Inhaled hyperosmolar agents for bronchiectasis

  • Conclusions changed
  • Review
  • Intervention

Authors


Abstract

Background

Mucus retention in the lungs is a prominent feature of bronchiectasis. The stagnant mucus becomes chronically colonised with bacteria, which elicit a host neutrophilic response. This fails to eliminate the bacteria, and the large concentration of host-derived protease may contribute to the airway damage. The sensation of retained mucus is itself a cause of suffering, and the failure to maintain airway sterility probably contributes to the frequent respiratory infections experienced by many patients.

Hypertonic saline inhalation is known to accelerate tracheobronchial clearance in many conditions, probably by inducing a liquid flux into the airway surface, which alters mucus rheology in a way favourable to mucociliary clearance. Inhaled dry powder mannitol has a similar effect. Such agents are an attractive approach to the problem of mucostasis, and deserve further clinical evaluation.

Objectives

To determine whether inhaled hyperosmolar substances are effective in the treatment of bronchiectasis.

Search methods

We searched the Cochrane Airways Group Specialised Register, trials registries, and the reference lists of included studies and review articles. Searches are current up to April 2014.

Selection criteria

Any randomised controlled trial (RCT) using hyperosmolar inhalation in patients with bronchiectasis not caused by cystic fibrosis.

Data collection and analysis

Two review authors assessed studies for suitability. We used standard methods recommended by The Cochrane Collaboration.

Main results

Eleven studies met the inclusion criteria of the review (1021 participants).

Five studies on 833 participants compared inhaled mannitol with placebo but poor outcome reporting meant we could pool very little data and most outcomes were reported by only one study. One 12-month trial on 461 participants provided results for exacerbations and demonstrated an advantage for mannitol in terms of time to first exacerbation (median time to exacerbation 165 versus 124 days for mannitol and placebo respectively (hazard ratio (HR) 0.78, 95% confidence interval (CI) 0.63 to 0.96, P = 0.022) and number of days on antibiotics for bronchiectasis exacerbations was significantly better with mannitol (risk ratio (RR) 0.76, 95%CI 0.58 to 1.00, P = 0.0496). However, exacerbation rate per year was not significantly different between mannitol and placebo (RR 0.92 95% CI 0.78 to 1.08). The quality of this evidence was rated as moderate. There was also an indication, from only three trials, again based on moderate quality evidence, that mannitol improves health-related quality of life (mean difference (MD) -2.05; 95% CI -3.69 to -0.40). An analysis of adverse events data, also based on moderate quality evidence, revealed no difference between mannitol and placebo (OR 0.96; 95% CI 0.61 to 1.51). Two additional small trials on 25 participants compared mannitol versus no treatment and the data from these studies were inconclusive.

Four studies (combined N = 113) compared hypertonic saline versus isotonic saline. On most outcomes there were conflicting results and the opportunities for the statistical aggregation of data from studies was very limited. It is not possible to draw robust conclusions for this comparison and judgments should be reserved until further data are available.

Authors' conclusions

There is an indication from a single, large, unpublished study that inhaled mannitol increases time to first exacerbation in patients with bronchiectasis. In patients with near normal lung function, spirometry does not change dramatically with mannitol and adverse events are not more frequent than placebo. Further investigation is required in a patient population with impaired lung function.

It is not possible to draw firm conclusions regarding the effect of nebulised hypertonic saline due to significant differences in the methodology, patient groups, and findings amongst the limited data available. The data suggest that it is unlikely to have benefit over isotonic saline in patients with milder disease, and hence future studies should test its use in those with more severe disease

Plain language summary

Inhaled hyperosmolar agents for bronchiectasis

Review question

We wanted to know if inhaled hyperosmolar agents - treatments which help people cough up sputum - may be helpful for people with bronchiectasis. We only included trials on people who had bronchiectasis and who did not have cystic fibrosis. Therefore we are unable to draw any conclusions for people with cystic fibrosis.

Background

Bronchiectasis is a lung condition that usually develops after a series of lung problems (such as childhood infections, problems in the lung structure, tuberculosis, and cystic fibrosis). A lot of mucus (phlegm) collects in the lungs, causing discomfort and the need to cough it up. The phlegm also collects bacteria, which can add to breathing difficulties and make people very ill by causing recurring lung infections that are difficult to clear with antibiotics. Breathing in (inhaling) hypertonic saline (salt solutions with a greater salt content than blood) liquids may help clear this mucus, as may the drug mannitol (inhaled in dry powder form). This is because the concentrated salt or sugar (mannitol) draws water into the mucus in the lung and makes it thinner and easier to cough out.

Study characteristics

We found 11 randomised controlled trials on 1021 participants that compared inhaled hyperosmolar agents versus no mucolytic treatment. Five studies compared inhaled mannitol versus placebo (with a total of 883 participants) and two very small studies (with a total of just 25 participants) compared inhaled mannitol with no treatment. We also found four studies (with a total of 113 participants) that compared hypertonic saline with isotonic (normal) saline.

Key results

For the comparison between mannitol and placebo only one study (a 12-month trial with 461 participants) provided information on the number of people who had an exacerbation (or flare up) over the course of a year. This study showed that people who were treated with mannitol had 8% fewer exacerbations on average compared with placebo. Overall, we felt the quality of this evidence was moderate and new trials would be likely to change either how effective we think the treatment is or how confident we are about it.

Three trials assessed the effect of mannitol on health-related quality of life, and again the quality of the evidence was rated as moderate. An analysis of adverse events data, also based on moderate quality evidence, revealed no difference between mannitol and placebo

The trials comparing hypertonic saline with isotonic saline had conflicting results for most of the outcomes of interest. Because we were unable to combine the data, it is not possible to draw robust conclusions for this comparison and judgments should be reserved until further data are available. Our analysis of adverse events between hypertonic saline versus isotonic saline showed no significant difference however this was based on a single study and the quality of the evidence was moderate.

Quality of the evidence

Details of how the patients in the trials were allocated to receive mannitol or not was clearly described in only one of the studies, and similarly only one of the hypertonic saline versus isotonic saline studies provided this information. The general lack of information on this point was considered carefully in the review in relation to our level of uncertainty in interpreting the results. Taking this into account, the quality of evidence was generally regarded as moderate both for the mannitol and hypertonic saline studies.

Summary of findings(Explanation)

Summary of findings for the main comparison. Inhaled mannitol versus placebo for bronchiectasis
  1. 1 (-1 limitations) One point deducted based on risk of bias assessment of trial NCT00669331

    2 (-1 inconsistency) One point deducted to reflect the variability among the included studies with respect to risk of bias

    3 (-1 limitations) One point deducted based on risk of bias assessment of trial Bilton 2013

Inhaled mannitol versus placebo for bronchiectasis

Patient or population: patients with bronchiectasis
Settings: community
Intervention: Inhaled mannitol

Comparison: placebo

OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of Participants
(studies)
Quality of the evidence
(GRADE)
Comments
Assumed riskCorresponding risk
Placebo Inhaled mannitol versus placebo

Exacerbations

Exacerbation rate per year.

Follow-up: 12 months

See commentSee comment

RR 0.92

(0.78 to 1.08)

461

(1 study)

⊕⊕⊕⊝

moderate 1

These data for this study are unpublished and were presented in summarised form the European Respiratory Society Conference 2013

Hospitalisations

Number of participants experiencing one of more hospitalisations.

Follow-up: 12 months

See commentSee comment

RR 0.61

(0.34, 1.09)

461

(1 study)

⊕⊕⊕⊝

moderate 1

These data for this study are unpublished and were presented in summarised form the European Respiratory Society Conference 2013

Adverse events

Follow-up: 12 weeks (Bilton 2013) and 12 months (NCT00669331)

894 per 1000 890 per 1000
(837 to 927)
OR 0.96
(0.61 to 1.51)
804
(2 studies)

⊕⊕⊕⊝

moderate 2

Bilton: "In both cohorts, most AEs were unrelated to study treatment. " There was also no significant difference between the two conditions with respect to SAEs OR 0.79 [0.52, 1.19]
MortalitySee commentSee commentSee commentSee commentSee commentNo deaths reported in any study

Health-related quality of life measured on the SGRQ

Scale from 0 to 100. Low on the scale is better.

Follow-up: 2 weeks (Daviskas 2004), 12 weeks (Bilton 2013) and 12 months (NCT00669331)

 The mean change from baseline in the SGRQ total score in the intervention groups was
-2.05 standard deviations higher
(-3.69 lower to -0.40 higher)

MD -2.05

(-3.69 to -0.40)

840

(3 studies)

⊕⊕⊕⊝

moderate 2

 

Symptoms

Bronchiectasis Symptoms Questionnaire Follow-up: 12 weeks

The mean symptom score was 28.4The mean symptom score was 1.20 lower (3.91 lower to 1.51 higher] MD -1.20 (-3.91 to 1.51)

343

(1 study)

⊕⊕⊕⊝

moderate 3

 

Lung function

FEV1 (L)

Follow-up: 12 weeks

The mean FEV1 was 1.95 LThe mean change in FEV1 was 0.03 L lower (0.10 lower to 0.16 higher) MD 0.03 (-0.10, 0.16)

343

(1 study)

⊕⊕⊕⊝

moderate 3

 
*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; MD: mean difference; OR: Odds ratio; SGRQ: St George's Respiratory Questionairre
GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.

Summary of findings 2 Hypertonic saline versus isotonic saline for bronchiectasis

Summary of findings 2. Hypertonic saline versus isotonic saline for bronchiectasis
  1. 1 A point was deducted to reflect the precision of the estimate as the data were contributed by only one study (with 40 participants)

Hypertonic saline versus isotonic saline for bronchiectasis 

Patient or population: patients with bronchiectasis
Settings: community
Intervention: Hypertonic saline

Control: Isotonic saline

 
OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of Participants
(studies)
Quality of the evidence
(GRADE)
Comments 
Assumed riskCorresponding risk 
Isotonic saline Hypertonic saline  

Frequency and duration of exacerbations

Follow-up: 3 months (Kellet 2011) and 12 months (Nicolson 2012)

See commentSee commentSee comment70 (2 studies) 

One cross-over study (Kellet 2011) showed a significant difference; the rate of exacerbations in the hypertonic saline group (2.14 exacerbations/year) was less than half the rate in the isotonic saline group (4.85 exacerbations/year), but no measure of dispersion was provided.

In Nicolson 2012 there were no significant differences between hypertonic saline versus isotonic saline on exacerbations (over 12 months), exacerbations requiring antibiotics, number of exacerbation days or number of exacerbation days requiring antibiotics. Median number of exacerbations (inter-quartile range) were 1 (0 to 4) for isotonic saline and 3 (0 to 6) for hypertonic saline (P = 0.24) .

 

Hospitalisations

Follow-up: 12 months

See commentSee commentSee comment40 (1 study) In Nicolson 2012 there was no significant difference between hypertonic saline versus isotonic saline with respect to hospital admissions (P = 0.34) or hospital days (P = 0.36). Four (10%) participants had hospital admissions, three in the isotonic saline group were in hospital for three, five and 61 days respectively and one participant in the hypertonic saline group was in hospital for 68 days. 

Adverse events

Follow-up: 12 months

0 per 1000 150 per 1000 OR 8.2
(0.4 to 169.9)
40
(1 study)
 

⊕⊕⊕⊝

moderate1

 

Mortality

Follow-up: 4 weeks Bradley 2011 and Kellett 2005, 3 months (Kellet 2011) and 12 months (Nicolson 2012)

See commentSee commentSee comment113 (4 studies) No deaths reported in any study 

Symptoms: cough, sputum volume and ease of expectoration, wheeze, dyspnoea

Follow-up: 4 weeks Bradley 2011, and Kellett 2005, 3 months (Kellet 2011) and 12 months (Nicolson 2012)

See commentSee commentSee comment113 (4 studies) A variety of symptom related outcomes were reported including sputum expectoration weight, sputum viscosity, ease of expectoration and cough. None of the data was reported in a manner sufficient for meta-analysis. 

Lung function

FEV1 (L)

Follow-up: 12 months

The mean FEV1 in the control group was 2.37 L.The mean FEV1 (L) - 12 months in the hypertonic saline group was
0.19 higher
(0.37 lower to 0.75 higher)
 40
(1 study)

⊕⊕⊕⊝

moderate1

  

Health-related quality of life. Measured on the SGRQ (total score reported)

Follow-up: 12 months

See commentSee comment

)
See comment40
(1 study)

⊕⊕⊕⊝

moderate1

SGRQ total was not reported for this study (Nicolson 2012) No significant difference in SGRQ symptoms, activity or impact domains. 
*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; OR: Odds ratio;
 
GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate.
Very low quality: We are very uncertain about the estimate.
 

Background

Description of the condition

Bronchiectasis is a disorder of the lungs characterised by irreversible dilatation of the bronchi. It is the end result of a number of different insults to the lungs. These include inherited diseases of the lungs’ defence mechanisms, such as ciliary dyskinesia and primary immunodeficiencies, congential structural lung abnormalities and acquired problems, such as the aftermath of tuberculosis or childhood respiratory infection, large airway obstruction and chronic inflammatory conditions (for example, rheumatoid arthritis). Many cases of bronchiectasis are idiopathic (Anwar 2013). Cystic fibrosis is the cause of the majority of cases of severe bronchiectasis (Sly 2013), however the management of cystic fibrosis-related bronchiectasis is considered separately in other reviews.

Impaired lung defences, changes in mucus consistency and mechanical obstruction of the bronchi all lead to a higher risk of lung infection and associated damage and dilatation of the bronchi. This dilatation in turn promotes mucus retention, leading to further and more protracted pulmonary infection and hence, a vicious cycle of disease progression.

The true population prevalence of bronchiectasis is unknown (Frey 2007; Shoemark 2007). Weycker 2005 suggested the prevalence could be as high as 270 cases per 100,000 in the over 75s. Bronchiectasis frequently complicates other respiratory diseases, such as chronic obstructive pulmonary disease (COPD) where its presence is an independent predictor of mortality (Martinez 2013).

Diagnosis is by high resolution computed tomography of the thorax (Bonavita 2012) in patients with symptoms of chronic or recurrent pulmonary suppuration.

Description of the intervention

Inhaled hyperosmolar agents are used in conditions, such as non-cystic fibrosis bronchiectasis, cystic fibrosis, COPD and asthma (Daviskas 1996; Elkins 2006a; Pavia 1978). The aim of treatment is to facilitate tracheo-bronchial mucous clearance. The first documented use of such a strategy was in 1978 using nebulised 1.21 M saline (approximately 7% saline) (Pavia 1978). Since then concentrations from 3% to 14.4% (Daviskas 1996; Robinson 1997) have been used in various studies. Various devices have been used to drive the nebulisation, from air-driven small-volume chambers, as typically used for delivering nebulised bronchodilators in asthma, through to low- and high-volume ultrasonic devices. The nebulisation process can be somewhat cumbersome and time consuming. Inhaled dry powder mannitol, which was developed initially as a bronchial challenge agent in asthma, is being increasingly used as an alternative. It is delivered from capsules via a purpose-designed inhaler device.

National guidelines on the diagnosis and management of bronchiectasis from several countries (BTS 2010; TSANZ 2010) generally ignore hyperosmolar agents on the grounds of insufficient evidence to make a positive recommendation. Anecdotally however hyperosmolar agents, particularly hypertonic saline, are increasingly prescribed in clinical practice.

How the intervention might work

The key clinical feature of bronchiectasis is chronic mucus hypersecretion, mucus retention or both. These are central to the vicious cycle of disease progression described above, causing the characteristic chronic productive cough and predisposing to recurrent lower respiratory tract infections. One of the central tenets of treatment for bronchiectasis is therefore the application of strategies to facilitate mucus clearance. This includes chest clearance manoeuvres instigated and taught by specialist respiratory physiotherapists (Pasteur 2010). Such clearance is however hindered both by the volume but also the viscosity of the sputum.

Whilst the specific clinically relevant mechanism has not been elucidated, hyperosmolar agents are used principally to reduce sputum viscosity and therefore ease chest clearance. When inhaled they are deposited in the airway, and the consequence of their hyperosmolarity in theory is that they draw fluid from the airway epithelium onto the mucosal surface and hence into contact with the mucous (Wills 2006). This then has the effect of both loosening the mucous and making it less viscous (Shibuya 2003; Wills 1997) which (particularly when used together with effective chest clearance techniques) will facilitate expectoration (Daviskas 2001; Kellett 2005).

One potential disadvantage of such agents occurs where there might be an asthmatic element to the bronchiectasis, such as in allergic bronchopulmonary aspergillosis. Inhaled hyperosmolar agents also have the effect of desiccating airway inflammatory cells, disrupting the cell membrane and causing release of bronchoconstricting and pro-inflammatory mediators, such as histamine and leukotrienes. This may cause significant acute bronchoconstriction, and hence it is recommended that a supervised test dose be administered with spirometry or peak expiratory flow (PEF) measurements before and after (Pasteur 2010). Those at risk of bronchial hyperresponsiveness should be pre-treated with an inhaled bronchodilator (Pasteur 2010).

Potential non-mucolytic mechanisms for the benefit of hypertonic saline could arise from its relative pro- and anti-inflammatory effects. Recent in vitro studies suggest that high concentrations of salt induce type 17 helper T lymphocytes, which in turn produce the antibacterial cytokines interleukins 17, 21 and 22 (Kleinewietfeld 2013; Wu 2013). It also may also have beneficial antiinflammatory effects by decreasing local interleukin 8 levels (Reeves 2011).

Why it is important to do this review

Clinically bronchiectasis manifests as chronic mucus production, along with a variable amount of airflow limitation. Mucus hypersecretion is a distressing symptom. Mucus plugging (obstruction of small- or medium-sized airways with mucus) can cause breathlessness and hypoxia. Inhaled hyperosmolar agents should in theory draw water into the airways, particularly the airway surface layer of fluid (Enderby 2007). This has the effect of improving ciliary function, of making sputum thinner, more watery and easier to cough out, and of stimulating cough (Elkins 2006), thus improving mucus clearance and the reducing risk of mucus plugging. If inhaled hyperosmolar agents are able to reduce these effects to a clinically significant degree in vivo, they would be useful drugs in the treatment of non-cystic fibrosis bronchiectasis.

Objectives

To determine whether inhaled hyperosmolar treatments are effective in the treatment of bronchiectasis.

Methods

Criteria for considering studies for this review

Types of studies

Any randomised trial comparing treated and untreated patients with bronchiectasis.

Types of participants

People with bronchiectasis of any cause except cystic fibrosis.

Types of interventions

Inhalation of any hyperosmolar substance, either single dose or more prolonged treatment.

Types of outcome measures

Primary outcomes
  • Frequency and duration of exacerbations

  • Hospitalisations

  • Adverse events

Secondary outcomes
  • Mortality

  • Health-related quality of life (measured on a validated questionnaire e.g. SF-36 and St George's Respiratory Questionnaire (SGRQ))

  • Serious adverse events

  • Symptoms: cough, sputum volume and ease of expectoration, wheeze, dyspnoea

  • Lung function

  • In vitro characteristics of sputum

  • Measurement of tracheobronchial clearance

Search methods for identification of studies

Electronic searches

We identified trials from the Cochrane Airways Group's Specialised Register (CAGR), which is maintained by the Trials Search Co-ordinator for the Group. The Register contains trial reports identified through systematic searches of bibliographic databases including the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, EMBASE, CINAHL, AMED, and PsycINFO, and handsearching of respiratory journals and meeting abstracts (please see Appendix 1 for further details). We searched all records in the CAGR using the search strategy in Appendix 2.

We also conducted a search of ClinicalTrials.gov (www.ClinicalTrials.gov) and the WHO trials portal (www.who.int/ictrp/en/). We searched all databases from their inception up to April 2014, with no restriction on language of publication.

Searching other resources

We searched all reference lists of available primary studies and review articles to identify other potentially relevant citations.

Data collection and analysis

Selection of studies

Review authors KS and IC independently screened the identified references using the abstract, title and MeSH headings, and independently assessed studies for potential relevance. At the next stage, using the full text of the potentially relevant studies, the same review authors (KS and IC) independently selected trials for inclusion in the review. Had there been disagreements, we planned to involve an independent third party adjudicator (SM); however, this was not necessary.

Data extraction and management

Data for included trials were extracted independently by two review authors (AH and IC) and entered into The Cochrane Collaboration's software program (Review Manager (RevMan)).

Assessment of risk of bias in included studies

Two review authors (IC and AH) assessed the trials with respect to selection bias, performance and detection bias, attrition bias, reporting bias and other potential sources of bias using The Cochrane Collaboration’s ’Risk of bias’ tool (Higgins 2011).

Measures of treatment effect

For dichotomous variables, we expressed data as odds ratios (OR) with 95% confidence intervals (CI). Data for continuous variables were reported as mean differences (MD) with 95% CIs or standardised mean differences (SMD) with 95%CIs in analyses where it was necessary to pool data from different measures.

Unit of analysis issues

The unit of analysis was the patient.

Dealing with missing data

We planned to contact authors if outcome data or information on trial design were missing; this was necessary in just one case Chandra 2008 and we are awaiting a response.

Assessment of heterogeneity

We tested heterogeneity among pooled estimates using the Der-Simonian and Laird method; and we considered P < 0.05 as the threshold for statistical significance. Heterogeneity was assessed at first using visual inspection of forest plots. The Chi2 test was similarly considered (P < 0.10) but interpreted with caution owing to the low power associated with this test. I2 was also considered and interpreted in relation to the following guidance (Higgins 2011):

  • 0% to 40%: might not be important;

  • 30% to 60%: may represent moderate heterogeneity;

  • 50% to 90%: may represent substantial heterogeneity;

  • 75% to 100%: considerable heterogeneity.

Where we encountered heterogeneity according to the above mentioned criteria, we applied the fixed- and random-effects models and commented on any differences, reporting the random-effects model in the review.

Assessment of reporting biases

We planned to examine publication bias using funnel plots if we had included an adequate number of trials (10 or more) aggregated in a single meta-analysis. We recognise that an asymmetric funnel plot can reflect heterogeneity, outcome reporting bias and small-study effects and may therefore not necessarily reflect publication bias.

Data synthesis

'Summary of findings' table

We created a 'Summary of findings' table using the following outcomes.

  • Frequency and duration of exacerbations

  • Hospitalisations

  • Adverse events

  • Mortality

  • Symptoms: cough, sputum volume and ease of expectoration, wheeze, dyspnoea

  • Lung function

  • Health-related quality of life (e.g. SF-36 and St George's Respiratory Questionnaire (SGRQ)

We used the five GRADE considerations (study limitations, consistency of effect, imprecision, indirectness and publication bias) to assess the quality of a body of evidence as it relates to the studies which contribute data to the meta-analyses for the prespecified outcomes. We used methods and recommendations described in Section 8.5 and Chapter 12 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011) using GRADEpro software. We justified all decisions to down- or up-grade the quality of studies using footnotes and included comments to aid reader's understanding of the review where necessary.

Subgroup analysis and investigation of heterogeneity

Subgroup and sensitivity analysis were performed by pooling absolute and relative data, in order to include sufficient studies at each time point. In these cases, we calculated individual and pooled statistics as MD and 95% CIs using a random-effects model. Subgroup analysis was performed using the following subgroups.

  • Dose (160 mg, 320 mg and 480 mg). This was feasible with the available data only the in inhaled mannitol versus no treatment comparison) As data were used in each subgroup from the same trial, the data from the subgroups were not aggregated.

  • Time (three, six and 12 months). This was feasible with the available data only in the hypertonic saline versus isotonic saline comparison. Again as data were used in each subgroup from the same trial, the data from the subgroups were not aggregated.

Sensitivity analysis

Sensitivity analyses were performed by comparing random-effects versus fixed-effect modelling.

Results

Description of studies

Results of the search

Thirty-six reports were identified in the searches up to April 2014 (Figure 1). They were independently evaluated against the inclusion criteria and 22 reports of 11 studies were judged as appropriate for inclusion (details are provided in Characteristics of included studies). Fourteen studies were excluded and they were added to the two identified in the previous version of the review to bring the total to 16.

Figure 1.

Study flow diagram.

Included studies

Eleven studies involving 1021 participants met our eligibility criteria; two (Daviskas 1999; Daviskas 2004) of the 11 were identified in the previous version of this review (Wills 2006). We found trials on three separate comparisons: inhaled mannitol versus placebo (five studies), inhaled mannitol versus no treatment (two studies) and hypertonic saline versus isotonic saline (four studies).

Inhaled mannitol versus placebo

Five studies on 883 participants (number of participants analysesd) compared inhaled mannitol with placebo (Bennoor 2012 (N = 50, Bilton 2013 (N = 343), Chandra 2008 (N = 12), Daviskas 2004 (N = 17), NCT00669331 (N = 461)). Only one of the five studies (Daviskas 2004) was of cross-over design and the others were parallel group trials.

Only one of the five trials was single blind (Bennoor 2012); the remaining four were double-blind studies (Bilton 2013; Chandra 2008; Daviskas 2004; NCT00669331). Study duration varied between 12 days and 12 months (Bennoor 2012 12 days, Bilton 2013 12 weeks, Chandra 2008 12 weeks, Daviskas 2004 two weeks and NCT00669331 12 months). There was also variation among the doses used in the studies. In Bilton 2013 the comparison was between inhaled 320 mg mannitol versus placebo twice daily. In Bennoor 2012 400 mg mannitol (10 capsules) or placebo was inhaled via rotahaler daily and in Daviskas 2004 the comparison was 400 mg inhaled mannitol versus placebo twice daily. In NCT00669331 mannitol 400 mg twice daily for 52 weeks was compared with matched control 10 capsules twice a day for 52 weeks. Chandra 2008 also used twice daily mannitol inhalation, however the dose is not specified in the trial report.

Inhaled mannitol versus no treatment

Two small cross-over studies involving 25 participants) reported this comparison (Daviskas 1999, N = 11; Daviskas 2008 N = 14. In Daviskas 1999 approximately 300 mg inhaled dry powder mannitol was delivered via approximately nine capsules and Daviskas 2008 used 160 mg, 320 mg or 480 mg of mannitol.

Hypertonic saline versus isotonic saline

Four studies with 113 participants compared hypertonic saline versus isotonic saline ( Bradley 2011 (N = 19), Kellett 2005 (N = 24), Kellet 2011 (N = 30), Nicolson 2012 (N = 40)). Three of the four studies were cross-over design (Bradley 2011; Kellett 2005; Kellet 2011) and Nicolson 2012 was a 12-month parallel group trial. Only one of the studies Kellet 2011 was single blind; the remaining three were double blind. There was considerable variation among the durations of these trials: Bradley 2011 was a 13-week trial with treatment for four weeks and a two-week washout between treatments,

Kellett 2005 used a single treatment of 10 to 20 minutes in a four-week trial; Kellet 2011 was a eight-month trial with treatments for three months and Nicolson 2012 ran for 12 months. There was relatively little variation among the interventions. Bradley 2011 used twice daily nebulised hypertonic saline (6%) masked with quinine sulfate for isotonic saline versus twice daily isotonic saline (0.9%) (nebulised) masked with quinine sulfate for hypertonic saline. In Kellett 2005 the comparison was between hyperosmolar saline - 7% after nebulised terbutaline versus normal saline after nebulised terbutaline. Kellet 2011 used 4 mL hypertonic saline (7%) once day versus 4 mL isotonic saline (0.9%) once day, and Nicolson 2012 compared hypertonic saline (6%) - 5 mL twice a day via a nebuliser for 12 months versus isotonic saline.

Excluded studies

Sixteen studies were excluded for the following reasons: 11 (69%) did not include a hyperosmolar agent, four (25%) were not randomised and the remaining study (6%) did not include patients with bronchiectasis (full details provided in Characteristics of excluded studies).

Risk of bias in included studies

Allocation

Two studies Kellett 2005; Nicolson 2012 were judged to be at low risk of bias in terms of random sequence generation, and the remaining nine studies were considered as unclear risk. In terms of allocation concealment only Nicolson 2012 was judged to be low risk of bias, and in the remaining 10 studies the risk was evaluated as unclear (Figure 2).

Figure 2.

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

Blinding

In terms of performance bias, three studies were judged as low risk (Bilton 2013, Kellett 2005; Nicolson 2012). In five studies the risk was considered to be unclear (Bradley 2011; Chandra 2008; Daviskas 1999; Daviskas 2004; NCT00669331) and in the remaining three the risk was evaluated as high (Bennoor 2012; Daviskas 2008; Kellet 2011). On detection bias, two studies were judged as low (Kellett 2005; Nicolson 2012). The level of risk was viewed as unclear in seven (Bilton 2013; Bradley 2011; Chandra 2008; Daviskas 2008; Daviskas 2004; Kellet 2011; NCT00669331). In the following two studies the risk was judged to be high (Bennoor 2012; Daviskas 1999).

Incomplete outcome data

Only five studies (Bilton 2013; Kellet 2011; Kellett 2005; NCT00669331; Nicolson 2012) were judged to be low risk in terms of attrition bias, and only one Bradley 2011, was judged to be in the high risk of bias category. The remaining five (Bennoor 2012; Chandra 2008; Daviskas 1999; Daviskas 2008; Daviskas 2004) were considered as unclear in this respect Figure 3.

Figure 3.

'Risk of bias' summary: review authors' judgements about each risk of bias item for each included study.

Selective reporting

Three studies (Bilton 2013; Daviskas 1999; Nicolson 2012) were evaluated as low risk in terms of reporting bas. The remaining eight (Bennoor 2012; Bradley 2011; Chandra 2008; Daviskas 2008; Daviskas 2004; Kellet 2011; Kellett 2005; NCT00669331) were considered as unclear on this measure.

Other potential sources of bias

In this category all 11 studies were evaluated as unclear.

Effects of interventions

See: Summary of findings for the main comparison Inhaled mannitol versus placebo for bronchiectasis; Summary of findings 2 Hypertonic saline versus isotonic saline for bronchiectasis

Inhaled mannitol versus placebo

Five studies (combined N = 883) compared mannitol versus placebo: Bennoor 2012 (50 participants), Chandra 2008 (12 participants), Bilton 2013 (362 participants randomised and 343 analysed), Daviskas 2004 (17 participants) and NCT00669331 (461 participants). Only one of the six studies Daviskas 2004 was a cross-over design and Bennoor 2012, Bilton 2013, Chandra 2008 and NCT00669331 were parallel group trials. Personal communication with Dr Bilton confirmed that people on placebo in NCT00669331 were given 50 mg of mannitol (intended as a non-therapeutic dose), and there is a possibility that this may have led to conservative estimates of the treatment effects. In most cases the results presented in the trial reports were not in a format that enabled the data to be entered into RevMan for the purposes of meta-analysis and are therefore reported narratively here. The quality of evidence for this comparison where GRADE assessments were applied was moderate (Summary of findings for the main comparison).

Exacerbations

The only data available from the included trials for this outcome were provided by NCT00669331, a C trial with 461 participants, and reported as exacerbation rate per year. There was an 8% reduction (risk ratio (RR) 0.92 95% confidence interval (CI) 0.78 to 1.08) in favour of mannitol, however this was not significant. These data for this study are unpublished and were presented at the European Respiratory Society Conference 2013.

However, the median time to first exacerbation was significantly different between the two groups (median: 165 versus 124 days for mannitol and placebo respectively (hazard ratio (HR) 0.78, 95% CI 0.63 to 0.96, P = 0.022, NCT00669331).

There was also a significant difference in favour of mannitol (RR 0.76, 95% CI 0.58 to 1.00, P = 0.0496) for the number of days on antibiotics for pulmonary exacerbations (NCT00669331). We acknowledge that time to first exacerbation and antibiotics for pulmonary exacerbations, per se, were not prespecified in our planned outcomes for this review. However, we are including this additional information as these outcomes are so very closely linked to our prespecified outcome of exacerbations.

Hospital admissions

Data for this outcome were provided by one trial on 461 participants (NCT00669331), as hospitalisations due to pulmonary exacerbations. There was no significant difference between the two groups (RR 0.61, 95% CI 0.34 to 1.09, P = 0.093).

Mortality

Only one study reported deaths (Bilton 2013). They reported 2/241 deaths in the mannitol-treated group, but stated that they were not related to study treatment.

Health-related quality of life

Three trials on 821 participants reported St George's Respiratory Questionnaire (SGRQ) data (Bilton 2013, Daviskas 2004, NCT00669331). The direction of scaling on the SGRQ indicates that a score of 100 represents worst possible quality of life and zero represents best possible quality of life. There was a significant difference in favour of mannitol on SGRQ total scores (mean difference (MD) -2.05; 95% CI -3.69 to -0.40 Analysis 1.1). There were considerable differences in duration of the three trials (Bilton 2013 12 weeks, Daviskas 2004 two weeks and NCT00669331 52 weeks), but low heterogeneity (I2 very near zero). The difference between the two groups is less than four points which is generally regarded as the minimum clinically important difference (MCID) for SGRQ total scores in COPD studies (Jones 2005) and the lower confidence limit is smaller than that threshold (-3.69); whether this MCID may also hold for bronchiectasis studies is unclear.

In separate analyses of component scores, data were available from only Bilton 2013 and Daviskas 2004. There were no significant differences between mannitol versus placebo on symptoms component (MD -1.42; 95% CI -5.12 to 2.28 Analysis 1.2) and activity component (MD -0.84; 95% CI -3.95 to 2.28 Analysis 1.3); however there was a significant difference in favour of mannitol on impact component scores (MD -3.16; 95% CI -5.77 to -0.55 Analysis 1.4). In view of the high level of heterogeneity (I² = 58%) on the impact component scores analysis we used a random-effects model and on that basis the significant difference was not observed (MD -2.21; 95% CI -5.92 to 1.49 Analysis 1.4); we therefore conclude that the significant effect on impact scores reported in Wills 2006 should be viewed with caution in light of the data now available from Bilton 2013.

However, in a recent parallel group study of 50 patients with stable bronchiectasis Bennoor 2012 reported a significant reduction from baseline in SGRQ symptom scores and impact scores but did not provide any summary data. There was no difference between the two groups with respect to SGRQ activity scores. With regard to total SGRQ scores and dyspnoea scores, there was a significant advantage in favour of mannitol (P < 0.001). These results were not reported in a format that enabled us to include the study in the meta-analyses.

Bilton 2013 reported no significant difference in cough score between mannitol and placebo for Leicester Cough Questionnaire (LCQ) data (MD 0.00; 95% CI -0.81 to 0.81, Analysis 1.5).

Symptoms

One trial on 343 participants (Bilton 2013) reported no significant difference between mannitol and placebo for Bronchiectasis Symptoms Questionnaire data (MD -1.20; 95% CI -3.91 to 1.51, Analysis 1.6).

In Daviskas 2004 symptom scores were reported as a domain of the SGRQ (see above).

Lung function

There was no significant difference in FEV1 % predicted between mannitol versus placebo (% predicted FEV1 2.70; 95% CI -8.53 to 13.93, Analysis 1.7) in Daviskas 2004 or in FEV1 (L) at 12 weeks (MD 0.03; 95% CI -0.10 to 0.16 Analysis 1.8). There were also no significant or clinically important differences in Chandra 2008 with respect to changes from baseline in FEV1 in people on either mannitol or placebo, and it was not possible to aggregate these data with those from other studies .

At 12 weeks, in Bilton 2013, there was no significant difference between mannitol versus placebo with respect to FVC (MD 0.04; 95% CI -0.15 to 0.23 Analysis 1.14) or FEF25-75 ((MD 0.07; 95% CI -0.09 to 0.23 Analysis 1.10). However, Bennoor 2012 reported a significant differential improvement in FVC (P < 0.039) and FEF25-75 (P < 0.045) on day 12; these results were not reported in a manner that enabled us to aggregate these data with other studies.

Sputum volume

Bilton 2013 reported a significant difference between the mannitol and placebo in the 24-hour sputum weight from baseline, noting a significant reduction in weight of sputum expectorated in the placebo group and not in those on mannitol (MD 4.32; 95% CI 1.60 to 7.04 Analysis 1.11), and this was interpreted by the authors in relation to the higher antibiotic use in the placebo group.

However in Bennoor 2012, there was no significant difference in sputum volume between the mannitol and placebo arms. No summary statistics were provided and it was therefore not possible to combine these results with those from other studies.

Antibiotic use

The difference between the mannitol and placebo groups in Bilton 2013 was not significant (odds ratio (OR) 0.72; 95% CI 0.46 to 1.14 Analysis 1.12).

Incremental shuttle walk test

Bilton 2013 reported no significant difference between mannitol versus placebo in the incremental shuttle walk test at week 12 (MD 18.70; 95% CI -20.50 to 57.90 Analysis 1.13).

Adverse events and serious adverse events

Bilton 2013 and NCT00669331 reported no significant difference between mannitol and placebo in either adverse events (OR 0.96; 95% CI 0.61 to 1.51 Analysis 1.9) or serious adverse events (OR 0.79; 95% CI 0.52 to 1.19 Analysis 1.15). The number of adverse events was high in Bilton (placebo: 80% versus mannitol: 82%), but most adverse events were unrelated to the treatment.

Bennoor 2012 reported that two patients in the mannitol group experienced dry mouths and four reported headache, fatigue and chest pain; it is unclear from the trial report whether the participants experiencing dry mouths were included among the four others and we are therefore reporting this narratively. No adverse events were reported in the placebo group.

Inhaled mannitol versus no treatment

Two small cross-over studies (N = 25) compared inhaled mannitol versus no treatment: Daviskas 1999 (11 participants) and Daviskas 2008 (14 participants). In Daviskas 2008 and Daviskas 1999 no treatment is referred to in the trial report as baseline (please see Characteristics of included studies). Data from both studies were not reported in a format that enabled us to enter them into a meta-analysis.

Exacerbations

There were no data available from the included trials for this outcome.

Hospital admissions

There were no data available from the included trials for this outcome.

Lung function

There were data available from only one of the included trials for this outcome Daviskas 2008. The radioaerosol deposition was consistent with the FEV1 % predicted being similar on all study days (76.7 (SD) 3.6; 76.5 (SD) 3.7; 77.0 (SD) 3.7; 76.0 (SD) 3.4 and 76.6 (SD) 3.7% for baseline, control, 160, 320 and 480 mg mannitol, respectively; P > 0.4).

Mortality

No deaths were reported.

Mucus clearance

Data were provided by two very small cross-over studies: Daviskas 1999 (11 participants) and Daviskas 2008 (14 participants). The latter suggests dose response effects (with 160, 320 and 480 mg mannitol) in mucus clearance from the whole right lung when compared to no treatment over 75 minutes; each of the three doses produced significantly more mucus clearance at that time point versus a no treatment condition, with 480 mg producing significantly more mucus clearance than 160 mg (P < 0.0001). The difference between treatment and no treatment was: 160 mg condition: 12.0% mean clearance P < 0.01, 320 mg condition: 16.8% mean clearance P < 0.001 and 480 mg condition: 21.9% mean clearance P < 0.001).

In Daviskas 1999, only one dose was used (reported as approximately 300 mg) and the difference between treatment and no treatment was 22.3% mean clearance over 75 minutes (P < 0.0001).

Data are included in Daviskas 1999 and Daviskas 2008 to provide further clarity in relation to mucus clearance in the right central, right intermediate and right peripheral areas. These findings need to be interpreted with caution in view of the very small sample, however - in addition to benefits in clearance in the whole right lung and intermediate and central areas - they do indicate that a significant advantage for mannitol in mucus clearance in the peripheral region over longer time periods (from 30 to 75 minutes after the start of the dose). At this stage, in view of the paucity of data, these impressions are in need of further clarification with larger studies and serve largely as hypotheses for further investigations.

Hypertonic saline versus isotonic saline

The quality of evidence for the comparison hypertonic saline versus isotonic saline was moderate (Summary of findings 2). Four studies (N = 113) compared hypertonic saline versus isotonic saline (Bradley 2011, N = 19, Kellett 2005 N = 24, Kellet 2011, N = 30, Nicolson 2012, N = 40). Three of the four studies Bradley 2011, Kellett 2005, Kellet 2011 were cross-over design and the data were not reported in a format that enabled us to include them in meta-analyses. Nicolson 2012 was a 12-month parallel group trial.

Exacerbations

One cross-over study (Kellet 2011) demonstrated a significant difference between hypertonic saline versus isotonic saline with respect to this outcome (P < 0.05). The rate of exacerbations in the hypertonic saline group (2.14 exacerbations/year) was less than half the rate in the isotonic saline group (4.85 exacerbations/year).

However this finding was not supported by Nicolson 2012, where there were no significant differences between hypertonic saline versus isotonic saline on exacerbations (over 12 months), exacerbations requiring antibiotics, number of exacerbation days or number of exacerbation days requiring antibiotics. Median number of exacerbations (inter-quartile range (IQR)) were 1 (zero to 4) for isotonic saline and 3 (zero to 6) for hypertonic saline (P = 0.24).

Hospital admissions

In Nicolson 2012 there was no significant difference between hypertonic saline versus isotonic saline with respect to hospital admissions (P = 0.34) or hospital days (P = 0.36). Four (10%) participants had hospital admissions, three in the isotonic saline group were in hospital for three, five and 61 days respectively and one participant in the hypertonic saline group was in hospital for 68 days.

Health-related quality of life

In Nicolson 2012 there was no significant difference between hypertonic saline versus isotonic saline with respect to SRGQ symptoms scores at three, six or 12 months Analysis 2.1; SGRQ activity scores at three, six or 12 months Analysis 2.2; or SGRQ impact scores at three, six or 12 months Analysis 2.3. Data for total SGRQ scores were not reported. In the same study there was no significant difference between hypertonic saline versus isotonic saline with respect to LCQ physical scores at three, six or 12 months Analysis 2.4; there was no significant difference between hypertonic saline versus isotonic saline with respect to LCQ psychological scores at six and 12 months, although there was a modest significant difference at three months (MD 0.90; 95% CI 0.12 to 1.68). At this time point however in the isotonic saline group the LCQ psychological domain score improved from 3.9 to 5.0 (a 1.1 unit change) and the hypertonic saline group score improved from 4.7 to 5.9 (1.2 unit change). The difference in change score (0.1 units) is therefore neither statistically nor clinically significant. Analysis 2.5. There were no differences between hypertonic saline versus isotonic saline with respect to LCQ social scores at three six or 12 months Analysis 2.6. Data for total LCQ scores were not reported.

However, an additional comparison between hypertonic saline versus isotonic saline is provided in a small cross-over study with 19 participants by Bradley 2011. Significant benefits in favour of hypertonic saline were observed in the physical domain of the LCQ ( -0.8 (SD 0.9); P = 0.01) and the respiratory symptoms domain of the Quality of Life Questionnaire-Bronchiectasis, QOL-B ( -11.6 (SD 17.7); P = 0.03) .

Kellet 2011 reported that significant changes were seen for St George’s Respiratory Quality of Life (global score) and in terms of the subscales of symptoms and impact, but did not reach statistical significance for the activity subscale. Graphs, but no data, were provided.

Mortality

No deaths were reported.

Lung function

Kellet 2011 reported a significant advantage for hypertonic saline versus isotonic saline in change from baseline in FEV1 % (MD 13.30; P < 0.01) and FVC (MD 10.51; P < 0.01).

Bradley 2011 report that hypertonic saline had an effect on FEV1; the effect size being an unspecified value between 0.01 and 0.14.

In terms of FEV (L), there was no significant benefit for hypertonic saline versus isotonic saline at three, six or twelve months in Nicolson 2012 (Analysis 2.7); similarly, there were no differences between the two groups at these time points on FVC Analysis 2.8 or FEF 25-75 Analysis 2.9. However there was a benefit for hypertonic saline versus isotonic saline in FVC % change from pre-treatment baseline in Kellet 2011 (MD 10.51; 95% CI 2.36 to 18.66 Analysis 2.10).

In Kellett 2005 there was no significant differences in FEV1 (P = 0.12) or FVC (P = 0.23) between hypertonic saline versus isotonic saline; these data were reported in a format that did not provide an opportunity for aggregation with other studies. The median, (inter-quartile) values were: FEV1 2.0, (1.25 to 2.40) hypertonic saline versus 1.85, (1.36 to 2.20) isotonic saline, and FVC 2.50, (1.79 to 3.08) hypertonic saline versus 2.55, (1.91 to 2.94) isotonic saline.

Sputum weight/volume

In Kellett 2005 sputum weights were significantly different between the isotonic saline median 3.17 g, (IQR 1.45 to 6.25) and hypertonic saline median 5.3 g, (IQR 2.97 to 9.33) arms. Ease of expectoration as measured on a 10 cm VAS scale was significantly lower with hypertonic saline, median 2.40 cm, (IQR 1.43 to 5.40) than isotonic saline median 5.20 cm, (IQR 2.75 to 8.38).

In Nicolson 2012 participants in both the hypertonic and isotonic saline groups improved over the course of the 12-month trial with only 15% having a positive sputum culture, in either group, at that point; these data were reported in a format that did not provide an opportunity for aggregation with other studies.

Bradley 2011 report that hypertonic saline had an effect on sputum; the effect size being an unspecified value between 0.01 and 0.14 .

Kellet 2011 reported "improvements in sputum viscosity and ease of expectoration (visual analogue scale)" in the conference abstract and abstract of the published trial report but gave no data.

Antibiotic use

In a cross-over study Kellet 2011 has demonstrated a significant difference between hypertonic saline versus isotonic saline with respect to this outcome (P < 0.05). The level of antibiotic use in the hypertonic saline group (2.4 courses per year) was less than half the level in the isotonic saline group (5.4 courses per year).

Nicolson 2012 reported no significant difference (P = 0.77) in the number of exacerbation days requiring antibiotics.

Cough

In Nicolson 2012, there were no significant differences between groups in cough frequency reported by VAS at three (P = 0.39), six (P = 0.17) or 12 months (P = 0.60).

Adverse events

There were reportedly no significant differences in the occurrence of adverse events between people in the hypertonic saline versus isotonic saline groups in Bradley 2011, but no data were reported.

Nicolson 2012 reported three adverse events in the hypertonic saline group: chest tightness during inhalation which resolved with treatment of an underlying acute exacerbation in one participant. An episode of hypertension was reported by another participant and this was resolved without further intervention, and was unrelated to the intervention in the view of the patient's physician. Both patients remained in the trial without incident to the completion of the study. An episode of rapid atrial fibrillation requiring attendance at the emergency department was experienced by a third participant and the saline inhalation was ceased prior to the three-month assessment. There were no adverse events in the isotonic saline group. The difference between the two groups was not significant (OR 8.20; 95% CI 0.40 to 169.90 Analysis 2.10).

Discussion

Summary of main results

In one relatively large trial, mannitol significantly extended the time to first bronchiectasis exacerbation from just over four months to around five and a half months. This translated into a trend towards a reduction in annual exacerbations with mannitol but did not reach conventional significance. The number of days patients spent taking antibiotics for exacerbations of bronchiectasis was also significantly reduced, though overall use of courses of antibiotics was probably unchanged.

The effect of mannitol on sputum volume is unclear. Bilton 2013 reported a reduction in sputum volume in the placebo group, whereas sputum production was maintained in the mannitol group. Bennoor 2012 reported no difference in sputum production between the mannitol and placebo arms. The clinical implications of sputum production differ in different patients. A reduction in expectorated sputum volume may reflect either a reduction in chronic bronchial sepsis or increased infection due to retained secretions. For some, but not all, patients sputum expectoration is a troubling symptom and changes in sputum expectoration need interpreting in this light.

Despite changes in the number of exacerbations, respiratory related quality of life was not clinically improved in those taking mannitol. These were, by and large, patients with long standing chronic illness and quality of life was not systematically recorded during exacerbations so this may be due to the timing of quality of life assessments. Alternatively it may be that such patients have adapted their lifestyles so that chest exacerbations do not dramatically impair their quality of life. Quality of life was generally assessed by means of the SGRQ. This includes questions on cough. The relationship between hyperosmolar therapy and cough may be complex. Hyperosmolar agents are directly irritant to the lungs and thus stimulate coughing in the short term, worsening some symptom elements of the SGRQ. However, they may also change the characteristics of the cough. Consequent thinner mucus might make cough less painful or embarrassing and thus improve these elements of the SGRQ and improved mucus clearance may reduce overall cough in the medium or long term. It is possible that quality of life is being modified by these therapies without overall improvement. Individual SGRQ question level data were not available.

On the whole, no dramatic effects of mannitol on lung function were noted. Potentially mannitol could have improved spirometry by reducing the amount of mucus plugging in the small airways (and thus increasing the effective small airway diameter). Mannitol can trigger bronchoconstriction so a detrimental effect on spirometry was also possible. With the exception of NCT00669331, patients in the included trials were selected as having relatively normal spirometry before inclusion as well as not having a large reduction in spirometry during a mannitol provocation test so large changes in spirometry would be surprising in these groups. A significant effect (good or bad) in those with abnormal baseline spirometry is not excluded by these studies. Spirometric data have not yet been reported for NCT00669331, though significant bronchospasm was not observed as an adverse outcome.

Inhaled mannitol did not appear to cause many adverse events. Lung function did not decrease as might have been expected and Bilton 2013 reported no significant difference in the rates of adverse events between mannitol and placebo. Bennoor 2012 described a number of symptoms in patients taking mannitol but severity and causality were unclear.

Two very small studies (Daviskas 1999; Daviskas 2008) looked at mannitol against no treatment. In one (Daviskas 2008), lung function was reported as unchanged, though the methodology used (radioisotope deposition not spirometry) was unusual. Both studies suggested a dose response effect with greater radioisotope evidence of sputum movement with increasing doses of mannitol. The small scale, preliminary nature of these studies should be noted, along with their use of a measurement technique not widely used in clinical practice.

Two of the four included studies that investigated the use of nebulised hypertonic versus isotonic saline (Kellet 2011; Nicolson 2012) contributed data on exacerbation rates, antibiotic use and/or hospitalisations. The studies of Bradley 2011 and Kellett 2005 were too short to assess an effect on exacerbation frequency, hospitalisations or antibiotic use, the former with a four-week treatment period, and the latter comparing the effect of four different single-dose treatment regimens.

The Nicolson 2012 and Kellet 2011 studies provide contrasting results. Whilst Nicholson found that treatment with both isotonic and hypertonic saline resulted in a very large drop in exacerbation frequency over the year of the study, there was no difference between the treatments. Likewise, there were no differences in number or length of hospital admissions. In contrast Kellet 2011 reported a statistically significant and clinically important reduction in exacerbations in the hypertonic saline group (annualised exacerbation rate 2.1 for hypertonic saline, 4.9 for isotonic saline), with a comparable reduction in use of antibiotics (annualised rate 2.4) versus isotonic saline (5.4). As noted below, these differences in impact between the two studies were also seen on many other shared outcomes, and likely to be due at least in part to significant methodological differences between the studies.

The strong impact of both treatments on outcomes in the Nicholson study merits further consideration. The pre-study exacerbation rate data were collected retrospectively and so relied presumably on patient recall. To patients the usual definition of an exacerbation would likely include the requirement for antibiotics and/or hospital admission, and probably last for more than a day. It is not possible therefore to directly relate this rate to the multiple definitions applied prospectively during the study period (i.e. the retrospective baseline definition of exacerbation was not the same as any of the prospective definitions). Nevertheless, the reported rates certainly imply a large benefit of study participation in both groups, also reflected in the very low requirement for antibiotics (median 0.5 courses per year in the isotonic saline group; 1.0 per year in the hypertonic saline group). The median “pre-study” exacerbation rate was five per annum, but during the study there were only a median one to two “exacerbation days requiring antibiotics” for the group as a whole. This fairly dramatic finding (and the similar findings in this study relating to quality of life as discussed below) could be explained by the benefit that could be obtained by participating in a study which provides relatively frequent medical review, and possibly more importantly the potential for increased self-awareness related to the disease process. Volunteers were all nebulising twice daily and filling in daily diaries reporting their symptoms, and 85% were using chest clearance techniques. It is also possible that retrospective recall overestimates exacerbation rate. The third potential contributing issue (applicable to all studies here) is that isotonic saline in fact is not really a placebo in this context and may have benefits by increasing airway hydration, with further benefit possibly conferred by the respiratory manoeuvres that participants were instructed to undertake during nebulisation. The Nicholson paper also did not select patients on the basis of features that could theoretically confer a differential benefit of hypertonic saline, such as frequent exacerbators or those with large volumes of sticky sputum. It is therefore possible that any potential signal due to “between-treatment” difference was overwhelmed by the benefit of study participation. One interesting finding of the Nicholson study was that at the end of the study far more patients in the hypertonic saline group chose to stay on their study medication long term than in the isotonic saline group. The exact numbers are not clear; it is stated in the results that six chose to remain on isotonic saline, whereas 23 stayed on hypertonic saline, but this latter is likely to be a typographical error as only 20 participants used hypertonic saline during the trial, and in the discussion section “nearly three-quarters” of the group were said to have remained on this treatment.

The Kellet 2011 study, which demonstrated significant benefits for hypertonic saline, is described as single blind, as the study medicines were presented in specific ampoules identifiable to the study team. In addition the adequacy of patient-blinding was not checked (as it was in the Nicholson study), and it seems very likely that volunteers could taste a clear difference between 7% and 0.9% saline, especially as all patients had a test dose of hypertonic saline at the beginning of the study, and with the cross-over design all patients received both treatments allowing direct comparison during the study. Another difference between the Nicolson 2012 and Kellet 2011 studies was that the former gave nebulised saline twice daily, the latter once only. The definition of exacerbation was different in the Kellet 2011 study and required contact with a primary or secondary care physician for a deterioration in bronchiectasis-related symptoms. Again, retrospective baseline exacerbation rate data were used, but not required for the primary treatment comparison because of the cross-over design. It is also worth noting that this study excluded patients colonised with Pseudomonas aeruginosa, for unexplained reasons. Such patients represent a subgroup in bronchiectasis with poor outcomes Evans 1996; Martinez 2007, and who might potentially benefit from mucolytic therapy, so their exclusion here is difficult to explain.

Nicholson and colleagues assessed health-related quality of life using the general respiratory SGRQ and cough-specific LCQ. As for exacerbations, a large clinically relevant improvement in all domains of both measures occurred during the study with both strengths of saline, with no difference between treatment groups. On the other hand, Kellet 2011 reported statistically and clinically significant improvements in SGRQ total scores and symptom and impact domains. Potential explanations for this are discussed above.

Bradley 2011 reported a statistically and clinically relevant improvement in the physical domain of the LCQ with hypertonic versus isotonic saline, as well as in the respiratory symptoms domain of the QOL-B, although this latter quality of life questionnaire has not been validated and the minimal important difference is not known. However, multiple comparisons were performed in a small number of participants in this feasibility study, with no adjustment for multiple testing, hence the probability of these being false-positive findings is not known.

In the Nicolson 2012 study, mean lung function was very good at baseline, for example mean FEV1 was 80% predicted in the isotonic saline group and 85% predicted in the hypertonic saline group, which may have lead to a ceiling effect explaining the minimal, although statistically significant, 90 mL improvement in both groups. No differences were seen between groups. Again a contrast is provided by the Kellet 2011 study, which investigated the use of hypertonic saline in bronchiectasis patients with more severe disease than Nicholson at least in terms of lung function, with mean FEV1 66% predicted. This paper reports a significant benefit of hypertonic over isotonic saline, with a 15% improvement in FEV1 following three months treatment with the former compared to a 2% improvement with the latter, with a similar-sized effect difference on FVC also.

The limited data available from the Bradley 2011 abstract do not enable the size of effect on FEV1 to be calculated, nor whether the change was statistically significant, whilst Kellett 2005 reported no improvement in lung function after a single dose of hypertonic saline versus isotonic saline.

No effects on sputum weight or volume were reported in the Nicholson paper, although the frequency of positive sputum cultures in both groups improved over the year of the study, with no between-group difference. Kellet 2011 reported unquantified benefits of hypertonic over isotonic saline on sputum viscosity and ease of expectoration in the abstract of their article, but no further results are given in the main results section. No meaningful interpretation of the effect of hypertonic saline on sputum is possible from the limited data available from the Bradley 2011 abstract.

Kellett 2005 looked at the effect of single dose isotonic or hypertonic saline in combination with chest clearance (active cycle of breathing technique) in patients who specifically reported “thick, sticky” sputum, and expectorating small volumes (less than 10 g per day), but who nevertheless had required at least one antibiotic course over the preceding six months. The outcomes of this study were particularly focused on the acute effect of first-dose of saline solutions on sputum parameters, and hence, whilst they provide useful data in this regard, they cannot be extrapolated to longer-term effects. In addition, the study’s objective was to perform a four-way comparison (the other two treatment schedules were chest clearance alone and nebulised bronchodilator plus chest clearance) and hence the comparison of outcomes between hypertonic and isotonic saline was performed post-hoc, with the variables grouped in multiple ways, and again with no adjustment for multiple-testing. With these qualifications, single dose hypertonic saline use was associated with increase sputum weight, decreased viscosity, and improved ease of expectoration when compared with isotonic saline.

In the Nicolson 2012 study, two patients reported side effects considered potentially due to hypertonic saline; one with transient chest tightness, which was relieved after treatment for an exacerbation, and another with atrial fibrillation which resulted in withdrawal from the study (NB all participants inhaled 200 mcg salbutamol prior to saline nebulisation). There were no adverse reactions noted in the isotonic saline group. Bradley 2011 reported no difference in side effects between the hypertonic and isotonic saline treatment groups. Two of the 32 patients screened for the Kellet 2011 study were excluded because of bronchial hyperresponsiveness to the test dose of hypertonic saline. Adverse events during the study period are not reported.

Overall completeness and applicability of evidence

Whilst the five studies (combined N = 883) comparing mannitol versus placebo provide more clarity in this area than was available in the previous version of this review, the opportunities for the aggregation of data were limited. Data comparing hypertonic saline with isotonic saline were, by comparison, relatively sparse and there were even more obvious limitations for meta-analyses. Over the last seven years since the last update of this review (Wills 2006), a number of new studies have been published, however one very recent large study comparing mannitol with placebo over a 12-month period NCT00669331 has yet to be published in full. The data available from that study at the time of completing this update are limited to those presented at a recent international conference.

Quality of the evidence

In terms of random sequence generation only two trials (Kellett 2005; Nicolson 2012) were considered to be at low risk of bias (selection bias); both trials compared hypertonic saline with isotonic saline, had adequate randomisation procedures, and all five trials comparing mannitol versus placebo (or no treatment) had issues with their randomisation procedure, which undermine our confidence in the results. The risk of bias for the remaining nine trials was evaluated as unclear as details of the random sequence generation were not described in the trial report. In terms of the blinding of participants and personnel, only three trials (Bilton 2013; Kellett 2005; Nicolson 2012) were judged to be at low risk of bias (performance bias), and only one of those trials compared mannitol with placebo (or no treatment).

Potential biases in the review process

The Cochrane Airways Group provides exceptional support in the identification of potentially relevant trials; however, concerns regarding study selection bias or publication bias are an issue in this review as with any other. We are mindful that a failure to identify unpublished trials may lead to an incomplete estimation of hyperosmolar agents. Even though an extensive search of the published literature, without language restrictions, for potentially relevant clinical trials was undertaken using a systematic search strategy to minimise the likelihood of bias, we acknowledge that additional unidentified trials may still be undetected. We note that improvements in the standardisation of reporting would facilitate the opportunities to draw comparisons among trials.

Agreements and disagreements with other studies or reviews

The availability of data in this update is considerable when compared to the previous version of this review Wills 2006 and the number of included studies has increased from two to 11. Whereas the previous review highlighted the paucity of evidence in this area, we now note that five studies (combined N = 883) compared mannitol versus placebo, enabling us to assess the clinical benefits of hyperosmolar agents over periods up to 12 months. Whilst a paucity of evidence remains in relation to some important outcomes, and we note that there is still a need for more information from randomised studies in relation to exacerbations and hospitalisations in particular, the additional nine trials available to this version of the review have provided considerably more clarity than was available hitherto in this important area.

Authors' conclusions

Implications for practice

Inhaled mannitol increases time to infective exacerbation in patients with bronchiectasis, without improving respiratory quality of life. It has a role in reducing exacerbations and antibiotic use in those with normal or mild to moderately impaired spirometry. Its safety has not been studied in patients with severe lung function impairment.

For hypertonic saline, the two full studies published to date that describe chronic dosing and hence could help guide clinical practice have produced conflicting results, which is likely due to their different patient groups and study designs. The magnitude of the effect size in the Kellet 2011 study on important endpoints including exacerbation rates, antibiotic use, lung function and quality of life suggest that there might be a clinically relevant effect of hypertonic saline in similar patients, despite the methodological concerns arising mainly from the potentially inadequate blinding. On the other hand, the results of Nicolson 2012 suggest that hypertonic saline does not confer any clear benefit over isotonic saline when it is introduced to treatment in those with physiologically mild disease.

Future larger scale studies should focus on targeting those with established disease and poor lung function, whilst addressing the potential bias associate with inadequate blinding.

Implications for research

Larger, more long-term studies into the use of inhaled hyperosmolar agents in bronchiectasis are now starting to appear. Future studies should focus on those with impaired spirometry and consider using health economy important measures such as hospitalisation rate and antibiotic usage. Long-term studies of hyperosmolar agents in patients with early bronchiectasis and relatively normal spirometry are also justified to see if these agents can prevent progression to impaired lung function.

Acknowledgements

The authors of the original version of this review acknowledged the support the Cochrane Airways Group staff, particularly Steve Milan, Toby Lasserson, Liz Arnold and Karen Blackhall for their help with the software and in trial identification, and their gratitude to Prof E Daviskas and Dr. S Anderson of the Royal Prince Alfred Hospital and Prof. P Cole of the Brompton Hospital for their personal communications.

In the 2013 update we would particularly like to acknowledge the contribution of Peter J Wills and Michael Greenstone, authors on the original version of this review and the excellent support and assistance received from Emma Welsh, Liz Stovold and Emma Jackson of the Cochrane Airways Review group, together with the greatly appreciated guidance received from Chris Cates (Cochrane Airways Review Group Co-ordinating Editor). We are indebted to Dr D Bilton for additional information. The support provided by librarians Judith Scammel, Jane Appleton and Hilary Garrett at St Georges University London is also greatly appreciated.

Data and analyses

Download statistical data

Comparison 1. Inhaled mannitol versus placebo
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Change from baseline in health-related quality of life (SGRQ). Total3 Mean Difference (Fixed, 95% CI)-2.05 [-3.69, -0.40]
1.1 2 Weeks1 Mean Difference (Fixed, 95% CI)-6.0 [-13.59, 1.59]
1.2 12 Weeks1 Mean Difference (Fixed, 95% CI)-1.27 [-3.69, 1.15]
1.3 12 months1 Mean Difference (Fixed, 95% CI)-2.4 [-4.76, -0.04]
2 Change from baseline in health-related quality of life (SGRQ). Symptoms2377Mean Difference (Fixed, 95% CI)-1.42 [-5.12, 2.28]
2.1 2 Weeks134Mean Difference (Fixed, 95% CI)-7.1 [-18.53, 4.33]
2.2 12 Weeks1343Mean Difference (Fixed, 95% CI)-0.76 [-4.67, 3.15]
3 Change from baseline in health-related quality of life (SGRQ). Activity2 Mean Difference (Fixed, 95% CI)-0.84 [-3.95, 2.28]
3.1 2 Weeks1 Mean Difference (Fixed, 95% CI)-1.2 [-7.53, 5.13]
3.2 12 Weeks1 Mean Difference (Fixed, 95% CI)-0.72 [-4.30, 2.86]
4 Change from baseline in health-related quality of life (SGRQ). Impact2 Mean Difference (Fixed, 95% CI)-3.16 [-5.77, -0.55]
4.1 2 Weeks1 Mean Difference (Fixed, 95% CI)-9.8 [-18.66, -0.94]
4.2 12 Weeks1 Mean Difference (Fixed, 95% CI)-2.53 [-5.26, 0.20]
5 Leicester Cough Questionnaire score after 12 weeks1 Mean Difference (IV, Fixed, 95% CI)Totals not selected
6 Bronchiectasis Symptoms Questionnaire score at week 121 Mean Difference (IV, Fixed, 95% CI)Totals not selected
7 Change in FEV1 (% predicted)1 % predicted FEV1 (Fixed, 95% CI)Totals not selected
8 FEV1 (L) at 12 weeks1 Mean Difference (IV, Fixed, 95% CI)Totals not selected
9 Adverse events2804Odds Ratio (M-H, Fixed, 95% CI)0.96 [0.61, 1.51]
10 FEF25-75(L/s) after 12 weeks1 Mean Difference (IV, Fixed, 95% CI)Totals not selected
11 Change in 24 hr sputum weight from baseline1 Mean Difference (IV, Fixed, 95% CI)Totals not selected
12 Antibiotic use1 Odds Ratio (M-H, Fixed, 95% CI)Totals not selected
13 Incremental Shuttle Walk (metres) at week 121 Mean Difference (IV, Fixed, 95% CI)Totals not selected
14 FVC (L) at 12 weeks1 Mean Difference (IV, Fixed, 95% CI)Totals not selected
15 Serious adverse events2804Odds Ratio (M-H, Fixed, 95% CI)0.79 [0.52, 1.19]
Analysis 1.1.

Comparison 1 Inhaled mannitol versus placebo, Outcome 1 Change from baseline in health-related quality of life (SGRQ). Total.

Analysis 1.2.

Comparison 1 Inhaled mannitol versus placebo, Outcome 2 Change from baseline in health-related quality of life (SGRQ). Symptoms.

Analysis 1.3.

Comparison 1 Inhaled mannitol versus placebo, Outcome 3 Change from baseline in health-related quality of life (SGRQ). Activity.

Analysis 1.4.

Comparison 1 Inhaled mannitol versus placebo, Outcome 4 Change from baseline in health-related quality of life (SGRQ). Impact.

Analysis 1.5.

Comparison 1 Inhaled mannitol versus placebo, Outcome 5 Leicester Cough Questionnaire score after 12 weeks.

Analysis 1.6.

Comparison 1 Inhaled mannitol versus placebo, Outcome 6 Bronchiectasis Symptoms Questionnaire score at week 12.

Analysis 1.7.

Comparison 1 Inhaled mannitol versus placebo, Outcome 7 Change in FEV1 (% predicted).

Analysis 1.8.

Comparison 1 Inhaled mannitol versus placebo, Outcome 8 FEV1 (L) at 12 weeks.

Analysis 1.9.

Comparison 1 Inhaled mannitol versus placebo, Outcome 9 Adverse events.

Analysis 1.10.

Comparison 1 Inhaled mannitol versus placebo, Outcome 10 FEF25-75(L/s) after 12 weeks.

Analysis 1.11.

Comparison 1 Inhaled mannitol versus placebo, Outcome 11 Change in 24 hr sputum weight from baseline.

Analysis 1.12.

Comparison 1 Inhaled mannitol versus placebo, Outcome 12 Antibiotic use.

Analysis 1.13.

Comparison 1 Inhaled mannitol versus placebo, Outcome 13 Incremental Shuttle Walk (metres) at week 12.

Analysis 1.14.

Comparison 1 Inhaled mannitol versus placebo, Outcome 14 FVC (L) at 12 weeks.

Analysis 1.15.

Comparison 1 Inhaled mannitol versus placebo, Outcome 15 Serious adverse events.

Comparison 2. Hypertonic saline versus isotonic saline
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 SGRQ Symptom1 Mean Difference (IV, Fixed, 95% CI)Subtotals only
1.1 3 months140Mean Difference (IV, Fixed, 95% CI)-8.5 [-20.48, 3.48]
1.2 6 months140Mean Difference (IV, Fixed, 95% CI)6.40 [-5.93, 18.73]
1.3 12 months140Mean Difference (IV, Fixed, 95% CI)1.60 [-12.53, 15.73]
2 SGRQ Activity1 Mean Difference (IV, Fixed, 95% CI)Subtotals only
2.1 3 months140Mean Difference (IV, Fixed, 95% CI)4.40 [-11.49, 20.29]
2.2 6 months140Mean Difference (IV, Fixed, 95% CI)2.60 [-11.86, 17.06]
2.3 12 months140Mean Difference (IV, Fixed, 95% CI)2.70 [-13.11, 18.51]
3 SGRQ Impact1 Mean Difference (IV, Fixed, 95% CI)Subtotals only
3.1 3 months140Mean Difference (IV, Fixed, 95% CI)-0.40 [-12.03, 11.23]
3.2 6 months140Mean Difference (IV, Fixed, 95% CI)-0.80 [-9.71, 8.11]
3.3 12 months140Mean Difference (IV, Fixed, 95% CI)4.40 [-7.50, 16.30]
4 Leicester Cough Questionnaire Physical1 Mean Difference (IV, Fixed, 95% CI)Subtotals only
4.1 3 months140Mean Difference (IV, Fixed, 95% CI)0.40 [-0.22, 1.02]
4.2 6 months140Mean Difference (IV, Fixed, 95% CI)-0.10 [-0.88, 0.68]
4.3 12 months140Mean Difference (IV, Fixed, 95% CI)-0.20 [-0.95, 0.55]
5 Leicester Cough Questionnaire Psychological1 Mean Difference (IV, Fixed, 95% CI)Subtotals only
5.1 3 months140Mean Difference (IV, Fixed, 95% CI)0.90 [0.12, 1.68]
5.2 6 months140Mean Difference (IV, Fixed, 95% CI)-0.10 [-0.85, 0.65]
5.3 12 months140Mean Difference (IV, Fixed, 95% CI)0.20 [-0.63, 1.03]
6 Leicester Cough Questionnaire Social1 Mean Difference (IV, Fixed, 95% CI)Subtotals only
6.1 3 months140Mean Difference (IV, Fixed, 95% CI)0.5 [-0.34, 1.34]
6.2 6 months140Mean Difference (IV, Fixed, 95% CI)-0.10 [-0.82, 0.62]
6.3 12 months140Mean Difference (IV, Fixed, 95% CI)-0.10 [-0.85, 0.65]
7 FEV1 (L)1 Mean Difference (IV, Fixed, 95% CI)Subtotals only
7.1 3 months140Mean Difference (IV, Fixed, 95% CI)0.13 [-0.43, 0.69]
7.2 6 months140Mean Difference (IV, Fixed, 95% CI)0.17 [-0.40, 0.74]
7.3 12 months140Mean Difference (IV, Fixed, 95% CI)0.19 [-0.37, 0.75]
8 FVC1 Mean Difference (IV, Fixed, 95% CI)Subtotals only
8.1 3 months140Mean Difference (IV, Fixed, 95% CI)0.05 [-0.63, 0.73]
8.2 6 months140Mean Difference (IV, Fixed, 95% CI)0.10 [-0.57, 0.77]
8.3 12 months140Mean Difference (IV, Fixed, 95% CI)0.11 [-0.57, 0.79]
9 FEF25-751 Mean Difference (IV, Fixed, 95% CI)Subtotals only
9.1 3 months140Mean Difference (IV, Fixed, 95% CI)0.12 [-0.52, 0.76]
9.2 6 months140Mean Difference (IV, Fixed, 95% CI)0.16 [-0.57, 0.89]
9.3 12 months140Mean Difference (IV, Fixed, 95% CI)0.29 [-0.35, 0.93]
10 Adverse events1 Odds Ratio (M-H, Fixed, 95% CI)Totals not selected
Analysis 2.1.

Comparison 2 Hypertonic saline versus isotonic saline, Outcome 1 SGRQ Symptom.

Analysis 2.2.

Comparison 2 Hypertonic saline versus isotonic saline, Outcome 2 SGRQ Activity.

Analysis 2.3.

Comparison 2 Hypertonic saline versus isotonic saline, Outcome 3 SGRQ Impact.

Analysis 2.4.

Comparison 2 Hypertonic saline versus isotonic saline, Outcome 4 Leicester Cough Questionnaire Physical.

Analysis 2.5.

Comparison 2 Hypertonic saline versus isotonic saline, Outcome 5 Leicester Cough Questionnaire Psychological.

Analysis 2.6.

Comparison 2 Hypertonic saline versus isotonic saline, Outcome 6 Leicester Cough Questionnaire Social.

Analysis 2.7.

Comparison 2 Hypertonic saline versus isotonic saline, Outcome 7 FEV1 (L).

Analysis 2.8.

Comparison 2 Hypertonic saline versus isotonic saline, Outcome 8 FVC.

Analysis 2.9.

Comparison 2 Hypertonic saline versus isotonic saline, Outcome 9 FEF25-75.

Analysis 2.10.

Comparison 2 Hypertonic saline versus isotonic saline, Outcome 10 Adverse events.

Appendices

Appendix 1. Sources and search methods for the Cochrane Airways Group Specialised Register (CAGR)

Electronic searches: core databases

Database Frequency of search
CENTRAL (The Cochrane Library)Monthly
MEDLINE (Ovid)Weekly
EMBASE (Ovid)Weekly
PsycINFO (Ovid)Monthly
CINAHL (EBSCO)Monthly
AMED (EBSCO)Monthly

 

Handsearches: core respiratory conference abstracts

Conference Years searched
American Academy of Allergy, Asthma and Immunology (AAAAI)2001 onwards
American Thoracic Society (ATS)2001 onwards
Asia Pacific Society of Respirology (APSR)2004 onwards
British Thoracic Society Winter Meeting (BTS)2000 onwards
Chest Meeting2003 onwards
European Respiratory Society (ERS)1992, 1994, 2000 onwards
International Primary Care Respiratory Group Congress (IPCRG)2002 onwards
Thoracic Society of Australia and New Zealand (TSANZ)1999 onwards

 

MEDLINE search strategy used to identify trials for the CAGR

Bronchiectasis search

1. exp Bronchiectasis/

2. bronchiect$.mp.

3. bronchoect$.mp.

4. kartagener$.mp.

5. (ciliary adj3 dyskinesia).mp.

6. (bronchial$ adj3 dilat$).mp.

7. or/1-6

Filter to identify RCTs

1. exp "clinical trial [publication type]"/

2. (randomised or randomised).ab,ti.

3. placebo.ab,ti.

4. dt.fs.

5. randomly.ab,ti.

6. trial.ab,ti.

7. groups.ab,ti.

8. or/1-7

9. Animals/

10. Humans/

11. 9 not (9 and 10)

12. 8 not 11

The MEDLINE strategy and RCT filter are adapted to identify trials in other electronic databases

Appendix 2. Search strategy for the Cochrane Airways Group Register

2013 update

#1 BRONCH:MISC1

#2 MeSH DESCRIPTOR Bronchiectasis Explode All

#3 bronchiect*

#4 #1 or #2 or #3

#5 MeSH DESCRIPTOR Mannitol Explode 1 2

#6 MeSH DESCRIPTOR Saline Solution, Hypertonic

#7 MeSH DESCRIPTOR Sodium Chloride

#8 "sodium chloride"

#9 NaCl

#10 mannitol

#11 osmolality

#12 osmolarity

#13 hyperosmolar

#14 #5 or #6 or #7 or #8 or #9 or #10 or #11 or #12 or #13

#15 #4 and #14

[In search line #1, MISC1 denotes the field in which the reference has been coded for condition, in this case, bronchiectasis]

Previous versions

"sodium chloride" OR NaCl OR mannitol OR osmolality OR osmolarity OR hyperosmolar

[Limited to bronchiectasis records]

What's new

DateEventDescription
2 April 2014New search has been performedNew literature search run.
2 April 2014New citation required and conclusions have changed

Review updated. The previous version of this review included two trials (28 participants), and now includes 11 trials (1021) participants.

'Summary of findings' tables added, review text redrafted, risk of bias updated.

History

Protocol first published: Issue 2, 2000
Review first published: Issue 1, 2001

DateEventDescription
28 July 2008AmendedConverted to new review format.
17 January 2006New citation required and conclusions have changedSubstantive amendment

Contributions of authors

In the original version of this review Wills 2006, the contribution of authors was as follows: PW: Initiation of the protocol, assessment of search results, quality assessment, data extraction, interpretation/discussion. MG: Editorial support, interpretation/discussion.

In the 2013 update SF, IC, KS and SM updated the background. KS and IC independently selected studies for inclusion. AH and IC independently extracted data. SM entered the data and AH checked data entry. The results, 'Risk of bias' and 'Summary of findings' sections were completed by SM, IC and AH. SM updated the methods section. AH, SF, IC, KS and SM completed the Discussion and Conclusions.

Chris Cates was the contact editor for this review and critically commented on it.

Declarations of interest

None known.

Sources of support

Internal sources

  • Hillingdon Hospital, UK.

  • NHS Research and Development, UK.

  • National Institute for Health Research, UK.

External sources

  • No sources of support supplied

Differences between protocol and review

In the 2013 update of this review we defined primary and secondary outcomes. We brought the review up to date with current methodological standards consistent with Higgins 2011. We split the outcomes into primary and secondary.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Bennoor 2012

MethodsSingle-blind, parallel group randomised controlled trial.
Participants50 patients with stable bronchiectasis (25 on mannitol and 25 on placebo). Details of severity not included in trial report.
Interventions400 mg mannitol (10 capsules) inhaled via rotahaler daily for twelve days versus 'similar amount' of empty capsules via rotahaler for 12 consecutive days.
Outcomes24 hour sputum volume, dyspnoea grade, FEV1, FVC, FEF25-75, SGRQ and adverse effects. Significance tests of change from baseline only. Data recorded at days 6 and 12.
Notes

12-day trial

Funding: Details not included in abstract.

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskProcess of randomisation not described 
Allocation concealment (selection bias)Unclear riskDetails not included in trial report
Blinding of participants and personnel (performance bias)
All outcomes
High riskSingle-blind trial
Blinding of outcome assessment (detection bias)
All outcomes
High riskSingle-blind trial
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskBy implication all randomised patients completed, but this is not explicitly stated
Selective reporting (reporting bias)Unclear riskFEV1 results not reported. Adverse events only reported in mannitol arm
Other biasUnclear riskNo other indication of bias

Bilton 2013

MethodsRandomised, double-blind, parallel-arm, placebo-controlled, multicentre clinical trial, 22 sites: UK, NZ, Australia
Participants

362 participants randomised (343 analysed) with bronchiectasis confirmed by high-resolution CT (HRCT) scan.

241 (231 analysed) in mannitol arm and 121 (112 analysed) in placebo.

Mean age 61.2 (10.4) mannitol and 62.3 (8.9) placebo.

Male 79 (34.2%) in mannitol and 40 (35.7%) in placebo.

Baseline FEV1 (L) 1.94 (0.55) in mannitol and 1.92 (0.56) in placebo. Baseline FEV 1 % predicted 74.9 (14.6) in mannitol and 74.6 (14.6) in placebo.

Inclusion criteria

  • Age 15-80 years

  • Bronchiectasis confirmed by HRCT scan

  • FEV 1 ≥ 50% predicted and ≥ 1.0 L

  • Clinically stable (for 2 weeks prior to study entry) and persistent cough

  • (present for the majority of days during 3 months prior to enrolment)

  • Chronic sputum (10 mL/day on the majority of days in the 3 months prior to enrolment)

  • Chronic chest congestion (chronic excessive accumulation of mucus)

  • Ability to perform techniques necessary to measure lung function

  • Provided written informed consent

Exclusion criteria

  • Bronchiectasis due to CF or focal endobronchial lesion

  • Terminally ill or listed for transplant

  • Previous mannitol trial, or participating in any other trial, or used HTS within 4 weeks prior to entry

  • Haemoptysis ( > 60 mL) in previous 6 months or IV antibiotics prescribed for an exacerbation within 4 weeks prior to entry

  • Active signs of asthma, malignancy, or TB, or uncontrolled hypertension

  • Smoking history (20 pack year or 1 cigarette/week within the previous 3 months

  • MI, CVA, or ocular/abdominal/chest/brain surgery within the previous 3 months

  • Pregnancy or lactation

  • Patient likely to develop bronchoconstriction, based on his/her clinical history (including medications) and a positive Aridol test, or the presence of any other condition likely to place the patient at risk

Interventions

320 mg mannitol supplied in 40 mg capsules. i.e. 8 x 40 mg capsules bid versus placebo.

Placebo capsules (10 mg) (Roquette) consisted of non-respirable (approximately 70 μm) United States Pharmacopeia/British Pharmacopoeia good manufacturing practice crystalline mannitol

Similar numbers of patients in each group (46.4% mannitol; 43.7% placebo) were receiving regular physiotherapy to clear secretions

Mannitol and placebo were delivered via a Plastiape dry powder inhaler device

Outcomes

Primary outcomes

  • Absolute change in wet sputum weight

  • Absolute change in SGRQ score from baseline

Secondary outcomes

  • Bronchiectasis symptoms using the Bronchiectasis Symptoms Questionnaire (BSQ) (a study-specific questionnaire)

  • Cough severity using the Leicester Cough Questionnaire (LCQ)

  • Lung function testing (including spirometry, multi-breath nitrogen washout, and carbon monoxide diffusing capacity)

  • Antibiotic use for flare of disease and pulmonary exacerbations (using a non-validated study-designed protocol definition based on the Fuchs definition)

  • High-resolution CT (HRCT) scan sub-study at seven centres (n = 82) using both the modified Bhalla scoring system and quantitative image analysis

  • Microbiology and inflammatory markers on sputum samples

  • Exercise capacity using the shuttle walking test

  • Safety (including adverse events [AEs], spirometry to assess potential bronchoconstriction, and emergence of new pathogens [qualitative microbiology])

Notes

12 weeks duration (plus open-label extension up to 52 weeks for safety assessment)

Funding: Pharmaxis Ltd

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskNo details given
Allocation concealment (selection bias)Unclear riskNo details given
Blinding of participants and personnel (performance bias)
All outcomes
Low riskPlacebo and all staff, patients, and caregivers were blinded to treatment assignment.
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo statement, although blinding may be implied by blinding of ‘caregivers’
Incomplete outcome data (attrition bias)
All outcomes
Low riskITT principles used, although not all patients randomised were analysed, but low attrition rates
Selective reporting (reporting bias)Low riskLow – results for all outcomes (some available in e-tables as supporting information rather than in main trial report)
Other biasUnclear riskNo other indication of bias

Bradley 2011

MethodsRandomised double-blind cross-over trial, single centre, Belfast
Participants

19 people with bronchiectasis (13 completed both treatment periods).

Mean age (SD), 61(11) . Males 9 (47.4%).

Baseline lung function; Mean % predicted FEV1 (SD): 57 (23). Daily sputum weight 17 (17) g)

Diagnostic criteria: Clinical and CT diagnosis

Inclusion criteria:

  • Non CF Bronchiectasis

  • 18-80 years

  • FEV1 less than 90%

  • Chronic sputum production

  • Clinically stable

  • Exclusion criteria:

  • Intolerance to HTS

  • Use of HTS or antibiotics 14 days prior to study

  • Clinically unstable

  • Smokers

  • Not smokers at entry, previous smoking not recorded

InterventionsTwice daily nebulised HTS (6%) masked with quinine sulfate for ITS versus twice daily ITS (0.9%) (nebulised) masked with quinine sulfate for HTS
Outcomes
  • 24 h sputum volume/weight

  • Lung function (FEV1),

  • Cough (Leicester Cough Questionnaire, LCQ)

  • Quality of life

  • Quality of Life (Questionnaire-Bronchiectasis, QOL-B)

  • Adverse events

Notes

4-week intervention with 2-week washout between treatments

Funding: Belfast Health and Social Care Trust

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskThe order of treatment was determined using concealed computerised randomised allocation performed by a statistician not involved in study conduct (EG).
Allocation concealment (selection bias)Unclear riskThe order of treatment was determined using concealed computerised randomised allocation performed by a statistician not involved in study conduct (EG).
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskDescribed as double blind and assessed as successful; A patient and co-ordinator questionnaire indicated that the study design and the use of quinine sulfate was successful in achieving double blinding and masking HTS/ITS
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskAssessment not mentioned explicitly.
Incomplete outcome data (attrition bias)
All outcomes
High riskOnly 13 of 19 completed both arms of the trial
Selective reporting (reporting bias)Unclear riskIncomplete reporting of domains of health-related quality of life scales.
Other biasUnclear riskNo other indication of bias

Chandra 2008

MethodsParallel group randomised double-blind, placebo-controlled trial
Participants

12 patients with bronchiectasis. Unclear in trial report how many participated in the two arms

Baseline lung function; Mean FEV1 L (SD): Mannitol 1.79 (0.59), Placebo 1.74 (0.45)

InterventionsTwice daily mannitol inhalation (dose not specified in trial report) versus placebo
OutcomesSpirometry, HRCT – derived measures mean lumen area (Ai), wall area (Aaw). Data recorded at baseline and 12 weeks.
Notes

A sub-study of a multicentre 12-week trial

Funding: Details not included in abstract.

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskNo details
Allocation concealment (selection bias)Unclear riskNo details
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskNo details; described as ‘double blind’ and ‘placebo’
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo details; described as double blind
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNo data
Selective reporting (reporting bias)Unclear riskThis report (conference abstract) gives limited details of outcomes included in main trial (of which this is a sub-study)
Other biasUnclear riskThis is a small sub-study of a larger trial

Daviskas 1999

MethodsControlled sequential cross-over trial of mannitol versus empty inhaler device versus no treatment. Randomisation between mannitol and no treatment. The mannitol treatment always preceded the empty inhaler device (referred to as control).
Participants

N = 11. Mean age 52 years (SD 7). 3 (27%) male. Baseline lung function; Mean % predicted FEV1 77 (SD 17).

Patients were recruited through their own physicians.

Inclusion criteria: Bronchiectasis diagnosed either with CT or bronchography.

Exclusion criteria: not explicit.

None of the patients had a history of smoking. None of the patients had cystic fibrosis and none was known to have any ciliary defects

Interventions

320 mg* inhaled dry powder mannitol delivered via 9 capsules versus empty inhaler device versus no treatment. *Approximately 300 mg mannitol from inhaler (mean = 320, SD = 81).

Visits were at least 48 hours apart. Inhaler device: Dinkihaler

OutcomesMucociliary clearance and cough
Notes

All patients withheld their medication, postural drainage, and exercise for at least 12 h before beginning the protocol on each day. First visit day (pre-treatment): ‘Assessment of airway responsiveness to dry powder mannitol.’ Five patients were receiving inhaled beta2-adrenergic agonists and glucocorticosteroids regularly, whereas five were not receiving any medication. All had recurrent infections treated with antibiotics. Two were receiving nebulised gentamicin regularly. Two patients were premedicated with nedocromil sodium 16 mg due to either severe airflow limitation or an asthmatic response to an initial mannitol challenge.   

Randomisation sequence unclear from methods; it appears that empty inhaler treatment period never preceded mannitol and it is unclear how this may have affected findings

Funding: National Health and Medical Research Council of Australia

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskNo details given of randomisation for mannitol versus no treatment
Allocation concealment (selection bias)Unclear riskNo details
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskMain comparison is with no treatment and there cannot be any blinding
Blinding of outcome assessment (detection bias)
All outcomes
High riskNo blinding for randomised comparison
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNo attrition implied although not stated
Selective reporting (reporting bias)Low riskNo explicit statement of outcomes, although mucus clearance is in sample size calculation and so presumably primary outcome.
Other biasUnclear riskNo other indication of bias

Daviskas 2004

MethodsRandomised, double-blind, cross-over trial.
Participants

17 participants with bronchiectasis.

Mean age: 59 years. Mean total SGRQ score: 31

Baseline lung function; Mean % predicted FEV1 (SD): 88.6 (4.4) at start of intervention period and 86.5 (3.9) at start of placebo period. Baseline FEF 25-75 % predicted (SD) 45.7 (5,3) at start of intervention period and 41.9 (5.0) at start of placebo period

Inclusion criteria not described.

Interventions

400 mg inhaled mannitol twice daily versus placebo twice daily.

Study duration: 2 x 2 weeks (and washout period 2 weeks)

OutcomesSGRQ, Spirometry, forced oscillation resistance and reactance
Notes

2 weeks of each treatment with 2 week washout period in between

Funding: Pharmaxis Ltd and Viasys Australia

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskUnclear (conference abstract) and no details given.
Allocation concealment (selection bias)Unclear riskUnclear (conference abstract) and no details given.
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskUnclear (conference abstract) and no details given.
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskDescribed as ‘placebo’ but no details given
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskUnclear (conference abstract) and no details given.
Selective reporting (reporting bias)Unclear riskUnclear (conference abstract) and no details given.
Other biasUnclear riskNo other indication of bias

Daviskas 2008

MethodsCross-over trial
Participants

14 people with stable bronchiectasis. Mean age: 63 years (SD: 6) years. Males 3/14 (21%). Baseline lung function; Mean % predicted FEV1 (SD): 77 (14)

Bronchiectasis was diagnosed with a high-resolution computed tomography (HRCT) scan. Based on the HRCT scan, all participants except one had extensive bronchiectasis involving at least the lower lobes, bilaterally. In one participant only, the bronchiectatic changes were localised on the lower left lobe and were reported as mild. None of the participants had a diagnosis of cystic fibrosis and all had long-standing symptoms of bronchiectasis. All participants had never been smokers

Inclusion criteria: An airway challenge with mannitol was performed in potential participants following an approved protocol in order to identify airway hyperresponsiveness to dry powder mannitol. Participants were eligible to proceed to subsequent visits if they had a decrease in FEV1 to ,15% of baseline FEV1 after the airway challenge with mannitol.

Interventions

Single dose of each intervention & control & no treatment. Treatment given for 15 minutes on each day

160 mg or 320 mg or 480 mg of mannitol. Mannitol was delivered in doses of 160 mg (four capsules of 40 mg),

320 mg (eight capsules of 40 mg) and 480 mg (12 capsules of 40 mg) versus breathing while at rest (called ‘baseline’ in the trial report but referred to as no treatment in this review). On the no treatment study day, participants rested in the sitting position during the intervention period

All participants withheld their regular medication for approx 20 hours on each study day

Dry powder mannitol (Pharmaxis Ltd, Frenchs Forest, Australia) was inhaled from capsules using the low-resistance dry powder inhaler (RS 01; Plastiape, Osnago, Italy).

Outcomes

Mucus clearance (%) during intervention over 15 minutes

Mucus clearance (%) post-intervention during 30 minutes

Mucus clearance (%) during 30 minutes after intervention in response to 100 requested voluntary coughs (i.e. 3 consecutive periods)

In each of: Whole Right lung;central, intermediate and peripheral regions of right lung

Data recorded during intervention and after each session as above

Notes

Cross-over (4 days; 3 doses and no treatment): Fifth day – control NOT randomised

Funding: National Health and Medical Research Council of Australia

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskStudy days 1–4 were randomised i.e. (no treatment and 3 doses of mannitol)
Allocation concealment (selection bias)Unclear riskNo details
Blinding of participants and personnel (performance bias)
All outcomes
High riskNo treatment day not blinded; unlikely that other arms blinded as numbers of capsules varies and no mention of placebos delivered in doses of 160 mg (four capsules of 40 mg), 320 mg (eight capsules of 40 mg) and 480 mg (12 capsules of 40 mg)
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskAs above. Interpretation of nuclear medicine imaging, conducted away from patient
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNo indication of participants failing to complete
Selective reporting (reporting bias)Unclear riskNo list of outcomes, although apparently complete reporting of mucus clearance
Other biasUnclear riskMeasurement of right lung sputum clearance with at least one patient with bronchiectasis limited to left lung

Kellet 2011

MethodsRandomised single-blind, two-period cross-over study. 4–week washout period
Participants

30 participants with bronchiectasis (28 completed study).
Mean age: 56.6 (14.6). Male: 14 (50%). Mean 2.6 exacerbations per year. Baseline lung function; Mean % predicted FEV1 (SD): 66.4 (26.1)

Diagnosis: confirmed by a high resolution CT scan (performed within the last 4 years).

Inclusion criteria:

  • Adult patients over the age of 18, with a pre-existent diagnosis of bronchiectasis, as confirmed by a high resolution CT scan

  • At screening patients were required to undertake a challenge to 7% HTS, to demonstrate tolerability as confirmed by a fall in FEV1 of less than 10% from pre-challenge levels.

Exclusion criteria:

  • Pseudomonas colonisation

  • Any severe medical condition that excluded the use of HTS

  • An expected broncho-constrictor response to a 7% challenge to HTS, such as evidence or history of bronchial reactivity, use of nebulised antibiotics or maintenance oral antibiotics or previous use of nebulised HTS

Interventions

4-week run in.

4 mL HTS (7%) once/day versus 4 mL ITS (0.9%) once/day

Other than exclusion criteria, participants were allowed to continue all chronic medication but were requested to keep a diary and report if medication changed during the course of the study.

Outcomes
  • Spirometry (FEV1 and FVC),

  • SGRQ,

  • Antibiotic use,

  • Annualised exacerbation rate (unscheduled primary or secondary care attendance).

  • Sputum viscosity

  • Ease of expectoration.

Data collected before and after each 3-month treatment

Notes

3-month trial (each arm)

Funding: Details not included in trial report

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskNo details
Allocation concealment (selection bias)Unclear riskNo details
Blinding of participants and personnel (performance bias)
All outcomes
High risk

Single blind

Reported as "At the time of performing this study we could not get a producer to provide normal saline in glass ampoules or HS in plastic ampoules. Patients were unaware of which solution was the active one. HS was provided in brown glass ampoules containing 4 mL of colourless solution. IS was provided in 4 mL plain plastic ampoules. Whilst we did not tell the subjects which solution was which, it is possible that trial volunteers could have done research to identify, which solutions are provided and for which there was existing information of benefit."

Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo details
Incomplete outcome data (attrition bias)
All outcomes
Low risk28/30 complete
Selective reporting (reporting bias)Unclear riskData for some outcomes. Supplementary material is graphical
Other biasUnclear riskNo other indication of bias

Kellett 2005

MethodsDouble-blind randomised cross-over trial. Four single sessions over a four-week period
Participants

24 participants with newly diagnosed stable bronchiectasis.

Male: 7 (29%) age 64 (3.8)

Female age 55 (3.6).

Diagnosis: Clinical and/or CT scan diagnosis of bronchiectasis

Inclusion criteria: Bronchiectatic patients who produce less than 10 g per day of sputum. Diagnosed by high resolution CT scanning, who had not previously had physiotherapy input were referred from chest physicians to take part in the study.

Exclusion criteria: Allergic broncho-pulmonary aspergillosis and cystic fibrosis phenotypes were excluded. Participants excluded if they had been prescribed a course of antibiotics in the last month or required antibiotics during the course of the study.

Interventions

Hyperosmolar saline - 7% after nebulised terbutaline versus normal saline after nebulised terbutaline 4 periods

(1) Active cycle breathing technique (ACBT) alone.

(2) Nebulised terbutaline followed by ACBT after 10 min.

(3) Nebulised terbutaline followed after 10 min by nebulised ITS (0.9%) then ACBT.

(4) Nebulised terbutaline followed after 10 min by nebulised HTS (7%) then ACBT.

Washout periods of one week

ACBT was performed immediately after nebulisation of 2 mL of HTS or ITS

Physiotherapy - active cycle breathing technique alone

Nebulised terbutaline was used in phases (2, 3, 4) as a premedication to minimise adverse effects.

Patients were able to continue all routine medication (inhaled/nebulised bronchodilators, long-acting bronchodilators, leukotriene antagonists, oral or inhaled steroids)

OutcomesSputum weight, sputum viscosity, ease of expectoration and lung function (FEV1 and FVC)
NotesFunding not specified
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskComputer determined random allocation
Allocation concealment (selection bias)Unclear riskNo details
Blinding of participants and personnel (performance bias)
All outcomes
Low riskThe nebuliser solutions for HTS and ITS were produced and used double blind. ITS and HTS solutions were produced by the pharmacy in blinded labelled sterile solutions.
Blinding of outcome assessment (detection bias)
All outcomes
Low riskThe nebuliser solutions for HTS and ITS were produced and used double blind. The physiotherapist performing challenges and sputum assessments remained blinded to ITS and HTS throughout the study.
Incomplete outcome data (attrition bias)
All outcomes
Low risk23 (of 24) patients in Table 2 of the trial report
Selective reporting (reporting bias)Unclear riskPaper gives data on all outcomes stated to be primary
Other biasUnclear riskNo other indication of bias

NCT00669331

MethodsPhase III multicentre, randomised, parallel group, controlled, double-blind study
Participants

People with non-cystic fibrosis bronchiectasis.

461 patients randomised and received study drug (ITT)

Mannitol ITT = 233 (Completers = 191 (82.0%))

Control ITT = 228 (Completers = 189 (82.9%))

Age (SD) Mannitol 59.3 yrs (14.1) Placebo 60.3 yrs (13.0) .

Males (SD) Mannitol 86 (36.9%) Placebo 86 (37.7%)

Historical exacerbation rate (SD) Mannitol 3.20/yr (1.38) Placebo 3.25/yr (1.40)

SGRQ score (SD) at baseline: Mannitol 53.0 (14.6) Placebo 52.2 (14.7)

FEV1%predicted (SD) at baseline: Mannitol 63.0% (13.6) Placebo 61.5% (13.4)

Inclusion criteria

  • Confirmed diagnosis of (non-CF) Bronchiectasis

  • 18-85 years of age

  • FEV1 40 - 85% predicted & ≥ 1 L

  • Sputum producers (screening ≥10 g)

  • ≥ 2 exacerbations in past year, and ≥ 4 in the past 2 years

  • SGRQ total score ≥30

  • Standard therapy continued (no HTS)

  • Stable condition on entry (including no major haemoptysis)

  • Pass Mannitol Tolerance Test (MTT) without pre-bronchodilator use

  • sputum production of ≥1 tablespoonful (15 mL) per day on the majority of days in the 3 months prior to screening

Exclusion criteria

  • Episode of haemoptysis (> 60 mL) in the previous 6 months

  • Not be taking rescue antibiotics for 4 weeks prior – be stable on entry

  • Have previously used Bronchitol

  • Usual excl criteria e.g. major surgery or heart problems

Countries included in the Study were: Argentina, Australia, Belgium, Chile, Germany, Netherlands, New Zealand, United Kingdom and USA.

84 sites screened at least one patient, with 78 sites randomising at least 1 patient

Interventions

Inhaled mannitol 400 mg bid (10 capsules twice a day) for 52 weeks versus matched control. The control was 50 mg mannitol, designed as placebo (non therapeutic dose)*** - communication from Dr Bilton, Royal Brompton Hospital, London UK.

(Protocol in http://ClinicalTrials.gov/show/NCT00669331)

Outcomes

Primary outcomes: antibiotic use for pulmonary exacerbations and health-related quality of Life as measured by the St. Georges Respiratory Questionnaire (SGRQ)

Secondary outcomes: exacerbations, sputum volume, daytime sleepiness scores, lung function (FEV1, FVC, FEV1/FVC, FEF25-75 values), safety profile, health-related costs, quality adjusted life years (QALYs), hospitalisations due to pulmonary exacerbations

Notes

Trial results presented at the 2013 European Respiratory Society Conference

Graded pulmonary exacerbation = worsening in respiratory signs and symptoms requiring a change in treatment*. Grading is based on the assessment of protocol pre-defined main and minor respiratory signs and symptoms**.

Type I: 3 main signs and symptoms

Type II: 2 main signs and symptoms

Type III: 1 main sign and symptoms, plus at least 1 minor sign and symptom

Main – Increased cough, increased sputum volume, Increased sputum purulence

Minor – URTI within previous 5 days, fever with no cause, increased wheezing, increased dyspnoea, increased respiratory rate, increased cardiac frequency, increased malaise, fatigue or lethargy.

**Adapted from AnthonisenNR, ManfredaJ, Warren CP, Hersfield ES, Harding GK, Nelson NA. Antibiotic therapy in exacerbations of chronic obstructive pulmonary disease. Ann Intern Med 1987;106:196-204

*Guidance for Industry Chronic Obstructive Pulmonary Disease: Developing Drugs for Treatment DRAFT GUIDANCE U.S. Department of Health and Human Services Food and Drug Administration Center for Drug Evaluation and Research (CDER) November 2007

***It is possible that this may have resulted in underestimates of any treatment effects

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskUnclear (results reported in oral conference presentation)
Allocation concealment (selection bias)Unclear riskUnclear (results reported in oral conference presentation
Blinding of participants and personnel (performance bias)
All outcomes
Unclear riskUnclear (results reported in oral conference presentation
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskUnclear (results reported in oral conference presentation .
Incomplete outcome data (attrition bias)
All outcomes
Low risk

Mannitol arm withdrawals n = 42 (18.0%) Due to AEs N = 16 (6.9%)

Placebo arm withdrawals n = 39 (17.1%) Due to AEs N = 10 (4.4%)

Selective reporting (reporting bias)Unclear riskResults reported in oral conference presentation do not reflect all outcomes listed in clinicaltrials.gov protocol for this trial and focus on exacerbations, health-related quality of life and adverse events
Other biasUnclear riskNo other indication of bias

Nicolson 2012

  1. a

    bid: twice a day; CF: cystic fibrosis; CT: computed tomography; CVA: cerebrovascular accident; FEF: forced expiratory flow; FEV1: forced expiratory volume in one second; FVC: forced vital capacity; HRCT: high-resolution computerised tomography; HTS: hypertonic saline; ITS: isotonic saline; ITT: intention-to-treat; IV: intravenous; MI: myocardial infarction; SD: standard deviation; SGRQ: St George's Respiratory Questionnaire; URTI: upper respiratory tract infection; VAS: visual analogue scale

MethodsRandomised, double-blind, parallel group trial
Participants

40 participants with mild bronchiectasis.

Intervention group mean age and SD 58 (15) years. Control group mean age and SD 56 (15) years. Males 7/20 (35%) in intervention group and 8/20 (40%) in control group.

Baseline lung function; Mean FEV1 L (SD): intervention 2.28 (0.87) and control 2.16 (0.75)

Baseline lung function; Mean % predicted FEV1 (SD): intervention 84.8 (20.5) and control 80.4 (21.1)

Inclusion criteria

  • Adults with bronchiectasis, confirmed by high resolution computed tomography

  • At least 2 respiratory exacerbations per year over the past 2 years producing sputum daily in a stable clinical state over 18 years of age

Exclusion criteria

  • Cystic fibrosis

  • Positive response to HTS challenge

  • FEV 1 decreased by ≥ 15%

  • FEV 1 ≤ 1L

Interventions

HTS (6%) 5 mL twice a day via a nebuliser for 12 months versus ITS

AeronebGo mouthpiece micropump nebuliser (Niche Medical, Perth, Australia). (described in relation to screening, but assumed same for intervention)

Usual treatment unchanged.

Outcomes
  • Number of pulmonary exacerbations

  • St George Respiratory questionnaire, Leicester Cough Questionnaire,

  • FEV1

  • FVC

  • FEF 25-75

  • Cough frequency (VAS)

  • Sputum colonisation

Data collected at randomisation, 3, 6, 12 months

An exacerbation was defined as recording at least three symptoms in one day for two or more consecutive days. Exacerbations requiring commencement of antibiotics and exacerbations based on symptoms only, were calculated.

Notes

12-month trial

Funded by Alfred Research Trust, Physiotherapy Research Foundation (Australian Physiotherapy Association), Alfred

Physiotherapy Research Grant and the Alfred Hospital Allergy, Immunology and Respiratory Medicine Department

Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskRandomised using a computer-generated randomisation sequence
Allocation concealment (selection bias)Low riskAllocation concealed using sealed opaque envelopes
Blinding of participants and personnel (performance bias)
All outcomes
Low riskThe sachets of saline, distributed by the clinical trials pharmacy, were of the same size and appearance. However the taste of the saline solutions was not masked by the addition of quinine (previously used as a blinding agent).
Blinding of outcome assessment (detection bias)
All outcomes
Low risk

Assessors, therapists, participants and their physicians were blinded to treatment allocation. A co-investigator, who had no role in treatment allocation or outcome assessment, was un-blinded to manage adverse events.

Fifty per cent of participants correctly identified the treatment they had been receiving and the treating physiotherapist identified the correct therapy in 58% of the participants at the final visit.

Incomplete outcome data (attrition bias)
All outcomes
Low riskN = 2 stopped HTS treatment, but they are included in the analysis
Selective reporting (reporting bias)Low riskThere are findings reported for all variables
Other biasUnclear riskNo other indication of bias

Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion
Briffa 2009Not an inhaled hyperosmolar agent. Trial of drug to reduce side effects of hyperosmolar agent, not of hyperosmolar agent per se.
Briffa 2011Trial of drugs to prevent adverse effects of hyperosmolar agent in bronchiectasis, not of hyperosmolar agent per se
Daviskas 2001Non-randomised before and after study. Eight patients with bronchiectasis not due to cystic fibrosis or ciliary immotility were subjected to mucociliary clearance studies with an inhaled radioaerosol on three successive days, with 330 mg mannitol administered on day 2.
Daviskas 2003Before and after study which studied the effect of 12 days' treatment of mannitol 400 mg daily on health status and lung function.
Daviskas 2010Not bronchiectasis.
Harrison 1983Not an inhaled hyperosmolar agent.
Murray 2011Not an inhaled hyperosmolar agent.
NCT00105183Not hyperosmolar agent, not bronchiectasis.
NCT00730977Single group design (no control group).
NCT00749866Not an inhaled hyperosmolar agent.
NCT01076491Single group design (no control group).
NCT01313624Not an inhaled hyperosmolar agent.
NCT01314716Not an inhaled hyperosmolar agent.
NCT01677403Not an inhaled hyperosmolar agent.
Serisier 2013Not an inhaled hyperosmolar agent.
Sutton 1988Not an inhaled hyperosmolar agent.

Ancillary