Mahler 2002.

Methods	Design: randomised, double‐blind, placebo‐controlled, parallel‐group, multi‐centre trial Duration: 6 months (+ 2 weeks run‐in period) Location: 65 centres in the United States
Participants	Population: 674 participants were randomised to salmeterol (160), fluticasone (168), salmeterol/fluticasone combination (165), and placebo (181) Baseline characteristics Age (mean years): salm 63.5, flut 64.4, salm/flut 61.9, pbo 64.0 % Male: salm 64.4, flut 61.3, salm/flut 62.4, pbo 75.1 % FEV1 predicted: salm 40, flut 41, salm/flut 41, pbo 41 Pack‐years (mean): salm 52.5, flut 54, salm/flut 55, pbo 60 Inclusion criteria: males and females aged 40 and older; history of at least 20 pack‐years; diagnosis of COPD; FEV1 < 65% of predicted but > 0.70 L, FEV1/FVC ratio < 70%; daily cough productive of sputum for 3 months of the year for 2 consecutive years and dyspnoea Exclusion criteria: current diagnosis of asthma; abnormal clinically significant ECG; moderate or severe exacerbation during the run‐in period; any significant medical disorder
Interventions	1. Salmeterol 50 bid (LABA) 2. Fluticasone 500 bid (ICS) 3. Salmeterol/fluticasone 50/500 bid (LABA/ICS) 4. Placebo (PBO) Inhaler device: Diskus dry powder Allowed co‐medications: Albuterol was allowed as needed, as were stable regimens of theophylline. Disallowed medications included oral corticosteroid use in the past 6 weeks, long‐term oxygen therapy, corticosteroids and all bronchodilators
Outcomes	Chronic Respiratory Disease Questionnaire, COPD exacerbations, AM pre‐dose and 2‐hour post‐dose FEV1, serial FEV1 over 12 hours, morning (AM) peak expiratory flow rate (PEFR), Chronic Bronchitis Symptoms Questionnaire (CBSQ), diary card data
Notes	Funding: GlaxoSmithKline Identifier(s): GSK SFCA3006
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was stratified by reversibility and investigative site to ensure a balance between treatment groups at each site and in terms of the number of reversible patients [no other details, industry sponsored]
Allocation concealment (selection bias)	Unclear risk	No details provided
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blind [presumed subject and investigator]
Blinding of outcome assessment (detection bias) All outcomes	Low risk	No details provided but outcomes not subject to detection bias
Incomplete outcome data (attrition bias) All outcomes	High risk	Withdrawal high and uneven between groups (salmeterol 28%, placebo 38%). Analyses based on the intent‐to‐treat (ITT) population consisted of all randomized subjects who had taken at least one dose of double‐blind study drug.
Selective reporting (reporting bias)	Low risk	All stated and expected outcomes were reported except FEV1 (secondary outcome)