Methods

Participants

This is a sub-cohort study of the Japan Environment and Children’s Study (JECS). The study design and recruitment methods have been described in detail previously.15,16 In brief, JECS is a large-scale birth cohort study involving 100,000 pairs of parents and children in 15 nationwide regions in Japan, wherein participants were observed from the foetal period to the age of 13 years old. Participants were recruited between January 2011 and March 2014; 103,060 pregnancies were registered, of which 100,304 were live births. For the sub-cohort study, 5,017 participants were randomly selected from those who met the eligibility criteria and consented to participate in the sub-cohort study.17 Data for this study were used from the dataset released in October 2019 (dataset jecs-ta-20190930). A schematic summarising the sample selection methodology is shown in Figure 1, and detailed study methodology is provided in Supporting Information. The JECS protocol was approved by the Institutional Review Board on Epidemiological Studies of the Ministry of the Environment (IRB number: 100406001) and by the Ethics Committees of all participating institutions. Written informed consent was obtained from all participants.

Questionnaire

Questionnaires were sent to the participants by post at the age of 12, 18, 24, and 36 months old, and the parents filled them out and sent them back. Those questionnaires were a modified version of ISAAC for children aged 6–7 years, with translation validation in Japanese.18-20 Okabe et al. investigated the validity of asthma symptoms based on ISAAC questionnaires involving 0-60-month-old infants by using asthma diagnosis made by allergists as criteria,21 wherein the sensitivity and specificity of asthma symptoms were 0.91 and 0.64 in 0-24-month-old children and 0.88 and 0.68 in 36-60-month-old children, respectively. In our study, asthma and wheezing were defined according to the definition by Okabe et al..21,22 The incidence rates of wheezing and asthma by age (0-12, 12-24, and 24-36 months old) were defined as a ratio of participants initially judged as having wheezing/asthma at each age compared to the that of all participants. Information on lower respiratory tract inflammation and respiratory syncytial virus (RSV) infection was obtained from the questionnaires at the age of 12, 18, and 24 months. Information concerning the presence of indoor pets and passive smoke after birth was obtained from the questionnaires at 18 months old.

Blood and urinal examination of mothers

Blood samples were obtained from mothers during the first trimester of pregnancy. Total serum IgE titres, assayed using ImmunoCAP (Thermo Fisher Scientific, Inc., Sweden), were analysed by an independent clinical laboratory. Mothers with high levels of total serum IgE were defined as those with a total IgE level of 214 IU/mL or above.23Maternal spot urine was collected at mid-pregnancy, and the urinary cotinine level was measured. Sample collection and measuring methods were described in detail previously.24 The cut-off values for active smoking and passive smoking during pregnancy were set at 36.8 and 0.31 μg/g creatinine, respectively.24

Blood examination of children

The mothers and children visited the test site when their children were 24 months old; blood samples were obtained from children. Total serum IgE titres, assayed using ImmunoCAP (Thermo Fisher Scientific, Inc., Sweden), were analysed by an independent clinical laboratory. As the cut-off value of high levels of total serum IgE in 24-month-old children is unclear, it was set at the upper tenth percentile of the total IgE level.

Exposure assessment (mite allergen/endotoxin)

Trained investigators visited homes when the children were 18 months old and collected house dust from mattresses on which the children slept. A 50-cm × 100-cm frame exclusive for sampling was placed on the mattress, and a vacuum cleaner (Model DC61, Dyson, Japan) was run over the area in the frame for 2 minutes (Supp Info Figure 1). The collected dust was frozen until analysis.Dermatophagoides pteronyssinus (Der p 1) andDermatophagoides farinae (Der f 1) in the dust were measured using an enzyme-linked immunosorbent assay kit (Indoor Biotechnologies Ltd, UK). Endotoxin levels in the dust were measured by kinetic chromogenic limulus amebocyte lysate assays (Kinetic-QCL), obtained from Lonza Japan (Tokyo). There is high structural similarity in the corresponding allergen between Der p 1 and Der f 1. Once sensitised, there is strong cross-reactivity even when sensitised with either type.25 Thus, we used the total exposure to both mite species Der 1 (Der p 1 + Der f 1) in the analysis of the association with health outcomes.

Statistical analysis

Dust mite allergen and endotoxin concentrations were expressed as ‘amount per m2 of dust sampling area’ and ‘amount per unit mass of dust’,26 and Spearman’s rank correlation coefficient was calculated for these exposure indices. The correlation coefficient for the ‘amount per gram of dust’ and ‘amount per m2 dust sampling area’ of Der 1 was 0.86, and that in endotoxin between the ‘amount per mg dust’ and ‘amount per m2 dust sampling area’ was 0.79, demonstrating a high correlation between these exposure indices. Thus, the ‘amount per m2 dust sampling area’ was analysed in the main analysis thereafter. Analysis of the ‘amount per unit mass of dust’ was performed as a sensitivity analysis. If the data were lower than those of the method detection limit, then a value of one-half was assigned. In the analysis of the association with health outcomes, the participants were classified into the first-to-fourth quartile (Q1-Q4) groups. Regarding the association between the dust mite allergen or endotoxin exposure level and health outcome, the odds ratios (ORs) of the exposure level in the Q2-Q4 groups in comparison with the Q1 group were estimated using a logistic regression model. Regarding health outcomes, the incidence rates of asthma and wheezing by age (0-12, 12-24, 24-36, and 0-36 months old) were selected for the main analysis. Secondary analysis was set to high levels of total serum IgE in children. High levels of maternal total serum IgE, active smoking/passive smoking based on the urinary cotinine level in mid-pregnancy, passive smoke after birth, children’s sex, lower respiratory tract inflammation, RSV infection, presence of indoor pets, and climatic region were regarded as covariates. Climatic region was classified into three areas based on mean temperature and mean humidity (Supp Info Figure 2). A trend test was then performed to confirm the quantitative relationship of health outcomes with the exposure level. Regarding continuous variable x giving 1, 2, 3, and 4 points to the Q1-Q4 groups as an explanatory variable of the statistical model, the significance level, p-value, of the OR of increase per unit x was calculated. All statistical analyses were performed using SAS ver9.4 (SAS Institute Inc., Cary, NC, USA).