Airway clearance techniques compared to no airway clearance techniques for cystic fibrosis

Abstract

Background

Cystic fibrosis (CF) is an inherited progressive life‐limiting disease characterised by the build‐up of abnormally thick, sticky mucus affecting mostly the lungs, pancreas, and digestive system. Airway clearance techniques (ACTs), traditionally referred to as chest physiotherapy, are recommended as part of a complex treatment programme for people with CF. The aim of an ACTs is to enhance mucociliary clearance and remove viscous secretions from the airways within the lung to prevent distal airway obstruction. This reduces the infective burden and associated inflammatory effects on the airway epithelia. 

There are a number of recognised ACTs, none of which have shown superiority in improving short‐term outcomes related to mucus transport. This systematic review, which has been updated regularly since it was first published in 2000, considers the efficacy of ACTs compared to not performing any ACT in adults and children with CF. It is important to continue to review this evidence, particularly the long‐term outcomes, given the recent introduction of highly effective modulator therapies and the improved health outcomes and potential changes to CF management associated with these drugs.

Objectives

To determine the effectiveness and acceptability of airway clearance techniques compared to no airway clearance techniques or cough alone in people with cystic fibrosis.

Search methods

We searched the Cochrane Cystic Fibrosis and Genetic Disorders Group Trials Register, which comprises references identified from comprehensive electronic database searches and handsearches of relevant journals and abstract books of conference proceedings, to 17 October 2022. We searched ongoing trials registers (Clinicaltrials.gov and the WHO International Clinical Trials Registry Platform) to 7 November 2022.

Selection criteria

We included randomised or quasi‐randomised studies that compared airway clearance techniques (chest physiotherapy) with no airway clearance techniques or spontaneous cough alone in people with CF.

Data collection and analysis

Both review authors independently assessed study eligibility, extracted data, and assessed the risk of bias of the included studies. We used GRADE methodology to assess the certainty of the evidence.

Main results

We included 11 cross‐over studies (153 participants) and one parallel study (41 participants). There were differences between studies in how the interventions were delivered, with several intervention groups combining more than one ACT. One study used autogenic drainage; five used conventional chest physiotherapy; nine used positive expiratory pressure (PEP), with one study varying the water pressure between arms; three studies used oscillating PEP; two used exercise; and two used high‐frequency chest wall oscillation (HFCWO). Of the 12 included studies, 10 were single‐treatment studies, and two delivered the intervention over two consecutive days (once daily in one study, twice daily in the second). This substantial heterogeneity in the treatment interventions precluded pooling of data for meta‐analysis. Blinding of participants, caregivers, and clinicians is impossible in airway clearance studies; we therefore judged all studies at unclear risk of performance bias. Lack of information in eight studies made assessment of risk of bias unclear for most other domains. 

We rated the certainty of evidence as low or very low due to the short‐term cross‐over trial design, small numbers of participants, and uncertain risk of bias across most or all domains.

Six studies (84 participants) reported no effect on pulmonary function variables following intervention; but one study (14 participants) reported an improvement in pulmonary function following the intervention in some of the treatment groups. Two studies reported lung clearance index: one (41 participants) found a variable response to treatment with HFCWO, whilst another (15 participants) found no effect on lung clearance index with PEP therapy (low‐certainty evidence). Five studies (55 participants) reported that ACTs, including coughing, increased radioactive tracer clearance compared to control, while a further study (eight participants) reported no improvement in radioactive tracer clearance when comparing PEP to control, although coughing was discouraged during the PEP intervention. We rated the certainty of evidence on the effect of ACTs on radioactive tracer clearance as very low.

Four studies (46 participants) investigated the weight of mucus cleared from the lungs and reported greater secretions during chest physiotherapy compared to a control. One study (18 participants) reported no differences in sputum weight (very low‐certainty evidence).

Authors' conclusions

The evidence from this review shows that ACTs may have short‐term effects on increasing mucus transport in people with CF. All included studies had short‐term follow‐up; consequently, we were unable to draw any conclusions on the long‐term effects of ACTs compared to no ACTs in people with CF.

The evidence in this review represents the use of airway clearance techniques in a CF population before widespread use of cystic fibrosis transmembrane conductance regulator (CFTR) modulators. Further research is needed to determine the effectiveness and acceptability of airway clearance in those treated with highly effective CFTR modulators.

Plain language summary

Airway clearance techniques compared to no airway clearance techniques for cystic fibrosis

Review question

What are the effects of using any airway clearance technique compared to not using an airway clearance technique for clearing excess mucus from the lungs of people with cystic fibrosis?

Background

The lungs of people with cystic fibrosis produce excess mucus. This leads to repeated infection and tissue damage in the lungs. It is important to clear the mucus using medicines and airway clearance techniques (physiotherapy). There are different airway clearance techniques for clearing mucus, some of which may include the use of mechanical devices. Daily physiotherapy takes a lot of time and trouble, so it is important to know if it works. We searched for studies where the people taking part had equal chances of being in the group using airway clearance techniques or the group with no airway clearance techniques. This is an update of a previously published review.

Search date

The evidence is current to 17 October 2022.

Study characteristics

We included 12 studies that enrolled 194 people with cystic fibrosis. The studies were very different and some looked at multiple treatments compared to no treatment. One study used autogenic drainage (a controlled breathing technique which uses different speeds and depths of exhaled breath to move mucus up the airways so it can be cleared by coughing); five studies used conventional chest physiotherapy (manual techniques of percussion and vibration applied to the chest wall, usually with the assistance of a physiotherapist or relative); nine used positive expiratory pressure (breathing out through a mask or mouthpiece against a resistance which causes pressure to build up in the lungs to move the mucus), and one of these varied pressure so used both standard and high‐pressure positive expiratory pressure; three studies used oscillating positive expiratory pressure (positive expiratory pressure combined with vibrations within the airway to loosen mucus); two used exercise (on a treadmill); and two used high‐frequency chest wall oscillation (high frequency vibrations applied outside the chest wall via an inflatable garment). We could not combine any results to analyse them statistically.

Key results

Summarising the findings of the 12 studies, we found limited evidence of a short‐term impact on lung function. Only one study reported an improvement in lung function in some of the treatment groups, whilst six other studies found no improvement. 

This review found that methods of clearing the airways may have short‐term benefits for moving mucus. Four studies found that the people using airway clearance techniques coughed up more sputum, but one study reported no difference with or without using an airway clearance technique. Five studies reported increased radioactive tracer clearance (a test in which people are imaged continuously after inhaling a radioaerosol to assess the time for it to be cleared from the lung) when using airway clearance, but one study of positive expiratory pressure found no difference. At present, there is no clear evidence to show the long‐term effects of performing airway clearance techniques on quality of life or survival.

Limitations of the evidence

We have little or very little confidence in the evidence, for several reasons. Most included studies had design problems, and in just under half of the studies, it was unclear whether all the results were reported. Also, in physiotherapy studies, the person receiving treatment and their physiotherapist know which treatment they are receiving, and this may affect some of the findings. For example, the amount of mucus coughed up and lung function tests (measured by half of the included studies) and a person's views on a particular technique (recorded in a quarter of the included studies) may be affected if a person is aware of which treatment they are receiving. Finally, it was not clear in most studies whether the individual was experienced with the technique they were using.

We were unable to find any studies looking at the effects of airway clearance techniques in people treated with the new cystic fibrosis transmembrane conductance regulator (CFTR) modulator treatments. Cystic fibrosis is caused by faulty proteins on the cell surface made by the mutated CFTR gene; these new medicines are designed to correct the function of the faulty proteins. 

Summary of findings

Summary of findings 1. Summary of findings for airway clearance techniques compared with no airway clearance techniques for cystic fibrosis.

Airway clearance techniques compared with no airway clearance techniques for cystic fibrosis
Patient or population: children and adults with cystic fibrosis Settings: outpatients Intervention: ACT (different combinations of PEP (standard or with varied water pressure), FET, postural drainage, conventional physiotherapy (manual percussion and vibrations), AD, HFCWO, exercise, and Flutter)a Comparison: no ACTb
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Assumed risk (no ACT)	Corresponding risk (ACT)
Lung function: FEV1 % predicted change from baseline   Follow‐up: 12 months	No studies reported FEV1 % predicted change from baseline at 12 months.					6 studies reported short‐term changes to FEV1 (L). Braggion 1995, van der Schans 1991 and Vandervoort 2022 reported no difference in FEV1 after treatment in either group. Jarad 2010 found a reduction in FEV1 after ACT but no difference in the control group. Changes were short‐lived and returned to baseline by the next day. Pfleger 1992 found an improvement in FEV1 after PEP (P < 0.05) and PEP‐AD (P < 0.02) but no difference in the AD, AP‐PEP, or control groups. Unpublished data from Dwyer 2017 showed no difference between Flutter and control (MD 0.02 L, 95% CI −0.04 to 0.08) or between exercise and control (MD 0.19 L, 95% CI −0.09 to 0.47).
Lung function: LCI   Follow‐up: after treatment	1 study reported that 15/20 participants in the intervention group showed an improvement in LCI after treatment, whereas the control group showed no change.   1 study reported no difference between the groups.		—	58 (2)	⊕⊕⊝⊝ Lowc	—
Exacerbation rate  (number of exacerbations/year)   Follow‐up: after treatment	No studies reported exacerbation rates.					—
Mucus transport rate (radio tracer clearance)   Follow‐up: after treatment	5 studies reported that ACTs, including coughing, increased radioactive tracer clearance compared to control.   1 study reported no difference in radioactive tracer clearance between groups.		—	55 (6)	⊕⊝⊝⊝ Very lowd,e	We were unable to enter any data into our analyses and we have reported narratively from the original papers. Results were reported in different ways and lacked SDs or CIs.
Expectorated secretions (sputum wet weight)   Follow‐up: after treatment	4 studies found a higher amount of expectorated secretions during ACTs compared to control.   1 study found no difference between groups.		—	47 (5)	⊕⊝⊝⊝ Very lowd,e	We were unable to enter any data into our analyses and we have reported narratively from the original papers. There were minimal data available to determine the true effect.
Exercise tolerance (MSWT)   Follow‐up: 12 months	No studies reported exercise tolerance.					—
Well‐being or QoL   Follow‐up: 12 months	No studies reported QoL.					—
*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% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).  ACT: airway clearance technique; AD: autogenic drainage; CI: confidence interval; FET: forced expiration technique; FEV1: forced expiratory volume in 1 second; HAT: hydroacoustic therapy; HFCWO: high‐frequency chest wall oscillation; LCI: lung clearance index; MD: mean difference; MSWT: modified shuttle walk test; QoL: quality of life; PEP: positive expiratory pressure; SD: standard deviation.
GRADE Working Group grades of evidence High certainty: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate certainty: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low certainty: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low certainty: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect.

^aInterventions were: PEP breathing alone (3 studies), with 1 of these varying the water pressure between the 2 active treatment arms (5 cmH₂O and 15 cm cmH₂O); PEP breathing alone followed by AD and AD followed by PEP breathing alone (1 study); PEP breathing with FET (5 studies); PEP with manual external chest compression by the physiotherapist (1 study); oscillating PEP (2 studies, 1 of which specified using Flutter); Flutter with FET (1 study); postural drainage alone (1 study); postural drainage in combination with FET (2 studies); postural drainage with vibrations, deep breathing, percussion, and FET (1 study); postural drainage with percussion and vibrations (2 studies); postural drainage with conventional physiotherapy (1 study); HFCWO with FET (1 study); HFCWO with cough (1 study); treadmill exercise (2 studies); AD alone (1 study).
^bControls were: resting breathing (3 studies); waiting without chest physiotherapy before being advised to cough (1 study); resting breathing with spontaneous coughing (6 studies); directed coughing (2 studies); placebo form of HAT (sitting in a bath with audible sounds, but without delivery of the acoustic waves thought to provide external thoracic oscillation therapy; 1 study).
^cDowngraded twice due to imprecision caused by low participant numbers and no SDs or CIs of the difference between groups.
^dDowngraded twice due to uncertain risk of bias across most or all domains in the included studies.
^eDowngraded once due to imprecision caused by very small sample size.

Background

Description of the condition

Cystic fibrosis (CF) is a common inherited life‐limiting disorder. Persistent infection and inflammation within the lungs are the major contributory factors to severe airway damage and loss of respiratory function over the years (Cantin 1995; Konstan 1997). Reduced hydration at the airway surface and dehydrated airway mucus disrupts the normal mucociliary transport mechanism and thereby leads to airway obstruction and mucus plugging (Zach 1990). For this reason, removal of airway secretions is an integral part of the management of CF. A variety of methods are used to help remove secretions from the lungs: some are physical, (i.e. airway clearance techniques (ACTs), traditionally referred to as chest physiotherapy). and some are chemical (i.e. medications and inhalation therapies). Treatment methods that improve mucus clearance are considered essential for optimising respiratory status and reducing the progression of lung disease.

Whilst CF remains life‐limiting, since the early 2010s, there has been a step‐change in the management of the disease with the development of highly effective cystic fibrosis transmembrane conductance regulator (CFTR) modulators, which target the underlying protein defect. Restoring CFTR function improves ion transport, correcting airway dehydration, reducing airway inflammation and infection in those who have responsive genetic mutations (Hisert 2017). Early clinical experience shows that these drugs significantly reduce sputum volume and viscosity and improve clearance of secretions in many individuals.

Description of the intervention

Prior to the introduction of CFTR modulators, people with CF began ACTs immediately after diagnosis and were advised to use them long‐term as part of daily standard care. Conventional chest physiotherapy involves gravity‐assisted postural drainage combined with percussion and vibration (usually requiring the assistance of a physiotherapist or relative) and forced expirations (i.e. huffing or coughing). These techniques are unpleasant, uncomfortable, and time‐consuming. Improvement in CF survival has led to a greater need for independence with treatments. Additionally, a better understanding of the physiological requirements necessary for effective airway clearance has led to the development of several self‐administered alternatives to conventional ACTs. Alternative ACTs can be performed in a sitting position or modified positioning and include the active cycle of breathing technique (ACBT), forced expiration technique (FET), autogenic drainage (AD), positive expiratory pressure (PEP), oscillatory PEP (Acapella, Aerobika, Cornet, Flutter, Quake), high‐frequency chest wall oscillation (HFCWO), and exercise. The Types of interventions section provides a definition of each method. Not only do these treatment methods help give people with CF more independence in managing their condition, but they also provide clinicians with different options depending on age, severity of disease, airway pathophysiology, individual preference, and likely adherence. 

How the intervention might work

Respiratory infections are the primary cause of morbidity and mortality in CF, and ACTs have played an important role in loosening and assisting the mucociliary transport of secretions proximally within the airways of people with CF. Airway clearance is important because it helps to prevent thick mucus from building up and obstructing distal airways, which in turn reduces the infective burden and associated inflammatory effects on the airway epithelia.

Conventional chest physiotherapy techniques rely on gravity through positioning and the additional mechanical effects of percussion and vibrations to enhance the mobilisation of secretions. More recent ACTs are based on two physiological premises: the ability to ventilate behind obstructed regions of the lung, and the ability to modulate expiratory airflow to achieve the minimum expiratory airflow bias necessary to mobilise secretions proximally (Mcllwaine 2017).

ACBT achieves airflow behind mucus obstruction during thoracic expansion exercises through the use of collateral ventilation and interdependence. These techniques use the equal pressure point during forced expiratory manoeuvres to generate expiratory airflow to mobilise secretions proximally. Periods of breathing control are interspersed throughout the cycle to minimise bronchospasm (Pryor 2019).

The AD technique adjusts rate, depth, and location of respiration whilst modulating inspiratory and expiratory airflow. This targets airflow behind secretions through collateral ventilation and pendelluft flow, and mobilises secretions proximally by using shearing forces generated by expiratory airflow, whilst minimising airway resistance (Mcllwaine 2014; Schoni 1989). 

It is possible to deliver PEP therapy either via a facemask or a mouthpiece. This generates pressures of 10 cmH₂O to 20 cmH₂O, improving airway clearance by increasing gas pressure behind mucus via collateral ventilation (Martin 1966), and temporarily increasing functional residual capacity (Groth 1985). The device is then removed and forced expiratory manoeuvres are utilised to mobilise secretions centrally. One alternative PEP technique is called high‐pressure PEP (Hi‐PEP), which uses pressures of 40 cmH₂O to 100 cmH₂O and may stabilise the airways by splinting them open during expiration and eliciting cough through the mask to clear secretions (Oberwaldner 1986). 

Oscillatory PEP devices have a similar physiological basis to PEP, allowing air to move behind secretions through collateral ventilation, with the addition of oscillations interrupting the expiratory airflow and causing vibration within the airway. This mechanical effect reduces the viscoelastic and spinnability properties of mucus, thereby improving mucus clearance (App 1998).

In HFCWO, oscillations external to the chest wall alter the rheology of mucus similarly to oscillatory PEP, but also enhance secretion clearance by creating an expiratory flow bias (King 1983) and increasing ciliary beat frequency (Hansen 1994). 

Studies have shown that exercise in people with CF improves the rate of mucus clearance from intermediate and peripheral lung regions and eases the expectoration of secretions by increasing ventilation and respiratory flow (Dwyer 2019; Dwyer 2011). 

Why it is important to do this review

The treatments recommended as standard care for managing CF are complex and time‐consuming. People with CF, their carers, and healthcare professionals have described airway clearance as one of the most burdensome therapies (Davies 2020; Sawicki 2009). The top research question identified in a priority setting exercise with the James Lind Alliance (JLA) was "What are the effective ways to simplify the treatment burden for people with CF?" (Rowbotham 2018). In this new era of CFTR modulator therapies, the associated benefits of this treatment to lung health, and the likely significant change in disease trajectory, it remains as important as ever to evaluate the evidence for airway clearance to help simplify treatments for people with CF. 

A series of Cochrane Reviews have evaluated the benefits of ACTs for people with CF. These address conventional physiotherapy (Main 2005), ACBT (McKoy 2016), AD (Burnham 2021), PEP therapy (McIlwaine 2019), oscillatory PEP (Morrison 2020), and exercise (Heinz 2022); no reviews have identified any single technique as being more or less efficacious than another. Updated versions of the conventional physiotherapy review and the ACBT review will be published later in 2023. This review is part of the ACT series and aims to determine the efficacy of all ACT interventions compared to no treatment or spontaneous coughing alone. It is the latest version of a review first published in 2000 and subsequently updated in 2013 and 2015 (van der Schans 2000; Warnock 2013; Warnock 2015).

Objectives

To determine the effectiveness and acceptability of airway clearance techniques compared to no airway clearance techniques or cough alone in people with cystic fibrosis. 

Methods

Criteria for considering studies for this review

Types of studies

Randomised controlled trials (RCTs) or quasi‐RCTs. We planned to analyse short‐term studies (less than seven days duration, including single‐treatment studies) separately from studies of longer duration.

Types of participants

People with CF, of any age, diagnosed on the basis of clinical criteria and sweat testing or genotype analysis.

Types of interventions

We compared ACTs to no ACTs or spontaneous coughing alone.

In existing literature and in practical terms, variation occurs in the application of specific techniques. For the purposes of the series of ACT reviews published by the Cochrane Cystic Fibrosis and Genetic Disorders (CFGD) Review Group, it is necessary to group these variations under their broader headings, as separate analysis of each variation would render the reviews unmanageable. We included the following interventions (where study authors described them as the primary intervention), with or without additional techniques. Two review authors independently categorised the physiotherapeutic interventions.

Conventional chest physiotherapy

This includes any combination of the following: postural drainage or modified postural drainage, percussion, chest vibrations, huffing, and directed coughing. It does not include exercise, PEP, or other mechanical devices.

Positive expiratory pressure (PEP) mask therapy

PEP is defined as breathing between 10 and 12 breaths through a PEP mask or mouthpiece, generating PEP of 10 cmH₂O to 25 cmH₂O, followed by expiration manoeuvres including cough to expectorate any mucus cleared.

High‐pressure PEP (Hi‐PEP) mask therapy

This is a modification of the PEP technique described above and includes a full‐forced expiration against a fixed mechanical resistance generating pressures of 40 cmH₂O to 100 cmH₂O.

Active cycle of breathing technique (ACBT)

ACBT includes breathing control, thoracic expansion exercises, and FET. 

Autogenic drainage (AD)

As described originally by Chevaillier in 1967 or modified versions thereof (Chevaillier 1984). This technique is characterised by breathing control aimed at achieving high expiratory airflows whilst keeping resistance in the airways as low as possible. By adjusting the depth and speed of respiration to target secretions and mobilise them proximally, people with CF can clear secretions independently with one or two FETs or coughs.

Airway oscillating devices

Airway oscillating devices include Acapella, Aerobika, Cornet, Flutter, and Quake. They generate an interrupted PEP resulting in repeated accelerations of expiratory airflow.

Mechanical percussive devices and external high‐frequency chest wall oscillatory devices

Mechanical percussive and external HFCWO devices such as The Vest and Hayek Oscillator provide external chest wall compressions. 

Exercise

In eligible RCTs, both aerobic exercise and strength training should have the sole purpose of improving mucus clearance as the primary intervention. 

Other adjunctive techniques

Additional interventions are widely used in clinical practice to augment existing ACTs. These include, but are not limited to, intermittent positive pressure breathing, non‐invasive ventilation, and mechanical insufflation‐exsufflation. Because these interventions constitute adjunctive therapies rather than primary interventions, we considered them to be outside the scope of this review.

Types of outcome measures

Primary outcomes

1. Pulmonary function tests (change from baseline)

   1. Forced expiratory volume in one second (FEV₁)

   2. Forced vital capacity (FVC)

   3. Lung clearance index (LCI)

2. Respiratory exacerbations

   1. Number of respiratory exacerbations per year

   2. Number of days in hospital per year

   3. Number of days of intravenous antibiotics per year 

Secondary outcomes

1. Other measures of pulmonary function (change from baseline)

   1. Total lung capacity (TLC)

   2. Functional residual capacity (FRC)

   3. Forced expiratory flow between 25% and 75% expired FVC (FEF_25-75)

2. Clearance of expectorated secretions (mucus, sputum, phlegm)

   1. Mucus transport rate (assessed by radioactive tracer clearance)

   2. Expectorated secretions (mucus, sputum, phlegm)*

      1. Dry weight 

      2. Wet weight

      3. Volume

3. Oxygen saturation (SpO₂; measured by pulse or transcutaneous oximetry)

4. Radiological ventilation scanning

5. Exercise tolerance (as measured by any standard exercise tests, e.g. cardiopulmonary exercise testing (CPET), modified shuttle walk test (MSWT), six‐minute walk test (6MWT), cycle ergonometry)

6. Well‐being or quality of life (QoL)

7. Participant preference

8. Adherence 

9. Nutritional status

   1. Height

   2. Weight

   3. Body composition

10. Adverse events (haemoptysis, pneumothorax, death)

*An increase in the amount of expectorated secretions as a short‐term effect of the intervention is considered beneficial.

Search methods for identification of studies

We applied no restrictions regarding language or publication status.

Electronic searches

We identified relevant studies from the CFGD Group's Cystic Fibrosis Trials Register using the term "airway clearance techniques".

The Cystic Fibrosis Trials Register is compiled from electronic searches of the Cochrane Central Register of Controlled Trials (CENTRAL; updated with each new issue of the Cochrane Library), weekly searches of MEDLINE, a search of Embase to 1995, and prospective handsearching of two journals: Pediatric Pulmonology and Journal of Cystic Fibrosis. Unpublished work is identified by searching the abstract books of three major CF 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, see the relevant sections of the CFGD Group's website.

Date of most recent search of the Group's Trials Register: 17 October 2022.

We also searched two online clinical trial registries: ClinicalTrials.gov (clinicaltrials.gov) and the WHO International Clinical Trials Registry Platform (ICTRP; trialsearch.who.int). See Appendix 1 for the search strategies used.

Date of latest search of trials registries: 7 November 2022.

Searching other resources

We reviewed the reference lists of the included studies and of other Cochrane publications relevant to this review. In addition, we contacted researchers to obtain additional information on relevant studies.

Data collection and analysis

Selection of studies

For the original review, two review authors from different centres independently assessed which studies to include. In the event of disagreement about inclusion of a study, they asked an independent author from a third centre to review the paper(s) in question. For updates since 2013, two new review authors (AG, LW) from the same centre have independently assessed studies for inclusion in the review. In case of any disagreement, they asked the review's contact editor to arbitrate.

Data extraction and management

Each review author independently extracted data on the outcomes listed in Types of outcome measures. We planned to use Cochrane Review Manager software to compile and analyse the data, but were only able to present a narrative summary (Review Manager 2014). For any unreported outcomes, we contacted the study authors to request additional data.

We planned to present short‐term studies (defined as having a duration of seven days or less) separately from longer‐term studies. We planned to group outcome data from longer‐term studies (more than seven days) into those measured at one month, three months, six months, 12 months, and every year thereafter. If studies had recorded outcome data at other time periods, we would have considered examining these as well.

For this update, we received individual participant data (IPD) from the authors of one study (Dwyer 2017). The CFGD Group's statistician prepared the data for analysis in Review Manager 5 (Review Manager 2014).

Assessment of risk of bias in included studies

For the original review, the review authors independently assessed the methodological quality of the included studies using a system described by Jadad and colleagues (Jadad 1996). In the event of disagreement about the quality score, they asked an independent author from a third centre to review the paper(s) in question. The review authors considered aspects such as generation of randomisation sequence, allocation concealment, degree of blinding, and whether data were reported completely.

From the 2013 review update onwards, we assessed the risk of bias of the included studies using the Cochrane risk of bias tool RoB 1 (Higgins 2017). Specifically, we judged each study as being at high, low, or unclear risk of bias for the following domains: sequence generation, allocation concealment, blinding (of participants, clinicians, and outcome assessors), incomplete outcome data, selective reporting, and other potential sources of bias.

Measures of treatment effect

For binary outcomes (adverse events), we planned to calculate odds ratios (ORs) and their 95% confidence intervals (CIs). For continuous outcomes (all other primary and secondary outcomes), we planned to record either the mean change from baseline for each group or mean post‐treatment or intervention values, and the standard deviation (SD) or standard error (SE) for each group.

Unit of analysis issues

We planned to analyse the data from cross‐over studies as recommended by Elbourne and colleagues; however this was not possible with the available data (Elbourne 2002). Instead, we described these data narratively.

Dealing with missing data

To enable an intention‐to‐treat analysis, we collected data on the number of participants with each outcome event by allocated group irrespective of compliance and whether the participant was later thought to be ineligible or otherwise excluded for treatment or follow‐up. Where there was evidence of missing data, we contacted the primary investigator for clarification.

Assessment of heterogeneity

When we are able to include a sufficient number of studies in a meta‐analysis, we plan to assess heterogeneity using the I²statistic (Higgins 2003). This measure describes the percentage of total variation across studies that is due to heterogeneity rather than chance (Higgins 2003). We will interpret the I²statistic according to the following thresholds, as suggested in the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2022).

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*

*The importance of the observed value of I² depends on magnitude and direction of effects and on strength of evidence for heterogeneity (e.g. P value from the Chi² test or a CI for I²).

Assessment of reporting biases

We assessed all included studies for potential reporting bias, including missing outcome values and relationships with sponsors.

We searched for the study protocols and compared these with the final publications to ensure all measured outcomes were reported. If a study protocol was unavailable, we compared the methods and results sections of each final publication to identify any discrepancies in outcome reporting.

Data synthesis

We were only able to present very limited data and were unable to generate any meta‐analyses for this review update. However, if we are able to perform meta‐analyses in future updates, we will combine the data using a fixed‐effect model if there is little or no heterogeneity (i.e. I²value below 50%). If the I² value is equal to or greater than 50%, we plan to use a random‐effects model.

Subgroup analysis and investigation of heterogeneity

If we identify a high degree of heterogeneity (i.e. I²value equal to or greater than 75%), we plan to investigate this by evaluating the pooled results of studies with long‐term interventions versus those with short‐term interventions. A series of separate physiotherapy reviews published by the Cochrane CFGD Group have included subgroup analyses examining the effects of specific interventions (Burnham 2021; Main 2005; McIlwaine 2019; McKoy 2016; Morrison 2020). 

Sensitivity analysis

If we are able to include and analyse more data in future updates of this review, we plan to test the robustness of our results by performing a sensitivity analysis of the data, comparing results with and without quasi‐randomised studies.

Summary of findings and assessment of the certainty of the evidence

In accordance with current Cochrane guidance, we included a summary of findings table in this update. We chose the following seven outcomes, which we consider the most important.

1. Lung function: FEV₁ % predicted change from baseline (at 12 months)

2. Lung function: LCI (after the intervention)

3. Exacerbation rate, assessed by number of respiratory exacerbations per year

4. Mucus transport rate, assessed by radioactive tracer clearance (after the intervention) 

5. Expectorated secretions, assessed by sputum wet weight (after the intervention)

6. Exercise tolerance, assessed by MSWT (at 12 months) 

7. Well‐being or QoL (at 12 months)

We used the GRADE approach to assess the certainty of the evidence for each outcome based on the risk of bias within the studies, relevance to our population of interest (indirectness), unexplained heterogeneity or inconsistency, imprecision of the results, and publication bias. We downgraded the evidence once for serious concerns and twice for very serious concerns.

Results

Description of studies

For additional information, see the Characteristics of included studies table, Characteristics of excluded studies table, and Characteristics of ongoing studies table.

Results of the search

The searches prior to 2018 identified 157 studies of ACTs. The review authors excluded 149 studies and included eight studies.

Further searches of the CFGD Group's CF Trials Register for this update identified an additional 179 citations. The search of the study registries identified 229 studies. This gave a total of 408 citations.

Following deduplication and removal of citations that were clearly irrelevant, we examined the full‐text articles of 23 potentially eligible studies (50 citations). We included four additional studies (seven citations) in this update (Dwyer 2017; Dwyer 2019; Grosse‐Onnebrink 2017; Vandervoort 2022). We excluded 17 new studies (ACTRN12605000535673; Corten 2020; Dacie 2015; Hansen 1990; Helper 2020; Horsley 2007; Hristara‐Papadopoulou 2007; Leemans 2020; McIlwaine 2001; McIlwaine 2014; NCT01266473; NCT03655249; NCT04010253; San Miguel Pagola 2020; Stanford 2019a; Varekojis 2003; Walicka‐Serzysko 2021) and identified two ongoing studies (six citations) that are still in the recruitment phase (NCT03760120; Stanford 2019b).

Figure 1 is a PRISMA diagram showing the flow of studies for this update.

1.

Study selection process for 2023 update.

We re‐evaluated the excluded studies list and removed 32 studies that we considered clearly irrelevant to this review, leaving 134 excluded studies.

Included studies

The latest version of this review includes 12 studies with a total of 194 participants (Braggion 1995; Dwyer 2017; Dwyer 2019; Elkins 2005; Falk 1993; Grosse‐Onnebrink 2017; Jarad 2010; Mortensen 1991; Pfleger 1992; Rossman 1982; van der Schans 1991; Vandervoort 2022).

Study design

One study (41 participants) had a parallel design (Grosse‐Onnebrink 2017), and 11 studies (153 participants) had a cross‐over design (Braggion 1995; Dwyer 2017; Dwyer 2019; Elkins 2005; Falk 1993; Jarad 2010; Mortensen 1991; Pfleger 1992; Rossman 1982; van der Schans 1991; Vandervoort 2022).

Two studies compared one active therapy to a control group with no intervention (Grosse‐Onnebrink 2017; Vandervoort 2022). Six studies compared two active therapies to control (Dwyer 2017; Dwyer 2019; Falk 1993; Jarad 2010; Mortensen 1991; van der Schans 1991). However, Jarad 2010 included one intervention group that did not receive a recognised ACT (hydro‐acoustic therapy (HAT)); we excluded this treatment arm from the review. One study compared three active therapies to control (Braggion 1995); and the remaining three studies compared four active therapies to control (Elkins 2005; Pfleger 1992; Rossman 1982).

Most studies delivered single daily treatments (Dwyer 2017; Dwyer 2019; Elkins 2005; Falk 1993; Grosse‐Onnebrink 2017; Mortensen 1991; Pfleger 1992; Rossman 1982; van der Schans 1991; Vandervoort 2022). Braggion 1995 administered therapy twice daily for two days, and Jarad 2010 repeated each intervention on two successive days.

Participants

Sample sizes ranged from six (Rossman 1982) to 41 (Grosse‐Onnebrink 2017). The mean age of participants ranged from seven years (Grosse‐Onnebrink 2017) to 51 years (Grosse‐Onnebrink 2017). Three studies did not report the sex of participants (Elkins 2005; Falk 1993; van der Schans 1991). One study included only males (Rossman 1982), six studies enrolled more males than females (Dwyer 2017; Dwyer 2019 Grosse‐Onnebrink 2017; Jarad 2010; Mortensen 1991; Vandervoort 2022), one study had equal numbers of males and females (Braggion 1995), and one study had more females than males (Pfleger 1992). The studies included people with a wide range of disease severity. In the eight studies reporting baseline mean FEV₁ % predicted, values ranged from 51% (SD 18%) in Dwyer 2017 to 94% (SD 16.8%) in Vandervoort 2022). Mortensen 1991 enrolled participants with median FEV₁ % predicted of 38.5% (range 26% to 101%). Rossman 1982 reported a range of FEV₁ % predicted of 12% to 77.7%. Jarad 2010 stated that all participants had FEV₁ % predicted of less than 80%; however, the study authors only reported mean baseline values of FEV₁ in litres, so the disease severity of those participants is not comparable with that of the other studies. Falk 1993 did not provide details on severity of disease of participants.

Vandervoort 2022 included a mixed population, recruiting people with a diagnosis of either primary cilia dyskinesia or CF; as the study authors reported the CF data separately in a subgroup analysis, we were able to include them in this review.

Interventions

The active interventions varied greatly across the included studies, with several studies having multiple treatment arms. 

Three studies assessed PEP breathing alone (Elkins 2005; Pfleger 1992; van der Schans 1991) with one of these varying the water pressure between the two active treatment arms (5 cmH₂O and 15 cm cmH₂O; van der Schans 1991). Pfleger 1992 also assessed PEP breathing followed by AD and AD followed by PEP. Five studies included an active group combining PEP breathing with FET (Braggion 1995; Dwyer 2019; Falk 1993; Mortensen 1991; Vandervoort 2022). Vandervoort 2022 combined PEP with manual external chest compression by the physiotherapist. Two studies assessed oscillating PEP alone (Elkins 2005; Jarad 2010); Jarad 2010 specified using a specific type of oscillating PEP (Flutter). One study combined Flutter with FET (Dwyer 2017). Five studies described assessing postural drainage (Braggion 1995; Elkins 2005; Falk 1993; Mortensen 1991; Rossman 1982), but only one of these assessed the technique alone (Rossman 1982). The studies also assessed postural drainage in combination with at least one further technique; two studies combined postural drainage with FET (Falk 1993; Mortensen 1991), one combined postural drainage with vibrations, deep breathing, percussion, and FET (Braggion 1995), two combined postural drainage with percussion and vibrations (Elkins 2005; Rossman 1982), and one with several ACTs grouped together as conventional physiotherapy (Rossman 1982). Two studies assessed HFCWO, one in combination with FET (Braggion 1995) and one in combination with cough (Grosse‐Onnebrink 2017). Two studies assessed exercise as an ACT in the form of 20 minutes of treadmill exercise (constant work rate equivalent to 60% peak oxygen consumption (VO₂)) (Dwyer 2017; Dwyer 2019). Only one study assessed AD (Pfleger 1992).

The control interventions also varied across studies. In three studies, the control was resting breathing (Dwyer 2017; Dwyer 2019; Vandervoort 2022), whilst Grosse‐Onnebrink 2017 described the control as waiting without chest physiotherapy before being advised to cough. The control involved resting breathing with spontaneous coughing in six studies (Braggion 1995; Elkins 2005, Falk 1993; Mortensen 1991; Pfleger 1992; Rossman 1982) and directed coughing in two studies (Rossman 1982; van der Schans 1991); Rossman 1982 had separate groups for spontaneous and directed cough. The control intervention in Jarad 2010 was a placebo form of HAT, which involved sitting in a bath with audible sounds, but without delivery of the acoustic waves thought to provide external thoracic oscillation therapy.

Outcome measures

The most common outcome, measured by six included studies, was radioactive tracer clearance (Dwyer 2019; Elkins 2005; Falk 1993; Mortensen 1991; Rossman 1982; van der Schans 1991). Five studies assessed sputum weight (Braggion 1995; Jarad 2010; Mortensen 1991; Pfleger 1992; Rossman 1982), two of which specified both wet and dry sputum weight (Braggion 1995; Jarad 2010). Five studies reported pulmonary function tests (Braggion 1995; Jarad 2010; Pfleger 1992; van der Schans 1991; Vandervoort 2022). Four of these reported FEV₁ and FVC (Braggion 1995; Jarad 2010; Pfleger 1992; Vandervoort 2022); Braggion 1995 and Jarad 2010 also reported FEF_25-75. Other reported pulmonary function tests were forced expiratory flow at 75% (FEF₇₅; Jarad 2010); residual volume as a fraction of TLC (RV/TLC) and airway resistance (Raw; Pfleger 1992); and TLC and FRC (van der Schans 1991). Two studies assessed LCI (Grosse‐Onnebrink 2017; Vandervoort 2022). Finally, four studies reported on participants' subjective assessment of the interventions (Braggion 1995; Dwyer 2017; Dwyer 2019; Jarad 2010).

Dwyer 2017 investigated changes in sputum properties following the intervention, namely sputum solids content (from which inferences of airway hydration are made) and sputum mechanical impedance. Dwyer 2017 also reported peak expiratory flow and peak expiratory to peak inspiratory airflow bias, and Dwyer 2019 reported cough frequency. These were not outcomes of interest for this review.

Excluded studies

Of the 134 excluded studies, 129 lacked a 'no treatment' or 'spontaneous coughing' control group. We excluded the remaining five studies for other reasons: Fauroux 1999 did not evaluate chest physiotherapy; ACTRN12605000535673 was an unpublished PhD thesis and the author was unable to provide any results to allow inclusion; Dacie 2015 and Horsley 2007 were published in abstract form only with insufficient detail on study design or outcomes, and we received no reply to our repeated requests for further information (if we receive responses in the future, we will reconsider this decision); and we excluded ACTRN12605000348651 after receiving confirmation from the study investigators that the trial was never completed.

Ongoing studies

Two studies identified through online searches of the clinical trials databases may meet our eligibility criteria. Stanford 2019b is registered as completed, but the lead investigator was unable to share results ahead of publication. NCT03760120 is listed as active but not recruiting

Stanford 2019b is a cross‐over RCT investigating outcome measures used in trials of ACTs and comparing single visits of no ACT to ACBT. The target sample size is 68 participants aged 16 years and over, both males and females. Investigators measured impulse oscillation system (IOS), LCI and spirometry before and after the resting session or ACBT. They also collected sputum throughout each session and for 30 minutes after completion of each session, and asked participants to complete questionnaires on the outcome measures at the end of each study visit. The review authors contacted the study investigators in August 2022, and they confirmed that data collection was complete, but they had yet to report.

NCT03760120 is also a cross‐over RCT, but investigating single sessions with a standard PEP mask versus a sham PEP mask to establish the effects on ventilation inhomogeneity in children and adolescents (aged five to 18 years) with CF (as measured by multiple breath washout tests). The study will also measure sputum weight and SpO₂. The target sample size was 18 but the trial registry entry states that 19 participants were randomised.

Risk of bias in included studies

The authors of the original review assessed methodological quality using the Jadad scoring system; these scores are recorded in Table 2. Details of the risk of bias assessments undertaken from the 2013 update onwards follow and are summarised in Figure 2.

1. Jadad scores for methodological quality (from the original review).

Study	Score
Braggion 1995	2
Falk 1993	1
Mortensen 1991	1
Pfleger 1992	2
Rossman 1982	1
van der Schans 1991	1

The maximal score according to Jadad is 5; however, 2 items are related to blinding of the investigator, and since blinding of the investigator is impossible in the case of airway clearance techniques, the maximal possible score for these studies is 3.

2.

Allocation

All 12 studies were described as randomised, but only three provided any details on the method of randomisation: Braggion 1995 used Latin square design, while Dwyer 2017 and Dwyer 2019 used computer‐generated randomisation. We judged these three studies at low risk of bias related to random sequence generation, and we judged the remaining studies at unclear risk of bias (Elkins 2005; Falk 1993; Grosse‐Onnebrink 2017; Jarad 2010; Mortensen 1991; Pfleger 1992; Rossman 1982; van der Schans 1991; Vandervoort 2022).

Three studies discussed allocation concealment: in Dwyer 2017 and Dwyer 2019 a person not involved in the interventions performed the randomisation and stored the result in sealed sequentially‐numbered opaque envelopes; and Grosse‐Onnebrink 2017 also described using sealed envelope assignment. We judged these three studies at low risk of bias related to allocation concealment and the remaining nine studies at unclear risk of bias (Braggion 1995; Elkins 2005; Falk 1993; Jarad 2010; Mortensen 1991; Pfleger 1992; Rossman 1982; van der Schans 1991; Vandervoort 2022).

Blinding

It is impossible to blind participants and caregivers or clinicians to physiotherapy interventions, but it is possible to blind the outcome assessors to the intervention. The authors of Mortensen 1991 described the study as single‐blind, which we assume refers to the outcome assessors being blinded. Dwyer 2017 and Dwyer 2019 clearly state that outcome data were analysed by blinded assessors. All outcome measures except well‐being and adherence to therapy are physiological data, so we do not consider the fact that participants, caregivers, or clinicians were not blinded as an important source of bias. We considered all studies at unclear risk of bias for this domain.

Incomplete outcome data

Seven studies reported dropouts or reported data for all participants, and we considered them to be at low risk of bias (Dwyer 2017; Dwyer 2019; Grosse‐Onnebrink 2017; Jarad 2010; Mortensen 1991; Pfleger 1992; Vandervoort 2022). Only one study reported data for all randomised participants (Mortensen 1991). In Dwyer 2017, one participant withdrew after the baseline visit and before the active interventions commencing, but the study authors did not give a reason. In Dwyer 2019, one participant withdrew after visit 2 due to a possible allergic reaction to the radioaerosol. In Grosse‐Onnebrink 2017, three participants (two in the intervention group and one in the control group) failed to meet quality control criteria for LCI testing, so did not complete the study and were excluded from all analyses. Jarad 2010 reported that one participant withdrew from the study due to time constraints. Pfleger 1992 excluded one participant due to respiratory infection. Vandervoort 2022 reported two excluded participants (one failed to attend and the second received a bronchodilator outside the study protocol), with two further participants being excluded from the analysis of the multiple breath washout results due to insufficient test repeatability. 

The remaining five studies made no reference to any dropouts (Braggion 1995; Elkins 2005; Falk 1993; Rossman 1982; van der Schans 1991) and we considered the risk of bias to be unclear. As all the studies measured treatment effects immediately after the intervention, we do not consider dropouts an important risk of bias.

Selective reporting

To assess selective reporting, we planned to compare the study protocols with each final publication. However, we were only able to obtain a study protocol for two studies (Dwyer 2017; Jarad 2010). Dwyer 2017 reported all outcomes prespecified in the protocol. Although lung function assessed by spirometry was only reported at baseline in the publication, the study authors subsequently shared these data for this Cochrane Review. The final publication from Dwyer 2017 also reported the additional outcome of cough frequency, which was not included in the published protocol. We judged this study at low risk of reporting bias. The protocol for Jarad 2010 stated secondary outcome data would be collected on SpO₂, respiratory rate, heart rate, and blood pressure during the interventions, and this was confirmed in the methods section of the published paper; however, the published paper made no further reference to these variables in the results or discussion sections, suggesting selective reporting. Consequently, we rated Jarad 2010 at high risk of reporting bias.

Eight studies without protocols were published in full, and we identified no discrepancies between the methods and results sections of the publications that would raise suspicion of selective reporting (low risk of bias; Braggion 1995; Dwyer 2019; Grosse‐Onnebrink 2017; Mortensen 1991; Pfleger 1992; Rossman 1982; van der Schans 1991; Vandervoort 2022). One of these reported data on expected outcomes; although the study authors only reported on baseline spirometry despite stating in the methods section that spirometry would be performed after completion of multiple breath washout tests (Grosse‐Onnebrink 2017). However, we did not consider this discrepancy to represent selective reporting and still judged the study at low risk of reporting bias.

Two studies were only available in abstract form, with insufficient information to assess selective reporting; we judged them at unclear risk of reporting bias (Falk 1993; Elkins 2005).

Other potential sources of bias

Five studies reported a source of funding (Dwyer 2017; Dwyer 2019 Grosse‐Onnebrink 2017; Mortensen 1991; Vandervoort 2022). However, as the interventions being studied in the remaining studies either do not require equipment or use widely available equipment, we did not consider that potential funding sources represented a significant risk of bias.

Jarad 2010 included sputum wet and dry weight as an outcome, but did not report the unit of measurement in either the protocol or published study. If we assume the measurement is in grams, then the participants appear to be non‐ or low‐sputum producers (sputum wet weight ranging from 0.0 g to 5.3 g during the interventions analysed in this review). This would likely have impacted on treatment efficacy and the ability to detect a difference between control and active treatment groups.

The efficacy of any physiotherapy technique may be influenced by the proficiency and familiarity of the individual with that technique. Therefore, naivety of individuals to some but not all interventions being studied may introduce a potential source of bias. Seven studies did not describe how experienced participants were with each intervention (Braggion 1995; Elkins 2005; Falk 1993; Mortensen 1991; Rossman 1982; van der Schans 1991; Vandervoort 2022). In Dwyer 2017, participants were taught to use the Flutter at visit 1, and those already using the device on a regular basis had their technique corrected if necessary; the report did not state how many participants had never used the Flutter before the start of the study. Similarly, Dwyer 2019 stated that participants were taught to use the PEP at visit 1, and those already familiar with the device had their technique corrected as necessary; however, the report did not specify the number of participants who had never used PEP before the start of the study. In Jarad 2010, 4/18 participants used the Flutter as their usual ACT, while the remainder of participants were naive to this treatment. Only Pfleger 1992 stated that all participants were trained in the techniques being studied during the six‐month period preceding commencement of the study.

In the Vandervoort 2022 study, there was a six‐month gap between the interventions. The study authors identified this as a limitation, stating that the gap was to reduce the study burden for the participants. They reported that baseline characteristics were stable between the study visits, but it is unclear whether this was the case for each participant.

Effects of interventions

See: Table 1

We included 11 short‐term cross‐over studies (less than seven days) and one study with a parallel design (Grosse‐Onnebrink 2017). No meta‐analysis was possible. Ten studies were single‐treatment studies (Dwyer 2017; Dwyer 2019; Elkins 2005; Falk 1993; Grosse‐Onnebrink 2017; Mortensen 1991; Pfleger 1992; Rossman 1982; van der Schans 1991; Vandervoort 2022); in Jarad 2010, each intervention was performed twice on successive days; and in Braggion 1995, each physiotherapy treatment was delivered four times over two days.

Primary outcomes

1. Pulmonary function tests

Six studies reported pulmonary function measures (Braggion 1995; Grosse‐Onnebrink 2017; Jarad 2010; Pfleger 1992; van der Schans 1991; Vandervoort 2022). In addition, investigators from Dwyer 2017 provided us with unpublished individual participant lung function data. 

a. Forced expiratory volume in one second (FEV₁)

Braggion 1995 reported no significant difference between any of the three interventions or control in FEV₁ measured 30 minutes after the intervention. Jarad 2010 reported a statistically significant reduction in FEV₁ (P = 0.028) in the Flutter group but not in the placebo group; the changes were short‐lived and values returned to baseline on the second study day (each intervention was repeated on two successive days). Pfleger 1992 measured FEV₁ at five different time points during each intervention session (where the intervention varied in a random order from session 2 to 5) and reported means and SDs at the end of the first and fifth intervention session, hence we were unable to enter the data in a meta‐analysis. The study authors found an improvement in FEV₁ from the first to the last time‐point analysis in the PEP group (P < 0.05) and PEP‐AD group (P < 0.02), but not in the control, AD, or AD‐PEP groups. van der Schans 1991 reported no differences in FEV₁ measures for the treatment or control, although some data were missing for this variable. Vandervoort 2022 also reported no differences in FEV₁ % predicted between the PEP and control group.

We analysed unpublished data from Dwyer 2017, which showed no difference in FEV₁ in either the Flutter group (mean difference (MD) 0.02 L, 95% CI −0.04 to 0.08; 1 study, 16 participants; Analysis 1.1) or the exercise group (MD 0.19 L, 95% CI −0.09 to 0.47; 1 study, 16 participants; Analysis 1.1) compared to control.

1.1. Analysis.

Comparison 1: Airway clearance techniques (ACTs) versus no ACTs, Outcome 1: FEV₁ (L) – change from baseline

b. Forced vital capacity (FVC)

Braggion 1995 reported no significant difference between any of the three treatments or control in FVC measured 30 minutes after the intervention. Jarad 2010 reported no statistically significant changes in FVC in either the Flutter or control groups. Pfleger 1992 measured FVC at five different time points during each intervention session (where the intervention varied in a random order from session 2 to 5) and reported means and SDs at the end of the first and fifth intervention session, hence we were unable to enter the data in a meta‐analysis. The study authors found an improvement in FVC comparing the first and last time‐point analysis in the PEP group (P < 0.01), the AD group (P < 0.05), and the PEP‐AD group (P < 0.01), but not in the control or AD‐PEP groups. Vandervoort 2022 reported no differences in FVC % predicted between the PEP and control group.

We analysed unpublished data from Dwyer 2017, which showed no difference in FVC in either the Flutter group (MD 0.2, 95% CI −0.10 to 0.50; 1 study, 16 participants; Analysis 1.2) or the exercise group (MD 0.14, 95% CI −0.13, 0.41; 1 study, 16 participants; Analysis 1.2) compared to control.

1.2. Analysis.

Comparison 1: Airway clearance techniques (ACTs) versus no ACTs, Outcome 2: FVC (L) – change from baseline

c. Lung clearance index (LCI)

Two studies used LCI as an outcome measure (Grosse‐Onnebrink 2017; Vandervoort 2022). Grosse‐Onnebrink 2017 reported improvement in 15/20 participants, demonstrated by a decrease in LCI by a median of 0.9 (range −0.45 to 3.47; P = 0.002), with five of these participants reaching the threshold for clinical relevance of the treatment effect. Conversely, LCI increased in 5/20 participants (a deterioration in condition), although none reached the threshold for clinical relevance. There was no difference in LCI during the control period. The study authors concluded that the active ACT (HFCWO) had an impact on LCI values, although this effect was heterogeneous across the population. Vandervoort 2022 found no difference in LCI when comparing PEP to control. 

We rated the evidence for this outcome as being of low certainty (Table 1).

2. Respiratory exacerbations

a. Number of respiratory exacerbations per year

No studies reported data on number of respiratory exacerbations per year.

b. Number of days in hospital per year

No studies reported data on number of days in hospital per year.

c. Number of days of intravenous antibiotics per year

No studies reported data on number of days of intravenous antibiotics per year.

Secondary outcomes

1. Other measures of pulmonary function (change from baseline)

a. Total lung capacity (TLC) 

van der Schans 1991 (eight participants) measured TLC after airway clearance with a PEP device and found no difference following the intervention. The study did observe some improvement in TLC during PEP breathing only, but these quickly returned to baseline immediately after the intervention.

b. Functional residual capacity (FRC)

Three studies (42 participants) measured FRC after airway clearance and found no difference between groups (Grosse‐Onnebrink 2017; van der Schans 1991; Vandervoort 2022). van der Schans 1991 did observe some improvement in FRC during PEP breathing only, but this quickly returned to baseline immediately after the intervention.

c. Forced expiratory flow between 25% and 75% expired forced vial capacity (FEF_25-75)

Two studies (34 participants) measured FEF_25-75 (Braggion 1995; Jarad 2010). Braggion 1995 reported no change in FEF_25-75 measured 30 minutes after the intervention between any of the three treatments or control. Jarad 2010 reported a statistically significant reduction in FEF_25-75 (P = 0.03) following Flutter, but not in the placebo group; the changes were short‐lived and values returned to baseline on the second study day (each intervention was repeated on two successive days).

2. Direct measures of mucus clearance

a. Mucus transport rate (assessed by radioactive tracer clearance)

Six studies reported radioactive tracer clearance (Dwyer 2019; Elkins 2005; Falk 1993; Mortensen 1991; Rossman 1982; van der Schans 1991). Five of these (55 participants) found that ACTs, including coughing, increased radioactive tracer clearance compared to the control period (Dwyer 2019; Elkins 2005; Falk 1993; Mortensen 1991; Rossman 1982). Dwyer 2019 found that both treadmill exercise and PEP increased the percentage of mucus clearance from the whole lung compared to control immediately after the intervention, with PEP being superior to the treadmill (PEP: MD 9.5%, P < 0.001; treadmill: MD 2.6%, P < 0.001). This study looked in detail at clearance from different lung regions, finding that whilst both interventions significantly increased clearance compared to control in peripheral regions (PEP: 8.0 % P < 0.001; treadmill: 6.7 %, P < 0.001) and intermediate regions (PEP: 4.9 %, P < 0.001; treadmill: 4.5%, P < 0.001), only PEP was effective at increasing clearance from the central region (PEP: 13.1%, P < 0.001; treadmill: 0.2%). The study authors considered this may be due to the increased cough frequency in the PEP group associated with the technique. Improvement in mucus clearance was observed primarily during the intervention in both groups, and there was no persistent benefit in the 60‐minute follow‐up period (Dwyer 2019). Elkins 2005 found the mean percentage of radioactivity cleared from each region of interest (% C30) was 8.4% greater during postural drainage with percussion and vibrations compared to control (95% CI 2.4 to 14.5; P = 0.017). Elkins 2005 also reported greater % C30 with PEP, oscillating PEP, and matched cough compared to control, but these did not reach statistical significance. Falk 1993 found approximately 6% clearance during the control measurement and 9% during ACTs. In Mortensen 1991, median clearance after 30 minutes during control was 7%, and median clearance during two different airway clearance sessions was 33% and 34%. Rossman 1982 found 32% radioactive tracer clearance during the control measurement and 40% to 46% during the different forms of ACT.

van der Schans 1991 (eight participants) reported no significant difference between two different airway clearance sessions of PEP breathing without coughing (clearance 10% and 6%) compared to a control period (clearance 8%). One possible reason for different outcomes between van der Schans 1991 and the other studies is that in van der Schans 1991, participants were requested not to cough, but coughing was encouraged in the other studies as a part of the treatment.

We rated the evidence for this outcome as being of very low certainty (Table 1).

b. Expectorated secretions

Five studies reported expectorated secretion weight or volume (Braggion 1995; Jarad 2010; Mortensen 1991; Pfleger 1992; Rossman 1982).

Four studies (46 participants) found a higher amount of expectorated secretions during ACTs compared to the control period (Braggion 1995; Mortensen 1991; Pfleger 1992; Rossman 1982). Braggion 1995 found a mean wet weight of expectorated secretions of 6 g during the control day and 23 g to 30 g during the airway clearance sessions. Mortensen 1991 reported medians and ranges, making comparisons between studies difficult, but did report a significantly larger sputum weight with FET plus PEP (median 8.6 g, range 3.5 to 19.9) and PD plus FET (median 8.0 g, range 2.3 to 13.9) compared to control (median 0.0 g, range 0.0 to 2.1) during the treatment period (P < 0.01). This increase in sputum weight was not sustained during follow‐up. Pfleger 1992 presented the mean (SD) weight of expectorated mucus for each treatment arm, but did not report specific data in the text. Data extracted from the graphs show that during spontaneous coughing, the mean weight of expectorated mucus was approximately 17 g, which was less than during the three forms of ACT (range 34 g to 45 g). Pfleger 1992 also reported that PEP alone produced the highest amount of sputum, followed by a combination of PEP and AD (in either order). AD alone produced the lowest volume of sputum. Rossman 1982 found a significantly higher volume of expectorated secretions during the different forms of ACT compared to the control session.

Jarad 2010 found no differences in wet or dry weight of expectorated sputum between the placebo or Flutter groups; there were no reported P values or absolute values of measurement for sputum weight, making it difficult to draw comparisons with other studies.

Dwyer 2017 did not assess sputum weight or volume, but did report some measures of sputum rheology. The clinical relevance of these measures is unclear, and we decided not to report them.

We rated the evidence for this outcome as being of very low certainty Table 1.

3. Oxygen saturation

No studies reported data on SpO₂.

4. Radiological ventilation scanning

No studies reported data on radiological ventilation scanning.

5. Exercise tolerance

No studies reported data on exercise tolerance.

6. Well‐being or quality of life

No studies reported data on well‐being or QoL.

7. Participant preference

One study included a participant questionnaire to assess the acceptability and preference of the interventions (Jarad 2010). The questionnaire addressed breathlessness during treatment, ease of clearance, relaxation, how pleasant therapy was to perform, and overall preference. The information from the completed questionnaires highlighted participant preference for placebo over Flutter. Half of participants (50.1%) rated the placebo treatment as slightly or much better than the Flutter regarding the ease at which phlegm is coughed up; 35.3% of participants reported that Flutter therapy was slightly or very unpleasant, with none stating that they would prefer to use Flutter for physiotherapy if available (Jarad 2010). When interpreting these results, it is important to consider that the placebo involved the participants sitting in a bath of warm water receiving a sham form of HAT. It is unclear whether this represents a true control, as the effect of sitting in a warm bath on secretion clearance and perceived ease of clearance is unknown.

Dwyer 2017 assessed self‐reported treatment efficacy. Participants rated ease of expectoration on a 10‐cm visual analogue scale, where 0 represented "very difficult to expectorate" and 10 represented "very easy to expectorate". The study found that treadmill exercise significantly improved ease of expectoration compared to control after 20 minutes' recovery (MD 1.3 cm, 95% CI 0.3 to 2.3). There was no difference between Flutter therapy and control for this outcome. In the same study, participants rated the sense of chest congestion on another 10‐cm visual analogue scale, where 0 represented "very congested" and 10 represented "very clear". There were no differences for this outcome between treadmill exercise and control; however, Flutter therapy was associated with a significant improvement in sense of chest congestion compared to control both immediately after the intervention (MD 0.8 cm, 95% CI 0.1 to 1.4) and following a 20‐minute rest period (MD 0.9 cm, 95% CI 0.2 to 1.7). Dwyer 2019 observed similar improvement in sense of chest congestion on the same scale following PEP therapy only, measured immediately after the intervention (MD 1.2 cm, 95% CI 0.1 to 2.3) and after a 60‐minute rest period (MD 1.8 cm, 95% CI 0.9 to 2.7).

8. Adherence

No studies reported data on adherence.

9. Nutritional status 

a. Height 

No studies reported the effect of interventions on participant height.

b. Weight 

No studies reported the effect of interventions on participant weight.

c. Body composition 

No studies reported the effect of interventions on participant body composition.

10. Adverse events

No studies reported data on adverse events.

Discussion

ACTs have been a mainstay of the respiratory management of people with CF for so long that it has been difficult for them, their parents, physiotherapists, and medical staff to consider a study design that incorporates a 'no treatment' control group for any length of time. Despite a reasonable degree of equipoise between physiotherapy and no treatment, many have argued that it would be unethical to recruit participants into a 'no treatment' group. This partly explains why there are no long‐term studies with this design.

With the introduction of new highly effective CFTR modulator treatments, the recommendation of daily airway clearance treatment for all is likely to become an active area of discussion for people with CF and their clinicians. People who may benefit least from daily ACTs include those with established lung damage and bronchiectasis, who may now report little or no sputum production, but also those started early on CFTR modulators, who may not develop lung disease. The populations included in this review were not receiving highly effective CFTR modulators; consequently, our results may not be applicable to people who do use these drugs.

Summary of main results

The heterogeneity of treatments and outcome measures reported in the included studies make it impossible to pool their results. In all 12 studies included in this review, it is unclear whether any of the changes in primary or secondary outcomes were sustained. 

Six studies reported a measure of pulmonary function and found no difference in effect between the treatment and control groups (Braggion 1995; Dwyer 2017; Jarad 2010; Pfleger 1992; van der Schans 1991; Vandervoort 2022). Only Pfleger 1992 observed any significant improvement following some of the interventions (and not following control). Vandervoort 2022 found no difference in LCI with airway clearance, whereas Grosse‐Onnebrink 2017 demonstrated a heterogeneous response. Given the small number of studies and variability in results, we are uncertain of the clinical relevance of LCI as an outcome in assessing airway clearance efficacy. We judged the evidence for the effect of ACTs on LCI to be of low certainty.

Six studies assessed radioactive tracer clearance (Dwyer 2019; Elkins 2005; Falk 1993; Mortensen 1991; Rossman 1982; van der Schans 1991). Five of these found positive effects associated with the active intervention groups compared to control, whereas van der Schans 1991, which included only eight participants, observed no difference (very low‐certainty evidence). The short‐term studies included in this review suggest that ACTs may increase mucus transport in people with CF. Specifically, four of the five studies that reported sputum weight found increased sputum weight with an intervention compared to control (Braggion 1995; Mortensen 1991; Pfleger 1992; Rossman 1982); whereas the remaining study reported no significant differences (Jarad 2010). We judged the evidence for this outcome to be of very low certainty.

Regarding participant preference, it is generally accepted that an individual's satisfaction or preference for an ACT is an important factor in treatment selection when considering likely adherence to treatment. Two studies addressed participant satisfaction, with participants in one study preferring exercise as a method of airway clearance over no active intervention (Dwyer 2017), but participants in the second study preferring the placebo over the intervention (Jarad 2010). In both studies, some of the participants were naive to the active intervention. We consider that reported participant preference from a single‐intervention study, where a new intervention introduces a 'novelty' element, could be misleading. Therefore, our results may not reflect likely satisfaction or preference for a technique carried out in the long term as part of daily management.

No study reported respiratory exacerbations, SpO₂, radiological ventilation scanning, exercise tolerance, well‐being or QoL, adherence, nutritional status, or adverse events.

Overall completeness and applicability of evidence

The UK CF registry data from 2021 shows that people with CF use a wide range of ACTs, with 59.7% using any form of PEP device (including oscillatory) and 59.9% using exercise (UK CF Trust 2021). Of the 12 studies included in this review, 10 evaluated a PEP device and two evaluated exercise as an intervention. Whilst other ACTs were represented in this review (AD and HFCWO), it was not possible to identify any studies comparing the commonly used active cycle of breathing techniques (ACBT) to a 'no treatment' group.

We identified only short‐term studies, where each intervention was repeated over one or two days. Due to the nature of CF, the long‐term clinical consequences of missing one or two treatments are unlikely to be significant. Therefore, the outcomes from these short‐term studies may not represent the true effect of the interventions performed in the long term and should be interpreted with caution. Despite this limitation, the included studies may provide a potentially useful signal regarding the efficacy of physiotherapy treatments.

Because of the nature of RCTs, all the interventions in the included studies were standardised across participants. This may not represent clinical practice, where physiotherapists advise individuals how to adapt or combine techniques (including frequency and duration of treatment) to create an individualised regimen balancing treatment effectiveness, acceptability, and likely adherence. The studies in this review do not provide sufficient evidence to support this approach in terms of clinical outcomes.

The studies included in this review recruited participants from both paediatric and adult populations (range seven to 51 years), and readers should take great care when extrapolating our findings to the younger paediatric population, particularly when considering the efficacy of routine ACT for asymptomatic screened babies. In 2008, the Association of Chartered Physiotherapists in Cystic Fibrosis (ACPCF) produced a guidance paper on the management of screened infants. This document provides a review of the evidence for airway clearance applicable to this population and consensus clinical opinion on this issue (Prasad 2008). In addition, there is a growing population of older adults with CF surviving well into the later decades of life who are not represented in the included studies.

The studies in this review included participants with a range of disease severity (12% to 102% predicted FEV₁). Not all studies reported baseline mean values, and it is therefore unclear if their participants reflect the current CF population, who have benefited from new therapies and associated improvements in lung health over recent decades. Only six studies were published from 2005, meaning it may be difficult to extrapolate the findings to the current population.

CF management is currently undergoing a period of significant change as researchers and clinical teams adapt to the development and increasingly widespread use of highly effective CFTR modulators worldwide. The introduction of these treatments will likely cause the role of daily ACTs for all to become an active area of discussion for people with CF and their clinicians. The long‐term effects of these new therapies are unknown. Some individuals with established lung damage and bronchiectasis may now report little or no sputum production, while in others, the early introduction of CFTR modulators may prevent the development of lung disease. 

This review aimed to inform practice to manage the condition of all people with CF, and the participants in the studies included in this review were not receiving CFTR modulators. It is not clear how applicable research from non‐treated populations will be for those individuals successfully treated with modulators in the short or long term. It is crucial to consider this gap in the evidence when considering ACT needs for people using CFTR modulators. Whilst modulator therapy has the potential to change the presentation and course of disease for treated individuals, there are people with CF who cannot use these drugs (due to ineligibility, lack or access, or intolerance), and our findings may be more applicable to them.

Quality of the evidence

All studies included few participants. Furthermore, there was substantial heterogeneity of treatments and outcome measures across studies, meaning we were unable to pool their results.

There are inherent risks of bias in physiotherapy studies. Because the interventions are physical (e.g. percussion, PEP, or postural drainage), a true placebo is not feasible. Similarly, participants and therapists cannot be blinded to the treatment being received. This partly explains the low quality scores of the included studies as reported using the Jadad scoring system in the original review, and the risk of bias assessments in the current version, as these methods place significant emphasis on blinding. 

We assessed the certainty of the evidence for this update using GRADE criteria (Table 1). We considered the evidence on the effect of ACTs on LCI to be of low certainty, mainly owing to imprecision caused by low participant numbers and no information on SDs or CIs of the difference between groups. We rated the certainty of the evidence on mucus transport rate and expectorated secretions as very low, owing to uncertain risk of bias across most or all domains in the included studies, and imprecision caused by small participant numbers.

Potential biases in the review process

We are unaware of potential bias in the review process. We undertook a comprehensive search of the literature with no time or language restrictions, and assessed studies independently against our eligibility criteria. We also worked independently when assessing risk of bias and when extracting data for inclusion in the final report to ensure accuracy of reporting.

Agreements and disagreements with other studies or reviews

The short‐term studies included in this review suggest that ACTs may increase mucus transport in people with CF (low‐ to very low‐certainty evidence). This finding supports the conclusion of an earlier review (Thomas 1995). These studies also suggest that in the short term, ACTs have no sustained effects on pulmonary function. Other Cochrane Reviews of physiotherapy interventions have assessed pulmonary function as an outcome in both short‐term and long‐term studies, and their short‐term results are similar to ours. (Burnham 2021; McIlwaine 2019; McKoy 2016; Morrison 2020).

Authors' conclusions

Implications for practice.

Short‐term cross‐over studies suggest that airway clearance regimens may have beneficial effects in people with cystic fibrosis (CF) in terms of improving mucus transport. However, we were unable to find any robust scientific evidence to support the hypothesis that airway clearance techniques have a long‐term beneficial effect in people with CF, nor to support the claim by some authors that these interventions are harmful.

Implications for research.

The gold standard for establishing efficacy of therapy is the randomised controlled trial comparing an intervention with a 'no treatment' group. However, there are several potential ethical considerations with regard to the withdrawal of an established and trusted treatment such as airway clearance techniques in people with CF, even in the absence of firm evidence. Conversely, it could be argued that in view of scant evidence to support the use of routine airway clearance for people with CF, and the fact that this intervention can be unpleasant, uncomfortable, and time‐consuming, a study with a 'no treatment' control group is justified in some circumstances. In a well population receiving cystic fibrosis transmembrane conductance regulator (CFTR) modulator therapy and with low sputum production, the ethical issues around withdrawing airway clearance may be considered less controversial. Therefore, we propose that future studies include control groups or control periods with sufficient numbers of participants. Ideally, such studies should have a parallel design and be conducted over months rather than days or weeks.

What's new

Date	Event	Description
12 April 2023	New search has been performed	A search of the Cystic Fibrosis and Genetic Disorders Review Group's Cystic Fibrosis Trials Register identified 179 references potentially eligible for inclusion in the review. Searches of online trials registers (ClinicalTrials.gov and the WHO ICTRP) identified 229 citations. After removing duplicates and eliminating clearly irrelevant citations, we assessed 57 citations. We included four new studies (seven references) (Dwyer 2017; Dwyer 2019; Grosse‐Onnebrink 2017; Vandervoort 2022).  We excluded 17 new studies (37 references) (ACTRN12605000535673; Corten 2020; Dacie 2015; Hansen 1990; Helper 2020; Horsley 2007; Hristara‐Papadopoulou 2007; Leemans 2020; McIlwaine 2001; McIlwaine 2014; NCT01266473; NCT03655249; NCT04010253; San Miguel Pagola 2020; Stanford 2019a; Varekojis 2003; Walicka‐Serzysko 2021) and added seven references to already excluded studies (Davies 2012; Fauroux 1999; Holland 2003; Reix 2009; Rodriguez 2013; Sontag 2010). We have listed two studies (six records) as ongoing (NCT03760120; Stanford 2019b).  As the list of excluded studies was extensive, we re‐evaluated it and removed 32 studies that we judged as being clearly irrelevant to the review, leaving a total of 134 studies listed as excluded.  In line with current Cochrane guidance, we added a summary of findings table to this update.
12 April 2023	New citation required but conclusions have not changed	To better reflect current terminology and practice, we changed the title of the review from 'Chest physiotherapy compared to no chest physiotherapy for cystic fibrosis'.

History

Protocol first published: Issue 1, 1999
Review first published: Issue 2, 2000

Date	Event	Description
25 November 2015	New search has been performed	A search of the Cystic Fibrosis and Genetic Disorders Review Group's Cystic Fibrosis Trials Register identified 29 potentially eligible new references. Eight references were additional references to four already excluded studies (Varekojis 2003; McIlwaine 2013; Parsons 1995; Reix 2009). The remaining 21 references were excluded as none of these studies included a no physiotherapy or spontaneous cough alone control group.
25 November 2015	New citation required but conclusions have not changed	One author, Cees van der Schans, has now stepped down from the review team. None of the newly identified references were eligible for inclusion in the review and hence our conclusions remain the same.
5 August 2013	New citation required but conclusions have not changed	A new review team have taken on this review at this update. Despite the inclusion of two new studies in this updated review, the conclusions remain the same.
5 August 2013	New search has been performed	A search of the Group's CF Trials Register identified 23 new references which were potentially eligible for inclusion in the review; two of these studies were assessed as suitable for inclusion (Elkins 2005; Jarad 2010) and the remaining 21 were excluded.
18 February 2009	New search has been performed	A search of the Group's Cystic Fibrosis Trials Register identified one additional reference to an already included study (Braggion 1995) and one to an already excluded study (Tannenbaum 2001).
18 February 2009	Amended	The Methods section has been updated in light of new guidance and functionality of RevMan 5.
12 November 2008	Amended	Converted to new review format.
20 February 2008	New search has been performed	The search of the Group's Cystic Fibrosis Trials Register identified one new reference which was the main paper to a previously excluded abstract (Lagerkvist 2006).
20 February 2008	Amended	The Plain Language Summary has been updated in line with guidance from The Cochrane Collaboration. Also, in a post hoc change and in line with Group guidelines, the outcome measures have been split into 'Primary outcomes' and 'Secondary outcomes'.
14 November 2006	New search has been performed	The search of the Group's Cystic Fibrosis Trials Register identified two new references. Both studies were excluded (Stites 2006; Warwick 2004).
14 November 2005	New search has been performed	The search of the Group's Cystic Fibrosis Trials Register identified four new references. One study identified was not eligible for inclusion in the review and has been added to the 'Excluded studies' section (Chatham 2004). The remaining three references were to three already excluded studies (Darbee 1990; Marks 1999; McIlwaine 1997).
18 May 2004	New search has been performed	Additional references (providing no additional information) have been added to the following already 'Included studies': Mortensen 1991; Falk 1993. Additional references have been added to the following already 'Excluded studies': Button 1997a; Costantini 1998; Orlik 2001. Three new studies have been added to 'Excluded studies': Hare 2002; Orlik 2000; Tannenbaum 2001.
14 August 2002	New search has been performed	Six crossover trials, previously cited in "Excluded Studies" have now been moved to the "Included Studies" section (Braggion 1995; Falk 1993; Mortensen 1991; Pfleger 1992; Rossman 1982; van der Schans 1991). Relevant changes to the text of the review have been made. Four new "Excluded Studies" have been incorporated into the review (Battistini 2001; Keller 2001; Pollard 2000; Orlick 2001). Additional references to studies already listed in "Excluded Studies" have been incorporated into the review within the following study ID's: Button 1997a; Gondor 1999; Grasso 2000; Marks 1999; Newhouse 1998).
9 February 2000	New citation required and conclusions have changed	Substantive amendment

Appendices

Appendix 1. Electronic search strategies

Database	Search strategy	Date searched
ClinicalTrials.gov (clinicaltrials.gov)	cystic fibrosis AND physiotherapy OR physical Therapy OR airway clearance techniques OR chest physiotherapy OR positive expiratory pressure OR oscillatory OR forced expiration technique OR active cycle of breathing OR autogenic drainage OR high frequency chest wall oscillation	7 November 2022
WHO ICTRP (trialsearch.who.int/)	cystic fibrosis AND physiotherapy OR physical Therapy OR airway clearance techniques OR positive expiratory pressure OR oscillatory OR forced expiration technique OR active cycle of breathing OR autogenic drainage OR high frequency chest wall oscillation	7 November 2022

 

Data and analyses

Comparison 1. Airway clearance techniques (ACTs) versus no ACTs.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 FEV1 (L) – change from baseline	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.1.1 Single intervention	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.2 FVC (L) – change from baseline	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.2.1 Single intervention	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Braggion 1995.

Study characteristics
Methods	Cross‐over RCT
Participants	16 people with CF Sex: 8 males; 8 females Age, mean (SD): 20.3 (4) years FEV1, mean (SD): 61.7 (17) % predicted
Interventions	Intervention 1: high‐frequency chest compression Intervention 2: postural drainage, breathing exercises, vibrations, manual percussion Intervention 3: PEP breathing Control: resting breathing with spontaneous coughing
Outcomes	Wet and dry weight expectorated mucus Pulmonary function tests (FVC, FEV1, FEF25-75) Subjective assessment
Notes	Measurement 30 minutes after intervention.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation by Latin square design (Williams 1949). To balance distribution between sexes, two 4 × 4 Latin squares were used for male participants and two for female participants.
Allocation concealment (selection bias)	Unclear risk	Not discussed.
Blinding (performance bias and detection bias) All outcomes	Unclear risk	Interventions did not allow participants or clinicians to be blinded; no mention of whether outcome assessors were blinded.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	No mention of any dropouts.
Selective reporting (reporting bias)	Low risk	Protocol not available for comparison, but data on expected outcomes reported. Immediate measurement after intervention.
Other bias	Unclear risk	Funding source not reported, but potential sources not thought to introduce a significant risk of bias. Unknown whether participants naive to all interventions prior to study.

Dwyer 2017.

Study characteristics
Methods	Cross‐over RCT
Participants	25 people with CF Age, mean (SD): 30 (8) years Sex: (15 males; 9 females) FEV1, mean (SD): 51 (SD 18) % predicted
Interventions	Intervention 1: treadmill exercise consisting of 20 minutes constant work rate equivalent to 60% peak VO2 (as assessed from visit 1) Intervention 2: Flutter Control: resting breathing for 20 minutes
Outcomes	Sputum properties – sputum solids content and mechanical impedance Peak expiratory airflow (reported but considered outside of review scope) Peak expiratory to peak inspiratory airflow bias (reported but considered outside of review scope) Number of coughs (reported but considered outside of review scope) VAS: subjective chest congestion / ease of expectoration
Notes	Airflow measured during interventions. Sputum collected immediately before and after the 20‐minute intervention, and after a further 20 minutes resting breathing or recovery. Additional time point of 5 minutes after intervention was also collected if participants spontaneously expectorated at this time.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated randomisation, administered by a person not involved in the interventions.
Allocation concealment (selection bias)	Low risk	Randomisation performed by a person not involved in the interventions and stored in sealed sequentially number opaque envelopes.
Blinding (performance bias and detection bias) All outcomes	Unclear risk	Not possible to blind participants or clinicians to the intervention. Outcome data analysed by blinded assessors.
Incomplete outcome data (attrition bias) All outcomes	Low risk	1 dropout reported following assessment visit prior to interventions commencing.
Selective reporting (reporting bias)	Low risk	Data on expected outcomes reported except for 1 secondary outcome (lung function), which was not reported after the intervention as described in the published protocol. All other outcomes collected before, during, immediately after, and 20 minutes after the interventions. Omission of post‐intervention lung function data not considered to significantly affect risk of bias given this was a secondary outcome.
Other bias	Unclear risk	Study was funded by the Australian Respiratory Council who had no input or contribution to the design of the study or collection, analysis, and interpretation of the data. Unknown how many participants were naive to Flutter before study.

Dwyer 2019.

Study characteristics
Methods	Cross‐over RCT
Participants	15 adults with mild‐to‐severe CF Age, mean (SD) 27 (9) years Sex: 10 males; 5 females FEV1, mean (SD): 65(23) % predicted
Interventions	Intervention 1: treadmill exercise consisting of 20 minutes at constant work rate equivalent to 60% peak VO2 (as assessed from visit 1) Intervention 2: PEP through device for 15 breaths followed by relaxed and deep breathing followed by FET or cough. Cycle repeated 6 times aiming for 10 cmH2O – 20 cmH2O pressure Control: resting breathing for 20 minutes
Outcomes	Mucus clearance measured using radioaerosol and dynamic imaging with a double‐headed gamma camera for whole right lung, central region, intermediate region, and peripheral region Number of coughs (reported but considered outside of review scope) VAS: subjective chest congestion
Notes	Radioaerosol and dynamic imaging procedure consisted of: Radioaerosol inhalation; dynamic imaging for 10 minutes to assess initial deposition and baseline clearance of the radioaerosol; 20‐minute intervention; and dynamic imaging over 60 minutes to assess post‐intervention mucus clearance. Coughs manually counted during 20‐minute intervention and 60‐minute follow‐up period. VAS completed at end of 60 minutes of post‐intervention monitoring.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated randomisation, administered by a person not involved in the interventions.
Allocation concealment (selection bias)	Low risk	Randomisation performed by a person not involved in the interventions and stored in sealed, sequentially numbered opaque envelopes.
Blinding (performance bias and detection bias) All outcomes	Unclear risk	Not possible to blind participants or clinicians to the intervention. Outcome data analysed by blinded assessors.
Incomplete outcome data (attrition bias) All outcomes	Low risk	1 dropout reported following visit 2 (possible allergic reaction to inhaled radioaerosol).
Selective reporting (reporting bias)	Low risk	Protocol not available for comparison. All outcomes prespecified in methods reported in results. Immediate measurement after intervention.
Other bias	Low risk	Funding source was reported as JJ Hooker Research Trust Fund Grant from the Australian Cystic Fibrosis Research Trust. Study reported the number of participants who were not naive to the PEP therapy (8/15).

Elkins 2005.

Study characteristics
Methods	Cross‐over RCT
Participants	12 adults with CF Age, mean (range): 25 (17–34) years Sex: numbers of males and females not stated FEV1, mean (range): 53 (16–88) % predicted
Interventions	Participants inhaled 99mTechnetium‐labelled sulphur colloid aerosol matching a target breathing pattern followed by 20 minutes of 1 of the following 4 interventions (randomised). Intervention1: postural drainage with percussion Intervention 2: PEP Intervention 3: oscillating PEP Intervention 4: matched cough (voluntary coughing to a maximum number of times coughed during previous interventions) Control: resting (sitting)
Outcomes	Mean % radioactive tracer clearance (10‐minute baseline dynamic SPECT scan after inhalation and scan at 90 min).
Notes	Study supported by National Health and Medical Research Council (NHMRC).
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Abstract states random order of the interventions but provides no details of randomisation method.
Allocation concealment (selection bias)	Unclear risk	Not discussed.
Blinding (performance bias and detection bias) All outcomes	Unclear risk	Not possible to blind participants or clinicians; no mention of whether outcome assessors were blinded.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	No mention of any dropouts.
Selective reporting (reporting bias)	Unclear risk	Protocol not available for comparison, and insufficient information to assess whether all outcomes reported.
Other bias	Unclear risk	Washout period unclear, not stated if interventions took place on same day or different days. Unknown whether participants were naive to all interventions prior to study.

Falk 1993.

Study characteristics
Methods	Cross‐over RCT
Participants	12 people with CF Age, sex, and disease severity not reported
Interventions	Intervention 1: FET, postural drainage Intervention 2: FET, PEP breathing Control: resting breathing with spontaneous coughing
Outcomes	Radioactive tracer clearance
Notes	Measurements at 30 minutes, 1 hour, 2 hours, and 24 hours after intervention.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Paper states random order of the interventions but provides no details of randomisation method.
Allocation concealment (selection bias)	Unclear risk	Not discussed.
Blinding (performance bias and detection bias) All outcomes	Unclear risk	Interventions did not allow participants or clinicians to be blinded; no mention of whether outcome assessors were blinded.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	No mention of any dropouts.
Selective reporting (reporting bias)	Unclear risk	Protocol not available for comparison, and insufficient information to assess whether all outcomes reported.
Other bias	Unclear risk	Funding source not reported, but potential sources not thought to introduce a significant risk of bias. Unknown whether participants naive to all interventions prior to study.

Grosse‐Onnebrink 2017.

Study characteristics
Methods	Parallel RCT
Participants	People hospitalised with infective exacerbation (defined by Fuchs criteria) and a confirmed diagnosis of CF. Mixed paediatric and adult population. 41 participants with baseline characteristics reported after randomisation.  Intervention group 20 participants Age, mean (SD): 19 (7.9) years  Sex: 15 males; 5 females FEV1, mean (SD): 50.87 (15.7) % predicted Control group 21 participants Age, mean (SD): 23.6 (11.8) years Sex: 13 males; 7 females FEV1, mean (SD): 57.36 (19.9) % predicted
Interventions	Intervention: 30 minutes of HFCWO followed by cough Control: 30 minutes of no intervention followed by cough
Outcomes	LCI and FRC assessed by MBW
Notes	Interventions were carried out on either day 2 or 3 after commencing intravenous antibiotics HFCWO settings were standardised based on a review of the literature and centre experience
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Paper states "randomly assigned (by sealed envelopes)" but provides no details of randomisation method.
Allocation concealment (selection bias)	Low risk	Sequence stored in sealed envelopes.
Blinding (performance bias and detection bias) All outcomes	Unclear risk	Interventions did not allow participants or clinicians to be blinded; no mention of whether outcome assessors were blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	All dropouts were reported and the reason given as failure to meet quality control criteria for multiple washout breath testing.
Selective reporting (reporting bias)	Low risk	Protocol not available for comparison, but data on expected outcomes were reported as expected.
Other bias	Low risk	Funding source identified as a WTZ research support service; authors disclosed no potential conflicts of interest.

Jarad 2010.

Study characteristics
Methods	Cross‐over RCT
Participants	19 adults with CF Age, mean (SD): 24 (4.8) years Sex: 11 males; 8 females Baseline FEV1 reported in L, so not comparable with other studies
Interventions	Intervention 1: HAT (this group excluded from the current review as not a recognised ACT) Intervention 2: Flutter Control: sitting in a bath with sham form of HAT
Outcomes	Expectorated sputum wet and dry weight FEV1, FVC, FEF25-75, FEF75 Participant questionnaire
Notes	Measurements 60 minutes after treatment.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Paper states treatment order was randomised but provides no details of randomisation method.
Allocation concealment (selection bias)	Unclear risk	Not discussed.
Blinding (performance bias and detection bias) All outcomes	Unclear risk	Interventions did not allow participants or clinicians to be blinded; no mention of whether outcome assessors were blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	There was 1 dropout reported following enrolment; this was due to time constraints. Complete data were presented for the remaining 18 participants.
Selective reporting (reporting bias)	High risk	Study protocol and methods section state SpO2, respiratory rate, HR and BP would be measured throughout the interventions, but there is no reference to these data in the results or discussion.
Other bias	Unclear risk	Funding source not reported, however paper states the authors had no conflict of interest. HAT group excluded from this analysis therefore potential bias from equipment provision not relevant. Unit of measurement for sputum weight not reported. 4/18 participants used Flutter as usual main physiotherapy method prior to study.

Mortensen 1991.

Study characteristics
Methods	Cross‐over RCT, interventions given on 3 occasions, each separated by 48 hours
Participants	10 people with CF Age, mean (SD): 20 (3.4) years Sex: 6 males; 4 females FEV1 median (range): 38.5 (26–101) % predicted
Interventions	20‐minute session of 1 of the following interventions immediately after ultrasonic nebulisation of 99mTC‐human albumin colloid. Intervention 1: postural drainage, FET, thoracic expansion exercises, relaxation Intervention 2: PEP breathing, FET Control: spontaneous coughing
Outcomes	Radioactive tracer clearance (measured every 30 minutes for 3 hours on each occasion) Sputum weight Penetration index (median and range) Retention at 24 hours (median and range) Number of huffs performed during treatment sessions (median and range) Number of cough manoeuvres (median and range) Only radioactive tracer clearance, sputum weight and FEV1 were reported in this review
Notes	Measurements at 30 minutes, 1 hour, and 24 hours after intervention Study approved by local ethical committee of Copenhagen
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Paper states random order of the interventions but provides no details of randomisation method.
Allocation concealment (selection bias)	Unclear risk	Not discussed
Blinding (performance bias and detection bias) All outcomes	Unclear risk	Described as single‐blind; as participants and caregivers could not be blinded due to type of intervention, we assumed that outcome assessors were blinded (but no detail of how this was achieved).
Incomplete outcome data (attrition bias) All outcomes	Low risk	No mention of any dropouts; data for all 10 participants present.
Selective reporting (reporting bias)	Low risk	Protocol not available for comparison, but data on expected outcomes reported. Immediate measurement after intervention.
Other bias	Unclear risk	Funding source reported. Unknown whether participants naive to all interventions prior to study.

Pfleger 1992.

Study characteristics
Methods	Cross‐over RCT
Participants	14 participants with CF Age, mean (range): 14 (9.8–22.4) years Sex: 5 males; 9 females FEV1, mean (SD) 53 (21) % predicted
Interventions	Intervention 1: PEP breathing Intervention 2: AD Intervention 3: PEP followed by AD Intervention 4: AD followed by PEP Control: spontaneous coughing
Outcomes	Pulmonary function tests (FVC, FEV1, RV/TLC, Raw) Weight of expectorated mucus
Notes	Measurements during and immediately after intervention Participants trained in interventions 6 months before commencement of the study
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Paper states random order of the interventions but provides no details of randomisation method.
Allocation concealment (selection bias)	Unclear risk	Not discussed.
Blinding (performance bias and detection bias) All outcomes	Unclear risk	Interventions did not allow participants or clinicians to be blinded. Assessor for sputum weight blinded, but no indication of whether assessment of other outcomes blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Abstract and paper state that 15 participants were randomly selected from local clinic, but provide data from 14, as 1 developed symptoms of acute respiratory viral infection during study and was excluded.
Selective reporting (reporting bias)	Low risk	Protocol not available for comparison, but data on expected outcomes reported. Immediate measurement after intervention.
Other bias	Low risk	Funding source not reported, however potential sources not thought to introduce a significant risk of bias.

Rossman 1982.

Study characteristics
Methods	Cross‐over RCT
Participants	6 participants with CF Age, mean (SD): 22.8 (5.6) years Sex: all male FEV1, range: 12%–77.7% predicted
Interventions	Intervention 1: postural drainage Intervention 2: postural drainage, mechanical percussion Intervention 3: regimented coughing Intervention 4: chest physiotherapy, breathing exercises, vibrations, manual percussion, postural drainage Control: spontaneous coughing
Outcomes	Radioactive tracer clearance Sputum weight
Notes	Measurements during and up to 2 hours after intervention
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Paper states random order of the interventions but provides no details of randomisation method.
Allocation concealment (selection bias)	Unclear risk	Not discussed.
Blinding (performance bias and detection bias) All outcomes	Unclear risk	Interventions did not allow participants or clinicians to be blinded; no mention of whether outcome assessors were blinded.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	No mention of any dropouts.
Selective reporting (reporting bias)	Low risk	Protocol not available for comparison, but data on expected outcomes reported. Immediate measurement after intervention.
Other bias	Unclear risk	Funding source not reported, however potential sources not thought to introduce a significant risk of bias. Unknown whether participants naive to all interventions prior to study.

van der Schans 1991.

Study characteristics
Methods	Cross‐over RCT
Participants	8 participants with CF Age, mean (SD): 16 (3) years Sex: not reported FEV1, mean (SD): 70 (24) % predicted
Interventions	Intervention 1: PEP breathing with a resistance of 5 cmH2O followed by 5 minutes of coughing Intervention 2: PEP breathing with a resistance of 15 cmH2O followed by 5 minutes of coughing Control: 5 minutes of directed coughing
Outcomes	Radioactive tracer clearance TLC FRC
Notes	Measurements during intervention
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Paper states random order of the interventions, but provides no details of randomisation method.
Allocation concealment (selection bias)	Unclear risk	Not discussed.
Blinding (performance bias and detection bias) All outcomes	Unclear risk	Interventions did not allow participants or clinicians to be blinded; no mention of whether outcome assessors were blinded.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	No dropouts mentioned.
Selective reporting (reporting bias)	Low risk	Protocol not available for comparison, but data on expected outcomes reported. Immediate measurement after intervention.
Other bias	Unclear risk	Funding source not reported. However, potential sources not thought to introduce a significant risk of bias. Unknown whether participants naive to all interventions prior to study.

Vandervoort 2022.

Study characteristics
Methods	Cross‐over RCT with a 6‐month gap between interventions
Participants	17 children with stable CF Age, mean (SD): 10.7 (2.8) years Sex: 11 males; 6 females FEV1, mean (SD): 94 (16.8) % predicted
Interventions	Intervention: PEP with 20 expirations (15 cmH2O) followed by ≥ 3 huffs and manual external percussion by physiotherapist (not defined at which stage of treatment) Control: resting breathing for 30 minutes
Outcomes	FEV1 FVC LCI, FRC and the index of convection‐dependent inhomogeneity multiplied by tidal volume assessed by MBW
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Paper states random order of the interventions but provides no details of randomisation method.
Allocation concealment (selection bias)	Unclear risk	Not discussed.
Blinding (performance bias and detection bias) All outcomes	Unclear risk	Interventions did not allow participants or clinicians to be blinded; no mention of whether outcome assessors were blinded.
Incomplete outcome data (attrition bias) All outcomes	Low risk	2 dropouts were accounted for prior to randomisation. Data from 2 further participants not included in MBW analysis due to insufficient test repeatability.
Selective reporting (reporting bias)	Low risk	Protocol not available for comparison, but data on expected outcomes reported. Immediate measurement after intervention.
Other bias	Unclear risk	Funding source declared and study authors disclosed no potential conflicts of interest.  There was a 6‐month gap between study visits, which is a sufficiently long washout for this type of study design. Baseline characteristics were reported as comparable at each visit confirming that the participants were stable; however, we considered the study to have unclear risk of bias.

AD: autogenic drainage; BP: blood pressure; CF: cystic fibrosis; FEF_25-75: forced expiratory flow between 25% and 75% expired forced vital capacity; FEF₇₅: forced expiratory flow at 75% expired forced vital capacity; FET: forced expiration technique; FEV₁: forced expiratory volume in 1 second; FRC: functional residual capacity; FVC: forced vital capacity; HAT: hydro‐acoustic therapy; HFCWO: high‐frequency chest wall oscillation; HR: heart rate; LCI: lung clearance index; MBW: multiple breath washout; PEP: positive expiratory pressure: Raw: airway resistance; RCT: randomised controlled trial; RV: residual volume; SD: standard deviation; SPECT: single‐photon emission computed tomography; SpO₂: blood oxygen saturation; TLC: total lung capacity; VAS: visual analogue score; VO₂: oxygen consumption.

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
ACTRN12605000348651	Study initially judged as potentially eligible for inclusion but principal investigator confirmed never completed.
ACTRN12605000535673	Unpublished PhD thesis; we were unable to obtain any results from the author.
App 1998	No control group without chest physiotherapy.
Arens 1994	No control group without chest physiotherapy.
Bain 1988	No control group without chest physiotherapy.
Baldwin 1994	No control group without chest physiotherapy.
Balestri 2004	No control group without chest physiotherapy.
Baran 1977	No control group without chest physiotherapy.
Battistini 2001	No control group without chest physiotherapy.
Bauer 1994	No control group without chest physiotherapy.
Bilton 1992	No control group without chest physiotherapy.
Blomquist 1986	No control group without chest physiotherapy.
Borka 2012	No control group without chest physiotherapy.
Braggion 1996	No control group without chest physiotherapy.
Button 1997	No control group without chest physiotherapy.
Castle 1994	No control group without chest physiotherapy.
Cerny 1989	No control group without chest physiotherapy.
Corten 2020	No control group without chest physiotherapy.
Costantini 1998	No control group without chest physiotherapy.
Dacie 2015	Insufficient detail in abstract to include, unsuccessful attempts to contact authors to obtain additional data.
Darbee 1990	No control group without chest physiotherapy.
Darbee 2005	No control group without chest physiotherapy.
Davidson 1988	No control group without chest physiotherapy.
Davies 2012	No control group without chest physiotherapy.
de Boeck 1984	No control group without chest physiotherapy.
Desmond 1983	No control group without chest physiotherapy.
Dosman 2003	No control group without chest physiotherapy.
Dunn 2013	No control group without chest physiotherapy.
Fainardi 2011	No control group without chest physiotherapy.
Falk 1984	No control group without chest physiotherapy.
Falk 1988	No control group without chest physiotherapy.
Fauroux 1999	Did not evaluate the outcomes identified.
Gaskin 1998	No control group without chest physiotherapy.
Giles 1995	No control group without chest physiotherapy.
Giles 1996	No control group without chest physiotherapy.
Gondor 1999	No control group without chest physiotherapy.
Gotz 1995	No control group without chest physiotherapy.
Grzincich 2008	No control group without chest physiotherapy.
Hansen 1990	No control group without chest physiotherapy.
Hare 2002	No control group without chest physiotherapy.
Hartsell 1978	No control group without chest physiotherapy.
Helper 2020	No control group without chest physiotherapy.
Hofmeyr 1986	No control group without chest physiotherapy.
Holland 2003	No control group without chest physiotherapy.
Homnick 1995	No control group without chest physiotherapy.
Homnick 1998	No control group without chest physiotherapy.
Horsley 2007	Abstract only; insufficient information on methodology.
Hristara‐Papadopoulou 2007	No control group without chest physiotherapy.
Jacobs 1981	No control group without chest physiotherapy.
Kerrebijn 1982	No control group without chest physiotherapy.
Kluft 1996	No control group without chest physiotherapy.
Kofler 1994	No control group without chest physiotherapy.
Kofler 1998	No control group without chest physiotherapy.
Konstan 1994	No control group without chest physiotherapy.
Kraig 1995	No control group without chest physiotherapy.
Lagerkvist 2006	No control group without chest physiotherapy.
Lannefors 1992	No control group without chest physiotherapy.
Leemans 2020	No control group without chest physiotherapy.
Lindemann 1992	No control group without chest physiotherapy.
Lorin 1971	No control group without chest physiotherapy.
Lyons 1992	No control group without chest physiotherapy.
Maayan 1989	No control group without chest physiotherapy.
Majaesic 1996	No control group without chest physiotherapy.
Marks 1999	No control group without chest physiotherapy.
Maxwell 1979	No control group without chest physiotherapy.
McCarren 2006	No control group without chest physiotherapy.
McIlwaine 1997	No control group without chest physiotherapy.
McIlwaine 2001	No control group without chest physiotherapy.
McIlwaine 2010	No control group without chest physiotherapy.
McIlwaine 2013	no control group without chest physiotherapy.
McIlwaine 2014	No control group without chest physiotherapy.
Miller 1995	No control group without chest physiotherapy.
Milne 2004	No control group without chest physiotherapy.
Morris 1982	No control group without chest physiotherapy.
Murphy 1983	No control group without chest physiotherapy.
Natale 1994	No control group without chest physiotherapy.
NCT01266473	No appropriate control group.
NCT03655249	No appropriate control group.
NCT04010253	No appropriate control group.
Newhouse 1998	No control group without chest physiotherapy.
Oberwaldner 1986	No control group without chest physiotherapy.
Oberwaldner 1991	No control group without chest physiotherapy.
Orlik 2000a	No control group without chest physiotherapy.
Orlik 2000b	No control group without chest physiotherapy.
Orlik 2001	No control group without chest physiotherapy.
Osman 2010	No control group without chest physiotherapy.
Padman 1999	No control group without chest physiotherapy.
Parreira 2008	No control group without chest physiotherapy.
Parsons 1995	No control group without chest physiotherapy.
Patel 2013	No control group without chest physiotherapy.
Phillips 2004	No control group without chest physiotherapy.
Pike 1999	No control group without chest physiotherapy.
Placidi 2006	No control group without chest physiotherapy.
Pollard 2000	No control group without chest physiotherapy.
Prasad 2005	No control group without chest physiotherapy.
Pryor 1979a	No control group without chest physiotherapy.
Pryor 1979b	No control group without chest physiotherapy.
Pryor 1981	No control group without chest physiotherapy.
Pryor 1990	No control group without chest physiotherapy.
Pryor 1994	No control group without chest physiotherapy.
Pryor 2010	No control group without chest physiotherapy.
Reisman 1988	No control group without chest physiotherapy.
Reix 2009	No control group without chest physiotherapy.
Rodriguez 2013	No control group without chest physiotherapy.
Salh 1989	No control group without chest physiotherapy.
Samuelson 1994	No control group without chest physiotherapy.
San Miguel Pagola 2020	No control group without chest physiotherapy.
Sanchez Riera 1999	No control group without chest physiotherapy.
Scherer 1998	No control group without chest physiotherapy.
Sontag 2010	No control group without chest physiotherapy.
Stanford 2019a	No control group without chest physiotherapy.
Steen 1991	No control group without chest physiotherapy.
Steven 1992	No control group without chest physiotherapy.
Tecklin 1976	No control group without chest physiotherapy.
Tonnesen 1982	No control group without chest physiotherapy.
Tugay 2000	No control group without chest physiotherapy.
Tyrrell 1986	No control group without chest physiotherapy.
van Asperen 1987	No control group without chest physiotherapy.
Van Ginderdeuren 2000	No control group without chest physiotherapy.
Van Ginderdeuren 2008	No control group without chest physiotherapy.
van Hengstum 1988	No control group without chest physiotherapy.
van Winden 1998	No control group without chest physiotherapy.
Vanlaethem 2008	No control group without chest physiotherapy.
Varekojis 2003	No control group without chest physiotherapy.
Verboon 1986	No control group without chest physiotherapy.
Walicka‐Serzysko 2021	Intervention outside of scope of this review and no appropriate control group.
Warwick 1990	No control group without chest physiotherapy.
Warwick 2004	No control group without chest physiotherapy.
Webber 1985	No control group without chest physiotherapy.
West 2010	No control group without chest physiotherapy.
Wheatley 2013	No control group without chest physiotherapy.
White 1997	No control group without chest physiotherapy.
Wilson 1995	No control group without chest physiotherapy.
Znotina 2000	No control group without chest physiotherapy.

Characteristics of ongoing studies [ordered by study ID]

NCT03760120.

Study name	Short term effects of physiotherapy on LCI (SPICy)
Methods	Cross‐over RCT
Participants	19 children aged 5–18 years, both boys and girls Inclusion criteria CF diagnosis Hospitalised for a scheduled intravenous antibiotics cycle and regularly followed‐up by the CF Centre Weight ≥ 15 kg FEV1 ≥ 40% predicted Ability to perform NMBW test Ability to perform spirometry Willing to adhere to protocol procedures In treatment with PEP mask
Interventions	Intervention: PEP  Control: sham PEP  Participants will undergo interventions on 2 subsequent mornings; 2 NMBWs before and after the intervention.
Outcomes	Effect of PEP therapy on lung heterogeneity measured by LCI, sputum weight, SpO2.
Starting date	Not currently recruiting.
Contact information	Carla Colombo – no contact information available.
Notes

Stanford 2019b.

Study name	Improving outcome measures for adult CF ACT trials
Methods	Cross‐over RCT Each participant will attend the research facility for 2 visits. Participants will be randomly assigned (computerised randomisation programme) to the order that they perform the study sessions.
Participants	Adults with CF, clinically stable.
Interventions	Intervention: a session of airway clearance utilising ACBT for a maximum of 60 minutes supervised by a specialist physiotherapist in adult CF. Control: a period of rest for up to 60 minutes in which participants sit in a comfortable position and can complete any non‐active tasks (e.g. reading, using internet)  The length of the ACBT and rest periods will be predefined by the individual participants' usual airway clearance regime and will be between 30 and 60 minutes, with both periods in both study visits matched.
Outcomes	Primary outcomes Change in LCI Change in FEV1 Electronic impedance tomography Change in impulse oscillation system Sputum wet weight Secondary outcomes Change in FVC Change in FEF at 25% of FVC Change in FEF at 50% of FVC Change in FEF at 75% of FVC SpO2 measurements Participant feedback questionnaire
Starting date	Recruitment status listed as active not recruiting (30 April 2018)
Contact information	Principal investigator: Nicholas J Simmonds, MD(Res) FRCP
Notes	Contacted authors August 2022.  Study data collection is complete but the authors have yet to report.

ACBT: active cycle of breathing technique; CF: cystic fibrosis; FEF: forced expiratory flow; FEV₁: forced expiratory volume in 1 second; FVC: forced vital capacity; HRCT: high‐resolution computer tomography; LCI: lung clearance index; NMBW: nitrogen multiple‐breath washout test; PEP: positive expiratory pressure; RCT: randomised controlled trial; SpO₂: blood oxygen saturation.

Differences between protocol and review

See van der Schans 1999 (protocol).

2013 update

We added the secondary outcome of participant preference during the 2013 review update as it is generally accepted that individual satisfaction or preference for airway clearance techniques is an important factor in treatment selection when considering likely adherence to treatment.

2015 update

We updated the Methods section in light of new guidance and functionality of RevMan (Review Manager 2014).

2023 update

We updated title of the review to reflect the use of the term 'airway clearance techniques' in preference to the traditional 'chest physiotherapy'.

We redefined the primary outcomes to better reflect outcomes used in recent CF literature. We added the primary outcome of lung clearance index (LCI) to reflect increasing use of this as an outcome measure in CF, while we moved the less commonly used forced expiratory flow between 25% and 75% expired FVC (FEF_25-75) to the secondary outcomes. 

We added a summary of findings table.

Contributions of authors

Original review

Ammani Prasad and Eleanor Main independently assessed studies for inclusion and assisted in writing of text. Cees van der Schans acted as guarantor of the review.

Updates from 2013

After Ammani Prasad and Eleanor Main stepped down from the review, Alison Gates and Louise Warnock joined the review team. These two review authors independently assessed studies for inclusion, re‐assessed the risk of bias of the included studies, and updated the text to include new studies.

Cees van der Schans commented on a draft of the updated review.

Updates from 2015

Cees van der Schans stepped down from the author team.

Sources of support

Internal sources

• No sources of support provided

External sources

• National Institute for Health & Care Research, UK

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

Declarations of interest

AG declares no known conflicts of interest.
LW declares no known conflicts of interest.

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