Aubier 1999

Methods	A randomised, double‐blind, double‐dummy, multi‐centre, parallel‐group study over 28 weeks from May 1996 to November 1997, at 55 centres in 3 countries (Germany, France, and the Netherlands). Run‐in 2 weeks and follow‐up 2 weeks
Participants	Population: 503 adolescents and adults (12 to 79 years) with asthma  Baseline characteristics: mean age 48 years; FEV₁ 73% predicted Concomitant ICS used by 100% of participants Inclusion criteria: at least 12 years old with a documented clinical history of reversible airways disease; received treatment with any inhaled corticosteroid continuously for 12 weeks before run‐in; FEV₁ % predicted between 50% and 100%. At the end of the 2‐week run‐in period, symptomatic (symptom score ≥ 2 on at least 4 of the last 7 consecutive days), with mean morning PEF > 50% and < 85% of the maximum PEF 15 minutes after administration of inhaled salbutamol 400 μg  Exclusion criteria: taking long‐acting beta₂‐agonists
Interventions	Fluticasone propionate and salmeterol 500/50 μg twice daily Fluticasone propionate 500 μg + salmeterol 50 μg twice daily (separate inhalers) Fluticasone propionate 500 μg twice daily Delivery was by Diskus device
Outcomes	Primary outcome: mean morning PEF during weeks 1 to 12 The paper reports: "The incidence of drug‐related adverse events was similar for the three treatments" Full SAE data from Web report. One death from bronchial carcinoma on salmeterol and fluticasone (separate inhalers). This death was not included in Jaeschke 2008b, as the participant stopped taking study medication to allow for elective surgery and died of surgical complications but was still included in the trial and had intended to restart treatment postoperatively
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	403/503 (80%) completed the study
Selective reporting (reporting bias)	Low risk	Full data on GlaxoSmithKline website