Donaldson 2020.

Study characteristics
Methods	RCT. Design: parallel, double‐blind. Location: single‐centre study, USA. Duration: 4 weeks.
Participants	Children aged ≥ 5 to <18 years with confirmed CF and stable lung disease (FEV1 > 60%). Age, mean (SD): HS 11.9 (3.3) years; IS 9.8 (3.6) years; overall  11.1 (3.5) years. Sex, number (%) female: HS  8/14 (57%); IS 4/9 (44%); overall 12/23 (52%). Disease status Post‐BD FVC % predicted, mean (SD): HS 101.4 (12.3); IS 100.0 (10.9); overall 100.8 (11.5). Post‐BD FEV1 % predicted, mean (SD): HS 97.0 (12.6); IS 93.2 (13.4); overall 95.5 (12.8). Post‐BD FEF25-75 % predicted, mean (SD): HS 91.1 (24.9); IS 86.2 (37.8); overall 89.7 (27.9). Taking inhaled antibiotics, number (%): HS 2 (14%); IS 1 (11%); overall 3 (13%). On chronic HS, number (%): HS 1 (7%); IS 3 (33%); overall 4 (17%). Exclusion criteria Individuals with unstable lung disease, who did not tolerate HS, or who were unable or unwilling to withhold prescribed HS for at least 2 weeks before the screening visit were excluded. 24 screened, 23 randomised, 20 completed treatment and follow‐up.
Interventions	Intervention: inhaled HS (6% NaCl, 4 mL) three times a day for 28 days. Control: inhaled IS (0.12% NaCl, 4 mL) three times a day for 28 days. Both treatments were given via an eFlow nebuliser.
Outcomes	Primary outcome Change in average whole lung MCC measured over 90 min (AveClr90) between baseline and 4 weeks of treatment. Secondary outcomes Change in FEV1 % predicted, CFQ‐R respiratory symptom domain score (Quittner 2009).
Notes	A sample size calculation was carried out to provide 80% power at a significance level of 0.05 to detect a treatment difference of 6.6% (absolute change in MCC).
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The randomisation schedule was generated and maintained by the investigational drug service.
Allocation concealment (selection bias)	Low risk	The randomisation schedule was generated and maintained by the investigational drug service.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Participants, care provider and investigator were all blinded to treatment allocation. Taste masking was not done, therefore it is unclear whether the participants were truly blinded.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Outcome assessors were also blinded to treatment allocation.
Incomplete outcome data (attrition bias) All outcomes	Low risk	24 participants were recruited but there was one screen failure. 23 participants were randomised and 20 participants completed the study. Sample size calculation was carried out.
Selective reporting (reporting bias)	Low risk	All outcomes listed in the methods and in the trial registration document are reported in the results.
Other bias	Low risk	No other risk of bias identified.