Prenatal phthalate exposure and anogenital distance in infants at 12 months.

This study was performed in the New York University Children Health and Environment Study (NYU CHES). A prospective birth cohort study that has been enrolling pregnant participants in NYC since 2016 ( Trasande et al. 2020 ). Eligibility criteria include being 18 years old or older, being fewer than 18 weeks of gestation, and intending to deliver at one of the three NYU-affiliated hospitals ( Trasande et al. 2020 ). Between 2016 and 2019, 2193 pregnant participants completed the first assessment of study ( Trasande et al. 2020 ). Measures of urinary phthalate metabolite concentrations were available in 1399 pregnant participants in early pregnancy, of whom 1276 had singleton live-born children. A total of 506 infants completed child postnatal visits at age 12 months and had AGD measurements available ( Figure S1 ). For 342 mother-child pairs (67.6%), we had a second measure of urinary phthalates in mid-pregnancy as well. To participate in the study, all participants provided their written informed consent. The NYU Grossman School of Medicine Institutional Review Board gave their approval to the research study. Phthalate metabolites were measured in spot urine samples collected throughout the early [mean = 10.7 weeks; standard deviation (SD) = 2.9] and mid-pregnancy [mean = 20.7; SD = 1.7]. These urine samples were gathered in polyethylene containers, aliquoted in tubes free of phthalates, and stored at −80°C till their chemical concentrations were assayed by the NYU Human and Environmental Exposure Analysis Laboratory (HEAL). For human monitoring, since phthalates are metabolized into a number of short-lived monoester metabolites that are mostly excreted in urine after absorption, urinary phthalate metabolite concentrations are regarded as reliable markers of exposure to parent phthalate compounds ( Kasper-Sonnenberg et al. 2025 ; Kato et al. 2004 ; Koch et al. 2012 ; Wang et al. 2019 ). To measure phthalates metabolites, glucoronidated and sulfated phthalate monoesters were enzymatically deconjugated, followed by Solid Phase Extraction (SPE). The extracted compound were subsequently analyzed using high-performance liquid chromatography electrospray ionization tandem mass spectrometry (HPLC-ESI-MS/MS), as outlined previously ( Gaylord et al. 2022 ; Liu et al. 2022 ). Phthalate metabolites that were detected in 50% or more of the samples (15 out of 22 measured) were included in this analysis. Values below the limit of detection (LOD) for these 15 phthalate metabolites were replaced by LOD/ 2 ( Hornung and Reed 1990 ). Based on their molecular weight, phthalate metabolites were categorized into two groups: low-molecular-weight (∑LMWP, <250Da) and high-molecular-weight (∑HMWP, ≥250Da). Additionally, we categorized HMWP metabolites according to the parent diester molecule: ∑DEHP, and ∑DiNP. To calculate a group compound of phthalates, this chemical must have 50% or more of the detection values in both periods of measurement. Using the following formula, we also determined the molar sum of metabolite components (nmol/L): formula = metabolite concentration n g m L x 1 molecular weight x 1 10 - 3 + metabolite concentration n g m L x 1 molecular weight x 1 10 - 3 + … We analyzed MCPP as an individual chemical, despite it having over 50% detection in both measurement periods, because it has been identified as a metabolite of several phthalates compound, including di-n-octyl phthalate (DnOP), DiNP and di-isodecyl phthalate ( Calafat et al. 2006 ; Gomez Ramos et al. 2016 ; Kasper-Sonnenberg et al. 2025 ). Phthalic acid (PA) was analyzed independently as a proxy of overall phthalate exposure ( Bang et al. 2011 ). Table 1 and S1 displays the concentrations (ng/mL) and detection rates for all phthalates metabolites analyzed in this study. Using urinary creatinine concentrations via the Boeniger method, we corrected for urinary dilution ( Boeniger et al. 1993 ; Kuiper et al. 2021 ). Briefly, phthalate concentrations were multiplied by the ratio of the study population’s batch-specific median creatinine value to the sample’s creatinine concentration. Among those with phthalate metabolite concentrations measured at two time points, coefficients of intraclass correlation (ICC) ranged from 0.19 to 0.25. To reduce measurement error due to metabolite variability and short half-lives, we calculated the average exposure using samples collected at <18 and 18–25 weeks—when two measurements were available. All individual phthalate metabolite and group concentrations were natural log-transformed. For the mixture analyses, standardized concentrations of all-natural log-transformed phthalate groups were used. AGD in infants was measured during visits at 12 months of age [median = 12.9 months; 95% range = 11.4–21.0]. For each measurement, trained examiners obtained three independent measures by using the caliper closed and zeroed out between each measurement to the nearest 0.1mm ( Sathyanarayana et al. 2015 ). Measurements were performed with the infants in the supine position with their legs up and apart and their hips flexed. Every AGD measurements is taken from the anus’s center to a genital landmark. In females, the landmarks were the base of posterior fourchette (AGDaf) and the anterior tip of the clitoral hood (AGDac). In males, the landmarks were the base of scrotum, where the skin transitions from rugate to smooth (AGDas), and the anterior base of the penis, where the pubic bone and penile tissue meet (AGDap). To measure penile width (PW), the diameter of penis’s base was measured while it was flaccid ( Sathyanarayana et al. 2015 ). To reduce measurement error, we included those participants who were measured three times in each AGD measure. Intra-rater ICC between the three measurements ranged between 0.87 and 0.99. We calculated the average of the three measurements for each AGD parameter. For 14% of the infants, a second examiner obtained independent measurement (intra-examiner ICC [0.89–0.92] and inter-examiner ICC [0.69–0.84]). Based on existing literature, potential confounders were chosen and visualized as a Directed Acyclic Graph (see Figure S2 ) ( Adibi et al. 2015 ; Cowell et al. 2023 ; Santos et al. 2021 ; Swan et al. 2015 ; Wenzel et al. 2018 ). Information on maternal age (years) and pre-pregnancy weight (kg) was retrieved from electronic health records, and missing data was supplemented using questionnaire data. During prenatal visits, information on maternal height (meters), race and ethnicity (Hispanic, non-Hispanic White, non-Hispanic Black, or non-Hispanic Asian/other/multiple), educational level (high school or less, some college but not degree, bachelor’s or associate’s degree, or postgraduate degree), partnership status (partnered or no), and alcohol intake during pregnancy (never, until pregnancy was known, or continued) was collected by questionnaires administered. Additionally, parity (nulliparous or primiparous/multiparous), health insurance (public or private), and missing data from questionnaire reports were filled in using information recorded in electronic health records. To assess environmental cigarette smoke exposure, we used urinary cotinine concentrations (below or above limit of detection), which were analyzed by HPLC-MS/MS and maternal questionnaire report of smoking behavior during pregnancy (mother smoked or did not smoke). Pre-pregnancy body mass index (BMI) (kg/m2) was computed using maternal weight and height. During the postnatal visits, a trained research assistant took measurements of the infant’s height and weight. The WHO growth charts for children age 0 to 60 months were used to estimate weight-for-length Z scores, specific to each sex ( Guevarra E 2019 ; World Health Organization 2006 ). Descriptive characteristics were determined for the entire study population, as well as female and male infants separately. Non-response analyses were further performed between participants and non-participants using student’s t-test, Mann-Whitney U test, or Chi-square test, as suitable. First, using multivariate linear regression models, we examined associations of natural log-transformed pregnancy-averaged phthalate exposure with infant AGD (mm). Then, we performed separate analyses for urinary phthalate concentration at <18 and 18–25 gestational weeks. Basic models were only adjusted for infant age at measurement and weight-for-length Z scores, whereas main models were additionally adjusted for maternal age, relationship status, race and ethnicity, maternal education, health insurance, alcohol drinking in pregnancy, environmental cigarette smoke in pregnancy, and pre-pregnancy BMI. For period-specific analysis based on gestational age period of exposure measurement (< 18 and 18–25 weeks of gestation), both basic and main models were further adjusted for gestational age at exposure measurement and batch. To investigate non-linear associations and dose-response relationships, pregnancy-averaged phthalate exposure was categorized into quartiles, and we used multivariate regression models to assess its relationship with infant AGD. Although prior studies suggested that maternal race and ethnicity may influence the relationship of prenatal exposure to phthalates with infant AGD, our sample size was too limited to conduct analyses stratified by racial and ethnic subgroups. To control for multiple comparisons, we used Benjamin-Hochberg correction with a false discovery rate (FDR) of less than 5% ( Benjamini and Hochberg 1995 ). Finally, we selected pregnancy-averaged phthalate exposure groups among females for mixture analysis to help reduce collinearity between individual phthalate metabolites. Group compound chemicals were standardized by calculating standard deviation scores to facilitate comparations between exposure. We employed partial-linear single-index (PLSI) models to explore the mixture effects of grouped phthalates on infant AGD, and quantified their contribution using the EPLSIM package ( Wang and Liu 2023 ; Wang et al. 2020 ). PLSI provides a unique parsimonious combination of chemicals as it incorporates a parametric component that allows us to quantify and rank the relative importance of each chemical, while simultaneously utilizing a flexible nonparametric link function to capture potential nonlinear associations and complex interactions between chemical mixture and infant AGD ( Wang and Liu 2023 ; Wang et al. 2020 ). Statistical analysis was performed using R version 4.3.0 ( R Core Team 2021 ).

Table 2 shows characteristics for all participants and for those with male and female children separately. Pregnant participants had a mean age of 31.8 years old, with a median pre-pregnancy BMI of 25.4 kg/m2. More than a third (38.5%) had a high school education or less, 43.1% were nulliparous, and 59.3% had public health insurance. More than half (58.5%) identified as Hispanic, 27.3% as non-Hispanic White, and 3.4% as non-Hispanic Black; 10.9% reported other or multiple races. Mothers of non-participant infants had lower pre-pregnancy BMI, were more likely to be nulliparous, were more likely to have consumed alcohol, and were less educated than mothers of infants included in the analysis ( Table S2 ). Table 1 and S1 presents the percent of values below the LOD for each phthalate metabolite and the median concentrations and interquartile range (IQR) for each group and metabolite. Variations in prenatal urinary phthalate metabolite concentrations were published previously ( Liu et al. 2022 ). Figure S3 illustrates Spearman correlation coefficients comparing phthalate metabolite concentrations during pregnancy. The median gestational age at delivery was 39.3 weeks and the mean birth weight was 3267 grams. In male infants, mean (SD) AGDas, AGDap and PW were 35.7mm (9.8), 75.3mm (12.2) and 14.7mm (2.7), respectively. In female infants, AGDaf and AGDac were 24.2mm (6.9) and 53.3mm (10.5), respectively ( Table 2 ). Table 3 shows the relationship between average prenatal exposure to phthalates and AGD in male infants. Each unit increase in natural log-transformed maternal average urinary concentrations of ∑HMWP, MECPP, and MEOHP was associated with −0.39mm (95%Confidence Interval (CI): −0.78, −0.01), −0.43mm (95%CI: −0.79, −0.06), and −0.40mm (95%CI: −0.73, −0.06) narrower PW, respectively. Nonetheless, after control for multiple testing, these associations did not remain significant. Table 4 displays the association between average prenatal exposure to phthalates and AGD in female infants. Each unit increase in natural log-transformed maternal average urinary concentrations of ∑DiNP was associated with 1.28mm (95%CI: 0.52, 2.03) longer AGDaf. Each unit increase in natural log-transformed average MECPP, MCiOP, and MCPP concentrations was associated with 1.39mm (95%CI: 0.28, 2.51), 1.36mm (95%CI: 0.56, 2.16), and 1.21mm (95%CI: 0.35, 2.07) longer AGDaf, respectively. After controlling for multiple comparisons, all associations remained statistically significant, except for MECPP. Each unit increase of natural log-transformed maternal average urinary concentrations of ∑HMWP, and ∑DEHP was associated with 2.18mm (95%CI: 0.55, 3.80), and 2.80mm (95%CI: 1.17, 4.44) longer AGDac, respectively. Furthermore, each unit increase of natural log-transformed average MEHP, MECPP, MEHHP, MEOHP, and MCPP concentrations was associated with 1.33mm (95%CI: 0.27, 2.39), 2.90mm (95%CI: 1.26, 4.54), 2.62mm (95%CI: 1.09, 4.16), 2.36mm (95%CI: 0.71, 4.01), and 1.87mm (95%CI: 0.57, 3.17) longer AGDac, respectively. After controlling for multiple comparisons, all associations remained statistically significant, except for MEHP. Tables 5 , and Figure 1 show overall non-linear positive associations between the mixture of grouped phthalates and female AGD. Both ∑DiNP and ∑DEHP contribute positively to explaining the variance in the AGDaf ( Table 5 , Figure 2A ) and in the AGDac ( Table 5 , Figure 2B ). Specifically, ∑DiNP accounts for 92.7% of the variance explained in the AGDaf ( Table 5 , Figure 2A ), and 99.1% of the variance explained in the AGDac ( Table 5 , Figure 2B ). The interaction plot showed significant interactions between the grouped metabolites ∑DiNP, ∑DEHP and ∑LMWP in relation to AGD in females, particularly in panels B and C ( Figure S4 and S5 ). Panel B shows interactions between ∑DiNP and ∑LMWP, and Panel C between ∑DiNP and ∑DEHP in association with AGD in female infants. Results of analysis of urinary phthalate concentrations at <18 weeks and 18–25 weeks separately were consistent with the main results using the averaged exposure ( Table S3 to S6 ). Modeling exposure as quartiles showed no indication of dose-response associations ( Table S7 and S8 ).

Our prospective study revealed no relationship between urinary concentrations of phthalate metabolites during pregnancy and AGD or PW in male infants. However, higher exposure to certain phthalates during pregnancy was associated with longer AGD in female infants. Specifically, our findings showed relationships of maternal urinary ∑DiNP metabolite concentrations during pregnancy with longer AGDaf, and ∑DEHP metabolite concentrations with longer AGDac. In the mixture analysis, associations with AGD among females were explained by exposure to ∑DiNP, and ∑DEHP. These findings suggest potential implications of exposure to endocrine-disrupting chemicals during critical periods on female sexual and reproductive development. Pregnant women are exposed daily to phthalates by inhalation, skin contact, ingestion, and intravenous medication ( Hauser et al. 2004 ; Lesseur et al. 2021 ; Mesquita et al. 2021 ). Phthalates have anti-androgenic, androgenic and estrogenic properties ( Harris et al. 1997 ; Inoshita et al. 2003 ; National Research Council 2008 ; Thibaut and Porte 2004 ), and can disrupt steroidogenesis ( Chen et al. 2013 ). Exposure to phthalates can alter maternal hormonal regulation and disrupt fetal sexual development in males (7–18 gestational weeks) and females (7–20 gestational weeks) ( Delbes et al. 2022 ; Engel et al. 2017 ; Gao et al. 2018 ; Katugampola et al. 2020 ; Lehraiki et al. 2009 ; Mogus et al. 2023 ; Rey and Picard 1998 ; Sharpe 2006 ; Skakkebaek et al. 2015 ; Skakkebæk et al. 2001 ; Zambrano et al. 2014 ). AGD is a sensitive biomarker of prenatal androgen and steroidogenic action during sexual development, and a key predictor of postnatal reproductive health outcomes ( Jain et al. 2018 ; Pan et al. 2021 ; Priskorn et al. 2018 ; Thankamony et al. 2009 ; Thankamony et al. 2016 ). Shorter AGD is linked to hypospadias, cryptorchidism, infertility, and prostate cancer in males ( Jain and Singal 2013 ; Singal et al. 2016 ; Thankamony et al. 2016 ), longer AGD is linked to menstrual cycle irregularities, endometriosis, and ovarian hyperandrogenism in females ( Mendiola et al. 2016 ; Mira-Escolano et al. 2014a ; Sánchez-Ferrer et al. 2017 ). The influence of prenatal phthalates exposure on AGD in males and females may be probably through different processes. Prior longitudinal investigations have revealed that prenatal phthalate exposure were linked to shorter AGD and narrower PW in male infants, with ∑DEHP the chemical most studied ( Adibi et al. 2015 ; Bornehag et al. 2015 ; Bustamante-Montes et al. 2013 ; Dorman et al. 2018 ; Martino-Andrade et al. 2016 ; Swan et al. 2015 ; Swan Shanna et al. 2005 ). Nonetheless, some studies found no association in male infants. A Danish study reported no relationship between phthalate exposure during pregnancy and AGD in male infants ( Jensen Tina et al. 2016 ). Another study in the United States reported racial differences in the associations; higher urine levels of MEHP during pregnancy was linked to shorter AGDap only in African-American mother-child pairs ( Wenzel et al. 2018 ). Participants in our study were mainly Hispanic (58%), followed by non-Hispanic White participants (27%). We found that prenatal exposure to DEHP metabolites (e.g., MEHP) was linked to narrower PW, but the association did not stay statistically significant after control for multiple comparisons, although the direction of the association was consistent with prior studies. Similar to our results, a prior study using NYU CHES data revealed no relationship between gestational phthalate exposure and fetus penile length ( Salvi et al. 2023 ). Among females’ infants, the four prior epidemiological studies that have been conducted were limited and had inconclusive findings ( Adibi et al. 2015 ; Arbuckle et al. 2018 ; Swan et al. 2015 ; Wenzel et al. 2018 ). We found that prenatal ∑DEHP, and ∑DiNP exposures were associated with longer AGD in females. Swan et al. and Adibi et al. found no relationship between first-trimester phthalate exposure and newborn AGD. However, the positive direction of their observed, particularly with metabolites of ∑DEHP and ∑DiNP, aligns with our findings ( Adibi et al. 2015 ; Swan et al. 2015 ). Arbuckle et al. in Canada found that first-trimester MEP exposure was linked to longer newborn AGDac, while MBzP was associated with shorter AGDac ( Arbuckle et al. 2018 ). Additionally, Wenzel et al. in the United States found that African-American pregnant participants (n=67) exposed to ∑LMWP during the second-trimester had newborns with shorter AGDac (feminizing effect), while White mothers (n=61) exposed to ∑LMWP had newborns with longer AGDac (masculinizing effect) ( Wenzel et al. 2018 ). Prior studies had only one measure of maternal urinary phthalate exposure; in most cases we were able to average two measurements, thereby reducing exposure misclassification. Differences between other studies and ours could be also explained by differences in the timing of exposure measurement, the racial and ethnic composition of the populations, sample sizes, and methodological approaches used to adjust for multiple comparisons. Prior animal models have shown that exposure to DiNP and/or DEHP for a period of ten days can lead to consequences in female reproductive health ( Chiang et al. 2020a ; b ; Laws et al. 2023 ). DiNP or DEHP, alone or mixed, have been linked to disruptions in estrous cyclicity, infertility, pregnancy loss and steroid hormones alteration later in life. Our findings cautiously suggest that there may be interactions between DiNP and LMWP, and DiNP and DEHP. This complex relationship between phthalate groups may indicate potential counteractive effects due to biological mechanisms. More studies are needed to understand the biological mechanisms by which DiNP and/or DEHP may interact and alter sexual development of the fetus. As most of the United States population is exposed to several phthalates, this has become a crucial problem of public health which required urgent solutions. Several potential mechanisms could explain our findings. First, phthalates disrupt normal androgen and estrogen signaling by binding to the AR, ER, luteinizing hormone receptor (Lhr), and follicle-stimulating hormone receptor (Fshr) during crucial windows of sexual development ( Engel et al. 2017 ; Lehraiki et al. 2009 ; Mogus et al. 2023 ; Pocar et al. 2012 ). Prenatal phthalate exposure interferes with the maternal-fetal ability to regulate sex hormones. Estrogens and androgens impact the ontogeny of key organs such as the brain, uterus, and gonads. By upregulating the hypothalamus, phthalates may change the female fetus’s sexual differentiation. Early pregnancy exposure to ∑DEHP and ∑LMWP has been positively linked to maternal estrogen (E1/E2) concentrations ( Sathyanarayana et al. 2017 ). Second-trimester ∑DiNP exposure has been inversely linked to maternal testosterone concentrations measured in hair samples during the early postpartum period, while third-trimester MBzP exposure has been positively associated with maternal hair cortisol, cortisone, 11-dehydrocorticosterone, and testosterone ( Mustieles et al. 2023 ). Moreover, prenatal phthalate exposure is negatively linked to cord-blood follicle-stimulating hormone concentrations among newborn males (MMP, MBP, MBzP, and ∑DEHP) and females (MMP, MBzP, and MEHP) ( Lu et al. 2024 ). Since the main pathway for masculinization is androgen signaling, females exposed to phthalates in utero may experience elevated androgenic signaling, contributing to the development of male-like traits, including longer AGD ( Hlisníková et al. 2020 ; Mira-Escolano et al. 2014b ). Moreover, phthalates may alter this balance by either mimicking estrogen or disrupting estrogenic pathways, which could enhance androgenic effects. Our study suggests that prenatal phthalates exposure may disrupt fetal sexual development in males and females through different mechanism. However, further research is needed to explore the underlying mechanisms behind of these associations. Second, phthalates may interfere with aromatase activity, the process by which androgens are converted to estrogens that is critical for sexual and hypothalamic/preoptic area (HPOA) development. Disruption of HPOA development may influence hormonal regulation and behavior later in life. Female rat pups prenatally exposed to phthalates show inhibition of aromatase activity in HPOA ( Andrade et al. 2006 ). Phthalates interfere with androgen (testosterone) and estrogen signaling, which is essential for the development of the fetal hypothalamus, amygdala, and insula. The interference with hypothalamus, amygdala, and insula development may result in impaired GnRH signaling, as well as impaired HPG-axis and Hypothalamic-Pituitary-Adrenal axis (HPA-axis) functioning, altering the sexual differentiation in brain and reproductive organ development ( Delbes et al. 2022 ; Gao et al. 2018 ; Katugampola et al. 2020 ; Rey and Picard 1998 ; Sharpe 2006 ; Skakkebaek et al. 2015 ; Skakkebæk et al. 2001 ; Zambrano et al. 2014 ). Thus, phthalates may disrupt the brain and reproductive development by interefering with the HPG-axis and HPA-axis, both directly and indirectly. Third, phthalates may influence gene expression. In animal studies, the ovaries of adult mouse progeny had reduced concentrations of mRNA for key steroidogenic enzymes and receptors when the dam’s diet throughout pregnancy was exposed to DEHP ( Pocar et al. 2012 ). In particular, the adult offspring’s ovaries showed lower levels of progesterone receptor, Fshr, Lhr, CYP19a1, and CYP17a1, ( Pocar et al. 2012 ). Phthalate exposure during pregnancy alters consequently ovarian steroidogenesis at several developmental stages ( Pocar et al. 2012 ). More studies should examine the underlying mechanisms by which phthalates influence fetal sexual development in females. This study had a prospective design, repeated measures of 15 phthalate metabolites during sexual development, and information on numerous potential confounders. Participants included in this study were comparable to those excluded in main characteristics ( Table S2 ). While our population was multiethnic, the study population was predominantly Hispanic and non-Hispanic White, limiting our ability to examine African-American participants. However, this study provides data on an ethnic group not as well represented in other studies (Hispanic population). Phthalate concentrations in our cohort study were equivalent to exposure in the common population according to Center for Disease Control and Prevention ( Environmental Protection Agency 2023 ). To reduce measurement error and increase precision in the exposure, we averaged phthalate metabolite concentrations measured at two time points. Specifically, during periods of susceptibility in the sexual development of the embryo and fetus ( Delbes et al. 2022 ; Engel et al. 2017 ; Gao et al. 2018 ; Katugampola et al. 2020 ; Lehraiki et al. 2009 ; Rey and Picard 1998 ; Sharpe 2006 ; Zambrano et al. 2014 ). However, the use of two-time points phthalate assessments may not represent the exposure period during pregnancy, so our findings should be interpreted with caution. Most phthalates detection rate in our study was about 70%, suggesting that fill in methods to handle values below LOD can be a good approach to minimize misclassification ( Seok et al. 2024 ). However, findings with metabolites with higher uncertainty should be cautiously interpreted. A good approach to replace values below LOD with lower uncertainty may be multiple imputations given the availability of data on predictors exposure, but unfortunately, we did not have data from questionnaires on dietary sources of phthalates or home assessment for exposure to phthalates. As some phthalate metabolites (i.e. MCPP) may be metabolites from different group compounds ( Calafat et al. 2006 ; Gomez Ramos et al. 2016 ; Kasper-Sonnenberg et al. 2025 ), we did not create group compounds using this metabolite. Additionally, to facilitate the replication of our study, we provided analysis for both the 15 individual metabolites and the group compounds. To minimize error in our AGD measures, we averaged three measurements taken by the same trained researcher ( Lee and Jacobs 2015 ). Additionally, we performed mixture analysis with grouped phthalate chemicals to enhance the robustness of our associations ( Carlin Danielle et al. 2013 ; Wang and Liu 2023 ). At the extreme concentration range, the limited number of participants may pose a statistical challenge, interpreting the chemical associations and slopes in the mixture analysis need caution. Finally, despite adjusting the associations between pregnancy phthalate exposure and infant AGD for potential confounders, residual confounding might still be present. In the past decade or so, DEHP has been regulated in Europe, Canada, and the United States, replaced by DiNP and DnOP in consumer products. Recently, the Environmental Protention Agency (EPA are considerating to prohibite the use of DiNP because of occupational risk in the United States. However, no further regulations exist, including no particular consideration for vulnerable populations ( Environmental Protection Agency 2024 ). Our findings add to evidence that prenatal DEHP, and DiNP exposure may disrupt with fetal sexual development in females, suggesting the need to revise the regulation of phthalates beyond DEHP. Phthalates and their metabolites are present in numerous products used during our daily lives, and our findings underscore the necessity of regulatory measures to limit phthalate exposure during crucial development windows. Although the spectrum and severity of negative reproductive outcomes associated with phthalates may be dependent on time or dosage of exposure in critical windows of sexual development, this study did not find evidence of a threshold impact or dose-response relationship between maternal urinary phthalate concentrations and infant AGD. Early pregnancy phthalate concentrations, mainly ∑DEHP, and ∑DiNP, were associated with longer AGD in female infants, suggesting the need for regulation of phthalates beyond DEHP. Phthalates might play a role in programming sexual development and reproductive function. More longitudinal studies in multiethnic cohorts are required to replicate our findings and comprehend the biological mechanisms underlying these associations.

Despite being recognized for their endocrine-disrupting properties, phthalates are regularly used as plasticizers in polyvinyl chloride plastics and consumer goods ( Hauser et al. 2004 ; Lesseur et al. 2021 ; Mesquita et al. 2021 ). Phthalates are shown to have anti-androgenic and estrogenic effects in animal and in vitro studies ( Mesquita et al. 2021 ; National Research Council 2008 ; Radke et al. 2018 ; Sathyanarayana et al. 2017 ). Exposure to phthalates in humans is associated with adverse reproductive health outcomes in males (e.g., reduced sperm function and infertility) and females (e.g., primary ovarian insufficiency and endometriosis) ( Mesquita et al. 2021 ; Radke et al. 2018 ). Epidemiological studies have also shown that prenatal phthalate exposure is associated with impaired fetal growth and preterm birth ( Cowell et al. 2023 ; Huang et al. 2014 ; Santos et al. 2021 ; Trasande et al. 2024 ; Welch et al. 2022 ), suggesting that developmental exposure to phthalates may have long-lasting health effects on offspring, including reproductive health ( Freire et al. 2024 ; Kim et al. 2024 ; Sathyanarayana et al. 2016 ). Phthalates can cross the placental barrier and subsequently interact with steroid receptors (i.e., androgen [AR] and estrogen receptors [ER]) during embryonic and fetal development ( Engel et al. 2017 ; Lehraiki et al. 2009 ; Mogus et al. 2023 ). Phthalate exposure may influence in utero androgen programming and disrupt hypothalamic, pituitary, and reproductive development that starts as early as 6–7 weeks of gestation ( Delbes et al. 2022 ; Gao et al. 2018 ; Katugampola et al. 2020 ; Rey and Picard 1998 ; Sharpe 2006 ; Skakkebaek et al. 2015 ; Skakkebæk et al. 2001 ; Zambrano et al. 2014 ). Growing literature proposes that prenatal exposure to phthalates may lead to changes in the sexual development of embryos and fetuses via dysregulation of the hypothalamic pituitary gonadal (HPG) axis ( Delbes et al. 2022 ; Rey and Picard 1998 ; Sharpe 2006 ; Skakkebaek et al. 2015 ; Skakkebæk et al. 2001 ). As sexual development predominantly occurs between 8 and 18 weeks of gestation in males and 8 and 20 weeks in females ( Delbes et al. 2022 ; Rey and Picard 1998 ; Sharpe 2006 ; Skakkebaek et al. 2015 ; Skakkebæk et al. 2001 ), phthalate exposure during this critical period may result in long-lasting effects on offspring sexual and reproductive outcomes. Anogenital distance (AGD), the distance between anus and genitals, is a key postnatal marker of in utero androgen exposure and reproductive development in animals and humans ( Jain et al. 2018 ; Priskorn et al. 2018 ; Thankamony et al. 2009 ; Thankamony et al. 2016 ). According to previous literature AGD is fixed in early pregnancy and unaffected by androgen levels during the postnatal period ( Dean and Sharpe 2013 ; Jain et al. 2018 ; Pan et al. 2021 ; Thankamony et al. 2016 ). Thus, AGD could be useful to detect several clinical implications, particularly for reproductive organs in both males and females ( Dean and Sharpe 2013 ; Jain et al. 2018 ; Pan et al. 2021 ; Thankamony et al. 2016 ). Phthalates are known for their antiandrogenic effects, for this reason several epidemiological studies primarily focused on the association between prenatal phthalate exposure and AGD in males. These previous studies have consistently reported that prenatal exposure to phthalates (i.e., di(2-ethylhexyl)phthalate [∑DEHP]), diisononyl phthalate [∑DiNP], and low molecular weight phthalates [∑LMWP]) is associated with shorter AGD in male newborns and infants ( Bornehag et al. 2015 ; Bustamante-Montes et al. 2013 ; Dorman et al. 2018 ; Martino-Andrade et al. 2016 ; Swan et al. 2015 ; Swan Shanna et al. 2005 ), but the majority of these studies did not include females ( Bornehag et al. 2015 ; Bustamante-Montes et al. 2013 ; Dorman et al. 2018 ; Martino-Andrade et al. 2016 ; Swan Shanna et al. 2005 ). Phthalates have also shown to have estrogenic and adrenergic effects ( Lu et al. 2024 ; Mustieles et al. 2023 ; Sathyanarayana et al. 2017 ), which suggest that females AGD could be also influenced by prenatal phthalate exposure. The few prospective studies that have included females have had inconclusive results ( Adibi et al. 2015 ; Swan et al. 2015 ; Wenzel et al. 2018 ). One study (n=373) did not find relationship between prenatal phthalate exposure and AGD among females ( Swan et al. 2015 ). Another study (n=275) reported that maternal urinary concentrations of mono-benzylphthalate (MBzP) in the first-trimester of pregnancy were linked to longer AGD in newborn females, but previous reports did not adjust for socioeconomic and lifestyle factors ( Adibi et al. 2015 ). A small study of 187 African-American mother-child pairs found that maternal urinary concentrations of monoisobutyl phthalate (MiBP) during the second-trimester were associated with longer AGD among females, but this association did not survive adjustment for confounders ( Wenzel et al. 2018 ). These inconsistent findings may be attributable to several methodological challenges, such as small sample size, measurement of non-persistent chemicals at a single time point, and lack of consideration for important confounders. Additionally, none of the earlier studies examined phthalates as mixture. Therefore, this study examined the association of prenatal phthalate exposure with AGD in male and female infants in a prospective birth cohort in New York City (NYC). We focused on phthalate metabolite measurements in early and mid-gestation to capture the critical period for genital and sexual development ( Bustamante-Montes et al. 2013 ; Delbes et al. 2022 ; Rey and Picard 1998 ; Sathyanarayana et al. 2017 ; Skakkebaek et al. 2015 ; Skakkebæk et al. 2001 ). We explored whether prenatal exposure to a mixture of phthalates exhibits additive or synergistic effects on AGD and the relative importance of phthalates in the mixture.