Robinson 1999.

Methods	Randomised controlled trial. Design: 4‐way cross‐over trial (no information on any washout period). Duration: single day for each intervention. Participants were initially randomised to receive either the mannitol or HS. On the second day they were randomised to either the remaining active or the control for the first day. On the third day they were randomised to receive either the remaining active or either of the controls. The final day was the remaining control. Location: Australia.
Participants	Total participants: n = 12, 5 males and 7 females. Age mean (SD) (range): 29.9 (9.4) (16 ‐ 46) years. BMI mean (SD) (range): 21.0.(1.8) (18 ‐ 24). 10/12 participants were colonised with Pseudomonas aeruginosa, 7/12 had Stapylocococcus aureus (including the 2 participants without Pseudomonas aeruginosa) and 4/12 had Aspergillus fumigatus.
Interventions	Pre‐treated with terbutaline 1000 mcg (turbulhaler). Group 1: single dose 7 mL HS 6%. Group 2: IS (0.9%) plus matched voluntary cough. Group 3: mannitol 300 mg (encapsulated dry powder). Group 4: empty capsules with matched voluntary coughs.
Outcomes	Sputum isotope % clearance at 30 minutes, sputum isotope clearance at 90 minutes, mucociliary clearance.
Notes	Isotope clearance was reported in this paper as occurring at 60 minutes. This is actually the same time period as the 90‐minute clearance reported in 1997 paper. The terminology had been changed. Lung function data are the mean values for all trial days: FEV1 % predicted mean (SD) (range): 60.2.(16.5) (42 ‐ 87). FVC % predicted mean (SD) (range): 78.8.(16.5) (47 ‐ 102). FEF25‐75 % predicted mean (SD) (range): 32.5.(21.1) (11 ‐ 77). No sample size calculation stated.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No detail provided.
Allocation concealment (selection bias)	Unclear risk	The allocation concealment process was not described.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Participants could discern taste.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	The trials were coded such that the investigators were blinded to the identity of the intervention at the time of data analysis.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	No discussion of whether an ITT analysis had been used or of any withdrawals.
Selective reporting (reporting bias)	Low risk	All outcomes stated in the methods were described in the results.
Other bias	Unclear risk	Washout period not clear. No sample size calculation stated. The time course of the effect of the interventions on the outcomes is not clear.