IntroductionInhaledcorticosteroids(ICS)arewidelyusedtotreatpatients with chronic obstructive pulmonary disease(COPD), either alone or in combination with a long-actingbeta agonist (LABA) [ 1]. Clinical trials of the effects ofthese drugs in COPD usually allow patients to be recruitedirrespective of their previous ICS and/or LABA use [ 2]. Inthe subset of patients who are using ICS/LABA treatments,these are withdrawn at the start of the run-in period prior torandomisation to ensure that all patients in the study are at asimilar baseline point [ 3]. However, it is probable that thewithdrawal of medication in some patients causes clinicaldeterioration, which may continue if they are assigned tothe placebo arm of the study. This is not clinicallyjustifiable in some patients with severe or very severeCOPD [Global Initiative for Chronic Obstructive LungDisease (GOLD) stage III IV) [ 1]. However, in stablemoderate COPD patients (GOLD stage II), ICS / LABAwithdrawal is likely to be better tolerated, as these patientshave better lung function. Nevertheless, it is possible that asignificant clinical deterioration could occur, even inmoderate COPD patients.We hypothesised that ICS/LABA withdrawal wouldcause significant deterioration in clinical status andincrease airway inflammation even in stable, moderateGOLD stage-II COPD patients. We conducted a pilotstudy in 14 COPD patients using an open, parallel groupdesign. Patients were randomised at a ratio of 3:2,stratified according to cu rrent smoking status, toeither withdrawal or continuation of inhaled fluticasonepropionate/salmeterol (SFC) for 6 weeks. Patients per-formed spirometry and sputum induction at random-isation (V0) and at 2-week intervals up to 6 weeks (V1,V2, V3). Inclusion criteria at screening were postbron-chodilator forced expiratory volume in 1 s (FEV