Types of studies

Randomised controlled trials (RCTs) or quasi‐RCTs. Two review authors assessed quasi‐RCTs and only included them if the allocation method was described and participants in the intervention groups were judged to be similar based on demographic characteristics at baseline.

All included trials are cross‐over trials and we have commented on the appropriateness of the analysis. We have only included data from these cross‐over trials where the investigators appropriately analysed their cross‐over data, or if first treatment phase data were available (Elbourne 2002). As short‐acting inhaled bronchodilators have an action lasting up to 8 hours, we only included cross‐over trials which had a washout period of 8 hours or more between treatments.

Types of participants

Children and adults with CF diagnosed by sweat test or genetic testing, with all stages and severity of lung disease, and with or without concomitant asthma.

Types of interventions

Short‐acting inhaled bronchodilators, delivered by any device, at any dose, at any frequency and for any duration compared to either placebo or another inhaled short‐acting bronchodilator in people with CF. This includes trials which looked at intermittent use and regular use.

The bronchodilators include:

1. short‐acting beta‐2 agonists (salbutamol (also known as albuterol), terbutaline and fenoterol); and

2. short‐acting muscarinic antagonists (ipratropium bromide).

We included single‐use trials (e.g. pre‐airway clearance therapy) and trials up to 7 days, short‐term trials up to 28 days, and longer‐term trials lasting for more than 28 days.

We did not include combined inhalers (with inhaled steroids or another bronchodilator) or long‐acting bronchodilators (e.g. tiotropium, salmeterol).

For this review 'inhaled' includes the use of pressurised MDIs, with or without a spacer, dry powder devices and nebulisers.

Types of outcome measures

We planned to assess the following outcome measures.

Electronic searches

We identified relevant trials from the Group's Cystic Fibrosis Trials Register using the terms: bronchodilator AND inhaled.

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 Unpublished work is identified by searching the abstract books of three major cystic fibrosis conferences: the International Cystic Fibrosis Conference, the European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

Date of last search: 28 March 2022.

We also searched the following trial registries and registers.

1. US National Institutes of Health Ongoing Trials Register ClinicalTrials.gov (www.clinicaltrials.gov; searched 12 April 2022).

2. Australian New Zealand Clinical Trials Registry (www.anzctr.org.au; searched 12 April 2022).

3. World Health Organization International Clinical Trials Registry Platform (WHO ICTRP) (trialsearch.who.int/; searched 12 April 2022).

For details of our trial registry search strategies, please see Appendix 2. 

Searching other resources

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

Selection of studies

For the original review of both short‐acting and long‐acting inhaled bronchodilators (Halfhide 2016), three review authors (Drs Halfhide, Evans and Couriel) independently selected the references to be included in the review from those identified by the search. For this new review, we included any trials from the original review where they fitted the current inclusion criteria.

Two review authors (SS, NR or CE) independently reviewed all articles on title and abstract and discarded any that clearly did not meet the inclusion criteria. We found the full text for each of the remaining references and two authors (SS, NR or CE) again filtered against the inclusion criteria. At this point, we added to the list of excluded trials any references that we discarded and gave reasons for this. We referred any disagreements between authors to a third review author (NR or CE depending on who did the original screening). We used Covidence for this process (Covidence 2019).

Data extraction and management

Two review authors (SS, NR or CE) independently extracted data from the identified trials using a specially designed data extraction form developed by the Cochrane Cystic Fibrosis and Genetic Disorders Review Group and adapted to this review. We set up a form in Covidence to allow independent data extraction by two review authors and the inbuilt comparison of responses (Covidence 2019).

We collected data on:

1. participant characteristics;

2. trial characteristics and trial design;

3. intervention and comparator;

4. outcome data ‐ we reported each outcome separately.

We extracted data for all trials identified for the review, whether they were included in the original review or not. We attempted to come to an agreement by discussion where there were discrepancies and where the issue still could not be resolved, we brought in a third review author to arbitrate (NR or CE depending on who did the original extraction).

We exported the extracted data from Covidence into the Review Manager 5.4 software for analysis (Covidence 2019; Review Manager 2020). We presented the data for beta‐2 agonists and muscarinic antagonists separately as the agents have different modes of action. Where possible we carried out the following comparisons.

1. Short‐acting inhaled beta‐2 agonists versus placebo, usual treatment or another short‐acting inhaled beta‐2 agonist

2. Short‐acting inhaled muscarinic antagonists versus placebo, usual treatment or another short‐acting inhaled muscarinic antagonist

3. Short‐acting inhaled beta‐2 agonists versus short‐acting inhaled muscarinic antagonists

We planned to present data at ≤ 7 days (including single‐dose trials), > 7 days and ≤ 28 days, > 28 days and ≤ 6 months, > 6 months and ≤ 1 year and annually thereafter to distinguish between single‐use or short‐term trials and longer‐term trials. We have only been able to analyse data from single‐dose trials and present results as such.

Assessment of risk of bias in included studies

Two review authors (SS, NR or CE) assessed all included trials for risk of bias regardless of whether they were previously assessed for the original version of the review or not.

We used the RoB 1 tool, as described in the Cochrane Handbook for Systematic Reviews of Interventions, to assess the risk of bias across the following six domains (Higgins 2017).

1. Sequence generation

2. Allocation concealment

3. Blinding (of participants, outcome assessors and trial personnel)

4. Incomplete outcome data

5. Selective reporting

6. Other potential sources of bias

We assessed each of the domains and ranked them as being at high, low or unclear risk of bias. We qualified our judgements in the risk of bias tables and have presented the information in a summary figure.

Measures of treatment effect

We have only been able to analyse data from three of our included trials, and for the remaining trials we have reported results from the original trial papers. Where data were available, we have reported measures of treatment effect for each outcome and each type of short‐acting inhaled bronchodilator separately, and followed the guidelines in the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2021).

For continuous data (FEV₁, FVC, FEF_25-75, lung clearance index, symptom scores, quality of life), we  recorded the mean change and standard deviation (SD) from baseline in each group. Where there was more than one trial, we calculated a pooled estimate of treatment effect using the mean difference (MD) and 95% confidence intervals (CIs). If there are more data available in future updates, where they have used different measurement scales or units of measurement (e.g. for symptom scores) we will use the standardised mean difference (SMD) to report the results. Where the trial authors only presented a baseline mean (SD) and a postintervention mean (SD) we were unable to calculate the SD of the change, and so we reported these results narratively.

For our primary outcome, we looked for a change of > 12% from baseline to show a meaningful effect (Barreiro 2004).

There were no dichotomous data to enter into our analyses, but if there are data to present in future updates (adverse events, number of exacerbations), we will calculate a pooled estimate of the treatment effects for each outcome across trials using the risk ratio (RR) and 95% CIs.

Unit of analysis issues

All of our included trials are cross‐over trials and we assessed each one to establish whether we could include any data in the analysis. Where we believe the trial authors have carried out an appropriate analysis for cross‐over trials, taking into account carry‐over effect and within‐person differences, we have included their results in our analyses (Hordvik 1996; Serisier 2000; Ziebach 2001). Where investigators have not analysed the data appropriately, no first‐phase data were available to analyse, or it is unclear how the investigators have analysed data, we have reported results narratively. This method means, however, that the advantage of using participants as their own controls is lost (Elbourne 2002). 

Where any of the trials have multiple arms, we have separated out the different arms and reported in the appropriate comparison. For example, in a trial reporting a muscarinic antagonist versus beta‐2 agonist versus placebo, we have analysed each treatment arm separately against placebo, and where appropriate, included in a meta‐analysis.

Dealing with missing data

The review authors attempted to contact the primary trial investigators for any missing data. We also assessed the use of intention‐to‐treat (ITT) analysis in the individual trials and the information given to explain the numbers of participants who dropped out of each arm of the trial.

Assessment of heterogeneity

Where we were able to include trials reporting the same outcome in a meta‐analysis, we tested for heterogeneity using visual inspection of the forest plots and the I² statistic. We looked for similar point estimates and overlapping CIs on the forest plots as an indication of low heterogeneity. We reported high levels of heterogeneity when point estimates varied and the CIs did not overlap. The I² statistic describes the percentage of total variation across trials which is due to heterogeneity rather than chance. We based our definitions of levels of heterogeneity on the levels given in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2021):

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

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

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

4. 75% to 100%: considerable heterogeneity.

We also reported on the magnitude and direction of the effects contributing to any heterogeneity found (Deeks 2021).

Assessment of reporting biases

We did not find sufficient trials with data to attempt to identify publication bias. If this changes at a future update, we will attempt to identify any publication bias in the included trials by generating a funnel plot. If there is asymmetry in the funnel plot, we will attempt to discover the reason, including reasons not related to publication bias (Page 2021).

We attempted to identify any selective reporting in the included publications by careful examination of the trial publications and consideration of reporting of both positive and negative effects of the intervention. We looked at competing interests of the authors to determine the role of external bias being introduced.

Data synthesis

We were only able to combine two trials in a meta‐analysis, and only produced three forest plots. If more data are available at a future update we will generate forest plots for each outcome. We will report each outcome on a separate forest plot and also separate out the different comparisons, e.g. short‐acting inhaled beta‐2 agonists versus placebo, short‐acting inhaled muscarinic antagonists versus placebo, etc.

Where there is low heterogeneity (I² < 40%) we will use a fixed‐effect model, but if heterogeneity > 40% we will use a random‐effects model (Deeks 2021).

Subgroup analysis and investigation of heterogeneity

We were unable to add any data into one meta‐analysis and therefore could not examine heterogeneity between trial results. In future updates if this becomes possible, where we find at least moderate statistical heterogeneity (I² > 50%) between the included trials, we will investigate the possible causes of this further by performing subgroup analyses looking at the effects of:

1. age;

2. severity of lung disease;

3. the presence of atopy; and

4. the presence of reversible airway obstruction (based on reversibility or methacholine challenge or exercise testing).

Sensitivity analysis

If we are able to include more data in a meta‐analysis in a future update we will carry out a sensitivity analysis to look at the effects of risk of bias on the trial results. We will look at the effect of adding in and taking out trials where there is high risk of bias. 

Summary of findings and assessment of the certainty of the evidence

We have presented a summary of findings table for each comparison that we have included in the review. We have selected the following outcomes to present, which we consider to be the most clinically relevant.

1. Change from baseline in FEV₁ % predicted

2. Quality of life (using a validated measure such as CFQ‐R (Quittner 2009), CFQoL (Gee 2000))

3. Number of exacerbations

4. Participant‐reported respiratory symptom score (using a validated score such as RSS, RSSQ (Goss 2007))

5. Adverse events (all treatment‐related adverse effects)

6. Change from baseline in FVC % predicted

7. Sputum volume (mL)

For each outcome we have reported the illustrative risk with and without the intervention and the magnitude of effect (RR or MD) (Schünemann 2021a). We also presented the number of trials contributing to the outcome and a grade for the overall certainty of the body of evidence. To do this we used the GRADE approach (Schünemann 2021b). We assessed the certainty of evidence based on the risk of bias within the trials contributing to the outcome, relevance to our population of interest (indirectness), unexplained heterogeneity or inconsistency, imprecision of the results and risk of publication bias. As a starting point we have labelled all of the evidence as high certainty, but we downgraded the evidence where there were issues in any of the aforementioned categories. We downgraded the evidence once where the risk was serious and twice where we deemed the risk to be very serious. We graded the body of evidence for each outcome as either high, moderate, low or very low certainty. We added reasons and supporting information for our decisions in the footnotes to the summary of findings tables.