Filling gaps in population estimates of phthalate exposure globally: A systematic review and meta-analysis of international biomonitoring data.

This systematic review and meta-analysis follows the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) guidelines. Our protocol was registered in the International Prospective Register of Systematic Reviews (PROSPERO) on February 27 th , 2024 (registration code: CRD42023481172). We searched PubMed, Ovid, and Web of Science for Boolean combinations of the following MeSH terms: “Phthalates”, “Mono-2-Ethylhexyl Phthalate”, “MEHP”, “Mono-2-ethyl-5-hydroxyhexy Phthalate”, “MEHHP”, “Mono-2-ethyl-5-oxohexyl Phthalate”, “MEOHP”, “Mono-2-ethyl-5-oxohexyl Phthalate”, “MEOHP”, “Mono-n-butyl Phthalate”, “MnBP”, “Mono(2-ethyl-5-carboxypentyl) Phthalate”, “MECPP”, “Mono(2-carboxymethylhexyl) Phthalate”, “MCMHP”, “Mono-carboxy-isooctyl Phthalate”, “MCOP”, “Mono-isononyl Phthalate”, “MiNP”, “Mono-oxo-isononyl Phthalate”, “MOiNP”, “Mono-hydroxy-isononyl Phthalate”, “OH-MiNP”, “Mono-carboxy-isononyl Phthalate”, “MCiNP”, “Mono(3-carboxypropyl) Phthalate”, “MCPP”, “Mono-octyl Phthalate”, “MOP”, “Monobenzyl Phthalate”, “MBzP”, “Mono-isodecyl Phthalate”, “MiDP”, “Mono-ethyl Phthalate”, “MEP”, “Mono-isobutyl Phthalate”, “MiBP”, “Di(2-ethylhexyl) Phthalate”, “DEHP”, “Diethyl Phthalate”, “DEP”, “Benzyl butyl phthalate”, “BBP”, “Diisononyl Phthalate”, “DiNP”, “Diisodecyl Phthalate”, “DiDP”, “Di-n-octyl phthalate”, “DnOP”, “Dibutyl Phthalate”, “DBP”, “Endocrine Disruptors”, “Endocrine Disruption”, “Endocrine Disrupting Chemical”, “EDs”, and “Xenoestrogens”. These terms were combined with relevant biological sampling vocabulary such as “Urine,” “Plasma,” “Serum,” “Human Tissue,” “Body Fluids,” “Saliva,” “Sweat,” and “Blood.” Additional combinations included “Concentrations,” “Levels,” “Biomonitoring,” “Exposure,” and “Human Exposures.” Following a similar format of a previously performed study, we selected studies that reported a population-level concentration measure (arithmetic mean [AM] and standard deviation [SD], geometric mean [GM] and 95% confidence interval [CI], or median and interquartile range [IQR]) of at least one of the urinary phthalate metabolites of interest, were written in English, and were published between January 1 st , 2000, and December 1 st , 2023. We excluded articles with non-human species and populations from locations where representative data are already available at the country level (i.e., the US, Canada, and Europe). Additionally, we excluded articles and data from cohorts that aimed to control or alter the participants’ phthalate exposure or cohorts restricted to participants with a particular health condition such as diabetes, infertility, asthma, endometriosis, or cancer (thyroid, breast, and prostate). We also excluded occupational studies, as we were interested in tracking routine exposure among the general population. If two or more analyses were based on the same cohort, the one with the highest number of participants was used. If the analyses had the same population size, we selected the article that reported on the greatest number of phthalate metabolites. In addition, we excluded articles that only provided details regarding diester phthalate molar concentrations instead of the individual metabolite urinary concentrations. Lastly, we excluded papers that lacked sufficient information to perform a regression analysis ( Acevedo et al., 2025 ). After removing duplicates, two reviewers (J.M.A. and M.S.R.) used Covidence (Melbourne, Australia) and EndNote X21 (Clarivate; Berkeley, California, US) to manage and review each article. They assessed each article’s title and abstract for relevancy and selected articles for a full-text screening. If an article’s eligibility was in question, a third reviewer (A.C.) was consulted to make the final decision. Upon completing the full-text screening, the reviewers extracted publication details such as author(s), publication year, sample size, and cohort characteristics (e.g., age range, geographical region, pregnant vs. non-pregnant population) from the eligible articles. The geographical areas were determined based on where the study’s sampling occurred (Latin America, Africa, Asia, or Australia; Table 1 ) (The World Bank, 2024). Due to the high quantity of analyses conducted among Asian populations, we used World Bank definitions to divide the region into sub-regions: Eastern and Pacific Asia (EPA) and Middle East and South Asia (MESA; Table 1 ) ( World Bank Group, 2024a ; World Bank Group, 2024b ; World Bank Group, 2024c ). We categorized study populations into the following age groups: “youth” (age <18 years), “adults” (18 years and older), and “Mixed” (age range includes both youth and adults). Additionally, we documented whether the analysis was conducted in a pregnancy cohort. Articles that provided phthalate measures for more than one subset of interest were allowed to contribute multiple data points to this analysis. For instance, articles providing measures for pregnant women and their children were recorded as two distinct observations (“youth” and “pregnant adults”). Along with publication and population details, outcome information was recorded, including the phthalate metabolite, chemical concentration, analytical method’s limit of detection (LOD), limit of quantification (LOQ), distribution percentiles, and sampling year. If a phthalate metabolite concentration was below the detectable values, we imputed the value as LOD/√2. In some cases, authors used LOQ instead of LOD to calculate phthalate metabolite concentrations below the detectable values. If a phthalate metabolite displayed a detection rate of 60% or less in majority of their cohorts, we excluded them from the analyses while still documenting how frequently they appeared in the literature. We selected the following phthalate metabolites for inclusion in our analysis based on how frequently they are in the literature review and how prevalent their parent compound is in industrial use: mono(2-carboxymethylhexyl) phthalate (MnBP); mono(2-ethyl-5-carboxypentyl) phthalate (MEHP); mono(2-ethyl-5-hydroxyhexyl) phthalate (MEHHP); mono(2-ethyl-5-oxohexyl) phthalate (MEOHP); mono(2-ethyl-5-carboxypentyl) phthalate (MECPP); mono(2-carboxymethylhexyl) phthalate (MCMHP); mono-carboxy-isooctyl phthalate (MCOP); mono-isononyl phthalate (MiNP); mono-oxo-isononyl phthalate (MOiNP); Mono-hydroxy-isononyl phthalate (OH-MiNP); mono-carboxy-isononyl phthalate (MCiNP); mono(3-carboxypropyl) phthalate (MCPP); mono-octyl phthalate (MOP); monobenzyl phthalate (MBzP); mono-isodecyl phthalate (MiDP); mono-ethyl phthalate (MEP); and mono-isobutyl phthalate (MiBP). While we ideally would have assessed exposure to the parent compounds or calculated the sum concentration of the parent compounds such as DEHP, studies did not consistently measure the same phthalate metabolites for each parent compound; thus, comparing across studies would have been impossible ( Supplement Table 1 ). We, therefore, present results for individual metabolites throughout, highlighting parent compounds when possible. Phthalate metabolite concentrations were documented as AM (± SD), GM (95% CI), and/or median (IQR or minimum/maximum). The sampling year was defined as the year urine samples were collected or the midpoint if the study was conducted over more than one year. While the majority of studies reported phthalate metabolite concentrations in their unadjusted state (ng/mL), some articles only provided phthalate concentrations adjusted for creatinine (ng/mg) or specific gravity (ng/mL). For studies that reported only specific gravity (SG)-adjusted phthalate concentrations, we applied a 1:1 ratio, using the SG value in place of the unadjusted phthalate concentration. To harmonize the studies that reported only creatinine-adjusted concentrations with those that reported creatinine-unadjusted concentrations, we used the following equation ( Park et al., 2016 ): P h t h a l a t e C o n c . ( n g m L ) = P h t h a l a t e C o n c . C r e a t i n i n e ( n g m g ) ∗ C r e a t i n i n e C o n c . ( m g m L ) Equation 1 : Adjustment for creatinine levels within urinary biomonitoring samples. We used previously published age-specific creatinine concentrations when applying this equation to approximate unadjusted concentrations ( Park et al., 2016 ). If a study reported a phthalate metabolite concentration for a population that included two or more of the reference source’s predefined age groups, we used the average creatinine level for the general population, 0.91 mg/mL ( Park et al., 2016 ) ( Supplemental Table 2 ). We chose to use AM phthalate concentrations in our analysis as a measure to assess the variability across different populations, regions, and age groups within the general population’s exposure to phthalates. AM concentration is simpler for many laboratories to obtain with less available resources, especially in LMIC regions. When the phthalate concentration’s AM was absent, we used either the study’s GM or median value as a substitute. We also estimated missing SD values for several studies by dividing the IQR by the empirical constant of 1.35 (SD = IQR /1.35) ( Higgins, 2008 ) or by using the median, maximum, and minimum values, following methods published previously ( Wan et al., 2014 ). Study-specific standard errors (SE) were required for our weighted regression models. If missing, we used the formula SE = SD/√n, with n being the total study population. When the median was substituted for the AM, we applied the formula SE = (1.253*SD)/√n to adjust for variability caused by the distribution of the sample median ( Hojo and Pearson, 1931 ). In bivariate analysis, we used ANOVA and t-tests to compare phthalate metabolite concentrations across international regions, age groups, and pregnant vs. non-pregnant populations. We used mixed-effects regression models to examine the change in urinary phthalate metabolite concentrations across the years in which the studies sampled their populations, adjusted for age group, pregnancy status, and region. We also performed a covariate-adjusted meta-regression model weighted for study-specific SE to enhance the precision of the estimated associations between time and each mean phthalate concentration ( DerSimonian and Laird, 1986 ). In addition, we stratified our meta-regression models by region to examine region-specific trends in phthalate exposure over time. We then repeated stratified covariate-adjusted meta-regression analyses including a quadratic term for time (year 2 ). By adding the quadratic term, we were able to evaluate whether there was a non-linear relationship between time and mean phthalate concentration and, if so, the shape of the exponential curve. To avoid multicollinearity when adding the quadratic term to the meta-regressions, we centered the sampling year variable by subtracting the study’s sampling year from the calculated average sampling year. To analyze heterogeneity, we calculated the pooled mean concentration and 95% CI for each phthalate globally and by region while also obtaining their Cochran’s Q and I 2 . If heterogeneity was high (70 to 100%), we implemented the trim-and-fill method to counteract the over- or underestimation resulting from heterogeneity ( Wang et al., 2012 ). We calculated the combined and region-specific predicted mean and SD exposure trend for each phthalate metabolite across five periods (2003, 2008, 2013, 2018, and 2023) to observe the shift in exposure over time. With the predicted mean and SD value, we also calculated the shift in estimated percentiles of concentration (10 th , 25 th , 50 th , 75 th , and 95 th ) using the inverse of the normal cumulative distribution at each time point. Lastly, to determine whether the use of AM as our primary measure significantly influenced the outcome (change in effect estimate >15%), we repeated our analysis excluding studies that reported only AM values. All hypothesis tests were two-sided, and we used a Type 1 error rate of 0.05 to determine statistical significance. Analyses were conducted using Stata statistical software (version 16.0, College Station, TX) and R (version 4.3.1, Vienna, Austria).

Our search yielded 27,759 articles ( Figure 1 ). After removing duplicates and conducting the initial title and abstract screening, we selected 1,928 articles for full-text screening. We excluded 1,714 articles during the process due to the articles failing to meet the eligibility criteria previously stated in section 2.3 and seen in Figure 1 . Additionally, we identified two secondary citations obtained from the literature that met the criteria for the study. Between the primary and secondary citations, we included a total of 216 articles with data from specified regions in our final analysis. Of the 216 articles in our final analysis, over 60% reported urinary concentrations (detectable or below the LOD) for at least one of the phthalate metabolites of DEHP (MEHP, MEHHP, MEOHP, MECPP, and MCMHP), BBP (MBzP), and DBP (MnBP). In contrast, fewer than 3% of articles provided enough information to report on a DiDP metabolite concentration (MCiNP and MiDP). MEHHP and MnBP were the most frequently reported phthalate metabolites across the review, having urinary concentrations reported in 191 articles. Nearly 60% of studies were conducted in the EPA region, and at least 60% of studies focused on non-pregnant populations. Among the age groups, adults (age 18 years and older) were the most frequently reported overall ( Supplemental Table 3 ). Due to the low detection rates (< 60%) reported by a majority of the cohorts that measured MiNP, MiDP, and MOP, we could not evaluate these metabolites further. Pregnancy, population type, and region were all associated with differences in urinary phthalate concentrations, supporting their inclusion in multivariable models. Regional differences were particularly notable for the metabolites of DEHP (MEHP and MEHHP), DiDP (MCiNP), DnOP (MCPP), DBP (MnBP), and BBP (MBzP) (p-value < 0.05), with the MESA region showing the highest concentrations across all of these metabolites, for example MEHP (57.95 ng/mL ± 64.02) and MnBP (160.48 ng/mL ± 120.93) compared with Australia, which had the lowest concentrations (MEHP: 4.00 ng/mL ± 0.14; MnBP: 28.35 ng/mL ± 9.40). Metabolites of DEHP (MEHP, MEOHP, MECPP, and MCMHP), DiNP (OH-MiNP), and DiBP (MiBP) displayed significant differences by age group (p-value < 0.05), with youth and mixed cohorts exhibiting higher concentrations compared with adults, particularly for DEHP metabolites (MEOHP: 34.01 ng/mL ± 23.61 vs. 19.64 ng/mL ± 69.23; MECPP: 51.91 ng/mL ± 36.33 vs. 19.73 ng/mL ± 15.31), and MiBP (42.92 ng/mL ± 58.57 vs. 23.68 ng/mL ± 25.87). Pregnant populations had the lowest concentrations of DEHP (MEOHP: 18.20 ng/mL ± 24.70; MECPP: 34.00 ng/mL ± 29.80) and DiBP (MiBP: 35.00 ng/mL ± 42.70) metabolites ( Table 2 ). When we performed mixed-effects regression models to examine the change in urinary phthalate metabolite concentrations over time, we found evidence of increasing exposure to the low molecular weight (LMW) phthalate, DBP as indicated by a significantly linear increase in MnBP over time across all three models (unadjusted, adjusted, and weighted for SE and adjusted for covariates; weighted model’s beta: 2.90 ng/mL/year; 95% CI: 1.35, 4.45). In contrast, MEP and the BBP metabolite MBzP displayed significant linear declines over time in weighted models (MEP beta: −1.23 ng/mL/year; 95% CI: −2.32, −0.14; MBzP beta: −0.28 ng/mL/year; 95% CI: −0.46, −0.11) ( Table 3 & Figure 2 ). We also observed linear declines over time in DEHP metabolites MEOHP, MEHHP, and MEHP concentrations in the weighted models (MEHP beta: −0.28 ng/mL/year; 95% CI: −0.56, −0.01; MEHHP beta: −0.64 ng/mL/year; 95% CI: −1.23, −0.05; MEOHP beta: −0.73 ng/mL/year; 95% CI: −1.11, −0.35) ( Table 4 & Figure 3 ). Upon adding a quadratic term for time to the weighted model, we observed a significant exponential increase in MnBP over time while no other LMW metabolite showed such a trend (beta: 0.65 ng/mL/year 2 ; 95% CI: 0.39, 0.92) ( Table 3 & Figure 2 ). Among high molecular weight (HMW) metabolites, the DEHP metabolite MEHHP showed a significant exponential decline over time with the addition of the quadratic term (beta: 0.10 ng/mL/year 2 ; 95% CI: 0.01, 0.20). Addition of the quadratic term also revealed a nonlinear increase in the DEHP metabolite MCMHP (beta: 0.33 ng/mL/year 2 ; 95% CI: 0.10, 0.55) and decrease in the DnOP metabolite MCPP (beta: −0.02 ng/mL/year 2 ; 95% CI: −0.03, −0.001) ( Table 4 & Figure 3 ). Region-specific linear and quadratic weighted models revealed similar temporal trends for the DBP metabolite MnBP, which increased exponentially (beta: 0.78 ng/mL/year 2 ; 95% CI: 0.50, 1.05) and the BBP metabolite MBzP, which decreased linearly (beta: −0.33 ng/mL/year; 95% CI: −0.50, −0.16) in the EPA region. Based on these values, we estimate that DBP exposure will increase by over 2-fold, while BBP exposure will decrease by over 2-fold by the year 2033 in the EPA region. The DEP metabolite MEP was the only metabolite to display a significant linear association between time and concentration in Latin America (beta: −10.62 ng/mL/year; 95% CI: −21.03, −0.2) and Africa (beta: −8.98 ng/mL/year; 95% CI: −9.69, −8.26). These MEP beta values suggest that DEP exposure will decrease 1.5-fold in Latin America and Africa by 2033 ( Table 5 and Supplement Figures 1 – 6 ). Significant linear and exponential declines were identified in the EPA region for DEHP metabolites, especially MEOHP (beta: −0.80 ng/mL/year; 95% CI: −1.17, −0.43) and MCMHP (beta: 0.32 ng/mL/year 2 ; 95% CI: 0.11, 0.53). We also noticed the mean concentration of the DnOP metabolite MCPP linearly decreased over time (beta: −0.22 ng/mL/year; 95% CI: −0.42, −0.02) in the EPA region and the concentration of the DiNP metabolite MCiNP exponentially decreased over time (beta: 0.16 ng/mL/year 2 ; 95% CI: 0.10, 0.22) in the EPA region. Like BBP, we can estimate that DiNP, following the MCiNP time trend, and DnOP exposure will decrease approximately 2-fold in the EPA region by 2033. Significant exponential declines in DEHP metabolite MCMHP (beta: 29.79 ng/mL/year 2 ; 95% CI: 24.73, 34.86) and MECPP (beta: 0.73 ng/mL/year 2 ; 95% CI: 0.05, 1.40) concentrations over time were also identified in the MESA region. Unlike the other parent compounds, DEHP exposure could not be determined entirely for 2033 due to the variability in urinary metabolite significance. While some DEHP metabolites (MEHHP and MEOHP) indicate an estimated 2-fold decrease in exposure, the majority of DEHP metabolites (MEHP, MECPP, and MCMHP) indicate a 2-fold or more increase in exposure by the year 2033. This is most notable for the DEHP metabolite MCMHP, showing an exponential decrease before 2015 but an exponential increase after and an approximately 4-fold increase by 2033 ( Table 6 and Supplement Figures 5 – 8 ). Further details on the estimates of exposure distribution throughout the years and their percentiles across all regions combined and by specific region are presented in Figures 2 and 3 ; Supplement Tables 4 - 6 ; and Supplement Figures 1 – 10 . Nearly all the metabolites’ measurements had high levels of heterogeneity (75 −100%) across the studies, as indicated by high Cochran’s Q and I 2 values. After implementing the Trim-and-Fill method to reduce publication bias, we noticed that almost all the phthalate metabolites’ mean concentrations internationally and based on region were either over- or underestimates of their mean concentrations before correction. Some concentrations such as MCMHP did not shift after these methods were implemented ( Supplement Tables 7 ). In sensitivity analyses restricted to studies that reported results as GM or medians, some beta coefficients shifted by 15% or more compared with the main regression results, indicating that relying on AM values influenced the observed temporal trends. MECPP and MCMHP concentrations, for example, failed to show significant non-linear time trends in the EPA region (p-value > 0.05). While some of the associations lost their statistical significance with the exclusion of the AM data, we noticed some of the betas were comparable to the main analysis. For example, DBP metabolite MnBP still displayed an exponential significant increase over time overall (beta: 0.40 ng/mL/year 2 ; 95% CI: 0.17, 0.62) and in the EPA region (beta: 0.51 ng/mL/year 2 ; 95% CI: 0.27, 0.75) while the DEHP metabolite MEOHP significantly declined linearly over time in the EPA region (beta: −1.10 ng/mL/year; 95% CI: −1.50, −0.70). We also observed additional trends when we excluded AM values. For example, DEHP metabolite MECPP and MCMHP displayed a declining linear trend over time internationally (MECPP: [beta: −1.19 ng/mL/year; 95% CI: −2.20, −0.18] and MCMHP: [beta: −0.71 ng/mL/year; 95% CI: −1.39, −0.03]) and in the EPA region (MECPP: [beta: −1.40 ng/mL/year; 95% CI: −2.59, −0.20] and MCMHP: [beta: −0.71 ng/mL/year; 95% CI: −1.40, −0.01])( Supplement Table 8 ).

In conclusion, this review reveals the widespread prevalence of exposure to some of the phthalates used in industry practices. Between 2000 and 2023, urinary metabolites of phthalates such as DEHP and DnOP declined across regions and within the EPA region. DEP exposure decreased in some regions, as evidenced by the significant decline of MEP over time. Our study shed some light on phthalate exposure trends over time and reveals gaps in biomonitoring data in various regions, particularly in regions with limited research infrastructure. Additional studies filling these gaps and estimating attributable disease burden and costs at regional and global levels are needed to show these EDCs’ impact on the public. In addition, increased discussion on regulations to lower or prevent the use and consumption of any form of phthalate is necessary in all regions to counteract the continuous growth of phthalate exposure. This, in turn, will contribute to aiding in the protection of human health through advocacy for measures to eliminate exposure to harmful environmental chemicals.

This systematic review is among the first to examine urinary phthalate metabolite concentration trends in non-occupationally exposed populations outside the US, Canada, and Europe. Throughout the period in which included studies were conducted (2000 – 2023), we observed significant differences in phthalate metabolites concentration across regions, age groups, and pregnancy status. Furthermore, we observed declining exposure to DEHP, BBP, and DEP both across and within regions. We also observed increases in DBP exposure, which was evident through the significant linear increase in its metabolite, MnBP, over time. Finally, we noticed that the results of sensitivity analyses which were restricted to GM or median values were not completely consistent with the main analysis findings. These differences indicate that caution is warranted when extrapolating results of analyses conducted using one measure to the other. When comparing our results to those seen in the US, Canada, and Europe, the majority of our temporal declines are consistent with what has been reported in these biomonitored regions, specifically in terms of DEHP and DnOP metabolites ( Health Canada, 2021 ; Kasper-Sonnenberg et al., 2025 ; Vogel et al., 2023 ; Zota et al., 2014 ). Prior studies analyzing the temporal trends in some of the regions evaluated in our review attribute the decline in phthalate exposure to the implementation of bans and regulations on phthalates ( Domínguez-Romero et al., 2023 ; Lyu et al., 2022 ). In the early 2000s, the Ministry of Health in Japan prohibited the use of DEHP, DiNP, and other phthalates in food packaging, children’s toys, and PVC gloves ( Lyu et al., 2022 ; Tsumura†* et al., 2003 ). In 2017, China restricted 16 phthalates in food products and packaging as well as followed several other countries like the US, Canada, Israel, Argentina, and Brazil in banning of children’s toys containing more than 1% by weight of DEHP, BBP, DBP, or other phthalates ( Wang and Qian, 2021 ). In the early 2000’s, the Ministry of Health in some Latin American countries banned the manufacture, import, export, and any other trade or charitable use of toys, flexible chewable baby merchandise, or plastic childcare products containing greater than 0.1% of various phthalates ( Tsang, 2008 ; Wu et al., 2020 ). These regulations are likely reasons for observed decreases in concentrations of these phthalate metabolites. Several prior studies have also identified significant shifts within the plasticizer market, particularly in the substitution of HMW phthalates ( Kasper-Sonnenberg et al., 2019 ; Lemke et al., 2021 ; Lessmann et al., 2019 ; Silva et al., 2019 ; Zota et al., 2014 ). DEHP, known for being the most used plasticizer globally, is increasingly being replaced by other phthalates such as DiNP and DiDP or non-ortho phthalate substitutes such as DEHTP, and non-phthalates such as cyclohexane-1,2-dicarboxylic acid-diisononyl ester (DINCH), and di(2-ethylhexyl) adipate (DEHA), which now represent approximately 30–60% of the US and Europe’s total plasticizer production. With these substitutions, urinary concentrations of DEHP metabolites such as MEHP, MEOHP, and others have diminished throughout the years ( Kasper-Sonnenberg et al., 2019 ; Lemke et al., 2021 ; Zota et al., 2014 ). A prior study performed in Mexico showed the effects of such substitutions by documenting changing concentrations of di(2-ethylhexyl) terephthalate (DEHTP), a well-documented alternative to DEHP. Between 2007 and 2010, DEHTP metabolite concentrations in the study’s participants increased while DEHP metabolite concentrations remained stagnant ( Wu et al., 2020 ). Prior studies performed in EPA regions and Saudi Arabia also showed similar outcomes, with alternative phthalate substitutes like DEHTP and DINCH increasing in concentration while concentrations of legacy phthalate like DEHP remain stagnant over time ( Ketema et al., 2023 ; Lee et al., 2021 ). While the MnBP trend we identified does not match the pattern seen in the US and Europe, a similar study analyzing the temporal and spatial trend of phthalate exposure in China between 2005 and 2020 does report similar findings ( Zhang et al., 2024 ). This study notes that European regions use HMW and eco-friendly phthalates like di-n-nonyl Phthalate (DNP) and didecyl phthalate (DDP) ( ECHA, 2013 ). In contrast, the main phthalates used in China, DEHP and DBP, comprised 40% of all phthalates used in 2019 ( Du and Stern, 2021 ), with common use in food packaging and processing ( Kang et al., 2023 ; Zhang et al., 2024 ). Ultra-processed foods and packaged foods, particularly seafood, are often packaged with plastic containing DEHP and DBP, from which phthalate leaches, contaminating the food product ( Alak et al., 2024 ; Tian et al., 2022 ). Given that Asian countries account for approximately two-thirds of global seafood production ( Hosomi et al., 2012 ), the EPA region could be more highly exposed to phthalates by consumption of seafood. These meals are also consumed at a higher rate by younger people rather than adults and pregnant individuals, thus increasing their risk of phthalate exposure as well ( Lin et al., 2011 ; Liu et al., 2022 ; Yang et al., 2019 ). Studies have also highlighted the impact of urbanization and agricultural practices on the public’s exposure to phthalates through the soil and air. Studies analyzing the soil in urban China indicate that high levels of urban litter and domestic trash, which can contain approximately 50% plastic content, offer an opportunity for phthalates to leach into the soil and groundwater ( Zeng et al., 2009 , 2020 ). The use of agriculture films, a standard tool used to protect crops, also contributes to phthalate exposure through the soil since they contain 10 to 60% plasticizers by weight to extend their longevity in the fields ( Ding et al., 2021 ; Zhou et al., 2020 ). One study analyzed the formula of 110 agricultural pesticides and found that phthalates accounted for 2 to 40% of the total organic solvent content ( Tang, 2020 ). Along with pesticides and agricultural films, fertilizers play a critical role in amplifying phthalate exposure rates. They are often packaged with high concentrations of plasticizers and create chemical deposits in the soil that accumulate over time ( Li et al., 2023 ; Mo et al., 2008 ; Zhang et al., 2015 ). These plastic additives and heightened urbanization contaminate soil, crops, and water sources, indirectly increasing the general public’s exposure to plasticizers ( Li et al., 2023 ). Several limitations should be considered when interpreting our results. First, most of the studies in our overall analysis originate from high-income countries, particularly in the EPA. The high number of studies from high-income countries may skew the combined analysis since high-income countries will have more funding to support stronger regulation and waste-management systems than LMICs. Second, several phthalate metabolites had insufficient articles for regional analyses, especially in regions with a high proportion of LMICs, such as Africa and Latin America. With such low representation for these countries, the data on each phthalate metabolite concentration may be skewed toward values documented in countries where more studies were conducted. Third, while humans are exposed to many different phthalates through industry practices, our review solely focuses on those that are most prevalent and most widely studied. This selection of phthalates does limit our understanding of potential temporal and geographic variation seen in different phthalate metabolites. Fourth, we acknowledge that many articles did not provide data on potential pre-analytical or analytical contamination, nor did they specify whether enzymes other than the K12 enzyme were used during sample preparation. The K12 enzyme is essential when quantifying concentrations of phthalate metabolites, as it facilitates their conversion into their free form, making them detectable through various analytical techniques like gas chromatography or mass spectrometry ( Blount et al., 2000 ; Koch and Calafat, 2009 ; Koch et al., 2018 ). Without knowledge regarding potential contamination or the use of appropriate enzymes, we could not systematically compare studies for their reliability and quality assurance or exclude those that did not meet a standard level of quality or reliability. For example, the substantially higher phthalate biomarker concentrations reported by studies in the MESA region could reflect contamination or use of incorrect enzymes. Fifth, our review evaluated exposure information by assessing concentrations of urinary phthalate metabolites, not the parent phthalate compounds. Not all laboratories provided the same metabolites for each parent compound, limiting our ability to extrapolate the parent compound exposure. Sixth, we applied creatinine concentrations from one study to populations across the globe, in which body composition and drinking habits could differ in a way that influences estimates of phthalate exposure. Finally, substituting GMs for phthalate metabolite AMs may have biased our results. This review is one of the few to analyze temporal and geographical differences in concentrations of various phthalate metabolites on a large international scale, including Latin America, Africa, and Australia. In addition, it is the first to examine differences in exposure levels by age and pregnancy status. Together, these strengths offer the opportunity to infer the general population’s exposure levels to parent compounds of the included phthalate metabolites. Our review examined individual phthalate metabolites instead of their parent compounds. Doing so allows us to approximate which parent phthalate compound is increasing and diminishing in use throughout the time and region when concentrations of all metabolites of a parent compound are not reported. Our review highlighted the lack of significant research in regions that include many LMICs. To the degree that data were available, the review provides insight into phthalate exposure in these regions, allowing for meaningful comparisons in future research with more extensively studied regions like the US and Europe. Finally, using the multi-reviewer approach throughout each stage of screening and data extraction processes minimized selection bias. Research documenting the adverse health outcomes associated with phthalate exposure is prevalent across the environmental and medical fields. However, over the past two decades, there have been fewer than five articles focused on phthalate exposure from Africa, a continent with 54 countries, and Australia, with a population of over 26 million. While many regions, particularly LMICs where populations are known to be disproportionately exposed, are understandably focusing their resources on environmental contamination caused by pesticides and heavy metals, these same populations are also highly exposed to plastic. This plastic waste contains phthalates among other chemicals, all of which warrant greater monitoring and research within these regions. With increased regulation of certain phthalate parent compounds such as DEHP, stronger assessment of DEHTP and other substitutes for DEHP, including non-phthalates like DINCH, are warranted. Furthermore, an increase in longitudinal or repeated panel studies is also crucial to understanding exposure levels, since measuring the same population over time will offer the best opportunity to document the influence of industrialization and changing policies and regulations regarding use of endocrine-disrupting chemicals (EDCs) within specific regions. In addition to increasing regulation and research on phthalates, a global biomonitoring program with the needed reliability and comparability will offer tremendous opportunities for research growth. Such a biomonitoring program would provide quality assessment, quality control, detection levels, and standards, which will allow researchers to perform advanced biomonitoring on various compounds, including not only EDCs, but metals, pesticides, and other environmental toxins, while minimizing the limitations seen across similar studies. Finally, studies on the attributable disease burden and costs associated with phthalate exposure at regional and international levels are needed. As previously stated, phthalates are associated with a plethora of adverse health outcomes, such as cancer and cardiovascular morbidity. Understanding the financial impact of phthalate-attributed health outcomes is essential for highlighting the economic losses and evaluating the tradeoffs of ongoing plasticizer use.

Phthalates were introduced in the 1920s and used as a replacement for volatile and odorous camphor in manufacturing plastics. In the 1950s, the development of di(2-ethylhexyl) phthalate (DEHP) quickly led to its adoption in commercial use and the production of polyvinyl chloride (PVC) for construction ( Koch and Calafat, 2009 ; Oehlmann et al., 2009 ). Because phthalates are not chemically bound to the polymer, allowing them to leach or outgas into their surroundings, exposure can occur through inhalation, skin contact, and oral consumption. Pharmacokinetic analysis has suggested that phthalates have a relatively short half-life of 4 to 24 hours, varying based on chemical type and body composition. Phthalates are identified as endocrine-disrupting chemicals (EDCs) and have been linked to reproductive toxicity with prolonged exposure, causing declines in fertility, reduced testis weight, and increased oxidative stress on reproductive organs ( Fréry et al., 2020 ; Koch and Angerer, 2011 ; Sedha et al., 2021 ). Long-term exposure to phthalates has also been associated with Type II diabetes ( Sun et al., 2014 ), neurodevelopment impairment ( Jankowska et al., 2019 ), obesity, cancer, and cardiovascular mortality ( Koch and Calafat, 2009 ; Trasande et al., 2022 ; Wang and Qian, 2021 ). Despite increasing recognition of the adverse health effects associated with phthalate exposure, commercial use has remained robust. From 1970 to 2006, phthalate production more than doubled, increasing from 1.8 to 4.3 million tons globally, and expanded into markets such as food packaging ( Perestrelo et al., 2021 ), personal care products ( Aldegunde-Louzao et al., 2024 ; Gao et al., 2020 ), medical supplies ( Wang and Kannan, 2023 ), and pesticides ( Habert et al., 2009 ; Rodgers et al., 2014 ). While originally more prevalent in developed nations, phthalate-containing products are now ubiquitous in low- and middle-income countries (LMICs), as well. In China, plastic production has grown throughout the years, with primary plastics production increasing from 13.8 million metric tons in 2000 to 144.4 million metric tons in 2022. With the demands for goods growing with the population and urbanization, China’s general population will most likely continue to increase in phthalate exposure in the coming years ( Luan et al., 2021 ). Like China, various countries in Latin America have experienced heightened exposure through the increased consumption of ultra-processed food and beverages containing traces of phthalates ( Matos et al., 2021 ). Furthermore, these countries’ use of Brazilian biofuel increases their exposure to side products like diisopentyl phthalate (DiPeP), which have adverse endocrine effects ( Bertoncello Souza et al., 2018 ; Rocha et al., 2017 ). In response to the ever-growing scientific evidence of the harmful impact of phthalates on human health, the United States (US), Canada, and Europe have implemented biomonitoring programs to evaluate trends in exposure. This data can be used to document the impact of regulatory actions. For example, the US, Canda, and Europe have banned any children’s toys or childcare articles, such as baby bottles and pacifiers, that contain more than 0.1% (1000 ppm) of benzyl butyl phthalate (BBP), dibutyl phthalate (DBP), DEHP, diisononyl phthalate (DiNP), and several other phthalates that present a specific adverse health risk to children ( Eichler et al., 2019 ; Renwick et al., 2024 ; Union, 2005 ; United States Consumer Product Safety Commission, 2019 ). These types of regulatory actions and increased awareness of the hazardous effects of phthalates have contributed to a decline in exposure by 10-fold or more in Europe and the US throughout the years ( Domínguez-Romero et al., 2023 ; Gerofke et al., 2024 ; Kasper-Sonnenberg et al., 2025 ; Koch et al., 2017 ; U.S. Environmental Protection Agency, 2017 ). While national biomonitoring surveys are occurring in the US, Canada, and Europe, exposure information is not readily available in other regions. We therefore performed a systematic review of the literature to extract measures of phthalate metabolite concentrations in non-occupationally exposed populations internationally, excluding locations where representative data are already available at the country level, and analyzed these data for trends over time.