Association between humidifier disinfectant exposure during infancy and subsequent neuropsychiatric outcomes during childhood: a nation-wide cross-sectional study

Background The purpose was to determine the association between infant exposure to humidifier disinfectant (HD) with neuropsychiatric problems in pre-school children. Methods A total of 2,150 children (age 4–11 months) were enrolled in the Panel Study of Korean Children (PSKC) study. The Korean version of the Child Behavior Checklist (CBCL) was used for assessments of neuropsychiatric problems. 1,113 children who participated in all the first to third PSKC studies and answered a question about HD exposure were finally enrolled. Results There were 717 (64.5%) children in non-HD group who were not exposed to HD and 396 (35.5%) in HD group with former exposure to HD. Exposure to HD was associated with total neuropsychiatric problems (adjusted odds ratio, aOR = 1.54, 95% CI = 1.15–2.06), being emotionally reactive (aOR = 1.55, 95% CI = 1.00–2.39), having attention problems (aOR = 1.96, 95% CI = 1.10–3.47), having oppositional defiant problems (aOR = 1.70, 95% CI = 1.07–2.71), and having attention deficit/hyperactivity problems (aOR = 11.57, 95% CI = 1.03–2.38). The risks for neuropsychiatric problems were clearly increased in boy, firstborn, and secondary smoker. Conclusions Exposure to HD during early childhood had a potential association with subsequent behavioral abnormalities. Supplementary Information The online version contains supplementary material available at 10.1186/s12887-021-02825-7.


Background
Chemical components to which the respiratory organs are sensitive can have systematic effects. Pre/postnatal exposure of chemicals may contribute to the development of several diseases, including allergic disease [1][2][3], eczema [4], low birth weight [5], preterm birth [6], and neurodevelopmental disorders [7,8]. In particular, previous studies showed that pollutants can affect the brain via neuroinflammatory pathways, through systemic or direct nose-to-brain routes, and lead to depression-like phenotypes in animal model [9]. In addition, there is evidence that prolonged exposures to these chemicals could lead to behavioral and social problems [8]. Patients with chronic pulmonary disorders who are exposed to chemical irritants may develop cognitive impairment, such as deficits in attention, concentration, and memory, and these are presumably mediated by hypoxemia and hypercapnia [10].
Exposure to humidifier disinfectant (HD) in South Koreathe 'humidifier disinfectant scandal'caused multiple social and health-related problems that are still ongoing [11][12][13]. Nation-wide epidemiologic and animal studies confirmed that the outbreak of interstitial lung disease (ILD) in children and adults was caused by HD exposure. The high mortality rate in infants (18-27%) led to discontinuation of HD sales [14]. The long-term prognoses of individuals who were exposed as preschool children are still uncertain. In addition, most of the previous studies have focused on the relationship between HD exposure and lung injury in adults and children [15][16][17]. However, we postulated that the extent of the toxic effects of HD would not be confined to the respiratory system. (we believe it is unlikely that the toxic effects of HD are limited to the respiratory system.) Although the association between various chemicals including air pollutants and neurodevelopment disease have been well known, the effect of HD is still limited. In a previous study of 135 individuals who were exposed to HD, 82 adults and 19 children had high risk for depression (self-evaluation), anxiety, anger, and posttraumatic stress disorder (PTSD) [18].
HD exposure is likely to lead to problems involving other organ systems, such as neuropsychiatric abnormalities, given that the nano-size of the toxic chemicals in HD can irritate cell membranes and migrate into the systemic circulation [19]. Because brain development is one of the most important processes during early infant development, we hypothesized that infants exposed during this critical time window would subsequently develop neurodevelopmental disabilities. To further investigate this matter, we examined children and families in the PSKC cohort which has data on HD exposure, multiple clinical and demographic factors, and developmental assessments. Therefore, we hypothesized that HD exposure during early childhood is associated with behavioral abnormalities, and that these increase the risk for neuropsychiatric diseases such as attention deficit hyperactivity disorder (ADHD) and developmental problems.

Study population
The Panel Study on Korean Children (PSKC) is a national representative, general population-based birth cohort participated children who born in 2008 and their parents. It collects and provides cross-sectional and longitudinal data at a national level [11,20] . The survey was explained to the mothers and went through the process of obtaining consent to participate in the survey. Through a stratified multi-stage sampling method, 2,150 newborns born between April and July 2008 and their parents were finally enrolled from 30 sampled hospitals with more than annual 500 deliveries across the country. The exclusion criteria were: (i) mother less than 18 years-old at birth; (ii) parents unable to communicate in the Korean language; (iii) critically ill mother postpartum; (iv) baby with severe illness; (v) mother with severe illness; (vi) baby planned for adoption; (vii) baby from a multi-fetal pregnancy; (viii) loss to follow-up; and (ix) non-responders to the questionnaire or failure to provide complete answers. Parents completed a selfanswered questionnaire that asked for a history of pediatric diseases, such as physician-diagnosed asthma, allergic rhinitis, and atopic dermatitis, and sociodemographic information including birth order, household income, parents' educational level, alcohol drinking history, and smoking history [21]. In addition, parents and child performed an investigator-administrated survey regarding their child's behavior every year from 2012 to 2014. Of 2,150 children who were enrolled, 1,113 children were included whose responses to all 3 CBCL questionnaires were complete and whose information on HD exposure was available (Fig. 1).

Ethical consideration
Informed consent was obtained from participants' parents or a legal guardian. All methods were carried out in accordance with relevant guidelines and regulations. The survey was approved by the IRB of the Korea Institute of Child Care and Education.

Behavioral/neuropsychiatric assessment
Behavioral and emotional problems of pre-school children were evaluated in the psychometric properties section of the Korean version of the Child Behavior Checklist for children aged 1.5-5 years (CBCL 1.5-5) from 2012 to 2014 [22,23]. The CBCL 1.5-5 test consists of 8 subscales: emotionally reactive, anxious/ depressed, somatic complaints, withdrawn, attention problems, aggressive behavior, sleep problems, and other problems. In addition, Diagnostic and Statistical Manual (DSM)-oriented scales were used to assess 5 scales: affective problems, anxiety problems, pervasive developmental problems, attention deficit/hyperactivity problems, and oppositional defiant problems.
The sum of the scores for individual items was converted to a T-score, in which a higher score indicated more severe problems. Internalizing problem was the sums of 4 subscales: emotionally reactive, anxious/depressed, somatic complaints, and withdrawn, and externalizing problem was the sums of 2 subscales: attention problems and aggressive behavior. The total problem was the sum of all subscales. The cut off score for internalizing, externalizing, and total problem was 60 and the cut off score for the other scales was 65 [24].

HD exposure
The status of HD exposure was estimated using a questionnaire in 2015. In this questionnaire, participants were asked whether they used a humidifier disinfectant, and the total duration of HD use. The period of use was classified as more than or less than 3 months [15].

Assessment of innate neurodevelopmental risk in newborns
The Korean version of the Ages and Stages Questionnaires (K-ASQ) was used to evaluate the baseline neurodevelopmental status of infants before HD exposure [25,26], which has good performance in screening children with biological and environmental risk factors [27,28]. This questionnaire was performed at the participants' 4 months of age.

Assessment of maternal depression
To rule out poor parenting from maternal post-partum depression as the cause of a child's poor mental health [29], the Kessler Depression scale (K6) was administered at delivery and the results were considered as a potential confounder in the multivariable analysis [30,31]. We defined a total score of 6 to 13 points as 'normal', 14 to 18 points as 'mild/moderate depression', and 19 to 30 points 'severe depression'.

Statistical analysis
The chi-square test was used to determine the significance of differences between the HD and non-HD groups in baseline demographic factors, pre-exposure risk assessments and the number of children reporting problems. Associations of HD exposure with behavior problems from age 3 to 5 years were evaluated as Odds ratios (ORs) and adjusted odds ratios (aORs) using generalized estimating equation (GEE) models with an unstructured correlation matrix. The potential confounding factors considered in this analysis were: gender, birth weight, BMI z score each year, maternal and paternal age at initial assessment, household income, maternal and paternal level of education, parental allergic history, depression of mother at delivery, and child's innate neurodevelopmental risks based on the K-ASQ [32]. All statistical analyses were performed using SPSS version 26.0 (IBM, Armonk, NY, USA), and P value less than 0.05 was considered statistically significant.

Demographic characteristics of the study population
We initially examined the baseline characteristics of 1,113 children from the PSKC study (Table 1). There were 551 girls (49.5%) and 562 boys (50.5%), and 396 children (35.6%) who were exposed to HD and 717 (64.4%) who were not exposed. Mothers of the non-HD group were approximately 7 months older than those of HD group (31.4 vs. 30.8 years; P = 0.046), but the two groups had no significant differences any other in baseline characteristics.

Assessments of innate neurodevelopmental risk in newborns and maternal depression
We assessed the presence of innate neurodevelopmental risk in the children and maternal depression soon after delivery (Supplementary Table 1). The HD and non-HD groups had no significant differences of the innate neurodevelopmental status in all five categories of the K-ASQ (communication, fine motor skills, gross motor skills, personal-social skills, and problem solving). In addition, there were no significant differences in maternal depression between two groups.
Behavioral/neuropsychometric properties of the HD and non-HD groups Table 2 shows the results of psychometric properties reported in first, second, and third CBCL test. In the first CBCL test, HD group had more children having attention problems than non-HD group. In the second CBCL test, children having total problems, emotionally reactive, withdrawal, sleep problems, affective problems, pervasive developmental problems, and attention deficit/ hyperactivity problems were more in HD group than non-HD group. In addition, in the third CBCL test, several neuropsychiatric problems were prominent in HD group, including emotionally reactive problems, anxious/ depressed problems, all externalizing problems, sleep problems, other problems, affective problems, and oppositional defiant problems. Table 3 showed the association of HD exposure with the CLBL and the DSM-oriented scales. The crude analysis showed that the HD group had statistical significant ORs for total problems, internalizing problems, being emotionally reactive, externalizing problems, attention problems, sleep problems, other problems, affective problems, and attention deficit/hyperactivity problems. After adjustment for confounders, the HD group had statistical significant aORs for total problems (aOR = 1.538, 95% CI = 1.148-2.059; P = 0.004), being emotionally reactive (aOR = 1.548, 95% CI = 1.004-2.387; P = 0.048), attention problems (aOR = 1.957, 95% CI = 1.104-3.469; P = 0.022), sleep problems (aOR = 1.637, 95% CI = 1.099-2.438; P = 0.015), oppositional defiant problems (aOR = 1.699, 95% CI = 1.065-2.709; P = 0.026) and attention deficit/hyperactivity problems (aOR = 1.566, 95% CI = 1.029-2.384; P = 0.036).

Subgroup analyses on association of HD exposure with psychometric outcomes
We examined the effect on neuropsychiatric abnormalities according to duration of HD exposure, gender, and the status of paternal smoking. First, Both durations (< 3 months and ≥ 3months) had positive association with total problems (aOR = 1.492; P = 0.049, aOR = 1.606; P = 0.044, respectively). The risk for being anxious/depressed problems (aOR = 1.953; P = 0.030), attention problems (aOR = 3.045; P = 0.004), and oppositional defiant problems (aOR = 2.178; P = 0.028) were statistically significant in long-term HD exposure group, but not in the short-term HD exposure group (Supplementary  Table 2).
Third, secondary smokers in HD group had statistically significant associations with total problem (OR = 1.622; P  Table 4).

Discussion
This is the nationwide cohort study to assess the association between exposure to HD at early childhood and neuropsychiatric or behavioral problems in preschool children. We showed that HD exposure was a potential risk for several categories of neuropsychiatric or behavioral problems in children, specifically being emotionally reactive and attention disorders. Moreover, the significance of the specific associations remained even after adjustment for confounding by gender, birth weight, BMI, maternal and paternal age, socio-economic status (household income), maternal and paternal level of education, parental allergic history, history of peripartum depression, and child's innate risk of developmental process. We also showed that exposure to HD for ≥ 3 months, in boys, and secondary smoker were potential risk factors for the association of exposure to HD with several behavioral abnormalities and neuropsychiatric disorders. In recent years, animal studies have reported that HD exposure led to harmful systemic diseases, including nervous system disorders. A recent study of zebrafish reported that some major components of HD, including polyhexamethylene guanidine (PHMG), oligo (2-[2- ethoxy]ethoxyethyl) guanidine chloride (PGH), and chloromethylisothiazolinone/methylisothiazolinone (CMIT/MIT), increased the levels of oxidative intermediates that are toxic to the brain [33]. In particular, this study found that 40 mg/L of PHMG, PGH, and CMIT/ MIT increased the levels of ROS in the optic tectum, a major part of the midbrain of vertebrates that is involved in visual processing, sensorimotor integration, and behavioral motor patterns during brain development [34].
Other in vitro research reported that 16 mg/L of PHMG, PGH, and CMIT/MIT decreased the viability of human fibroblasts by 32%, 49%, and 41%, respectively. In addition, CMIT/MIT was reported to have toxicity in cultured neurons through zinc-dependent process, and this toxic effect required the activation of p44/42 extracellular signal-regulated kinase (ERK) via a 12lipoxygenase-mediated pathway [35]. These molecular events lead death of neurons and adverse neuropsychiatric outcome. This is one possible mechanism by which HD may adversely affect neurons in infants. Our study has some limitations. First, it was based on questionnaires, and therefore may be susceptible to recall bias regarding HD exposure. However, the short intervals between the CBCL tests and the social hot issue of lung disease from HD exposure might help their parents to remember whether to use HD and minimize this bias. In addition, the rate of HD use in our study was comparable to those reported in previous studies [11]. Second, many factors can affect the developmental process in children. Although our study adjusted for a broad range of confounding factors, other factors not considered in our analysis may have influenced the identified associations. For example, genetic predisposition may have an effect, but this was outside the scope of this study. Third, the alpha error could not be ruled out. A further analysis using Holm methods, including total problems, internalizing problems, externalizing problems, sleep problems, other problems, and DSMoriented scales, was carried out that total problem was statistically significantly associated with HD exposure (P = 0.040). However, the results need to be interpreted carefully. Lastly, the underlying mechanism that exposure to HD at early childhood is associated with certain neuropsychiatric problems is unclear. In addition, because we were not able to access the precise data about the types of the HD, the association with behavioral/neuropsychiatric problems according to the type of the HD could not be analyzed. However, major types of HD including PHMG, PGH, CMIT/MIT cause neuronal toxicity in animal models [34,35], so future animal and human studies are needed to examine the effects of these chemicals and confirm a causal relationship of HD exposure with neuropsychiatric problems. ORs and aORs were calculated using logistic regression using a generalized estimating equation model, with the non-HD group as the reference b Adjusted for the following confounding factors: sex, birth weight, BMI z-score each year, maternal and paternal ages, household income, maternal and paternal education levels, parental allergic history, mother's depression during delivery, and child's baseline neurodevelopmental properties prior to HD exposure (K-ASQ) P values less than 0.05 are in bold Our study showed that exposure to HD at early childhood is a potential risk for behavioral abnormalities in preschool age children. Thus, we suggest that clinicians consider the assessment of neuropsychiatric status of children with risk factors.

Conclusion
Exposure to HD at early childhood was the potential risk for subsequent behavioral abnormalities, and longer exposure, male, and secondary smoke were the factors to increased risk for certain neuropsychiatric outcomes. The use of HD in young age should be avoided in light of its association with behavioral abnormalities.