VESTRI 2016

Methods	A randomised, double‐blind, multi‐centre, parallel‐group study carried out over 26 weeks
Participants	Population: 6208 boys or girls aged 4 to 11 attending daycare or school, with persistent asthma and a history of asthma exacerbation in the previous 12 months, with no exacerbation occurring during the 30 days before randomisation Baseline characteristics: mean age was 7.6 years; 61.8% were male and 64.3% were white Inclusion criteria: consistent use of asthma medication during the 4 weeks before enrolment, and a history of 1 to 4 asthma exacerbations in the past 12 months Exclusion criteria: history of life‐threatening asthma or unstable asthma; taking high‐dose ICS or ICS/LABA; concurrent respiratory disease, respiratory infection, exercise‐induced asthma as the only asthma‐related diagnosis; asthma exacerbation in previous 4 weeks or > 4 exacerbations in previous 12 months; asthma hospitalisation in previous 4 weeks or > 2 asthma hospitalisations in previous 12 months; clinically significant uncontrolled disease; neurological or psychiatric disease or drug or alcohol abuse (in the patient or the parent/guardian); use of investigational medication; drug allergy; severe hypersensitivity to cow's milk proteins; concomitant medications that could significantly affect the course of asthma; use of potent cytochrome P450 3A4 inhibitors; affiliation with investigator’s site; child in care
Interventions	• Fluticasone propionate (100 or 250 μg) twice daily via Diskus • Fluticasone propionate and salmeterol (a fixed‐dose combination of 100 μg fluticasone propionate plus salmeterol at a dose of 50 μg or a fixed‐dose combination of fluticasone at a dose of 250 μg plus salmeterol at a dose of 50 μg) twice daily via Diskus
Outcomes	Primary endpoint: first serious asthma‐related event (defined as hospitalisation, intubation, or death)
Notes	Sponsored by GSK
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Central randomisation procedure
Allocation concealment (selection bias)	Low risk	Centralised Registration and Medication Ordering System (RAMOS‐NG)
Blinding (performance bias and detection bias) All outcomes	Low risk	Double‐blind with respect to salmeterol but not to dose of ICS
Independent Assessment of causation (detection bias) Asthma‐related events	Low risk	Independent adjudication committee for asthma outcomes
Incomplete outcome data (attrition bias) All outcomes	Low risk	Only 2 children in each group failed to complete the trial
Selective reporting (reporting bias)	Low risk	Data were extracted for all review outcomes