Study questionnaire
The questionnaire involves sections asking for information regarding the child as well as its parents and grandparents. Regarding the child, the questionnaire captures the following information: birth of the child (date of birth, birth weight, type of delivery), medical history in the first year of life and during follow-up in the subsequent years of life (asthma symptoms, diagnosis of airway diseases or atopic dermatitis, intake of medication), breastfeeding (duration of breastfeeding, type of milk used as supplements), and living conditions (number of persons in the household, whereabouts of the child during the day). In each family, the questionnaire is only completed for one child. Those participants who already have more than one child are asked to complete the questionnaire for the youngest one.
The maternal section of the questionnaire asks for the mother’s general health (asthma diagnosis and asthma symptoms, allergies, atopic dermatitis), medical history during pregnancy (infections, gestational diabetes, intake of medication), and smoking. The paternal section is addressed to those fathers who had not participated in SOLAR II (i.e., the mother of the child is a member of the original cohort) and collects information on symptoms or diagnosis of asthma, bronchitis, hay fever, or atopic dermatitis, as well as intake of medication and smoking.
In addition, demographic data (age, education, occupational status, number of children) are recorded for that parent who did not take part in SOLAR II. Moreover, life-time prevalence of asthma, bronchitis, hay fever, and atopic dermatitis among the grandparents of the children is assessed by the questionnaire.
When the children are in primary school age (six to nine years of age), the parents will be asked to complete the questionnaire again. By then, it will be possible to perform a first reliable diagnosis of asthma and atopic dermatitis.
All sections are based on validated questionnaires that have been successfully used in previous studies such as the German Multicentre Allergy Study [
56] and the PASTURE project (Protection against Allergy - STUdy in Rural Environments) [
57].
Clinical examination
In addition to the questionnaire data, we aim to invite the children and their parents to a clinical examination when the children are in primary school age. This will involve lung function test and blood sampling. Lung function measurements will be conducted in accordance with the guidelines of the American Thoracic Society [
58]. Blood samples will be used for epigenetic analyses and for measurement of total immunoglobulin E (IgE) levels as a test for allergic sensitisation. For analysis of IgE levels, we will use the same procedure as in SOLAR II [
42].
Regarding epigenetics, peripheral CD-4+ T cells will be isolated using magnetic cell sorting and isolation kits. DNA will be stored frozen at −80 °C and will be isolated with the FlexiGene DNA Kit (Qiagen NV, Venlo, Netherlands). For array-based DNA methylation analysis, 800 ng of genomic DNA will be treated with sodium bisulfite using the EZ-96 DNA Methylation Kit (Zymo Research Corporation, Irvine, USA). The Infinium HumanMethylation450 BeadChip Kit (Illumina Inc., San Diego, USA) will then be used for measurement of genome-wide DNA methylation. After genome-wide amplification of 4 μl of bisulfite-treated DNA and enzymatic fragmentation, the samples will be added to the chip. The fraction or percent methylation will be reported as Beta-value. This value is expressed as a continuous variable between 0 and 1 corresponding to the ratio of methylated molecules to the sum of methylated and unmethylated molecules [
59].
Data management
Data will be stored and handled at the Institute and Outpatient Clinic for Occupational, Social and Environmental Medicine, University Hospital of Munich (LMU), Munich, Germany. Data from the paper version of the questionnaire will be entered into a password-protected MS Access database (Microsoft Corporation, Redmond, USA). Double data entry will be used to detect and reduce errors. For further analysis, these data will be combined with the questionnaire data that participants directly entered into the SurveyMonkey database. To make full use of all the data collected since 1995, the ACROSSOLAR dataset will also be merged with the datasets of the previous studies (ISAAC II and SOLAR I and II) using a unique study ID.
Statistical analysis
First, descriptive statistics will outline the characteristics of the study sample. Bivariate analyses will then evaluate potential associations between the exposure and outcome variables that are included in the study questionnaire. Next, statistical methods for analysis of longitudinal data such as Poisson regression and logistic regression models will be used for further examination of possible associations [
60,
61]. Potential confounders will be included in the model if they change the effect estimates for the association between the exposure and outcome variables by at least ten percent (change-in-estimate criterion) [
62]. To address the problem of missing data, we will run multiple imputations using the R-package Amelia II (The R Foundation for Statistical Computing, Vienna, Austria) which performs expectation-maximization with bootstrapping algorithms [
63‐
65].
Regarding epigenetic analyses, CpG sites that are within 50 bp to known SNPs or in allosomal positions will be removed. For each chip, colour bias adjustment using quantile normalisation and background correction based on a negative control will be performed using the R-package lumi [
66]. Further analyses will be carried out with a random intercepts mixed model approach.