ≤ 0.01), with the ratio of MECPP:MEHP in the first 6 months (14.2) nearly double compared to remaining age groups (range: 5.05–8.73). Similarly, MECPP:MCMHP, which compares carboxylated metabolites across hexyl and ethyl oxidative pathways, was substantially higher in the first 6 months (43.6) compared to the remaining age groups (range: 5.38–8.67). For DEHTP metabolites, we also found an increased ratio of the carboxylated metabolite (MECPTP:MEHHTP) in the first 6 months of life—nearly three times the level (67.4) as compared to the remaining age groups (range: 12.0–24.4). We saw no age-related variability in the ratio of DINCH metabolites

Figure 3.

Box and whisker plots of ratios of specific gravity adjusted metabolite concentrations by age group of secondary oxidative metabolites of DEHP compared to hydrolysis metabolite (A), across secondary oxidative metabolites of DEHP (B), and carboxylated to oxidative metabolites of DINCH and DEHTP (C) from 438 urine samples collected among 187 participants in the UNC BCP.

¹Ratios calculated using specific gravity adjusted metabolite concentrations: SGmeti = meti * (GM(SGi) – 1 / SGi – 1), with values below the LOD imputed as LOD√2.

²p-values from tests of trend were calculated using continuous age groups categories as the independent variable in a linear mixed model with the ratio as the dependent variable.

4. Discussion

We report urinary concentrations of phthalate metabolites and replacement plasticizers from infants and young children enrolled in a longitudinal study between 2017 and 2020. We found detection frequencies of most phthalate metabolites were high (>80%) and metabolites of replacement plasticizers DINCH (58–60%) and DEHTP (>97%) were also widely detected. Metabolites of DiNP were infrequently detected (<10%). We found that SG-adjusted concentrations of MnBP, MCPP, MiBP, MEP, and MBzP were inversely associated with age, and usually highest in the first year of life. For DEHP and DEHTP, the ratio of carboxylated metabolites, MECPP and MECPTP, respectively, to other common metabolites, was substantially higher in the first 6 months of life, indicating potential differences in metabolism according to age.

While information from nationally representative biomonitoring data is lacking, several small cohort studies have been conducted to measure concentrations of metabolites of phthalate and replacement plasticizers from urine samples collected in infancy and early childhood (sTable 9). Previous studies of non-hospitalized US children under 5 years of age have found widespread exposure to a range of phthalate metabolites (Balalian et al., 2019; Brock et al., 2002; Watkins et al., 2014). Studies in other high-income nations corroborate findings of widespread exposure to phthalates in infancy and early childhood (Arbuckle et al., 2016; Becker et al., 2009; Carlstedt et al., 2013; Chen et al., 2017; Enke et al., 2013; Frederiksen et al., 2014; Huang et al., 2015; Kim et al., 2017; Lin et al., 2011; Liu et al., 2020; Navaranjan et al., 2020; Polanska et al., 2014; Völkel et al., 2014). However, much of the extant literature stems from studies with sample collection before 2010, which may not accurately reflect the current profile of phthalate exposure as many compounds have been phased out and replaced in recent years (Engel et al., 2021).

Additionally, neonates in hospitalized settings have been shown to be highly exposed to some phthalates, particularly DEHP, due to the presence of these chemicals in medical devices (Calafat et al., 2004; Latini et al., 2010). Comparisons of metabolite concentrations in participants based on their estimated intensity/duration of medical equipment use have found up to 14x higher levels of DEHP metabolites in the high exposure group (Green et al., 2005; Weuve et al., 2006). Bernard et al. (2023) found that DEHP metabolites among neonates were 16x lower after discharge compared to during their stay in the NICU (Bernard et al., 2023).

While early childhood phthalate exposure has been extensively studied, information on replacement plasticizers is sparse. Hammel et al. (2019) found widespread exposure to DINCH (>85%) and DEHTP metabolites (100%) among US children aged 3–6 months in samples collected between 2014–2016 (Hammel et al., 2019). Frederikson et al. (2022) found similarly high levels of exposure to DINCH and DEHTP from samples collected between 2016 and 2018 in Danish children aged 0–1 years. Similarly, we detected DINCH metabolites in most samples, and DEHTP metabolites in nearly all samples.

Population-based biomonitoring of children under 3 years is not available, but above age 3, NHANES biomonitoring in the general US population indicates younger age groups typically have higher concentrations of phthalate metabolites, such as MBP, MiBP, MEHP, MECPP, MEHHP, MEOHP, and MCOP (Centers for Disease Control and Prevention, 2019). Few US studies have performed longitudinal urine sampling in infancy and early childhood. Watkins et al. (2014) found that MEP decreased and MCOP increased between 1 and 5 years of age but observed few other age-related trends in metabolite levels, although these values were not adjusted for urinary dilution (Watkins et al., 2014). Similarly, Balalian et al. (2019) reported relatively consistent concentrations for all metabolites measured at 3 and 5 years (Balalian et al., 2019). In contrast, we found higher concentrations of MnBP, MCPP MiBP, MEP, and MBzP in infancy compared to later years. Our ability to identify age-related trends may be due in part to the accelerated cohort design of BCP, where children were enrolled over multiple years at different baseline ages, and then followed forward for up to 24 months (Howell et al., 2019). This design allowed us to better separate age from birth cohort effects on exposure.

Studies in other countries have produced mixed results regarding age-related trends. Lin et al. (2011) found higher concentrations of MnBP and DEHP metabolites at 2 years compared to 5 years in Taiwanese children (Lin et al., 2011). However, others report that concentrations of phthalates generally increased with age across infancy and early childhood (Carlstedt et al., 2013; Frederiksen et al., 2014; Kim et al., 2017; Navaranjan et al., 2020; Völkel et al., 2014). For example, Carlstedt et al. (2013) found higher concentrations of MEOHP and MEHHP at 6 months compared to 2 months, but lower concentrations of MEP, among Swedish children (Carlstedt et al., 2013), and Frederikson et al. (2014) found increasing urinary concentrations of MBP, MiBP, MEP, and MBzP, in Finnish infants between birth and 14 months (Frederiksen et al., 2014). These higher concentrations are often attributed to changes in dietary patterns as mixed diets are introduced, or developmental changes that lead to increased contact with environmental sources of exposure to phthalates (Carwile et al., 2022; Hammel et al., 2019; ‘t Mannetje et al., 2021; Watkins et al., 2014). While previous studies have indicated that exposure to phthalates from breastfeeding is likely low (Fromme et al., 2011; Hines et al., 2009; Kim et al., 2020), a study by Frederikson et al. (Frederiksen et al., 2022) showed that exclusively breast-fed infants had higher exposure to phthalates compared to time periods after the introduction of a mixed diet.

Differences in profiles of DEHP metabolites between infants/neonates as compared to adult populations has been previously reported (Enke et al., 2013; Frederiksen et al., 2014). Enke et al. (2013) compared metabolite ratios between newborns aged 2–5 days and pregnant women, finding MECPP:MEHHP and MECPP:MEOHP were over 5x higher in newborns (Enke et al., 2013). Frederickson et al. (2014) similarly found that MECPP decreased across the first year of life whereas the proportion of secondary hexyl metabolites (MEHHP and MEOHP) increased (Frederiksen et al., 2014). These metabolic differences are attributable to the immaturity of infant metabolism resulting in more limited detoxification abilities. DEHP metabolism is a two-step process of initial hydrolysis to MEHP and secondary oxidation performed by CYP2C9 and CYP2C19. In-vitro studies show that both CYP2C9 and CYP2C19 mediate oxidation of MEHP to MEHHP while only CYP2C9 is involved in oxidation to MECPP (Choi et al., 2012). Th higher proportion of MECPP metabolites in infancy likely due to their age-dependent expression and functional activity, with studies showing higher expression and catalytic efficient of CYP2C19 and lower expression of CYP2C9 for infants (Zane et al., 2018). We also found a higher proportion of MECPP relative to other metabolites in the first year of life, and a much higher MECPP:MCMHP ratio in the first 6 months indicating an age-based change in the preference for the hexyl oxidative pathway compared to the ethyl oxidative pathway. We also report an increased ratio of carboxylated metabolites of DEHTP (MECPTP:MEHHTP) in the first 6 months of life. Furthermore, our results indicate that these patterns do not persist beyond the first year of life, as ratios of carboxylated metabolites to remaining common metabolites stabilized after the first year.

Strengths of this study include longitudinal collection of samples across infancy and early childhood, allowing examination of age-related trends in metabolite concentrations. Most participants (66%) had multiple samples, including over 10% of participants with >5 samples, allowing examination of reliability across age groups. In addition, this study’s recent enrollment period (2017–2020) provides information on more recent exposure patterns of phthalates and replacement plasticizers. Our study also has limitations. The UNC BCP was designed to characterize normative brain development, and due to study protocol constraints, most participants lived in close proximity to UNC. Thus, enrollment tended to over-represent white participants of higher socioeconomic and educational backgrounds. The lack of racial/ethnic and socioeconomic diversity in this study population negatively impacts the generalizability of our results to more diverse populations, as people of color (Welch et al., 2023) and children in lower income families (Navaranjan et al., 2020), have been shown to have higher exposure to several phthalates.

5. Conclusions

We found widespread exposure to a range of phthalates and replacement plasticizers across infancy and early childhood, with the highest concentrations of several metabolites observed in the first year of life. Moreover, we found evidence that metabolism of DEHP and DEHTP may differ in the first six months, indicated by higher proportions of carboxylated metabolites. Future studies should characterize the nature and impact of phthalate metabolism differences in early life.