Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Jul 7.
Published in final edited form as: Environ Res. 2026 Jun 6;305(Pt 1):124940. doi: 10.1016/j.envres.2026.124940

Prenatal exposure to organophosphate ester flame retardants and plasticizers and maternal immune responses in three ECHO cohorts

Jiwon Oh a, Judy Van de Water b,c, Thomas G O’Connor d, Kristin Scheible e,f, Anne P Starling g,h,i, Dana M Dabelea h,i, Raina N Fichorova j, Rebecca J Schmidt a,c, Irva Hertz-Picciotto a,c, Deborah H Bennett a
PMCID: PMC13333603  NIHMSID: NIHMS2191787  PMID: 42250839

Abstract

Organophosphate ester (OPE) flame retardants and plasticizers may alter maternal immune responses; yet epidemiologic evidence in pregnant populations remains limited. We pooled data from 410 pregnant participants across three Environmental influences on Child Health Outcomes (ECHO) Cohort sites. During mid- to late-pregnancy, spot urine was collected for OPE metabolite quantification, followed by plasma/serum collection for immune markers on the same day or thereafter. Twelve immune markers available in at least two sites were included. We estimated associations between log2-transformed OPE metabolite concentrations and log2-transformed, site-specific z-scored immune marker levels using linear mixed-effects regression with site as a random effect for pooled analyses, and linear regression for site-specific analyses. We conducted exploratory analyses to evaluate effect modification by race/ethnicity, pre-pregnancy obesity, and fetal sex. We observed associations between bis(1,3-dichloro-2-propyl) phosphate and elevated interleukin (IL)-8 levels in the pooled sample (β per doubling=0.06, 95% confidence interval [CI]: 0.02–0.11) and in both contributing sites. Associations between other OPE metabolites and immune markers were not consistently observed across study sites. Participants with a race/ethnicity other than non-Hispanic White showed stronger associations of diphenyl phosphate (β=0.16, 95% CI: 0.04–0.28) and bis(2-chloroethyl) phosphate (β=0.04, 95% CI: −0.01–0.09) with higher tumor necrosis factor-α. Sex-dimorphic associations for IL-6 were observed, with directions varying by OPE metabolite. This U.S. multi-site study suggests certain OPE exposures may be positively associated with maternal IL-8 levels. Such immune responses could alter the intrauterine environment, warranting further investigation into the downstream effects on child health, given the limited current evidence.

Keywords: organophosphate esters, pregnancy, cytokines, chemokines

1. INTRODUCTION

Organophosphate esters (OPEs) became commonly applied as flame retardants and plasticizers after polybrominated diphenyl ethers (PBDEs) were phased out (Blum et al. 2019). OPEs are used in a variety of consumer products and industrial processes, including building materials, textiles, electronics, food packaging, and children’s products (Huang et al. 2025). Human exposure is widespread via dietary intake, dust ingestion, inhalation, and dermal contact (Hou et al. 2016; Wei et al. 2015). Consequently, several OPE metabolites have been frequently detected in the urine of both the U.S. general (Huang et al. 2024; Ospina et al. 2018) and pregnant populations (Oh et al. 2024). OPE exposure during pregnancy is of particular concern because it can affect the pregnant woman and the developing fetus, with potential long-term health impacts (Doherty et al. 2019; Li et al. 2023).

Exposure to OPEs has been associated with disrupted endocrine function (Hu et al. 2023), as well as altered immune responses and inflammation (Huang et al. 2025). Experimental studies have shown that OPE exposure can lead to cytokine dysregulation and inflammatory responses. In mice, increased expression of inflammatory cytokines has also been observed in the brain following exposure to triphenyl phosphate (TPHP) (Liu et al. 2020) and to tris(1,3-dichloro-2-propyl) phosphate (TDCPP) (Zhong et al. 2020), in the lung following exposure to tri-n-butyl phosphate (TnBP) and tris(2-butoxyethyl) phosphate (TBOEP) (Meng et al. 2022), and in the intestine after exposure to TPHP (Peng et al. 2023). More relevant to the prenatal window, prenatal exposure to an environmentally relevant OPE mixture in rats altered cytokine and chemokine signaling in dams and offspring, particularly in females, and reduced splenic monocyte-macrophages in male offspring (Giles et al. 2025). Maternal TPHP exposure in mice increased placental expression of TNF- α and nuclear factor kappa B (NF-κB) and reduced learning and memory in male offspring (Hong et al. 2022).

Epidemiologic evidence regarding associations between OPE exposure and immune markers in pregnant populations remains sparse. In the only recent study of 65 pregnant women, diphenyl phosphate (DPHP) was associated with elevated maternal serum levels of C-C chemokine ligand 4 (CCL4) and C-X-C chemokine ligand 8 (CXCL8, also known as interleukin [IL]-8), while conversely showing an inverse association with CCL3 in cord blood (Nguyen et al. 2025). Given existing evidence in pregnant populations linking OPE exposure to oxidative stress, a possible trigger of inflammation (Ait Bamai et al. 2019; Yao et al. 2021), and linking other flame retardants and plasticizers, such as PBDEs, phthalates, and phenols, to inflammatory cytokines (Ferguson et al. 2016; Ferguson et al. 2014; Kelley et al. 2019; Nguyen et al. 2025; Watkins et al. 2015; Zota et al. 2018), these associations warrant further investigation in larger-scale studies.

The present study examined associations between urinary OPE metabolite concentrations and immune markers measured in serum or plasma in pregnant participants from three Environmental influences on Child Health Outcomes (ECHO) sites. We also assessed whether these associations were modified by maternal race and ethnicity, pre-pregnancy obesity, and fetal sex. As changes in maternal immune profiles during pregnancy can affect maternal health, fetal development and subsequent child health (Goldstein et al. 2020; Racicot et al. 2014; Yockey and Iwasaki 2018), understanding the immunological effects of prenatal OPE exposure is critical.

2. METHODS

Study Population

This study analyzed data from the ECHO Cohort, funded by the National Institutes of Health (NIH), which combines ongoing and new pregnancy and pediatric cohort sites across the United States (U.S.) (Knapp et al. 2023). We obtained access to the de-identified ECHO-wide Cohort–3rd Release data through the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) Data and Specimen Hub (DASH). Study protocols received approval from either the single ECHO Institutional Review Borad (IRB) or the site’ local IRBs, and participants provided written informed consent prior to data collection. Details on enrollment, data collection protocol, and data harmonization are available elsewhere (Gillman and Blaisdell 2018; Knapp et al. 2023).

Among the initial 4,808 ECHO pregnant participants whose urinary OPE metabolite concentrations were measured using the OPE panel (Oh et al. 2024), we first excluded participants missing a urine collection date (n=193). We then excluded participants who lacked immune marker data measured in serum or plasma (n=3,677). To ensure temporality, we further excluded participants whose blood samples for immune markers were collected prior to their urine sample, restricting the sample to 616 participants (821 samples). To account for the acute nature of immune markers, we excluded all but the one blood sample collected closest to the urine collection date for participants with multiple measurements (616 samples). After excluding one site with a small sample size (n=2), we excluded participants who lacked a complete set of immune marker measurements (n=204). This resulted in a final analytic sample of 410 pregnant participants from three ECHO sites: 164 from Healthy Start, 74 from MARBLES/EARLI (Markers of Autism Risk in Babies—Learning Early Signs/Early ASD Risk Longitudinal Investigation), and 172 from Rochester (Figure S1).

The Rochester site (“ECHO-UPSIDE”) recruited women with normal-risk pregnancies before 14 weeks’ gestation between 2015 and 2019 at outpatient obstetric clinics affiliated with the University of Rochester (O’Connor et al. 2021). Additional eligibility criteria included age 18 years or older, singleton pregnancy, no known substance abuse problems or history of psychotic illness, and English proficiency. Maternal urine and blood samples included in this study were collected during gestation at 18‒35 weeks (median: 21) and 18‒35 weeks (median: 21), respectively, with nearly 98% of paired samples obtained on the same day.

The MARBLES study enrolled pregnant women in Northern California during 2006–2018 (Hertz-Picciotto et al. 2018), and the EARLI study recruited women during 2009–2012 across multiple sites, including Southeast Pennsylvania, Northeast Maryland, and Northern California (Newschaffer et al. 2012). Both are autism enriched-risk pregnancy cohorts, recruiting pregnant women with a prior biological child diagnosed with autism or whose current pregnancy has an older half-sibling. Other inclusion criteria were age 18 years or older, ability to communicate in English, and residence within 2.5 hours of study clinics. Maternal urine and blood samples included in this study were collected during gestation at 16‒35 weeks (median: 28) and 17‒39 weeks (median: 30), respectively, with half of the samples collected on the same day.

The Healthy Start study recruited pregnant women prior to 24 weeks of gestation from obstetric clinics at the University of Colorado Hospital between 2009 and 2014 (Friedman et al. 2021). Exclusion criteria included age under 16 years, multiple gestations, and a history of stillbirth or extremely preterm birth (<25 weeks’ gestation), as well as pre-existing diabetes, asthma, cancer, or psychiatric illness. Maternal urine and blood samples included in this study were collected at the same study visit between 20‒33 weeks’ gestation (median: 26).

Measurements of Urinary OPE Metabolites

A single maternal urine sample per participant collected during the second or third trimester was shipped on dry ice by each site to the Human Health Exposure Analysis Resource (HHEAR) laboratory at New York University Grossman School of Medicine. Nine OPE metabolites were quantified using analytical methods described in detail elsewhere (Oh et al. 2024; Wang et al. 2019). In brief, 0.5 mL urine aliquots were spiked with 1 ng of deuterated internal standard mixtures and extracted using solid-phase extraction cartridges (Phenomenex). OPE metabolites were then identified and quantified by high-performance liquid chromatography (ExionLC system; SCIEX), coupled to an AB SCIEX QTRAP 5500+ triple quadrupole mass spectrometer (Applied Biosystems).

The nine quantified OPE metabolites were bis(butoxyethyl) phosphate (BBOEP; a metabolite of TBOEP), bis(2-chloroethyl) phosphate (BCETP; a metabolite of tris(2-chloroethyl) phosphate), bis(1-chloro-2-propyl) phosphate (BCPP; a metabolite of tris(1-chloro- 2-propyl) phosphate), bis(1,3-dichloro-2-propyl) phosphate (BDCPP; a metabolite of TDCPP), bis(2-ethylhexyl) phosphate (BEHP; a metabolite of a metabolite of tris(2- ethylhexyl) phosphate), bis(2-methylphenyl) phosphate (BMPP; a metabolite of tris(2-methylphenyl) phosphate), a composite of dibutyl phosphate and di-isobutyl phosphate (DBUP/DIBP; metabolites of tributyl phosphate and its isomer tri-isobutyl phosphate), DPHP (a metabolite of TPHP), and dipropyl phosphate (DPRP; a metabolite of tripropyl phosphate). DBUP/DIBP was reported as a composite due to co-elution, which precluded separate quantification. Limits of detection (LODs) ranged from 0.01 to 0.04 ng/mL by metabolite, and values below the LOD were imputed as LOD divided by the square root of two (Hornung and Reed 1990). Six OPE metabolites detected in more than 60% of the overall pooled sample were included in the statistical analyses. This included BCETP and BCPP despite its 56–57% detection in one site (Rochester), as it exceeded the 60% cutoff in the other sites and in the overall sample.

For quality control (QC), HHEAR urine QC pools, standard reference materials (SRM 3672 and SRM 3673; National Institute of Standards and Technology), and urine pool samples spiked with reference standard were analyzed in every batch (Kannan et al. 2021). Trace-level signals detected in procedural blanks were subtracted from all samples. Mean recoveries from spiked pools ranged from 70% to 133%, with coefficients of variation (CVs) of 9–19% (Oh et al. 2024). CVs for the HHEAR urine QC pools and SRMs ranged from 12% to 31%, except for BBOEP in SRM 3672 (40%). Additionally, masked duplicate aliquots (i.e., 191 selected from the full sample set analyzed using the OPE panel) were tested, with the medians of relative percent differences ranging from 6% to 27% for six OPE metabolites detected in at least 60% of samples (Oh et al. 2024).

Measurements of Blood Immune Markers

Immune marker sets, laboratories, and measurement timing differed across the sites. To enable multi-site analyses, we included twelve immune markers measured by at least two sites: inflammatory cytokines (tumor necrosis factor alpha [TNF-α] and interleukin [IL]-6–both measured in all three sites–and IL-1β–measured in two), an inflammatory chemokine (IL-8), T-helper type 1 (Th1)-related cytokines (IL-2, IL-12p70), T-helper type 2 (Th2)/regulatory cytokines (IL-4, IL-5, IL-10), hematopoietic/growth factors (IL-7, granulocyte macrophage colony-stimulating factor [GM-CSF]), and an acute phase protein (C-reactive protein [CRP]). LODs varied by panel and batch. Marker values below the LOD were imputed as one-half the LOD (Uh et al. 2008). All immune markers were detected in more than 70% of samples in each site, except for IL-2 (66%) and IL-5 (59%) in MARBLES/EARLI.

The Rochester site measured maternal plasma cytokines and chemokines at the University of Rochester Medical Center using the Milliplex MAP Magnetic Bead Panel (Millipore) (Serrano et al. 2024). In MARBLES and EARLI, maternal plasma cytokines and chemokines were quantified at the University of California, Davis with multiplex bead-based kits (Millipore MILLIPLEX MAP Human Cytokine/Chemokine Kit) (Volk et al. 2020). The Healthy Start study measured three immune markers in maternal serum collected during the mid-pregnancy visit at the University of Colorado Clinical and Translational Sciences Institute Core Laboratories: TNF-α using enzyme-linked immunosorbent assay (R&D Systems), IL-6 using Luminex MAP technology (R&D Systems), and CRP using immunoturbidimetric methodology (Beckman Coulter) (Friedman et al. 2021). All assays were performed according to the manufacturer’s protocols.

Covariates

Covariates were selected a priori based on a directed acyclic graph established using DAGitty (Textor et al. 2016) and on data availability across the three sites (Figure S2). The adjustment set for the models evaluating associations between OPE exposure and immune markers included maternal education (high school/some college or less, bachelor’s/graduate/professional degree), age at delivery (years), pre-pregnancy body mass index (BMI; kg/m2), parity (nulliparous, parous), and race/ethnicity (self-reported; non-Hispanic White, other [including Hispanic, non-Hispanic Black, Asian, and multiple/other races]). We grouped participants who were not non-Hispanic White into a single racial/ethnic minority category to avoid small cell counts within each site and to account for social contexts that may influence environmental exposures and immune responses. Additionally, fetal sex (male, female), and gestational week at blood sample collection (weeks), and immune marker batch ID were included as precision variables because they can influence immune marker concentrations. Information regarding preeclampsia, gestational diabetes, and maternal infections (including influenza, sexually transmitted infections, genitourinary tract infections, pneumonia, skin infections, and toxoplasmosis) was abstracted from medical records or obtained via self-report. Although these conditions may act as mediators, their temporal relationship to sample collection varied across participants. Therefore, they were included only in sensitivity analyses.

Statistical Analysis

Participant characteristics, OPE metabolite concentrations, and immune marker levels were summarized by site. Baseline characteristics of included and excluded participants were compared using chi-squared tests. Before statistical analyses, OPE metabolite concentrations were corrected for urinary dilution using specific gravity or creatinine (Kuiper et al. 2022). For each site, metabolite concentrations were multiplied by the ratio of the site-specific median dilution value to the individual sample’s dilution value; for specific gravity, the value of specific gravity minus one was used when computing this ratio. Spearman correlation coefficients among dilution-adjusted OPE metabolites were calculated by site.

Immune marker levels, as well as the correlations among these markers, varied by site, reflecting differences in participant characteristics and limited convergence across analytical platforms. Therefore, to enable multi-site analyses and to reduce bias from technical heterogeneity, immune markers were log2-transformed and standardized (z-scored) within each cohort prior to pooling (Maaniitty et al. 2024). This approach prioritized the relative distribution within each site (Breen et al. 2011), while mitigating the impact of platform-specific measurement scales. Dilution-adjusted OPE metabolite concentrations were log2-transformed to satisfy the normality of model residuals and attenuate the influence of outliers. We first conducted pooled analyses using linear mixed-effects regression models with a random intercept for site to estimate associations between each OPE metabolite and each immune marker. We then estimated site-specific associations using linear regression models. Missing covariate data (<10% within each site) were imputed by site using multiple imputation by chained equations, including exposures, outcomes, and covariates in the imputation models. Effect estimates from 20 imputed datasets were combined using Rubin’s rules (Rubin 1987).

A series of sensitivity analyses were conducted to assess the robustness of the findings from pooled and site-specific analyses. First, we additionally adjusted for any infections during pregnancy in the primary models. Second, we repeated the primary analyses after excluding participants with preeclampsia and/or gestational diabetes. Third, we dichotomized the immune markers at the site-specific 75th percentile and categorized them as ‘high’ or ‘low’. Utilizing the 75th percentile ensured sufficient case counts in smaller cohorts (i.e., MARBLES/EARLI) when investigating associations with elevated inflammatory profiles. The dichotomized variables were then modeled them as outcomes using Poisson mixed-effects models with a random intercept for site for pooled analyses, or Poisson regression models for site-specific analyses.

For the two immune markers measured in all three sites (TNF-α and IL-6), we conducted exploratory analysis of effect modification by maternal race and ethnicity (non-Hispanic White vs. other), pre-pregnancy obesity (BMI ≥30 vs. <30), and fetal sex (male vs. female) in the pooled sample, given prior evidence that these factors modify associations of other flame retardants and endocrine disrupting chemicals (EDCs) with immune markers in pregnant women (Han et al. 2025; Jin et al. 2023; Lee et al. 2023; Siwakoti et al. 2024; Zota et al. 2018). We stratified the pooled analyses by each potential effect modifier and tested interaction terms between each OPE metabolite and the modifier.

Statistical significance was defined as a raw p-value <0.05 for all primary, site-specific, and effect modification analyses. To account for multiple testing, we applied the Benjamini–Hochberg false discovery rate (FDR) procedure, with q-values <0.20 regarded as statistically significant. All statistical analyses were conducted using R version 4.1.3 (R Foundation for Statistical Computing).

3. RESULTS

Participant Characteristics

MARBLES/EARLI participants notably differed from those in Healthy Start and Rochester sites across several characteristics (Table 1). Compared with participants in Healthy Start and Rochester, MARBLES/EARLI pregnant participants were older (median 34.2 years vs. 30.3 and 30.8 years), had a higher prevalence of pre-pregnancy obesity (42% vs. 17% and 28%), were more likely to be parous (88% vs. 49% and 65%), were more often from racial/ethnic groups other than non-Hispanic White (46% vs. 30% and 33%), were more likely to have a high school/college education or less (52% vs. 34% and 35%), and more frequently experienced any infections during pregnancy (22% vs. 14% and 5%). Compared with the excluded ECHO participants, those included in the analytic sample were more likely to be parous, non-Hispanic White, have a bachelor’s degree or higher, and report no infections during pregnancy (Table S1).

Table 1.

Participant characteristics among pregnant participants from three ECHO sites.

Characteristics Frequency (%)a or median (25th, 75th percentiles)
All
(n=410)		 Healthy Start (n=164) MARBLES/EARLI
(n=74)		 Rochester
(n=172)
Maternal age at delivery (years) 31.1 (27.8, 34.3) 30.3 (26.5, 33.5) 34.2 (31.1, 37.4) 30.8 (27.8, 33.2)
 <30 166 (40) 78 (48) 13 (18) 75 (44)
 30–34 157 (38) 57 (35) 29 (39) 71 (41)
 ≥35 87 (21) 29 (18) 32 (43) 26 (15)
Pre-pregnancy body mass index (kg/m2) 25.5 (22.2, 30.6) 24.0 (20.7, 27.5) 27.7 (23.6, 33.2) 25.8 (23.1, 32.9)
 Normal/underweight 182 (45) 92 (56) 22 (30) 68 (41)
 Overweight 116 (29) 43 (26) 21 (28) 52 (31)
 Obese 107 (26) 29 (18) 31 (42) 47 (28)
 Missing 5 0 0 5
Parity
 Nulliparous 151 (37) 83 (51) 8 (12) 60 (35)
 Parous 254 (63) 81 (49) 61 (88) 112 (65)
 Missing 5 0 5 0
Maternal race/ethnicity
 Non-Hispanic White 268 (66) 114 (70) 40 (54) 114 (67)
 Other racesb 139 (34) 50 (30) 34 (46) 55 (33)
 Missing 3 0 0 3
Maternal education
 Less than high school/some college 145 (37) 55 (34) 36 (52) 54 (35)
 Bachelor’s/graduate/professional degree 244 (63) 109 (66) 33 (48) 102 (65)
 Missing 16 0 0 16
Any infections during pregnancy
 No 362 (92) 141 (86) 57 (78) 164 (95)
 Yes 46 (8) 23 (14) 16 (22) 8 (5)
 Missing 1 0 1 0
Preeclampsia
 No 383 (95) 157 (96) 63 (93) 163 (95)
 Yes 21 (5) 7 (4) 5 (7) 9 (5)
 Missing 6 0 6 0
Gestational diabetes
 No 388 (94) 151 (96) 65 (93) 166 (97)
 Yes 23 (6) 7 (4) 5 (7) 11 (3)
 Missing 10 6 4 0
Fetal sex
 Female 180 (44) 68 (41) 29 (39) 83 (48)
 Male 230 (56) 96 (59) 45 (61) 89 (52)
Gestational week at urine collection (week) 24 (21, 28) 27 (25, 29) 28 (24, 30) 21 (20, 22)
Gestational week at blood collection (week) 24 (21, 29) 27 (25, 29) 30 (27, 34) 21 (20, 22)
Open in a new tab
a

Percentage was calculated without missing observations.

b

Other races include Asian, native Hawaiian or other Pacific Islanders, American Indian or Alaska native, and multiple races.

Abbreviations: ECHO, Environmental influences on Child Health Outcomes.

OPE Metabolite Concentrations and Maternal Immune Marker Levels

Dilution-adjusted concentrations of most OPE metabolites were broadly comparable across sites (Table 2). DPHP and DBUP/DIBP were detected in nearly all samples (>98%), and DPHP and BDCPP had higher median concentrations than the other metabolites. MARBLES/EARLI participants had higher median BDCPP (1.47 vs. 0.91 ng/mL) and BCPP (0.56 vs. 0.33 ng/mL) concentrations compared with Healthy Start, and higher median BCETP concentrations compared with Rochester (1.03 vs. 0.40 ng/mL). Within each site, OPE metabolites were weakly correlated with one another (Spearman correlation coefficients [rsp]=−0.25 to 0.35) (Figure S3).

Table 2.

Distributions of dilution-adjusted OPE metabolite concentrations and immune markers among pregnant participants from three ECHO sites.

Healthy Start (n=164) MARBLES/EARLI (n=74) Rochester (n=172)
LOD % >LOD p50 (p25, p75) LODa % >LOD p50 (p25, p75) LOD % >LOD p50 (p25, p75)
OPE metabolites (ng/mL) b
DPHP 0.03 100 1.18 (0.73, 2.28) 0.03 99 1.33 (0.87, 2.00) 0.03 100 1.26 (0.78, 2.15)
DBUP/DIBP 0.04 94 0.16 (0.11, 0.22) 0.04 95 0.21 (0.14, 0.32) 0.04 91 0.23 (0.15, 0.35)
BDCPP 0.02 91 0.91 (0.55, 2.03) 0.02 91 1.47 (0.57, 2.31) 0.02 92 1.51 (0.64, 2.72)
BCETP 0.02 84 0.82 (0.29, 1.91) 0.02 91 1.03 (0.40, 2.83) 0.02 57 0.40 (<LOD, 1.34)
BBOEP 0.02 65 0.05 (<LOD, 0.08) 0.02 69 0.06 (<LOD, 0.12) 0.02 66 0.06 (<LOD, 0.14)
BCPP 0.02 62 0.33 (<LOD, 0.87) 0.02 64 0.56 (<LOD, 1.34) 0.02 56 0.40 (<LOD, 1.34)
Immune markers (pg/mL)
TNF-α 0.16 98 1.22 (0.87. 1.94) 0.70–3.61 100 18.0 (13.6, 26.1) 0.16 100 2.50 (1.89, 3.12)
IL-6 0.16 100 1.19 (0.93, 1.78) 0.90–2.60 81 6.20 (2.13, 12.0) 0.11 100 1.57 (1.10, 2.71)
IL-1β - - - 0.04–0.80 99 1.60 (1.10, 4.06) 0.14 100 1.32 (1.00, 1.94)
IL-8 - - - 0.40–2.40 100 12.4 (4.33, 19.5) 0.13 100 2.20 (1.64, 4.01)
IL-2 - - - 0.32–1.72 66 3.06 (0.50, 15.0) 0.19 100 2.74 (1.97, 3.84)
IL-12p70 - - - 0.60–2.32 99 10.2 (4.77, 15.3) 0.15 100 2.15 (1.64, 2.87)
IL-4 - - - 0.20–4.50 89 2.84 (2.25, 10.9) 1.12 100 16.1 (11.6, 26.4)
IL-5 - - - 0.20–3.12 59 1.56 (0.76, 2.83) 0.12 100 3.86 (2.86, 5.90)
IL-10 - - - 0.04–2.27 89 13.5 (7.68, 17.8) 0.56 100 11.5 (8.52, 15.9)
IL-7 - - - 0.72–1.88 96 7.82 (5.04, 11.5) 0.42 100 5.22 (3.99, 6.56)
GM-CSF - - - 0.68–7.50 72 20.9 (3.75, 131.6) 0.35 100 29.9 (18.4, 50.5)
CRP (mg/L) 0.20 100 4.28 (2.52, 7.09) - - - 0.001 100 17.0 (7.39, 33.1)
Open in a new tab
a

LODs for MARBLES/EARLI varied by batch.

b

Three OPE metabolites detected in less than 60% of samples were excluded from the table.

Abbreviations: CI, confidence interval; BBOEP, bis(butoxyethyl) phosphate; BCETP, bis(2-chloroethyl) phosphate; BCPP, bis(1-chloro-2-propyl) phosphate; BDCPP, bis(1,3-dichloro-2-propyl) phosphate; CRP, C-reactive protein; DBUP/DIBP, composite of dibutyl phosphate and di-isobutyl phosphate; DPHP, diphenyl phosphate; ECHO, Environmental influences on Child Health Outcomes; GM-CSF, granulocyte-macrophage colony-stimulating factor; IL, interleukin; LOD, limit of detection; OPE, organophosphate ester; p25, 25th percentile; p50, 50th percentile; p75; 75th percentile TNF-α, tumor necrosis factor alpha.

Several blood immune marker levels differed markedly across sites (Table 2); however, absolute levels were not directly compared due to differences in population characteristics and assay platforms. Within both the MARBLES/EARLI and Rochester sites, IL-6 and IL-8 were positively and strongly correlated (rsp=0.79), and GM-CSF, IL-2, IL-10, and IL-12p70 were positively and moderately correlated with one another (rsp=0.43–0.66) (Figure S4). Many other markers also showed positive, weak-to-moderate correlations.

Associations between OPE Metabolite Concentrations and Maternal Immune Marker Levels

In the pooled sample, higher BDCPP was associated with elevated IL-8 levels (β per doubling=0.06, 95% confidence interval [CI]: 0.02–0.11), although this association remained not significant after FDR correction (q-value=0.38) (Figure 1 and Table S2). This association was consistent across contributing sites (Rochester: β=0.06, 95% CI: 0.00–0.12, q-value=0.24; MARBLES/EARLI: β=0.09, 95% CI: 0.00–0.18, q-value=0.74), noting that IL-8 was not measured in Healthy Start (Figure 2 and Table S3). Furthermore, several positive associations in the pooled sample were found between BDCPP and IL-4 (β=0.06, 95% CI: 0.01–0.11), BBOEP and IL-7 (β=0.08, 95% CI: 0.00–0.16), and BCPP and IL-10 levels (β=0.04, 95% CI: 0.00–0.07) (all q-values >0.38).

Figure 1.

Figure 1.

Open in a new tab

Associations between dilution-adjusted, log2-transformed OPE metabolite concentrations and log2-transformed, scaled maternal immune marker levels in the pooled sample. Point estimates represent betas with 95% confidence intervals (error bars); detailed values are provided in Tables S1. Red colors indicate estimates with a raw p-value <0.05 (none of them had a false discovery rate-corrected q-value <0.2). Models were adjusted for gestational week at blood sample collection, batch ID, maternal age at delivery, pre-pregnancy body mass index, parity, maternal race/ethnicity, maternal education, and fetal sex.

Abbreviations: CI, confidence interval; BBOEP, bis(butoxyethyl) phosphate; BCETP, bis(2-chloroethyl) phosphate; BCPP, bis(1-chloro-2-propyl) phosphate; BDCPP, bis(1,3-dichloro-2-propyl) phosphate; CRP, C-reactive protein; DBUP/DIBP, composite of dibutyl phosphate and di-isobutyl phosphate; DPHP, diphenyl phosphate; GM-CSF, granulocyte-macrophage colony-stimulating factor; IL, interleukin; OPE, organophosphate ester; Th, T-helper; TNF-α, tumor necrosis factor alpha.

Figure 2.

Figure 2.

Open in a new tab

Site-specific associations between dilution-adjusted, log2-transformed OPE metabolite concentrations and log2-transformed, scaled maternal immune marker levels. Point estimates represent betas with 95% confidence intervals (error bars); detailed values are provided in Tables S1. Filled points indicate estimates with a raw p-value <0.05, and asterisks indicate estimates with a false discovery rate-corrected q-value <0.2. Models were adjusted for gestational week at blood sample collection, batch ID, maternal age at delivery, pre-pregnancy body mass index, parity, maternal race/ethnicity, maternal education, and fetal sex.

Abbreviations: CI, confidence interval; BBOEP, bis(butoxyethyl) phosphate; BCETP, bis(2-chloroethyl) phosphate; BCPP, bis(1-chloro-2-propyl) phosphate; BDCPP, bis(1,3-dichloro-2-propyl) phosphate; CRP, C-reactive protein; DBUP/DIBP, composite of dibutyl phosphate and di-isobutyl phosphate; DPHP, diphenyl phosphate; GM-CSF, granulocyte-macrophage colony-stimulating factor; IL, interleukin; OPE, organophosphate ester; Th, T-helper; TNF-α, tumor necrosis factor alpha.

In site-specific analyses, the association between BDCPP and GM-CSF in MARBLES/EARLI was the most robust (β=0.18, 95% CI: 0.07–0.28, q-value=0.11) (Figure 2 and Table S3). In the Rochester site, we observed several marginal trends (all q-values=0.24). For example, BDCPP was positively associated with IL-4, IL-6, and IL-8 (β range: 0.06 to 0.07), and BCPP with GM-CSF, IL-7, IL-10, and IL-12p70 (all βs=0.05). In contrast, DPHP and DBUP/DIBP were inversely associated with IL-1β and IL-2 (β range: −0.15 to −0.11). In Healthy Start, DPHP was associated with higher TNF-ɑ (β=0.11), while BBOEP was associated with lower TNF-ɑ (β=−0.11) (all q-values=0.45). However, none of these site-specific associations were consistently observed across other study sites (Figure 2 and Table S3).

The overall associations remained consistent when additionally adjusting for prenatal infections (Figure S5). After excluding participants with preeclampsia and/or gestational diabetes, results were generally similar, though the positive association between BDCPP and IL-8 in the pooled sample became stronger (β=0.08, 95% CI: 0.03–0.13, q-value=0.12) (Figure S6). When using dichotomized immune marker outcomes, only the positive association between BDCPP and IL-8 was observed in the pooled sample (relative risk [RR]=1.23, 95% CI: 1.05–1.46, q-value=0.59) (Figure S7). In the Rochester site, the associations of BDCPP with IL-8 (RR=1.33, 95% CI: 1.05–1.70, q-value=0.71) and IL-6 (RR=1.30, 95% CI: 1.02–1.65, q-value=0.77) persisted from the primary analysis, while other site-specific associations were attenuated to the null.

Effect Modification Analyses

We found evidence of effect modification of associations for TNF-α by race/ethnicity (p-interaction<0.05) (Table 3). DPHP and BCETP were positively associated with elevated TNF-α levels among participants from racial/ethnic groups other than non-Hispanic White (DPHP: β=0.16, 95% CI: 0.04–0.28; BCETP: β=0.04, 95% CI: −0.01–0.09), whereas non-Hispanic White participants showed non-significant inverse associations. Fetal sex also modified associations for IL-6. The association between DPHP and IL-6 was negative among those carrying male fetuses (β=−0.18, 95% CI: −0.30– −0.07), while those carrying female fetuses showed null associations (p-interaction=0.03). In contrast, the association between BCPP and IL-6 showed the opposite pattern, with a positive association among participants carrying male fetuses (β=0.04, 95% CI: 0.00–0.07) (p-interaction=0.05). No effect modification by pre-pregnancy BMI was detected.

Table 3.

Associations between dilution-adjusted, log2-transformed OPE metabolite concentrations and log2-transformed, scaled maternal immune marker levels in the pooled sample of three sites, stratified by potential effect modifiers.

TNF-α IL-6
Race/ethnicity NH White (n=271) Other (n=139) NH White (n=271) Other (n=139) p int
β (95% CI) p β (95% CI) p p int β (95% CI) p β (95% CI) p
DPHP −0.06 (−0.15, 0.04) 0.24 0.16 (0.04, 0.28) 0.01 0.01 −0.04 (−0.13, 0.05) 0.39 0.00 (−0.11, 0.12) 0.93 0.55
DBUP/DIBP −0.05 (−0.16, 0.05) 0.33 0.02 (−0.15, 0.18) 0.84 0.54 −0.04 (−0.14, 0.07) 0.47 −0.19 (−0.35, −0.03) 0.02 0.11
BDCPP −0.03 (−0.08, 0.02) 0.21 −0.01 (−0.10, 0.07) 0.75 0.74 −0.02 (−0.06, 0.03) 0.47 0.02 (−0.07, 0.11) 0.71 0.53
BCETP −0.02 (−0.07, 0.02) 0.26 0.04 (−0.01, 0.09) 0.10 0.05 0.01 (−0.03, 0.06) 0.48 0.04 (−0.01, 0.08) 0.14 0.28
BBOEP −0.06 (−0.15, 0.02) 0.15 0.02 (−0.11, 0.14) 0.81 0.36 −0.01 (−0.09, 0.07) 0.75 0.02 (−0.10, 0.13) 0.75 0.75
BCPP 0.02 (−0.02, 0.06) 0.38 −0.03 (−0.08, 0.02) 0.22 0.17 0.01 (−0.03,0.05) 0.51 0.00 (−0.05, 0.05) 0.94 0.61
Pre-pregnancyobesity Obese (n=109) Non-Obese (n=301) Obese (n=109) Non-Obese (n=301) p int
β (95% CI) p β (95% CI) p p int β (95% CI) p β (95% CI) p
DPHP −0.02 (−0.16, 0.13) 0.82 0.05 (−0.04, 0.13) 0.27 0.55 −0.07 (−0.22, 0.08) 0.37 0.00 (−0.08, 0.07) 0.92 0.45
DBUP/DIBP −0.02 (−0.20, 0.15) 0.81 −0.04 (−0.15, 0.07) 0.47 0.90 −0.07 (−0.28, 0.13) 0.47 −0.05 (−0.15, 0.05) 0.34 0.38
BDCPP −0.05 (−0.12, 0.03 0.22 −0.02 (−0.07 0.03) 0.45 0.82 −0.05 (−0.13, 0.03) 0.20 0.00 (−0.04, 0.05) 0.91 0.72
BCETP −0.04 (−0.10, 0.01) 0.13 0.02 (−0.02, 0.06) 0.35 0.77 0.01 (−0.05, 0.07) 0.71 0.03 (−0.01, 0.07) 0.10 0.88
BBOEP −0.06 (−0.19, 0.07) 0.36 −0.03 (−0.11, 0.05) 0.51 0.82 0.00 (−0.14, 0.14) 1.00 0.01 (−0.06, 0.09) 0.73 0.20
BCPP 0.02 (−0.04, 0.07) 0.55 −0.01 (−0.05, 0.03) 0.64 0.62 0.01 (−0.05, 0.07) 0.71 0.01 (−0.03, 0.04) 0.73 0.55
Fetal sex Male (n=230) Female (n=180) Male (n=230) Female (n=180) p int
β (95% CI) p β (95% CI) p p int β (95% CI) p β (95% CI) p
DPHP 0.01 (−0.10, 0.11) 0.91 0.06 (−0.04, 0.17) 0.25 0.57 −0.09 (−0.19, 0.01) 0.08 0.04 (−0.07, 0.14) 0.47 0.08
DBUP/DIBP −0.08 (−0.21, 0.05) 0.22 0.05 (−0.08, 0.18) 0.48 0.11 −0.18 (−0.30, −0.07) 0.002 0.01 (−0.12, 0.14) 0.92 0.03
BDCPP −0.02 (−0.08, 0.04) 0.59 −0.04 (−0.10, 0.02) 0.18 0.77 −0.03 (−0.09, 0.02) 0.23 0.01 (−0.04, 0.07) 0.62 0.25
BCETP 0.03 (−0.02, 0.07) 0.21 −0.03 (−0.07, 0.02) 0.31 0.08 0.02 (−0.02, 0.06) 0.44 0.04 (−0.01, 0.09) 0.12 0.60
BBOEP −0.08 (−0.18, 0.01) 0.08 0.02 (−0.08, 0.13) 0.65 0.10 0.01 (−0.08, 0.09) 0.90 −0.03 (−0.13, 0.08) 0.63 0.74
BCPP −0.02 (−0.06, 0.02) 0.38 0.02 (−0.03, 0.06 0.50 0.18 0.04 (0.00, 0.07) 0.05 −0.02 (−0.07, 0.03) 0.39 0.05
Open in a new tab

Note: OPE metabolite concentrations were urinary dilution adjusted and log2-transformed, and immune marker levels were log2-transformed and z-scored within each site prior to pooling. Models included site as a random effect and were adjusted for gestational week at blood sample collection, batch ID, maternal age at delivery, pre-pregnancy body mass index, parity, maternal race/ethnicity, maternal education, and fetal sex.

Abbreviations: CI, confidence interval; BBOEP, bis(butoxyethyl) phosphate; BCETP, bis(2-chloroethyl) phosphate; BCPP, bis(1-chloro-2-propyl) phosphate; BDCPP, bis(1,3-dichloro-2-propyl) phosphate; DBUP/DIBP, composite of dibutyl phosphate and di-isobutyl phosphate; DPHP, diphenyl phosphate; IL, interleukin; NH, non-Hispanic; OPE, organophosphate ester; TNF-α, tumor necrosis factor alpha.

4. DISCUSSION

In our multi-site analysis, higher prenatal BDCPP concentrations were associated with elevated IL-8 levels. Although this association was observed in two contributing sites and the pooled sample, IL-8 was not measured in the third site, limiting comparability across all three sites. Therefore, this finding warrant confirmation in future studies. Several associations between other OPE metabolites and immune markers were not replicated across sites. In exploratory analyses, we identified effect modification by race/ethnicity for associations between several OPE metabolites and TNF-α, with more pronounced positive associations observed among racial/ethnic groups other than non-Hispanic White. Additionally, several associations for IL-6 differed by fetal sex, although the directions of these associations varied by OPE metabolite.

The significant association between BDCPP and higher IL-8 levels likely reflects constant, ongoing exposure to its parent compound, TDCPP, which is ubiquitous in the indoor environment and consumer products (Huang et al. 2025). This is evidenced by the moderate-to-strong reproducibility of BDCPP with intraclass correlation coefficients ranging from 0.5 to 0.6 in repeated urine samples of pregnant women (Bommarito et al. 2023; Hoffman et al. 2014; Romano et al. 2017). This pseudo-persistent exposure may provide a sustained stimulus sufficient to induce a systemic inflammatory response.

The pro-inflammatory chemokine IL-8 is released in response to environmental and endogenous stimuli and primarily recruits neutrophils to sites of inflammation (Alharbi and Alhujaily 2024). Mechanistically, TDCPP has shown immunotoxicity in vivo and in vitro models using various tissues and cells (Alharbi and Alhujaily 2024; Canbaz et al. 2017; Li et al. 2025; Li et al. 2020; Wang et al. 2025; Xie et al. 2025; Zhong et al. 2020). For example, in spleen tissue, TDCPP exposure has been shown to activate the NF-κB pathway, which is involved in the regulation of immune responses to infection and inflammation (Cao et al. 2023). Experimental studies also reported that TDCPP exposure in lung cells caused cellular, mitochondrial, and DNA damage (Yuan et al. 2021) while increasing reactive oxygen species and IL-8 levels (Pyambri et al. 2025). Oxidative stress was identified as a possible initial mechanistic process for these outcomes. While IL-8 is a necessary effector or innate immunity and plays a role in maintaining a healthy pregnancy by regulating not just infection but also embryo development and angiogenesis (Vilotić et al. 2022), its supraphysiologic production influenced by OPE exposure can contribute to cellular stress responses and chronic inflammation.

In humans, only one prior study examined these associations in 65 Californian pregnant women, reporting slightly different results (Nguyen et al. 2025). That study observed positive associations between DPHP and higher maternal serum levels of IL-8 and CCL4, along with a marginal positive association between BDCPP and CCL4, but not with IL-8. Although none of our contributing sites measured the chemokine CCL4, the inconsistent results across studies and our site-specific findings may be attributed to differences in sample size, OPE metabolite concentrations, immune marker levels, timing of urine and blood sample collection, and participant characteristics.

We identified that participants from racial/ethnic groups other than non-Hispanic White were potentially more vulnerable to the effects of OPE exposure on TNF-α, a key pro-inflammatory cytokine triggering and regulating inflammatory responses. Although effect modification for prenatal OPE exposure by race/ethnicity is understudied, one study has reported stronger associations between per- and polyfluoroalkyl substances and TNF-α or IL-6 among African American participants (Siwakoti et al. 2024). Race and ethnicity should be considered a possible proxy for unmeasured social and environmental factors, such as structural racism and environmental and health disparities, which may influence differences in OPE exposure and immune marker baseline levels, rather than serving as a direct biological variable. Certain racial/ethnic groups likely experience more chronic stress from these social and environmental factors, potentially altering inflammatory pathways that overlap with OPE effects, contributing to adverse health outcomes (Barrett and Padula 2019; Clougherty et al. 2014).

We also observed sex-dimorphic trends in associations for IL-6, a pro-inflammatory signaling molecule: mothers carrying male fetuses showed inverse associations with DBUP/DIBP and positive associations with BCPP, whereas mothers carrying female fetuses exhibited the null or opposite pattern. As supported by sex-dimorphic associations reported for other EDCs (Lee et al. 2023; Siwakoti et al. 2024), these sex-specific responses to environmental chemicals are biologically relevant. Fetal sex impacts not only the immunity at the placenta and maternal-fetal interface but also influences the stimulation and modulation of the maternal immune system during pregnancy, leading to differences in circulating cytokine levels (Baines and West 2023). Fetal sex has also been shown to modify OPE exposure and sex steroids hormone levels in pregnant women (Hansel et al. 2025), which can further alter immune responses (Wan et al. 2026). Experimental evidence supports these sex-specific effects of prenatal exposure to an OPE mixture on immune status (Giles et al. 2025). In that study, male offspring showed a reduction in splenic monocyte–macrophage populations, while female offspring exhibited more pronounced cytokine changes, possibly reflecting compensatory inflammation. Therefore, future studies are needed to confirm our sex-specific associations while testing a broader panel of cytokines and chemokines in the analysis.

A major strength of this study is the combination of three U.S. cohort sites with distinct profiles in immune marker levels (likely stemming from both biological and methodological variations) and participant characteristics to examine understudied associations between OPE exposure and inflammatory profiles in pregnant women. We ensured high data quality by using OPE metabolite concentrations measured in a single laboratory, which minimized potential measurement error. Furthermore, while specific manufacturers varied, the primary method for quantifying the multiplexed cytokines and chemokines across all three sites relied on Luminex xMAP technology, establishing platform similarity for these specific markers. The ECHO Cohort’s rich dataset also allowed us to adjust for a comprehensive covariate set, assess effect modification by key maternal and fetal factors.

Several limitations must be acknowledged. First, we relied on OPE metabolite concentrations measured in a single spot urine sample per participant. While certain OPE metabolites show fair to good reproducibility in pregnant women’s urine likely due to relatively stable exposure sources, such as the home environment (Bommarito et al. 2023; Hoffman et al. 2014; Romano et al. 2017), repeated measurements of OPE metabolites would have better captured individuals’ OPE exposure given their short biological half-lives. Second, we also relied on immune marker levels measured at a single time point during mid- to late-pregnancy. Immune marker levels naturally change significantly as pregnancy proceeds (Meyyazhagan et al. 2023). For example, maternal serum levels of IL-8 have been shown to decrease as pregnancy progresses but to increase toward the end of gestation in non-complicated pregnancy (Vilotić et al. 2022). Therefore, further studies employing repeated measurements of immune markers are warranted to characterize their trajectories. Third, immune markers play critical roles in establishing and maintaining a healthy pregnancy (Meyyazhagan et al. 2023), and our data set lacked levels of anti-inflammatory decoy molecules that are essential for the assessment of the homeostatic balance (Fichorova 2004). Because the biological impact of pro-inflammatory signaling is often modulated by these regulatory counterparts, elevated levels of selected cytokines or chemokines do not necessarily indicate a pathological state, and our findings should be interpreted within the context of population health.

Fourth, although we utilized z-score standardization and included random effects for site to enable pooled analyses, inter-site comparability of immune markers may still be limited by technical factors. Future studies would benefit from more integrated multi-site designs, such as analyzing all biospecimens on a single platform within the same laboratory, to minimize batch effects and enhance biological comparability across cohort sites. Fifth, the potential for temporal misalignment cannot be excluded. While 100% of the Healthy Start and 98% of the Rochester blood samples were collected on the same day as the urine samples, only half of the MARBLES/EARLI samples were obtained on the same day. This discrepancy in sampling intervals may have introduced a degree of measurement error, suggesting a cautious interpretation of the pooled associations. Sixth, MARBLES and EARLI are autism-enriched cohorts, which may limit direct comparability with populations at lower risk. Although there is currently no definitive evidence that mothers of children with autism exhibit fundamentally different immune profiles during pregnancy, we cannot exclude the possibility that underlying familial or genetic factors associated with neurodevelopmental risk could influence maternal immune regulation. Consequently, the generalizability of our pooled findings to lower-risk populations should be interpreted with caution. Seventh, the generalizability of our findings may be limited due to demographic differences between our analytic sample and the broader ECHO Cohort or the general U.S. population. Lastly, there may be unmeasured or residual confounding factors that affect immune markers, such as diet or recent subclinical infections, which could potentially bias the observed associations. Future studies would benefit from a more focused collection of data on diet, infection history, other environmental exposures, and longitudinal perinatal as well as childhood exposures to better characterize the immunological impact of OPEs.

5. CONCLUSION

This study supports the concept of a positive association between prenatal TDCPP exposure (as measured by its metabolite BDCPP) and maternal inflammation. It identifies IL-8, a chemokine that primarily recruits neutrophils to the inflammation site, as a potential target for future experimental and larger population-based observational studies of mechanisms underlying the observed association. We found evidence of effect modification by race/ethnicity and fetal sex, identifying subgroups potentially susceptible to associations between certain OPEs and key mediators of inflammation. However, given the exploratory nature of these analyses, further investigation is necessary to establish the robustness and generalizability. Given the known temporal variability of both OPE metabolite concentrations and immune marker levels during gestation, our findings should be confirmed by future studies leveraging repeated measurements of exposure and immune markers. Furthermore, research investigating how OPE-related immune alterations impact child health is warranted to elucidate the mechanisms by which prenatal exposure influences long-term health.

Supplementary Material

1

HIGHLIGHTS.

  • This study included mother-child pairs prospectively followed at three ECHO sites

  • Prenatal urinary OPE metabolites and blood immune markers were measured by each site

  • Prenatal BDCPP was associated with elevated maternal interleukin-8 levels

  • Race/ethnicity and fetal sex modified associations of OPEs with inflammatory markers

Acknowledgments

The authors wish to thank our ECHO colleagues; the medical, nursing, and program staff; and the children and families participating in the ECHO Cohort.

Research reported in this publication was supported by the Environmental influences on Child Health Outcomes (ECHO) Program, Office of the Director, National Institutes of Health, under F32OD037666 (Jiwon Oh), UG3/UH3OD023365 (Irva Hertz-Picciotto; Deborah H. Bennett), UG3/UH3OD023248 (Dana M. Dabelea), R01DK076648 (Dana M. Dabelea), R00ES025817 (Anne P. Starling), UG3/UH3OD023342 (Kristen Lyall; Rebecca J. Schmidt), U24ES028533 (Rebecca J. Schmidt), UG3/UH3OD023349 (Thomas G. O’Connor), U2CES026542-01 (HHEAR Kannan Kurunthachalam).

Role of Funder Statement

The sponsor, NIH, participated in the overall design and implementation of the ECHO Program, which was funded as a cooperative agreement between NIH and grant awardees. The sponsor approved the Steering Committee-developed ECHO protocol and its amendments including COVID-19 measures. The sponsor had no access to the central database, which was housed at the ECHO Data Analysis Center. Data management and site monitoring were performed by the ECHO Data Analysis Center and Coordinating Center. All analyses for scientific publication were performed by the study statistician, independently of the sponsor. The lead author wrote all drafts of the manuscript and made revisions based on co-authors without input from the sponsor. The study sponsor did not review or approve the manuscript for submission to the journal.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Declaration of interests

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

Rebecca J. Schmidt reports a relationship with Linus Biotechnology Inc that includes: consulting or advisory. Deborah H. Bennett reports a relationship with Linus Biotechnology Inc that includes: consulting or advisory. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Conflicts of Interest

RJS consults for Beasley Law, Inc. RJS and DHB consult for Linus Biotechnology, Inc. No other authors have any conflicts to report.

Data Availability Statement

Select de-identified data from the ECHO Program are available through NICHD’s Data and Specimen Hub (DASH). Information on study data not available on DASH can be found on the ECHO study DASH webpage.

References

  1. Ait Bamai Y, Bastiaensen M, Araki A, Goudarzi H, Konno S, Ito S, Miyashita C, Yao Y, Covaci A, Kishi R, 2019. Multiple exposures to organophosphate flame retardants alter urinary oxidative stress biomarkers among children: The Hokkaido Study. Environment international. 131, 105003. [DOI] [PubMed] [Google Scholar]
  2. Alharbi A, Alhujaily M, 2024. Molecular Mechanism of Indoor Exposure to Airborne Halogenated Flame Retardants TCIPP (Tris (1, 3-Dichloro-2-Propyl) Phosphate) and TCEP Tris (2-chloroethyl) Phosphate and Their Hazardous Effects on Biological Systems. Metabolites. 14, 697. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Baines KJ, West RC, 2023. Sex differences in innate and adaptive immunity impact fetal, placental, and maternal health. Biology of reproduction. 109, 256–270. [DOI] [PubMed] [Google Scholar]
  4. Barrett ES, Padula AM, 2019. Joint impact of synthetic chemical and non-chemical stressors on children’s health. Current environmental health reports. 6, 225–235. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Blum A, Behl M, Birnbaum LS, Diamond ML, Phillips A, Singla V, Sipes NS, Stapleton HM, Venier M, 2019. Organophosphate ester flame retardants: are they a regrettable substitution for polybrominated diphenyl ethers? Environmental science & technology letters. 6, 638–649. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bommarito P, Friedman A, Welch B, Cantonwine D, Ospina M, Calafat A, Meeker J, McElrath T, Ferguson K, 2023. Temporal trends and predictors of gestational exposure to organophosphate ester flame retardants and plasticizers. Environment international. 180, 108194. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Breen EC, Reynolds SM, Cox C, Jacobson LP, Magpantay L, Mulder CB, Dibben O, Margolick JB, Bream JH, Sambrano E, 2011. Multisite comparison of high-sensitivity multiplex cytokine assays. Clinical and Vaccine Immunology. 18, 1229–1242. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Canbaz D, Logiantara A, van Ree R, van Rijt LS, 2017. Immunotoxicity of organophosphate flame retardants TPHP and TDCIPP on murine dendritic cells in vitro. Chemosphere. 177, 56–64. [DOI] [PubMed] [Google Scholar]
  9. Cao L, Wei L, Du Q, Su Y, Ye S, Liu K, 2023. Spleen toxicity of organophosphorus flame retardant tdcpp in mice and the related mechanisms. Toxics. 11, 231. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Clougherty JE, Shmool JL, Kubzansky LD, 2014. The role of non-chemical stressors in mediating socioeconomic susceptibility to environmental chemicals. Current Environmental Health Reports. 1, 302–313. [Google Scholar]
  11. Doherty BT, Hammel SC, Daniels JL, Stapleton HM, Hoffman K, 2019. Organophosphate esters: are these flame retardants and plasticizers affecting children’s health? Current environmental health reports. 6, 201–213. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Ferguson KK, Cantonwine DE, McElrath TF, Mukherjee B, Meeker JD, 2016. Repeated measures analysis of associations between urinary bisphenol-A concentrations and biomarkers of inflammation and oxidative stress in pregnancy. Reproductive Toxicology. 66, 93–98. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Ferguson KK, Cantonwine DE, Rivera-Gonzlez LO, Loch-Caruso R, Mukherjee B, Anzalota Del Toro LV, Jimnez-Vlez B, Calafat AM, Ye X, Alshawabkeh AN, 2014. Urinary phthalate metabolite associations with biomarkers of inflammation and oxidative stress across pregnancy in Puerto Rico. Environmental science & technology. 48, 7018–7025. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Fichorova RN, 2004. Guiding the vaginal microbicide trials with biomarkers of inflammation. JAIDS Journal of Acquired Immune Deficiency Syndromes. 37, S184–S193. [PMC free article] [PubMed] [Google Scholar]
  15. Friedman C, Dabelea D, Thomas DS, Peel JL, Adgate JL, Magzamen S, Martenies SE, Allshouse WB, Starling AP, 2021. Exposure to ambient air pollution during pregnancy and inflammatory biomarkers in maternal and umbilical cord blood: The Healthy Start study. Environmental research. 197, 111165. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Giles BH, Subramaniam NK, Caruana V, Kukolj N, Robaire B, Mann KK, 2025. A mixture of organophosphate esters sex-specifically impacts monocyte–macrophages in Sprague-Dawley rats. Toxicological Sciences. kfaf078. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Gillman MW, Blaisdell CJ, 2018. Environmental influences on child health outcomes, a research program of the National Institutes of Health. Current opinion in pediatrics. 30, 260–262. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Goldstein JA, Gallagher K, Beck C, Kumar R, Gernand AD, 2020. Maternal-fetal inflammation in the placenta and the developmental origins of health and disease. Frontiers in immunology. 11, 531543. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Han H, Grill S, Shen X, Williams PL, James-Todd T, Ford JB, Rexrode KM, Calafat AM, Chavarro JE, Hauser R, 2025. Association of urinary phthalate metabolite concentrations with inflammatory biomarker levels among pregnant women. Environmental Research. 121911. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Hansel MC, Lubina KA, Roepke TA, Ohman-Strickland P, Kannan K, Wang C, Miller RK, O’Connor TG, Rivera-Núñez Z, Barrett ES, 2025. Maternal organophosphate esters and sex steroid hormones in mid-pregnancy. Environmental research. 271, 121063. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Hertz-Picciotto I, Schmidt RJ, Walker CK, Bennett DH, Oliver M, Shedd-Wise KM, LaSalle JM, Giulivi C, Puschner B, Thomas J, 2018. A prospective study of environmental exposures and early biomarkers in autism spectrum disorder: design, protocols, and preliminary data from the MARBLES study. Environmental health perspectives. 126, 117004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Hoffman K, Daniels JL, Stapleton HM, 2014. Urinary metabolites of organophosphate flame retardants and their variability in pregnant women. Environment international. 63, 169–172. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Hong J, Lu X, Wang J, Jiang M, Liu Q, Lin J, Sun W, Zhang J, Shi Y, Liu X, 2022. Triphenyl phosphate disturbs placental tryptophan metabolism and induces neurobehavior abnormal in male offspring. Ecotoxicology and Environmental Safety. 243, 113978. [DOI] [PubMed] [Google Scholar]
  24. Hornung RW, Reed LD, 1990. Estimation of average concentration in the presence of nondetectable values. Applied occupational and environmental hygiene. 5, 46–51. [Google Scholar]
  25. Hou R, Xu Y, Wang Z, 2016. Review of OPFRs in animals and humans: Absorption, bioaccumulation, metabolism, and internal exposure research. Chemosphere. 153, 78–90. [DOI] [PubMed] [Google Scholar]
  26. Hu W, Gao P, Wang L, Hu J, 2023. Endocrine disrupting toxicity of aryl organophosphate esters and mode of action. Critical Reviews in Environmental Science and Technology. 53, 1–18. [Google Scholar]
  27. Huang M, Zeng L, Wang C, Zhou X, Peng Y, Shi C, Wang S, Li Y, Barceló D, Li H, 2025. A comprehensive and quantitative comparison of organophosphate esters: Characteristics, applications, environmental occurrence, toxicity, and health risks. Critical Reviews in Environmental Science and Technology. 55, 310–333. [Google Scholar]
  28. Huang Y-S, Shi H-Z, Huang X, Pan Y-M, Wang Y-C, Gao Z-J, Jiang P-Y, Yang W-Y, 2024. Urinary concentrations of organophosphate flame-retardant metabolites in the US population. JAMA network open. 7, e2435484–e2435484. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Jin S, Cui S, Huang X, Li Z, Han Y, Cui T, Su Y, Xiong W, Zhang X, 2023. BMI-specific inflammatory response to phthalate exposure in early pregnancy: findings from the TMCHESC study. Environmental Science and Pollution Research. 30, 123383–123395. [DOI] [PubMed] [Google Scholar]
  30. Kannan K, Stathis A, Mazzella MJ, Andra SS, Barr DB, Hecht SS, Merrill LS, Galusha AL, Parsons PJ, 2021. Quality assurance and harmonization for targeted biomonitoring measurements of environmental organic chemicals across the Children’s Health Exposure Analysis Resource laboratory network. International journal of hygiene and environmental health. 234, 113741. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Kelley AS, Banker M, Goodrich JM, Dolinoy DC, Burant C, Domino SE, Smith YR, Song PX, Padmanabhan V, 2019. Early pregnancy exposure to endocrine disrupting chemical mixtures are associated with inflammatory changes in maternal and neonatal circulation. Scientific reports. 9, 5422. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Knapp EA, Kress AM, Parker CB, Page GP, McArthur K, Gachigi KK, Alshawabkeh AN, Aschner JL, Bastain TM, Breton CV, 2023. The environmental influences on child health outcomes (ECHO)-wide cohort. American journal of epidemiology. 192, 1249–1263. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Kuiper JR, O’Brien KM, Welch BM, Barrett ES, Nguyen RH, Sathyanarayana S, Milne GL, Swan SH, Ferguson KK, Buckley JP, 2022. Combining Urinary Biomarker Data From Studies With Different Measures of Urinary Dilution. Epidemiology. 33, 533–540. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Lee CW, Cathey AL, Watkins DJ, Rosario-Pabón ZY, Vélez-Vega CM, Alshawabkeh AN, Cordero JF, Meeker JD, 2023. Associations of urinary phthalate metabolites and inflammatory biomarkers among pregnant women in Puerto Rico. Science of the Total Environment. 854, 158773. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Li M-R, Zhou G-R, Wang Z-Y, Li B-Y, Men S-H, Yan Z-G, 2025. TDCPP promotes apoptosis and inhibits the calcium signaling pathway in human neural stem cells. Science of The Total Environment. 961, 178429. [DOI] [PubMed] [Google Scholar]
  36. Li X, Li N, Rao K, Huang Q, Ma M, 2020. In vitro immunotoxicity of organophosphate flame retardants in human THP-1-derived macrophages. Environmental Science & Technology. 54, 8900–8908. [DOI] [PubMed] [Google Scholar]
  37. Li Y, Wang X, Zhu Q, Xu Y, Fu Q, Wang T, Liao C, Jiang G, 2023. Organophosphate flame retardants in pregnant women: sources, occurrence, and potential risks to pregnancy outcomes. Environmental Science & Technology. 57, 7109–7128. [DOI] [PubMed] [Google Scholar]
  38. Liu X, Zhao X, Wang Y, Hong J, Shi M, Pfaff D, Guo L, Tang H, 2020. Triphenyl phosphate permeates the blood brain barrier and induces neurotoxicity in mouse brain. Chemosphere. 252, 126470. [DOI] [PubMed] [Google Scholar]
  39. Maaniitty E, Jalkanen J, Sinisilta S, Gunn J, Vasankari T, Biancari F, Jalkanen S, Airaksinen KJ, Hollmen M, Kiviniemi T, 2024. Differential circulating cytokine profiles in acute coronary syndrome versus stable coronary artery disease. Scientific Reports. 14, 17269. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Meng Y, Xu X, Xie G, Zhang Y, Chen S, Qiu Y, Zhu Z, Zhang H, Yin D, 2022. Alkyl organophosphate flame retardants (OPFRs) induce lung inflammation and aggravate OVA-simulated asthmatic response via the NF-кB signaling pathway. Environment international. 163, 107209. [DOI] [PubMed] [Google Scholar]
  41. Meyyazhagan A, Kuchi Bhotla H, Pappuswamy M, Tsibizova V, Al Qasem M, Di Renzo GC, 2023. Cytokine see-saw across pregnancy, its related complexities and consequences. International Journal of Gynecology & Obstetrics. 160, 516–525. [DOI] [PubMed] [Google Scholar]
  42. Newschaffer CJ, Croen LA, Fallin MD, Hertz-Picciotto I, Nguyen DV, Lee NL, Berry CA, Farzadegan H, Hess HN, Landa RJ, 2012. Infant siblings and the investigation of autism risk factors. Journal of neurodevelopmental disorders. 4, 7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Nguyen V, Sehgal N, Li L, Goin DE, Morello-Frosch R, Woodruff TJ, Varshavsky J, Park J-S, Gaw SL, Robinson JF, 2025. Associations between brominated and organophosphate flame retardant levels and inflammatory biomarkers in maternal blood, cord blood, and placental tissue at mid-gestation. Science of The Total Environment. 1001, 180501. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. O’Connor T, Best M, Brunner J, Ciesla AA, Cunning A, Kapula N, Kautz A, Khoury L, Macomber A, Meng Y, 2021. Cohort profile: Understanding Pregnancy Signals and Infant Development (UPSIDE): a pregnancy cohort study on prenatal exposure mechanisms for child health. BMJ open. 11, e044798. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Oh J, Buckley JP, Li X, Gachigi KK, Kannan K, Lyu W, Ames JL, Barrett ES, Bastain TM, Breton CV, 2024. Associations of Organophosphate Ester Flame Retardant Exposures during Pregnancy with Gestational Duration and Fetal Growth: The Environmental influences on Child Health Outcomes (ECHO) Program. Environmental Health Perspectives. 132, 017004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Ospina M, Jayatilaka NK, Wong L-Y, Restrepo P, Calafat AM, 2018. Exposure to organophosphate flame retardant chemicals in the US general population: Data from the 2013–2014 National Health and Nutrition Examination Survey. Environment international. 110, 32–41. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Peng C, Zhang X, Chen Y, Wang L, 2023. Toxicity assessment of organophosphate flame retardant triphenyl phosphate (TPHP) on intestines in mice. Ecotoxicology and Environmental Safety. 268, 115685. [DOI] [PubMed] [Google Scholar]
  48. Pyambri M, Pavlidou A, Lacorte S, Jaumot J, Bedia C, 2025. Phenotypic and lipidomic alterations in lung cells induced by organophosphate flame retardants. Chemistry and physics of lipids. 270, 105508. [DOI] [PubMed] [Google Scholar]
  49. Racicot K, Kwon JY, Aldo P, Silasi M, Mor G, 2014. Understanding the complexity of the immune system during pregnancy. American journal of reproductive immunology. 72, 107–116. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Romano ME, Hawley NL, Eliot M, Calafat AM, Jayatilaka NK, Kelsey K, McGarvey S, Phipps MG, Savitz DA, Werner EF, 2017. Variability and predictors of urinary concentrations of organophosphate flame retardant metabolites among pregnant women in Rhode Island. Environmental Health. 16, 40. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Rubin DB, 1987. Multiple Imputation for Nonresponse in Surveys. John Wiley & Sons Inc. New York. [Google Scholar]
  52. Serrano J, Womack S, Yount C, Chowdhury SF, Arnold M, Brunner J, Duberstein Z, Barrett ES, Scheible K, Miller RK, 2024. Prenatal maternal immune activation predicts observed fearfulness in infancy. Developmental psychology. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Siwakoti RC, Harris SM, Ferguson KK, Hao W, Cantonwine DE, Mukherjee B, McElrath TF, Meeker JD, 2024. Prenatal exposure to per-and polyfluoroalkyl substances (PFAS) and their influence on inflammatory biomarkers in pregnancy: findings from the LIFECODES cohort. Environment international. 194, 109145. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Textor J, Van der Zander B, Gilthorpe MS, Liśkiewicz M, Ellison GT, 2016. Robust causal inference using directed acyclic graphs: the R package ‘dagitty’. International journal of epidemiology. 45, 1887–1894. [DOI] [PubMed] [Google Scholar]
  55. Uh H-W, Hartgers FC, Yazdanbakhsh M, Houwing-Duistermaat JJ, 2008. Evaluation of regression methods when immunological measurements are constrained by detection limits. BMC immunology. 9, 1–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Vilotić A, Nacka-Aleksić M, Pirković A, Bojić-Trbojević Ž, Dekanski D, Jovanović Krivokuća M, 2022. IL-6 and IL-8: an overview of their roles in healthy and pathological pregnancies. International journal of molecular sciences. 23, 14574. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Volk HE, Park B, Hollingue C, Jones KL, Ashwood P, Windham GC, Lurman F, Alexeeff SE, Kharrazi M, Pearl M, 2020. Maternal immune response and air pollution exposure during pregnancy: insights from the Early Markers for Autism (EMA) study. Journal of neurodevelopmental disorders. 12, 42. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Wan B, Chen W, Shan M, Chi L, 2026. Immune modulation by sex hormones: unveiling the complex interplay and regulatory networks. Frontiers in Immunology. 17, 1748097. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Wang S, Shi C, Lai R, Wang C, Chen T, Zeng L, Peng Y, Huang M, Zheng Y, Gao Y, 2025. Immunotoxicity and cancer risk exacerbation induced by tris (1, 3-dichloro-2-propyl) phosphate exposure: insights from macrophage inflammatory responses and in vivo tumor models. Environment International. 109903. [DOI] [PubMed] [Google Scholar]
  60. Wang Y, Li W, Martínez-Moral MP, Sun H, Kannan K, 2019. Metabolites of organophosphate esters in urine from the United States: concentrations, temporal variability, and exposure assessment. Environment international. 122, 213–221. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Watkins DJ, Ferguson KK, Del Toro LVA, Alshawabkeh AN, Cordero JF, Meeker JD, 2015. Associations between urinary phenol and paraben concentrations and markers of oxidative stress and inflammation among pregnant women in Puerto Rico. International journal of hygiene and environmental health. 218, 212–219. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Wei G-L, Li D-Q, Zhuo M-N, Liao Y-S, Xie Z-Y, Guo T-L, Li J-J, Zhang S-Y, Liang Z-Q, 2015. Organophosphorus flame retardants and plasticizers: sources, occurrence, toxicity and human exposure. Environmental Pollution. 196, 29–46. [DOI] [PubMed] [Google Scholar]
  63. Xie G, Yang Y, Zhang X, Chen Y, Cao X, Zhang J, Jiang Y, Yang W, 2025. Tris (1, 3-dichloro-2-propyl) phosphate (TDCPP) aggravates Parkinson’s disease neurotoxicity through ferroptosis-related oxidative stress and neuroinflammation. Brain Research. 149805. [DOI] [PubMed] [Google Scholar]
  64. Yao Y, Li M, Pan L, Duan Y, Duan X, Li Y, Sun H, 2021. Exposure to organophosphate ester flame retardants and plasticizers during pregnancy: thyroid endocrine disruption and mediation role of oxidative stress. Environment international. 146, 106215. [DOI] [PubMed] [Google Scholar]
  65. Yockey LJ, Iwasaki A, 2018. Interferons and proinflammatory cytokines in pregnancy and fetal development. Immunity. 49, 397–412. [DOI] [PMC free article] [PubMed] [Google Scholar]
  66. Yuan S, Ma M, Zhu X, Han Y, Rao K, Wang Z, 2021. Cytotoxicity, mitochondrial impairment, DNA damage and associated mechanisms induced by tris (1, 3-dichloro-2-propyl) phosphate and tris (2-butoxyethyl) phosphate in A549 cells. Science of The Total Environment. 787, 147668. [Google Scholar]
  67. Zhong X, Wu J, Ke W, Yu Y, Ji D, Kang J, Qiu J, Wang C, Yu P, Wei Y, 2020. Neonatal exposure to organophosphorus flame retardant TDCPP elicits neurotoxicity in mouse hippocampus via microglia-mediated inflammation in vivo and in vitro. Archives of Toxicology. 94, 541–552. [DOI] [PubMed] [Google Scholar]
  68. Zota AR, Geller RJ, Romano LE, Coleman-Phox K, Adler NE, Parry E, Wang M, Park J-S, Elmi AF, Laraia BA, 2018. Association between persistent endocrine-disrupting chemicals (PBDEs, OH-PBDEs, PCBs, and PFASs) and biomarkers of inflammation and cellular aging during pregnancy and postpartum. Environment international. 115, 9–20. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS2191787-supplement-1.docx (2.9MB, docx)

Data Availability Statement

Select de-identified data from the ECHO Program are available through NICHD’s Data and Specimen Hub (DASH). Information on study data not available on DASH can be found on the ECHO study DASH webpage.

RESOURCES