Assael 2013.

Study characteristics
Methods	Open‐label RCT (randomised 1:1). Parallel design. Duration: 6 months ‐ 3 cycles of 28 days on and 28 days off. Multicentre: 91 centres in Europe and USA.
Participants	273 participants with documented CF diagnosis, aged 6 years and over, positive P aeruginosa sputum culture in previous 3 months, FEV1 75% or less at screening. Age (mean (SD)): AZLI 25.8 (9.1) years; TIS 25.1 (9.0) years. Gender split n (% males): AZLI 68 (50%); TIS 66 (50%). CFTR genotype (n (%)): homozygous for Δ508 ‐ AZLI 64 (54.7) and TIS 60 (54.1); heterozygous for Δ508 ‐ AZLI 36 (30.8) and TIS 30 (27.0); unidentified/other ‐ AZLI 17 (14.5) and TIS 21 (18.9). FEV1 % predicted at screening (n (%)): > 50% to ≤ 75% ‐ AZLI 76 (55.9) and TIS 75 (56.8); ≤ 50% ‐ AZLI 60 (44.1) and TIS 57 (43.2); < 25% ‐ AZLI 6 (4.4) and TIS 6 (4.5). Baseline FEV1 % predicted baseline values (mean (SD)): AZLI 52.3 (15.6); TIS 52.2 (14.6). Baseline FEV1 (L) (mean (SD)): AZLI 1.8 (0.6); TIS 1.8 (0.6). Baseline CFQ‐R respiratory symptom score (mean (SD)): AZLI 62.9 (20.4); TIS 58.0 (20.8).
Interventions	Intervention 1 (n = 136): AZLI 75 mg 3x daily via PARI eFlow electronic nebuliser. Intervention 2 (n = 132): TIS 300 mg 2x daily via PARI LC Plus nebuliser with compressor or via another nebuliser compatible with country‐specified labelling.
Outcomes	Primary outcomes: non‐inferiority of AZLI for relative change in FEV1 % predicted at day 28; superiority of AZLI for relative change in FEV1 % predicted over 3 treatment cycles. Secondary outcomes: time to need for additional IV antibiotics; CFQ‐R change from baseline for respiratory symptom scores; number of respiratory hospitalisations; additional anti‐pseudomonal use; change in weight from baseline; change in P aeruginosa sputum density; treatment satisfaction questionnaire.
Notes	26‐week active comparator period followed by an optional 24‐week open‐label single‐arm extension period for 3 months. Additional antibiotics allowed for diagnosis of pulmonary exacerbations.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Paper states randomised using ratio of 1:1; supplementary paper confirms randomisation by interactive voice/web response system using a code generated by Gilead.
Allocation concealment (selection bias)	Low risk	Interactive voice/web response system used.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Open‐label design, first treatment given 2x daily and the other 3x daily.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Open‐label design. An independent data safety board monitored adverse events during the trial, and an independent, blinded review committee identified respiratory hospitalisations and IV and inhaled anti‐pseudomonal antibiotics used for respiratory events. Unclear as to how a blinded monitoring committee might affect results.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Statistical analyses were performed on the ITT population: randomised participants receiving at least 1 dose of trial drug. 268/273 randomised participants received at least 1 dose: AZLI 136/137 or TNS 132/136. 233/273 participants (85.3%) completed the active‐comparator period. Reasons for withdrawal given for all dropouts.
Selective reporting (reporting bias)	Low risk	Primary and secondary outcomes reported with 95% CIs (relative and actual change in FEV1 % predicted). No protocol available but all outcomes listed in the methods are reported in the results.
Other bias	Unclear risk	The trial was funded by Gilead Sciences. 4 authors had consultancies with them and 3 authors are shareholders and/or employees.