Association between environmental mercury exposure and allergic disorders in Korean children: Korean National Environmental Health Survey (KoNEHS) cycles 3–4 (2015–2020)

Although previous studies have suggested potential adverse effects of mercury on a child’s immune system, the associations have been inconsistent. We aimed to determine the association between urinary mercury levels and allergic diseases in Korean children with high mercury exposure. Data from 853 and 710 children aged 6–11 years in the Korean National Environmental Health Survey (KoNEHS) cycle 3 (2015–2017) and cycle 4 (2018–2020) were analyzed. We examined the association between mercury exposure and the prevalence of atopic dermatitis (AD), asthma, allergic rhinitis (AR), and allergic multimorbidity. After adjusting for all covariates, the urinary mercury level was positively associated with AD in the 2015–2017 study (OR = 1.34, 95% CI = 1.01, 1.79) and AR in 2018–2020 study (OR = 1.46, 95% CI = 1.01, 2.10). Pooled effects showed OR of 1.34 (95% CI = 1.01, 1.79) for AD and 1.47 (95% CI = 1.01, 2.12) for allergic multimorbidity. The association with allergic multimorbidity was greater in boys (OR = 1.88, 95% CI = 1.01, 3.49) than in girls (OR = 1.25, 95% CI = 0.73, 2.14). These results suggest that environmental mercury exposure may exacerbate symptoms of atopic dermatitis and allergic multimorbidity in children.


Allergic outcomes
Information on current symptoms of atopic dermatitis, asthma, and allergic rhinitis was collected through a structured questionnaire.Since the participants were children, the guardians of the research participants completed the questionnaire, and trained surveyors checked the questionnaire for the guardians.To determine the current presence of allergies, participants who had received a physician's diagnosis were asked the following two questions: "Do you currently have symptoms of atopic dermatitis/asthma/allergic rhinitis?"and "Are you currently taking medication for atopic dermatitis/asthma/allergic rhinitis?"Individuals who answered "yes" to either of these two questions were considered to have current allergic diseases.As allergic disorders tend to coexist with other symptoms 19 , the presence of multimorbidity was noted if a participant had two or more of the three diseases of atopic dermatitis, asthma, and allergic rhinitis at the same time 20 .

Covariates
To control for potential confounding factors, the covariates included sex, age, parental education level, monthly household income, body mass index (BMI), urinary cotinine level, and urinary creatinine concentration.All data were collected through KoNEHS surveys, physical examinations, and experimental measurements.Parental education level was classified into three stages (college or associated degree, university, and graduate school) according to the highest educational qualification level of the mother or father; monthly household income was classified into four stages (low, lower-middle, upper-middle, and high).BMI was calculated by dividing the weight (kg) by the square of the height (m 2 ).The urinary cotinine level was used as a biomarker for smoking exposure 21 , and the urinary creatinine concentration was used to adjust for the urinary mercury concentration.

Statistical analysis
As the KoNEHS surveyed based on a two-stage proportionally stratified sampling design, we conducted the statistical analysis using the complex sampling design, which considered sampling unit, strata, and sampling weights 17 .Because KoNEHS does not provide integrated weights for each cycle or data to calculate them, the data from the 3rd and 4th cycles were analyzed separately.Due to the left-skewed distribution, urinary mercury, urinary cotinine, and urinary creatinine concentrations were natural log-transformed before the analysis.Age, urinary creatinine concentration, urinary cotinine concentration, and BMI were used as continuous variables, whereas sex, parental education level, and monthly household income were used as categorical variables.
To compare the average mercury levels among demographic groups, independent t-tests and Wald F tests were used.Logistic regression analysis was used to confirm the association between mercury concentration and allergic disorders, and odds ratios (ORs) and 95% confidence intervals (CIs) were calculated.Furthermore, multinomial logistic regression analysis was used to confirm the relationship between urinary mercury concentration and the multimorbidity of each disease.The multimorbidity of allergic disorders was grouped into "none," "one symptom," and "two or three symptoms," depending on the individual's symptom count.The regression model was adjusted for sex, age, parental education level, monthly household income, BMI, urinary cotinine concentration, and urinary creatinine concentration.A stratified analysis was conducted to compare the magnitude of the association by sex.To calculate the overall effect size in the 3rd and 4th cycles, fixed-effect models were calculated using the Mantel-Haenzel method in the R "meta" package.
Statistical analyses were performed using SPSS 26.0 and R 4.2.2, and statistical significance was defined as p < 0.05.

Results
Table 1 shows the demographic and clinical characteristics of the study participants of the 3rd and 4th cycles.The average age by cycle was 8.50 years and 8.53 years, respectively, with about half of the participants being male in both periods.The prevalence rates of allergic disorders in the 3rd and 4th cycles were 9.3% and 8.9% for atopic dermatitis, 0.8% and 0.8% for asthma, and 33.1% and 32.1% for allergic rhinitis, respectively, with no significant difference by cycle.
Figure 1 shows the urinary mercury levels according to the characteristics of the study participants.The geometric mean of the total urinary mercury concentration was 0.401 µg/L in the 3rd cycle and 0.399 µg/L in the 4th cycle, and there was no significant difference by sex (male vs. female: 0.40 µg/L vs. 0.41 µg/L in the 3rd cycle, 0.41 µg/L vs. 0.39 µg/L in the 4th cycle).In both cycles, mercury concentrations were significantly higher when urinary cotinine and urinary creatinine concentrations were higher (p < 0.05).In addition, participants with atopic dermatitis symptoms had higher mercury concentrations than those without atopic dermatitis.There were directional differences in asthma and allergic rhinitis according to the cycle, but they were insignificant.
Table 2 shows the association between urinary mercury concentration and allergic symptoms.As a result of multivariable logistic regression analyses, after adjusting for creatinine (Model 1), urinary mercury showed a positive correlation between atopic dermatitis in cycle 3 (OR [95% CI] = 1.36 [1.02, 1.81]) and cycle 4 OR ([95% CI] = 1.21 [0.64, 2.29]).After adjusting for all covariates in the 3rd and 4th cycles (Model 3), urinary mercury positively correlated with atopic dermatitis, asthma, and multimorbidity.Similar results were observed in pooled Table 1.Demographic and clinical characteristics of study participants.AM arithmetic mean, BMI body mass index, GM geometric mean, GSD geometric standard deviation, SD standard deviation. 1BMI was categorized into three groups: underweight or normal, which was defined as < 85th percentile; overweight, which was 85th-95th percentile; and obese, which was ≥ 95th percentile 22 .www.nature.com/scientificreports/effects using two-cycle data, and significant positive correlations were found for atopic dermatitis and multimorbidity (Fig. 2).Table 3 shows the results of subgroup analyses to determine the relationship between urinary mercury concentration and allergic disease symptoms according to sex.For atopic dermatitis and multimorbidity, the pooled effects of the 3rd and 4th cycles showed a greater association in males than in female participants.The positive association between multimorbidity and mercury was significant only in boys.

Discussion
In this study, we investigated the association between urinary mercury concentrations and allergic diseases in children aged 6-11 years who participated in the KoNEHS cycles 3 and 4.After adjusting for all covariates, the urinary mercury level was positively associated with atopic dermatitis in 3rd cycle and with allergic rhinitis in 4th cycle.When we calculated the pooled effects using data from cycles 3 and 4, the OR for atopic dermatitis was www.nature.com/scientificreports/1.34 (95% CI = 1.01, 1.79), and the OR for allergic multimorbidity was 1.47 (95% CI = 1.01, 2.12).Additionally, the association with allergic multimorbidity was greater in boys than in girls.
The results of this study are consistent with the findings of previous studies on the effects of mercury exposure on allergic diseases.In studies from Korea, serum mercury levels were reported to have a positive correlation with the risk of atopic dermatitis in adults and infants 7,12 , and a significant positive correlation was observed Table 2. Association of urinary mercury level with current allergic symptoms.CI confidence interval, OR odds ratio. 1 Multimorbidity was defined as the presence of two or more symptoms, and the statistics were calculated by comparing this group and the group with no symptoms.Model 1 was adjusted for urinary creatinine level.Model 2 was adjusted for all covariates in Model 1 and further adjusted for sex and age.Model 3 was adjusted for all covariates in Model 2 and further adjusted for household income, parental education, BMI, and urinary cotinine level.www.nature.com/scientificreports/with asthma in a longitudinal study of children 11 .However, a cross-sectional study in Germany and the United States reported no association with asthma 6,13 , which may be explained by the higher average body mercury levels in Korea compared to those in Germany and the United States 24 .The average mercury levels in children in this study were 0.40 µg/L for the 3rd and 4th cycles, which were higher than those in the United States NHANES (2015-2016) (0.25 µg/L for 3-11-year-olds) and Germany GerES V (2014-2017) (0.068 µg/L for 5-9-year-olds) in similar periods and age groups.Previous studies have attributed this difference in mercury levels to Korean eating habits, which include frequent consumption of grains and seafood 25,26 and exposure to air pollution 27 .For these reasons, serum mercury levels are also high at 2.75 µg/L in adults based on the 3rd cycle of KoNEHS, which is similar to or lower than those in Japan and Taiwan, where similar diets are shared 28,29 .
Conversely, this study did not observe a consistent association with allergic rhinitis.Unlike atopic dermatitis and asthma, few studies have observed an association of allergic rhinitis with mercury, and no association was found in a study that examined overall allergic diseases 20 .Allergic rhinitis is classified into seasonal allergic rhinitis, affected by outdoor allergens such as pollen and perennial allergic rhinitis, affected by indoor allergens such as mold and animal hair.Allergic rhinitis is known to be affected by various environmental factors such as season, region, and lifestyle 30,31 .Moreover, in previous studies investigating the association between harmful environmental factors and allergic rhinitis, a significant negative association was observed for substances such as bisphenol S, vanadium, and 3,5,6-trichloro-2-pyridinol [32][33][34] .Considering these findings, the lack of consistent associations observed in this study may be due to differences in variables of the study environment that were not examined.
Various prior studies have reported on the biological mechanisms of low-level mercury exposure and changes in the immune system.Mercury exposure induces the proliferation of T and B lymphocytes by increasing the expression of major histocompatibility complex (MHC) class II molecules on antigen-presenting cells, and it increases serum immunoglobulin (Ig) G1 and IgE associated with allergic sensitization 9,35 .Mercury can also increase the levels of inflammatory cytokines such as interleukin-4 (IL-4), interleukin-13 (IL-13), and tumor necrosis factor-α (TNF-α) by stimulating T helper 2 (Th2) cells and worsen IgE-dependent diseases by interfering with the regulation of the immune system 8,36,37 .Sensitized Th2 cells can trigger allergic diseases when moving to the respiratory tract and various organs, and the atopic march may occur in which atopic dermatitis, asthma, and allergic rhinitis progress in sequence 38 .In this study, a significant positive association between mercury and multimorbidity was observed, and our results are supported by the fact that the association between each allergic disease has been frequently reported in several prior studies 39,40 .
Furthermore, this study showed a stronger association between mercury exposure and allergic diseases in boys, which is consistent with the findings of previous studies 11,41 .Although the mechanism by which allergic diseases have different effects by sex is not clearly understood, it is generally known that allergic diseases are more common in boys than in girls during infancy but more common in women after puberty 42 .This is thought to be because boys have a stronger allergic sensitization response than girls at a young age 43 .Moreover, boys are known to have slower lung function development than girls, so they are relatively vulnerable to the response of allergens 44 .After puberty, estrogen stimulates mast cells and eosinophils to promote allergic reactions, whereas androgens inhibit the differentiation of some T cell subsets, resulting in immunosuppressive effects [45][46][47] .Therefore, mercury exposure in boys appears to increase their susceptibility to allergic diseases, but follow-up studies are needed to confirm the results of this study.
This study has several limitations.First, since this study is a cross-sectional study, it is difficult to prove the causality of the result that urinary mercury concentration in children affects allergic disorders.Therefore, followup studies targeting children are necessary.Second, the evaluation of parameters such as allergic symptoms based on the parents' questionnaire responses may not reflect the actual disease status well.However, their responses are based on the diagnosis from experienced doctors, and information on allergic disorders of children based on parental surveys is relatively accurate 48 .In addition, information bias was minimized in this study by determining Table 3. Association of urinary mercury level with current allergic symptoms according to sex.CI confidence interval, OR odds ratio. 1 Statistics could not be calculated for empty cells because only rare cases of asthma were found in this sex subgroup. 2Multimorbidity was defined as the presence of two or more symptoms, and the statistics were calculated by comparing this group and the group with no symptoms.All models were adjusted for urinary creatinine level, age, sex, urinary cotinine level, household income, parental education, and body mass index.www.nature.com/scientificreports/ the presence of allergic disorders based on medication use and treatment.Third, due to the inherent limitations of using secondary data, data on long-term indicators such as hair mercury concentration and information about parental allergy history were not included in the analysis.However, the urinary mercury concentration is an indicator of chronic exposure to mercury 13 and has been widely used as an indicator that can easily be obtained from a large number of samples 49 .In addition, even if some confounding variables were not considered, this study was based on KoNEHS, a large national bio-monitoring survey of Korean children.Therefore, our study provides sufficient statistical power to identify associations that may be less evident and thus provides reliable epidemiological evidence that mercury exposure in Korean children is associated with allergic diseases.

Figure 2 .
Figure 2. Pooled effects with data from the Korean National Environmental Health Survey (KoNEHS) cycle 3 and cycle 4.All models were adjusted for urinary creatinine level, age, sex, urinary cotinine level, household income, parental education, and body mass index.CI confidence interval, OR odds ratio.