The Association between Maternal Urinary Phthalate Concentrations and Blood Pressure in Pregnancy: A Systematic Review and Meta-Analysis

Phthalates are commonly found in a wide range of environments and have been linked to several negative health outcomes. While earlier research indicated a potential connection between phthalate exposure and blood pressure (BP) during pregnancy, the results of these studies remain inconclusive. The objective of this meta-analysis was to elucidate the relationship between phthalate exposure and BP in pregnancy. A comprehensive literature search was carried out with PubMed, EMBASE, and Web of Science, and pertinent studies published up until 5 March 2023 were reviewed. Random-effects models were utilized to consolidate the findings of continuous outcomes, such as diastolic and systolic BP, as well as the binary outcomes of hypertensive disorders of pregnancy (HDP). The present study included a total of 10 studies. First-trimester MBP exposure exhibited a positive association with mean systolic and diastolic BP during both the second and third trimesters (β = 1.05, 95% CI: 0.27, 1.83, I2 = 93%; β = 0.40, 95% CI: 0.05, 0.74, I2 = 71%, respectively). Second-trimester monobenzyl phthalate (MBzP) exposure was positively associated with systolic and diastolic BP in the third trimester (β = 0.57, 95% CI: 0.01, 1.13, I2 = 0; β = 0.70, 95% CI: 0.27, 1.13, I2 = 0, respectively). Conversely, first-trimester mono-2-ethylhexyl phthalate (MEHP) exposure demonstrated a negative association with mean systolic and diastolic BP during the second and third trimesters (β = −0.32, 95% CI: −0.60, −0.05, I2 = 0; β = −0.32, 95% CI: −0.60, −0.05, I2 = 0, respectively). Additionally, monoethyl phthalate (MEP) exposure was found to be associated with an increased risk of HDP (OR = 1.12, 95% CI: 1.02, 1.23, I2 = 26%). Our study found that several phthalate metabolites were associated with increased systolic and diastolic BP, as well as the risk of HDP across pregnancies. Nevertheless, given the limited number of studies analyzed, additional research is essential to corroborate these findings and elucidate the molecular mechanisms linking phthalates to BP changes during pregnancy.


Introduction
Hypertensive disorders of pregnancy (HDP) rank among the top contributors to maternal and neonatal mortality and morbidity [1,2]. Hypertension can either be chronic or develop during pregnancy and may be accompanied by other comorbidities, potentially resulting in pre-eclampsia (PE) and preterm birth (PTB) [3]. Factors known to heighten

Materials and Methods
The present study adhered to PRISMA guidelines (http://www.prisma-statement.org/ (accessed on 5 March 2023)) (Supplementary Material, Table S1), and was not registered with PROSPERO in advance.

Search Strategy
We conducted a search for studies published up to 5 March 2023 using three databases: Embase, PubMed, and Web of Science. Relevant keywords were determined based on the PECO framework [18]. To generate search terms, we combined keywords that represented exposure and outcome components within the PECO framework. The following search terms were employed: (gestational phthalate exposure OR phthalates OR urinary phthalate metabolites OR maternal urinary phthalate concentrations) AND (blood pressure OR cardiometabolic indices OR hypertensive diseases of pregnancy OR pregnancy-induced hypertension OR maternal hemodynamics OR gestational hypertensive disorders OR gestational hypertension OR preeclampsia). These terms were used to identify English-language original articles reporting the effects of phthalate exposure on BP during pregnancy or risk of HDP. A comprehensive list of search terms utilized across various databases can be found in Supplementary Materials, Table S2.
No restrictions were placed on study designs during the search process. Relevant studies were also identified by screening reviews and reference lists of search results. After removing duplicates, titles and abstracts underwent an initial screening, followed by a full-text review. Two authors (MYZ and JCQ) independently assessed all articles, with their findings reaching a consensus. In cases of disagreement, a third reviewer (CYH) examined the article and made the final decision.

Inclusion and Exclusion Criterion
In alignment with the PECO elements, we applied the following inclusion criteria: (1) the population (P) consisted of pregnant women; (2) exposure (E) was determined by phthalate exposure measured from biological samples; (3) the comparator (C) involved a group with lower exposure levels; (4) the outcomes (O) included BP during pregnancy and risk of HDP, which were verified by physicians, medical records, or according to the International Classification of Diseases codes; and (5) the study was accessible as a full text and in English. The exclusion criteria included: (1) studies focusing on the general population; (2) studies concentrating on occupational exposure; and (3) studies without a reference group.

Data Extraction and Quality Assessment
Two authors (MYZ and JCQ) independently carried out the data extraction and quality assessment. From each eligible study, we collected the source (author name and publication year), location, study period, study design, phthalates examined, exposure assessment, exposure period, outcomes, covariates adjusted for, and main findings. Additionally, we contacted several corresponding study authors to request supplementary results not included in their published articles but relevant to our research. Additionally, if necessary, we contacted several corresponding study authors to request supplementary results not included in their published articles but relevant to our research.
In our systematic review and meta-analysis, we assessed the quality of the incorporated studies by employing the Newcastle-Ottawa Scale (NOS) specifically designed for observational research [19]. The NOS contains eight elements, organized into three distinct categories: (I) selection of study groups; (II) group comparability; and (III) interest-specific exposure or outcomes. Individual categories receive a score, with the highest possible being four, two, and three stars, adding up to a maximum of nine stars. Utilizing this scale, we categorized studies as high-quality (≥7 stars), moderate (4-6 stars), or low-quality (≤3 stars) [20]. Any differences in determining the quality of the studies were settled through achieving a consensus.

Statistical Analysis
We carried out a meta-analysis employing a random-effects model to accommodate variations both within and across studies. By utilizing the reported odds ratios (ORs) and their associated 95% confidence intervals (95% CIs), we calculated the combined ORs for the binary outcome of HDP. Additionally, coefficients from linear regressions were combined in a meta-analysis to estimate the effects of phthalate exposure on continuous outcomes, including diastolic and systolic BP. We performed separate meta-analyses for each trimester-specific exposure (phthalate metabolite exposure during the first, second, and second trimesters, or the full pregnancy mean) and outcomes (diastolic BP and systolic BP) when at least three studies were available. Given the limited number of studies addressing various HDP types, we included all research on HDP (e.g., GH and PE as distinct outcomes or a mix of these conditions), regardless of the employed definition, in order to provide an adequate basis for random-effects meta-analyses.
For the purpose of facilitating effect estimate comparisons across studies, we transformed the documented ORs and 95% CIs to align with a standardized exposure increment (per natural log-unit increase). We assessed statistical heterogeneity among the selected studies using the Q-test and the I 2 statistic (range: 0-100%). I 2 values were classified as low, medium, or high based on values of 25, 50, or 75%, respectively [21]. We employed funnel plot asymmetry and Egger's test to evaluate publication bias. If fewer than five studies were included, publication bias was not assessed. We adopted a p-value of <0.05 for Egger's test. If the p-value was less than 0.05, publication bias was present, and vice versa [22]. For analyses with >3 studies, sensitivity analyses were conducted by systematically excluding one study at a time and recalculating summary effect sizes. All analyses were performed using Stata software version 16.1 (StataCorp, Texas, USA). All p-values were two-sided with a significance level set at 0.05.

Results
The PRISMA flowchart, illustrating our literature search process, can be found in Figure 1. Our study encompassed a total of 379 studies, incorporating the outcomes of keyword searches. Following an in-depth examination of the full texts, 43 articles were deemed potentially suitable for inclusion in the meta-analysis. Finally, the present study included a total of 10 studies for the meta-analysis. Of the 10 included studies, 6 were performed in North America [9,13,14,[23][24][25], 2 were from Europe [15,26], and the remaining 2 were from Asia [16,17]. The studies incorporated in our analysis were published between 2015 and 2021, with participant numbers ranging from 152 to 3273. Among these, nine were cohort studies, while one was a cross-sectional study. Each study assessed phthalate exposure using maternal urinary samples as the measurement method.

Study Characteristics
The majority (N = 7) of the included studies documented preeclampsia [
The Newcastle-Ottawa Scale (NOS), ranging from 0 to 9 stars, was utilized to evaluate the quality of the studies, with a higher number of stars indicating superior quality. As shown in Table 2, all studies were assessed as high-quality (>7) using the NOS tool.

Main Meta-Analysis Findings
The meta-analysis synthesized outcomes regarding the associations between multiple different phthalates, BP during pregnancy, and HDP risk, which are presented in Table 3 below.

Phthalates Exposure during Pregnancy and Risk of HDP
The collective findings indicate that trimester-specific phthalate exposure during pregnancy was not associated with HDP, with the exception of monoethyl phthalate (MEP) exposure. The combined effect estimates ranged from 1.01 to 1.18, with a low 95% confidence interval (CI) spanning from 0.86 to 1.02 and a high 95% CI extending from 1.09 to 1.41. However, almost all these combined effect estimates were not statistically significant. The only exception was MEP exposure, which was found to be associated with an increased risk of HDP (OR = 1.12, 95% CI = 1.02 to 1.23, I 2 = 26%).  Notes: The NOS has eight items grouped in three domains: (I) selection of study groups; (II) comparability of the groups; and (III) exposure or outcomes of interest. Each domain is scored with a maximum of four, two, and three stars, respectively; thus, the total score for the scale sums up to nine stars. Based on this scale, the studies were classified as high quality (≥7 stars), moderate (4-6 stars), and low quality (≤3 stars).

Sensitivity Analysis and Publication Bias
The sensitivity analysis results demonstrated that the combined effect estimates remained generally stable when individual studies were excluded. Due to the limited number of studies included for each exposure and outcome combination, funnel plots were not conducted. Nonetheless, for all exposures, the p-values from Egger's tests were not statistically significant (p > 0.05), suggesting no evidence of publication bias. However, given the relatively small number of studies in our meta-analysis, these test results should be interpreted cautiously, as they may lack the power to detect publication bias in meta-analyses with a limited number of studies [27].

Discussion
In this meta-analysis, we synthesized data from 10 human epidemiological studies to investigate the connection between phthalate exposure, BP, and HDP risk. Our findings suggest that, aside from MBP, MBzP, and MEHP, the majority of phthalate metabolites show no association with BP. Notably, only MEP exposure was associated with a heightened risk of HDP.
The impact of phthalate exposure on BP during pregnancy and HDP risk has attracted considerable interest in recent years. Soomro et al. found that prenatal exposure to phthalates, particularly MBP and MEP, could play a vital role in PIH [26]. Werner et al. proposed that early mid-pregnancy exposure to MBzP correlated with increased diastolic BP and HDP [13].  reported an association between phthalate metabolite exposure and decreased systolic and diastolic BP during pregnancy, particularly in the second trimester [15]. This observation is consistent with results showing an association between MEHP exposure and reduced systolic and diastolic BP. Although evidence on gestational BP changes related to phthalate exposure is somewhat conflicting, there is consistent support from prior studies involving children and nonpregnant adults that DEHP [28,29], MBP (a metabolite of DBP) [23,28], and MBzP [30,31] correlate with BP changes.
Numerous factors could account for the discrepancies observed in the results of prior studies. Inconsistencies between studies investigating the same phthalate might arise from inadequate statistical power, variation in outcome assessment, divergent analysis methods, or differing study populations. While the majority of studies assessed phthalate levels during one trimester, the timing of exposure evaluation was not consistent across studies. It is well-established that phthalate plasma/serum concentrations decrease as pregnancy progresses due to the expansion of plasma volume and the distribution of phthalates into the fetal compartment [13].
In terms of outcomes, three studies exclusively examined BP during pregnancy, one study focused on PE, and others investigated GH or other HDPs. Only a single study explored the association with repeated BP measurements throughout pregnancy [17]. Recently, women participating in the Programming Research in Obesity, Growth, Environment, and Social Stressors (PROGRESS) study were followed for up to 72 months postpartum to assess the associations between gestational phthalate exposure and postpartum outcomes. Wu et al. (2021) found that combined phthalate mixture levels were positively associated with elevated postpartum BP up to 72 months postpartum. This finding highlights the significance of examining women's health beyond pregnancy when evaluating the potential effects of phthalate exposure [25].
The precise mechanisms underlying phthalate-induced elevated BP remain uncertain; however, current research provides some understanding. Phthalate metabolites function as agonists for PPARγ (peroxisome proliferator-activated receptor gamma), known to inhibit the renin-angiotensin-aldosterone system, an essential BP regulator [32,33]. Phthalate exposure has been associated with heightened oxidative stress in pregnant women [34], potentially affecting the release of circulating angiogenic factors. These factors not only serve as predictors for PE [35,36] but are also linked to PIH [37]. Furthermore, phthalates may impact BP by modifying thyroid hormone levels during pregnancy. A previous study found a connection between urinary phthalates and decreased serum thyroxine in pregnant women [38], which has been demonstrated to raise the risk of PE [39]. Additionally, phthalate-triggered inflammatory responses might play a significant role in GH. An earlier study among pregnant women uncovered a relationship between specific urinary phthalate metabolites and increased levels of inflammatory cytokines [40]. This heightened production of inflammatory cytokines could have a crucial role in GH or PE [41].

Strengths and Limitations
Meta-analyses can offer more accurate conclusions than individual studies by increasing statistical power, which is particularly crucial for rare outcomes. In our meta-analysis, we identified several significant relationships between phthalates and BP during pregnancy and risk of HDP, even though many individual studies failed to do so. The robustness of our findings was evident as the exclusion of individual studies typically had minimal impact on the results. Additionally, our meta-analysis showed no significant signs of small-study bias.
However, our study has several limitations that should be considered when interpreting the results. First, while we included all available studies in three databases examining the effect of phthalates on BP during pregnancy and the risk of HDP up to 2023, insufficient data on this research topic prevented us from conducting subgroup analyses based on study design, location, and offspring gender. Second, the studies were conducted in various geographic regions and populations, leading to differences in phthalate exposure. This factor may further affect the measured concentrations of these phthalates at different times. Furthermore, the range of exposure varied among the included studies, which could influence the meta-analysis results. Third, we exclusively included studies that assessed phthalate levels in urine samples to maintain consistency and comparability in our analysis. Although phthalate metabolites in urine are considered suitable biomarkers for short-term exposure to parent compounds due to their rapid excretion and short half-life in humans [42], urinary measurements may not accurately reflect long-term exposure. Lastly, some studies combined GH and PE, preventing us from including these studies in separate meta-analyses for HDP as a distinct outcome.

Recommendations for Future Study
While some epidemiological evidence and potential mechanisms suggest that phthalates may be risk factors for elevated BP during pregnancy and HDP, additional research is necessary to verify this relationship and guide medical recommendations and environmental policies. Such research faces challenges, including logistical issues related to sample collection, timing, and the selection of specific phthalate metabolites for investigation. Acquiring well-timed samples from pregnant women and confirming their HDP diagnosis is both difficult and costly. Prior studies have indicated variations or interactions influenced by factors such as parity and fetal sex [13,43]. To develop a comprehensive understanding of HDP risk, these factors, among others, should be considered. Future research should, at a minimum, assess the modifying effects of fetal sex, parity, and race/ethnicity, and include the timing of phthalate measurements during pregnancy as a covariate in analyses. Accounting for essential confounders, such as maternal age, smoking status, and pre-pregnancy BMI, is crucial, as is exploring the modifying influence of diet and physical activity [44,45]. Investigating the relationship between phthalates and HDP in diverse racial and ethnic populations is essential to evaluate risk and develop targeted prevention strategies.
To fully understand the effects of phthalates on health outcomes, future research must take into account their cumulative burden. Methods for exposure assessment and strategies for categorizing phthalates during risk assessment are being developed [46,47]. As epidemiological studies increasingly measure multiple phthalate metabolites, developing methods for modeling exposure mixtures will become increasingly important. Available statistical methods include, but are not limited to, BKMR [48], environmental risk score [49], supervised principal component analysis followed by classification and regression tree [50], toxicant score [51], joint WQS regression [52,53], and machine learning methods such as lasso or adaptive elastic net [50]. When selecting the appropriate method, the correlation structure between phthalates is crucial. For instance, joint WQS displays strong sensitivity and specificity for identifying predictors within a correlated mixture, while other methods might be more suitable for non-correlated or weakly correlated exposures [54,55].

Conclusions
Our meta-analysis determined that there was a strong association between several phthalate metabolites (MBP, MBzP, and MEHP) and BP during pregnancy, while MEP was found to be associated with an increased risk of HDP. However, these results should be interpreted cautiously due to the relatively small number of included studies. Given the considerable maternal and fetal morbidity and mortality associated with HDP, this study has significant public health implications. Nonetheless, additional research is required to clarify the underlying mechanisms.

Conflicts of Interest:
The authors declare no conflict of interest.