Childhood atopic disorders in relation to placental changes—A systematic review and meta‐analysis

Fetal programming may arise from prenatal exposure and increase the risk of diseases later in life, potentially mediated by the placenta. The objective of this systematic review was to summarize and critically evaluate publications describing associations between human placental changes and risk of atopic disorders during childhood. The review adhered to the Preferred Reporting Items for Systematic Reviews and Meta‐analysis guidelines. The inclusion criteria were original research articles or case reports written in English describing a human placental change in relation to disease occurring in offspring during childhood. The MEDLINE and EMBASE databases were searched for eligible studies. Risk of bias (RoB) was assessed using the ROBINS‐I tool. The results were pooled both in a narrative way and by a meta‐analysis. Nineteen studies were included (n = 12,997 participants). All studies had an overall serious RoB, and publication bias could not be completely ruled out. However, five studies showed that histological chorioamnionitis in preterm‐born children was associated with asthma‐related problems (pooled odds ratio = 3.25 (95% confidence interval = 2.22–4.75)). In term‐born children, a large placenta (≥750 g) increased the risk of being prescribed anti‐asthma medications during the first year of life. Placental histone acetylation, DNA methylation, and gene expression differences were found to be associated with different atopic disorders in term‐born children. There is some evidence supporting the idea that the placenta can mediate an increased risk of atopic disorders in children. However, further studies are needed to validate the findings, properly control for confounders, and examine potential mechanisms.


| INTRODUC TI ON
Atopic disorders, including food allergies, allergic rhinitis, atopic dermatitis (AD), and asthma, are among the most common chronic disorders in children living in developed countries.They originate through a complex crosstalk between genetic susceptibility and environmental factors.The pathogenesis involves a shift of the immune system from a T helper (Th) 1 to a Th2-dominant response. 1e development of the immune system begins in fetal life and continues in the neonatal period, when the immune response is characterized by Th2 dominance.However, with increasing postnatal contact with different pathogens, the immune response normally shifts toward a Th1-type response. 2Atopic disorders can manifest early in infancy, and specific phenotypic characteristics have been found in healthy neonates who later develop an atopic disorder. 3,4These findings indicate that the origin of atopic disorders may already be found in fetal life. 5The developmental origins of health and disease (DOHaD) theory, first proposed by Barker,6 suggests that fetal programming of developing organs in response to prenatal exposures increases susceptibility to diseases later in life. 7This theory also holds true for atopic disorders, as several prenatal exposures have been linked to later development of atopy. 8,9e offspring of women who live on farms during pregnancy have a reduced risk of atopic disorders, 10 and cord blood shows stronger Th1 responses. 11Maternal stress during pregnancy has been associated with a Th2-dominant response and an increased risk of wheezing in offspring, 12,13 and different atopic manifestations in offspring. 14Maternal stressful events during pregnancy, such as divorce, have also been associated with an increased risk of wheezing, allergy, asthma, and AD in offspring. 15,16Another well-studied prenatal exposure linked to an increased risk of atopic disorders in offspring, particularly wheezing and asthma, is maternal smoking during pregnancy. 17One mechanism involved in fetal programming may be epigenetic changes, 7,8 which have been reported in placentas exposed to ambient air pollution and in placentas from women who live on farms. 18,19Differences in cord blood cytokines and immune cells have also been found between neonates who later develop an allergic disease and healthy controls.For example, detectable levels of interleukin 4 and interferonγ in cord blood were associated with reduced incidences of asthma and atopy during childhood. 20creased expression of CCL22, which encodes a Th2-associated chemokine, in cord blood was found in neonates who later developed allergy. 21e maternal environment and exposures during pregnancy affect the fetus through the placenta, which is a transient, fetal organ that works in close contact with maternal tissue.The placenta is responsible for the delivery of oxygen and nutrients to the fetus and for the removal of waste products. 22,23It produces a wide range of hormones and cytokines crucial for pregnancy maintenance and fetal growth and development.One important function is to prevent immunological rejection of the fetus, which is semiallogeneic to the pregnant woman.The placenta accomplishes this task by shifting the immunological milieu toward Th2 dominance. 5,21The placenta also functions as the fetal immune system by transferring immunoglobulin G from the mother 24 and creating a barrier that protects the fetus from dangerous antigens. 257][28][29] However, even though comprehensive reviews have been published on pre-and perinatal factors that may program the fetus to an increased risk of later atopic disorders, 8,9,16,30 no systematic review to date has summarized and critically evaluated the evidence of a link between placental variations and the risk of atopic disorders during childhood.The aim of this study was to perform such a review and to investigate whether gestational age at birth (preterm or term) influences associations between placental changes and childhood atopic disorders.

| Eligibility criteria
The study was reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) guidelines 31 and registered at the International Prospective Register of Systematic Reviews (PROSPERO, CRD42021276562), where the protocol can be accessed.Inclusion criteria were original research articles or case reports written in English, published on or after the 1st of January 2000, describing a placental variable in relation to

Key message
Atopic disorders may originate in utero.The placenta determines the fetal environment.This systematic review evaluates the evidence of an increased risk of atopic disorders in children exposed to placental changes.Nineteen studies were included.Five of them studying pretermborn children found an increased risk of asthma-related disorders after exposure to histological chorioamnionitis.
In term-born children, associations were found between high placental weight and an increased risk of anti-asthma medication prescriptions during the first year of life and between DNA methylation, histone acetylation, gene expression patterns, and atopic disorders.The placenta may play a role in programming susceptibility for atopic disorders but more research is needed.
offspring disorders.The disorder had to be diagnosed after the neonatal period but before 19 years of age.The neonatal period was defined as the first 28 days of life corrected for prematurity. 32In the database searches, eligible offspring disorders were atopic and neuropsychological disorders and tumors since our initial overall aim was to investigate evidence for all these three outcomes.We have recently published a review of neuropsychological outcomes. 33In this specific review, only atopic disorders were eligible, defined as allergy, hypersensitivity, sensitization, asthma, recurrent wheezing, allergic rhinitis, and AD.All types of placental variables were accepted, for example, weight, epigenetic modifications, and gene expression; however, the variable had to be linked to the placental disc.Exclusion criteria were animal studies, studies of placental abruption due to its large impact on neonatal life, 34 and studies of twin-twin transfusion syndrome due to difficulties in discriminating effects of the placental abnormality from those of blood flow differences to fetuses. 35Combined fetal and placental variables, such as fetal-to-placental weight ratio and indirect placental variables, were not accepted.

| Search methods
We searched the EMBASE and MEDLINE databases.The search strategy built by librarians included MeSH terms and free text words (Table S1).The first search was performed in June 2021, and the second in May 2022.All duplicates were removed.An additional search for eligible studies was performed by checking the bibliographies of included studies and relevant review articles.

| Selection process
After each database search, all titles and abstracts were screened independently by two reviewers (first search: ML and JPK; second search: ML and FC).The full text of all studies that were selected by at least one reviewer was read independently by the same two reviewers as above.Disagreements were settled through discussion.ML screened bibliographies to find additional studies.

| Data collection and extraction
Two reviewers (ZB and ML) extracted the following data from included studies: authors, publication year, study design, setting, study population, primary aim, primary outcome, placental variable (exposure), atopic disorder (outcome), age at outcome assessment, statistical methods used, results of the association analysis, confounders controlled for, and funding sources.All data on relevant exposures and outcomes were collected, and no assumed data were used to replace missing or unclear information.

| Risk of bias assessment
To assess risk of bias (RoB), the preferred Cochrane tool for nonrandomized studies, ROBINS-I, 36 was used with adjustment for exposures instead of interventions.The assessment was performed independently by two reviewers (ZB and ML), focusing on the aim of this review and not on the aim of each individual study.4][45][46][47][48] All studies failing to control for these covariates were judged to have a serious RoB in the domain of confounding in accordance with instructions. 36To ensure consistency among studies, the following criteria were used: >20% failure to include eligible subjects was considered to indicate a serious RoB in the "Selection of participants" domain and >20% failure to collect outcome data was considered to indicate a serious RoB in the "Missing data" domain.A corresponding failure of 10%-20% was considered to indicate a moderate RoB, and a failure of <10% to indicate a low RoB in these domains, as described previously. 33Any disagreements were resolved by discussion.

| Synthesis methods
Studies were grouped by type of exposure and outcome and by type of study population (preterm or term).Studies with the same type of exposure, outcome, and effect measure were included in a metaanalysis.A narrative synthesis was otherwise used.

| Statistical methods
The correlation between sample and effect sizes was investigated by Spearman's rank test.A comparison of studies with and without significant associations was performed by Mann-Whitney U test, using IBM SPSS Statistics version 29.0 (IBM Corp., Armonk, NY, USA).The degree of heterogeneity between studies included in the meta-analysis was assessed using the I 2 statistic, in which an I 2 statistic >30% was considered to indicate moderate heterogeneity, and an I 2 statistic >50% was considered to indicate substantial heterogeneity.In addition, a p value <.05 from the noncentral χ 2 test for heterogeneity was considered statistically significant.The fixed-effects model was used to pool effects from individual studies because no heterogeneity was found, and the result was illustrated in a forest plot.For the pooled result, a 95% confidence interval (CI) that did not include 1 was considered statistically significant.Risk of publication bias was presented using a funnel plot. 49The meta-analysis and plots were performed and

| Ethical considerations
No ethical approval was needed, as only published results were analyzed.

| Study selection
The database searches yielded 3252 studies, of which 40 were deemed potentially eligible to this review after screening titles and abstracts.The 40 studies were read in full text and 17 of them met the inclusion criteria.By screening bibliographies, two additional studies were found, leading to a final sample of 19 studies (n = 12,997 participants) (Figure 1).

| Study characteristics and results of individual studies
1][52][53] One study included both preterm and term-born children but only sampled the placentas of preterm neonates consequently. 546][57][58][59][60][61][62] One study 51 was a follow-up of another study. 50Several studies examined the impact of environmental exposures on atopic disorder risk in offspring in relation to placental changes.These exposures included an anthroposophic lifestyle, 56,57 maternal anxiety or distress, 59,60 air pollutants, 61 vitamin D levels, 61,63 and maternal smoking together with vitamin C supplementation 64,65 (Table S2).
Histopathological disorders, mainly acute histological chorioamnionitis (HCA), were the most prevalent placental change.Gene expression and epigenetic changes were also prevalent.The most common outcomes were asthma and wheezing, followed by AD, sensitization to allergens, and allergic rhinitis (Table 1).The outcomes were assessed at varying ages (0.5-14 years) but most often before 3 years of age (Table 2).

| Risk of bias in studies
All studies had an overall serious RoB, mostly due to not controlling for every important covariate (Figure 2).

| Results of synthesis
Acute histological chorioamnionitis was the only studied placental change in preterm-born children, while many different placental changes were studied in term-born children, such as histopathological lesions, weight, and gene expression.In contrast, no corresponding difference between preterm and term-born children was found concerning outcome (Table 2).
Placental epigenetic and gene expression changes showed variable results.Significant associations were found only in subgroups, 56,61 for a minority of investigated CpGs 65 or for only one of several investigated genes. 58,60,61Another study only included a minority of the participants in the gene expression analysis. 63wever, associations were found between histone acetylation at allergy-related immune regulatory genes and allergic sensitization, 56 as well as between genome-wide DNA methylation and wheeze 65 or AD. 60,61Differential gene expression were found for LRCH1, 61 MMP9, and TLR7 in relation to AD, 66 GR-P in relation to allergic sensitization, 58 ORMDL3, 66 HLA-DRB1, and HLA-DRB5 in relation to asthma, 63 and KITL1 in relation to allergic rhinitis 66 (Table 2).
However, sample sizes and effect sizes in these 10 studies did not correlate (rho = .176,p = .626),and there was no significant difference in sample size between studies reporting and studies not reporting an association.

Placental changes
Weight, volume, and/or surface area 4

Outcomes
Wheezing, asthma, and respiratory health 12

Food allergy 2
Sensitization to allergens 5 Abbreviation: 11β-HSD2, 11β hydroxysteroid dehydrogenase type 2.  Transcriptome analyzes to confirm the methylation findings identified a significant difference in the placental expression of AHR, but no differences related to DPP10 or HLA-DRB1 were found.

TA B L E 1
In children with high prenatal PM

| DISCUSS ION
To the best of our knowledge, this is the first systematic review on associations between placental changes and risk of atopic disorders.
Only 19 studies could be included, even though atopic disorders are common, the performed search was broad, the placenta is easily accessible after delivery, and the interest in the DOHaD theory.The studies comprised approximately 13,000 children, of whom 9700 came from one single study, 67 pointing to the scarcity of evidence in this research field.However, we found a consistent association between HCA and an increased risk of asthma or asthma-related conditions in preterm-born children.In term-born neonates, a large placenta (≥750 g) increased the risk of prescription of anti-asthma medications during the first year of life according to one single study, 67 and placental epigenetic and gene expression changes were associated with atopic disorders.
Chorioamnionitis is a well-known cause of preterm birth.Due to immature lungs, preterm neonates are at risk of developing respiratory distress syndrome, bronchopulmonary dysplasia, and chronic lung disease (CLD) of prematurity. 69Bronchopulmonary dysplasia is associated with increased prescription of anti-asthma medications 70 and asthma-like symptoms in children. 71Acute histological chorioamnionitis may further increase the risk of neonatal pulmonary morbidity in addition to the risk imposed by prematurity alone, 69 and clinically diagnosed chorioamnionitis has been associated with an increased risk of childhood asthma irrespective of prematurity. 72However, associations between histologically proven placental inflammation and childhood respiratory morbidity are not as clear. 73The pathogenesis linking HCA to later respiratory morbidity is thought to include a fetal inflammatory response to cytokines and chemokines released in the presence of HCA. 74This response may damage fetal lungs and airways and persists in the neonatal period as systemic inflammation. 75The result found here of an increased risk of asthma and asthmalinked symptoms in preterm-born children exposed to HCA [50][51][52][53][54]76 agrees with these previous results. Bess damage to fetal lungs and airways, which is more deleterious to preterm-born neonates, HCA may also affect the developing immune system, 75 which may also increase the risk for later atopic disorders.The effect on the immune system seems to be stronger if the insult occurs earlier during fetal life.75 Since HCA is a strong risk factor for preterm birth, these factors may explain why only preterm-born children exposed to HCA were found to have an increased risk of atopic disorders in this review.Term-born children exposed to HCA did not show this increased risk (see Table 2).[55][56][57]63   A positive API was defined as early frequent wheezing (>3 episodes during the first 3 years of life) and at least one of two major criteria (physician-diagnosed asthma or physician-diagnosed eczema) or two of three minor criteria (physician-diagnosed allergic rhinitis, wheezing apart from colds or eosinophilia).

TABLE 2 (Continued)
gene expression, and epigenetic changes.The association between a large placenta and an increased risk of prescription of anti-asthma medications found in one study 67 has similarities with previous reports.Larger placentas have been associated with increased blood pressure in adults 77 and with psychiatric disturbances in boys. 78large placenta may be secondary to maternal diabetes, 79 which has been associated with childhood atopic disorders. 80Placental overgrowth may be a compensatory mechanism to overcome poor nutritional flow to the fetus to maintain normal growth or to maintain sufficient nutritional flow when the demand is high due to increased fetal growth, as in the case of maternal diabetes.However, fetal organ development may be impaired in these situations, increasing the risk of later diseases.On the other hand, prescription of anti-asthma medications is only a proxy for asthma and so far, there are no adequate studies on associations between placental weight and childhood asthma.Further research is needed.ing the whole genome or epigenome, 60,61,[63][64][65] while the others used a targeted approach investigating a few selected genes. 56,58,66fferences in approach between the included studies may partly explain the diversity in the genetic findings.
Many of the genes reported here to be associated with offspring atopic disorders (see Table 2) are involved in the immune system or F I G U R E 2 Results of the risk of bias assessments of the included studies using the ROBINS-I tool.
Other genes are involved in basic biological processes with some associations to development (HDAC4, 92 C18orf1, 93 KITL1, 94 and MMP9 89 ) or atopic disorders (PRKCA, 95 DPP10, 96 and ORMDL3 97 ).None of these genes has been examined in animal or in vitro studies yet, as far as we know.Such studies are needed to prove causation between their expression in placenta and the risk of atopic disorders in offspring and to further understand the effects of the genes in this context.In addition, the gene expression findings need to be validated in other human populations and at the protein level in placenta.
There are several limitations of the studies included in this review.Only observational studies were included, and they cannot establish causation.On the other hand, randomized controlled trials in humans investigating our research question can hardly be performed for ethical and other reasons.In observational studies, it is important to control for potential confounding variables.However, many of the included studies did not control for any covariate.Two studies controlled for post-exposure variables (CLD and wheezing) that might have been affected by the exposure, 50,51 which is not appropriate and might have introduced bias. 36None of the studies controlled for all variables deemed important (smoking, sex, GA, and BW).The sample sizes were most often small (11-439 children), which probably reduced the ability to find significant associations, as the potential influence of the placenta on the risk of developing a complex disease, such as atopic disorders, later in life is likely to be rather small.Small sample sizes also hindered proper controlling for confounders.
The strengths of this review include the strict adherence to the PRISMA guidelines, 31 the broad search including all types of placental and atopic disorders, and the recognition of several important confounders.
The limitations of this review consist of the lack of consideration of other important covariates, such as shared genetics, mode of delivery, and environmental exposures other than smoking.However, some of the included studies did control for these covariates, and one study investigated the effects of air pollution. 61Other limitations are the exclusion of combined fetal-placental variables, such as the fetal-to-placental weight ratio, which might have been interesting to study, as this ratio was found to be associated with AD in girls. 62The exclusion of publications prior to the year 2000 and the exclusion of research letters might have excluded some important findings.Inclusion of cord blood findings could have deepened our understanding of the impact of the prenatal period on later risk of atopic disorders.][100][101][102] The tool used for the RoB assessment was mainly constructed for cohort studies, making it somewhat difficult to use in case-control studies.Furthermore, publication bias could not be completely excluded.

F I G U R E 4
Meta-analysis of the associations found between histological chorioamnionitis and asthma-related disorders in preterm-born children.
F I G U R E 3 Forest plot of all 10 studies presenting the association between a placental variable and an atopic disorder as an odds ratio.FIRS, fetal inflammatory response syndrome; GR, glucocorticoid receptor; HCA, histological chorioamnionitis; HSD-2, 11β hydroxysteroid dehydrogenase type 2.
constructed in Stata version 18.0 (Stata Corp, College Station, TX, USA) and R version 4.3.0(R Foundation for Statistical Computing, Vienna, Austria).

2. 5
exposure and low CB vitamin D levels, the following placental gene expression differences (FC) were found between children with and without AD (n = 12): AHR: -1.900 (p = 0.073) OBSCN: -1.164 (p = 0.146) LRCH1: -1.529 (p = 0.003) NMRAL1: -1.016 (p = 0.817) ZNF101: -1.077 (p = 0.487) PCDHB8: -1.009 (p = 0.912) NPR2: -1.084 (p = 0.667) glutathione abundance, glutathione to glutathione disulfide ratios, nor 11β-HSD2 abundances in placenta differed significantly between offspring with and without AD None Rindsjö, 2015 58 Term, only females n = 28-29 2 years Expression of genes encoding the GR (three isoforms) and 11β-HSD2 Sensitization to food and aeroallergens OR (95% CI) for an association between sensitization and: GRα at the maternal side of the placenta: 0.50 (0.063-1.824)GRα at the fetal side: 0.38 (0.043-2.058)GRβ at the maternal side: 0.33 (0.046-1.653)GRβ at the fetal side: 2.46 (0.731-10.907)GR-P at the maternal side: 0.04 (0.0009-0.678)GR-P at the fetal side: 1.15 (0.265-4.564) 11β-HSD2 at the maternal side: 0.52 (0.168-1.441) 11β-HSD2 at the fetal side: 0.79 (0.242-2.195)None, but controlling for parental sensitization did not substantially change the ORs (exact data not shown) six placentas (33%) from children that later developed asthma had chronic chorioamnionitis compared to none of the 21 children that had not developed asthma by age 3 years (p = 0.043).No other histopathological differences nor morphometric differences were found in the placentas from children that later developed asthma or not.Placentas from children that later developed asthma (n = 5) had lower expression of HLA-DRB1 and HLA-DRB5 compared to placentas from children that had not developed asthma by age 3 years (n = 6).But no other gene expression differences were found c Epigenetic processes regulate gene transcription, are influenced by the environment and essential to normal development.Thus, they are relevant when studying mechanisms for fetal programming.Increased histone acetylation generally increases gene transcription,while increased DNA methylation typically decreases transcription.Some of the studies included here investigating gene expression, either directly or indirectly through epigenetic markers, examined an environmental exposure, such as air pollution, 61 maternal smoking,65 anxiety,60 or vitamin D sufficiency,63 in relation to later risk of atopic disorders.Most studies used an explorative approach search- Flowchart of the inclusion of studies in the review.

Preterm or term-born study population a Numbers b Age at outcome assessment Placental change Outcome Associations found between placental change and offspring outcome Confounders controlled for
Results from the included studies sorted by type and size of the study population.
TA B L E 2

Source Preterm or term-born study population a Numbers b Age at outcome assessment Placental change Outcome Associations found between placental change and offspring outcome Confounders controlled for
NoneTA B L E 2 (Continued)

with eczema had significantly lower expression of MMP9 compared to boys without eczema, and the same was true for girls. Girls with eczema had significantly lower expression of TLR7 compared to girls without eczema, but
no corresponding difference was seen in boys.

Source Preterm or term-born study population a Numbers b Age at outcome assessment Placental change Outcome Associations found between placental change and offspring outcome Confounders controlled for
a In relation to the association reported here.b Numbers of children included in the association analysis between placental change and atopic disorder.