Ind 2003

Methods	A randomised, double‐blind, double‐dummy, multi‐centre, parallel‐group study over 28 weeks from January 1995 to December 1996, at 99 centres in Canada, Denmark, Iceland, Ireland, Italy, and the United Kingdom. Run‐in 4 weeks
Participants	Population: 502 adolescents and adults (16 to 75 years) with asthma poorly controlled on current ICS   Baseline characteristics: mean age 45 years; FEV₁ 2.3 L Concomitant ICS used by 100% of participants   Inclusion criteria: currently receiving ICS at a dose of 1000 to 1600 μg daily of inhaled BDP or equivalent; asthma poorly controlled (demonstrated by PEF < 85% of maximal achievable PEF after inhaling 400 μg salbutamol) and had experienced at least 2 exacerbations of asthma in the last year that required a change in asthma therapy. Therefore, over the last 10 days of the baseline period, had to demonstrate an average morning PEF < 90% of maximal achievable PEF measured at screening and diurnal variation in PEF ≥ 15%; had to have asthma symptoms on ≥ 4 of the last 7 days or nights of the baseline period Exclusion criteria: receiving continuous OCS; having any serious uncontrolled systemic disease or participation deemed unsuitable by the physician; had to demonstrate a period variation in PEF ≥ 15% (highest evening value ‐ lowest morning value as a percentage of highest PEF) over last 10 days and/or nights of the run‐in period and to have suboptimal PEF, with average PEF over last 10 days of the run‐in not exceeding 90% of post‐bronchodilator PEF (measured at visit 1)
Interventions	• Fluticasone propionate 250 μg + salmeterol 50 μg twice daily (in separate inhalers) • Fluticasone propionate 250 μg twice daily • Fluticasone propionate 500 μg twice daily Delivery was MDI (fluticasone propionate 500 arm not used in this review)
Outcomes	Primary efficacy variables: mean morning PEF; incidence and severity of asthma exacerbations No SAE information found in paper publication. Full SAE data on web report. One fatal pneumothorax on salmeterol and fluticasone (separate inhalers)
Notes	Sponsored by GSK
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Not reported
Allocation concealment (selection bias)	Unclear risk	Not reported
Blinding (performance bias and detection bias) All outcomes	Low risk	Double‐blind, double‐dummy
Independent Assessment of causation (detection bias) Asthma‐related events	High risk	Causation of SAEs not independently assessed
Incomplete outcome data (attrition bias) All outcomes	Low risk	432/502 (86% completed study)
Selective reporting (reporting bias)	Low risk	Full data on GlaxoSmithKline website