Maternal exposure to di-2-ethylhexylphthalate and adverse delivery outcomes: A systematic review.

Phthalates are industrial chemicals extensively used in a variety of consumer products, including plastic food containers, cosmetics/beauty products, toys, and certain medical products such as blood bags and pharmaceutical coatings [ 1 ]. Because of their wide-spread use and biological effects in animals, phthalates were included in the list of regulated (priority) pollutants by the US Environmental Protection Agency and the European Union [ 2 ]. Humans are exposed to phthalates through ingestion, inhalation, and dermal contact as well as via parenteral route when using medical devices [ 1 ]. Phthalates have short biological half-lives (6-12 hours), metabolize quickly, do not bioaccumulate, and are primarily excreted in urine [ 1 , 3 ]. Secondary phthalate metabolites are detected in 100% of the samples from general US population with wide variation [ 4 , 5 ]. Further, higher levels of phthalates in younger women as compared to men of the same age have been also reported possibly reflective of their potential exposure from cosmetic products [ 6 ]. Di-2-ethylhexyl phthalate (DEHP) is the most common phthalate that the general population is exposed to ubiquitously mainly through ingestion [ 7 , 8 ]. DEHP is rapidly hydrolyzed in the intestine to the corresponding monoesters (mono-(2-ethyl-hexyl) phthalate, MEHP) [ 7 , 9 , 10 ]. These monoesters are considered the biologically active metabolites and their use as biomarkers of DEHP exposure minimizes accidental contamination from parent compound [ 11 - 13 ]. In addition, urinary concentrations integrate exposures from multiple routes thus accounting for the total exposure [ 14 ]. Upon absorption, these monoesters undergo further hydroxylation and oxidation ( Figure 1 ) [ 7 ]. A greater proportion of the dose of DEHP is represented by the more downstream metabolites, including mono-(2-ethyl-5-hydroxyhexyl) phthalate (MEHHP), mono-(2-ethyl-5-oxohexyl) phthalate (MEOHP), and mono-(2-ethyl-5-carboxypentyl) phthalate (MECPP) [ 15 , 16 ]. In addition, it has been previously shown that a higher ratio of MEHP to MEHHP or MEHP to MEOHP is associated with potentially greater endocrine disrupting capacity [ 7 ]. Animal studies have found a variety of adverse effects from exposure to phthalates, including DEHP. Most severe of these effects were noted for reproductive system and normal development. In animal studies, reduction in testosterone levels following administration of DEHP in male animals resulted in underdevelopment of various androgen-dependent tissues and testicular abnormalities including reduced anogenital distance, agenesis of the gubernacular cords and sex accessory tissues, undescended testis, epididymal agenesis, testicular atrophy, and others [ 13 , 17 , 18 ]. Some of these effects resemble testicular dysgenesis syndrome in humans [ 14 ]. In female animals, reproductive effects from phthalates included altered serum estradiol levels, advanced or delayed onset of puberty, increased ovarian and uterine weights, and deficits in growing follicles and corpora lutea [ 14 ]. Other effects observed in one or more animal species included changes in hepatic structure and function, including liver cancer, changes in kidney function, and disruption of thyroid signaling, immune functions, and metabolic homeostasis [ 13 , 14 , 19 - 25 ]. The evidence on the association of phthalates with adverse effects in humans is limited. Previous studies suggested an association of exposure to phthalates with the risk of premature thelarche [ 26 , 27 ], higher risk of endometriosis [ 14 , 28 , 29 ], low sperm quality [ 11 , 30 , 31 ], reduced testosterone levels [ 32 - 34 ], obesity, diabetes, and possibly breast cancer [ 12 , 35 - 41 ]. Given the variety of these effects and ubiquitous exposure, phthalates were included on the list of endocrine-disrupting compounds with high exposure concern, evidence of endocrine disruption, and highest priority for research [ 42 ]. Moreover, phthalates have been recently classified by the International Agency for Research on Cancer as possible carcinogens to humans [ 43 ]. Several biological mechanisms were suggested for reproductive and developmental toxicity of DEHP. The primary metabolite of DEHP, MEHP, is a well-known ligand for the peroxisome proliferator-activated receptor (PPAR) family [ 14 , 44 ], is a mitochondrial toxicant and disruptor of lipid and glucose metabolism [ 14 , 44 - 46 ], and is the most potent DEHP metabolite in its toxicity [ 14 , 47 , 48 ]. Even though some studies suggested differences in susceptibility to the toxic effects of peroxisome proliferators across species with lower potential in humans, the basis for species differences in peroxisome proliferation and carcinogenesis by phthalate esters has not been fully described [ 13 , 48 ]. In addition, phthalates were also found to reduce the expression of insulin-like factor 3 (insl3) gene involved in the initial stages of testicular descent [ 14 ]. In females, DEHP-induced activation of PPAR resulted in dysregulation of aromatase activity and decreased estradiol production in rat granulosa cells [ 14 ]. Finally, a growing body of evidence suggests that, in addition to endocrine-disrupting effects on reproductive system, DEHP exhibits pro-inflammatory properties [ 49 - 53 ] and is associated with thyroid dysfunction [ 54 - 57 ]. Importantly, inflammation, oxidative stress, and hypothyroidism, all have been associated with adverse pregnancy outcomes in previous studies in humans [ 58 - 61 ]. The role of prenatal exposures with endocrine disrupting potential, including phthalates, on pregnancy outcomes is poorly understood. Adverse pregnancy outcomes, including preterm delivery, short gestational age, and abnormal birth weight, remain a public health concern [ 62 , 63 ]. These outcomes are associated with an increased risk of morbidity and mortality in the first year of life [ 64 ] as well as long-term health consequences in childhood and adulthood, such as neurodevelopmental disability, an increased risk of behavioral problems, hypertension, type 2 diabetes, cardiovascular disease, obesity, psychiatric disorders, and cancer [ 65 - 70 ]. The purpose of this systematic review was to summarize published studies on the association of exposure to DEHP, the most common phthalate, with preterm delivery, gestational age, and birthweight in humans and to identify methodological gaps that need to be addressed in future studies.

An electronic search was performed using PubMed Central (U.S. National Institutes of Health [NIH]), BioMed Central, and Toxnet with the cutoff date of July 31, 2015. Bibliographies of the articles identified in the electronic searches were then searched manually for additional relevant references. We used any combination of the key words/terms “DEHP” ‘MEHP’, ‘MEHHP’, ‘MEOHP, ‘MECPP’, ‘DEHP metabolites’ with “gestational age”, “preterm delivery”, “birth weight”, and “birthweight” to identify relevant publications. Study selection was accomplished by first applying the following inclusion criteria: (1) accessible in full-text manuscript and (2) published in English. We then excluded studies that did not measure DEHP metabolites in biological specimens to objectively characterize the exposure (referred to as exposure biomarkers or direct exposure assessment method). Our search yielded 117 manuscripts, from which 17 studies were relevant to the topic and met the eligibility criteria ( Figure 2 ). Two studies were further excluded due to the absence of objective exposure assessment (biomarkers of exposure) [ 71 , 72 ]. Two articles reported the results on the exactly same study population [ 73 , 74 ] and thus only one was included in the review [ 73 ]. From each selected article, we extracted the data on epidemiologic design features including study type, sample size, characteristics of the study population, exposure assessment approaches (type and timing of biological specimen, measured DEHP metabolites), outcome assessment method, statistical analysis methods, and results for each of the studies outcomes. We examined the evidence of the association between DEHP exposure and the birth outcomes across the studies while grouping them by the type of the outcome and phthalate analyte.

In Table 1 , we summarize the key characteristics of the 15 studies included in this review. Most of the studies were prospective cohorts (9 studies or 56%), three studies utilized a nested case-control design, one study was case-control, and two studies were cross-sectional. Most of the studies simultaneously evaluated more than one adverse outcome, totaling 4 studies for preterm delivery, 9 studies for gestational age, and 10 studies for birthweight. The mean sample size across the studies was 307 with a median of 283 women. Five of the studies included racially/ethnically diverse study populations and the remaining 10 studies were limited to a single race/ethnicity (5 Asian, 3 Caucasian, and 2 Hispanic). Most of the studies (13 out of 15) used medical records to verify the pregnancy outcomes. All of the studies, except one, measured one or more DEHP metabolite in a biological sample; one study assessed parent DEHP concentrations only [ 75 ]. Selected studies additionally assessed associations of the birth outcomes with sum of DEHP metabolites (MEHP, MEHHP, MEOHP, and MECCP) or % MEHP (ratio of MEHP to the sum of MEHHP, MEOHP, and MECCP). Most of the studies (11 out of 15) measured phthalates in urine (spot urine or repeated samples). Other types of biological specimens included umbilical cord blood, maternal blood, meconium, and placental tissue. The details of the study designs and methods are summarized in Table 2 . Preterm delivery was defined as <37 weeks of gestation across all the studies. The results of these studies by the type of DEHP biomarker are presented in Figure 3 . Across four studies that evaluated the association of DEHP with preterm delivery, one study [ 76 ] reported significant inverse associations with MEHP, MEHHP, and MEOHP, and three studies reported significant positive associations with MEHP, MECPP, and parent DEHP concentrations [ 34 , 73 , 75 ]. Figure 4 summarizes the results of the studies on the association of gestational age with DEHP. Four of the nine studies found inverse associations of all individual metabolites, ΣDEHP, and parent DEHP with gestational age and two studies reported positive associations with MEHP, MEHHP, and MEOHP. The results of the studies on the association of birthweight with phthalates are presented in Figure 5a (findings reported as regression coefficients) and 5b (findings reported as odds ratios), with the exception of three studies that did not report risk estimates [ 77 - 79 ]. Among ten studies, two reported significant inverse associations of MEHHP or DEHP with birthweight [ 75 , 80 ] and two studies found positive associations with MEHP, MEHHP, and MEOHP [ 81 , 82 ].

The evidence on the association of phthalates with adverse birth outcomes remains inconsistent. We systematically reviewed 15 published studies that investigated the association of the most common phthalate, DEHP, with preterm delivery, gestational age, and birthweight. We identified important methodological concerns (discussed below) related to different aspects of study design which could have resulted in selection bias and exposure misclassification and contributed to null findings and biased associations in the identified studies. We further discussed their implications and suggested some strategies for the future studies. The median sample size across the studies was relatively small (283 women) and given that the relative increase in the risk associated with environmental exposures is usually low, typically below 1.5 [ 83 ], these studies were likely underpowered to detect significant associations. In addition, some of the studies were focused on high-risk populations thus limiting generalizability of their findings. For example, Philippat et al. included pregnant women who gave birth to infants with malformations and their matched controls [ 84 ]. Another study by Huang et al. included women who had indication for amniocentesis due to advanced maternal age and abnormal screening [ 78 ]. This study subsequently reported a significant inverse association of DEHP with birthweight and gestational age and a higher risk of preterm delivery in women with higher DEHP levels, which should be interpreted with caution. The majority of the identified studies were limited to a single race/ethnicity. Previous studies show disparities in distribution of selected phthalates among reproductive aged and pregnant women [ 85 , 86 ] with higher levels noted in non-Hispanic Black and Hispanic women, though the ethnic disparities in DEHP or other high molecular weight phthalates among women are inconsistent with most studies reporting no differences and a few reporting higher levels in Caucasians [ 87 ]. Further, a recent study showed that non-Hispanic Blacks and Hispanic individuals had higher levels of %MEHP (the ratio of MEHP to the sum of the secondary metabolites), reflective of a slower MEHP conversion rate and possibly, higher potential for adverse effects from DEHP exposure [ 88 ]. Previous studies also demonstrated pronounced racial disparities in occurrence of adverse pregnancy outcomes including preterm birth and low birthweight [ 89 , 90 ]. Finally, racial/ethnic differences in the associations of phthalates with adverse health outcomes such as diabetes have been also reported [ 91 ]. The above listed disparities make it important to examine these associations in racially diverse populations and to identify population subgroups that might be more susceptible to adverse effects of phthalates on pregnancy outcomes. The studies varied with respect to the different aspects of exposure assessment, including the type of biological specimen, timing of sample collection, and the number of samples per woman. Use of urine samples is the standard approach for biomonitoring of phthalates, including DEHP. Urine samples offer several advantages as compared to other types of biospecimens, including ease of sample collection, higher metabolite concentrations, and reduced potential for contamination by the parent compound and its subsequent conversion into metabolites as the result of post-collection enzymatic activity in the blood samples leading to hydrolysis of extraneous phthalate diesters to their monoesters and, subsequently, exposure misclassification [ 1 , 7 , 8 , 11 , 16 , 92 ]. In the studies measuring concentrations in umbilical cord blood, contamination by DEHP in the sampling and analytic equipment cannot be excluded [ 93 ]. Furthermore, previous studies suggest that cord blood samples may not be reflective of maternal exposure levels during pregnancy (correlation between maternal and cord blood levels ranging between 0.1 and 0.5) [ 16 ]. Whenever possible, the use of urinary samples for exposure assessment should be preferred. Even though DEHP is rapidly metabolized within hours (half-life 6-12 hours) [ 1 , 3 ], previous studies suggest that a single urine sample can accurately reflect phthalate exposure over the previous 3 months [ 14 , 94 , 95 ]. However, previous reports suggest that metabolite concentrations have equivocal reproducibility in pregnant women, especially during the last 6 weeks of pregnancy (intraclass correlation for DEHP metabolites ranging from 0.30 to 0.36; for ΣDEHP =0.08) [ 96 , 97 ]. Thus, the use of a single sample in the studies assessing the exposure during the 3 rd trimester (majority of the studies) could result in significant exposure misclassification. Collection of repeated samples (one per pregnancy trimester) to account for intra-individual variation in DEHP biomarker levels is strongly recommended. Further, standardization for the time of day may be needed to account for within-individual variability [ 11 , 16 ]. Finally, it was previously reported that urinary creatinine excretion may be affected by individual characteristics, including age, muscle mass, and race, as well as lifestyle factors, such as diet and physical activity [ 16 ]. Creatinine excretion may vary during the course of pregnancy and specific gravity may be more effective in correcting for urine dilution later in pregnancy as it is not influenced by individual factors, thus reducing the chances for exposure misclassification [ 16 , 97 , 98 ]. It remains unclear which window of susceptibility during pregnancy would be more relevant for the potential effect of phthalates on adverse birth outcomes. A recent report indicates that both early and late exposures during pregnancy could have implications for preterm birth [ 74 ]. Previous studies on the associations of other environmental exposures with birth outcomes suggest that exposures during the 2 nd and 3 rd trimester could have greater impact on birth outcomes, as the result of more rapid fetal weight gain in the 3 rd trimester as well as disrupting effects of DEHP on parturition [ 76 , 99 , 100 ]. In some of the included studies, the samples were collected in the 1 st trimester while others have varied sample collection timing [ 77 , 78 , 80 , 84 ]. In addition, some of the effects of phthalates on selected outcomes, such as birthweight, could potentially result from “fetal programming” earlier in pregnancy leading to long-term effects on structure, physiology and metabolism [ 101 - 103 ]. For example, undernutrition in the 1 st trimester of pregnancy has been linked to increased birth weight [ 102 ] suggesting possible importance of this crucial period for metabolic programming. Thus, exposure assessment during different windows of susceptibility is warranted for better understanding of the effects of DEHP exposure on birth outcomes. It has been previously shown that a higher ratio of MEHP to MEHHP or MEHP to MEOHP is associated with a greater physiologic effect and potentially greater endocrine disrupting capacity as compared to individual metabolites [ 7 ]. It was also suggested that MEHP, but not other metabolites, can disturb energy metabolism of fat cells [ 104 ], a mechanism which could potentially lead to changes in birthweight. Only one study attempted to account for individual differences in phthalate metabolism by examining the ratio of MEHP concentration to the sum of MEHHP, MEOHP, and MECPP (% MEHP). Using the ratios of metabolites rather than individual metabolite concentrations in future studies could help to account better for different metabolic patterns. Finally, even though the evidence from animal studies suggests transplacental transfer of MEHP [ 105 ], findings from the studies in humans have been conflicting [ 14 , 106 ]. Thus, the results of the studies on maternal MEHP concentrations and birthweight should be interpreted with caution. Previous studies suggest that several maternal factors and environmental exposures can affect birthweight and duration of pregnancy. Maternal weight gain during the 2 nd and 3 rd trimester, history of diabetes and gestational diabetes, and maternal smoking and alcohol use increase the risk of adverse birth outcomes [ 107 - 113 ]. However, only a few of the studies in this review accounted for possible confounding effect of these risk factors and some of the findings could be explained in part by the residual confounding effects. Adjustment for these risk factors in future studies is warranted.

We found no consistent evidence of the association of phthalates with preterm delivery, gestational age, and birthweight across the studies included in this review, which might be explained by the heterogeneity of the studies. Given the aforementioned methodological gaps that likely contributed to the findings, addressing these concerns in more thorough investigation of these associations is warranted to advance our scientific knowledge on the potential effects of DEHP exposure on birth outcomes.