Baseline characteristics
The Charlson Comorbidity Index (22) was used as a marker for the overall comorbidity burden based on International Classification of Diseases (ICD)-10 codes diagnoses recorded from in- or outpatient hospital contacts (23) with exclusion of asthma diagnoses (ICD10 J45-J46). Specific asthma-related comorbidities were chosen from the existing literature (5, 24-26) and estimated by presence of hospital-given diagnoses or use of relevant medication dispensed from public pharmacies in Denmark. Additional details on baseline comorbidities including specification of applied ICD-10 codes and Anatomical Therapeutic Chemical (ATC)-codes are available in Online Supplement Table E2 .
Statistical analyses
Descriptive statistics were presented as number, median and interquartile range (IQR) for continuous variables, and as frequencies and percentages for categorical variables. The annual period prevalence proportion of OCS users was defined as the number of individuals filling at least one OCS prescription per calendar year per 100 individuals in the study population, stratified according to OCS-usage groups (high-use and low-use), and by sex and age categories (18-25 years, 26-35 years, 36-45 years). Lorenz curves (27) for the years 1999, 2009 and 2018 were computed to assess the skewness of OCS consumptions among the prevalent OCS users, ranking all users in order by the amount of consumed OCS. The Gini coefficient, where 0 indicates total equality in consumption among users and 1 indicates total inequality, was calculated as a single measure of skewness in consumption of OCS among the users.
Stata Version 16 (StataCorp, College Station, TX, USA) was used in the analyses.