Statistical analysis
Longitudinal data analysis was used to assess the change in PFTs over time. Quantitative data is expressed as median and interquartile range, whereas qualitative data is described by frequency (composition ratio). Generalized linear mixed models were assessed with age at PFTs as the fixed effects, while random effects were specified at the level of the individual. All decreases and increases in PFT parameters were described as changes in percentage of predicted for height. At the initial PFTs, we used a linear mixed model to determine whether bronchodilator responses were affected by patient age at the time of diagnosis, or by allergy factors (including a history of wheezing, atopic dermatitis and family asthma history). All data were analyzed using R software (v3.5.3, Auckland, New Zealand). The p value<0.05 was considered to be statistically significant.