This study was conducted using the STROBE checklist.
Study design and participants
In this prospective cohort study we used data from Japan Environment and Children’s Study. The JECS is a study of the environment that can affect the development and health in Japan’s children.28. The details of the JECS are described previously29. All participants were recruited between January 2011 – March 2014. The “jecsan-20180131” dataset was used in the present study. It was released in March 2018. We excluded stillbirths, abortions, missing values and multiple birth patients.
Data collection
From a questionnaire completed by the parents, data regarding the first year of OM history, breastfeeding during the first 6 month, daycare, pneumococcal and Hib vaccination were collected. The data regarding live births, stillbirths, singleton and multiple births, sex and gestational weeks, maternal age, birth weight, and whether the child was diagnosed as having trisomy 21 were extracted from medical records by doctors, midwives/nurses, research coordinators, and other health care professionals.
C1y questionnaire gathered the primary outcome. This asked caregivers to tell us when their children were 1 years old. Have you been diagnosed by a doctor with OM? The outcome variable was binary: no diagnosis of OM as 0 or diagnosis with OM at 1. In this questionnaire, we did not distinguish between acute and chronic otitis media. This questionnaire did not include data on how OM was treated.
Exposure factors included maternal occupational use of insecticide for more than half an hour from conception to the first or second trimesters. The question was “What was the frequency of occupational insecticide use for more than half of a day during pregnancy?” This information was derived from the questionnaire for mother M-T1 questionnaire that was filled in the first trimester. The M-T2 questionnaire was filled in in the second/third.30. These variables were categorized as follows: 1 to 13, 13 times per months as 2, 2 to 2, 16 times per week and every day as 3.
We chose the covariates below as they were known risk factors for OM as discussed in the introduction. Although there was no evidence of OM in maternal infertility treatment, primary ciliary dysfunction, which can cause both chronic OM as well as infertility, was reported to be an inherited condition.31. We therefore selected maternal infertility treatments as a covariate. Other covariates include sex, gestational hours, birth weight and living with older sibling at 6 months. We also included smoking around children. These included maternal and paternal smoking habits as well as maternal exposure to secondhand smoke and household smoking at the young age of one month. We focused on maternal occupation, particularly farmers. However, there were very few farmers. Instead, we used major occupational classes. We counted the number of mothers in each major maternal occupational category.
The binomial variables were as follows: sex of child: male as 1, females as 2, living with siblings at 6 month, daycare use at a one year old, maternal history, chronic OM, history, Hib vaccine and pneumococcal vaccine, children with trisomy 21 history of fertility treatment: No, yes, 1.
We also classified categorical variable as: gestational week32,3337 to 41 weeks as 1,37 weeks as 2,42weeks, as 3, and 2.5 -4kgs as 2,4kgs. As 4. Maternal age: 3034 years as 1, 2529years, as 2, and 35years respectively. Household smoking habits at the age 1 month: No one as 1. Someone smoking in a faraway place as 2, or smoking at a nearby baby’s location as 1. Maternal occupational groups: Full-time homemaker as 1, professional, technicians as 2, clerical and support workers as 3, and service workers at 4. We used the minimum category number to control all variables.
Statistic analysis
Stata version 15 software was used to perform all statistical analyses (StataCorp College Station, TX USA). First, we compared patient characteristics among the OM and non-OM histories groups for the first year of children. Second, we used logistic regression analysis for logistic regression analysis to determine the relationship between OM in the first year and maternal insecticide usage. This was done from conception to the first trimester, and again from the second/third trimesters. We also performed sensitivity analysis using the multiple imputation method because there were many missing values. We also performed sub-group analysis to determine the relationship between OM and maternal insecticide use from conception to the first trimester and again from the second/third trimesters. This was done using logistic regression analysis in three subgroups. These groups included whether children were born to a sibling, whether they attended daycare as children turned one, and whether their mother had a history chronically OM. Significance was defined to be p0.05.
Ethical considerations
The Ministry of the Environments Institutional Review Board on Epidemiological Studies and the Ethics Committees of all participating organizations reviewed and approved the JECS Protocol. We confirmed that all research was performed in accordance with relevant guidelines/regulations. Before participating in the current study, all participants gave written informed consent.
