SOLO.

Methods	Randomised, open‐label, parallel group study over a 6‐month period. In the two‐week run‐in patients used current asthma treatment (pre‐study ICS +/‐ LABA). Unknown number of centres in Canada.
Participants	Population:  1538 asthmatic adults aged 12 years and over with asthma  on at least 400 µg/day ICS and symptomatic unless  also on LABA (74% of those randomised were on LABA and ICS). Inclusion Criteria: Aged 12 years or more and asthma diagnosis for a minimum of three months. Previous treatment with ICS for at least 3 months (at least 400µg/day) with at least 3 inhalations of relief medication in the last 7 days of run‐in, or concurrent use of LABA. Patients with a smoking history of over 10 pack‐years or exacerbation requiring a change in asthma treatment in the past 14 days were not included; nor were patients already using SiT. Baseline Characteristics: Mean age:  40 years. FEV1 not measured but PEF 94% predicted. Mean ICS dose at enrolment 569 µg/day, and 74% were also using LABA. Hospital admission for asthma in the past year:unknown. Course of oral steroids for asthma in past year: unknown.
Interventions	Budesonide/formoterol 200/6 µg one inhalation twice daily (400 µg budesonide/day), as maintenance and reliever. Current best practice.
Outcomes	Primary outcome: Time to first severe exacerbation Secondary outcomes Number of severe exacerbations Mean use of as‐needed treatment Prescribed asthma medications PEF ACQ. A subgroup used induced sputum eosinophils as a primary outcome. Definition of severe exacerbation Hospitalisation, or ER visit or course of oral corticosteroids for at least 3 days due to asthma
Notes	Study D5890L00004 is now reported on the AstraZeneca trials web site (for 2013 update).
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Randomization code assigned from a computer generated randomisation schedule" Demoly 2009
Allocation concealment (selection bias)	Low risk	"Patients were randomised, strictly sequentially...using coded envelopes. When a patient had been randomised, the envelope was opened and the code was revealed." Demoly 2009
Blinding (performance bias and detection bias) All outcomes	High risk	Open‐label study
Incomplete outcome data (attrition bias) All outcomes	Low risk	1400/1538 (91%) completed the study
Selective reporting (reporting bias)	Low risk	Data have been obtained for all primary outcome measures