Association Between Prenatal Exposure to Metals and Atopic Dermatitis Among Children Aged 4 Years in Taiwan

Key Points Question Is prenatal exposure to arsenic and other metals associated with the occurrence of atopic dermatitis in young children? Findings In this cohort study of 586 mother and child pairs, prenatal arsenic exposure was associated with approximately 2.4 times the occurrence of atopic dermatitis in children at age 4 years. In addition, coexposure to arsenic and cadmium was associated with increased odds of developing atopic dermatitis. Meaning The results of this study suggest that prevention of inorganic arsenic and cadmium exposure among pregnant women could reduce the risk of atopic dermatitis and other allergic diseases in children.


Introduction
The increased prevalence and severity of allergic symptoms have become an important issue for the global disease burden. 1 Approximately 20% of children and 3% of adults have been diagnosed with atopic dermatitis worldwide. 2A recent study found that adults with active or severe atopic dermatitis had an increased risk of all-cause death and certain causes of death, such as infectious and respiratory diseases. 3Atopic dermatitis begins to emerge at age 1 to 2 years 4 and is regarded as the initial stage in a series of childhood allergic diseases, such as atopic asthma. 5e risk factors associated with atopic dermatitis include genetic factors, obesity, climate factors, and environmental pollutants. 2 Previous studies have reported an association between exposure to heavy metals, including cadmium and lead and allergic diseases in adults. 6,7Another recent study found that prenatal inorganic arsenic exposure was associated with allergic airway inflammation in children up to age 14 years. 8Among infants, exposure to lead and chromium was associated with both atopic dermatitis symptoms and severity. 9However, the association between multiple metal exposure during pregnancy and the risk of atopic dermatitis in young children remains unknown.
Environmental exposures during gestational development might increase susceptibility to disease in later life. 10The mechanism of disease may involve activation of the aryl hydrocarbon receptor and the NFE2L2 gene. 11Findings from studies of the general population have also suggested that allergic sensitization may be associated with increases in disease susceptibility. 12,13Fetal and early postnatal exposures to arsenic, cadmium, lead, mercury, and chromium may be important factors associated with immunotoxic effects. 14,15Because exposure to chemicals in the environment is complex and part of the exposome associated with health and disease, 16 this birth cohort study aimed to investigate the association between prenatal exposure to multiple metals and the risk of atopic dermatitis in young children.

Study Population
The participants in this study were enrolled in the Taiwan Maternal and Infant Cohort Study (TMICS), a birth cohort study conducted at 9 hospitals in northern, central, southern, and eastern Taiwan.A total of 1152 pregnant women were enrolled in the initial TMICS from October 2012 to May 2015.
Mothers and children from the central and eastern regions were later invited via telephone to participate in follow-up interviews conducted from August 2016 to January 2019.In total, 566 participants were unavailable for follow-up because they moved away from the area without leaving contact information or had scheduling conflicts, and 586 children were successfully followed up at age 4 years.A total of 216 children were excluded from the final analysis because they were missing data on (1) maternal urinary metal concentrations (n = 137); (2)

Data and Specimen Collection
Structured questionnaires, including items about demographic characteristics, lifestyle, dietary patterns, and residential environment, were administered to all participating women by trained interviewers during the third trimester of pregnancy.Physical measurements (eg, height, weight, and blood pressure) and urine samples were obtained, with urine samples collected in glass bottles and stored at −20 °C until analysis.Follow-up interviews were conducted in person at participants' homes or participating hospitals.The presence of atopic dermatitis among children was defined as a positive response to the question, "Has your child ever had atopic dermatitis diagnosed by a physician?"during the questionnaire interview at age 4 years.

Assessment of Metals in Maternal Urine
Frozen urinary specimens were moved to a 4 °C refrigerator for 2 to 4 hours, then defrosted at room temperature (26 °C-28 °C).Urinary metal concentrations were quantified using an inductively coupled plasma mass spectrometer (Agilent 7700x ICP-MS; Agilent Technologies) with a quadrupole mass filter and autosampler.For quality control, we used standard reference materials, spike recovery rate, and blank and duplicate samples.We performed the quality control regimen for every 10 samples, and the control spike recovery rates ranged from 89.0% to 114.5%.The limits of detection were 0.399 μg/L for arsenic, 0.066 μg/L for cadmium, 0.022 μg/L for lead, 0.016 μg/L for cobalt, 0.225 μg/L for copper, 0.090 μg/L for nickel, 0.008 μg/L for thallium, and 0.417 μg/L for zinc.For instances in which the outcome was below the limit of detection, concentrations were estimated as the limit of detection divided by the square root of 2. Metal concentrations in maternal urine were divided by urine creatinine for urine volume correction (μg/g creatinine).

Estimation of Total Inorganic Arsenic in Maternal Urine
The urinary concentration of inorganic arsenic represents the toxic form of total arsenic exposure, and the urinary concentration of arsenic species (arsenite, arsenate, monomethylarsonate, and dimethylarsenate) provides a better estimate of inorganic arsenic exposure than the concentration in blood.We calculated total inorganic arsenic as the sum of arsenite, arsenate, monomethylarsonate, and dimethylarsenate.We estimated total inorganic arsenic using a linear regression analysis of data from a previous TMICS pilot study 8 derived from 336 women for whom both total urinary arsenic and total inorganic arsenic concentrations were available (with estimated total arsenic level calculated as the sum of 21.35241 and 0.14493 multiplied by the arsenic level).
Estimated total inorganic arsenic concentrations were highly correlated with total inorganic arsenic concentrations (r = 0.61; P < .001)(eTable 2 in Supplement 1).Therefore, we used estimated total inorganic arsenic when assessing total inorganic arsenic concentrations.

Statistical Analysis
Mean values with SDs were used to summarize continuous variables, and 2-sided unpaired t tests were used to compare the differences between groups.Median values with IQRs, mean values with SDs, and Mann-Whitney U tests were used to summarize the concentrations of maternal urinary metals and compare the differences between groups.We used frequencies with percentages to summarize categorical variables and χ 2 tests to comparison groups.Maternal urinary metal concentrations were log 2 -transformed for normality.Logistic regression analyses were used to estimate the association between prenatal exposure to 8 metals (arsenic, cadmium, lead, cobalt, copper, nickel, thallium, and zinc) and the risk of atopic dermatitis in children.We calculated Q values to assess the multiple comparison effect and reduce type 1 error probability. 17 adjusted for the child's sex (male or female), the presence of parental allergies (yes or no), geographic area (central or eastern), exposure to tobacco smoke during pregnancy and among children at age 4 years (yes or no), and maternal educational level (<12 years, 13-16 years, or >16 years).In sensitivity analyses, we used a directed acyclic graph to select the minimum sufficient factor adjustment for the model. 18We also calculated propensity scores for potential imbalances in exposure groups, and we reduced bias by using a process resembling randomization. 19Previous studies have reported that parental allergies might be associated with 30% of the risk of asthma in children 20 and the risk of atopic dermatitis in infants. 21In addition, maternal allergies may be associated with a higher risk of atopic dermatitis among children compared with paternal allergies. 22erefore, we stratified groups by maternal history of allergic diseases (including allergic dermatitis, asthma, and rhinitis) and included only participants with complete data on maternal allergies.
In the present study, maternal urinary metal concentrations were highly correlated among metals (eFigure 1 in Supplement 1); thus, we performed a generalized weighted quantile sum (WQS) regression analysis to evaluate the mixture effect of coexposure to metals. 23The data were randomly divided into training (40%) and validation (60%) sets.We used a bootstrap method on the training data set to quantify the WQS index.The validation data set was then used to test the accuracy of the WQS index.The possible selection threshold of metals was defined as the weight of each metal that was greater than the quotient (0.125) of the total sum of weights in the WQS index (sum = 1) divided by the number of metals (n = 8).We used SAS software, version 9.

Results
Among 370 children included in the final analysis, the mean (SD) age was

Prenatal Exposure to Metals
In the fully adjusted model, every doubled increase in maternal urinary estimated total inorganic arsenic was significantly associated with 2.42-fold (95% CI, 1.33-4.39;P = .003)higher odds of atopic dermatitis in children at age 4 years (

Discussion
This cohort study found that prenatal exposure to arsenic was positively associated with the risk of atopic dermatitis among children aged 4 years in Taiwan, and the odds of developing atopic a Data missing for 14 participants.
b Data missing for 44 participants.
c Data missing for 109 participants.
d Data missing for 20 participants.
e Data missing for 106 participants.
f Data missing for 202 participants.
g Calculated as the sum of 21.35241 and 0.14493 multiplied by the arsenic level.dermatitis were approximately 2.4-fold higher when estimated total inorganic arsenic was evaluated.
We also observed a significant association between estimated total inorganic arsenic concentrations and the development of atopic dermatitis among children who had mothers with allergies.The low exposure to tobacco smoke among children with atopic dermatitis in this study may be associated with altered behavior among parents attempting to protect their children with allergic diseases.
The pathophysiological features of atopic dermatitis include epidermal barrier dysfunction, skin inflammation, interaction between the nervous and immune systems, genetic susceptibility, and the effect of microbiota. 24The polarization of T cells in atopic dermatitis is considered biphasic, with a predominant response from helper T (T H ) 2 cells in the acute phase and from T H 1 cells in the chronic phase, and T H 2 cells have been suggested to have important roles in cutaneous allergic inflammation as an integral part of atopic dermatitis. 25A previous study reported that newborns with the specific epigenetic pattern of T H 2 cell-skewed immunity may have increased responses to common environmental allergens later in life. 26Exposure to arsenic during pregnancy has been associated with a DNA methylation pattern in infants 27 as well as changes in children's immunoglobulin E levels. 28though metals other than arsenic (ie, cadmium, lead, cobalt, copper, nickel, thallium, and zinc) were examined in this study, these metals were not associated with the incidence of atopic dermatitis in children at age 4 years; however, an association was found between prenatal coexposure to these metals (including arsenic) and atopic dermatitis in children.The exposure to cadmium was secondary to arsenic in the WQS index.Cadmium is also a well-established toxic metal that can detrimentally impact many organs, including the bones, kidneys, and pancreas. 29Kim et al 30 observed an association between prenatal cadmium exposure and atopic dermatitis in infants aged 6 months, and elevated cadmium concentrations in cord blood were associated with an increased risk of atopic dermatitis.
Hanson et al 31 treated pregnant mice with 10 ppm of cadmium chloride as an environmentally relevant dose.They found that cadmium exposure may induce decreases in the production of interferon gamma and CD8 + CD223 + T cells in the spleens of their offspring at age 7 weeks.A recent study reported that cadmium concentrations mediated by in-cell viability had differentially functional effects on in vitro murine macrophages and mast cells after cadmium exposure. 32Mast cells and macrophages have been suggested to play a role in atopic dermatitis-associated inflammation and pruritus.For example, interleukin 31 is derived from T H 2 cells, macrophages, mast cells, and mast cells in atopic dermatitis lesions with higher interleukin 31 immunoreactivity than normal skin. 33Interleukin 23, secreted by dendritic cells and macrophages, has been documented to be an important cytokine to control inflammation in peripheral tissues.Skin tissues of patients with atopic dermatitis have substantially increased activation of the T H 17/interleukin 23 and T H 22 axes and keratinocyte production. 34Thus, prenatal exposure to cadmium may be associated with atopic dermatitis in early childhood through disruption of immune cell development in the fetus.After stratifying the analysis by maternal history of allergic diseases, we also found a significant association between maternal arsenic exposure and atopic dermatitis in children at age 4 years.One plausible explanation for this association may be heritability, including genetic and epigenetic factors.
Children with mothers who had allergies had greater odds of developing atopic dermatitis, which suggests a genetic role and an association with the in utero environment, which could shape neonatal immunity and, consequently, the development of atopic dermatitis in early childhood. 357][38] Therefore, improving the environment and reducing these potential sources are important steps to reducing the risk of exposure to metals in this susceptible population.

Strengths and Limitations
This study has strengths.Its main strength is its birth cohort design, in which urinary samples were collected from pregnant women in the third trimester and used to estimate the likelihood that children would develop atopic dermatitis 3 or 4 years later.The use of urine sampling could have increased the participation rate among pregnant women because urine sampling is less invasive than venous blood sampling.Moreover, blood arsenic concentrations only reflect recent exposure and cannot be used as a reliable estimate of chronic exposure. 39Our participants are children aged 4 years, and atopic dermatitis is the most commonly observed allergy that can be well diagnosed.The results were similar to those obtained when the analysis was adjusted for the presence of different urinary metals at age 4 years (data not shown), probably owing to the different exposure scenarios.
We used well-designed methods for data collection.We examined the association between coexposure to a mixture of metals and atopic dermatitis because of the complex consequences of metal exposure for human health, and some urinary metal concentrations were highly correlated. 40is study also has limitations.First, there may be insufficient power in the statistical analyses because of the relatively small sample.However, the bootstrap method used in the generalized WQS model was applied to the training data set to reduce the partial effects of insufficient power.Second, we collected maternal urine samples in glass bottles; thus, we only measured total arsenic concentration and did not separately measure the concentrations of individual arsenic species.Therefore, total inorganic arsenic values were not measured directly but estimated using an equation derived from data on total arsenic and total inorganic arsenic concentrations from a previous birth cohort study among pregnant women residing in the same geographic area.Two basic metabolic processes of inorganic arsenic are oxidation-reduction and methylation reactions.Approximately 70% of absorbed organic and inorganic arsenic is excreted from the kidneys via urine. 41ird, we used parent-reported atopic dermatitis in this study without obtaining data from a complete clinical examination conducted in parallel by a physician.The reliance on parental reporting might produce underestimation of the results through nondifferential misclassification because the women did not have knowledge of their urinary metal concentrations in the third trimester; thus, they would not have been likely to overreport their children's atopic dermatitis (vs people who may overreport symptoms or disease when they realize that they have had high exposure).Moreover, the parent-reported presence of atopic dermatitis in children was based on "Has your child ever had atopic dermatitis diagnosed by a physician?"This question may be more reliable than, "Has your child ever had atopic dermatitis?" Fourth, we reported observations among a birth cohort of children aged 4 years residing in the central and eastern regions of Taiwan; thus, external validity may be low, and results may not be generalizable to all children aged 4 years in Taiwan.Fifth, the possibility of residual or unmeasured confounding from genetic factors and exposure to other environmental chemicals cannot be eliminated.Epigenetic factors, such as modifications to genomic DNA and posttranscriptional micro-RNA regulation, have also been associated with the development of atopic dermatitis. 42In addition, we adjusted the model for the presence of parental allergies and further stratified the analysis by maternal allergic status to reduce partial confounding from genetic factors.

Conclusions
This birth cohort study found an association between prenatal coexposure to metals and atopic dermatitis in children at age 4 years.Arsenic followed by cadmium accounted for most of the metal mixture index associated with atopic dermatitis.Prevention of exposure to inorganic arsenic and cadmium during pregnancy may be helpful for the control of atopic dermatitis and other potential allergic diseases in young children.Environmental improvements and reductions in contact with sources of exposure are important to reduce the risk of metal exposure among this susceptible population.
allergic diseases in children at age 4 years (n = 54); or (3) exposure to environmental tobacco smoke among children at age 4 years, maternal educational level or allergic diseases, or child's sex (n = 25).The final analysis of follow-up JAMA Network Open | Allergy Prenatal Exposure to Metals and Atopic Dermatitis in Children Aged 4 Years in Taiwan JAMA Network Open.2021;4(10):e2131327. doi:10.1001/jamanetworkopen.2021.31327(Reprinted) October 27, 2021 2/14 Downloaded From: https://jamanetwork.com/ on 09/20/2023 data included 370 mother and child pairs (Figure 1; eTable 1 in Supplement 1).The study was approved by the ethics committees of the National Health Research Institutes and the participating hospitals in Taiwan.All participating women provided written informed consent during their third trimester of pregnancy (for participation in the original study), and additional written informed consent was obtained before follow-up interviews.This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cohort studies.

Table 1 .
Characteristics of Children in Birth CohortPrenatal Exposure to Metals and Atopic Dermatitis in Children Aged 4 Years in Taiwan

Table 1 .
Characteristics of Children in Birth Cohort (continued)