Perinatal Exposure to Environmental Endocrine Disruptors in the Emergence of Neurodevelopmental Psychiatric Diseases: A Systematic Review

Background: Exposure to endocrine disruptors is on the rise, with new compounds regularly incriminated. In animals and humans, this exposure during critical developmental windows has been associated with various developmental abnormalities, including the emergence of psychiatric disorders. We aimed to review the association between perinatal endocrine disruptor exposure and neurodevelopmental disorders in humans, focusing on cognitive and psychiatric disorders. Methods: We performed a systematic review with key words referring to the fields of neurodevelopment and endocrine disruptors. We reviewed 896 titles, choosing studies on the basis of titles and abstracts. We searched through the methodology sections to find perinatal exposure and neurodevelopmental disorders, following the categories indicated in the Diagnostic and Statistic Manual of Mental Disorders (5th edition). References in some studies brought us to a total of 47 studies included here. Results: Convergent studies report an association between exposure to endocrine disruptors and autism spectrum disorder, attention-deficit hyperactivity disorder, global developmental delay, intellectual disability, communication disorders and unspecified neurodevelopmental disorders. Conclusion: Sufficient data exist to report that exposure to some endocrine disruptors is a risk factor for the emergence of neurodevelopmental disorders. Studying endocrine disruptor exposure in humans is still associated with some limits that are difficult to overcome.


Background
The notion of endocrine disruptor compounds (EDCs) was first used at the Wingspread Conference Center in 1994 [1]. It describes any exogenous chemical compound able to interfere with the endocrine system, resulting in adverse effects on the health of an individual and/or his/her offspring. They may not only be medicines, but also pesticides, compounds used in the plastic industry, mass consumption goods, industrial products or pollutants. Since the early nineties, some studies have reported permanent effects of EDCs on development. In 2012, the Global Assessment of the State-of-Science of Endocrine Disruptors issued by the World Health Organization and the United Nation Environment Program concluded that "different EDCs can act together and result in an increased risk of adverse effects on human health and wildlife."

Materials and Methods
Our eligibility criteria did not screen participants on age. Children and adults were both included since the neurodevelopmental window can be very large, with symptoms appearing years later. Studies followed a prospective or retrospective observational design (due to obvious ethical questions). They may, or may not, have had a control group. Indeed, some EDCs being ubiquitous in our environment, comparisons could only be assessed according to high vs. low exposure. The psychiatric assessment needed to follow a standardized method. We reviewed studies that had been published until 1 December 2018, exposing us to publication and selective reporting within studies biases. We screened the PubMed and Embase databases. Regarding exposure, we searched the terms: "endocrine disruptors," "bisphenol," "phthalates," "pollutants," "polybrominated diphenyl ethers" (PBDE), "hexaclorobenzene" (HCB), "polychlorinated biphenyls" (PCB), "metal," "solvents," "polycyclic aromatized hydrocarbons" (PHA), "chlorpyfiros" (CPH), and "organohalogens compounds" (OHC). Regarding the outcomes we used the terms: "neurodevelopment," "attention deficit disorder hyperactivity" (ADHD), "autism spectrum disorder" (ASD), "intellectual disabilities" (ID), "learning disorders" (LD), "developmental delay," "language" and "behavior". We first reviewed titles and abstracts. We then searched for the methodology to find our eligibility criteria.

Results
We reviewed 864 titles and found 32 other titles through references in screened articles. We chose studies based on title and abstract. A total of 770 studies that did not concern neurodevelopment were excluded. We then carefully searched for the methodology to find human exposure, perinatal exposure and neurodevelopmental disorders assessed with a standardized method. A total of 79 studies that did not meet these criteria were excluded, which brought us to a total of 47 studies included here (PRISMA chart in Figure 1). 3 "polycyclic aromatized hydrocarbons" (PHA), "chlorpyfiros" (CPH), and "organohalogens compounds" (OHC). Regarding the outcomes we used the terms: "neurodevelopment," "attention deficit disorder hyperactivity" (ADHD), "autism spectrum disorder" (ASD), "intellectual disabilities" (ID), "learning disorders" (LD), "developmental delay," "language" and "behavior". We first reviewed titles and abstracts. We then searched for the methodology to find our eligibility criteria.

Results
We reviewed 864 titles and found 32 other titles through references in screened articles. We chose studies based on title and abstract. A total of 770 studies that did not concern neurodevelopment were excluded. We then carefully searched for the methodology to find human exposure, perinatal exposure and neurodevelopmental disorders assessed with a standardized method. A total of 79 studies that did not meet these criteria were excluded, which brought us to a total of 47 studies included here (PRISMA chart in Figure 1).

EDCs Exposure and ASD
ASD is characterized by persistent deficits in social communication and social interaction across multiple contexts, including deficits in social reciprocity, nonverbal communicative behaviors used

EDCs Exposure and ASD
ASD is characterized by persistent deficits in social communication and social interaction across multiple contexts, including deficits in social reciprocity, nonverbal communicative behaviors used for social interaction, and skills in developing, maintaining, and understanding relationships. In addition to social communication deficits, the diagnosis of ASD requires the restricted or repetitive patterns of behavior, interests, or activities. There has been a large increase in the diagnosis of ASD in recent decades [23]. One of the main explanations could be the use of better tools designed to diagnose the disease. As of today, the diagnosis itself is also better understood and recognized. The model of a multifactorial disease with complex genetic factors interacting with environmental factors has been advanced. Based on that idea, we can hypothesize that changes in the environment could, at least partially, explain the recent increase in the number of cases.
The results of ten studies [24][25][26][27][28][29][30][31][32][33], published between 2006 and 2017, are shown in Table 1. ASD was assessed as a clinical category in six studies while four others investigated autistic traits through social competence evaluation. We noticed that the first six studies have very different designs, which make them difficult to compare. Windham and Nishijo both studied prenatal exposure with follow-up studies [24,25]. Nishijo's cohort included younger children and the diagnosis was made using a scale, not through a clinical examination, unlike other studies. To these six studies, we added four more studies that found an association between EDCs exposure and autistic traits, measured using scales about social competence [26][27][28][29].
Five authors found significant associations between an ASD diagnosis and exposure to phthalates, air pollutants, 2,3,7,8-Tetrachlorodibenzodioxin (TCDD) and BPA [24][25][26][27][28]. Only Rahbar et al. did not find any association between an ASD diagnosis and postnatal exposure to different EDCs (dioxins, dibenzofurans, BPA and phthalates), though this cohort was small and a few patients were under the range of detection threshold for each EDC [33]. The high TCDD groups showed higher ASRS scores than the mild-TCDD groups (p = 0.042). No differences in neurodevelopmental scores. Negative associations between PCDD levels and SRS scores in the whole group (p < 0.05) in girls and in one subscale (social motivation) in boys. For PCB, associations with one subscale for the whole group (autistic mannerisms).

EDCs Exposure and ADHD
ADHD is a neurodevelopmental disorder defined by impaired levels of attention, disorganization, and/or hyperactivity-impulsivity. As for ASD, diagnoses of ADHD have been increasing lately. Population surveys suggest that ADHD occurs in most cultures in about 5% of children and about 2.5% of adults. Changes in our environment, in association with better tools for diagnosis, could explain this observation [34].
Three studies using the same cohort from Minorca (Spain) found an association between the diagnosis and PBDE, HCB and nitrogen dioxide [26,27,37]. Another study found an association with PBDE exposure (mostly postnatal) [38].
Regarding BPA, prenatal exposure was not found to be associated with ADHD [37], although postnatally, the results seem to differ according to two studies [39,40].
Once again, one main difference among all studies is the way ADHD is diagnosed. In some, the outcome is an ADHD diagnosis, as defined in the DSM, while in others, the outcome is defined as "ADHD-traits." It has been suggested that ADHD psychopathology can be viewed dimensionally, with symptoms distributed continuously in the general population [46,47]. Thus, healthy individuals may present high and low levels of ADHD-traits, altering their functioning, but not display enough symptoms to suffer from an ADHD, in the clinical sense.

EDCs Exposure and Global Developmental Delay (GDD)
Global developmental delay is diagnosed when an individual fails to meet expected milestones in several areas of intellectual functioning. The diagnosis is made with systematic assessment of intellectual functioning such as the Bayley Scale of Infant Development (BSID) or the Gesell Developmental Schedules (GDS). These tests are chosen according to the individual's age. The results of seven studies, published between 2006 and 2012, are shown in Table 3 [48][49][50][51][52][53][54]. Most of these studies present a positive association between prenatal exposure to EDCs and a decrease in mental and/or psychomotor index. Two studies found an association with phthalates [48,51], and two studies also found an association with polycyclic aromatic hydrocarbon (PAH) [50,53]. Noteworthy, Perera et al. presented recent results in favor of a reversible decrease of developmental quotient after a source of PAH was removed [54]. Maternal serum PBDE10 = 28.7 ng/g Child serum (7 years

EDCs Exposure and Intellectual Disability
Intellectual disability is characterized by deficits in general mental abilities. The deficit results in the impairment of adaptive functioning in the conceptual, social and practical domains. It is typically measured using psychometry valid tests. Those tests need to be adapted to the child's age. The results of five studies, published between 2009 and 2017, are shown in Table 4 [55][56][57][58][59].
Two studies from two different cohorts found convergent results in favor of an association between exposure to PAH and intellectual quotient (IQ) alterations. We notice that they used two different methods to assess the IQ, and in one study [57], the exposure was 10 times higher than in the other [58]. Two studies searching for an association with PCB exposure reported inconsistent results, but again, the exposure itself and PCB concentrations differed [55,56].

Exposure to EDCs and Communication Disorder
Disorders of communication include deficits in language, speech and communication. They are assessed with standardized measures that must take into account the individual's cultural and linguistic context. Till et al. found a significant association between high prenatal exposure to organic solvents and difficulties in receptive (p = 0.25) and expressive (p = 0.44) language as well as graphomotor abilities (p = 0.04) when compared to low exposure [60]. The results are summarized in Table 5. Association of high exposure with difficulties in receptive (p = 0.025) and expressive (p = 0.044) language as with graphomotor abilities (p = 0.004) when compared to low exposure.

ED Exposure and Behavior/Unspecified Disorders
Here, we included symptoms (mostly behavioral symptoms) that cause impairment in social, occupational, or other important areas of functioning, that do not meet the full criteria for any of the disorders in the neurodevelopmental disorders diagnostic class ("unspecified neurodevelopmental disorders" as defined in DSM-5).
The results of thirteen studies published between 1994 and 2017 are shown in Table 6. Six studied perinatal exposure to BPA. Among these, three studies used the Health Outcome and Measures of the Environment (HOME) cohort and showed no associations between behavior and BPA exposure at 5 weeks [61], but found an association in girls, aged 2 and 3 years old [62,63]. In older children (aged 10 years old), three other studies found an association between urinary BPA concentration and behavior [64][65][66].
Five studies examined the consequence of prenatal exposure to phthalates. At 5 days, no association was found [67]. At 5 weeks, authors found an association with decreased regulation, decreased handling, and nonoptimal reflexes [61]. In older children (3-9 years old), all three studies found an association with several behavioral abnormalities [68][69][70]. Finally, two studies found an association with prenatal exposure to PCB and behavior [71,72]. Behavior (Rutter's Child Behavior Scale and Werry Weiss Peter's Activity Scale) PCB (prenatal) Interrogation (contaminated oil) n = 115 (followed 6 years) + 115 controls Exposed children found to have scores 7% to 43% (mean = 23%) higher than the control children on the Rutter scale at every time point.
BPA (pre and postnatal) Urinary BPA = 1.2 ng/mL n = 239 (followed 3 years) In girls, association of prenatal exposure to BPA in with BASC-2 and BRIEF-P scores (increased 9 to 12 points). We observed a significant interaction between maternal urinary BPA and sex for several behaviors (externalizing, aggression, anxiety disorder, oppositional/defiant disorder and conduct disorder traits), but no significant associations between BPA and scores on any CBCL scales. Behavior BPA was associated with relationship problems at 3 years and hyperactivity/inattention at 5. MnBP was associated with internalizing behavior, relationship problem and emotional symptoms at 3. MPzP was associated with internalizing problems and relationship problems at 3.

Discussion
The aim of this review was to present an overview of the known effects of EDCs on the emergence of cognitive and psychiatric neurodevelopmental disorders. We chose clinical approach and presented our results with categories based on DSM classification, aware that it may not corroborate the pathophysiological processes underlying each disease. We believe this different approach may be useful to clinicians, presenting our results in a more clinical and cohesive manner and providing them guidance in going through patients' medical histories. Numerous recent studies have shown that exposure to EDC can be linked to specific or less specific neurodevelopmental disorders. However, a number of issues make it difficult to characterize the relation between the exposure and the outcome.

•
First, the exposure itself can be difficult to define in terms of duration and moment of exposure.
The specific measurement of one exposure may be difficult to obtain. As of today, more than 800 products are considered as EDCs, some being commonly found in our environment. We are probably mostly studying mixture effects, and it may be difficult to isolate one specific exposure [74]. Long-lived EDCs are easier to study than many other EDCs (such as BPA and phthalates) that are quickly metabolized in humans and rapidly degraded in the environment. Some papers have tried to overcome this issue in animals, but it has yet to be done in humans [75]. In humans, evidence is mostly based on correlations between concentrations of chemicals and outcomes, assessed at one point during an individual's life, which is not a real-world situation. At some very specific times, even in very low concentrations, any exogenous EDCs may exceed the body's natural endogenous hormone levels, changing target cells that are sensitive to hormones. Thus, even extremely low dosages of EDCs can alter biological outcomes. Moreover, studies showed that, due to nonmonotonic dose-response curves, the effects of low doses cannot be predicted by the effects observed at high doses [76]. • Second, the pathophysiological processes involved with EDCs have not been clearly elucidated yet. If, by definition, they are all able to alter the physiological endocrine system, their pathophysiology may differ or overlap. EDCs interfere with the endocrine system in multiple ways, making it difficult to link one pathway to one symptom. They can act directly on the behavior and development via sexual hormone perturbations (for example, the anti-androgenic effect of phthalates [77]), but they could interfere directly with the neuronal development as well; for instance, interactions with the vasopressin system [78] and oxytocin [79] have been described, among others. • Third, the outcome measurement is most often based on neuropsychological assets that may not correlate to a clinical diagnosis, but more to a 'profile.' The fact that diagnosis in the developmental disorder categories co-occur frequently makes it difficult to determine the specificity of the cause-effect relationship [80].
Under the 'neurodevelopmental disorders' section, the DSM-5 includes disorders such as intellectual disabilities, communication disorders, ASD, ADHD, specific learning disorders and motor disorders. Those disorders are mainly diagnosed in children. We know now that other psychiatric disorders, such as schizophrenia, diagnosed years later, are also considered as neurodevelopmental disorders. We could assume that perinatal exposure to EDCs also has an impact in these 'adult' disorders [81]. Although research in the field is still scarce [82], an effort in that direction now seems crucial.
Finally, the expanding work around EDCs is recently taking a new turn with the notion of epigenetic transgenerational inheritance and of all its possible consequences [82]. Indeed, it has been reported that prenatal exposure to an EDC may not only cause adult late-onset disease, but may affect future generations through the germline [83]. Transgenerational behavioral phenotypes of development exposure have indeed been shown for two EDCs, namely vinclozin [84] and BPA [85]. These transgenerational effects may be supported by molecular alterations to the germline, promoting effects on subsequent generations. As reported in a recent communication by Skinner, recent work suggests that endocrine disruptors may alter the epigenetic programming of the germline, these changes being transmitted across generations in the absence of direct exposure [86].

Conclusions
Sufficient data exist to report that exposure to some EDCs participates in the emergence of neurodevelopmental disorders. This may account for the growing incidence of neurodevelopmental disorders observed in the general population. Despite the current effort in the field, the specificity of this link in terms of exposure and outcome will be difficult to assess. Indeed, the measurements of exposure, use of biomarkers and neuropsychological tools used for outcome all show an extreme variability in the literature, making it difficult to accumulate evidence and identify pathophysiological processes. While EDCs are increasingly prevalent in our environment, more cohorts in humans and experimental studies in animals are urgently needed to bring a more comprehensive picture of the long-term neurodevelopmental consequences of EDCs, as well as to define the precise windows of vulnerability.