AUSTRI 2016

Methods	A randomised, double‐blind, multi‐centre, parallel‐group study carried out over 26 weeks
Participants	Population: 11,679 adults and older children (> 12 years) with persistent asthma; all participants had a history of severe asthma exacerbation in the past year but no events in the past month Baseline characteristics: mean age 43.4 years Inclusion criteria: main further inclusion criteria were the following: peak expiratory flow ≥ 50% of predicted normal value, use of daily medication for asthma control and ≥ 1 asthma exacerbation requiring treatment or hospitalisation within 12 months before randomisation Exclusion criteria: main exclusion criteria were the following: history of life‐threatening asthma, concurrent respiratory disease, > 10‐pack‐year smoking history, respiratory infection, unstable asthma status
Interventions	• Fluticasone propionate and salmeterol (at a dose of 100 μg of fluticasone and 50 μg of salmeterol, 250 μg and 50 μg, or 500 μg and 50 μg, respectively) twice daily • Fluticasone propionate (at a dose of 100 μg, 250 μg, or 500 μg) twice daily
Outcomes	Primary outcome: time until first serious asthma‐related adverse outcome
Notes	Sponsored by GSK
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was performed with the use of an interactive voice‐response system with stratification
Allocation concealment (selection bias)	Low risk	Randomisation was performed with the use of an interactive voice‐response system with stratification
Blinding (performance bias and detection bias) All outcomes	Low risk	Double‐blind with respect to salmeterol but not dose of fluticasone
Independent Assessment of causation (detection bias) Asthma‐related events	Low risk	Independent assessment of safety outcomes
Incomplete outcome data (attrition bias) All outcomes	Low risk	For mortality, the status of all participants who took at least 1 dose of treatment was assessed after 6 months
Selective reporting (reporting bias)	Low risk	Data were extracted for all outcomes of the review