Nebulised hypertonic saline for cystic fibrosis

Abstract

Background

Hypertonic saline enhances mucociliary clearance and may lessen the destructive inflammatory process in the airways. This is an update of a previously published review.

Objectives

To investigate efficacy and tolerability of nebulised hypertonic saline treatment in people with cystic fibrosis (CF) compared to placebo or other treatments that enhance mucociliary clearance.

Search methods

We searched the Cochrane Cystic Fibrosis and Genetic Disorders Group's Cystic Fibrosis Trials Register, comprising references identified from comprehensive electronic database searches, handsearches of relevant journals and abstract books of conference proceedings. We also searched ongoing trials databases.

Most recent search: 25 April 2022.

Selection criteria

We included randomised and quasi‐randomised controlled trials assessing hypertonic saline compared to placebo or other mucolytic therapy, for any duration or dose regimen in people with CF (any age or disease severity).

Data collection and analysis

Two authors independently reviewed all identified trials and data, and assessed trial quality. We assessed the certainty of the evidence using GRADE. For cross‐over trials we stipulated a one‐week washout period. We planned to use results from a paired analysis in the review, but this was only possible in one trial. For other cross‐over trials, we chose to treat the trials as if they were parallel.

Main results

We included 24 trials (1318 participants, aged one month to 56 years); we excluded 29 trials, two trials are ongoing and six are awaiting classification. We judged 15 of the 24 included trials to have a high risk of bias due to participants' ability to discern the taste of the solutions.

Hypertonic saline 3% to 7% versus placebo (stable disease)

We are uncertain whether the regular use of nebulised hypertonic saline in stable lung disease leads to an improvement in forced expiratory volume in one second (FEV₁) % predicted at four weeks, (mean difference (MD) 3.30%, 95% confidence interval (CI) 0.71 to 5.89; 4 trials, 246 participants; very low‐certainty evidence). In preschool children we found no difference in lung clearance index (LCI) at four weeks, but a small improvement after 48 weeks of treatment with hypertonic saline compared to isotonic saline (MD ‐0.60, 95% CI ‐1.00 to ‐0.19; 2 trials, 192 participants). We are also uncertain whether hypertonic saline made a difference to mucociliary clearance, pulmonary exacerbations or adverse events compared to placebo.

Hypertonic saline versus control (acute exacerbation)

Two trials compared hypertonic saline to control, but only one provided data. There may be little or no difference in lung function measured by FEV₁ % predicted after hypertonic saline compared to isotonic saline (MD 5.10%, 95% CI ‐14.67 to 24.87; 1 trial, 130 participants). Neither trial reported any deaths or measures of sputum clearance. There were no serious adverse events.

Hypertonic saline versus rhDNase

Three trials compared a similar dose of hypertonic saline to recombinant deoxyribonuclease (rhDNase); two trials (61 participants) provided data for inclusion in the review. We are uncertain whether there was an effect of hypertonic saline on FEV₁ % predicted after three weeks (MD 1.60%, 95% CI ‐7.96 to 11.16; 1 trial, 14 participants; very low‐certainty evidence). At three months, rhDNase may lead to a greater increase in FEV₁ % predicted than hypertonic saline (5 mL twice daily) at 12 weeks in participants with moderate to severe lung disease (MD 8.00%, 95% CI 2.00 to 14.00; low‐certainty evidence). We are uncertain whether adverse events differed between the two treatments. No deaths were reported.

Hypertonic saline versus amiloride

One trial (12 participants) compared hypertonic saline to amiloride but did not report on most of our outcomes. The trial found that there was no difference between treatments in measures of sputum clearance (very low‐certainty evidence).

Hypertonic saline compared with sodium‐2‐mercaptoethane sulphonate (Mistabron®)

One trial (29 participants) compared hypertonic saline to sodium‐2‐mercaptoethane sulphonate. The trial did not measure our primary outcomes. There was no difference between treatments in any measures of sputum clearance, courses of antibiotics or adverse events (very low‐certainty evidence).

Hypertonic saline versus mannitol

One trial (12 participants) compared hypertonic saline to mannitol, but did not report lung function at relevant time points for this review; there were no differences in sputum clearance, but mannitol was reported to be more 'irritating' (very low‐certainty evidence).

Hypertonic saline versus xylitol

Two trials compared hypertonic saline to xylitol, but we are uncertain whether there is any difference in FEV₁ % predicted or median time to exacerbation between groups (very low‐certainty evidence). No other outcomes were reported in the review.

Hypertonic saline 7% versus hypertonic saline 3%

We are uncertain whether there was an improvement in FEV₁ % predicted after treatment with 7% hypertonic saline compared with 3% (very low‐certainty evidence).

Authors' conclusions

We are very uncertain if regular use of nebulised hypertonic saline by adults and children over the age of 12 years with CF results in an improvement in lung function after four weeks (three trials; very low‐certainty evidence); there was no difference seen at 48 weeks (one trial; low‐certainty evidence). Hypertonic saline improved LCI modestly in children under the age of six years.

Evidence from one small cross‐over trial in children indicates that rhDNase may lead to better lung function than hypertonic saline at three months; qualifying this, we highlight that while the study did demonstrate that the improvement in FEV₁ was greater with daily rhDNase, there were no differences seen in any of the secondary outcomes.

Hypertonic saline does appear to be an effective adjunct to physiotherapy during acute exacerbations of lung disease in adults. However, for the outcomes assessed, the certainty of the evidence ranged from very low to low at best, according to the GRADE criteria.

The role of hypertonic saline in conjunction with cystic fibrosis transmembrane conductance regulator (CFTR) modulator therapy now needs to be considered, and future research needs to focus on this aspect.

Plain language summary

Hypertonic saline (salt water with at least 3% salt) nebulised as a fine mist through a mask or mouthpiece for cystic fibrosis

Review question

Is inhaling hypertonic saline (salt water with at least 3% salt) as a mist through a mask or mouthpiece better for improving mucus clearance in the lungs of people with cystic fibrosis (CF) than a placebo (a mist with no or very little salt) or other agents?

Background

People with CF produce large amounts of thick mucus which is difficult to clear and blocks up their airways. Chest physiotherapy or medication, e.g. hypertonic saline, or both combined, are used to try and clear this mucus from the airways. Hypertonic saline is water with a concentration of 3% to 7% salt and is inhaled as a fine mist. This is an update of an earlier review.

Key messages

• We are uncertain whether inhaling nebulised hypertonic saline regularly improves lung function compared to placebo.

• Nebulised hypertonic saline does seem to work well as an add‐on to physiotherapy.

What did we do?

We searched for studies that looked at the use of nebulised hypertonic saline compared to either a placebo or a different type of treatment for clearing mucus from the lungs. We compared the size and methods of the studies and stated how confident we were in the results.

What did we find?

We included 24 trials with 1318 people with CF aged between one month and 56 years. Two thirds of the trials compared hypertonic saline to a placebo (a dummy treatment); the remaining trials compared hypertonic saline to another type of mucus clearing treatment (including mannitol; rhDNase (Pulmozyme®); amiloride; Mistabron®; xylitol); and one trial compared 7% hypertonic saline with 3% hypertonic saline. Trials assessed different concentrations of hypertonic saline with different nebulisers and different treatment schedules; the most common treatment was twice‐daily 7% hypertonic saline and the most common nebuliser was ultrasonic. Most trials treated people with a bronchodilator to widen the airways before giving the hypertonic saline.

Main results

Hypertonic saline 3% to 7% versus placebo

We are not sure whether hypertonic saline leads to an improvement in lung function in stable disease after four weeks. Two trials showed that there may be a small improvement in lung function (measured using the lung clearance index) with hypertonic saline compared to placebo in preschool children. We are also unsure whether hypertonic saline makes a difference to clearing mucus from the lungs, exacerbations or side effects compared to placebo.

During exacerbations, we found that there may be little or no difference in lung function after hypertonic saline compared to placebo. The trials did not report any serious side effects and there were no deaths.

One study compared 7% hypertonic saline with a lower concentration of hypertonic saline (3%); we are uncertain whether the higher concentration improved lung function.

Hypertonic saline versus mucus mobilising treatments

Three trials compared hypertonic saline with rhDNase and found that rhDNase may lead to an improvement in lung function compared to hypertonic saline after three months. We are unsure whether there is any difference in side effects.

One trial compared hypertonic saline to amiloride and a further trial compared hypertonic saline to sodium‐2‐mercaptoethane sulphonate (Mistabron®), but neither of the trials gave information about the effect of the treatments on lung function.

Similarly, a trial comparing hypertonic saline with mannitol did not give information about the effects on lung function, but they did report that there was no difference between treatments in clearing mucus from the lungs. People taking mannitol said it was more irritating than hypertonic saline.

Two trials compared hypertonic saline with xylitol, but we are unsure if there is any difference in lung function and none of our other outcomes were measured.

What are the limitations of this evidence?

We are not confident in the evidence from these trials. There is a high risk that people knew which treatment they were receiving in half the trials as they could taste the difference between the solutions.

Other factors that made us unsure of the results were the small numbers of people taking part in the trials combined with a wide variation in results; also, some trials limited participants to those who could tolerate hypertonic saline or to certain age groups.

How up to date is this evidence?

The evidence is current to 25 April 2022.

Summary of findings

Summary of findings 1. Hypertonic saline 3% to 7% versus isotonic saline for cystic fibrosis (stable lung disease).

Hypertonic saline 3% to 7% versus isotonic saline for cystic fibrosis (stable lung disease)
Patient or population: adults and children with cystic fibrosis (stable lung disease) Settings: outpatients Intervention: hypertonic saline 3% to 7% Comparison: isotonic saline
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of participants (trials)	Certainty of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Isotonic saline	Hypertonic saline 3% to 7%
FEV1 (% predicted) change from baseline, short‐term   Follow‐up: 4 weeks	The mean change in FEV1 (% predicted) ranged from ‐1.42 to 2.8 in the isotonic saline groups.	The mean change in FEV1 (% predicted) was 3.30 higher (0.71 higher to 5.89 higher) in the hypertonic saline group.	NA	246 (4 trials)a	⊕⊝⊝⊝ very lowb,c,d,e	1 trial had a cross‐over design.
FEV1 (% predicted) change from baseline, long‐term   Follow‐up: 48 weeks	The mean change in FEV1 (% predicted) was 2.44 in the isotonic saline group.	The mean change in FEV1 (% predicted) was 2.31 higher (2.72 lower to 7.34 higher) in the hypertonic saline group.	NA	134 (1 trial)	⊕⊕⊝⊝ lowb,e	The included trial also measured change in FEV1 (% predicted) at: 12 weeks, MD 4.10 (95% CI ‐0.08 to 8.28); 24 weeks, MD 5.37 (95% CI 1.03 to 9.71); and 36 weeks, MD 3.63 (95% CI ‐1.56 to 8.82).
LCI   Follow‐up: 4 weeks	The mean LCI was 8.89 in the isotonic saline group.	The mean LCI was 1.03 lower (2.76 lower to 0.70 higher) in the hypertonic saline group.	NA	10 (1 trial)	⊕⊝⊝⊝ very lowf,g	Trial had a cross‐over design. 2 trials reported change from baseline in LCI at 12 months in preschool children and found a slight difference between groups favouring hypertonic saline MD ‐0.60 (95% CI ‐1.00 to ‐0.19).
Mortality	Outcome not reported.		NA	NA	NA
Measures of sputum clearance   Follow‐up: up to 24 hours	The trials used radio‐labelled aerosol clearance and an 'area under the curve' measure to assess mucociliary clearance. Both measures significantly favoured treatment with hypertonic saline.		NA	80 (4 trials)	⊕⊝⊝⊝ very lowb,e,f	All trials had a cross‐over design.
Pulmonary exacerbations   Follow‐up: up to 48 weeks	1 trial showed that there were fewer exacerbations per year requiring intravenous antibiotic therapy in the hypertonic saline group than in the isotonic saline group and that the interval during which participants remained free of exacerbations was also significantly longer in the hypertonic saline group.   The second trial found no difference in the mean number of exacerbations per year or hospitalisation rates between the hypertonic saline group and the controls.   2 further trials reported the rate of exacerbations in preschool children and found no difference between groups.		NA	607 (4 trials)	⊕⊕⊝⊝ very lowe,h,i
Adverse events   Follow‐up: up to 48 weeks	There were no differences between treatment groups in most adverse events including cough,chest tightness, pharyngitis, haemoptysis, sinusitis, sneezing, tonsillitis and vomiting. Fever was slightly more common in the isotonic saline group whilst rhinorrhoea was more common in the hypertonic saline group.		NA	513 (3 trials)	⊕⊕⊝⊝ lowb,e	A further trial reported adverse events at this time point, but reported number of events rather than number of participants experiencing adverse events and we have not included these data in the analysis. Adverse events were also reported in shorter‐term trials. Single‐dose trials: there was little difference in the drop in FEV1 within 5 minutes of receiving hypertonic saline (MD 5.20%, 95% CI ‐0.59 to 10.99; 1 trial, 12 participants) but participants reported throat irritation, shortness of breath and chest tightness related to inhalation of hypertonic saline. Medium‐term trials: 3 trials reported at 2 to 4 weeks and no differences were reported for cough, hoarseness, chest pain, pharyngitis, haemoptysis, wheezing, or nasal congestion. 1 trial reported more adverse events in the hypertonic saline group for fever, rhinorrhoea, malaise and ear infections (P = 0.003). The same trial reported increased sputum production in the hypertonic saline group whilst the remaining 2 trials did not report this adverse event.
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; FEV1: forced expiratory volume in 1 second;LCI: lung clearance index; MD: mean difference; NA: not applicable.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

^a1 trial (n = 19) was of a cross‐over design.
^bDowngraded once due to risk of bias arising from participants being able to discern the taste of the intervention and limited information about trial methods.
^cDowngraded once due to imprecision caused by wide CIs.
^dDowngraded once due to inconsistency as there was substantial heterogeneity (I² = 51%) which may have originated from different age groups recruited in the trials or different baseline levels of lung function.
^eDowngraded once due to indirectness as the results only apply to those who can tolerate hypertonic saline.
^fDowngraded twice due to imprecision: cross‐over trial analysed as a parallel trial (due to available data) which is likely to over‐estimate the within‐trial variability and increase imprecision, and small number of participants.
^gDowngraded once due to applicability: results apply only to those who can tolerate hypertonic saline and the trial only included children aged 6 to 18 years, so results may not apply to adults.
^hDowngraded once due to risk of bias: 1 trial was at high risk of detection bias as participants could discern the taste of the intervention.
ⁱDowngraded once for inconsistency as there was heterogeneity in the results of the trials. Three out of the four trials reported no difference whilst one trial found a slight difference in favour of hypertonic saline.

Summary of findings 2. Hypertonic saline 3% to 7% versus isotonic saline for cystic fibrosis (during acute exacerbations of lung disease).

Hypertonic saline 3% to 7% versus isotonic saline for cystic fibrosis (during acute exacerbations of lung disease)
Patient or population: adults and children with cystic fibrosis (during acute exacerbations of lung disease) Settings: hospitalised patients and outpatients Intervention: hypertonic saline 3% to 7% Comparison: isotonic saline
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of participants (trials)	Certainty of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Isotonic saline	Hypertonic saline 3% to 7%
FEV1 (% predicted) change from baseline, short‐term   Follow‐up: approximately 14 days (at time of hospital discharge)	The mean % change in FEV1 (% predicted) was 32.3% in the isotonic saline group.	The mean % change in FEV1 (% predicted) was 5.10% higher (14.67% lower to 24.87% higher) in the hypertonic saline 3% to 7% group.	NA	130 (1 trial)	⊕⊕⊝⊝ lowa,b
FEV1 (% predicted) change from baseline, long‐term   Follow‐up: NA	Outcome not reported.		NA	NA	NA
LCI   Follow‐up: NA	Outcome not reported.		NA	NA	NA
Mortality   Follow‐up: NA	No deaths were reported in either trial.		NA	142 (2 trials)	⊕⊕⊝⊝ lowb,c	1 trial had a cross‐over design.
Measures of sputum clearance   Follow‐up: NA	Outcome not reported.		NA	NA	NA
Pulmonary exacerbations   Follow‐up: up to 1 year	There was no significant difference between the groups in time until the next pulmonary exacerbation requiring hospitalisation.		HR 0.86 (95% CI 0.57 to 1.30)	132 (1 trial)	⊕⊕⊝⊝ lowa,b
Adverse events   Follow‐up: up to 1 year	Adverse events reported were cough and wheeze. No serious adverse events were reported.		NA	142 (2 trials)	⊕⊝⊝⊝ very lowb,c,d	1 trial had a cross‐over design.
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; FEV1: forced expiratory volume in 1 second; HR: hazard ratio;LCI: lung clearance index; MD: mean difference; NA: not applicable.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

^aDowngraded once due to risk of bias: high risk of selection bias due to sequential allocation.
^bDowngraded once due to applicability: results apply only to those who can tolerate hypertonic saline, and the trial included only adults so results may not apply to children.
^cDowngraded once due to risk of bias: first trial was at high risk of detection bias as participants could discern the taste of the intervention, second trial was at high risk of selection bias due to sequential allocation.
^dDowngraded once due to imprecision: no numerical data provided and small sample size.

Summary of findings 3. Hypertonic saline compared with rhDNase with for cystic fibrosis.

Hypertonic saline compared with rhDNase with for cystic fibrosis
Patient or population: adults and children with cystic fibrosis Settings: outpatients Intervention: hypertonic saline (daily) Comparison: rhDNase (daily)a
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of participants (trials)	Certainty of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	rhDNase	Hypertonic saline
FEV1 (% predicted) change from baseline, short‐term   Follow‐up: 3 weeks	The mean change from baseline in FEV1 (% predicted) was 1.6% higher (7.96% lower to 11.16% higher) in the hypertonic saline group compared to the daily rhDNase group.b		NA	14 (1 trial)	⊕⊝⊝⊝ very lowc,d,e	Trial had a cross‐over design.   No significant difference in the primary outcome (lung function) at this time‐point, with improvements only in secondary outcomes.
FEV1 (% predicted) change from baseline, long‐term   Follow‐up: 3 months	The mean change from baseline in FEV1 (% predicted) was 8% higher (2% higher to 14% higher) in the hypertonic saline group compared to the daily rhDNase group.b		NA	47 (1 trial)	⊕⊝⊝⊝ very lowb,f,g	Trial had a cross‐over design. An additional cross‐over trial of 18 participants found no difference between treatments in FEV1 after 10 weeks (no data presented).
LCI	Outcome not reported.		NA	NA	NA
Mortality	Outcome not reported.		NA	NA	NA
Measures of sputum clearance	Outcome not reported.		NA	NA	NA
Pulmonary exacerbations   Follow‐up: NA	15 episodes occurred during treatment with hypertonic saline and 18 with daily rhDNase. There was no statistical difference between treatments (see comment).		NA	47 (1 trial)	⊕⊝⊝⊝ very lowb,f,g	Trial had a cross‐over design. Number of episodes reported rather than the number of participants with exacerbations (leading to a unit of analysis issue) so data not entered into the analysis.
Adverse events   Follow‐up: 3 months	Increased cough was reported in 13 participants using hypertonic saline and 17 on daily rhDNase. There were similar rates of other adverse events between treatment arms (see comment).		NA	47 (1 trial)	⊕⊝⊝⊝ very lowb,f,g	Trial had a cross‐over design, so data not entered into analysis.
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; FEV1: forced expiratory volume in 1 second; LCI: lung clearance index; MD: mean difference; NA: not applicable.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

^aAn alternate day rhDNase group was also included in 1 trial (Suri 2001), but to allow a comparison across the trials, only results from the rhDNase daily group are presented in the tables.
^bData analysed as MD between treatment groups via generic inverse variance due to cross‐over design of the trial; therefore, an estimate of the assumed risk is not available.
^cDowngraded once due to risk of bias: high risk of detection bias as participants could discern the taste of the intervention and limited information was provided about the methodological design of the trial.
^dDowngraded once due to applicability: results apply only to those who can tolerate hypertonic saline.
^eDowngraded once due to imprecision: cross‐over trial analysed as a parallel trial due to available data, this approach is likely to over‐estimate the within study variability and increase imprecision; also small sample size.
^fDowngraded once due to applicability: results apply only to those who can tolerate hypertonic saline, and the trial included only participants under the age of 18 so results may not apply to adults.
^gDowngraded once due to imprecision: small sample size.

Summary of findings 4. Hypertonic saline compared with amiloride for cystic fibrosis.

Hypertonic saline compared with amiloride for cystic fibrosis
Patient or population: adults and children with cystic fibrosis Settings: outpatients Intervention: hypertonic saline Comparison: amiloride
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of Participants (studies)	Certainty of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Amiloride	Hypertonic saline
FEV1: change from baseline, short‐term	Outcome not reported.
FEV1: change from baseline, long‐term	Outcome not reported.
LCI	Outcome not reported.
Mortality	Outcome not reported.
Measures of sputum clearance   Follow‐up: 60 minutes	There was no significant difference between treatment groups.		NA	12 (1 trial)	⊕⊝⊝⊝ very lowa,b,c	Trial had cross‐over design.
Pulmonary exacerbations	Outcome not reported.
Adverse events	Outcome not reported.
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; FEV1: forced expiratory volume in 1 second; LCI: lung clearance index; NA: not applicable.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

^aDowngraded once due to risk of bias: high risk of detection bias as participants could discern the taste of the intervention and limited information was provided about the trial methods (including whether a washout period was used).
^bDowngraded once due to applicability: results apply only to those who can tolerate hypertonic saline, and the trial included only adults so results may not apply to children.
^cDowngraded once due to imprecision: no numerical data provided and small sample size.

Summary of findings 5. Hypertonic saline compared with sodium‐2‐mercaptoethane sulphonate (Mistabron®) for cystic fibrosis.

Hypertonic saline compared with sodium‐2‐mercaptoethane sulphonate (Mistabron®) for cystic fibrosis
Patient or population: adults and children with cystic fibrosis Settings: outpatients Intervention: hypertonic saline Comparison: sodium‐2‐mercaptoethane sulphonate
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of participants (trials)	Certainty of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Sodium‐2‐mercaptoethane sulphonate	Hypertonic saline
FEV1: short‐term	Outcome not reported.
FEV1: long‐term	Outcome not reported.
LCI	Outcome not reported.
Mortality	Outcome not reported.
Measures of sputum clearance   Follow‐up: 2 months	No significant difference in sputum volume, colour or cough frequency between the groups.		NA	29 (1 trial)	⊕⊝⊝⊝ very lowa,b,c	Trial had cross‐over design.
Pulmonary exacerbations   Follow‐up: 2 months	See comment.		NA	29 (1 trial)	⊕⊝⊝⊝ very lowa,b,c	Trial had cross‐over design. The only information provided relevant to this outcome was that there was no change in the number of courses of antibiotics prescribed.
Adverse events   Follow‐up: 2 months	See comment.		NA	29 (1 trial)	⊕⊝⊝⊝ very lowa,b,c	Trial had cross‐over design. Participants in both treatment groups described coughing at the beginning of their inhalations. No serious adverse events occurred during the trial.
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; FEV1: forced expiratory volume in 1 second; LCI: lung clearance index; NA: not applicable.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

^aDowngraded once due to risk of bias: high risk of detection bias as participants could discern the taste of the intervention and limited information was provided about the trial design.
^bDowngraded once due to applicability: results apply only to those who can tolerate hypertonic saline and the trial included only children aged 6 to 15 years so results may not apply to other age groups.
^cDowngraded once due to imprecision: no numerical data provided and small sample size.

Summary of findings 6. Hypertonic saline compared with mannitol for cystic fibrosis.

Hypertonic saline compared with mannitol for cystic fibrosis
Patient or population: adults and children with cystic fibrosis Settings: outpatients Intervention: hypertonic saline Comparison: mannitol
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of participants (trials)	Certainty of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Mannitol	Hypertonic saline
FEV1: short‐term   Follow‐up: up to 95 minutes	See comment.		NA	12 (1 trial)	⊕⊝⊝⊝ very lowa,b,c	Trial had cross‐over design. FEV1 was assessed in the included trial at 5 minutes and 95 minutes postintervention. These very short‐term time‐points are not of clinical relevance to this review. Change from baseline within‐groups was reported but no between‐group data.
FEV1: long‐term	Outcome not reported.
LCI	Outcome not reported.
Mortality	Outcome not reported.
Measures of sputum clearance   Follow‐up: up to 95 minutes	There was no significant difference between treatment groups for matched voluntary cough.		NA	12 (1 trial)	⊕⊝⊝⊝ very lowa,b,d	Trial had cross‐over design.
Pulmonary exacerbations	Outcome not reported.
Adverse events   Follow‐up: up to 95 minutes	See comment.		NA	12 (1 trial)	⊕⊝⊝⊝ very lowa,b,d	Trial had cross‐over design. Mannitol was considered to be a more 'irritating' treatment than other treatments (4‐armed trial); no specific data given.
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; FEV1: forced expiratory volume in 1 second; LCI: lung clearance index; NA: not applicable.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

^aDowngraded once due to risk of bias: high risk of detection bias as participants could discern the taste of the intervention and no washout period was used.
^bDowngraded once due to applicability: results apply only to those who can tolerate hypertonic saline, and the trial included only participants over the age of 16 so results may not apply to younger children.
^cDowngraded once due to applicability: the outcome measured only at very short‐term time points (minutes after intervention), which are not of clinical relevance to this review.
^dDowngraded once due to imprecision: no numerical data provided and small sample size.

Summary of findings 7. Hypertonic saline 7% versus xylitol for cystic fibrosis.

Hypertonic saline 7% versus xylitol for cystic fibrosis
Patient or population: adults with cystic fibrosis Settings: outpatients Intervention: hypertonic saline 7% Comparison: aerosolised xylitol
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of participants (trials)	Certainty of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Xylitol	Hypertonic saline 7%
FEV1 (% predicted) change from baseline, short‐term   Follow‐up: 2 weeks	Mean change in FEV1 % predicted ranged from ‐0.1 to 8.8.	Mean change in FEV1 % predicted was 1.89% higher in the hypertonic saline (0.8% lower to 4.58% higher).	NA	89 (2)	⊕⊝⊝⊝ very lowa,b,c	1 trial had a cross‐over design.
FEV1 (% predicted) change from baseline, long‐term	This outcome was not reported.
LCI	This outcome was not reported.
Mortality	See comments.					No deaths were reported in either trial.
Measures of sputum clearance	This outcome was not reported.
Pulmonary exacerbations   Follow‐up: 6 months	There was no difference in median time to exacerbation between groups.		NA	59 (1)	⊕⊕⊝⊝ lowb,c	Data were presented as a graph and we have reported narratively
Adverse events: serious adverse events   Follow‐up: 2 weeks	There was no difference between groups in relation to serious adverse events, RR 4.07 (95% CI 0.47 to 35.34).		NA	89 (2)	⊕⊝⊝⊝ very lowa,c,d	There were no differences between treatment groups in specific adverse events including gastro‐intestinal disorders, hepato‐biliary disorders, bronchospasm, haemoptysis or vascular disorders.
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; FEV1: forced expiratory volume in 1 second;LCI: lung clearance index; MD: mean difference; NA: not applicable; RR: risk ratio.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

^aDowngraded once due to risk of bias due to trial methods being unclear in one trial and the risk of participants being able to discern the taste of the interventions due to their sweet and salty taste.
^bDowngraded once due to imprecision caused by small participant numbers.
^cDowngraded once due to indirectness as the trials were only carried out in adults and the results may not be applicable to children.
^dDowngraded once due to imprecision caused by low participant numbers, low event rates and wide CIs.

Summary of findings 8. Hypertonic saline 7% versus hypertonic saline 3% for cystic fibrosis.

Hypertonic saline 7% versus hypertonic saline 3% for cystic fibrosis
Patient or population: children with cystic fibrosis Settings: outpatients Intervention: hypertonic saline 7% Comparison: hypertonic saline 3%
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of participants (trials)	Certainty of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Hypertonic saline 3%	Hypertonic saline 7%
FEV1 (% predicted) change from baseline, short‐term   Follow‐up: 28 days	Mean (SD) change in FEV1 % predicted was 12.53% (20.04).	Mean change in FEV1 % predicted was 13.00% lower in the hypertonic saline 7% group (25.20% lower to 0.73% lower).	NA	30 (1)	⊕⊝⊝⊝ very lowa, b, c	The trial was seriously underpowered and was carried out as a pilot study (sample size calculation stated that 394 participants were needed and the pilot trial included 30).
FEV1 (% predicted) change from baseline, long‐term	This outcome was not reported.
LCI	This outcome was not reported.
Mortality	This outcome was not reported.
Measures of sputum clearance	This outcome was not reported.
Pulmonary exacerbations	This outcome was not reported.
Adverse events	This outcome was not reported.
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; FEV1: forced expiratory volume in 1 second;LCI: lung clearance index; MD: mean difference; NA: not applicable; SD: standard deviation.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate: the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited: the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate: the true effect is likely to be substantially different from the estimate of effect.

^aDowngraded once due to risk of bias due to the trial being seriously underpowered.
^bDowngraded once due to imprecision caused by small participant numbers.
^cDowngraded once due to indirectness as the trials were only carried out in children and the results may not be applicable to adults.

Background

Description of the condition

Cystic fibrosis (CF) is the most common life‐limiting autosomal recessive genetic disorder in populations of Northern European descent (Bobadilla  2002). In 1989, the gene responsible was identified on the long arm of chromosome 7 (Kerem 1989). This gene encodes for a protein named the cystic fibrosis transmembrane conductance regulator (CFTR), which functions as a chloride channel on the surface of epithelial cells. The altered CFTR is thought to result in defects of electrolyte transport which then cause increased water reabsorption across respiratory epithelia. This may lead to dehydration of the airway surface liquid, which in turn may prevent normal clearance of mucus (Davis 1996), although the precise mechanism by which CFTR causes abnormal mucus is still unknown.

Description of the intervention

Improvement of sputum clearance is a major therapeutic aim in CF. Treatments to improve mucus clearance in CF include chest physiotherapy, with and without the addition of agents that enhance mucus clearance. Treatment with nebulised recombinant deoxyribonuclease (rhDNase) has been widely accepted to be of benefit in CF (Yang 2021), and is thought to exert its major effect by enhancing sputum clearance. However, treatment with rhDNase is relatively expensive and its use in most countries is restricted as a consequence. Hypertonic saline may represent a potential alternative or supplementary therapy to improve mucociliary clearance in the context of long‐term maintenance therapy or during times of acute worsening of lung disease in CF.

How the intervention might work

In vitro deposition of hypertonic saline onto the airway surface improves mucus clearance. Dasgupta demonstrated that the addition of 3% hypertonic saline improved measures of sputum clearance and that hypertonic saline had a greater effect on mucus clearance in vitro than rhDNase (Dasgupta 1995). The postulated molecular mechanism of this effect is as follows:

1. hypertonic saline breaks the ionic bonds within the mucus gel, which could reduce the degree of cross‐linking and entanglements and lower viscosity and elasticity (Ziment 1978);

2. with chronic infection the mucin macromolecules develop fixed negative charges, causing increased repulsion; the addition of hypertonic saline increases the ionic concentration of the mucus and causes a conformational change by shielding the negative charges and thereby reducing repulsion ‐ this would result in a more compact mucus macromolecule that would allow more effective clearance (Robinson 1997);

3. in addition hypertonic saline induces an osmotic flow of water into the mucus layer, rehydrating secretions and thereby improving mucus rheology (Robinson 1997).

Why it is important to do this review

In the long term, improvement in mucociliary function may reduce bacterial load and chronic inflammation within the airways and therefore reduce the decline in lung function that is consequent to this. Hypertonic saline is easy and inexpensive to produce. Therefore, it is important to determine if nebulised hypertonic saline improves outcomes in CF, and to determine the frequency of adverse effects. This is an update of a previously published review (Wark 1999; Wark 2000; Wark 2003; Wark 2009; Wark 2018).

Objectives

To investigate efficacy and tolerability of nebulised hypertonic saline treatment in people with CF compared to placebo or other treatments that enhance mucociliary clearance.

Methods

Criteria for considering studies for this review

Types of studies

We included controlled clinical trials, both random allocation and quasi‐random allocation (e.g. where there is alternate allocation to treatment and control groups).

Types of participants

People of all ages and of both sexes with CF diagnosed clinically or by sweat and genetic testing, including all degrees of disease severity, were eligible.

Types of interventions

We included studies of nebulised hypertonic saline (defined as any concentration of saline greater than or equal to 3% delivered via a mask or mouthpiece with a nebuliser pump) compared to either placebo or usual treatment or any other mucus‐mobilising treatments (including, but not limited to, physical airway clearance techniques and medications which demonstrate improved mucus clearance e.g. rhDNase) or other concentrations of hypertonic saline. The minimum treatment duration considered in this review was a single dose. We planned to consider trials comparing hypertonic saline used in conjunction with another intervention if the comparator group also received the second intervention.

Types of outcome measures

We planned to assess the following outcome measures.

Primary outcomes

1. Lung function (absolute change and change in per cent (%) predicted)

   1. forced expiratory volume at one second (FEV₁)

   2. forced vital capacity (FVC)

   3. lung volume (residual volume (RV) and total lung capacity (TLC))

   4. FEV_0.5

   5. lung clearance index (LCI)

2. Mortality

Secondary outcomes

1. Measures of sputum clearance (including measures of mucociliary clearance)

2. Measures of exercise capacity

3. Measures of quality of life (QoL) and symptoms

4. Pulmonary exacerbations (where a clear definition is described demonstrating an increase in symptoms or a decline in pulmonary function)

   1. frequency

   2. admission to hospital

   3. duration of hospital stay (post hoc change)

   4. outpatient treatments (hospital in the home, unscheduled visits to the doctor)

   5. use of antibiotics, either intravenous, oral or inhalational

5. Medication delivery time (minutes)

6. Cost of treatment

7. Adherence to treatment with hypertonic saline along with other treatments after hypertonic saline is added

8. Bacteriology in pulmonary secretions, including sputum culture, culture from cough swab or bronchial lavage (post hoc change)

9. Adverse effects such as bronchospasm, cough and acute decline in pulmonary function (acute decline was limited to the immediate phase of receiving treatment with hypertonic saline to within the first three hours, and described separately to longer‐term lung function data as it represents acute bronchospasm provoked by hypertonic saline)

Search methods for identification of studies

We searched for all relevant published and unpublished trials without restrictions on language, year or publication status.

Electronic searches

We identified relevant trials from the Cochrane Cystic Fibrosis and Genetic Disorders Group's Cystic Fibrosis Trials Register using the term: hypertonic saline.

The Cystic Fibrosis Trials Register is compiled from electronic searches of the Cochrane Central Register of Controlled Trials (CENTRAL) (updated each new issue of the Cochrane Library), weekly searches of MEDLINE, a search of Embase to 1995 and the prospective handsearching of two journals ‐ Pediatric Pulmonology and the Journal of Cystic Fibrosis. Unpublished work is identified by searching the abstract books of three major cystic fibrosis conferences: the International Cystic Fibrosis Conference; the European Cystic Fibrosis Conference and the North American Cystic Fibrosis Conference. For full details of all searching activities for the register, please see the relevant sections of the Group's website.

Date of the most recent search of the Group's Cystic Fibrosis Trials Register: 25 April 2022.

We also searched the following trials registries:

1. US National Institutes of Health Ongoing Trials Register Clinicaltrials.gov (www.clinicaltrials.gov; searched 18 May 2022);

2. World Health Organization International Clinical Trials Registry Platform (WHO ICTRP) (apps.who.int/trialsearch; searched 18 May 2022).

For details of our search strategies, please see Appendix 1.

Searching other resources

We checked the bibliographies of included trials and any relevant systematic reviews identified for further references to relevant trials.

Data collection and analysis

Selection of studies

The authors (PW, VMM, SS) independently selected the abstracts found during the searches. They then discussed potential and excluded abstracts to reach consensus. If trials were only in abstract form, the review authors contacted the trial authors for additional information. Two review authors then independently reviewed the full trials and, by consensus with the third author, included them if they were suitable or excluded them, documenting reasons for exclusion.

Data extraction and management

Two authors (PW, VMM or SS) independently extracted data on trial characteristics and results using standard data acquisition forms. The authors entered the data into the Review Manager software (RevMan 2014).

The authors considered data reported up to and including three months to be short term and data reported at over three months to be long term.

The authors obtained additional data for one trial from the original investigators (Dentice 2016). Where an author of this Cochrane Review was a co‐author on an included trial, a third party performed the data extraction and assessment of quality (both risk of bias and GRADE) for that trial. This occurred when both the current review authors were co‐investigators in the National Hypertonic Saline in Cystic Fibrosis Study trial, and Ashley Jones and a second person from the editorial base extracted the data and assessed the risk of bias (Elkins 2006a). This was also the case when one author (PW) was an author on one further trial, when VP and a second person from the editorial base extracted the data and assessed the risk of bias (Dentice 2016).

Assessment of risk of bias in included studies

Two authors assessed the risk of bias of each trial using the Cochrane risk of bias tool described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2017) (PW, VMM or SS). In particular, they examined details of the generation of allocation sequence, the concealment of treatment allocation schedule, whether the trial was blinded, whether intention‐to‐treat (ITT) analyses were possible from available data, and if the number of participants lost to follow‐up or subsequently excluded from the trial was recorded.

Measures of treatment effect

If the review authors found that trials did not use an ITT analysis, then they sought data on the number of participants with each outcome event, by allocated treated group, irrespective of adherence and whether the participant was later thought to be ineligible or otherwise excluded from treatment or follow‐up. Regarding dichotomous outcome measures, currently none of the trials report on mortality. For adverse event data, the authors have calculated a pooled estimate of the treatment effect for each outcome across the studies and determined the risk ratio (RR). For the outcome of an improvement of over 10% in FEV₁ reported by two cross‐over trials, the authors used the generic inverse variance to analyse the data and present the odds ratio (OR).

For continuous outcomes, the authors recorded either a mean change from baseline for each group or mean post‐treatment or post‐intervention values and standard deviation (SD) for each group. They calculated a pooled estimate of treatment effect for each of these individually by calculating the mean difference (MD) and 95% confidence intervals (CIs) where appropriate. Where the SD was not reported or available to use, we used the MD and 95% CI for each group to calculate the SDs.

The review authors reported costs of treatment narratively.

Unit of analysis issues

Where trials measured data longitudinally, the authors based the analysis on the final time point results. Methods do exist to carry out a meta‐analysis of aggregate longitudinal data, where individual patient data (IPD) are not available, but these are not available at the moment in RevMan.

For trials with a cross‐over design, at least one week was required to allow sufficient washout of effect, at least for the measures of short‐term outcomes. For these trials of cross‐over design, the authors planned to carry out the analysis using results from a paired analysis, as recommended by Elbourne 2002. This was only possible in one trial (Suri 2001). For the remaining cross‐over trials, the data that were provided in the trial report were not sufficient to carry out this type of analysis. For these trials, we chose to ignore the cross‐over design and treat the results from the two periods as if they were independent (Adde 2004; Amin 2010; Ballmann 1998; Chadwick 1997; Laube 2009; Riedler 1996; Robinson 1996; Robinson 1997; Robinson 1999; Weller 1980). Elbourne reported that using this approach is conservative, due to the fact that it ignores the within‐patient correlation (Elbourne 2002).

Dealing with missing data

The authors originally planned to include missing participants due to dropouts in an ITT analysis. The authors attempted to obtain any missing statistics (such as SDs or correlation coefficients) from the trial authors, or they obtained the original data and determined the statistics. The authors were only able to obtain additional data from two trials (Adde 2004; Dentice 2016). The review authors made two attempts to contact trial authors for missing data before accepting that the additional data would not be made available. However, if the trial authors contact us with data in the future, we will add the information to the review at the following update.

Assessment of heterogeneity

The authors tested for heterogeneity between studies using a standard Chi² test and I² statistic (Higgins 2003). The Chi² test is a statistical test for heterogeneity, whereas I² assesses the quantity of inconsistency across studies in the meta‐analysis. The authors accepted a P value of below 0.1. They used the following I² ranges to interpret heterogeneity:

1. 0% to 40%: might not be important;

2. 30% to 60%: may represent moderate heterogeneity;

3. 50% to 90%: may represent substantial heterogeneity;

4. 75% to 100%: considerable heterogeneity.

Assessment of reporting biases

Due to the chronic nature of the disease, in many CF trials investigators collect data longitudinally at different time points throughout the course of the trial. In all the trials the authors examined when data were collected during the trial and also which data were reported in the trial publication. If it appeared that time points were missing that the review authors would expect to have been reported (based on clinical and biologic plausibility), the review authors would have reported this. The authors also planned to assess publication bias by constructing funnel plots if they had been able to include a sufficient number of trials.

Data synthesis

The authors have used fixed‐effect analyses in this review. For future updates, when appropriate, where between‐trial variability is statistically significant, the authors plan to carry out random‐effects analyses.

Subgroup analysis and investigation of heterogeneity

For future updates, where possible, the authors plan to investigate heterogeneity using subgroup analysis if the I² statistic is over 40%; they will consider the following subgroups:

1. strength of hypertonic saline (comparing a concentration of 3% to 7% versus a concentration greater than 7%);

2. volume of hypertonic saline (less than 5 mL versus 5 mL to 10 mL versus more than 10 mL).

Sensitivity analysis

For a future update, when possible, the authors plan to perform a sensitivity analysis based on risk of bias of the trials, excluding those with a high risk of performance bias and including and excluding quasi‐randomised trials.

Summary of findings and assessment of the certainty of the evidence

In a post hoc change in line with current Cochrane guidance, at the 2018 update we added a summary of findings table for each comparison presented in the review (Table 1; Table 2; Table 3; Table 4; Table 5; Table 6; Table 7; Table 8). We selected the following seven outcomes to report (chosen based on relevance to clinicians and consumers).

1. FEV₁ (short‐term change (up to and including three months))

2. FEV₁ (long‐term change (longer than three months))

3. LCI

4. Mortality

5. Measures of sputum clearance

6. Pulmonary exacerbations

7. Adverse events

We determined the certainty of the evidence using the GRADE approach; and downgraded evidence in the presence of a high risk of bias in at least one trial, indirectness of the evidence, unexplained heterogeneity or inconsistency, imprecision of results, high probability of publication bias. We downgraded evidence by one level if we considered the limitation to be serious and by two levels if very serious.

Results

Description of studies

The trials included in this review were heterogenous in terms of age, severity of underlying lung disease, colonisation of micro‐organisms, other interventions, as well as the dose, timing and delivery of hypertonic saline.

Results of the search

The searches identified 61 potentially eligible trials (159 references). We included 24 trials (84 references) (Adde 2004; Amin 2010; Amin 2016; Ballmann 1998; Cardinale 2003; Chadwick 1997; Dentice 2016: Donaldson 2020; Elkins 2006a; Eng 1996; Gupta 2012; Laube 2009; Mainz 2015; NCT01355796; PRESIS 2019; Riedler 1996; Robinson 1996; Robinson 1997; Robinson 1999; Rosenfeld 2012; SHIP 2019; Singh 2020; Suri 2001; Weller 1980) and excluded 29 trials (59 references) (ACTRN12619001681145; ACTRN12621000855820; Aquino 2012; Brivio 2016; Brown 2010; Buonpensiero 2010; Corcoran 2017; DeCono 2008; Dentice 2012; Donaldson 2006; Elkins 2006b; EUCTR2007‐002707‐40‐BE; Genkova 1998; Grasemann 2013; IRCT20180307038994N1; IRCT20191112045413N2; IRCT20201017049055N1; King 1997; Kobylyansky 2000; NCT01094704; Nenna 2017; O'Neill 2017; Ros 2012; Ruiz de Valbuena Maiz 2012; San Miguel‐Pagola 2016; Tiddens 2022; Van Ginderdeuren 2008; Van Ginderdeuren 2011; Vanlaethem 2008). There are six trials (11 references) currently listed as 'Awaiting classification' until more information is available to allow a judgement regarding eligibility (Balinotti 2015; Dwyer 2013; Hofmann 1997; Lennox 2016; Palacio 2014; SIMPLIFY 2022) and two trials are ongoing (ISRCTN14081521; NCT02343445). The study flow chart is presented as a figure (Figure 1).

1.

Study flow chart

Included studies

There were 24 trials which met the inclusion criteria, with a total of 1318 participants (Adde 2004; Amin 2010; Amin 2016: Ballmann 1998; Cardinale 2003; Chadwick 1997; Dentice 2016: Donaldson 2020; Elkins 2006a; Eng 1996; Gupta 2012; Laube 2009; Mainz 2015; NCT01355796; PRESIS 2019; Riedler 1996; Robinson 1996; Robinson 1997; Robinson 1999; Rosenfeld 2012; SHIP 2019; Singh 2020; Suri 2001; Weller 1980). Of these, 18 were published as full papers, one was published only as a trial registration document (with results) (NCT01355796) and four were reported in abstract form only (Adde 2004; Cardinale 2003; Chadwick 1997; Laube 2009), but additional data were provided by one of these investigators (Adde 2004).

Trial design

There were 10 trials of parallel design (Cardinale 2003; Dentice 2016; Donaldson 2020; Elkins 2006a; Eng 1996; Gupta 2012; PRESIS 2019; Rosenfeld 2012; SHIP 2019; Singh 2020). There were 14 trials that were of cross‐over design (Adde 2004; Amin 2010; Amin 2016; Ballmann 1998; Chadwick 1997; Laube 2009; Mainz 2015, NCT01355796; Riedler 1996; Robinson 1996; Robinson 1997; Robinson 1999; Suri 2001; Weller 1980) and two of these had a four‐arm cross‐over design (Robinson 1997; Robinson 1999). A washout period was not stated in four cross‐over trials (Chadwick 1997; Laube 2009; Robinson 1996; Robinson 1997); there was no washout (interventions given on single days consecutively) in one trial (Mainz 2015). Where there was a washout period described, this ranged from one week (Amin 2016; NCT01355796) up to eight weeks (Weller 1980).

The number of participants varied between trials from 10 (Riedler 1996; Robinson 1997) to 321 (Rosenfeld 2012).

In total 10 trials were multicentre (Amin 2010; Amin 2016; Dentice 2016; Elkins 2006a; Eng 1996; Mainz 2015; PRESIS 2019; Rosenfeld 2012; SHIP 2019; Suri 2001) and seven trials were single centre (Adde 2004; Donaldson 2020; Gupta 2012; Laube 2009; NCT01355796; Riedler 1996; Singh 2020); it was unclear whether the remaining seven trials were multicentre or single centre. Seven trials were run in Australia (Dentice 2016; Elkins 2006a; Eng 1996; Riedler 1996; Robinson 1996; Robinson 1997; Robinson 1999). Six trials were run in Europe ‐ three trials were run in Germany (Ballmann 1998; Mainz 2015; PRESIS 2019), two in the UK (Suri 2001; Weller 1980) and one in Italy (Cardinale 2003). Two trials were run in Canada (Amin 2010; Amin 2016) and three trials in the USA (Donaldson 2020; Laube 2009; NCT01355796); a further two trials were run in centres across both Canada and the USA (Rosenfeld 2012; SHIP 2019). One trial was run in Brazil (Adde 2004), one in India (Gupta 2012), and one trial did not clearly state where it was run (Chadwick 1997).

Participants

The ages of participants ranged from one month (PRESIS 2019) to 56 years (Amin 2016), but details of age were not given in three studies (Ballmann 1998; Cardinale 2003; Chadwick 1997). Most studies only recruited participants over the age of five or six years, but three studies included only infants or preschool children, or a mix of both (PRESIS 2019; Rosenfeld 2012; SHIP 2019). Most of the studies recruited approximately equal numbers of males and females, although five studies included a disproportionate number of females (over 60% female) (Adde 2004; Amin 2010; Amin 2016; Riedler 1996; Suri 2001). One study included disproportionately fewer females overall and particularly in the intervention group (13.3%) (Gupta 2012).

Fifteen trials stated the diagnostic criteria for CF in the participants, which confirmed CF on the basis of a positive sweat chloride test or the presence of two common genetic mutations (Amin 2016; Dentice 2016; Donaldson 2020; Elkins 2006a; Eng 1996; Gupta 2012; Laube 2009; Mainz 2015; NCT01355796; PRESIS 2019; Rosenfeld 2012; SHIP 2019; Singh 2020; Suri 2001; Weller 1980). The remaining nine trials only stated that the participants had CF.

Selection by tolerance for hypertonic saline

Four trials stated they tested for tolerance to hypertonic saline (Dentice 2016; Elkins 2006a; Rosenfeld 2012; SHIP 2019). Two trials excluded those who were intolerant to their test dose of hypertonic saline (Rosenfeld 2012; SHIP 2019), and Elkins 2006a excluded participants who demonstrated bronchial reactivity following hypertonic saline defined by a fall in FEV₁ of 15% following tolerability testing. Three trials excluded those that had known intolerance of hypertonic saline at the screening visit (Donaldson 2020; NCT01355796; PRESIS 2019). Four trials stated that prior use of hypertonic saline was an exclusion criterion (Amin 2016; Ballmann 1998; Elkins 2006a; Rosenfeld 2012; Singh 2020). Additionally, two trials excluded participants who had previously used rhDNase (Ballmann 1998; Suri 2001).

Baseline microbiology

Baseline sputum microbiology was stated in 13 trials (Amin 2010; Amin 2016; Ballmann 1998; Elkins 2006a; Laube 2009; Mainz 2015; Riedler 1996; Robinson 1997; Robinson 1999; Rosenfeld 2012; Singh 2020; Suri 2001; Weller 1980). Weller 1980 mentioned bacterial growth, but no details were given. The Dentice 2016 trial measured change in bacterial density for Pseudomonas aeruginosa and Staphylococcus aureus, but did not state baseline microbiology.

P aeruginosa

The presence of P aeruginosa was described in 12 trials. In one trial, seven (37%) participants were described as colonised with P aeruginosa (Amin 2010), and 10 out of 12 participants in the later Robinson trial were colonised with P aeruginosa (Robinson 1999). In the later Amin trial, 17 out of 18 (94.4%) had a positive sputum culture for P aeruginosa (Amin 2016). In the Laube 2009 trial, P aeruginosa was cultured in 17% of participants. In the Rosenfeld 2012 trial 60 participants (38%) and 69 participants (42.3%) were colonised with P aeruginosa in the hypertonic saline group and isotonic saline groups, respectively. Ballmann 1998 reported that three out of the 14 participants were chronically colonised with P aeruginosa. Elkins 2006a reported the presence of P aeruginosa in 79 of the 83 participants in the hypertonic saline group and 78 of the 81 control participants. In each of two further trials, all 10 participants had P aeruginosa in their sputum (Riedler 1996; Robinson 1997). In the Suri 2001 trial, 48% of participants had P aeruginosa. Mainz 2015 reported the presence of P aeruginosa in 23 (33%) participants. Singh 2020 reported median log colony forming units (CFU) (min to max) at baseline but found no difference between the groups.

S aureus

The presence of S aureus was described in six trials. Laube 2009 reported that S aureus was cultured in 42% of participants. Elkins 2006a reported that S aureus was present in 44 of the 83 participants in the hypertonic saline group and 47of the 81 control participants. Robinson reported that 5 out of 10 participants in the 1997 trial and 7 out of 12 in the 1999 trial had S aureus (including two who also had P aeruginosa) (Robinson 1997; Robinson 1999). Suri 2001 reported that 39% of participants were colonised with S aureus. Singh 2020 reported median log CFU (min to max) at baseline, but found no difference between the groups.

Other pathogens

Three studies excluded participants if they were colonised with Burkholderia cepacia complex (Amin 2010; Elkins 2006a; Suri 2001). Amin 2010 further excluded any participant who had positive sputum cultures for non‐tuberculosis mycobacteria in the past year. Three trials reported that no participants in either group had B cepacia (Riedler 1996; Robinson 1997; Rosenfeld 2012).

Robinson also reported in the 1999 trial that 4 out of 12 participants had Aspergillus fumigatus (Robinson 1999), and Suri 2001 reported that 2% of participants were infected with Stenotrophomonas maltophilia. Mainz 2015 reported a wide range of pathogens.

Baseline clinical severity

Most trials recruited participants with stable disease; in one of the Robinson trials, it was clearly stated that people with CF who were clinically unstable (defined as an exacerbation in the previous four weeks) were excluded (Robinson 1999). Three further trials excluded participants who were experiencing or had recently experienced an acute respiratory exacerbation (Ballmann 1998; Rosenfeld 2012; SHIP 2019). Rosenfeld 2012 also excluded any participant with a secondary chronic lung condition not related to their CF, or other major organ dysfunction. Three trials required participants to be clinically stable (Donaldson 2020; Elkins 2006a; NCT01355796). Amin 2010 included only participants with a baseline FEV₁ of greater than 80% predicted and a room air oxyhaemoglobin saturation of greater than 90%. In the Eng 1996 trial, participants were required to have an FEV₁ greater than 20% predicted at baseline and to be on stable medications for the previous 14 days, and the participants in the later Amin trial were required to have an FEV1 of ≥ 40% predicted (Amin 2016). In the trial by Laube 2009, the children had FEV₁ and FVC greater than 90% predicted. In two trials by Robinson, participants needed to be in a stable clinical condition without any change to their medications (Robinson 1996; Robinson 1997); only the 1997 Robinson trial included a participant with an FEV₁ % predicted of less than 30% (Robinson 1997). Suri 2001 required participants to have an FEV₁ less than 70% predicted (people with CF with at least moderate lung disease) and be clinically stable with no exacerbations or change in medications in the last 14 days (consequently, these participants have more severe lung disease at baseline, mean FEV₁ % predicted 48% (range 14 to 77%)). Weller 1980 stated that all participants received routine treatment for five years. Others used mean FEV₁ as a % predicted value or FVC as a % predicted value to assess disease severity (Adde 2004; Cardinale 2003; Chadwick 1997). Mainz 2015 recruited participants with clinical symptoms of rhinosinusitis, but did not state the clinical severity of lung disease or whether they were stable clinically at the recruitment.

Three trials recruited participants who were experiencing an acute exacerbation (Dentice 2016; Riedler 1996; Singh 2020). Dentice 2016 enrolled participants with a confirmed diagnosis of CF within 24 hours of a hospital admission for management of a pulmonary exacerbation (defined as at least 4 out of 12 criteria described by Fuchs 1994) for a minimum of seven days). The Riedler 1996 trial selected 10 consecutive adolescents admitted with an exacerbation of their lung disease who all had productive coughs.

Interventions

An ultrasonic nebuliser was used to deliver hypertonic saline in nine trials (Amin 2010; Amin 2016; Eng 1996; Riedler 1996; Robinson 1996; Robinson 1997; Robinson 1999; Rosenfeld 2012; Suri 2001); while eight trials used a high‐output jet nebuliser (Adde 2004; Ballmann 1998; Dentice 2016; Elkins 2006a; Gupta 2012; Laube 2009; Singh 2020; Weller 1980). One trial used a hybrid ulltrasonic / jet nebulier, Pari LC Sprint Sinus nebuliser (Mainz 2015). One trial used an eFlow nebuliser (Donaldson 2020), and two used a PARI LC Sprint junior nebuliser (PRESIS 2019; SHIP 2019). One trial stated only that the inhalation was aerosolised (NCT01355796), and two trials did not state the type of nebuliser used (Cardinale 2003; Chadwick 1997).

Different concentrations of hypertonic saline were used in the trials, ranging from 3.5% to 7%, and this is outlined in detail in the tables (Characteristics of included studies).

In 17 trials, isotonic (0.9%) saline was used as a control (Amin 2010; Amin 2016; Cardinale 2003; Chadwick 1997; Dentice 2016; Donaldson 2020; Elkins 2006a; Eng 1996; Laube 2009; Mainz 2015; PRESIS 2019; Riedler 1996; Robinson 1996; Robinson 1997; Robinson 1999; Rosenfeld 2012; SHIP 2019). Four of these trials compared hypertonic saline 7% with isotonic saline 0.9% twice daily (Amin 2010; Elkins 2006a; Rosenfeld 2012; SHIP 2019) and two trials compared hypertonic saline 6% to isotonic saline twice daily (Eng 1996; PRESIS 2019) and a further trial compared hypertonic saline 6% to isotonic saline three times a day (Donaldson 2020). Two trials added quinine sulphate (0.25 mg/mL) to both solutions to mask the taste (Dentice 2016; Elkins 2006a). One trial administered 6% hypertonic saline or isotonic saline once per day, approximately 1 mL to each nostril (Mainz 2015). Six trials used a single administration of nebulised hypertonic saline compared to isotonic saline (Amin 2016; Laube 2009; Riedler 1996; Robinson 1996; Robinson 1997; Robinson 1999;). In the 1996 trial, Robinson compared a single administration of nebulised hypertonic saline (7%), amiloride (0.3% in 0.12% NaCl) and a combination of amiloride and hypertonic saline to isotonic saline (0.9%) (Robinson 1996), while in the 1997 trial Robinson compared differing concentrations of nebulised hypertonic saline (3%, 7%, and 12%) with isotonic saline and voluntary cough (Robinson 1997). In the 1999 trial, Robinson compared hypertonic saline 6% to 0.9% isotonic saline with matched voluntary cough, mannitol 300 mg, and placebo capsules with matched voluntary cough (Robinson 1999). In the Laube 2009 trial, participants attended for two visits at least one week apart. They received either 5 mL 0.12% isotonic saline or 5 mL 7% hypertonic saline, with the order of treatment randomised. Two trials did not state the frequency of nebulisation (Cardinale 2003; Chadwick 1997).

One trial compared 7% hypertonic saline with 3% hypertonic saline twice a day (Gupta 2012).

Three trials compared hypertonic saline to rhDNase (Adde 2004; Ballmann 1998; Suri 2001). Adde 2004 used a regimen of hypertonic saline 6% (10 mL) compared to 2.5 mg rhDNase twice daily; Ballmann 1998 compared nebulised 5.75% saline (10 mL) to 2.5 mg rhDNase twice daily; and Suri 2001 compared hypertonic saline 7% (5 mL) twice daily to rhDNase 2.5 mg daily or to rhDNase 2.5 mg alternate daily.

Weller 1980 compared hypertonic saline 7% (3 mL) to Mistabron® 20% (a mucolytic agent).

Two trials compared hypertonic saline 7% with inhaled xylitol twice a day (NCT01355796; Singh 2020).

Additional treatments were also used in association with the hypertonic saline. With the exception of three trials (Chadwick 1997; Mainz 2015; Weller 1980), all trials pretreated participants with short‐acting beta‐agonists. In the Suri 2001 trial, the only pretreated participants were those who were already using bronchodilators or whose FEV₁ fell by more than 15% after the test dose of hypertonic saline. Pretreatment was not stated by Laube 2009.

In the Eng 1996 trial participants performed physiotherapy at home and received hypertonic saline or isotonic saline prior to their regular physiotherapy session. The place of chest physiotherapy is likely to be an important contributor to mucolytic therapy, but its role as a confounder was not addressed.

Two trials used hypertonic saline or isotonic saline as an adjunct to physiotherapy and an exercise programme while hospitalised for a pulmonary exacerbation when all participants also received intravenous antibiotics (Dentice 2016; Riedler 1996). In one trial participants received hypertonic saline 7%, three times a day (Dentice 2016), and in the second trial they received hypertonic saline 6% as a single treatment (Riedler 1996).

Outcomes

Lung function was the most common outcome measured, but this was reported in a number of ways. Fifteen trials reported on FEV₁ (Adde 2004; Amin 2010; Amin 2016; Ballmann 1998; Dentice 2016; Donaldson 2020; Elkins 2006a; Eng 1996; Gupta 2012; NCT01355796; Riedler 1996; Robinson 1996; SHIP 2019; Singh 2020; Suri 2001); eight trials reported on FVC (Amin 2010; Dentice 2016; Elkins 2006a; Eng 1996; Gupta 2012; Riedler 1996; Suri 2001; Weller 1980); and four reported on FEF_25-75 (Amin 2010; Amin 2016; Riedler 1996; Rosenfeld 2012). Four trials measured LCI (Amin 2010; Amin 2016; PRESIS 2019; SHIP 2019). A number of trials reported on less common measures of lung function (see the tables for further details (Characteristics of included studies) and one trial simply reported on general lung function (Cardinale 2003). Mucociliary clearance was reported in six trials (Donaldson 2020; Laube 2009; Riedler 1996; Robinson 1996; Robinson 1997; Robinson 1999) and two reported on sputum production (Riedler 1996; Weller 1980). Pulmonary exacerbations were reported in eight trials (Dentice 2016; Elkins 2006a; NCT01355796; PRESIS 2019; Rosenfeld 2012; SHIP 2019; Singh 2020; Suri 2001) and Rosenfeld 2012 additionally reported additional antibiotics for all causes. Four trials reported sputum cultures (Adde 2004; PRESIS 2019; Rosenfeld 2012; SHIP 2019) and three trials reported changes in quantitative microbiology (Dentice 2016; NCT01355796; Singh 2020). A further seven trials reported symptom scores (Adde 2004; Dentice 2016; Eng 1996; NCT01355796; SHIP 2019; Singh 2020; Weller 1980), and two reported on satisfaction or preference (Adde 2004; Ballmann 1998). One trial reported on sinus and nasal symptoms (Mainz 2015). Tolerability was only reported on by one trial (Rosenfeld 2012), and one trial reported on nebulisation time (Ballmann 1998). Linked to this, two trials reported adherence to treatment (Rosenfeld 2012; Suri 2001). Ten trials reported on QoL (Amin 2010; Dentice 2016; Donaldson 2020; Elkins 2006a; NCT01355796; PRESIS 2019; Rosenfeld 2012; SHIP 2019; Singh 2020; Suri 2001), and eight trials on adverse events (Cardinale 2003; Dentice 2016; Eng 1996; Mainz 2015; NCT01355796; PRESIS 2019; Rosenfeld 2012; Singh 2020). Two trials reported data relating to cost in comparison to rhDNase (Ballmann 1998; Suri 2001).

Excluded studies

We excluded 29 trials from the review (ACTRN12619001681145; ACTRN12621000855820; Aquino 2012; Brivio 2016; Brown 2010; Buonpensiero 2010; Corcoran 2017; DeCono 2008; Dentice 2012; Donaldson 2006; Elkins 2006b; EUCTR2007‐002707‐40‐BE; Genkova 1998; Grasemann 2013; IRCT20180307038994N1; IRCT20191112045413N2; IRCT20201017049055N1; King 1997; Kobylyansky 2000; NCT01094704; Nenna 2017; O'Neill 2017; Ros 2012; Ruiz de Valbuena Maiz 2012; San Miguel‐Pagola 2016; Tiddens 2022; Van Ginderdeuren 2008; Van Ginderdeuren 2011; Vanlaethem 2008).

Five trials were not randomised in design and were therefore excluded (DeCono 2008; EUCTR2007‐002707‐40‐BE; IRCT20180307038994N1; IRCT20201017049055N1; NCT01094704) and three were excluded as there was no comparison group (Dentice 2012; Genkova 1998; IRCT20191112045413N2). Two trials were cross‐over trials which had an insufficient washout period (Aquino 2012; Nenna 2017). One trial was excluded as it was undertaken in a non‐CF population (Kobylyansky 2000) and a further trial was performed in vitro (King 1997). Three trials were excluded as they studied hypertonic saline in conjunction with other therapies, but did not include a comparator group without hypertonic saline (Brivio 2016; Buonpensiero 2010; Ros 2012). One trial compared hypertonic saline with or without pretreatment with amiloride (Donaldson 2006). Six trials did not compare hypertonic saline to control, instead they compared different airway clearance techniques used in conjunction with hypertonic saline (ACTRN12619001681145; ACTRN12621000855820; San Miguel‐Pagola 2016; Van Ginderdeuren 2008; Van Ginderdeuren 2011; Vanlaethem 2008) and one did not use hypertonic saline but rather compared L‐arginine with isotonic saline (Grasemann 2013). Four trials compared different formulations or timings of hypertonic saline, or both (Brown 2010; Elkins 2006b; O'Neill 2017; Ruiz de Valbuena Maiz 2012) and one trial looked at the effect of hypertonic saline on structural lung disease via MRI (Tiddens 2022). We excluded the final trial because the hypertonic saline was given overnight via nasal cannulae rather than by facemask or nebuliser (Corcoran 2017).

Studies awaiting classification

There are six trials currently listed as 'Awaiting classification' until more information is available to allow a judgement regarding eligibility (Balinotti 2015; Dwyer 2013; Hofmann 1997; Lennox 2016; Palacio 2014; SIMPLIFY 2022).

One trial is a prospective, double‐blind randomised controlled trial (RCT) that recruited 19 participants (Balinotti 2015). The intervention group received hypertonic saline 3% plus 0.25 mg/kg salbutamol twice daily compared to the control group, which received isotonic saline 0.9% plus 0.25 mg/kg salbutamol twice daily for a year. The stated outcomes are maximal flow at functional residual capacity (change from baseline), pulmonary exacerbations, respiratory rate, nutritional status and adverse events. This trial is only published as an abstract at present.

The Dwyer 2013 trial is a three‐armed RCT of parallel design with a duration of 16 weeks. Participants were randomised to either inhalation of 4 mL of nebulised 6% hypertonic saline plus 0.25 mg/mL quinine sulphate twice a day, inhalation of 4 mL of nebulised 3% hypertonic saline plus 0.25 mg/mL quinine sulphate twice a day or the control intervention of inhalation of 4 mL of nebulised 0.9% isotonic saline pus 0.25mg/mL quinine sulphate twice a day. The primary outcome was lung function measured by the change in FEV₁ % predicted, and secondary outcomes included other measures of lung function, QoL, exercise capacity, exacerbations, adverse events and sputum bacterial diversity.

The Hofmann 1997 trial is an open‐label, single‐dose, RCT that included 20 participants. Participants were randomised to either amiloride in hypertonic saline 5.58% or amiloride in isotonic saline, and the primary outcome was change in ion flow by nasal potential difference. We have been unable to find any further information than an abstract.

A further trial listed as awaiting classification is a trial registration document for a cross‐over RCT comparing inhaled solution of 8.4% hypertonic bicarbonate by nebuliser with inhaled solution of 7% hypertonic saline by nebuliser (Lennox 2016). The primary outcome was the change in exhaled breath condensate pH, and secondary outcomes included change in expectorated sputum and change in spirometry. All outcomes were measured at four hours and after two inhalations on a single day. We have been unable to find any further information or publications relating to the trial .

The fifth trial is an open‐label, parallel group, RCT of parallel design which ran for 24 weeks (Palacio 2014). Investigators enrolled 27 children aged three to six years and randomised them to either hypertonic saline 7% nebulised twice a day or isotonic saline 0.9% nebulised twice a day. The primary outcome was lung function with secondary outcomes measuring respiratory symptoms, new isolation of P aeruginosa, rate of exacerbations, adverse events and adherence.

The final trial listed as awaiting assessment is a multicentre stopping trial where participants are randomised to stopping treatment or continuing (SIMPLIFY 2022). Investigators randomised clinically stable children (aged 12 years or over) and adults, who have been taking triple therapy (Kaftrio™) (elexacaftor/tezacaftor/ivacaftor) for at least 90 days, to either discontinuing hypertonic saline therapy or continuing it over a period of six weeks. The primary outcome is absolute change in FEV₁ % predicted at week 6; secondary outcomes are adverse events and absolute change in respiratory symptoms.

Ongoing studies

The review authors identified two ongoing trials from trials registries (ISRCTN14081521; NCT02343445). Both were described as randomised controlled trials of parallel design. 

One three‐arm trial is comparing P‐1037 solution for inhalation in either hypertonic saline 4.2% or in 0.17% isotonic saline to placebo (0.17% saline). The trial is recruiting children aged 12 or over who have been clinically stable for at least two weeks before enrolment. All treatments are inhaled twice daily over a period of 15 days. The primary outcomes are adverse events and FEV₁ change from pre‐dosing to one hour post‐dosing. Secondary outcomes measure FEV₁ (absolute change from baseline to Day 15); FVC (absolute change from baseline to Day 15); Cystic Fibrosis Questionnaire ‐ Revised (CFQ‐R) (Quittner 2009); and forced expiratory flow at 25% to 75% of FVC (FEF_25%-75%) (absolute change from baseline to Day 15) (NCT02343445).

The second ongoing trial is a stopping trial similar to the SIMPLIFY 2022 trial above and known as CF STORM (ISRCTN14081521). The CF STORM trial includes adults and children aged 12 or over with CF who are established on triple therapy (Kaftrio™) for more than three months. Participants are randomised to either continue muco‐active therapy (includes rhDNase and hypertonic saline) or discontinuing it for 52 weeks. When we are able to include this trial we will need to obtain IPD to look at the effects of stopping hypertonic saline versus continuing as opposed to those who stopped rhDNase. The primary outcome is change in FEV₁ % predicted at 52 weeks, whilst secondary outcomes report change in FEF_25-75 % predicted at 52 weeks; need for extra antibiotic treatment (number of courses and number of days of extra oral, IV, or nebulised antibiotics) at 52 weeks; need for extra chronic medications; number and proportion of positive respiratory cultures for significant pathogens; hospital admission; change in nutritional status; number of pulmonary exacerbations; change in QoL using CFQ‐R; adverse events; and costs to the UK National Health Service (ISRCTN14081521).

Risk of bias in included studies

Please refer to the risk of bias figures (Figure 2; Figure 3).

2.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies

3.

Risk of bias summary table: review authors' judgements for each risk of bias item for each study

Allocation

Generation of the randomisation sequence

No details of the randomisation process were provided by 10 trials (Adde 2004; Cardinale 2003; Chadwick 1997; Laube 2009; Mainz 2015; NCT01355796; Robinson 1996; Robinson 1997; Robinson 1999; Weller 1980). For nine of these we judged the risk of bias to be unclear; however, additional data received from Dr Adde confirmed that a random numbers table was used to generate the randomisation sequence (Adde 2004), as such we judged this trial to have a low risk of bias.

We also judged a further 13 trials to have a low risk of bias (Amin 2010; Amin 2016; Ballmann 1998; Donaldson 2020; Elkins 2006a; Eng 1996; Gupta 2012; PRESIS 2019; Riedler 1996; Rosenfeld 2012; SHIP 2019; Singh 2020; Suri 2001). Computer‐generated randomisation lists were used in seven trials (Amin 2010; Amin 2016; Donaldson 2020; Elkins 2006a; Gupta 2012; PRESIS 2019; Singh 2020), two trials used random permuted‐block allocation (Rosenfeld 2012; SHIP 2019), and one reported that participants drew lots to decide treatment (Ballmann 1998). Eng 1996 stated the use of random number tables. Riedler 1996 used a coin toss to randomise participants. Suri 2001 used telephone randomisation to an independent trials co‐ordinating unit, stratified by hospital and balanced after each group of 12 children.

We judged there to be a high risk of bias in the Dentice 2016 trial as participants were enrolled sequentially upon admission to hospital.

Allocation concealment

In 11 trials, investigators did not report any details regarding methods of allocation concealment (Adde 2004; Ballmann 1998; Cardinale 2003; Chadwick 1997; Eng 1996; Laube 2009; Mainz 2015; NCT01355796; Robinson 1996; Robinson 1997; Weller 1980). We judged 10 of these to have an unclear risk of bias. Additional data received from Dr Adde confirmed that the sequence of treatment was put into numbered envelopes which were kept in the hospital pharmacy and not opened until after participants were recruited (Adde 2004); thus we judged this trial to have a low risk of bias.

We judged a further 12 trials to have a low risk of bias (Amin 2010; Amin 2016; Donaldson 2020; Elkins 2006a; Gupta 2012; PRESIS 2019; Riedler 1996; Robinson 1999; Rosenfeld 2012; SHIP 2019; Singh 2020; Suri 2001). Ten trials concealed the allocation sequence either by using investigators off‐site, investigators not otherwise involved in the trial or a secure website (Amin 2010; Amin 2016; Gupta 2012; Elkins 2006a; Gupta 2012; PRESIS 2019; Rosenfeld 2012; SHIP 2019; Singh 2020; Suri 2001). Randomisation was coded such that investigators were blinded to the identity of the intervention at the time of analysis in one trial (Robinson 1999), and in another each participant was assigned to order of treatment by a coin toss (Riedler 1996).

One trial used sequential alternate allocation, and thus we did not judge the allocation to be adequately concealed (high risk of bias) (Dentice 2016).

Blinding

Five trials  were described as "blinded" and we judged them to be at a low risk of bias (Dentice 2016; Elkins 2006a; Gupta 2012; PRESIS 2019; SHIP 2019); two of these trials reported adding quinine sulphate to the solutions to mask the taste  (Dentice 2016; Elkins 2006a).

We judged four trials to have an unclear risk of bias (Amin 2016; Cardinale 2003; Donaldson 2020; Singh 2020). Cardinale 2003 did not provide any detail to allow us to make a decision as to whether the participants were blinded, whilst the remaining three trials were described as blinded, but admitted that participants may have been able to discern which group they were in due to the taste and duration of nebulisation for the different interventions (Amin 2016; Donaldson 2020; Singh 2020).

We judged the remaining 15 trials to have a high risk of bias due to the discernible taste of hypertonic saline (Adde 2004; Amin 2010; Ballmann 1998; Chadwick 1997; Eng 1996; Laube 2009; Mainz 2015; NCT01355796; Riedler 1996; Robinson 1996; Robinson 1997; Robinson 1999; Rosenfeld 2012; Suri 2001; Weller 1980). The Adde 2004 trial attempted to blind participants by adding quinine sulphate to the solutions to mask the taste, but additional information stated that participants were well aware of their allocation as the taste of hypertonic saline could not be masked. Chadwick 1997 described their trial as single‐blind and did not address the issue of the taste of hypertonic saline, and one trial was open‐label (NCT01355796).

Regarding the blinding of the investigators, 11 trials reported that researchers were blinded and so had a low risk of bias (Amin 2010; Amin 2016; Dentice 2016; Donaldson 2020; Elkins 2006a; Eng 1996; Gupta 2012; PRESIS 2019; Robinson 1999; Rosenfeld 2012; SHIP 2019). A lack of information led us to judge 12 trials as having an unclear risk of bias (Adde 2004; Ballmann 1998; Cardinale 2003; Chadwick 1997; Laube 2009; Mainz 2015; Riedler 1996; Robinson 1996; Robinson 1997; Singh 2020; Suri 2001; Weller 1980). Riedler 1996 and Chadwick 1997 were described as single‐blind and stated that participants could discern the taste of hypertonic saline, thus implying that the researchers were blinded, and Weller 1980 was described as double‐blind, although the participants could discern the taste, but there were no definite statements on the blinding of trial investigators. We judged one trial to be at high risk of bias with regard to blinding of investigators as the trial was open‐label (NCT01355796).

Incomplete outcome data

We judged six trials to have an unclear risk of bias due to incomplete outcome data (Ballmann 1998; Chadwick 1997; Riedler 1996; Robinson 1997; Robinson 1999; Suri 2001). Ballmann 1998 provided no information about whether an ITT analysis was used; three trials did not state as whether an ITT approach had been used and did not describe any withdrawals (Riedler 1996; Robinson 1997; Robinson 1999). The Chadwick 1997 trial had no description of dropouts. Suri 2001 provided details of withdrawals with reasons and used an ITT analysis, but an additional report of airway inflammatory changes following treatment stated that only 28 of the 48 participants were able to perform induced sputum and be included for these outcomes.

We judged 17 trials to be at low risk (Adde 2004; Amin 2010; Amin 2016; Dentice 2016; Donaldson 2020; Elkins 2006a; Eng 1996; Gupta 2012; Laube 2009; Mainz 2015; NCT01355796; PRESIS 2019; Robinson 1996; Rosenfeld 2012; SHIP 2019; Singh 2020; Weller 1980). Three trials stated there were no withdrawals (Dentice 2016; NCT01355796; Robinson 1996). The remaining trials gave details of withdrawals and the reasons for these. Adde 2004 did not report withdrawals; additional data provided by the trial investigators stated that one participant (not included in the analysis) had to stop treatment with hypertonic saline due to severe dyspnoea during its nebulisation. Amin 2010 described details of missing data for three participants; two due to uninterpretable LCI results and one due to an inability to adhere to the trial protocol. Amin 2016 trial excluded three participants but analysed the remaining 18 participants on an ITT basis. Similarly, Donaldson 2020 lost three of the 23 randomised participants, but full reasons were given. In the Elkins 2006a trial, two participants (one from each group) withdrew voluntarily after randomisation and before the first dose; a clear description of withdrawals after randomisation was given by group and with reasons, with a total of 82 participants in the hypertonic saline group and 80 in control group included in an ITT analysis. Gupta 2012 reported that only one participant in the 3% hypertonic saline group was lost to follow‐up. Laube 2009 described two dropouts following randomisation and accounted for this by replacing them. Mainz 2015 detailed six dropouts due to non‐adherence. Only two participants withdrew from the PRESIS 2019 trial, one in each group, with full reasons given. Rosenfeld 2012 gave details (with reasons) of withdrawals, which were in roughly equal numbers across groups, and reported an ITT analysis of 158 participants in the hypertonic saline group and 163 in the control group. In the SHIP 2019 trial, 10 out of 150 participants did not complete the trial; this was less than 15% and investigators supplied full reasons. Only one participant dropped out of the Singh 2020 trial, the remaining 59 participants completed the trial and were included in the analysis. Two trials gave details of withdrawals, but did not state whether an ITT analysis had been performed (Eng 1996; Weller 1980). In the Eng 1996 trial, six participants withdrew in total (three from each group) with reasons given. In the Weller 1980 trial there was a clear description of dropouts and withdrawals.

Cardinale 2003 stated that no adverse events were reported with hypertonic saline, but did not give any information for the placebo group. We judged this trial to be at high risk of bias for this domain.

Selective reporting

Three trials were reported only in abstract form, thus limiting the availability of data concerning the outcomes which were planned to have been reported; and we judged these to have an unclear risk of bias (Adde 2004; Cardinale 2003; Chadwick 1997). A further two trials did not report all the outcomes listed in the methods, and we also deemed these to be at an unclear risk of bias (Gupta 2012; PRESIS 2019). In the remaining 19 trials, the outcomes stated in the 'Methods' section were reported in the 'Results' section (Amin 2010; Amin 2016; Ballmann 1998; Dentice 2016; Donaldson 2020; Elkins 2006a; Eng 1996; Laube 2009; Mainz 2015; NCT01355796; Riedler 1996; Robinson 1996; Robinson 1997; Robinson 1999; Rosenfeld 2012; SHIP 2019; Singh 2020; Suri 2001; Weller 1980).

Other potential sources of bias

We found descriptions of sample size calculations in 13 trials (Amin 2010; Amin 2016; Dentice 2016; Donaldson 2020; Elkins 2006a; Eng 1996; Gupta 2012; Laube 2009; PRESIS 2019; Rosenfeld 2012; SHIP 2019; Singh 2020; Suri 2001); however, the remaining 11 trials did not undertake such calculations (Adde 2004; Ballmann 1998; Cardinale 2003; Chadwick 1997; Mainz 2015; NCT01355796; Riedler 1996; Robinson 1996; Robinson 1997; Robinson 1999; Weller 1980).

In total 14 trials used a cross‐over design, but six of these either did not use or did not state a washout period between treatment arms (Chadwick 1997; Laube 2009; Riedler 1996; Robinson 1996; Robinson 1997; Robinson 1999), and we judged the risk of bias to be unclear. Eight trials did report the washout period; the stated washout periods were one week (Amin 2016; NCT01355796), two weeks (Adde 2004; Suri 2001), three weeks (Ballmann 1998), four weeks (Amin 2010; Mainz 2015) and eight weeks (Weller 1980). We judged these eight trials to be at low risk of bias.

Effects of interventions

See: Table 1; Table 2; Table 3; Table 4; Table 5; Table 6; Table 7; Table 8

We graded the certainty of the evidence for those outcomes included in the summary of findings tables. For the definitions of these gradings, please refer to the summary of findings tables (Table 1; Table 2; Table 3; Table 4; Table 5; Table 6; Table 7; Table 8).

Hypertonic saline 3% to 7% versus isotonic saline in stable lung disease

This comparison included 15 trials (n = 931) (Amin 2010; Amin 2016; Cardinale 2003; Chadwick 1997; Donaldson 2020; Elkins 2006a; Eng 1996; Laube 2009; Mainz 2015; PRESIS 2019; Robinson 1996; Robinson 1997; Robinson 1999; Rosenfeld 2012; SHIP 2019). A summary of the results and our judgements with regard to the certainty of the evidence can be found in the table (Table 1).

In the analysis, due to data limitations, we have entered and analysed data from cross‐over trials as if they were from parallel trials. One trial assessed children with stable lung disease and chronic sinuitis and looked at the effect of the intervention on sinus symptoms (Mainz 2015). 

We report results from one trial narratively; this trial has been published as an abstract only, and we have been unable to ascertain the duration of the trial or intervention (Chadwick 1997).

Primary outcomes

1. Lung function

a. FEV₁

Four trials (246 participants) examined the effect of hypertonic saline 3% to 7% compared to isotonic saline on the change in FEV₁ % predicted after two to four weeks treatment (Amin 2010; Donaldson 2020; Elkins 2006a; Eng 1996). We are uncertain if there was an improvement in FEV₁ % predicted in the hypertonic saline group compared to isotonic saline (MD 3.30%, 95% CI 0.71 to 5.89; 4 trials, 246 participants; very low‐certainty evidence; Analysis 1.1).

1.1. Analysis.

Comparison 1: Hypertonic saline 3% to 7% versus isotonic saline, Outcome 1: Change in FEV₁ (% predicted)

Amin 2010 selected 20 children aged 6 to 18 years old with essentially normal lung function as measured by spirometry with an FEV₁ greater than 80% predicted and found no difference compared to isotonic saline (MD ‐0.42%, 95% CI ‐7.45 to 6.61; Analysis 1.1).

The two earlier trials recruited participants with greater impairment in lung function (Elkins 2006a; Eng 1996). When the data from these two trials were pooled and analysed, there was an improvement in FEV₁ % predicted at four weeks compared to control (MD 4.15%, 95% CI 1.14 to 7.16; 2 trials, 205 participants; Analysis 1.1). The I² value at this time point for all participants is 51% (according to our definition this may just represent substantial heterogeneity); however, if the Eng 1996 trial is removed from the meta‐analysis I² reverts to zero. The participants in the Eng 1996 trial had a lower FEV₁ % predicted at baseline than in the other trials, which may account for this difference. Similarly, for the analysis of data for participants aged over 14 years, which combines data for just two trials (Elkins 2006a; Eng 1996), the lower baseline lung function values of the participants in the Eng 1996 trial could account for this heterogeneity.

Only Elkins 2006a examined the mean % change in FEV₁ % predicted at time points after four weeks (4, 12, 24, 36 and 48 weeks). Trial results failed to demonstrate a benefit over isotonic saline at any time‐point, other than at 24 weeks (MD 5.37%, 95% CI 1.03 to 9.71; 1 trial, 140 participants; low‐certainty evidence).

Cardinale 2003 stated that the trial measured lung function, but did not report any results other than to state there was no difference in lung function results.

Mainz 2015 also measured FEV₁% predicted at day 1 and day 29, with no differences seen between the groups. However, as hypertonic saline was not delivered to the airways, these outcomes are not included in our analysis (Analysis 1.1).

One trial was a single‐dose trial which measured FEV₁ % predicted 24 hours after the single dose and found no difference between treatments (MD 0.30, 95% CI ‐3.48 to 4.08; 1 trial, 18 participants; Analysis 1.1) (Amin 2016).

b. FVC

We analysed data from three trials for the mean change in FVC % predicted at four weeks (Amin 2010; Elkins 2006a; Eng 1996). When the data were pooled they did not demonstrate a difference in FVC (MD 1.07%, 95% CI ‐1.63 to 3.78; 3 trials, 225 participants; Analysis 1.2).

1.2. Analysis.

Comparison 1: Hypertonic saline 3% to 7% versus isotonic saline, Outcome 2: Change in FVC (% predicted)

Again, only Elkins 2006a measured the change in FVC % predicted at 48 weeks (134 participants). There was no difference between groups, although the control group showed no improvement over this period of time (Analysis 1.2).

c. Lung volumes

These were not reported as outcomes in any of the included trials.

d. Change in FEV_0.5 and FEV_0.75

Rosenfeld 2012 assessed infant lung function in a subset of participants (73 out of 158), which included FEV_0.5, and found a small difference (but with wide CIs) following treatment with hypertonic saline compared to isotonic saline at 48 weeks (MD 41.00 mL, 95% CI 0.96 to 81.04; Analysis 1.3).

1.3. Analysis.

Comparison 1: Hypertonic saline 3% to 7% versus isotonic saline, Outcome 3: Mean change in FEV_0.5 (mL)

A further trial (150 participants) using infant lung function measured change in FEV_0.75 L/min (SHIP 2019); investigators found no difference at 12 weeks, 24 weeks, 36 weeks or 48 weeks (Analysis 1.4).

1.4. Analysis.

Comparison 1: Hypertonic saline 3% to 7% versus isotonic saline, Outcome 4: Mean change in FEV_0.75 (L)

e. LCI

Four trials used LCI as their primary outcome (Amin 2010; Amin 2016; PRESIS 2019; SHIP 2019). In the Amin 2010 trial, the original trial investigators' analysis found hypertonic saline improved LCI compared to isotonic saline at four weeks, although in our analysis we are uncertain whether there was a difference between groups (MD ‐1.03, 95% CI ‐2.76 to 0.70; 1 trial, 19 participants; very low‐certainty evidence; Analysis 1.5).

1.5. Analysis.

Comparison 1: Hypertonic saline 3% to 7% versus isotonic saline, Outcome 5: LCI

The Amin 2016 trial was a single‐dose study which reported change from baseline in LCI and found a treatment effect of ‐0.60 (standard error (SE) 0.32) between groups.

The PRESIS 2019 and SHIP 2019 trials both reported change in LCI in a preschool population at 12 months. Our analysis shows a small improvement in LCI favouring the hypertonic saline group where a fall in LCI indicates better lung function (MD ‐0.60, 95% CI ‐1.00 to ‐0.19; 2 trials, 190 participants; Analysis 1.6). 

1.6. Analysis.

Comparison 1: Hypertonic saline 3% to 7% versus isotonic saline, Outcome 6: LCI change from baseline

2. Mortality

Only one trial (23 participants) reported all cause mortality as an outcome, and there were no deaths in either group (Donaldson 2020).

Secondary outcomes

1. Measures of sputum clearance

Four trials used radio‐labelled aerosol clearance to assess mucociliary clearance (Laube 2009; Robinson 1996; Robinson 1997; Robinson 1999), and we judged the certainty of the evidence to be very low. In this method the participant was given the radio‐labelled aerosol from an ultrasonic nebuliser and serial lung scans were performed. Two of the Robinson trials showed that hypertonic saline increased radioisotope clearance compared to isotonic saline controls, P < 0.05 (Robinson 1996) and P < 0.01 (Robinson 1997). The Robinson 1997 trial showed that increasing concentrations of hypertonic saline also had an effect, with a difference between hypertonic saline 3% and hypertonic saline 12% favouring the higher concentration; but no difference between hypertonic saline 7% and hypertonic saline 12% was reported (Robinson 1997). A comparison was made between three trials for per cent isotope clearance at 60 to 120 minutes (Laube 2009; Robinson 1997; Robinson 1999); we are uncertain whether treatment with hypertonic saline improved isotope clearance (MD 6.14%, 95% CI 2.56 to 9.72; 3 trials, 68 participants; very low‐certainty evidence; Analysis 1.7).

1.7. Analysis.

Comparison 1: Hypertonic saline 3% to 7% versus isotonic saline, Outcome 7: Radiolabelled isotope clearance (%)

Two of the Robinson trials reported measuring mucociliary clearance as 'area under the curve' (AUC), where the lower the value of AUC is, the faster the clearance, (Robinson 1996; Robinson 1997) and showed that hypertonic saline 7% and hypertonic saline 12% improved mucociliary clearance compared to isotonic saline. In the Robinson 1996 trial, the results for AUC showed hypertonic saline and hypertonic saline with amiloride were significantly different from cough, isotonic saline and amiloride alone. Combined analysis of the two trials favoured treatment (MD ‐212.06, 95% CI ‐271.64 to ‐152.48; 2 cross‐over trials, 22 participants; Analysis 1.8).

1.8. Analysis.

Comparison 1: Hypertonic saline 3% to 7% versus isotonic saline, Outcome 8: Mucociliary clearance measured as area under the curve

The participants in the Robinson trials had moderate to severe airflow obstruction; those in the Robinson 1996 trial had a mean FEV₁ % predicted of 60.8% (range 27% to 112%); those in the Robinson 1997 trial had a mean FEV₁ % predicted of 52% (range 31% to 84%); and in the Robinson 1999 trial, the mean FEV₁ was 60% (range 27% to 112%). This differed from the Laube 2009 trial, who only included participants with normal lung function (greater than 90% predicted FEV₁ and FVC). In this trial there was no difference in mucociliary clearance measured at 20, 60, 90 minutes and 24 hours.

One trial used gamma scintigraphy to measure the mucociliary clearance rate (% clearance) after treatment at 28 days (MD 5.12%, 95% CI 1.07 to 9.17; 1 trial, 20 participants; Analysis 1.9) (Donaldson 2020).

1.9. Analysis.

Comparison 1: Hypertonic saline 3% to 7% versus isotonic saline, Outcome 9: Change in mucocilliary clearance rate (% clearance)

2. Measures of exercise capacity

Eng 1996 demonstrated an improvement in exercise tolerance; the paper used a visual analogue scale (VAS) and during week 1 reported a mean (SD) improvement with hypertonic saline of 2.05 (1.3) and with isotonic saline of 1.7 (1.25) (P = 0.015); during week 2 the mean (SD) rise with hypertonic saline was 2.76 (1.45) and with isotonic saline 1.75 (1.6) (P = 0.02) (Eng 1996). When analysed, these data favour hypertonic saline both at week 1 (MD 0.88, 95% CI 0.19 to 1.57; 1 trial, 53 participants; Analysis 1.10) and at week 2 (MD 1.01, 95% CI 0.18 to 1.84; 1 trial, 53 participants; Analysis 1.10). 

1.10. Analysis.

Comparison 1: Hypertonic saline 3% to 7% versus isotonic saline, Outcome 10: Exercise capacity (using a subjective visual analogue score)

3. Measures of QoL and symptom scores

Five trials (678 participants) assessed health‐related QoL (Amin 2010; Donaldson 2020; Elkins 2006a; Rosenfeld 2012; SHIP 2019). One trial assessed the CFQ‐R for both parent and participant (Amin 2010), and two trials used the CFQ‐R for parents only (Rosenfeld 2012; SHIP 2019). Donaldson 2020 used the standard CFQ‐R respiratory domain at four weeks, and the Elkins 2006a trial measured QoL using the SF‐36 questionnaire and the Cystic Fibrosis Questionnaire (CFQ) for adults and for parents.

The CFQ‐R domain for parents or participants was assessed in three trials (365 participants) (Amin 2010; Elkins 2006a; Rosenfeld 2012), and demonstrated no difference between groups (Analysis 1.11). There is moderate heterogeneity for CFQ Parent data (three trials) according to our definition above; the I² value is 47%. This value reverts to zero if the data from the Rosenfeld 2012 trial are removed. This trial recruited young children under the age of five years in contrast to the other two trials, which may account for the heterogeneity (Rosenfeld 2012). The SHIP 2019 trial also used the parent‐reported CFQ‐R respiratory domain and found no difference between groups at 48 weeks (adjusted treatment difference 0.1 (95% CI ‐5.7 to 5.8), taken directly from the trial paper).

1.11. Analysis.

Comparison 1: Hypertonic saline 3% to 7% versus isotonic saline, Outcome 11: Quality of life (change from baseline)

Elkins 2006a reported a significantly higher (better) score in the mental health domain of the SF‐36 in participants over 14 years (MD 7.77, 95% CI 1.86 to 13.68; 1 trial, 91 participants; Analysis 1.11); however, the overall score for the CFQ and SF‐36 was no different in those under 14 years of age (MD 2.84, 95% CI ‐7.90 to 13.58; 1 trial, 91 participants; Analysis 1.11) (Elkins 2006a). There were better results in the domains of role (P = 0.04), emotion (P = 0.03) and health (P = 0.01) in the CFQ Adult compared to the control group. In participants under the age of 14, the digestion domain was better (P = 0.02) in the control group compared to the hypertonic saline group using the CFQ for parents. This trial also measured absenteeism from work and school; and participants in the hypertonic saline group experienced fewer days off work, school or days they were unable to participate in usual activity; seven days in the hypertonic saline group as compared to 24 days in the control group (P < 0.001) (Elkins 2006a).

Donaldson 2020 reported CFQ‐R respiratory domain at four weeks, and again found no difference between groups (Analysis 1.11).

Two trials assessed symptoms using a VAS with 10 cm scales ranging from ‐5 cm to +5 cm (Eng 1996; Riedler 1996). One trial found improvements in symptoms for quality of sleep and feeling of cleared chest measured after one and two weeks of treatment with hypertonic saline 6% (Eng 1996). The Riedler trial looked at a similar VAS for feeling of cleared chest alone, which was measured four days after treatment and demonstrated a difference in their first block of 10 participants favouring hypertonic saline 6% compared to isotonic saline (Riedler 1996). The results of the two trials were pooled and demonstrated a result favouring treatment (MD 0.97, 95% CI 0.35 to 1.60; 2 trials, 72 participants; Analysis 1.12).

1.12. Analysis.

Comparison 1: Hypertonic saline 3% to 7% versus isotonic saline, Outcome 12: Feeling of cleared chest (using a subjective visual analogue scale)

One trial compared hypertonic saline versus isotonic saline delivered via a pulsating nebuliser (designed to enhance aerosol deposition into the nose and sinus cavities) for improving nasal and sinus symptoms, assessed using the validated Sino‐Nasal Outcome Test‐20 (SNOT‐20) symptom score (Mainz 2015). This tool is a disease‐specific, health‐related, 20‐item QoL measure for people with rhinosinusitis focusing on rhinogenous as well as on general discomforts. Scores range between 0 and 5 for each item, with higher scores indicating a greater health burden. On the first day of treatment, the hypertonic saline group described worsened symptom scores, mean (SD) 23.0 (10.4) compared to 24.8 (11.0) in the isotonic saline group (P < 0.005). By day 29 however, there was no difference between the groups; hypertonic saline 20.7 (10.1) and isotonic saline 19.4 (9.6) (Mainz 2015).

4. Pulmonary exacerbations

a. frequency

Four trials (607 participants) reported on this outcome but, although the definition of exacerbations was comparable, we elected not to pool the data as the age difference between the participants in the trials was so great (Elkins 2006a; PRESIS 2019; Rosenfeld 2012; SHIP 2019). We judged the certainty of the evidence to be very low.

Elkins 2006a reported pulmonary exacerbations as a secondary outcome. The trial found there were fewer exacerbations per year requiring intravenous antibiotic therapy in the hypertonic saline group than in the control group (MD ‐0.46, 95% CI ‐0.82 to ‐0.10; Analysis 1.13) (Figure 4). It was also reported that the mean number of days on which participants met this definition of an exacerbation was 17 days in the control group and six days in the hypertonic saline group (difference 11 days, 95% CI 3 to 19; P = 0.02). The interval during which participants remained free of exacerbations was also significantly longer in the hypertonic saline group than in the control group (P = 0.03). The Elkins 2006a trial also found that there was no difference in the mean number of visits not requiring antibiotics (MD ‐0.25, 95% ‐0.84 to 0.34; Analysis 1.13) (Figure 4).

1.13. Analysis.

Comparison 1: Hypertonic saline 3% to 7% versus isotonic saline, Outcome 13: Average number of pulmonary exacerbations

4.

Forest plot of comparison: 1 Hypertonic saline 3% to 7% versus isotonic saline, outcome: 1.13 Average number of exacerbations

Elkins 2006a stated in the original trial report that exacerbations, defined according to signs and symptoms alone, regardless of treatment, were also less frequent in the hypertonic saline group compared to control: 1.32 per participant and 2.74 per participant respectively (difference 1.42, 95% CI 0.86 to 1.99; P < 0.001). The mean number of days during which participants met criteria for a symptom‐defined exacerbation was 69 days in the control group and 22 days in the hypertonic saline group (difference 47 days, 95% CI 30 to 63; P < 0.001). The time participants remained free of exacerbations was significantly longer in the hypertonic saline group (P < 0.001), with a 48‐week exacerbation‐free survival rate of 41% in the hypertonic saline group and 16% in the control group.

Rosenfeld 2012 used protocol‐defined exacerbations as their primary outcome. They found no difference in the mean (SD) number of exacerbations; those with hypertonic saline experienced 2.3 (1.69) events per year, compared to isotonic saline 2.3 (1.15) events per year (Analysis 1.13).

The PRESIS 2019 and SHIP 2019 trials both reported the rate of pulmonary exacerbations and also time to first exacerbation in preschool children and infants. PRESIS 2019 reported the exacerbation rate in each group and found no difference (hypertonic saline group 1.1 (95% CI 0.0 to 2.1); isotonic saline group 1.2 (95% CI 0.4 to 1.9); P = 0.86). The probability of remaining free of pulmonary exacerbations in the hypertonic saline group compared to the isotonic saline group was 0.78, i.e. no difference.

The SHIP 2019 trial reported time to first pulmonary exacerbation as a hazard ratio (HR) at 48 weeks and again found no difference (HR 0.94, 95% CI 0.65 to 1.36; P = 0.63).

b. admission to hospital

Elkins 2006a stated that there was no difference reported in hospitalisation rates between the hypertonic saline group and the controls (Analysis 1.14).

1.14. Analysis.

Comparison 1: Hypertonic saline 3% to 7% versus isotonic saline, Outcome 14: Average number of hospital admissions per participant

c. duration of hospital stay (post hoc change)

No trial reported the duration of hospitalisation due to pulmonary exacerbations.

d. outpatient treatments (hospital in the home, unscheduled visits to the doctor)

No trial reported details of outpatient treatments due to pulmonary exacerbations.

5. Medication delivery time

No trials reported on this as an outcome.

6. Cost

No trials reported on this as an outcome.

7. Adherence

Four trials (655 participants) judged treatment adherence by the number of returned ampoules or vials (Amin 2010; Elkins 2006a; Rosenfeld 2012; SHIP 2019). 

We were only able to analyse data from one trial; SHIP 2019 reported the mean adherence rate in each group and found no difference between groups (MD ‐0.03, 95% CI ‐0.07 to 0.01; 1 trial, 150 participants; Analysis 1.15).

1.15. Analysis.

Comparison 1: Hypertonic saline 3% to 7% versus isotonic saline, Outcome 15: Mean adherence rate

We report the results from three trials narratively. In the cross‐over trial by Amin 2010, adherence was 95.3% for the hypertonic saline period and 84.5% for the isotonic saline period; this was not statistically significant (P = 0.26). Elkins 2006a reported adherence as 63% in the control group compared to 64% in the hypertonic saline group. Rosenfeld 2012 reported no difference in adherence between those receiving hypertonic saline and isotonic saline, with an overall mean adherence of 75.2% (95% CI 72.2% to 78.2%). 

8. Bacteriology

Four trials (607 participants) reported on bacteriology; we were able to analyse data from two trials (Elkins 2006a; PRESIS 2019) and report narrative results from two trials (Rosenfeld 2012; SHIP 2019).

Elkins 2006a measured the bacterial load of sputum; but when analysed, there was no difference between groups in the concentration of P aeruginosa or S aureus from baseline to 48 weeks (Analysis 1.16). The PRESIS 2019 trial reported on time (years) to first isolation of a CF pathogen and found no difference for either S aureus,H influenzae or P aeruginosa (Analysis 1.17).

1.16. Analysis.

Comparison 1: Hypertonic saline 3% to 7% versus isotonic saline, Outcome 16: Change in log10 colony forming units (CFU)/g from baseline at final visit

1.17. Analysis.

Comparison 1: Hypertonic saline 3% to 7% versus isotonic saline, Outcome 17: Mean time (years) to first detection of CF pathogens

Rosenfeld 2012 found no new bacterial pathogens, and the pathogens that were identified did not differ between the groups. The SHIP 2019 trial reported that changes in microbiology during the study did not differ significantly between the two groups.

9. Adverse events

Adverse events were reported in ten trials (825 participants) (Amin 2010; Amin 2016; Chadwick 1997; Donaldson 2020; Elkins 2006a; Eng 1996; PRESIS 2019; Robinson 1999; Rosenfeld 2012; SHIP 2019). Where appropriate data are available, we have presented these in the analyses (Analysis 1.18; Analysis 1.19; Analysis 1.20). We deemed the certainty of the evidence to be low.

1.18. Analysis.

Comparison 1: Hypertonic saline 3% to 7% versus isotonic saline, Outcome 18: Adverse events: acute fall in lung function

1.19. Analysis.

Comparison 1: Hypertonic saline 3% to 7% versus isotonic saline, Outcome 19: Adverse events 2 to 4 weeks

1.20. Analysis.

Comparison 1: Hypertonic saline 3% to 7% versus isotonic saline, Outcome 20: Adverse events 48 to 52 weeks

We report adverse events below split by time frame, but we have been unable to ascertain the time frame for one trial (Chadwick 1997). This trial demonstrated that participants with an FEV₁ % predicted of 40% to 70% at baseline experienced a significant fall in FEV₁ following isotonic saline, while none of the participants fell significantly with hypertonic saline (Chadwick 1997). 

Single‐dose trials

Two single‐dose trials (n = 32) reported adverse events (Amin 2016; Robinson 1999).

Amin 2016 (20 participants) reported three adverse events. In one participant FEV₁ dropped by more than 20% after inhaling hypertonic saline and the participant discontinued with the trial. Two participants reported shortness of breath and chest tightness related to inhalation of hypertonic saline; both were given salbutamol and continued on the trial.

In the Robinson 1999 trial, despite participants being pretreated with terbutaline, there was a tendency for those who received hypertonic saline to have a larger fall in FEV₁ % predicted within five minutes of receiving hypertonic saline, although the difference was not statistically significant (Analysis 1.18). Participants who received hypertonic saline described higher scores for throat irritation on a VAS compared to isotonic saline control, though the number describing throat irritation was not stated. Frequency of cough in those treated with hypertonic saline could not be directly compared to the days when they received the isotonic saline control as to they were encouraged to cough on the control days to match the active day's cough so as not to confound the results of the mucociliary clearance data (Robinson 1999).

Two to four weeks

Three trials (n = 101) of two to four weeks' duration reported adverse events (Amin 2010; Donaldson 2020; Eng 1996).

Amin 2010 reported the overall number of adverse events and adverse events related to the trial drugs, rather than the number of participants experiencing an adverse event. There were no significant differences in cough, hoarseness or chest pain (P = 0.17); however, there were significantly more overall adverse events in the hypertonic saline group for the symptoms of increased sputum production, fever, rhinorrhoea, malaise and ear infections (P = 0.0035). Amin 2010 further reported a mean (SD) drop in FEV₁ % predicted after the first inhalation of hypertonic saline 116 (140) mL and after isotonic saline 41 (88) mL; when analysed, the data showed no difference between groups (Analysis 1.18).

We were able to combine data for two trials (Donaldson 2020; Eng 1996). No differences were found between groups for any adverse event, cough, pharyngitis, chest tightness, haemoptysis, wheezing, nasal congestion or increased sputum production (Analysis 1.19).

48 to 52 weeks

Four trials (n = 677) reported adverse events at 48 to 52 weeks (Elkins 2006a; PRESIS 2019; Rosenfeld 2012; SHIP 2019).

The Elkins 2006a trial reported the number of adverse events (not participants experiencing adverse events) as pulmonary exacerbations (see above), chest pain, gastro‐intestinal symptoms, headache, joint pain, pharyngitis and tonsillitis; these were significantly fewer in the hypertonic saline group. Adverse drug reactions were also reported, events that in the opinion of the investigator were related directly to the trial medication. These were significantly higher in the hypertonic saline group (n = 14) than the control group (n = 1), P = 0.01, and included cough, chest tightness, pharyngitis, haemoptysis, sinusitis, sneezing, tonsillitis and vomiting (Elkins 2006a).

We were able to combine data for the remaining three trials (Analysis 1.20).

There was no difference between groups with regard to any adverse events, serious adverse events, cough, chest tightness, vomiting, wheezing, nasal congestion, ear infection, abdominal distension, or diarrhoea (Analysis 1.20). We graded the certainty of evidence as low.

Fever was slightly more common in the isotonic saline group (RR 0.73, 95% CI 0.53 to 1.00; 3 trials, 513 participants; low‐certainty evidence; Analysis 1.20), whilst rhinorrhoea was more common in the hypertonic saline groups (RR 1.70, 95% CI 1.04 to 2.79; 1 trials, 42 participants; low‐certainty evidence; Analysis 1.20).

Hypertonic saline 3% to 7% versus isotonic saline during acute exacerbations of lung disease

Two trials (n = 142) assessed the effect of hypertonic saline during an acute exacerbation of lung disease (Dentice 2016; Riedler 1996). One was of parallel design, and we were able to enter data into the analysis (Dentice 2016). The second trial used a cross‐over design and, due to data limitations, we have reported the results narratively (Riedler 1996). A summary of the results and our judgements in regard to the certainty of the evidence can be found in Table 2.

Primary outcomes

1. Lung function

a. FEV₁

Only the Dentice 2016 trial assessed the change in lung function, and this was as a secondary outcome. Spirometry was performed daily, and the investigators reported that FEV₁ was higher in the hypertonic saline arm using a mixed‐effects model in the first 10 days of treatment. We were able to assess the change from baseline in FEV₁ % predicted after the authors supplied the raw data. Hypertonic saline may make little or no difference to the change in FEV1 % predicted compared to isotonic saline at treatment day 7, the 99 participants reporting data at treatment day 10, or the 130 participants at time of discharge (low‐certainty evidence) (Analysis 2.1).  Dentice also showed that participants treated with hypertonic saline were more likely to return to their pre‐exacerbation FEV₁ than those treated with isotonic saline, 75% versus 57% (number needed to treat for an additional beneficial outcome (NNTB) was 6 (95% CI 3 to 65)) (Dentice 2016).

2.1. Analysis.

Comparison 2: Hypertonic saline 3% to 7% versus isotonic saline in acute lung disease, Outcome 1: Change in FEV₁ from baseline (% predicted)

b. FVC

Only Dentice 2016 measured the change in FVC, as a secondary outcome and in a similar manner to FEV₁. We were able to assess the change from baseline in FVC % predicted after the authors supplied the raw data. Again, there was no evidence of any difference between groups when assessed at treatment day 7 (130 participants), at day 10 (99 participants) or at discharge (130 participants), although significance was nearly reached at this time point, MD 6.1% (95% CI ‐0.68 to 12.88) (Analysis 2.2).

2.2. Analysis.

Comparison 2: Hypertonic saline 3% to 7% versus isotonic saline in acute lung disease, Outcome 2: Change in FVC from baseline (% predicted)

2. Mortality

There were no deaths reported in either short‐term trial (Dentice 2016; Riedler 1996) (low‐certainty evidence).

Secondary outcomes

1. Measures of sputum clearance

This outcome was not reported in either trial (Dentice 2016; Riedler 1996).

2. Exercise tolerance

Dentice 2016 assessed exercise tolerance using the shuttle walk test, comparing the groups at day 7; the results just failed to demonstrate a difference between groups (MD 46.00 m, 95% CI ‐14.81 to 106.81; 1 trial, 132 participants; Analysis 2.3).

2.3. Analysis.

Comparison 2: Hypertonic saline 3% to 7% versus isotonic saline in acute lung disease, Outcome 3: Shuttle walk test

3. Measures of QoL and symptom score

Riedler 1996 used a VAS to report a feeling of cleared chest in a short‐term cross‐over trial. The paper reported median (interquartile range) scores, which we were unable to analyse and so report narratively. The score was significantly higher after hypertonic saline, median (range) 2.0 (0.0 to 3.0) compared to after isotonic saline 0.5 (‐2.0 to 2.3) (P = 0.04), but this was not the case after participants had crossed over to the alternate arm, when the median (range) score was 2.0 (1.0 to 3.0) for hypertonic saline and 1.0 (0.0 to 3.0) for isotonic saline (P = 0.463) (Riedler 1996).

Dentice 2016 also used a VAS (100 mm scale) to measure changes in chest congestion, sleep disturbance and dyspnoea (both daily and at the time of discharge) and analysed the data using a mixed‐effect model. At discharge, the hypertonic saline group had significantly less severe sleep disturbance by 15 mm (95% CI 6 to 23), chest congestion by 9 mm (95% CI 4 to 14) and dyspnoea by 6 mm (95% CI 1 to 12). Dentice 2016 also measured QoL using the SF‐36 and the CFQ at day 7 and at discharge, but reported no differences between the groups.

4. Pulmonary exacerbations

Dentice 2016 assessed the length of time until the next pulmonary exacerbation requiring hospitalisation and duration of hospitalisation.

b. admission to hospital

We are uncertain whether there is any difference between treatment groups in the length of time until the next pulmonary exacerbation requiring hospitalisation, HR 0.86 (95% CI 0.57 to 1.30). We assessed this evidence as low certainty (Dentice 2016).

c. duration of hospital stay

This was the primary outcome in the Dentice 2016 trial. The length of stay was 12 days in the hypertonic saline group and 13 days in the isotonic saline group. The paper reported this with an MD of one day (95% CI 0 to 2 days), P = 0.07. The mean estimate of one day was below the two‐day difference nominated in the sample size calculation.

5. Medication delivery time

This outcome was not assessed by either trial (Dentice 2016; Riedler 1996).

6. Cost of treatment

This outcome was not assessed by either trial (Dentice 2016; Riedler 1996).

7. Adherence

In Dentice 2016, both treatment arms demonstrated very good adherence; only 6% of the hypertonic saline group and 14% of the control group were less than 75% adherent to the allocated intervention. Riedler 1996 did not report adherence.

8. Bacteriology

Dentice 2016 found no evidence of any difference between the groups in terms of bacterial density when comparing bacterial cultures taken on admission with cultures taken at day 7. There was no evidence of any difference between the groups for participants who were positive for P aeruginosa on admission and who were negative on day 7 (10% in the hypertonic saline group and 6% in the isotonic saline group). Clearance of S aureus was higher in both groups, but still did not show any evidence of a difference between groups: 25% in the hypertonic group and 24% in the control group. This outcome was not reported by Riedler 1996.

9. Adverse events

Both trials reported on adverse events, but we judged the certainty of the evidence to be very low. 

Dentice 2016 reported that no participants had an acute fall greater than 15% in FEV₁ or oxygen desaturation after their first dose of hypertonic saline (with salbutamol before treatment). There were reports of mild cough and wheeze that resolved in 15 minutes; but there were no serious adverse events reported in either group.

Riedler 1996 reported that most participants coughed "substantially more" when inhaling hypertonic saline compared to isotonic saline, especially during the first few minutes of inhalation, but this usually resolved before the end of the inhalation period.

Hypertonic saline versus mucus mobilising treatments

Hypertonic saline versus rhDNase

Three trials were eligible for inclusion in this comparison (n = 80) (Adde 2004; Ballmann 1998; Suri 2001). All three were cross‐over trials; due to data limitations we analysed two of these as if they were parallel trials (Adde 2004; Ballmann 1998), but we were able to analyse the Suri 2001 data using the generic inverse variance method. A summary of the main results and judgements on the certainty of the evidence are presented in Table 3.

Primary outcomes

1. Lung function

a. FEV₁

All three trials reported on this outcome (Adde 2004; Ballmann 1998; Suri 2001). Suri 2001 measured the mean increase in FEV₁ % predicted from baseline comparing hypertonic saline 3% 5 mL (3% increase) to daily rhDNase (16% increase) and alternate‐daily rhDNase (14% increase). We present comparisons between daily hypertonic saline and daily rhDNase in this review; these were reported by Ballmann 1998 and Suri 2001. The results have not been pooled because the duration of the interventions in the two trials was very different; three weeks for the Ballmann 1998 trial and three months for the Suri 2001 trial.

After three weeks, we are uncertain if there is a difference between hypertonic saline and rhDNase (very low‐certainty evidence; 1 trial; Analysis 3.1) (Ballmann 1998). After three months, there was a lower change in FEV₁ % predicted after hypertonic saline compared to rhDNase, but we are uncertain of this result as the certainty of evidence is very low (MD 8.00%, 95% CI 2.00 to 14.00; 1 trial, 47 participants; very low‐certainty evidence; Analysis 3.1) (Suri 2001). Both Ballmann 1998 and Suri 2001 compared the number of participants who improved their FEV₁ by 10% or more from baseline after treatment. At three weeks, Ballmann 1998 reported in the paper that those treated with hypertonic saline were less likely to increase their FEV₁ by 10%, but our analysis does not show a statistical difference between treatments; this was also true for the Suri 2001 trial at three months (Analysis 3.2).

3.1. Analysis.

Comparison 3: Hypertonic saline versus rhDNase, Outcome 1: Change in FEV₁ (% predicted)

3.2. Analysis.

Comparison 3: Hypertonic saline versus rhDNase, Outcome 2: Improvement in FEV₁ >10%

Adde 2004 reported the change in FEV₁ % predicted and found no difference between treatments. The results have not been presented in a meta‐analysis as they did not report an MD and standard error.

b. FVC

Suri 2001 reported that there was no difference between daily rhDNase and hypertonic saline (Analysis 3.3). 

3.3. Analysis.

Comparison 3: Hypertonic saline versus rhDNase, Outcome 3: Change in FVC (% predicted)

2. Mortality

None of the trials reported on this outcome.

Secondary outcomes

1. Measures of sputum clearance

None of the trials reported on this outcome.

2. Measures of exercise capacity

Suri measured exercise tolerance using a three‐minute step test at the end of each treatment period. As part of the step test, the changes in the saturation of haemoglobin with oxygen in arterial blood (SaO2), the VAS score and the 'fifteen count breathlessness score' (FCS) were recorded (Suri 2001). They reported no differences between the groups for either oxygen saturation, VAS for breathlessness, or FCS (Analysis 3.4; Analysis 3.5; Analysis 3.6). 

3.4. Analysis.

Comparison 3: Hypertonic saline versus rhDNase, Outcome 4: Exercise tolerance ‐ oxygen saturation

3.5. Analysis.

Comparison 3: Hypertonic saline versus rhDNase, Outcome 5: Exercise tolerance ‐ VAS for breathlessness

3.6. Analysis.

Comparison 3: Hypertonic saline versus rhDNase, Outcome 6: Exercise tolerance ‐ FCS

3. Measures of QoL and symptom scores

Two trials reported on this outcome; however, the results have not been pooled as the outcome measures were not standardised (Adde 2004; Suri 2001). Suri 2001 assessed symptoms using the quality of well‐being self‐administered form 1.04. They reported there was no difference in scores between the groups (Analysis 3.7). Adde 2004 assessed symptom scores using a five‐point Likert scale and also reported no difference between groups.

3.7. Analysis.

Comparison 3: Hypertonic saline versus rhDNase, Outcome 7: Mean percentage change in quality of life score

4. Pulmonary exacerbations

Suri 2001 described pulmonary exacerbations during the trial, with 15 episodes occurring during treatment with hypertonic saline and 18 with daily rhDNase. We are uncertain if there is any difference between treatments (very low‐certainty evidence). 

5. Medication delivery time

Ballmann 1998 compared delivery time in minutes between hypertonic saline 5.85% 10 mL twice‐daily and rhDNase 2.5 mg twice‐daily and found that hypertonic saline took longer to nebulise (MD ‐31.00 minutes, 95% CI ‐37.56 to ‐24.44; 1 trial, 14 participants (cross‐over design with 14 participants in each arm); Analysis 3.8). The large difference in nebulisation time relates to the difference in volumes nebulised using the same Pari master or Pari LL nebuliser (Ballmann 1998). 

3.8. Analysis.

Comparison 3: Hypertonic saline versus rhDNase, Outcome 8: Delivery time (minutes)

6. Cost

Two trials compared the cost of treatment between rhDNase and hypertonic saline (Ballmann 1998; Suri 2001). 

Ballmann 1998 compared one month of hypertonic saline treatment with rhDNase (Deutschmark (DM) 2427) to hypertonic saline (DM 86). 

Suri 2001 compared total healthcare cost for the treatments incorporating not just drug cost but also admission, outpatient review, cost of investigations and the cost of utilising community resources. Suri 2001 investigated the mean cost difference between daily rhDNase and hypertonic saline and alternate‐day rhDNase at 12 weeks. As reported in the original paper, the drug cost per day was reported to be GBP 0.38 for hypertonic saline, GBP 20.39 for once‐daily rhDNase and GBP 10.20 for alternate‐day rhDNase (Suri 2001). The average total cost of an occupied bed per day ranged from GBP 280 to GBP 397 (Suri 2001). The mean annual drug cost of daily rhDNase was GBP 1755 compared with GBP 37 for hypertonic saline and the MD in the total health service cost between daily rhDNase and hypertonic saline was GBP 1409.00 (95% CI GBP 440.00 to GBP 2318.00). The MD in total cost between daily rhDNase and alternate‐day rhDNase was GBP 513.00 (95% CI GBP ‐546.00 to GBP 1510.00) (Suri 2001). 

7. Adherence

Only Suri 2001 assessed adherence and reported the number of returned treatment packs. Those on rhDNase had compliance rates of 84%, with those on hypertonic saline having 93% compliance.

8. Bacteriology

Suri 2001 assessed sputum microbiology throughout the trial and did not identify any new pathogens acquired during the course of the trial amongst individuals.

Adde 2004 compared P aeruginosa growth and found no difference in bacterial load using hypertonic saline compared to rhDNase.

9. Adverse events

Based on data from one trial we are uncertain if there is any difference in adverse events between groups as we judged the certainty of the evidence to be very low. Three participants had to withdraw because of a fall of 15% or greater in FEV₁ after receiving hypertonic saline despite pretreatment with bronchodilators. Increased cough was very common with all treatments and reported in 13 participants using hypertonic saline, 17 on daily rhDNase and 23 on alternate day rhDNase (very low‐certainty evidence) (Suri 2001).

Hypertonic saline versus amiloride

One trial was eligible for inclusion in this comparison (n = 12) (Robinson 1996). The main results and the certainty of the evidence are presented in Table 4.

Primary outcomes

1. Lung function

The included trial did not report on this as an outcome. 

2. Mortality

The included trial did not report on this as an outcome.

Secondary outcomes

1. Measures of sputum clearance

The included trial looked at the effect of amiloride with hypertonic saline and amiloride alone compared to isotonic saline (Robinson 1996). We are uncertain whether there is any difference with amiloride plus hypertonic saline or amiloride alone compared to isotonic saline (very low‐certainty evidence).

2. Measures of exercise capacity

The included trial did not report on this as an outcome. 

3. Measures of QoL and symptom scores

The included trial did not report on this as an outcome. 

4. Pulmonary exacerbations

The included trial did not report on this as an outcome. 

5. Medication delivery time

The included trial did not report on this as an outcome. 

6. Cost

The included trial did not report on this as an outcome. 

7. Adherence

The included trial did not report on this as an outcome. 

8. Bacteriology

The included trial did not report on this as an outcome.

9. Adverse events

The included trial did not report on this as an outcome.

Hypertonic saline versus sodium‐2‐mercaptoethane sulphonate (Mistabron®)

One trial was eligible for inclusion in this comparison (n = 29) (Weller 1980). Only 3 mL of 7% hypertonic saline was used in the Weller 1980 trial. We were unable to enter data into the graphs, because SDs were not reported in the paper. The main results and the judgements on the certainty of the evidence are presented in Table 5.

Primary outcomes

1. Lung function

Weller 1980 compared sodium‐2‐mercaptoethane sulphonate 20% 3 mL twice‐daily to hypertonic saline 7% 3 mL twice‐daily. Participants were divided into sputum producers and non‐sputum producers. The sputum producers who were given sodium‐2‐mercaptoethane sulphonate increased peak expiratory flow (PEF) (change from baseline of 7 L/min) compared to hypertonic saline (change from baseline of ‐2 L/min, P < 0.02). There was no significant difference in PEF in the non‐sputum producers. Furthermore, FVC was not significantly different in either group. The V_max 50% vital capacity (VC) increased in the sputum producers with sodium‐2‐mercaptoethane sulphonate (+10) compared to hypertonic saline (0; P < 0.005). In the non‐sputum producers, hypertonic saline improved V_max 50% VC (+14) compared to sodium‐2‐mercaptoethane sulphonate (‐5), but this was not significant. In the sputum producers group residual volume (RV) and total lung capacity (TLC) improved with hypertonic saline (+1) compared to sodium‐2‐mercaptoethane sulphonate (‐5; P < 0.05). In the non‐sputum producers group, sodium‐2‐mercaptoethane sulphonate had no effect (0) on RV or TLC, whilst hypertonic saline had some effect (‐6), again this did not reach statistical significance. 

2. Mortality

The included trial did not report on this as an outcome.

Secondary outcomes

1. Measures of sputum clearance

We are uncertain whether there is any difference in sputum volume, colour or cough frequency between the groups (very low‐certainty evidence) (Weller 1980). 

2. Measures of exercise capacity

The included trial did not report on this as an outcome. 

3. Measures of QoL and symptom scores

The included trial did not report on this as an outcome. 

4. Pulmonary exacerbations

There was no change in sputum bacteriology or the number of courses of antibiotics prescribed (Weller 1980) (very low‐certainty evidence). 

5. Delivery time

The included trial did not report on this as an outcome. 

6. Cost

The included trial did not report on this as an outcome. 

7. Adherence

The included trial did not report on this as an outcome.

8. Bacteriology

The included trial did not report on this as an outcome.

9. Adverse events

In the Weller 1980 trial, the group given sodium‐2‐ mercaptoethane sulphonate and hypertonic saline described coughing at the beginning of their inhalations; no other serious adverse events occurred (very low‐certainty evidence).

Hypertonic saline versus mannitol

One trial was eligible for inclusion in this comparison (n = 12) (Robinson 1999). In the analysis, due to data limitations, data from this cross‐over trial have been entered and analysed as if they were from a parallel trial. The main results and the certainty judgements on the evidence are presented in Table 6.

Primary outcomes

1. Lung function

The change from the pre‐intervention FEV₁ was reported at five minutes after inhalation and again 95 minutes later (Robinson 1999). The paper reported the mean (standard error (SE)) value for each intervention and their respective controls at each time‐point, but only reported the level of significance compared to control. At five minutes post inhalation the mean (SE) change in FEV₁ for hypertonic saline was ‐5.8 (1.2) and for mannitol ‐7.3 (2.5); at the later time point the mean (SE) difference was ‐2.0 (0.7) for hypertonic saline and ‐1.8 (2.7) for mannitol (Robinson 1999) (very low‐certainty evidence). 

2. Mortality

The included trial did not report on this as an outcome. 

Secondary outcomes

1. Measures of sputum clearance

In the included trial, we are uncertain whether there was any difference between mannitol at 300 mg and hypertonic saline for matched voluntary cough (Robinson 1999) (very low‐certainty evidence). 

2. Measures of exercise capacity

The included trial did not report on this as an outcome. 

3. Measures of QoL and symptom scores

In the 1999 trial, Robinson used a VAS to assess the need to cough and reported that there was no difference between hypertonic saline 6% and mannitol 300 mg (Robinson 1999).

4. Pulmonary exacerbations

The included trial did not report on this as an outcome. 

5. Medication delivery time

Robinson compared the time taken to nebulise hypertonic saline 6% (4.4 mL) to mannitol 300 mg and found hypertonic saline took less time (MD ‐6.10 min, 95% CI ‐7.32 to ‐4.88; 1 trial, 12 participants; Analysis 4.1) (Robinson 1999). 

4.1. Analysis.

Comparison 4: Hypertonic saline versus mannitol, Outcome 1: Delivery time (mins)

6. Cost

The included trial did not report on this as an outcome.

7. Adherence

The included trial did not report on this as an outcome.

8. Bacteriology

The included trial did not report on this as an outcome.

9. Adverse events

In the 1999 Robinson trial, mannitol was regarded as more irritating than the control on VAS. While FEV₁ fell significantly five minutes after treatment with both mannitol and hypertonic saline 6% compared to control (P = 0.004), by 95 minutes we are uncertain whether there is any difference between the groups (Robinson 1999) (very low‐certainty evidence).

Hypertonic saline 7% versus aerosolised xylitol

Two trials (89 participants) are included in this comparison (NCT01355796; Singh 2020). A summary of the results and our judgements with regard to the certainty of the evidence can be found in the Table 7.

Primary outcomes

1. Lung function

a. FEV₁

Both of the included trials reported change in FEV₁ % predicted at 14 days and we are uncertain whether there is any difference between groups MD (very low‐certainty evidence; Analysis 5.1) (NCT01355796; Singh 2020).

5.1. Analysis.

Comparison 5: Hypertonic saline 7% versus xylitol, Outcome 1: FEV₁ % predicted mean change from baseline

b. FVC

Only Singh 2020 reported on FVC and again found no difference between groups (Analysis 5.2).

5.2. Analysis.

Comparison 5: Hypertonic saline 7% versus xylitol, Outcome 2: FVC % predicted mean change from baseline

2. Mortality

No deaths were reported in either group in either trial (NCT01355796; Singh 2020).

Secondary outcomes

1. Measures of sputum clearance

Neither of the trials reported on this outcome.

2. Exercise tolerance

Neither of the trials reported on this outcome.

3. Measures of QoL and symptom score

The Singh 2020 trial reported change in CFQ‐R score but did not provide sufficient data for the analysis; the authors reported that there were no differences in any domain between the two groups (Singh 2020).

4. Pulmonary exacerbations

The Singh 2020 trial reported median time to next pulmonary exacerbation and stated that there was no difference between groups after participants were followed up for six months (low‐certainty evidence).

5. Medication delivery time

Neither of the trials reported this outcome.

6. Cost of treatment

Neither of the trials reported this outcome.

7. Adherence

The Singh 2020 trial reported on the mean number of doses taken and found no difference between the hypertonic saline group or the xylitol group (Analysis 5.3).

5.3. Analysis.

Comparison 5: Hypertonic saline 7% versus xylitol, Outcome 3: Adherence to treatment ‐ mean number of doses taken

8. Bacteriology

Both trials reported on bacteriology outcomes, but only in participants who were able to produce sputum. The NCT01355796 trial found a slight reduction in the sputum density of P aeruginosa in the xylitol group compared to the hypertonic saline (MD 2.91 log CFU/mL, 95% CI 0.26 to 5.56; 1 cross‐over trial, 27 participants; Analysis 5.4) but when the results were combined with data from the Singh 2020 trial, there was no difference between groups. This outcome was only measured in participants who were able to produce sputum.

5.4. Analysis.

Comparison 5: Hypertonic saline 7% versus xylitol, Outcome 4: Density of colonisation of pathogens ‐ change from baseline log CFU/mL

The Singh 2020 trial also reported on P aeruginosa mucoid and S aureus. Investigators found that there was a greater reduction in sputum density in the hypertonic saline group compared to the xylitol group (MD ‐1.80 log CFU/mL, 95% CI ‐3.48 to ‐0.12; 1 trial, 40 participants; Analysis 5.4), conversely there was no difference between groups in the change in sputum density of S aureus (Analysis 5.4) (Singh 2020).

9. Adverse events

Both trials reported the number of participants experiencing adverse events during the trial period. No participants in the xylitol group from either trial reported serious adverse events compared to three out of 69 participants in the hypertonic saline groups (one participant had pneumonia (NCT01355796) and two participants were diagnosed with distal intestinal obstruction syndrome, but this was deemed to be unrelated to the study (Singh 2020)). When we entered the data into our analysis, we found no differences between groups for any adverse event, serious adverse events, gastrointestinal disorders, hepatobiliary disorders, bronchospasm, haemoptysis, or vascular disorders (very low‐certainty evidence; Analysis 5.5).

5.5. Analysis.

Comparison 5: Hypertonic saline 7% versus xylitol, Outcome 5: Adverse events

Hypertonic saline 7% versus hypertonic saline 3%

One trial (30 children) are included in this comparison (Gupta 2012). A summary of the results and our judgements in regard to the certainty of the evidence can be found in the tables (Table 8).

Primary outcomes

1. Lung function

a. FEV₁

There was no evidence of a difference in FEV₁ % predicted endpoint values between groups at 14 days and 28 days; only data for 28 days are entered in the analysis (Analysis 6.1). The mean percentage change from baseline in FEV₁ % predicted was greater in the 3% hypertonic saline group compared to the 7% hypertonic saline group although the CIs were wide (MD ‐13.00% predicted, 95% CI ‐25.27 to ‐0.73; 1 trial, 30 participants; very low‐certainty evidence; Analysis 6.2) (Gupta 2012).

6.1. Analysis.

Comparison 6: Hypertonic saline 7% versus hypertonic saline 3%, Outcome 1: FEV₁ % predicted (absolute value)

6.2. Analysis.

Comparison 6: Hypertonic saline 7% versus hypertonic saline 3%, Outcome 2: FEV₁ % predicted mean percentage change from baseline

b. FVC

Similarly, there was no evidence of a difference in FVC % predicted endpoint values at either 14 or 28 days; again, only data for 28 days were entered in the analysis (Analysis 6.3). There was a greater improvement in mean change from baseline in the 3% hypertonic saline group at 28 days, but again, CIs were wide (MD ‐15.22% predicted, 95% CI ‐29.36 to ‐1.08; 1 trial, 30 participants; Analysis 6.4) (Gupta 2012).

6.3. Analysis.

Comparison 6: Hypertonic saline 7% versus hypertonic saline 3%, Outcome 3: FVC % predicted (absolute value)

6.4. Analysis.

Comparison 6: Hypertonic saline 7% versus hypertonic saline 3%, Outcome 4: FVC % predicted mean percentage change from baseline

2. Mortality

Mortality was not reported in this trial.

Secondary outcomes

1. Measures of sputum clearance

The trial did not report on this outcome.

2. Exercise tolerance

Exercise capacity was measured via peak expiratory flow rate (PEFR), % oxygen level in arterial blood (SaO₂) and heart rate during a three‐minute step test (Gupta 2012). No differences were seen between groups for any of the outcomes; PEFR, % SaO₂ or heart rate (beats per minute (BPM)) (Analysis 6.5; Analysis 6.6; Analysis 6.7) .

6.5. Analysis.

Comparison 6: Hypertonic saline 7% versus hypertonic saline 3%, Outcome 5: Exercise capacity ‐ PEFR (L/min) via a 3‐minute step test

6.6. Analysis.

Comparison 6: Hypertonic saline 7% versus hypertonic saline 3%, Outcome 6: Exercise capacity ‐ SaO₂ via a 3‐minute step test

6.7. Analysis.

Comparison 6: Hypertonic saline 7% versus hypertonic saline 3%, Outcome 7: Exercise capacity ‐ heart rate (bpm) via a 3‐minute step test

3. Measures of QoL and symptom score

This trial did not report on QoL, but did report symptoms of breathlessness (Gupta 2012). Participants were asked to rate perception of breathlessness during a three‐minute step test using a VAS where 0 was no breathlessness and 10 was severe breathlessness. Data showed no difference between groups (Analysis 6.8).

6.8. Analysis.

Comparison 6: Hypertonic saline 7% versus hypertonic saline 3%, Outcome 8: Breathlessness via a 3‐minute step test

Investigators also measured breathlessness using a 15‐count scale, where the children were asked to take a deep breath and then count out loud from one to 15. The number of breaths taken during the exercise was recorded. Mean ranks of the two groups were comparable, and no differences were seen (Gupta 2012).

4. Pulmonary exacerbations

This outcome was not reported.

5. Medication delivery time

This outcome was not reported.

6. Cost of treatment

The authors of the trial noted that both of the treatments are cost‐effective, but the 3% strength is easily commercially available (Gupta 2012).

7. Adherence

Participants were asked to return all unused containers and to keep a diary to assess adherence, but results were not reported (Gupta 2012).

8. Bacteriology

This outcome was not reported

9. Adverse events

The trial did not report adverse events.

Discussion

Summary of main results

Hypertonic saline versus control (stable disease)

We included 15 trials comparing hypertonic saline to control in participants with stable lung disease (Amin 2010; Amin 2016; Cardinale 2003; Chadwick 1997; Donaldson 2020; Elkins 2006a; Eng 1996; Laube 2009; Mainz 2015; PRESIS 2019; Robinson 1996; Robinson 1997; Robinson 1999; Rosenfeld 2012; SHIP 2019). We are uncertain whether the regular use of nebulised hypertonic saline by adults and children over the age of 12 years with CF and stable lung disease leads to an improvement in lung function (FEV₁ % predicted) after four weeks of treatment (Amin 2010; Elkins 2006a; Eng 1996), or whether there is any difference in lung function after 48 weeks of treatment (Elkins 2006a). There were mixed results for other measures of lung function. Four trials in preschool children used LCI as their primary outcome and measure of lung function (Amin 2010; Amin 2016; PRESIS 2019; SHIP 2019). Our analysis of one paediatric trial showed no difference in LCI at four weeks, although the original investigators reported a difference in favour of hypertonic saline (very low‐certainty evidence) (Amin 2010). In preschool children, treatment for 48 weeks with hypertonic saline did show a small improvement in LCI favouring the hypertonic saline group (MD ‐0.60, 95% CI ‐1.00 to ‐0.19; 2 trials, 190 participants; Analysis 1.6) (PRESIS 2019; SHIP 2019). No trial reported on mortality.

We are uncertain whether hypertonic saline improved mucociliary clearance compared to control (very low‐certainty evidence) (Laube 2009; Robinson 1996; Robinson 1997; Robinson 1999). Two trials assessed symptom improvement after short‐term treatment using simple VAS and found an improvement in feelings of better chest clearance, exercise tolerance and quality of sleep (Eng 1996; Riedler 1996). In the long‐term trials (48 weeks), Elkins 2006a showed treatment may improve some aspects of QoL in adults but not in children, while Rosenfeld 2012 showed no improvement in parent‐reported QoL scores. Elkins 2006a also reported decreased absenteeism from work or school. When delivered by a Pari Sinus nebuliser, hypertonic saline did not improve nasal and sinus symptoms in one trial (Amin 2010).

Hypertonic saline may make little or no difference to pulmonary exacerbations compared to isotonic saline. One multicentre trial (164 adults) showed that nebulised hypertonic saline reduced the frequency of pulmonary exacerbations (Elkins 2006a). However, in a paediatric multicentre trial (321 children under six years of age with stable lung disease) hypertonic saline did not reduce the frequency of pulmonary exacerbations when compared to placebo (Rosenfeld 2012). Our analysis of the adult trial demonstrated a reduction in exacerbations requiring antibiotics, though no difference was seen in hospitalisations (Elkins 2006a). This finding has been felt to be important enough to result in an increasing uptake in treatment with hypertonic saline and led Rosenfeld to determine if early intervention in young children (under six years of age) would be beneficial. The Rosenfeld 2012 trial was powered to assess an impact on well‐defined pulmonary exacerbations, but no benefit was seen after 48 weeks of treatment. Two trials in preschool children found that there may be no difference in rate of exacerbations between groups (PRESIS 2019; SHIP 2019).

No trial assessed medication delivery time or the cost of treatment. Limited reporting from three trials indicated no difference between hypertonic saline or control for adherence (Amin 2010; Elkins 2006a; Rosenfeld 2012).

Nine trials reported details of adverse events (Amin 2010; Amin 2016; Chadwick 1997; Elkins 2006a; Eng 1996; PRESIS 2019; Robinson 1999; Rosenfeld 2012; SHIP 2019). 

Two single‐dose trials reported no difference between groups in fall in FEV₁ after inhalation of hypertonic saline but did report throat irritation, shortness of breath and chest tightness related to inhalation of hypertonic saline (Amin 2016; Robinson 1999).

Three trials reported at two to four weeks, two of which found no differences between groups for cough, hoarseness, chest pain, pharyngitis, haemoptysis, wheezing, or nasal congestion (Donaldson 2020; Eng 1996). One trial reported more adverse events in the hypertonic saline group for fever, rhinorrhoea, malaise and ear infections (P = 0.003) (Amin 2010).

Four trials reported at 48 weeks and found no differences between treatment groups for most adverse events including cough, chest tightness, pharyngitis, haemoptysis, sinusitis, sneezing, tonsillitis and vomiting. Fever was slightly more common in the isotonic saline group whilst rhinorrhoea was more common in the hypertonic saline group (Elkins 2006a; PRESIS 2019; Rosenfeld 2012; SHIP 2019).

Hypertonic saline versus control (acute exacerbation)

Two trials assessed hypertonic saline in participants hospitalised with an acute exacerbation (Dentice 2016; Riedler 1996). Hypertonic saline remains the only reported mucus clearing therapy that has been used in the context of acute exacerbations (Dentice 2016; Riedler 1996). In adults admitted to hospital with an acute exacerbation of CF lung disease, hypertonic saline did not lead to a greater improvement in lung function as measured by FEV₁ % predicted (MD 5.10%, 95% CI ‐14.67 to 24.87; 1 trial, 130 participants; low‐certainty evidence) and FVC % predicted, although a greater proportion of those treated with hypertonic saline regained their pre‐exacerbation FEV₁ (Dentice 2016). No deaths were reported in either trial (low‐certainty evidence). Neither trial assessed measures of sputum clearance. Hypertonic saline did lead to improvements in symptoms and QoL as well as a modest reduction in length of hospital stay, although it did not lengthen the time until the next exacerbation (low‐certainty evidence). Neither trial assessed medication delivery time nor the cost of treatment. Only one trial reported on adherence and assessed this as very good in both treatment arms (Dentice 2016). The same trial reported no difference in any measures of bacteriology. Even though these trials treated participants with acute exacerbations of lung disease, Dentice 2016 reported there were no serious adverse events and the intervention was well‐tolerated; Riedler 1996 reported that initially most participants coughed "substantially more" when inhaling hypertonic saline compared to isotonic saline, but this usually resolved before the end of the inhalation period (very low‐certainty evidence).

Hypertonic saline versus rhDNase

Three trials compared hypertonic saline to rhDNase (Adde 2004; Ballmann 1998; Suri 2001). In comparison with rhDNase, there was no difference between treatments at three weeks (very low‐certainty evidence) (Ballmann 1998), but at three months hypertonic saline was less likely to result in an improvement in lung function in those with stable lung disease (very low‐certainty evidence) (Suri 2001). Neither trial assessed LCI, mortality or measures of sputum clearance. One trial reported at 12 weeks on the change in exercise tolerance, dyspnoea, oxygen saturation during exercise and symptom score and found no differences between those treated with rhDNase and hypertonic saline (Suri 2001). In the same trial, all treatment arms experienced a high frequency of pulmonary exacerbations, but this may be a reflection of the severity of the groups' underlying lung disease (very low‐certainty evidence). In terms of cost, both Ballmann 1998 and Suri 2001 found hypertonic saline to be less expensive compared to rhDNase. One trial found that hypertonic saline took significantly longer than rhDNase to administer (Ballmann 1998); this may have implications for adherence to treatment. Suri 2001 specifically examined adherence, which appeared high and comparable to rhDNase in the three‐month trial. Only one trial assessed adverse events and found similar rates of events between groups (Suri 2001) (very low‐certainty evidence). Acute bronchospasm remains a concern with hypertonic saline. Despite pretreatment of participants with bronchodilators, three were excluded from the Suri 2001 trial due to a fall in FEV₁ greater than 15%.

Hypertonic saline versus amiloride

One small trial evaluated this comparison, but did not report on many of our outcomes (Robinson 1996). The trial did report that there was no difference between hypertonic saline and amiloride in terms of sputum clearance (very low‐certainty evidence).

Hypertonic saline compared with sodium‐2‐mercaptoethane sulphonate (Mistabron®)

One small trial evaluated this comparison, but did not provide data which we were able to analyse (Weller 1980). The trial did not report FEV₁, but reported mixed results for other lung function measurements. There was no difference between groups in measures of sputum clearance, sputum bacteriology or the number of courses of antibiotics prescribed (very low‐certainty evidence). All participants described coughing at the beginning of their inhalations; no other serious adverse events occurred (very low‐certainty evidence) (Weller 1980).

Hypertonic saline versus mannitol

One cross‐over trial compared hypertonic saline to mannitol (Robinson 1999). Investigators assessed FEV₁ at up to 95 minutes, which is not of clinical relevance to this review (very low‐certainty evidence), and reported within‐group changes from baseline but no data for between‐group comparisons. There was no difference in sputum clearance between groups (very low‐certainty evidence). The trial also reported no difference between groups in symptoms (the need to cough) and that hypertonic saline took less time to nebulise than mannitol. However, mannitol was considered to be more irritating than hypertonic saline (very low‐certainty evidence). No other outcomes from this review were reported.

Hypertonic saline versus xylitol

Two trials (89 participants) compared hypertonic saline to aerosolised xylitol, one of which was a cross‐over trial (NCT01355796) and the other a parallel trial (Singh 2020). There was no difference between groups in FEV₁ % predicted at 14 days (very low‐certainty evidence) or median time to exacerbation at six months (very low‐certainty evidence).

We also found very low‐certainty evidence for no difference in adverse events between groups.

Hypertonic saline 7% versus hypertonic saline 3%

One trial (30 participants) compared 7% hypertonic saline with 3% hypertonic saline and found that mean percentage change from baseline in FEV₁ % predicted was greater in the 3% group compared to the 7% group (very low‐certainty evidence). No other outcomes were reported.

Overall completeness and applicability of evidence

The rationale behind the use of hypertonic saline is the inherent defect in the CFTR that results in abnormal airway mucus and reduced mucus clearance. This in turn is likely to predispose to recurring infection, which continues to have the greatest impact on mortality and morbidity in CF. Treatment to improve mucociliary clearance has been proposed to at least retard this progressive destructive process, to provide an adjunct to physical therapies and to reduce the reliance on antimicrobial use.

A series of small proof‐of‐concept trials included in this review initially showed that hypertonic saline resulted in an improvement in measures of mucociliary clearance over isotonic saline and cough alone (Riedler 1996; Robinson 1996; Robinson 1997; Robinson 1999). They showed that a dose of 7% hypertonic saline was more effective than 3%, but there was no significant advantage gained by increasing the dose to 12% or by adding amiloride. While they investigated the dose‐response effect of varying concentrations of saline, they did not look at the impact of the volume nebulised.

The evidence available for hypertonic saline in stable CF lung disease has increased since this review was first published in 1999. However, the conclusions remain limited to a few relatively large multicentre trials that have examined the issue in heterogenous patient populations; and unfortunately there are no data on the effects of hypertonic saline in combination with rhDNase. The Elkins 2006a trial, while not demonstrating a sustained improvement in lung function, remains promising by demonstrating a reduction in exacerbation frequency and a small improvement in QoL. This effect was not seen in children without airflow limitation as measured by spirometry (FEV₁ over 80% predicted) (Amin 2010). It did improve mucociliary clearance and ventilation inhomogeneity as measured by the LCI in those with mild lung disease, though the long‐term implications of this finding are unclear. When delivered following a bronchodilator it is an inexpensive additional therapy for people with CF, which does not appear to be associated significant adverse effects (Amin 2010). In younger children hypertonic saline did not improve measures of infant lung function or reduce pulmonary exacerbations (Rosenfeld 2012).

We chose lung function as our primary outcome because of its relationship in the long term with mortality in CF and by inference to see if hypertonic saline would alter this decline (Courtney 2007). Elkins 2006a performed the largest trial in adults and children over six years of age, to address whether hypertonic saline (7%) compared to isotonic saline (0.9%) would improve FEV₁ after 48 weeks of treatment. The trial failed to demonstrate a significant improvement in lung function over 48 weeks, although our pooled analysis of both the smaller trial by Eng 1996 together with the Elkins 2006a trial did demonstrate a small improvement in FEV₁ after four weeks of treatment.

One limitation of using lung function markers such FEV₁ is that, despite their association with long‐term outcomes, they may be relatively insensitive to small changes in CF, especially over time given the progressive nature of the disease. Also, those with intact lung function who may benefit most in the long term from hypertonic saline would be less likely to demonstrate an improvement. With this in mind, Amin 2010 used the LCI, which is a novel measure of ventilation inhomogeneity using an inert gas multiple‐breath washout technique and which has been validated in CF and shown to be more sensitive in detecting early airway disease (Gustafsson 2008). Amin proposed that in those with only mild airflow limitation this would be more sensitive to change with hypertonic saline, and demonstrated this after four weeks of treatment, although there was no improvement seen with FEV₁ or even improvement in symptom scores (Amin 2010). It remains to be determined if this change will result in significant long‐term improvement in health for individuals with CF or how this may relate to the progression of CF lung disease.

It is important to realise that before participants started the trials, they were pretreated with salbutamol and lung function was measured to determine whether they could tolerate hypertonic saline; only those who did tolerate it were allowed to carry on. This means that the results of the trials only apply to those people with CF who can tolerate hypertonic saline.

Ballmann 1998 compared hypertonic saline 5.75% 10 mL twice‐daily to nebulised rhDNase 2.5 mg and found that, in both groups, FEV₁ improved to a similar degree in three weeks. In the Suri 2001 trial, twice‐daily 5 mL hypertonic saline 7% was compared to daily and alternate‐day rhDNase 2.5 mg. When treated with either rhDNase regimen, participants had a significant improvement in lung function from baseline, but when treated with hypertonic saline there was an increase of only 3% from baseline, less than seen in the Ballmann 1998 and Eng 1996 trials. Both of these trials used 10 mL of hypertonic saline, compared to 5 mL as used by Suri 2001 and Weller 1980. This raises the possibility that the effectiveness of treatment may also depend on the total volume of saline nebulised and this may account for the lower effect size seen by Suri 2001. In addition, Suri 2001 found a wide variation in response to treatment with both rhDNase and hypertonic saline, with over 50% of participants demonstrating a more than 10% increase in FEV₁ with rhDNase and 35% of participants with hypertonic saline. Thus, despite the overall reduced effect seen with hypertonic saline on lung function in individuals, both rhDNase and hypertonic saline have the potential to substantially improve lung function in the medium term. The wide variation seen in response to both hypertonic saline and rhDNase raises the possibility that there may be subgroups of people with CF who are more likely to respond to efforts to improve mucociliary clearance. It also suggests that some individuals may respond better to one treatment compared to the other, and physicians may wish to consider this particularly in individuals who fail to respond to rhDNase.

Quality of the evidence

Using GRADE, we judged the certainty of the evidence from this review to be of very low to low certainty, depending on the outcome measured (Table 1; Table 2; Table 3; Table 4; Table 5; Table 6).

We do not think that the way the trials were designed affected the results, with even the cross‐over trials allowing for adequate washout at least in terms of lung function change. We judged that all participants had equal chances of being in either of the treatment groups. However, in all but two trials (Dentice 2016; Elkins 2006a), the taste was not masked and participants would be able to identify the hypertonic saline. The participants who dropped out of the trials appear to have been accounted for and were unlikely to influence the results.

Potential biases in the review process

Two review authors independently extracted data using predefined data extraction sheets, and then a consensus was reached in regard to inclusion or exclusion. However, two authors (PW and VM) were involved in one trial (Elkins 2006a) and one author (PW) was involved in a further trial (Dentice 2016). Where one or more author(s) of this Cochrane Review was a co‐author on an included trial, a third party performed the data extraction and assessment of quality for that trial.

Agreements and disagreements with other studies or reviews

The first therapy used to enhance mucus clearance was rhDNase; it is effective in improving lung function in CF and has been adopted in most countries (Ramsey 1994). A separate systematic review found that, when compared with placebo, rhDNase improved lung function in people with CF in trials lasting from one month to two years (Yang 2021). There was a decrease in pulmonary exacerbations in trials of six months or longer. Voice alteration and rash were the only adverse events reported with increased frequency in randomised controlled trials. However, investigators also concluded that there was not enough evidence to firmly conclude if rhDNase is superior to hyperosmolar agents such as hypertonic saline in improving lung function. Nebulised rhDNase is a relatively expensive treatment, and in many countries its use is restricted to those who have degrees of impairment of lung function and who demonstrate an improvement in pulmonary function tests during a trial period (Ramsey 1994).

Similar to the authors of the Cochrane Review of rhDNase (Yang 2021), we do not believe there is sufficient evidence to conclusively show superiority of rhDNase over hypertonic saline; however, the effect size does appear to be greater for the former.

The other hyperosmolar agent that has been adapted for use as a treatment in CF is inhaled mannitol. This treatment is the subject of a Cochrane Review which found that, at least for up to six months, mannitol led to an improvement in lung function (Nevitt 2020). The review did not find that it improved QoL, but it did reduce exacerbation frequency. While mannitol was assessed against control, against rhDNase and in combination with rhDNase, it was not assessed against hypertonic saline.

Authors' conclusions

Implications for practice.

There is low to very low‐certainty evidence showing the benefits of hypertonic saline compared to placebo in those over the age of 12 years with cystic fibrosis (CF) in terms of improvement in lung function, at least in the short term, but it should not be used in preference to dornase alfa (rhDNase). At this stage the benefit appears to be a modest reduction in frequency of pulmonary exacerbations, and an improvement in chest symptoms, though evidence does not exist to say in whom it works best. In children under the age of 12 years we have not found sufficient evidence to justify its routine use. In those over the age of six years, an improvement in exacerbation frequency was not seen, and in this cohort of children with milder lung disease, the modest benefits seen with hypertonic saline in earlier trials could not be detected. One small cross‐over trial in 10 children reported a small improvement, but this was not seen in our analysis (very low‐certainty evidence).

In preschool children, hypertonic saline led to an improvement in lung clearance index (LCI). It has been argued that, in this age group, LCI is a surrogate clinical predictor of early lung disease; it is a test that can readily be performed in the clinic and is more responsive to change than other parameters of infant lung function ‐ an effect that is consistent with our analysis. The effect size in change in LCI, while positive, was also modest. The implication that this may have for longer‐term outcomes such as decline in lung function is unclear.

An important consideration now is the place of therapy with hypertonic saline in the context of cystic fibrosis transmembrane conductance regulator (CFTR) modulator use. Especially in those individuals with minimal bronchiectasis and sputum production, the place for additional inhalational therapies to improve mucus clearance is now uncertain.

The variation in response seen in individuals to both rhDNase and hypertonic saline raises the possibility that certain individuals will respond better to one agent compared to the other. In order to assess the individual response to therapy, it would therefore be reasonable to conduct a modified N of 1 trial in people with CF where the person acts as their own control, on and off treatment. Suitable individuals to consider therapy are those in the clinic who fail to respond, are ineligible or are unable to tolerate rhDNase; with hypertonic saline used as an alternative agent to increase mucociliary clearance. Unfortunately, there are no data on the effects of hypertonic saline in combination with rhDNase.

In the majority of trials hypertonic saline was used after pretreatment with bronchodilators and as an adjunct to chest physiotherapy; in both cases this may be important to ensure its efficacy. When delivered following a bronchodilator, hypertonic saline is an inexpensive and safe therapy for people with CF.

Implications for research.

The effect hypertonic saline has on mucus clearance, and consequently short‐term improvements in lung function, is relatively small in all age groups. The most relevant consideration now is the place of therapy with hypertonic saline in people with CF on CFTR modulators. Important questions now to consider would include: what is their benefit, if any, in the context of modulator use? Who should be targeted in therapy? And when should hypertonic saline be considered? Finally, when can mucus clearance therapies be safely withdrawn? Considering the clear benefit of modulator therapy in those who are suitable, these are now the most important questions to consider regarding the use of all mucus clearance therapies.

Future assessments of efficacy for agents that improve mucociliary clearance should be assessed in robust longer‐term randomised controlled trial designs. These should consider alternative primary outcomes, such as pulmonary exacerbations or validated quality of life (QoL) measures, as these appear to be more responsive to change for this intervention than conventional lung function tests. Measurement of pulmonary exacerbation rates are difficult and would benefit from tighter clinical definitions, especially around milder exacerbations. The emergence of more sensitive measures of early disease, such as LCI or chest computer tomography (CT), may provide novel outcomes that are a more sensitive index to change, as has been the case in preschool children. However, their role in terms of clinical outcomes in the long term remains less certain. It may be possible to obtain these data in retrospect from national data registries. Equally, it will be important to determine the impact that CFTR modulators have on the use of these therapies, their uptake and adherence.

There is also a good case for undertaking equivalence trials comparing hypertonic saline to other interventions, as hypertonic saline is likely to remain less costly, especially where new treatments tend to be associated with increased costs.

Consideration should be given to defining whether there are groups of individuals who will respond better to hypertonic saline or other mucociliary clearance agents, especially rhDNase, to better tailor treatment. Any future trials of hypertonic saline should determine who will benefit most from this intervention. Trials that combine mucociliary agents or physical therapy, or both, also need to be considered to assess efficacy.

What's new

Date	Event	Description
14 June 2023	New citation required but conclusions have not changed	A new author has joined the review team (Sherie Smith). Seven new trials have been included at this update (353 participants). Four trials were added to the existing comparison of hypertonic saline 3% to 7% versus isotonic saline in stable lung disease (Amin 2016; Donaldson 2020; PRESIS 2019; SHIP 2019). We have included two new comparisons: hypertonic saline compared to xylitol (two trials, 90 participants; NCT01355796; Singh 2020) and 7% hypertonic saline compared to 3% hypertonic saline (one trial, 30 participants; Gupta 2012). Despite these changes, our conclusions remain the same.
14 June 2023	New search has been performed	A search of the Cochrane Cystic Fibrosis and Genetic Disorders Review Group's Cystic Fibrosis Trials Register identified 23 new references potentially eligible for inclusion in this updated review. Included studies Two additional references were to two already included studies (Gupta 2012; Mainz 2015). Five studies previously listed as awaiting classification have been included at this update (Amin 2010; Donaldson 2020; PRESIS 2019; SHIP 2019; Singh 2020) with nine additional references added to four of these: two references were added to the Donaldson study (Donaldson 2020); four references were added to the PRESIS study (PRESIS 2019); two references were added to the SHIP study (SHIP 2019); and one reference was added to the Singh study (Singh 2020).  On further investigation a trial registry entry (NCT02276898) previously listed as ongoing was identified as the protocol for an included study (Amin 2016). Excluded studies Two additional references were added to two already excluded studies (Aquino 2012; San Miguel‐Pagola 2016). Two additional references were added to two studies previously listed as awaiting classification and now excluded at this update (Brown 2010; Corcoran 2017).  Six additional references were added to the SHIP‐CT study previously listed as ongoing and moved to excluded studies at this update (Tiddens 2022). Studies awaiting classification One new study (two references) is listed as awaiting classification (SIMPLIFY 2022) Trial Registry searches Seven trials were identified from the trial registry searches; four of these were excluded (ACTRN12619001681145; ACTRN12621000855820; IRCT20191112045413N2; IRCT20201017049055N1); one has been listed as ongoing (ISRCTN14081521), one entry is listed under the study already listed as awaiting classification (SIMPLIFY 2022) and one was deleted immediately as it was not relevant.

History

Protocol first published: Issue 3, 1998
Review first published: Issue 2, 1999

Date	Event	Description
11 September 2018	New citation required and conclusions have changed	The conclusions have been amended in light of the new evidence. The addition of new data has allowed us to conclude that hypertonic saline appears to be an effective adjunct to physiotherapy during acute exacerbations of lung disease in adults.
11 September 2018	New search has been performed	A major update of the review was conducted with 90 new references identified from searches of the Cochrane Cystic Fibrosis and Genetic Disorders Trials Register, clinicaltrials.gov and the WHO ICTRP. New trials Five important new trials (24 references) have been included (Amin 2010; Dentice 2016; Laube 2009; Mainz 2015; Rosenfeld 2012). This increased the number of participants in the trial from 442 to 966. Nine new trials (27 references) have also been excluded (Brivio 2016; Buonpensiero 2010; DeCono 2008; Dentice 2012; EUCTR2007‐002707‐40‐BE; Grasemann 2013; IRCT20180307038994N1; NCT01094704; O'Neill 2017; Ros 2012; San Miguel‐Pagola 2016; Van Ginderdeuren 2008; Van Ginderdeuren 2011). 16 newly identified trials (21 references) are currently awaiting assessment until further information is available to allow inclusion or exclusion (Amin 2016a; Balinotti 2015; Brown 2010a; Corcoran 2017a; Donaldson 2013a; Dwyer 2013; Hofmann 1997; Singh 2019a; NCT01355796a; NCT01377792a; NCT01619657; SHIP 2019a; Lennox 2016; Nenna 2017a; Palacio 2014; PRESIS 2018a). Three new trials (four references) are listed as ongoing (NCT02276898; NCT02343445; NCT02950883a). Previously identified trials 11 additional references have been added to five already included trials (Elkins 2006a; Eng 1996; Robinson 1997; Robinson 1999; Weller 1980). One additional reference has been added to an already excluded trial (Donaldson 2006). Two trials (two references) which were previously listed as 'Awaiting classification' have been excluded (Elkins 2006b; Vanlaethem 2008). An additional reference to the Elkins trial was identified in the latest searches. One trial (one reference) previously excluded has been moved to 'Awaiting classification' pending further information (Hofmann 1997).
26 April 2010	Amended	Contact details updated.
31 December 2008	New citation required and conclusions have changed	A large multi‐centre trial has been included in the review, resulting in the conclusions being updated (Elkins 2006a).
31 December 2008	New search has been performed	A search of the Group's trials register identified two new eligible trials (Cardinale 2003; Elkins 2006a). These trials have now been included in the review. A trial previously listed in Studies awaiting classification has now been included in the review (Adde 2004). Two studies identified by the searches have now been excluded from the review (Donaldson 2006; Kobylyansky 2000). A further reference to the already included Suri trial has been added (Suri 2001). Two trials have been added to Studies awaiting classification (Elkins 2006b; Vanlaethem 2008).
12 November 2008	Amended	Converted to new review format.
2 May 2005	New citation required but conclusions have not changed	No new trials have been included. One trial is currently awaiting assessment (Adde 2004). Individual patient data have kindly been provided by Dr Adde and will be included in a future update of the review. Further data have been included for the Suri trial (Suri 2001). The layout of the analysis has been changed from the previous updates to ensure clarity for the reader.
2 October 2003	New search has been performed	Six additional references to the already included Suri trial have been added (Suri 2001). This did not add greatly to the data already present. Two of the references looked at the cost of treatment with HS and rhDNase and additional data has been added. One other reference looked at the effect of rhDNase and HS on airway inflammation and the data has now been included. The other references did not add to the data already presented. One additional reference to the already included Ballmann trial has been added (Ballmann 1998). No additional data has been included in the review. One additional reference to the already included Riedler trial has been added (Riedler 1996). No additional data has been included in the review.
3 November 2002	New search has been performed	An additional trial was found and incorporated in the review (Suri 2001). This was a relatively large clinical trial comparing hypertonic saline and rhDNase. Significant changes have been made to the review.
3 November 2001	New search has been performed	With this update significant changes to style were made particularly in the order of the outcomes and the presentation of the results. The Suri trial was added and this contained a large amount of additional information particularly concerning the effect of hypertonic saline versus DNase on lung function (Suri 2001).

Appendices

Appendix 1. Electronic search strategies

Database or Resource	Strategy
Clinicaltrials.gov	[Advanced Search Form] Other Terms: hypertonic Study Type: Interventional Studies Condition/ Disease: cystic fibrosis
WHO ICTRP	cystic fibrosis AND hypertonic

Data and analyses

Comparison 1. Hypertonic saline 3% to 7% versus isotonic saline.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Change in FEV1 (% predicted)	5		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
1.1.1 At 2 to 4 weeks (all participants)	4	246	Mean Difference (IV, Fixed, 95% CI)	3.30 [0.71, 5.89]
1.1.2 At 2 to 4 weeks (participants aged 14 years and over)	2	205	Mean Difference (IV, Fixed, 95% CI)	4.15 [1.14, 7.16]
1.1.3 At 2 to 4 weeks (children FEV1 > 80% predicted)	1	20	Mean Difference (IV, Fixed, 95% CI)	‐0.42 [‐7.45, 6.61]
1.1.4 At 12 weeks	1	149	Mean Difference (IV, Fixed, 95% CI)	4.10 [‐0.08, 8.28]
1.1.5 At 24 weeks	1	140	Mean Difference (IV, Fixed, 95% CI)	5.37 [1.03, 9.71]
1.1.6 At 36 weeks	1	134	Mean Difference (IV, Fixed, 95% CI)	3.63 [‐1.56, 8.82]
1.1.7 At 48 weeks	1	134	Mean Difference (IV, Fixed, 95% CI)	2.31 [‐2.72, 7.34]
1.1.8 24 hours after single dose	1	36	Mean Difference (IV, Fixed, 95% CI)	0.30 [‐3.48, 4.08]
1.2 Change in FVC (% predicted)	3		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
1.2.1 At 2 to 4 weeks	3	225	Mean Difference (IV, Fixed, 95% CI)	1.07 [‐1.63, 3.78]
1.2.2 At 12 weeks	1	141	Mean Difference (IV, Fixed, 95% CI)	4.56 [0.79, 8.33]
1.2.3 At 24 weeks	1	140	Mean Difference (IV, Fixed, 95% CI)	3.64 [0.17, 7.11]
1.2.4 At 36 weeks	1	134	Mean Difference (IV, Fixed, 95% CI)	3.40 [‐0.82, 7.62]
1.2.5 At 48 weeks	1	134	Mean Difference (IV, Fixed, 95% CI)	2.76 [‐1.09, 6.61]
1.3 Mean change in FEV0.5 (mL)	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.3.1 At 48 weeks	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.4 Mean change in FEV0.75 (L)	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
1.4.1 12 weeks	1	150	Mean Difference (IV, Fixed, 95% CI)	0.02 [‐0.03, 0.07]
1.4.2 24 weeks	1	150	Mean Difference (IV, Fixed, 95% CI)	0.02 [‐0.04, 0.08]
1.4.3 36 weeks	1	150	Mean Difference (IV, Fixed, 95% CI)	0.04 [‐0.01, 0.09]
1.4.4 48 weeks	1	150	Mean Difference (IV, Fixed, 95% CI)	0.01 [‐0.04, 0.06]
1.5 LCI	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.5.1 At 4 weeks	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.6 LCI change from baseline	2	190	Mean Difference (IV, Fixed, 95% CI)	‐0.60 [‐1.00, ‐0.19]
1.6.1 At 12 months	2	190	Mean Difference (IV, Fixed, 95% CI)	‐0.60 [‐1.00, ‐0.19]
1.7 Radiolabelled isotope clearance (%)	3		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
1.7.1 At 60 mins	3	68	Mean Difference (IV, Fixed, 95% CI)	6.14 [2.56, 9.72]
1.8 Mucociliary clearance measured as area under the curve	2		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
1.8.1 Single dose	2	44	Mean Difference (IV, Fixed, 95% CI)	‐212.06 [‐271.64, ‐152.48]
1.9 Change in mucocilliary clearance rate (% clearance)	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.9.1 At 4 weeks	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.10 Exercise capacity (using a subjective visual analogue score)	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
1.10.1 At Week 1	1	52	Mean Difference (IV, Fixed, 95% CI)	0.88 [0.19, 1.57]
1.10.2 At Week 2	1	52	Mean Difference (IV, Fixed, 95% CI)	1.01 [0.18, 1.84]
1.11 Quality of life (change from baseline)	4		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
1.11.1 CFQ parent	3	365	Mean Difference (IV, Fixed, 95% CI)	1.62 [‐1.69, 4.92]
1.11.2 CFQ 14+	1	91	Mean Difference (IV, Fixed, 95% CI)	7.77 [1.86, 13.68]
1.11.3 SF36	1	91	Mean Difference (IV, Fixed, 95% CI)	2.84 [‐7.90, 13.58]
1.11.4 CFQ‐R respiratory domain at 4 weeks	1	20	Mean Difference (IV, Fixed, 95% CI)	7.40 [‐2.45, 17.25]
1.12 Feeling of cleared chest (using a subjective visual analogue scale)	2		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
1.12.1 At up to Week 1	2	72	Mean Difference (IV, Fixed, 95% CI)	0.97 [0.35, 1.60]
1.13 Average number of pulmonary exacerbations	2		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.13.1 Number requiring antibiotics	2		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.13.2 Number of visits not requiring antibiotics	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.14 Average number of hospital admissions per participant	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.14.1 At 48 weeks	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.15 Mean adherence rate	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
1.16 Change in log10 colony forming units (CFU)/g from baseline at final visit	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.16.1 P. aeruginosa	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.16.2 S. aureus	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.17 Mean time (years) to first detection of CF pathogens	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.17.1 First detection of S aureus	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.17.2 First detection of H influenzae	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.17.3 First isolation of P aeruginosa	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.18 Adverse events: acute fall in lung function	2		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.18.1 Acute fall in FEV₁ % predicted	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.18.2 Acute fall in FEV₁ mL	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.19 Adverse events 2 to 4 weeks	2		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
1.19.1 Any adverse event	1	23	Risk Ratio (M‐H, Fixed, 95% CI)	0.81 [0.59, 1.10]
1.19.2 Cough	2	75	Risk Ratio (M‐H, Fixed, 95% CI)	0.80 [0.41, 1.58]
1.19.3 Pharyngitis	2	75	Risk Ratio (M‐H, Fixed, 95% CI)	3.08 [0.36, 26.35]
1.19.4 Chest tightness	1	52	Risk Ratio (M‐H, Fixed, 95% CI)	2.79 [0.12, 65.38]
1.19.5 Haemoptysis	1	52	Risk Ratio (M‐H, Fixed, 95% CI)	0.93 [0.21, 4.17]
1.19.6 Wheezing	1	23	Risk Ratio (M‐H, Fixed, 95% CI)	0.22 [0.01, 4.93]
1.19.7 Nasal congestion	1	23	Risk Ratio (M‐H, Fixed, 95% CI)	0.43 [0.09, 2.08]
1.19.8 Increased sputum production	1	33	Risk Ratio (M‐H, Fixed, 95% CI)	1.36 [0.09, 19.88]
1.20 Adverse events 48 to 52 weeks	3		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
1.20.1 All adverse events	2	192	Risk Ratio (M‐H, Fixed, 95% CI)	1.00 [0.92, 1.09]
1.20.2 Serious adverse events	2	192	Risk Ratio (M‐H, Fixed, 95% CI)	0.85 [0.42, 1.69]
1.20.3 Cough	3	513	Risk Ratio (M‐H, Fixed, 95% CI)	1.02 [0.87, 1.19]
1.20.4 Chest tightness	1	321	Risk Ratio (M‐H, Fixed, 95% CI)	0.21 [0.01, 4.26]
1.20.5 Vomiting	1	321	Risk Ratio (M‐H, Fixed, 95% CI)	0.52 [0.13, 2.03]
1.20.6 Wheezing	1	321	Risk Ratio (M‐H, Fixed, 95% CI)	2.06 [0.38, 11.11]
1.20.7 Fever	3	513	Risk Ratio (M‐H, Fixed, 95% CI)	0.73 [0.53, 1.00]
1.20.8 Nasal congestion	2	471	Risk Ratio (M‐H, Fixed, 95% CI)	0.90 [0.63, 1.27]
1.20.9 Ear infection	2	363	Risk Ratio (M‐H, Fixed, 95% CI)	1.27 [0.74, 2.17]
1.20.10 Acquisition of Burkholderia cepacia	1	321	Risk Ratio (M‐H, Fixed, 95% CI)	1.03 [0.15, 7.23]
1.20.11 Abdominal distension / flatulence	1	42	Risk Ratio (M‐H, Fixed, 95% CI)	2.00 [0.71, 5.64]
1.20.12 Rhinorrhoea	1	42	Risk Ratio (M‐H, Fixed, 95% CI)	1.70 [1.04, 2.79]
1.20.13 Diarrhoea	1	42	Risk Ratio (M‐H, Fixed, 95% CI)	0.83 [0.30, 2.31]

Comparison 2. Hypertonic saline 3% to 7% versus isotonic saline in acute lung disease.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
2.1 Change in FEV1 from baseline (% predicted)	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
2.1.1 At day 7	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
2.1.2 At day 10	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
2.1.3 At discharge	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
2.2 Change in FVC from baseline (% predicted)	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
2.2.1 At day 7	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
2.2.2 At day 10	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
2.2.3 At discharge	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
2.3 Shuttle walk test	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
2.3.1 At day 7	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

Comparison 3. Hypertonic saline versus rhDNase.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
3.1 Change in FEV1 (% predicted)	2		Treatment difference (IV, Fixed, 95% CI)	Totals not selected
3.1.1 At 3 weeks	1		Treatment difference (IV, Fixed, 95% CI)	Totals not selected
3.1.2 At 3 months	1		Treatment difference (IV, Fixed, 95% CI)	Totals not selected
3.2 Improvement in FEV1 >10%	2		Odds Ratio (IV, Fixed, 95% CI)	Totals not selected
3.2.1 At 3 weeks	1		Odds Ratio (IV, Fixed, 95% CI)	Totals not selected
3.2.2 At 3 months	1		Odds Ratio (IV, Fixed, 95% CI)	Totals not selected
3.3 Change in FVC (% predicted)	1		Treatment difference (IV, Fixed, 95% CI)	Totals not selected
3.3.1 At 3 months	1		Treatment difference (IV, Fixed, 95% CI)	Totals not selected
3.4 Exercise tolerance ‐ oxygen saturation	1		Treatment difference (IV, Fixed, 95% CI)	Totals not selected
3.4.1 At 3 months	1		Treatment difference (IV, Fixed, 95% CI)	Totals not selected
3.5 Exercise tolerance ‐ VAS for breathlessness	1		Treatment difference (IV, Fixed, 95% CI)	Totals not selected
3.5.1 At 3 months	1		Treatment difference (IV, Fixed, 95% CI)	Totals not selected
3.6 Exercise tolerance ‐ FCS	1		Treatment difference (IV, Fixed, 95% CI)	Totals not selected
3.6.1 At 3 months	1		Treatment difference (IV, Fixed, 95% CI)	Totals not selected
3.7 Mean percentage change in quality of life score	1		Treatment difference (IV, Fixed, 95% CI)	Totals not selected
3.7.1 At 3 months	1		Treatment difference (IV, Fixed, 95% CI)	Totals not selected
3.8 Delivery time (minutes)	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

Comparison 4. Hypertonic saline versus mannitol.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
4.1 Delivery time (mins)	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

Comparison 5. Hypertonic saline 7% versus xylitol.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
5.1 FEV1 % predicted mean change from baseline	2		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
5.1.1 At 14 days	2	119	Mean Difference (IV, Fixed, 95% CI)	1.89 [‐0.80, 4.58]
5.2 FVC % predicted mean change from baseline	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
5.2.1 At 14 days	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
5.3 Adherence to treatment ‐ mean number of doses taken	1		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
5.4 Density of colonisation of pathogens ‐ change from baseline log CFU/mL	2		Mean Difference (IV, Fixed, 95% CI)	Subtotals only
5.4.1 P aeruginosa	2	94	Mean Difference (IV, Fixed, 95% CI)	0.62 [‐0.90, 2.14]
5.4.2 P aeruginosa mucoid	1	40	Mean Difference (IV, Fixed, 95% CI)	‐1.80 [‐3.48, ‐0.12]
5.4.3 S aureus	1	40	Mean Difference (IV, Fixed, 95% CI)	0.50 [‐1.03, 2.03]
5.5 Adverse events	2		Risk Ratio (M‐H, Fixed, 95% CI)	Subtotals only
5.5.1 Any adverse event	1	60	Risk Ratio (M‐H, Fixed, 95% CI)	1.00 [0.07, 15.26]
5.5.2 Serious adverse events	2	119	Risk Ratio (M‐H, Fixed, 95% CI)	4.07 [0.47, 35.34]
5.5.3 Gastrointestinal disorders (bowel obstruction)	1	59	Risk Ratio (M‐H, Fixed, 95% CI)	7.23 [0.39, 134.16]
5.5.4 Hepatobiliary disorders (abnormal lab results)	1	59	Risk Ratio (M‐H, Fixed, 95% CI)	1.29 [0.38, 4.34]
5.5.5 Bronchospasm	1	59	Risk Ratio (M‐H, Fixed, 95% CI)	0.21 [0.01, 4.13]
5.5.6 Haemoptysis	1	59	Risk Ratio (M‐H, Fixed, 95% CI)	2.07 [0.20, 21.60]
5.5.7 Vascular disorders	1	59	Risk Ratio (M‐H, Fixed, 95% CI)	0.34 [0.01, 8.13]

Comparison 6. Hypertonic saline 7% versus hypertonic saline 3%.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
6.1 FEV1 % predicted (absolute value)	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
6.1.1 At 28 days (endpoint)	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
6.2 FEV1 % predicted mean percentage change from baseline	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
6.2.1 At 28 days	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
6.3 FVC % predicted (absolute value)	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
6.3.1 At 28 days (endpoint value)	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
6.4 FVC % predicted mean percentage change from baseline	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
6.4.1 At 28 days	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
6.5 Exercise capacity ‐ PEFR (L/min) via a 3‐minute step test	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
6.5.1 PEFR endpoint values at 28 days	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
6.6 Exercise capacity ‐ SaO2 via a 3‐minute step test	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
6.6.1 SaO2 % at 28 days	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
6.7 Exercise capacity ‐ heart rate (bpm) via a 3‐minute step test	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
6.7.1 At 28 days	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
6.8 Breathlessness via a 3‐minute step test	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
6.8.1 VAS of perceived breathlessness at 28 days	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Adde 2004.

Study characteristics
Methods	RCT. Design: open label, cross‐over with 4 weeks in each treatment arm and a 2‐week washout period in between. Location: Brazil (single centre).
Participants	Total participants: n = 18, 5 male and 13 female. CF diagnosis: not stated. Age: mean (SD) 14.8 (4.8) years, range 8.7 to 25.8 years. Baseline characteristics Airway colonisation: Pseudomonas aeruginosa = 17 Staphylococcus aureus = 12 Haemophylus influenzae = 1 Shwachman score: median (range) 65 (55 – 90). Pancreatic insufficiency: n = 17. Previous use of rhDNase: n = 15. Continuous inhaled gentamicin therapy: n = 11.
Interventions	Group 1: HS 6% 10 mL. Group 2: rhDNase 2.5 mg 2x daily. Both trial drugs were delivered by a Pari LC Plus nebulizer with a Proneb® compressor, by a mouthpiece. The first inhalation was done in the hospital, for technique supervision, checking of immediate side effects, post‐medication PFT and measurement of nebulization time. Salbutamol (400 mcg) was given prior to inhalation, twice daily, in both arms of the trial. All the other treatment for CF was unchanged. While in the rhDNase arm of the trial the participant was asked to do a normal saline nebulization (5 mL) at another time during the day.
Outcomes	Change in FEV1, sputum culture bacterial growth, in vitro studies of mucus, symptom score, satisfaction with treatment.
Notes	Abstract only, but some additional information from authors. No sample size calculation stated.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	No detail provided in abstract, but additional information from authors confirmed use of a random numbers table.
Allocation concealment (selection bias)	Low risk	No detail provided in abstract, but additional information from authors confirmed that for each potential participant the selected sequence of treatment was defined and written in a piece of paper which was then put into numbered envelopes that were kept in the hospital pharmacy. After participants were recruited and informed consent obtained the first envelope was opened and read to see the allocated treatment.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Double blind, participant blinding was attempted by masking the taste of the solutions with quinine sulphate but additional info states "as participants were well aware of the medication flavour of HS this could not be masked".
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not discussed.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Did not report withdrawals. Additional data provided stated that 1 participant (not included in the analysis) had to stop treatment with HS due to severe dyspnoea during its nebulization.
Selective reporting (reporting bias)	Unclear risk	Abstract only.
Other bias	Low risk	2‐week washout period in between treatment arms. No sample size calculation stated.

Amin 2010.

Study characteristics
Methods	RCT. Design: cross‐over trial with 4 weeks treatment in each arm and 4‐week washout period. Location: Canada (multicentre).
Participants	Total participants: n = 20 randomised, 1 excluded from analysis. 7 males and 12 females. CF diagnosis as defined by two or more clinical features of CF and a documented sweat chloride > 60 mEq/L by quantitative pilocarpine iontophoresis test or a genotype showing two well characterised disease causing mutations. Mean (SD) age at baseline = 10.6 (3.1) years. Baseline characteristics Pseudomonas aeruginosa +ve: n = 7 BMI: mean (SD) 17.0 (3.0) Pancreatic insufficiency: 84% Genotype DF508/DF508 homozygous: 42% DF508 compound heterozygous: 21% Lung function: FVC % predicted: mean (SD) 101 (11.3), range (81 to 121) FEV1 % predicted: mean (SD) 96 (12), range (80 to 118) FEF25‐75% predicted: mean (SD) 84 (24), range (53 to 120)
Interventions	Group 1: 4 mL HS 7% 2x daily. Group 2: 4 mL IS 0.9% 2x daily. The solutions were administered using the PARI LC Star nebuliser. 2x 100 mg puffs of salbutamol (Ventolin) were administered before each inhalation of study solution using a holding chamber.
Outcomes	LCI, CFQ‐R (Quittner 2009), FEV1, FVC, FEF 25-75.
Notes	Investigators calculated the sample size required for testing using HS as the main exposure variable and the LCI as the primary outcome variable.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Concealed computer‐generated randomisation.
Allocation concealment (selection bias)	Low risk	Randomisation concealed as performed by a research pharmacist not otherwise involved in the trial.
Blinding of participants and personnel (performance bias) All outcomes	High risk	The solutions were indistinguishable from each other in appearance but participants could discern a difference in taste between solutions. High risk with this cross‐over design.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Blinding of researchers.
Incomplete outcome data (attrition bias) All outcomes	Low risk	20 recruited, details of missing data for 3 participants described (below), complete cross‐over data were therefore available for 17 participants. The LCI results of 1 participant failed to meet the quality control criteria for all 4 trial visits and were therefore excluded from the analysis. 1 participant receiving IS withdrew from the trial after completion of the initial 4‐week trial period because of difficulties complying with the trial protocol. 1 participant had uninterpretable LCI data at visit 2.
Selective reporting (reporting bias)	Low risk	All outcomes stated in the 'Methods' section reported in the 'Results' section.
Other bias	Low risk	4‐week washout period. Investigators calculated the sample size.

Amin 2016.

Study characteristics
Methods	RCT. Cross‐over design (1‐week washout between treatments). Duration: single inhalation of each treatment. Multicentre (2 centres) in Canada.
Participants	242 people with CF identified for the study (134 from St Michael's Hospital and 108 from the Hospital for Sick Children). 113 (47%) were excluded because they were currently using HS. 129 were approached of which 21 (16%) people consented and were enrolled in the trial. 21 participants randomised; 16 completed measurements at all of the time points; 18 participants contributed to the ITT analysis and 3 participants were excluded. Inclusion criteria: confirmed diagnosis of CF; informed consent and verbal assent (as appropriate); at least 6 years of age at enrolment; able to perform reproducible spirometry meeting American Thoracic Society standards; pre‐bronchodilator FEV1% predicted ≥ 40% predicted; able to perform reproducible LCI maneuvers at screening. Exclusion criteria: airway cultures yielding Burkholderia cepacia complex in the previous 2 years or non‐tuberculous mycobacteria in the past year; oral corticosteroid use; oxygen supplementation; lung transplantation; intravenous antibiotics or oral quinolones within 14 days of enrolment; or investigational drugs within 30 days of enrolment; use of hypertonic saline (7%) < 4 weeks before screening or outside of the study protocol; participation in any therapeutic clinical study < 4 weeks or, 5 half‐lives, whichever is longer, before screening; smoking < 3 months before screening; presence of a condition or abnormality that in the opinion of the site investigator would compromise the safety of the participant or the quality of the data. Participants who experienced a drop in FEV1 of 20% or more after study drug administration were also excluded from further study participation. Age median (range): 14 years (7 ‐ 56). Gender n (%): 11 (61.1%) females. LCI median (range): 122 (6.5 ‐ 21.1). FEV1% predicted median (range): 86.9 (47.4 ‐ 102.8). FVC % predicted median (range): 97.9 (61.9 ‐ 121.8). FEF25%-75% predicted median (range): 67.1 (19.5 ‐ 99.2). Delta F508 homozygous n (%): 16 (88.9%). Pseudomonas aeruginosa positive sputum culture n (%): 17 (94.4%).
Interventions	Intervention: single inhalation of 4 mL HS (7%). Control: single inhalation of 4 mL IS (0.9%). The solutions were administered using the PARI LC Star nebuliser (Pari, Midlothian, Virginia, USA). 2x 100 ug puffs of salbutamol were administered before each inhalation of study solution using a holding chamber (Aerochamber Max, Trudell, London, Canada).
Outcomes	MBW and spirometry were performed at each visit ‐ at baseline, 1, 2, 4 and 24 h post inhalation. Primary outcome LCI change from baseline to 24 h post inhalation Secondary outcomes LCI SF6 change from baseline to 24 h post inhalation Spirometry (FEV1, FEF25‐75) change from baseline to 24 h post inhalation Changes in LCI N2, LCI SF6 and spirometric outcomes over 24 h (i.e. slope analysis of all time points) Adverse events ClinicalTrials protocol also stated that FVC % predicted would be measured
Notes	All tests were performed in the research pulmonary function laboratory at the Hospital for Sick Children, Toronto Canada between April 2012 and September 2014. The trial was stopped because of difficulty with recruitment. Dr. Felix Ratjen has acted as a consultant for Novartis Pharmaceuticals. Funding support as well as the HS used in the study was provided by Novartis Pharmaceuticals.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Participants were randomised to HS first or IS first via a concealed, computer‐generated randomisation performed by a research pharmacist who was not otherwise involved in the study.
Allocation concealment (selection bias)	Low risk	The allocation was concealed via the computer‐generated randomisation procedure.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Personnel and participants were blinded to intervention.The solutions were indistinguishable from each other in appearance, but not in taste.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Outcome assessors were blinded and the primary efficacy analysis was also carried out blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	21 participants were randomised, 3 were excluded and 18 were analysed on an ITT basis. Sample size calculation was undertaken (target enrolment was 24 participants).
Selective reporting (reporting bias)	Low risk	All outcomes listed in the methods and trial registration document are reported in the results with the exception of FVC but this does not affect the overall results of the study.
Other bias	Low risk	No other risk of bias identified.

Ballmann 1998.

Study characteristics
Methods	RCT. Design: cross‐over with 2 treatment periods of 3 weeks each and a washout period of 3 weeks in between. Location: Germany.
Participants	Total participants: n = 14, 8 males and 6 females. FEV1 % predicted had to be greater than 40%. Baseline characteristics FEV1 % predicted: mean (SD) 75.6% (14%). Pseudomonas aeruginosa: chronic colonisation: n = 9; free: n = 3; intermittent: n = 2.
Interventions	Pre‐treated salbutamol 200 mcg MDI inhaled (2 puffs). Group 1: 10 mL HS 5.85% 2x daily. Group 2: 2 mL Pulmozyme 2.5 mg 2x daily. Treatment was given via a Pari Master with Pari LL Routine medication not altered during trial.
Outcomes	Change in FEV1 as a % of predicted, nebulisation time, comparison of cost (in Deutschmarks), preference.
Notes	There was a 3‐week washout between interventions. No sample size calculation stated.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Participants randomised into groups of 4 and drew lots to start with HS or rhDNase.
Allocation concealment (selection bias)	Unclear risk	No details provided as to how lots concealed allocation.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Taste of HS and difference in volume made blinding not possible.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not discussed.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	No information provided whether an ITT was used.
Selective reporting (reporting bias)	Low risk	All outcomes stated in the 'Methods' section reported in the 'Results' section.
Other bias	Low risk	Washout period of 3 weeks. No sample size calculation stated.

Cardinale 2003.

Study characteristics
Methods	RCT. Design: parallel, placebo‐controlled. Location: Italy.
Participants	Total participants: n = 25. Treatment group: n = 12. Placebo group: n = 13. Baseline characteristics: stable disease.
Interventions	Group 1: HS 7%. Group 2: IS 0.09%. Frequency of nebulization not stated.
Outcomes	Change in lung function and exhaled nitric oxide measures, cough, sputum production. Measurements at baseline and 2 weeks.
Notes	Abstract only. 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	No detail provided.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	No detail provided.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No detail provided.
Incomplete outcome data (attrition bias) All outcomes	High risk	Cardinale stated that no adverse events were reported with HS, but no information given for IS group.
Selective reporting (reporting bias)	Unclear risk	Abstract only, no detail provided.
Other bias	Unclear risk	No sample size calculation stated.

Chadwick 1997.

Study characteristics
Methods	RCT Design: cross‐over with 3 arms. Duration of intervention is not stated.
Participants	Total participants: n = 15. Groups stratified according to FEV1 (over 70% and 40% ‐ 70%).
Interventions	Group 1: IS. Group 2: HS 3.5%. Group 3: hypotonic saline.
Outcomes	Change in FEV1 % predicted, nebulisation.
Notes	Abstract only. 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	No detail provided.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Stated as single blind, possibly assessors who were blinded due to difficulties in masking taste of intervention, but not clear.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Stated as single blind, possibly assessors who were blinded due to difficulties in masking taste of intervention, but not clear.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	No description of any dropouts given.
Selective reporting (reporting bias)	Unclear risk	Abstract only, no detail provided.
Other bias	Unclear risk	No sample size calculation stated. No washout period stated.

Dentice 2016.

Study characteristics
Methods	RCT. Design: parallel. Duration: 3‐day study which began towards the end of hospitalisation for an exacerbation (approximately 14 days) with follow‐up for 1 year. If re‐admission within that year, participant invited to repeat 3‐day study. Location: Australian multicentre trial.
Participants	132 adults with CF admitted to hospital with a respiratory exacerbation. Age: mean (SD) 28 (9) years. 49% were female. FEV1 % predicted: mean (SD) 48 (20)%.
Interventions	Group 1: 3x daily nebulisation of 4 mL HS 7%. Group 2: IS 0.12% (taste‐masked). Treatment was administered via Pari LC with jet nebuliser Interventions given immediately before or during physiotherapy.
Outcomes	Length of hospital stay, lung function, oxygenation, bacterial load, symptom scores, QoL, exercise tolerance, time to relapse.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Done sequentially upon admission.
Allocation concealment (selection bias)	High risk	Allocation was sequential.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Taste of intervention and control masked by quinine.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Investigators blinded to intervention.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No drop outs.
Selective reporting (reporting bias)	Low risk	All outcomes reported.
Other bias	Low risk	None apparent.

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.

Elkins 2006a.

Study characteristics
Methods	RCT. Design: parallel, double‐blind. Location: 16 adult or paediatric hospitals in Australia. Duration: 48 weeks. The trial was conducted between September 2000 and November 2003.
Participants	Total participants: n = 164, 93 males and 71 females, aged over 6 years. HS Group (n = 83) Age, mean (SD): 18.4 (9.3) years. Sex: 46% female. BMI, mean (SD): 19.9 (3.9). FEV1, mean (SD), range: 73 (21), 40 – 132. FVC, mean (SD), range: 85 (18), 45 ‐ 127. FEF25%-75%, mean (SD), range: 56 (34), 11 ‐ 155. Sputum: Pseudomonas aeruginosa: 79%; Staphylococcus aureus: 44%. Control Group (n = 81) Age, mean (SD): 18.7 (9.2) years. Sex: 42% female. BMI, mean (SD): 20.1 (3.6). FEV1, mean (SD), range: 76(21), 40 – 127. FVC, mean (SD), range: 88 (18), 44 ‐ 137. FEF25%-75%, mean (SD), range: 61 (35), 10 ‐ 151. Sputum: Pseudomonas aeruginosa: 78%; Staphylococcus aureus: 47%.
Interventions	Group 1: 4 mL HS 7% 2x daily. Group 2: 4 mL IS 0.09% 2x daily. Solutions were prepared by Pfizer, quinine sulphate (0.25 mg per mL) was added as a taste‐masking agent. Solutions were nebulized with a Pari LC PLUS jet nebulizer and a Pari Proneb Turbo compressor. A bronchodilator was administered before each inhalation of the trial solution. All other standard care was maintained throughout the trial.
Outcomes	Mean change in FEV1 and FVC at 4, 12, 36 and 48 weeks. QOL and pulmonary exacerbations.
Notes	Sample size calculation undertaken.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Concealed computer randomisation with minimisation algorithm to balance for age, FEV1 and long‐term treatment with rhDNase, use of physiotherapy and trial centre.
Allocation concealment (selection bias)	Low risk	Randomisation performed by a person not otherwise involved in the trial.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blind, participant blinding was achieved by masking the taste of the solutions with quinine sulphate. Participants and their clinicians, remained unaware of the treatment assignments throughout the trial.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	The research assistants and the trial coordinator remained unaware of the treatment assignments throughout the trial.
Incomplete outcome data (attrition bias) All outcomes	Low risk	2 participants (1 from each group) withdrew voluntarily after randomisation and before first dose. 82 in HS group and 80 in control group included in ITT analysis. Withdrawals described as follows: HS Group: 15 in total withdrawn: 7 lost to follow‐up (2 owing to time constraints; 2 owing to insufficient perceived benefit from trial solution; 2 owing to adverse reaction to trial solution (cough); 1 provided no reason) and 8 stopped inhalations but continued visits (4 had adverse reaction to trial solution; 1 had cough and vomiting; 1 had pharyngitis and wheezing; 1 had voice changes; 1 had chest tightness; 2 could not tolerate taste of trial solution; 1 had insufficient benefit from trial solution; 1 lost interest). Control Group: 17 in total withdrawn: 10 lost to follow‐up (5 owing to time constraints; 3 owing to insufficient perceived benefit from trial solution; 1 failed to attend; 1 provided no reason) and 7 stopped inhalations but continued visits (3 owing to time constraints; 2 had adverse reaction to trial solution (tonsillitis in 1 and lethargy in 1); 1 had insufficient benefit from trial solution; 1 provided no reason.
Selective reporting (reporting bias)	Low risk	All outcomes stated in the methods were described in the results.
Other bias	Low risk	Sample size calculation undertaken, no other potential bias identified.

Eng 1996.

Study characteristics
Methods	RCT. Design: open‐label, parallel. Duration: 2 weeks. Location: 2 centres in Australia (1 children's hospital and 1 adult unit).
Participants	Total participants: n = 58 randomised, 6 withdrew during trial. Inclusion criteria Diagnosis of CF with positive sweat chloride test. Able to do pulmonary function tests. Cough and daily sputum production. Regular chest physiotherapy at home. Reasonable distance from clinic. On stable medications regimen for last 14 days. Exclusion criteria > 20% fall in FEV1 at baseline assessment. Exacerbation of CF in last 4 weeks requiring admission to hospital. Exacerbation requiring admission to hospital during trial period. HS Group (n = 27) Sex split: 18 males, 9 females. Age, mean (range): 16.1 (7 ‐ 25) years. Height, mean (SD): 155 (17.7) cm. Weight, mean (SD): 47 (14.5) kg. FEV1 % predicted, mean (SD): 50.0 (9.7). FVC % predicted, mean (SD): 73.5 (15.9). IS Group (n = 25) Sex split: 13 males, 12 females. Age, mean (range): 16.7 (8 to 36) years. Height, mean (SD): 150 (19.6) cm. Weight, mean (SD): 43 (16.2) kg. FEV1 % predicted, mean (SD): 53.7 (7.8). FVC % predicted, mean (SD):77.2 (9.8).
Interventions	Pre‐treated salbutamol 600 mcg MDI and volumatic spacer device. Group 1: 10 mL HS 6% 2x daily. Group 2: 10 mL IS 2x daily. Inhaled using ultrasonic nebuliser (Omron NE‐U 07). First and last inhalations administered in clinic under a doctor's supervision.
Outcomes	Mean change in FEV1 at 2 weeks Mean change FVC at 2 weeks VAS for cleared chest at 1 and 2 weeks VAS for dyspnoea, fatigue, appetite, exercise tolerance, quality of sleep, general well‐being; adverse effects (increased cough, haemoptysis, chest tightness and pharyngitis)
Notes	Sample size calculation undertaken.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Random number tables used.
Allocation concealment (selection bias)	Unclear risk	No detail provided.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Open‐label, participants not told which group they were in, but not able to disguise trial drug due to salty taste.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Technician measuring lung function was blinded to treatment assignment.
Incomplete outcome data (attrition bias) All outcomes	Low risk	6 participants withdrew in total (3 from each group) due to: non‐compliance with clinic visit (1 in IS group); exacerbation of respiratory infection requiring hospital admission (2 IS group, 1 HS group); irritating cough during inhalation (1 HS group); increased haemoptysis which occurred 3 hours after first treatment (1 HS group). No specific mention of whether an ITT analysis had been performed.
Selective reporting (reporting bias)	Low risk	All outcomes stated in the methods were described in the results.
Other bias	Low risk	Sample size calculation undertaken, no other bias identified.

Gupta 2012.

Study characteristics
Methods	RCT. Design: double‐blind, parallel design Duration: 28 days. Location: single‐centre study conducted in India.
Participants	Children with CF (diagnosed based on 2 abnormal sweat test results (sweat chloride > 60 mEq) in the presence of suggestive clinical features) who are being followed up in clinic, aged between 6 and 16 years. 36 assessed for eligibility 31 randomised, 30 participants analysed (1 participant lost to follow‐up). Age, mean (SD): 3% HS group 10.6 (2.87) years; 7% HS group 10.87 (3.64) years. Sex: 3% HS group, 9 male (60%) and 6 female (40%); 7% HS group 13 male (86.67%) and 2 female (13.33%). CF mutation homozygous delta F508: 3% HS 3 (20%); 7% HS 1 (6.67%). CF mutation heterozygous mutation: 3% HS 0 (0%); 7% HS 2 (13.33%). Other CF mutations: 3% HS 9 (60.00%); 7% HS 7 (46.67%). Mutation analysis not done: 3% HS 3 (20%); 7% HS 5 (33.33%). FEV1 % predicted, mean (SD): 3% HS 50.2% (20.28); 7% HS 57.87% (26.93). FVC % predicted, mean (SD): 3% HS 55.07 (13.69); 7% HS 60.33 (22.84).
Interventions	Intervention: sterile 7% HS. Comparator: sterile 3% HS. Both solutions were made available in transparent collapsible bags of equal volume. Solutions were prepared by the Department of Pharmacology, All India Institute of Medical Sciences, with technical assistance from Baxter Pharmaceuticals Limited as per Good Manufacturing Practices guidelines, using Indian Pharmacopeia grade sodium chloride and sterile, double‐distilled water. Inhalation of 5 mL of the test drug, after taking a bronchodilator, twice a day by a jet nebuliser for 28 days.
Outcomes	Primary outcome Improvement in FEV1 (absolute % predicted) Secondary outcome Improvement in FVC Functional capacity (3‐min step test) PEFR Oxygen saturation Heart rate VAS 15‐count score, pre‐ and post‐ 3‐min step test
Notes	Sample size calculation was carried out and with less than 10% difference, 5% error and power of 80%, the required number of participants was 394. This sample size was not possible and so a pilot study with 15 participants in each group was carried out.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The random sequence was generated using a computer program by a person not involved in the trial.
Allocation concealment (selection bias)	Low risk	The intervention solutions were sequentially numbered as per the randomisation list by another person not involved in the trial.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Participants and investigators were blinded. The collapsible bags were similar in appearance and taste was not discernable.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Outcome assessors were blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	31 participants were randomised and there was only 1 lost to follow‐up in the 3% HS group. This meant that there were 15 in each arm and all of these participants were followed up.
Selective reporting (reporting bias)	Unclear risk	In the methods the authors say that they will be measuring adherence by the return of unused saline and keeping a diary but these are not mentioned in the results.
Other bias	Unclear risk	A sample size calculation was carried out with an alpha error of 5% and power of 80% which estimated that a sample size of 394 was needed. This was not possible and so the authors conducted a pilot study instead.

Laube 2009.

Study characteristics
Methods	RCT. Design: double‐blind, cross‐over. Duration: single dose of each treatment, washout period not stated. Location: USA (single centre).
Participants	Total participants: n = 12. Age, median (range): 10.5 (8.9 ‐ 12.4) years. Sex: 5 males. Normal pulmonary function (FEV1 and FVC > 90% of predicted values).
Interventions	Group 1: HS 7%. Group 2: 0.12% saline. Treatment followed by radio‐labelled isotopes.
Outcomes	Mucociliary clearance at 20, 60 and 90 minutes and 24 hours.
Notes	Mearsurements of MMC following inhalation of 0.12% saline were compared to 9 healthy adult controls. Abstract only. No sample size calculation stated.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The method of randomisation is not stated.
Allocation concealment (selection bias)	Unclear risk	Not stated in methods.
Blinding of participants and personnel (performance bias) All outcomes	High risk	No attempt was made to blind to taste of the solution.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not stated in methods.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Low risk as dropouts were replaced.
Selective reporting (reporting bias)	Low risk	All reported outcomes were given.
Other bias	Low risk	Sample size was estimated from data based on adolescents

Mainz 2015.

Study characteristics
Methods	RCT. Design: double‐blind, cross‐over design. Duration: 1st arm of 28 days followed by a 28‐day washout period and then alternative treatment for 28 days. Location: multicentre in Germany.
Participants	69 people with CF.
Interventions	Group 1: HS 6% via PariSinusTM. Group 2: IS via PariSinusTM.
Outcomes	Primary outcome Sinonasal outcome test (SNOT‐20) upper airway symptoms or disease‐specific QOL. Secondary outcomes Rhinoscopy, rhinomanometry, cytokines in nasal lavage.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Stated but not defined.
Allocation concealment (selection bias)	Unclear risk	This was stated but not described.
Blinding of participants and personnel (performance bias) All outcomes	High risk	No attempt was made to blind the taste of the HS solution.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Stated but not described in detail.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All drop outs clearly defined.
Selective reporting (reporting bias)	Low risk	All outcomes that were stated were described.
Other bias	Low risk	No other major source of bias detected.

NCT01355796.

Study characteristics
Methods	RCT. Design: open‐label cross‐over trial Duration: 2 weeks. Location: single‐centre trial carried out in the USA.
Participants	30 participants recruited and randomised. Inclusion criteria: documented diagnosis of CF (medical record evidence of 2 identified CFTR mutations or a positive sweat chloride test or nasal voltage difference, and 1 or more clinical findings of CF); 16 years or older; FEV1 over 30% predicted; oxygen saturation at least 90% on room air; clinically stable, without evidence of pulmonary exacerbation for at least 2 weeks prior to screening (defined as use of oral or intravenous antibiotics for CF exacerbation); use of effective contraception in women; ability to provide written informed consent and assent; successful completion of the trial doses of study drugs. Exclusion criteria: pregnancy; haemoptysis more than 100 mL within the last 30 days; change in chronic medication within the last 30 days; history of elevated serum creatinine (at least 2 mg/dL) within 30 days or at screening; history of lung and other solid organ transplantation; wait‐listed for lung or other solid organ transplant; known intolerance to inhaled HS. Baseline characteristics for whole study cohort Age, mean (SD): 31 (12.56) years (4 participants were aged 18 years or younger and 26 participants were aged between 18 and 65). Sex: 15  females (50%); 15 males (50%). Disease status, FEV1 % predicted, mean (interquartile range): 71% predicted (64 to 78).
Interventions	Intervention 1: aerosolised xylitol (5 mL) 2x daily for 14 days. Intervention 2: aerosolised 7% HS (4 mL) 2x daily for 14 days. Washout period of 7 days between treatments.
Outcomes	Primary outcomes FEV1 change from baseline Adverse events Respiratory symptom score Secondary outcomes Density of colonisation per gram of sputum Time to next exacerbation Sputum cytokines CFQ‐R (Quittner 2009) All outcomes measured at 98 days.
Notes	Sponsored by University of Iowa. Collaborators: Ann & Robert H Lurie Children's Hospital of Chicago Northwestern University. U01HL102288 (US NIH Grant/Contract).
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No information available on method of sequence generation.
Allocation concealment (selection bias)	Unclear risk	No information given describing allocation concealment.
Blinding of participants and personnel (performance bias) All outcomes	High risk	This was an open‐label trial and therefore no blinding. It is likely that some outcomes will be affected by this more than others. For example, knowledge of allocation may influence participant reported outcomes such as respiratory symptom score and adverse events. It is less likely to affect density of colonisation or time to next exacerbation.
Blinding of outcome assessment (detection bias) All outcomes	High risk	Open‐label trial and outcome assessors were not blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All participants completed the trial and were included in the analysis.
Selective reporting (reporting bias)	Low risk	We only have results that were posted to the trial registration document but all outcomes listed in the methods are reported in the results.
Other bias	Low risk	No other risk of bias identified. No details of sample size calculation.

PRESIS 2019.

Study characteristics
Methods	RCT. Design: parallel assignment. Duration: 52 weeks. Location: multicentre trial conducted at 5 CF centres in Germany within the clinical trial network of the German Centre for Lung Research.
Participants	42 infants with CF were enrolled. Inclusion criteria: confirmed diagnosis of CF established in neonatal period either via CF newborn screening or because of symptoms typical for CF (e.g. meconium ileus), positive family history or positive prenatal screening and fulfilling at least 1 of 3 criteria (sweat chloride ≥ 60mEq/L, 2 CF causing mutations of CFTR gene alterations of transepithelial potential difference of nasal or rectal epithelia typical for CF); age at enrolment is 0 to 4 months; participant's and parent's ability to comply with medication use, trial visits, and trial procedures as judged by the investigator (therefore parents have to understand the character of the study and individual consequences); participation is voluntary so only participants, whose parents or legal guardians gave written consent, are included. Exclusion criteria: born < 30 weeks gestation; prolonged mechanical ventilation in the first 3 months of life; a significant medical disease or condition other than CF likely to interfere with the child's ability to complete the entire protocol; previous major surgery except for meconium ileus; other major organ dysfunction, excluding pancreatic or hepatic dysfunction or another condition due to CF; physical findings that would compromise the safety of the participant or the quality of the trial data as determined by investigator; history of adverse reaction to sedation; known hypersensitivity to treatment; participation in other interventional trials at the same time. Criteria, which lead to a displacement of the procedures in sedation until the child has recovered: clinically significant upper airway obstruction as determined by investigator (e.g. severe laryngomalacia, markedly enlarged tonsils, significant snoring, diagnosed obstructive sleep apnoea); acute intercurrent respiratory infection, defined as an increase in cough, wheezing, or respiratory rate with onset in 2 weeks preceding visit; oxygen saturation < 95% before initial pulmonary function test or initial MRI; severe gastroesophageal reflux, defined as persistent frequent emesis despite anti‐reflux therapy. 42 participants enrolled (originally aimed for 40) aged up to 4 months. Baseline characteristics for all randomised participants (n = 42) Age, mean (SD) years: HS 0.26 (0.08); IS 0.26 (0.07). Sex: HS 10 male and 11 female; IS 10 male and 11 female. CFTR genotype F508del/F508del, n (%): HS 11 (52.4); IS 11 (52.4). CFTR genotype F508del/other, n (%): HS 6 (28.6); IS 8 (38.1). CFTR genotype other/other, n (%): HS 4 (19.0); IS 20 (95.2). Pancreatic insufficiency, n (%): HS 17 (81.0); IS 20 (95.2).
Interventions	Intervention group: 4 mL HS 6% (MucoClear® 6%) administered via inhalation 2x daily for 52 weeks. Control group: 4 mL IS 0.9% administered via inhalation 2x daily for 52 weeks. Both interventions delivered using the PARI LC SPRINT® Junior nebulizer with a baby bend, size‐adapted PARI® Baby face mask size 0 ‐ 3, connection tubing (2.2 m) and a PARI JuniorBOY® SX compressor.
Outcomes	Primary outcome Number of participants with adverse events and serious adverse events at end of trial. Secondary outcomes Rate of protocol‐defined pulmonary exacerbations requiring treatment with oral, inhaled or intravenous antibiotics Time to first pulmonary exacerbation Change from baseline in proportion of children with morphological or functional changes, or both, due to CF lung disease according to MRI chest score and CXR Chrispin‐Norman score (at end of trial) Change in extent and severity of bronchial dilatation after MRI and CXR scores at end of trial Proportion of children with impairments in lung function determined via multiple breath washout at baseline, after 3, 6, 9, and 12 months of inhalation Severity of impairment in lung function test at baseline, after 3, 6, 9, and 12 months Health‐related QOL as assessed by scores from the CFQ‐R (German version), administered quarterly Change in anthropometric and basic respiratory parameters (weight, height, BMI, weight‐for‐height, resting respiratory rate, and room air oxygen saturation) at end of trial Proportion of participants with new isolation of CF pathogen from clinically collected respiratory cultures among participants from whom Pseudomonas aeruginosa or other CF pathogens were not isolated from respiratory cultures prior to enrolment Time to first isolation of a CF pathogen
Notes	Principal investigator: Marcus A. Mall, MD, Heidelberg University, Germany. Collaborator: German Center for Lung Research. Start date: June 2012. Final data collection: November 2016. Supported by grants from the German Federal Ministry of Education and Research (82DZL10106, 82DZL10201, 82DZL10401, and 82DZL10501) and the Dietmar Hopp Foundation. Study solutions and inhalation devices were provided by PARI GmbH, Starnberg, Germany.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation process was based on a randomisation list prepared by a biostatistician of the Coordination Center for Clinical Studies (KKS) Heidelberg not involved in the study. The randomisation list contained subsequent randomisation numbers automatically generated by a dedicated randomisation web service (randomizer.at) for the multicentre setting.
Allocation concealment (selection bias)	Low risk	Randomisation and allocation all done by an independent biostatistician via a web generation service.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Blinding of participants, care providers, and investigators. HS and IS inhalation solutions were supplied by Pari GmbH, Starnberg, Germany, as identically packaged 4 mL plastic ampullae. Blinding was done by the hospital pharmacy at Heidelberg University Hospital.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Outcome assessors were blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Only 2 participants withdrew, 1 in each group with reasons given (the participant in the HS group was withdrawn at 30 weeks due to unblinding because of an exacerbation; the participant in the IS group was withdrawn by the parents as they declined further sedation for study procedures).
Selective reporting (reporting bias)	Unclear risk	QOL is listed in the trial registration document but is not reported in the results.
Other bias	Low risk	No other risk of bias identified. Sample size calculation carried out and met.

Riedler 1996.

Study characteristics
Methods	RCT. Design: cross‐over. Duration: first treatment on day 1, alternative treatment on second day. Location: single centre in Australia.
Participants	Total participants: n = 10 Age, mean (range): 16.5 (13 ‐ 20) years. Sex: 3 males and 7 females. FEV1 as % predicted, median (range): 53.5 (41 ‐ 73) FVC % predicted, median (range): 72 (15 ‐ 85). Participants were recruited as they were admitted with exacerbations of their lung disease with cough productive of tenacious sputum.
Interventions	Pre‐treated with 4 mL nebulised salbutamol 5 mg via jet nebuliser. Group 1: HS 6%. Group 2: IS. Single treatment via ultrasonic nebuliser (Timeter Compuneb MP500) for 10 min.
Outcomes	Sputum weight, VAS to assess feeling of cleared chest, spirometry (FEV1, FVC, FEF25-75, PEF).
Notes	7 participants were treated for a second block of treatment, but it was not defined who these were. Sputum expectoration and score changes not distributed normally. No sample size calculation stated.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Assigned by coin toss.
Allocation concealment (selection bias)	Low risk	Each participant assigned by coin toss to order of treatment ‐ no one could foresee allocation.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Single‐blinded, taste could be discerned by participants.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Single‐blinded, possibly assessors who were blinded due to difficulties in masking taste of intervention, but not clear.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Did not discuss whether an ITT approach had been used and there was no description of withdrawals.
Selective reporting (reporting bias)	Low risk	All outcomes stated in the methods were described in the results.
Other bias	Unclear risk	No washout period. No sample size calculation stated.

Robinson 1996.

Study characteristics
Methods	RCT. Design: cross‐over trial with the order of the 4 interventions being randomised. The last 9 participants underwent a cough study day, this was not included on the randomisation order and was always performed on the last day. Duration: single dose of each intervention, study days generally a week apart. Location: Australia.
Participants	Total participants: n = 12. Age, mean (SEM), range: 21.9 (3.0), 18 ‐ 28 years. Sex: 9 males and 3 females. Height, mean (SEM): 173.4 (11.6) cm. Weight, mean (SEM): 64.7 (11.3) kg. Stable disease.
Interventions	Pre‐treated with nebulised salbutamol 5 mg in 2.5 mL saline via ultrasonic nebuliser (Omron NE‐U06). Single inhalations of: Group 1: 7 mL HS 7%. Group 2: 7 mL amiloride 3 mg/mL. Group 3: 7 mL HS plus 7 mL amiloride 3 mg/mL. Group 4: 7 mL IS 0.9%. Voluntary cough, single episode. All done 1 week apart (control group 2).
Outcomes	Sputum isotope clearance 60 minutes, mucociliary clearance rate, change in FEV1.
Notes	Participants acted as own controls. Spirometry measures were taken immediately after inhalation and are not a long‐term outcome measure. Paper reported lung function (mean values for all trial days): FEV1 % predicted, mean (SEM), range: 60.8 (29.7),27 ‐ 112; FVC % predicted, mean (SEM), range: 77.4 (22.4), 44 ‐ 118; FEF25-75 % predicted, mean (SEM), range: 41.1 (37.5), 9 ‐ 104. No sample size calculation stated.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	No details provided.
Allocation concealment (selection bias)	Unclear risk	No details provided.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Participants and all but 1 of investigators blinded to solutions, but participants may have been able to discern taste of intervention and work out the longer nebulization times for the combination of amiloride and HS (although no reference made to this).
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	All but 1 of investigators blinded to solutions, but not clear which of investigators knew and whether they were assessing outcomes.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No withdrawals or dropouts.
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.

Robinson 1997.

Study characteristics
Methods	RCT. Design: cross‐over with 4 arms (no information on any washout period). Duration: each intervention given on 1 study day. Location: Australia.
Participants	Total participants: n = 10 Age, mean (SD), range: 22.1 (3.8), 19 ‐ 28 years. Sex: 7 males and 3 females. FEV1 % predicted, mean (SD), range: 52.0% (6.7), 31 ‐ 84% predicted. All participants were chronically colonised Pseudomonas aeruginosa. In addition, 5 of the participants had Staphylococcus aureus. Participants in a stable clinical condition and baseline medications were not altered throughout the trial period.
Interventions	Pre‐treated with nebulised salbutamol 5 mg in 2.5 mL saline via an ultrasonic nebuliser. Group 1: HS 3% single dose. Group 2: HS 7% single dose. Group 3: HS 12% single dose. Group 4: Voluntary cough and IS combined as the control. Each participant took part in each arm.
Outcomes	Sputum isotope % clearance at 30 minutes, sputum isotope clearance at 90 minutes, MCC.
Notes	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)	High risk	There was no random allocation to treatment.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Participants could discern taste for intervention.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not discussed.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	No discussion of whether an intention‐to‐treat 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.

Robinson 1999.

Study characteristics
Methods	RCT. 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 Age, mean (SD), range: 29.9 (9.4), 16 ‐ 46 years. Sex: 5 males and 7 females. 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. Treatments were administered via an Omron‐NE‐U06 domiciliary ultrasonic nebuliser.
Outcomes	Sputum isotope % clearance at 30 minutes, sputum isotope clearance at 90 minutes, MCC.
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.

Rosenfeld 2012.

Study characteristics
Methods	RCT. Design: double‐blind, parallel. Duration: 48 weeks. Location: 30 centres in USA and Canada.
Participants	Total participants: n = 321 (176 males, 145 females) aged 4 ‐ 60 months with an established diagnosis of CF (details of diagnosis given in supplementary paper). Age, mean (SD): HS group 2.2 (1.4) years; control group 2.3 (1.5) years. Sex: HS group 84 males (53%); control group 92 males (56%). Weight, mean (SD); HS group 12.2 (4.1) kg; control group 12.5 (4.1) kg. Weight percentile, mean (SD): HS group 39.7 (28.1); control group 43.0 (29.1). Height, mean (SD): HS group 84.8 (14.8) cm; control group 85.7 (15.0) cm. Height percentile, mean (SD): HS group 36.9 (27.0); control group 39.9 (28.1). Positive respiratory culture (Pseudomonas aeruginosa isolated from respiratory culture at or at any time prior to randomisation. For other organisms, positive culture at or within 24 months prior to randomisation): Pseudomonas aeruginosa: HS group 60 (38.0%); control group 69 (42.3%); Staphylococcus aureus: HS group 98 (62.0%); control group 124 (76.1%); MRSA: HS group 5 (3.2%); control group 11 (6.8%); Stenotrophomonas maltophilia: 25 (15.8%); 35 (21.5%); Achromobacter xylosoxidans: HS group 4 (2.5%); control group 3 (1.8%).
Interventions	Pretreatment: all participants received albuterol or levalbuterol prior to each trial drug dose ‐ 2 puffs via MDI via a valved holding chamber with face mask or by nebulizer (distinct from the nebulizer used to administer the trial drug) and PARI Proneb® Ultra compressor. Group 1 (n = 158): HS 7% 2x daily. Group 2 (n = 163): IS 0.9% 2x daily. Both treatments administered via Proneb Ultra compressor with a Sprint Jr nebulizer equipped with a Baby facemask or mouthpiece ‐ participants under 36 months used a facemask and those over 36 months used a mouthpiece, but this was individualized as developmentally appropriate.
Outcomes	Pulmonary exacerbation rate (events per person‐year; defined as treatment with oral, inhaled, or intravenous antibiotics for 1 or more prespecified signs and symptoms) Number of treatment days and number of courses of antibiotics Time to first exacerbation Lung function (FEV 0.5, FEF 75, FEF25%-75%, FRC, RV/TLC) CFQ‐R (Quittner 2009) ‐ Parent and reported Change in height and weight Change in resting respiratory rate, pulse oximetry and parent‐reported cough Rate of intolerance to the test dose of HS at enrolment Adverse events and withdrawal rates Treatment‐emergent respiratory cultures positive for CF pathogens detected through clinical cultures performed at each site’s microbiology laboratory Adherence to treatment was assessed by  the number of used drug vials returned the Treatment Adherence Questionnaire completed quarterly the weekly parent questionnaire.
Notes	Clinicaltrials.gov Identifier: NCT00709280. Trial visits occurred at enrolment or randomization and 4, 12, 24, 36, and 48 weeks after randomization. At the enrolment visit, after pretreatment with albuterol or levalbuterol, all participants were evaluated for intolerance to a test dose of 7% HS according to predefined criteria. Participants who tolerated the test dose were randomized. Sample size calculation was undertaken.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Participants were randomised 1:1 based on random permuted blocks, stratified by age and site (4 to 29 months, 30 to 60 months) via a secure website.
Allocation concealment (selection bias)	Low risk	Randomisation was done via a secure website.
Blinding of participants and personnel (performance bias) All outcomes	High risk	HS (Hyper‐Sal; PARI Respiratory Equipment) and 0.9% IS supplied by Catalent Pharma Solutions as identically packaged 4 mL blow‐fill‐seal plastic ampoules, but taste could be discerned as different.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Researchers blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	ITT analysis of 158 in HS group and 163 in control group. Details of withdrawals given as follows: HS group: 15 withdrew in total (follow‐up range, 2 to 44 weeks): 5 lost to follow‐up; 4 treatment burden; 2 intolerant to trial drug; 1 time constraints; 3 other. Control group: 14 withdrew in total (follow‐up range, 0 ‐ 42 weeks): 2 lost to follow‐up; 3 treatment burden; 2 insufficient perceived benefit from trial drug; 1 intolerant to trial drug; 1 time constraints; 1 other adverse event; 4 other.
Selective reporting (reporting bias)	Low risk	All outcomes stated in the methods were described in the results.
Other bias	Low risk	Sample size calculation was done. No other bias identified

SHIP 2019.

Study characteristics
Methods	RCT. Design: parallel design. Duration: 48 weeks. Location: multicentre in USA and Canada.
Participants	150 children were enrolled Inclusion criteria: confirmed diagnosis of CF; written informed consent by parent or legal guardian; age ≥ 36 months and ≤72 months at screening visit; able to comply with medication use, trial visits and trial procedures as judged by the site investigator; able to perform technically acceptable multiple breath washout measurements at the screening and enrolment visits. Exclusion criteria: acute intercurrent respiratory infection, defined as an increase in cough, wheezing, or respiratory rate with onset within 3 weeks preceding screening or enrolment visit; acute wheezing at screening or enrolment visit; oxygen saturation < 95% (< 90% in centres located above 4000 feet elevation) at screening or enrolment visit; physical findings that would compromise the safety of the participant or the quality of the trial data as determined by site investigator; investigational drug use within 30 days prior to at screening or enrolment visit; treatment with inhaled HS at any concentration within 30 days prior to screening or enrolment visit; chronic lung disease not related to CF; inability to tolerate first treatment dose at the enrolment visit. Baseline characteristics Age, mean (SD): HS 4.5 (1.0) years; IS 4.4 (0.9) years. Age group 36 to 54 months, n (%): HS 44 (58%); IS 43 (58%). Age group 55 to 72 months, n (%): HS 32 (42%); IS 31 (42%). Sex, n (%): HS 36 male (47%) and 40 female (53%); IS: 33 male (45%) and 41 female (55%). CFTR genotype homozygous delta F508, n (%): HS 38 (50%); IS 40 (54%). CFTR genotype compound heterozygote delta F508, n (%): HS 33 (43%); IS 28 (38%). CFTR genotype other, n (%): HS 5 (7%); IS 6 (8%). LCI2-5 mean (SD): HS 9.3 (2.0); IS 9.4 (2.0). FEV0.75 % predicted mean (SD): HS 90.6 (12.4); IS 95.0 (16.2).
Interventions	Intervention: 4 mL 7% HS via inhalation 2x daily. Control: 4 mL 0.9% HS via inhalation 2x daily. The delivery system for both groups is a PARI Sprint Junior nebulizer with a PARI Baby facemask or mouthpiece driven by a PARI Vios® compressor. (PARI Respiratory Equipment, Midlothian, VA, USA)
Outcomes	Primary outcomes Change from baseline in LCI measured by N2 multiple breath washout Secondary outcomes Change from baseline in FEV0.75 measured by preschool spirometry Protocol‐defined pulmonary exacerbation rate Health‐related QOL measured by the modified parent‐reported CFQ‐R for preschoolers Respiratory signs as measured by the Cystic Fibrosis Respiratory Sign Diary for ages 0 ‐ 6 (CFRSD0‐6) Treatment‐emergent CF respiratory pathogens from clinical respiratory cultures
Notes	Principal investigators: Stephanie Davis, MD Indiana University; Richard A Kronmal, PhD University of Washington; Felix Ratjen, MD, PhD, FRCPC Hospital for Sick Kids, Toronto; Margaret Rosenfeld, MD, MPH Seattle Children's Hospital. Study start date: March 2015. Estimated primary completion date: August 2018. Estimated Study completion date: February 2019.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Random permuted block randomisation was done separately for participants aged 36 to 54 months and 55 to 72 months. The code was generated via a web‐based data entry system.
Allocation concealment (selection bias)	Low risk	A web‐based data entry system was used to allocate the random code. Site pharmacists were provided with unique kit numbers for dispensation.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Participants, care providers, and investigator were blinded to allocation.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Outcome assessors were blinded to allocation.
Incomplete outcome data (attrition bias) All outcomes	Low risk	10 participants did not complete the study. HS group: 3 withdrew (1 intolerant of study drug, 1 perceived treatment burden to be too great, 1 lost to follow‐up). IS group: 7 withdrew (3 lost to follow‐up, 2 time constraints, 1 perceived treatment burden to be too great, 1 perceived insufficient benefit from study drug). Sample size was calculated assuming ⍺ of 0.05 with 80% power. They estimated that they would need a sample size of 128 but due to the uncertainty of the treatment effect size they increased the sample size to 150.
Selective reporting (reporting bias)	Low risk	All outcomes listed in the trial registration document and in the methods were reported in the results.
Other bias	Unclear risk	Those performing the statistical analysis were not blinded.

Singh 2020.

Study characteristics
Methods	RCT. Design: parallel design. Duration; 14 days. Location: USA. Setting: hospital admission.
Participants	63 adults enrolled and 60 were randomised to treatment. Inclusion criteria: diagnosed with CF (medical record evidence of CFTR mutation or sweat chloride test or nasal voltage difference, and 1 or more clinical findings of CF); 12 years or older; FEV1 > 30% predicted (within the last 14 days and oxygen saturation > 90% on FiO2 ≤ 50%); admitted for an exacerbation; use of effective contraception in women; written informed consent. Exclusion criteria: pregnancy; history of asthma based on methacholine challenge or bronchial hyper‐responsiveness on PFT; haemoptysis more than 60 mL within the last 30 days; use of any investigational study drug within the last 30 days; initiation of hypertonic saline within the last 30 days; a serum creatinine 2 mg/dL or more; active malignancy in the last year; antibiotics for CF exacerbation as an outpatient in the last 2 weeks; Burkholderia cepacia colonisation; waiting list for lung transplant; lack of FEV1 data from the last 14 days; previous participation in this study. Baseline characteristics Age, mean (SD): HS 30.4 (12.8) years; Xylitol 29.7 (12.3) years; overall 30.1 (12.4) years. Sex, female/male (n); HS 12/17; Xylitol 15/15; overall 27/32. FEV1 % predicted, mean (SD): HS 53.3 (18.8); Xylitol 58.8 (20.1); overall 56.1 (19.5). CFTR genotype homozygous F508del, n (%): HS 17 (59); Xylitol 18 (60);  Azithromycin use,  n (%): HS 28 (97); Xylitol 28 (93). rhDNase use, n (%): HS 14 (48); Xylitol 13 (43). Baseline Pseudomonas aeruginosa, median log CFU (min to max): HS 6.2 (0 to 7.3); Xylitol 2.0 (0 to 7.6). Baseline Staphyloccocus aureus, median log CFU (min to max): HS 4.7 (0 to 7.1); Xylitol 3.5 (0 to 7.0).
Interventions	Intervention: aerosolized 7% HS (5 mL) 2x daily. Control: 15% xylitol (5 mL) 2x daily. Duration of treatment was 2 weeks and was administered via a Pari LC nebulizer with Vios compressor.
Outcomes	Primary outcomes FEV1 (change from baseline) Adverse events Respiratory symptom score Secondary outcomes Density of colonisation per g of sputum Time to next exacerbation Sputum cytokines CFQ‐R (Quittner 2009) Outcomes were reported at baseline and day 14.
Notes	Principal Investigator: Joseph Zabner, M.D. Study Director: Lakshmi Durairaj, M.D. Study Chair: Jan L Launspach, R.N., CCRC. Study start date: June 2009. Estimated primary completion date: January 2018. Estimated study completion date: January 2018.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Participants were randomised in a 1:1 ratio in 5 blocks of 12 each using a random number generator.
Allocation concealment (selection bias)	Low risk	Participants were assigned to either xylitol or HS by a research pharmacist from the investigational drug services at the Iowa Hospitals and Clinics based on a sequence generated by a statistician.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Care provider and investigator were blinded to allocation. Blinding of participants could not be assured because of the sweet and salty tastes of the medications.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	It is unclear whether the outcome assessors were blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	60 participants were randomised and only 1 participant withdrew themselves. 59 participants completed the study and were included in the analysis. Sample size calculation showed that 30 participants per group would give 80% power.
Selective reporting (reporting bias)	Low risk	All outcomes listed in the trial registration document and the methods were reported in the results.
Other bias	Low risk	No other risk of bias identified.

Suri 2001.

Study characteristics
Methods	RCT Design: open‐label, cross‐over with 3 treatment arms; 48 children were randomised, 8 to each of the 6 possible treatment orders. Duration: each arm lasted 12 weeks with a 2‐week washout period between treatments. Location: multicentre in UK (2 institutions).
Participants	Total participants recruited: n = 48, but 1 withdrew without commencing treatment, so trial population is 47. Age, mean (SD): 12.6 (2.8) years, range (7.3 ‐ 17 years). Sex: 19 (40%) males; 28 (60%) females. Weight, mean (SD): 40·0 (12·6) kg, range (18·8 ‐ 77 kg). FEV1 (L), mean (SD): 1.18 (0.47), range (0.44 ‐ 2.34 L). FEV1 % predicted, mean (SD): 48% (15), range (14% ‐ 77%). FVC % predicted, mean (SD): 68% (22), range (20% ‐ 112%). Lung microbiology (number of children with 3 positive cultures of the organism in the previous year): Pseudomonas aeruginosa: 22 (48%); Staphylococcus aureus: 18 (39%); Stenotrophomonas maltophilia: 1 (2%).
Interventions	Pre‐treated with BDs. All treatments were administered with a Sidestream nebuliser and Porta‐Neb compressor (Medic‐Aid, Bognor Regis, UK). Group 1: 5 mL HS 7% 2x daily immediately before the participant's regular physiotherapy session. Group 2: rhDNase 2.5 mg 1x daily at least 1 hour before physiotherapy. Group 3: 2.5 mg alternate daily at least 1 hour before physiotherapy. Routine medication and physiotherapy were continued throughout the trial.
Outcomes	Primary outcome % change in FEV1 from baseline Secondary outcomes FVC Number of pulmonary exacerbations (defined as a previously outlined protocol for respiratory‐tract infections) Weight gain Exercise tolerance (3‐min step test and oxygen saturation monitored) QOL (quality of well‐being scale self‐administered form 1.04, filled out by the participant and guardian together) Total healthcare cost (hospital and community‐health‐service perspective, so participants' costs excluded from analysis; resources included covered hospital admissions (inpatient, outpatient, and day case), radiological investigations, blood tests, drug use, and the use of community services (visits to general practitioners, district nurses, and physiotherapists)) Adherence (count of unused bottles of HS and empty vials of rhDNase; each participant was also given a diary to record the treatment doses taken)
Notes	Sample size calculation undertaken. Before starting the HS treatment period, each participant received a test dose of HS in hospital so that he or she could be monitored for bronchoconstriction.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Telephone randomisation to an independent trials coordinating unit, stratified by hospital and balanced after each group of 12 children.
Allocation concealment (selection bias)	Low risk	Telephone randomisation to an independent trials coordinating unit,
Blinding of participants and personnel (performance bias) All outcomes	High risk	No attempt to conceal taste, paper states masking impossible because HS can easily be distinguished from rhDNase by its salty taste and timing in relation to physiotherapy.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No statement on blinding of outcome assessors.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	An ITT approach was used within this trial, 43 children are included in the comparison of daily and alternate‐day rhDNase, and 40 in the comparison of daily rhDNase and hypertonic saline. Details of withdrawals as follows: 1 girl withdrew almost immediately after randomisation without starting the first treatment due to illness; 8 children were unable to complete all 3 treatment periods; 4 participants had severe declines in their pulmonary status and required long courses of intravenous antibiotics (2 of them were taking alternate‐day rhDNase, 1 HS, and the other daily rhDNase). In the additional report of airway inflammatory changes following treatment, only 28 of the 48 participants were able to perform induced sputum and be included. It is not evident that any attempts were made to adjust for missing data from those participants unable to do induced sputum.
Selective reporting (reporting bias)	Low risk	All outcomes stated in the methods were described in the results.
Other bias	Low risk	We judged a 2‐week washout period to be low risk. Sample size calculation undertaken.

Weller 1980.

Study characteristics
Methods	RCT. Design: double‐blind, cross‐over with 2 arms. Duration: 2‐month baseline periods preceding and following 2x 8‐week treatment periods. Location: UK.
Participants	Total participants recruited: n = 29. Diagnosis was confirmed in all by history, examination, and a sweat test. 27 (13 males, 14 females) completed the trial aged 6 to 15 years (mean 10.7 years); of these 27 participants, 22 were chronic sputum producers. Baseline characteristics Age, mean (SD); sputum producers 10.9 (2.1) years, range (6.5 ‐ 15); sputum non‐producers 9.8 (2.8) years range (6.1 ‐ 12.25). FVC % predicted, mean (SD): sputum producers 76% (15.9), range (38 ‐ 101); sputum non‐producers 88% (10.1), range (76 ‐ 104). PEFR % predicted, mean (SD): sputum producers 80% (21.5), range (27 ‐ 113); sputum non‐producers 96% (11.6), range (78 ‐ 107).
Interventions	No reported pretreatment. Group 1: 3 mL sodium‐2‐mercaptoethane sulphonate (Mistabron) 20% 2x daily after physiotherapy. Group 2: 3 mL HS 7% 2x daily after physiotherapy. Inhalations via a Wright nebuliser operated by an air compressor (Aerolyser Electric Inhaler, Aerosol Products (Colchester) Ltd) producing a flow of 8 L/min.
Outcomes	PEFR, FVC, V max 50% VC, RV/TLC. Diary card to record sputum volume, sputum colour, and cough frequency. At monthly clinic visits: sputum cultures, pulmonary function tests (PEFR, FVC, Vmax50%VC, RV/TLC). At beginning and end of trial: CXRs(Chrispin Norman score), full blood count, liver function tests and plasma electrolytes including creatinine.
Notes	Participants divided into sputum producers and non‐sputum producers. Sample size calculation not undertaken.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Stated that order of treatment was randomised, but no details given as to randomisation process.
Allocation concealment (selection bias)	Unclear risk	No detail provided.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Described as double‐blind, but participants could discern difference in taste of interventions.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Described as double‐blind, but not clear whether outcome assessors were one of the parties blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Clear description of dropouts and withdrawals (see below) but it was not stated if an ITT analysis had been performed. 2 participants withdrawn from trial: 1 due to an acute exacerbation during the second treatment period (saline inhalations) for which she received other inhalational treatment (antibiotics and bronchodilators); and 1 did not take her inhalations correctly and was admitted to hospital early in the trial, during saline therapy, both for more intensive medical treatment and for psychiatric reasons (also received other inhalational treatment).
Selective reporting (reporting bias)	Low risk	All outcomes stated in the methods were described in the results.
Other bias	Low risk	Washout period: 2‐month baseline periods preceding and following 2x 8‐week treatment periods. Sample size calculation not undertaken.

BD: bronchodilator
BMI: body mass index
CF: cystic fibrosis
CFQ‐R: Cystic Fibrosis Questionnaire ‐ Revised
CFTR: cystic fibrosis transmembrane conductance regulator
CFU: colony forming units
CXR: chest x‐ray
FEF_25%-75%: mid‐peak expiratory flow
FEV₁: forced expiratory volume at 1 second
FEV_0.75: forced expiratory volume at 0.75 seconds
FiO2: fraction of inspired oxygen
FVC: forced vital capacity
HA: hyaluronic acid
HS: hypertonic saline
IS: isotonic saline
ITT: intention‐to‐treat
LCI: lung clearance index
LCI_2-5: number of lung volume (functional residual capacity) turnovers needed to clear the tracer gas to one‐fortieth of the starting concentration
MCC: mucociliary clearance
MDI: metered dose inhaler
MRI: magnetic resonance imaging
N2: nitrogen
NaCl: sodium chloride
PEFR: peak expiratory flow rate
PFT: pulmonary function test
QOL: quality of life
rhDNase: deoxyribonuclease
RV: residual volume
SEM: standard error of the mean
SF6: sulfur hexafluoride
TLC: total lung capacity
VAS: visual analogue scale
VC: vital capacity

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
ACTRN12619001681145	The intervention is the MetaNeb device and not HS.
ACTRN12621000855820	The intervention is the MetaNeb device and not HS; non‐randomised.
Aquino 2012	Cross‐over trial which had an inadequate washout period between treatments (48 hours).
Brivio 2016	Comparison of HS to HS plus HA (no control without HS).
Brown 2010	The intervention and comparator are both 7% HS, but measured at different time points.
Buonpensiero 2010	Comparison of HS to HS plus HA (no control without HS).
Corcoran 2017	HS was given via nasal cannulae overnight rather than by facemask and nebuliser.
DeCono 2008	Not randomised.
Dentice 2012	Did not compare HS to a control group. Instead the trial sought out the optimum timing of physiotherapy to HS use, before, during or after.
Donaldson 2006	Did not compare to a control group that did not include HS, comparison of HS with or without pretreatment with amiloride.
Elkins 2006b	Comparison of dose frequency of HS; all participants received HS, but randomised to 2 or 4 times daily.
EUCTR2007‐002707‐40‐BE	Not randomised.
Genkova 1998	Did not compare to a control treatment; did not report any results.
Grasemann 2013	The trial did not include an arm with HS. It compared inhaled L‐arginine with isotonic saline as a control.
IRCT20180307038994N1	Not randomised.
IRCT20191112045413N2	Single‐arm trial with no comparator
IRCT20201017049055N1	Not randomised.
King 1997	In vitro trial only.
Kobylyansky 2000	Trial was not performed in a CF population.
NCT01094704	Not randomised; reporting results at different time points after a single dose of HS.
Nenna 2017	Cross‐over trial with no washout period.
O'Neill 2017	Comparison of timing of HS in reference to airway clearance. No comparator group that did not include HS.
Ros 2012	Comparison of 2 formulations of 7% HS, with or without HA.
Ruiz de Valbuena Maiz 2012	The study is comparing different volumes of HS rather than different strengths.
San Miguel‐Pagola 2016	Comparison of 2 arms both using HS with and without physiotherapy interventions.
Tiddens 2022	This study looks at the effect of HS on structural lung disease via MRI.
Van Ginderdeuren 2008	Did not compare HS to a control group. Instead the trial sought to evaluate the efficacy of an autogenic drainage treatment combined with HS.
Van Ginderdeuren 2011	Did not compare HS to a control group. Instead the trial sought to evaluate the efficacy of an autogenic drainage treatment combined with HS.
Vanlaethem 2008	Comparison of HS in conjunction with airway clearance techniques.

CF: cystic fibrosis
HA: hyaluronic acid
HS: hypertonic saline
MRI: magnetic resonance imaging

Characteristics of studies awaiting classification [ordered by study ID]

Balinotti 2015.

Methods	Prospective, double‐blind RCT. Duration: 1 year.
Participants	Children over 6 years of age with a diagnosis of CF made by 6 months of age. 19 participants randomised: 3 excluded, 9 participants in HS group, 7 participants in IS group.
Interventions	Intervention: HS 3% plus 0.25 mg/kg salbutamol 2x daily. Control: IS 0.9% plus 0.25 mg/kg salbutamol 2x daily.
Outcomes	Primary outcome: maximal flow at functional residual capacity (change from baseline). Secondary outcomes: pulmonary exacerbations, respiratory rate, nutritional status, AEs.
Notes

Dwyer 2013.

Methods	RCT. Parallel design. Duration: 16 weeks. Location: Australia. Sequence generation: random number generation with block allocation. Treatment allocation is stratified for: age, sex, FEV1. Allocation concealment: random allocation to 1 of 3 groups by the Trial Pharmacist at the trial co‐ordinating centre. Treatment allocation recorded at the Trial Pharmacy and the random allocation lists, randomisation procedure and the unblinded treatment allocation is to be concealed from all other trial staff and the participant. Blinded.
Participants	Inclusion criteria: informed consent, diagnosis of CF (positive sweat test or genotyping), best FEV1 in the previous six months > 20% of predicted normal value FEV1 > 85% of best in the previous 6 months, no non‐routine antibiotics in the last 14 days, minimum age 6 years, both sexes eligible. Exclusion criteria: colonisation with Burkholderia cepacia, major haemoptysis within the last 12 months, pregnant or lactating females, investigational drugs within the last 30 days, previous lung transplant, hypertonic saline within the last 14 days, inhaled mannitol within the last 14 days.
Interventions	Intervention 1: 2 x daily inhalation of 4 mL of nebulised 6% HS + 0.25 mg/mL quinine sulphate. Intervention 2: 2x daily inhalation of 4 mL of nebulised 3% HS + 0.25 mg/mL quinine sulphate. Control: 2x daily inhalation of 4 mL of nebulised 0.9% IS + 0.25mg/mL quinine sulphate.
Outcomes	Primary outcome Lung function as measured by the change in FEV1 % predicted (measured at baseline, Week 1, Week 4, Week 8 and Week 16 (end of trial)). Secondary outcomes Lung function as measured by the change in FVC % predicted (measured at baseline, Week 1, Week 4, Week 8 and Week 16 (end of trial)). Lung function as measured by the change in FEF25%-75% predicted (measured at baseline, Week 1, Week 4, Week 8 and Week 16 (end of trial)). QoL as measured by the CFQ‐R (Quittner 2009) (measured at baseline, Week 1, Week 4, Week 8 and Week 16 (end of trial)). QoL as measured by the Medical Outcomes Survey SF‐36 (measured at baseline, Week 1, Week 4, Week 8 and Week 16 (end of trial)). Exercise capacity as measured by the total distance covered in the MST‐25 (measured at baseline, Week 4 and Week 16 (end of trial)). Exercise capacity as measured by the total exercise time in the EST‐25 (measured at baseline, Week 4 and Week 16 (end of trial)). Sputum bacterial diversity as measured by the acquisition or loss of bacterial organisms in expectorated sputum as measured by routine microscopy culture and sensitivity (measured at baseline and Week 16 (end of trial)). Tolerability of nebulised trial solution as measured by participant on a 10‐point visual analogue scale (measured at baseline, Week 1, Week 4, Week 8 and Week 16 (end of trial)). Medication use as measured by number of doses of each prescribed medication (measured at baseline and weekly during trial (Week 1 to Week 16)). Pulmonary exacerbations as measured by the Fuchs exacerbation criteria (Fuchs 1994) (measured weekly during trial (Week 1 to Week 16)). AEs (such as intolerable cough, sore throat, bronchospasm, haemoptysis, nausea, pulmonary exacerbation) as measured by presence of new symptoms, likelihood of being related to trial solution, severity of symptoms and time to resolution of symptoms (measured weekly during trial (Week 1 to Week 16)). Adherence to nebulisation of trial solution as measured weekly by self‐report in diary and count of returned unused ampoules of trial solution at the end of the trial (week 16) Administration time of nebulised trial solution as measured by stop watch from start to completion of 1 dose of trial inhalation solution (measured at baseline, Week 1, Week 4, Week 8 and Week 16 (end of trial)).
Notes	Saline at lower tonicity in cystic fibrosis (SALTI‐CF) trial. Funding source: Australian Cystic Fibrosis Research Trust.

Hofmann 1997.

Methods	Open‐label RCT.  Single dose.
Participants	20 adults with CF.
Interventions	Intervention: amiloride in HS 5.58%. Control: amiloride in IS.
Outcomes	Change in nasal potential difference.
Notes	Trial to determine the effects of HS and amiloride on change in ion flow by nasal potential difference. Abstract only.

Lennox 2016.

Methods	RCT (phase I). Cross‐over design. Location: USA.
Participants	Inclusion criteria: aged 12 years and over; FEV1 greater than 50% predicted; able to spontaneously expectorate sputum (with or without chest physiotherapy); stable disease as defined by clinician assessment and no use of IV antibiotics in the past 4 weeks, no changes in CF‐related medications in the 4 weeks prior to study screening and SpO2 < 94% on room air or use of supplemental oxygen. Exclusion criteria: reactive airway disease; use of inhaled hypertonic saline in the past 28 days; use of IV antibiotics in the past 4 weeks; changes in CF‐related medications in the 4 weeks prior to screening; SpO2 < 94% on room air or use of supplemental oxygen; presence of untreated gastroesophageal reflux disease or residual acid reflux symptoms more than 3 times per week; pregnant or nursing females. 12 participants enrolled.
Interventions	Hypertonic bicarbonate group: inhaled solution of 8.4% hypertonic bicarbonate by nebuliser. HS group: inhaled solution of 7% HS by nebuliser.
Outcomes	Primary outcomes: change in exhaled breath condensate pH at 4 hours; change in pH after inhalation of 2 doses on 1 day. Secondary outcomes: change in expectorated sputum at 4 hours; change in sputum wet‐to‐dry ratio after inhalation of 2 doses on 1 day; change in spirometry at 4 hours; FEV1 before and after inhalation of 2 doses on 1 day.
Notes	Principal Investigator: Joseph M PIlewski, MD, University of Pittsburgh. Collaborators: Cystic Fibrosis Foundation Therapeutics. Start date: August 2014. Completion date: July 2016. Listed retrospectively on clinicaltrials.gov in 2018. No publication to date.

Palacio 2014.

Methods	Open‐label RCT. Parallel assignment. Location: 2 centres in Argentina. Duration: 24 weeks.
Participants	27 children aged 3 to 6 years with CF. 21 completed trial.
Interventions	Intervention 1: HS 7% nebulised 2x daily. Control: IS 0.9%, nebulised 2x daily. Salbutamol given prior to each dose.
Outcomes	Primary outcome Lung function Secondary outcomes Respiratory symptoms Anthropometric measures New isolation of Pseudomonas aeruginosa Rate of exacerbations AEs Adherence
Notes	We are waiting for further data to judge whether this should be included.

SIMPLIFY 2022.

Methods	Parallel RCT; stopping trial. Multicentre trial (80 centres in the USA). Duration of trial: 6 weeks.
Participants	870 adults and children aged 12 or over Diagnosed with CF FEV1 ≥ 70 % predicted at the Screening Visit if under 18 years old, and ≥ 60 % predicted at Screening Visit if 18 years or older. Clinically stable. Current treatment with elexacaftor/tezacaftor/ivacaftor for at least 90 days prior to and including the Screening Visit and willing to continue daily use for the duration of the trial. Currently taking HS (at least 3%) or DNase (or both) for at least 90 days prior to and including the Screening Visit and willing to continue daily use for the 2‐week screening period.
Interventions	Arm 1: discontinue HS therapy. Arm 2: continue HS therapy The concentration of HS is according to clinical prescription (e.g. 7% sodium chloride or 3.5% sodium chloride) and at least 3%. 2 further arms look at continuing or discontinuing DNase. We will only be looking at HS.
Outcomes	Absolute change in FEV1 % predicted at week 6 Adverse events Absolute change in respiratory symptoms
Notes	This is a stopping trial.

AEs: adverse events
CF: cystic fibrosis
CFQ‐R: cystic fibrosis questionnaire ‐ revised
EST‐25: Endurance Shuttle Test‐25
FEF_25%-75%: mid peak expiratory flow
FEV₁: forced expiratory volume at one second
FVC: forced vital capacity
HS: hypertonic saline
IS: isotonic saline
IV: intravenous.
MST‐25: Modified Shuttle Test‐25
QoL: quality of life
RCT: randomised controlled trial
SF‐36: Short Form‐36
SpO2: peripheral capillary oxygen saturation

Characteristics of ongoing studies [ordered by study ID]

ISRCTN14081521.

Study name	A randomised open‐label trial to assess change in respiratory function for people with cystic fibrosis (pwCF) established on triple combination therapy (Kaftrio™) after rationalisation of nebulised mucoactive therapies (the CF STORM trial)
Methods	Open‐label parallel RCT. Participants established on triple therapy (Kaftrio™) for more than 3 months will be randomly allocated to either stopping or continuing their nebulised muco‐active treatment (DNase or HS, or both). Randomisation and allocation is via a web‐based system. Duration of trial: 12 months.
Participants	Adults and children aged 12 years or over. Inclusion criteria Clear diagnosis and clinical features of CF. One or two Phe508del variants. Established on daily muco‐active nebulised therapy for at least 3 months. No need for extra antibiotics in the preceding 6 weeks. Able to perform spirometry.
Interventions	Arm 1: stopping nebulised muco‐active therapy (DNase or HS, or both) Arm 2: continuing nebulised muco‐active therapy
Outcomes	Change in FEV1 % predicted at 52 weeks Change in FEF25-75 % predicted at 52 weeks Need for extra antibiotic treatment (number of courses and number of days of extra oral, IV, or nebulised antibiotics) at 52 weeks Need for extra chronic medications Number and proportion of positive respiratory cultures for significant pathogens Hospital admission Change in nutritional status Number of pulmonary exacerbations  Change in QoL using CFQ‐R (Quittner 2009) Adverse events Costs to the UK National Health Service
Starting date	Date of first enrolment: 1 April 2021
Contact information	Professor Kevin Southern (kwsouth@liverpool.ac.uk) Institute in the Park, University of Liverpool, Alder Hey Children’s Hospital, Eaton Road, L12 2AP, Liverpool, UK
Notes	This is a stopping trial and includes DNase as well as HS. When we are able to include the trial we will need IPD for those participants that stopped or continued HS.

NCT02343445.

Study name	Clearing lungs with ENAC inhibition in cystic fibrosis (CLEAN‐CF)
Methods	Blinded (participant, care provider, investigator, outcomes assessor) RCT. Parallel design. 3 arms. Duration: 15 days. Multicentre (33 locations) in the USA.
Participants	Actual enrollment: 142 participants, both genders. Inclusion criteria: aged 12 years or older; diagnosis of CF as determined by the 1997 CF consensus criteria; non‐smoker; FEV1 at Screening Visit 1 between 40% and 90%; stable regimen of CF medications and chest physiotherapy for the 28 days prior to screening; willing to discontinue use of HS for the duration of the trial; clinically stable for at least 2 weeks; all females of child‐bearing potential must have a negative serum pregnancy test and if sexually active must agree to practice a highly effective form of contraception throughout the trial and for 28 days after the last dose of trial medication. Exclusion criteria: history of any organ transplantation or any significant disease or disorder; use of diuretics (including amiloride) or renin‐angiotensin antihypertensive drugs or trimethoprim in the 28 days prior to Screening; history of significant intolerance to inhaled HS, as determined by the investigator; known hypersensitivity to the trial drug or amiloride; any clinically significant laboratory abnormalities at Screening Visit 1 as judged by the investigator (or any of the following: potassium ≥ 5 mEq/L; abnormal renal function; abnormal liver function, defined as ≥ 3 x upper limit of normal; haemoglobin level < 10.0 g/dL); female who is pregnant or lactating; history of sputum or throat swab culture yielding Burkholderia species or Mycobacterium abscessus within 2 years of screening; previous participation in an investigational trial involving administration of any investigational compound or use of an investigational device with 28 days prior to Screening; currently being treated with any ivacaftor containing regimen.
Interventions	Intervention 1: P‐1037 solution for inhalation, 85 μg (28.3 μg/mL) in HS (4.2%). Intervention 2: P‐1037 solution for inhalation, 85 μg (28.3 μg/mL) in 0.17% saline. Control: placebo (0.17% saline). All treatments were inhaled 2x daily.
Outcomes	Primary outcomes Adverse events related to P‐1037 in treatment groups. FEV1 (change from pre‐dosing to 1 hour post‐dosing). Secondary outcomes FEV1 (absolute change from baseline to Day 15). FVC (absolute change from baseline to Day 15). CFQ‐R. FEF25%-75% (absolute change from baseline to Day 15).
Starting date	Trial start date: April 2015. Primary Completion Date: February 2016. Study Completion Date: February 2016.
Contact information	Vertex Pharmaceuticals Incorporated.
Notes	Sponsors and Collaborators: Vertex Pharmaceuticals Incorporated, Parion Sciences.

CF: cystic fibrosis
CFQ‐R: cystic fibrosis questionnaire ‐ revised
DNase: dornase alfa
ENAC: epithelial sodium channels 
FEF_25%-75%: mid‐peak forced expiratory flow rate
FEV₁: forced expiratory volume in one second
FVC: forced vital capacity
HS: hypertonic saline
IPD: individual patient data
IV: intravenous
QoL: quality of life
RCT: randomised controlled trial

Differences between protocol and review

Update 2018

The updated review now includes forced expiratory volume in 0.5 seconds (FEV_0.5)and the lung clearance index (LCI) which were both added as additional outcomes. We also added the outcome of duration of hospital stay due to pulmonary exacerbations and bacteriology. We feel these new outcomes are useful and of interest to clinicians and patients alike.

We have removed the outcome forced expiratory flow at 25–75% of forced vital capacity (FEF_25-75) from the review, since this outcome is not deemed clinically important in this population and is not a reliable measure of lung function. We have also removed the outcome of exhaled nitric oxide as we do not think this is an important outcome for people with cystic fibrosis.

Summary of findings tables have been added to the review.

Update 2023

No further differences between the protocol and review. 

Contributions of authors

For all versions up to 2019, two authors (PW and VMM) selected the trials that were included in this review and each author independently assessed the methodological quality (risk of bias) for each trial. PW and VMM wrote the review.

At the 2023 update, Sherie Smith has been added to the author team and led on study screening, data extraction, data analysis and writing up of the update.

All authors contributed to and approved the final version.

Peter Wark acts as guarantor of the review.

Sources of support

Internal sources

• No sources of support provided

External sources

• National Institute for Health Research, UK

  This systematic review was supported by the National Institute for Health and Care Research (NIHR), via Cochrane Infrastructure funding to the Cochrane Cystic Fibrosis and Genetic Disorders Group.

Declarations of interest

Peter Wark: I declare my position of Fiduciary Officer for Cystic Fibrosis Australia (the main patient body in Australia). I was an investigator and an author on one trial included in the review (Dentice 2016) and a co‐investigator on a further trial, the National Hypertonic Saline in Cystic Fibrosis Study trial (Elkins 2006a).

Vanessa M McDonald has participated in educational symposia funded by GlaxoSmithKline, AstraZeneca, and Menarini, and has participated on advisory boards for GlaxoSmithKline, Novartis, AstraZeneca, and Menarini (unrelated to the present work). She was a co‐investigator in the National Hypertonic Saline in Cystic Fibrosis Study trial (Elkins 2006a).

Sherie Smith has no conflicts of interest to declare. 

Peter Wark and Vanessa M McDonald did not make inclusion decisions or assess risk of bias for their own studies (up to 2019 these judgements were undertaken by editorial base staff and verified by the clinical contact editor). Sherie Smith reassessed and verified these decisions for the 2023 update.

New search for studies and content updated (no change to conclusions)