Raphael 2017

Methods	A randomised, double‐blind, multi‐centre, parallel‐group, placebo‐controlled phase 3 study carried out over 12 weeks
Participants	Population: 787 adults and older children (> 12 years old) with persistent asthma previously treated with low‐ or mid‐dose inhaled corticosteroids (ICSs) or ICS/long‐acting beta‐agonists were enrolled in this study; 647 were randomised Baseline characteristics: mean age was 41.5 years; 71% of participants were receiving inhaled glucocorticoids at baseline, and 29% were receiving combined ICS/LABA; 44% of the cohort was male and 80% white Inclusion criteria: at the screening visit, patients were required to have FEV₁ 40% and 85% of predicted value for age, height, sex, and race, as per National Health and Nutrition Examination Survey III reference values. Previous treatment had to include low‐ or mid‐dose ICS or ICS/LABA for at least 1 month before consent was obtained (those taking ICS/LABA were required to have a pre‐screening visit to change to a comparable dose of ICS monotherapy). All patients were required to be able to replace their current SABA with albuterol/salbutamol HFA MDI inhalation aerosol at the screening visit for use as required for the duration of the study. Participants were to withhold all inhaled SABA bronchodilators for at least 6 hours before all study visits. In addition, participants had to demonstrate reversibility of disease (15% reversibility (all participants) and 200‐mL increase (participants 18 years old) from baseline FEV₁) within 30 minutes following SABA administration at the screening visit. Exclusion criteria: history of life‐threatening asthma exacerbation; asthma exacerbation requiring systemic corticosteroids 30 days before the screening visit, any hospitalisation for asthma 2 months before the screening visit
Interventions	• Fluticasone propionate and salmeterol (50 or 100 fluticasone propionate/12.5 μg salmeterol) twice daily • Fluticasone propionate (50 or 100 μg) twice daily A novel inhalation‐driven device (multi‐dose dry powder inhaler (MDPI); Teva Pharmaceuticals, Inc., Frazer, PA) was used to administer all doses of the interventions
Outcomes	Safety was assessed by monitoring of vital signs, physical and oropharyngeal examinations, ECGs, concomitant medication usage, and adverse events (AEs). All AEs were coded using the Medical Dictionary for Research Activities, version 17.10, preferred terms. Participants who demonstrated oropharyngeal signs consistent with oral candidiasis were to have a culture to confirm the diagnosis. SAEs were defined as an AE occurring at any dose that resulted in death, a life‐threatening AE, inpatient hospitalisation or prolongation of existing hospitalisation, persistent or significant disability or incapacity, or a congenital anomaly or birth defect. Asthma exacerbations, defined as any worsening of asthma requiring an emergency department visit or hospitalisation, were documented
Notes	Sponsored by Teva Pharmaceuticals
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Study states: "At the randomisation visit, patients were randomized to one of the five treatment groups in equal proportions" but does not specify how they were assigned to each treatment group"
Allocation concealment (selection bias)	Unclear risk	No details
Blinding (performance bias and detection bias) All outcomes	Unclear risk	Reported to be carried out in a double‐blind manner, but further information detailing how this was maintained throughout the study was not readily available in this publication It appears that blinding was maintained throughout the intervention and follow‐up period, but again it is unclear how this was done
Independent Assessment of causation (detection bias) Asthma‐related events	High risk	No clearly documented independent assessment of causation
Incomplete outcome data (attrition bias) All outcomes	Low risk	Reasons for participant withdrawal were clearly stated and discussed in the Results section: "AEs were the most frequent reason for study withdrawal, occurring in 12 patients overall, including six patients in the placebo group. Another 10 patients withdrew due to disease progression or lack of efficacy, including six in the placebo group" with overall attrition rates relatively low in both arms of the study ‐ 6.5% for fluticasone propionate and 4.2% for FS, respectively, from a relatively large population (N = 787). Safety data were reported for all included participants bar 6 people, with details as to why they were not reported made clear
Selective reporting (reporting bias)	Low risk	Study was registered with an NCT number and had clearly prespecified primary and secondary outcomes. Data for safety outcomes were clearly reported and were easily accessible