Chest physiotherapy compared to no chest physiotherapy for cystic fibrosis

Louise Warnock; Alison Gates

doi:10.1002/14651858.CD001401.pub3

. 2015 Dec 21;2015(12):CD001401. doi: 10.1002/14651858.CD001401.pub3

Chest physiotherapy compared to no chest physiotherapy for cystic fibrosis

Louise Warnock ^1,^✉, Alison Gates ¹

Editor: Cochrane Cystic Fibrosis and Genetic Disorders Group

PMCID: PMC6768986 PMID: 26688006

Abstract

Background

Chest physiotherapy is widely used in people with cystic fibrosis in order to clear mucus from the airways. This is an updated version of previously published reviews.

Objectives

To determine the effectiveness and acceptability of chest physiotherapy compared to no treatment or spontaneous cough alone to improve mucus clearance in 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.

Date of the most recent search of the Group's Cystic Fibrosis Trials Register: 02 June 2015.

Selection criteria

Randomised or quasi‐randomised clinical studies in which a form of chest physiotherapy (airway clearance technique) were taken for consideration in people with cystic fibrosis compared with either no physiotherapy treatment or spontaneous cough alone.

Data collection and analysis

Both authors independently assessed study eligibility, extracted data and assessed the risk of bias in the included studies. There was heterogeneity in the published outcomes, with variable reporting which meant pooling of the data for meta‐analysis was not possible.

Main results

The searches identified 157 studies, of which eight cross‐over studies (data from 96 participants) met the inclusion criteria. There were differences between studies in the way that interventions were delivered, with several of the intervention groups combining more than one treatment modality. One included study looked at autogenic drainage, six considered conventional chest physiotherapy, three considered oscillating positive expiratory pressure, seven considered positive expiratory pressure and one considered high pressure positive expiratory pressure. Of the eight studies, six were single‐treatment studies and in two, the treatment intervention was performed over two consecutive days (once daily in one, twice daily in the other). This enormous heterogeneity in the treatment interventions prevented any meta‐analyses from being performed. Blinding of participants, caregivers or clinicians in airway clearance studies is impossible; therefore this was not considered as a high risk of bias in the included studies. Lack of protocol data made assessment of risk of bias unclear for the majority of other criteria.

Four studies, involving 28 participants, reported a higher amount of expectorated secretions during chest physiotherapy as compared to a control. One study, involving 18 participants, reported no significant differences in sputum weight. In five studies radioactive tracer clearance was used as an outcome variable. In three of these (28 participants) it was reported that chest physiotherapy, including coughing, increased radioactive tracer clearance as compared to the control period. One study (12 participants) reported increased radioactive tracer clearance associated with all interventions compared to control, although this was only reported to have reached significance for postural drainage with percussion and vibrations; and the remaining study (eight participants) reported no significant difference in radioactive tracer clearance between chest physiotherapy, without coughing, compared to the control period. Three studies, involving 42 participants reported no significant effect on pulmonary function variables following intervention; but one further study did report significant improvement in pulmonary function following the intervention in some of the treatment groups.

Authors' conclusions

The results of this review show that airway clearance techniques have short‐term effects in the terms of increasing mucus transport. No evidence was found on which to draw conclusions concerning the long‐term effects.

Plain language summary

Chest physiotherapy compared to no chest physiotherapy for cystic fibrosis

Review question

We reviewed the evidence about the effect of using chest physiotherapy compared to no physiotherapy 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 drugs and chest physiotherapy. Physiotherapy clears mucus by different techniques or by using mechanical devices or both. 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 chest physiotherapy or the group with no chest physiotherapy. This is an update of previously published reviews.

Search date

The evidence is current to: 02 June 2015.

Study characteristics

We included eight studies in the review reporting results from a total of 96 people with cystic fibrosis. All the studies were very different and some looked at multiple treatments compared to no treatment. One study looked at autogenic drainage, six considered conventional chest physiotherapy, three considered oscillating positive expiratory pressure, seven considered positive expiratory pressure and one considered high pressure positive expiratory pressure. We could not combine the results to analyse them statistically.

Key results

Summarising the findings of these eight studies, we found that methods of clearing the airways have short‐term benefits for moving mucus. Three studies measured sputum which had been coughed up and found those people using chest physiotherapy coughed up more sputum; four studies measured radioactive tracer clearance and found increased clearance with chest physiotherapy. Only one study reported an improvement in lung function in some of the treatment groups; but three other studies who reported this outcome did not find any significant effect from chest physiotherapy. At present there is no clear evidence of long‐term effects in chest clearance, quality of life or survival with chest physiotherapy.

Quality of the evidence

Most of the included studies had some design problems which may affect confidence in the results. In just under half of the studies it was not clear as to whether all of the results were reported.

In physiotherapy studies the person and their physiotherapist will know which treatment they are receiving and this may affect some of the findings. Half the studies looked at amount of sputum coughed up and lung function testing, with a quarter asking the person's views on the treatment and these results may have been affected by being aware of the treatment. In all of the studies it was not clear if the person was experienced in carrying out the treatment. This may affect how well they were able to do the treatment which could affect confidence in the results.

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). Excessive production of thick mucus may overwhelm the normal mucus transport mechanisms and thereby lead to airway obstruction and mucus plugging (Zach 1990). Removal of airway secretions is therefore an integral part of the management of CF. A variety of methods are used to help remove secretions from the lungs, some physical, i.e. chest physiotherapy, and some chemical, i.e. medications and inhalation therapies. Treatment methods which improve mucus clearance are considered essential in optimising respiratory status and reducing the progression of lung disease.

Description of the intervention

Chest physiotherapy has, for a long time, played an important role in assisting the clearance of airway secretions and is usually commenced as soon as the diagnosis of CF is made. However, the performance of chest physiotherapy may be unpleasant, uncomfortable, and time‐consuming. Early chest physiotherapy relied on techniques for which the assistance of another person, such as a physiotherapist or relative, was needed and which included postural drainage, percussion, vibration, and shaking performed by an assistant and huffing or coughing. More recently, several self‐administered alternatives to these conventional techniques have been developed. These include the active cycle of breathing techniques (ACBT), forced expiration technique (FET), autogenic drainage (AD), positive expiratory pressure (PEP), flutter, high frequency chest compression (HFCC) and exercise. We have defined all of these methods under the interventions below. These methods of treatment help to give the individual with CF more independence in their management.

Why it is important to do this review

Despite the expansion of treatment modalities, there remains little evidence supporting their efficacy (Prasad 1998; van der Schans 1996). A previous meta‐analysis concluded that standard chest physiotherapy resulted in more mucus (phlegm or sputum) expectoration than no treatment in people with CF (Thomas 1995a). Similarly a review into airway clearance techniques used in the management of non‐CF related bronchiectasis concluded that there may be improvements in sputum expectoration, selected measures of lung function and health‐related quality of life (Lee 2013). Lee, however, also highlighted that data to establish long term efficacy of airway clearance techniques in this population were lacking (Lee 2013). In contrast to these findings a further review by Osadnik suggested the benefits achieved from airway clearance techniques for people with COPD may only confer small benefits in some of the measured clinical outcomes (Osadnik 2012). However, COPD is less frequently associated with the excessive sputum production as often manifested in CF or non‐CF related bronchiectasis.

This review is an update of previously published reviews (van der Schans 2000; Warnock 2013). It compares the efficacy of any of the above interventions as compared to no treatment or spontaneous coughing alone. Subsequent reviews will aim to determine whether a specific type of treatment offers any advantage over others.

Objectives

To determine the effectiveness of chest physiotherapy (airway clearance) compared to no treatment or cough alone in people with CF. This review does not address all possible comparisons between the multiple treatment techniques available for people with CF. This review is the first in a series of reviews which will compare the efficacy of different treatment modalities.

The following hypotheses will be tested: chest physiotherapy, whatever the type of intervention, is more:

effective than no chest physiotherapy;
effective than spontaneous coughing alone;
acceptable than no chest physiotherapy;
acceptable than spontaneous coughing alone.

Methods

Criteria for considering studies for this review

Types of studies

Randomised or quasi‐randomised clinical studies.

Short‐term studies (less than seven days duration, including single treatment studies) will be analysed 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

Chest physiotherapy of any type (see below) compared to no chest physiotherapy or spontaneous coughing alone.

In existing literature and in practical terms, variation occurs in the application of specific techniques. For the purposes of this series of reviews, it is necessary to group these variations under their broader headings. Separate analysis of each variation would render the reviews unmanageable. The following interventions aim to improve mucus transport or facilitate expectoration:

Conventional chest physiotherapy

This will include any combination of the following: postural drainage; percussion; chest shaking; huffing; and directed coughing. It should not include the use of exercise, PEP or other mechanical devices.

Positive expiratory pressure (PEP) mask therapy

As described by the authors to be the primary intervention, with or without additional techniques. PEP is defined as breathing with a positive expiratory pressure of 10 to 25 cmH₂0.

High pressure PEP (hPEP) mask therapy

As described by the authors to be the primary intervention, with or without additional techniques. It is a modification of the above PEP technique but includes a full forced expiration against a fixed mechanical resistance.

Active cycle of breathing techniques (ACBT)

This includes relaxation or breathing control, forced expiration technique (FET), thoracic expansion exercises and may include postural drainage or chest clapping.

Autogenic Drainage (AD)

As described originally by Chevalier or modified versions thereof. The authors should have identified AD to be the primary intervention, with or without additional techniques.

Exercise

With the sole purpose of improving mucus clearance as the primary intervention, with or without additional techniques.

Oscillating devices

Oscillating devices including flutter or cornet, thoracic oscillation, and oral oscillation. Flutter or cornet as described by the authors to be the primary intervention, with or without additional techniques. These devices produce an oscillatory PEP effect. Thoracic oscillation as defined by the authors to be the primary intervention, with or without additional techniques, to provide oscillation to the chest wall. Oral oscillation as defined by the authors to be the primary intervention, with or without additional techniques, to provide oscillation to the airways via the mouth.

Two authors independently categorised the physiotherapeutic interventions.

Types of outcome measures

Primary outcomes

Expectorated secretions (mucus, sputum, phlegm), dry or wet weight, or volume (an increase in the amount of expectorated secretions as a short‐term effect of the intervention is considered as beneficial)
Mucus transport rate (assessed by radioactive tracer clearance)
Pulmonary function tests (post‐intervention objective change from baseline compared to control)
1. forced expiratory volume in one second (FEV₁)
2. forced vital capacity (FVC)
3. forced expiratory flow between 25% and 75% expired FVC (FEF_25‐75)

Secondary outcomes

Oxygen saturation measured by pulse or transcutaneous oximetry
Total lung capacity (TLC) and functional residual capacity (FRC) (objective change from baseline compared to control)
Radiological ventilation scanning
Subjective perception of well‐being, ability to participate in activities of daily living
Therapy compliance
Objective change in exercise tolerance;
Nutritional status (assessed by growth, weight, body composition)
Number of respiratory exacerbations per year
Number of days in hospital per year
Number of days of intravenous antibiotics per year
Cost of intervention
1. equipment
2. duration
Deaths
Patient preference (post hoc change)

Search methods for identification of studies

Electronic searches

Relevant studies were identified from the Group's Cystic Fibrosis Trials Register using the terms: airway clearance techniques.

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

Date of the most recent search of the Group's Trials Register: 02 June 2015.

Data collection and analysis

Selection of studies

For the original review two authors from different centres independently assessed which studies should be included. 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 authors (AG, LW) from the same centre have independently assessed studies for inclusion in the review; if there was any disagreement they asked the third author (CvdS) to arbitrate.

Data extraction and management

Each author independently extracted data on the outcome measures listed above. Review authors planned to use the Cochrane Review Manager software to compile and analyse the data, but were only able to present a narrative summary (Review Manager 2011).

The authors planned to present short‐term studies (defined as having a duration of seven days or less) separately to longer‐term studies. The authors planned to group outcome data from longer‐term studies (more than seven days) into those measured at one, three, six, 12 months and annually thereafter. If outcome data are recorded at other time periods, then the authors planned to consider examining these as well.

Assessment of risk of bias in included studies

In the original review, the authors independently assessed the methodological quality of the included studies using a system as described by Jadad (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 authors considered aspects such as generation of randomisation sequence, the concealment of this sequence, degree of blinding and whether data were reported completely.

For the 2013 update of the review, the new author team assessed the risk of bias of the included studies according to the methods described in Chapter 8 of The Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). Specifically, the authors judged there to be either a high, low or unclear risk of bias from sequence generation, concealment of sequence allocation, blinding (from participants, clinicians and outcome assessors), missing data and reporting biases.

Measures of treatment effect

For continuous outcomes, the authors 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. In the case of binary outcomes, the authors planned to calculate the odds ratios (OR) and their 95% confidence intervals (CIs).

Unit of analysis issues

If studies had a cross‐over design, the authors planned to analyse the data from these as recommended by Elbourne; however this was not possible with the data currently available (Elbourne 2002).

Dealing with missing data

In order to allow an intention‐to‐treat analysis, the authors collected data on the number of participants with each outcome event by allocated treated group irrespective of compliance and whether or not the participant was later thought to be ineligible or otherwise excluded for treatment or follow up. Where there was evidence of missing data the authors contacted the primary investigator for clarification.

Assessment of heterogeneity

When the authors are able to include a sufficient number of studies in a meta‐analysis, they plan to assess heterogeneity using the I² statistic (Higgins 2003). This measure describes the percentage of total variation across studies that are due to heterogeneity rather than by chance (Higgins 2003). The values of I² lie between 0% and 100%, and a simplified categorization of heterogeneity that we plan to use is of low (I² value of 25%), moderate (I² value of 50%), and high (I² value of 75%) (Higgins 2003).

Assessment of reporting biases

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

Where possible the authors sought the study protocols and compared these with the final publications to ensure all measured outcomes were reported. If a study protocol was not available, the review authors compared the 'Methods' and 'Results' sections of each final publication to identify any discrepancies in outcome reporting.

Data synthesis

The authors were not able to present data in a meta‐analysis for this version of the review. However, if in future they are able to perform a meta‐analysis they will combine the data using a fixed‐effect model if there is little or no heterogeneity (i.e. I² is less than 25%). If there is a moderate or high degree of heterogeneity (i.e. I² is equal to or greater than 50%), they plan to use a random‐effects model.

Subgroup analysis and investigation of heterogeneity

If the authors identify a high degree of heterogeneity (i.e. I² is equal to or greater than 75%), they plan to investigate this using subgroup analyses including long‐term versus short‐term interventions. Subgroup analyses looking at the effects of specific interventions has been carried out in a series of separate physiotherapy reviews published by the Cochrane Cystic Fibrosis and Genetic Disorders Review Group (Elkins 2006; Main 2005; Morrison 2009; McKoy 2012 ).

Sensitivity analysis

The authors plan to test the robustness of their results by performing a sensitivity analysis of the data comparing results with and without quasi‐randomised studies.

Results

Description of studies

Results of the search

Of the 157 studies of airway clearance techniques identified by the literature searches, 149 were excluded and eight studies were included.

Included studies

Study design

The eight included studies were all cross‐over in design and all included a control period (Braggion 1995; Elkins 2005; Falk 1993; Jarad 2010; Mortensen 1991; Pfleger 1992; Rossman 1982; van der Schans 1991).

Four studies compared two active therapies to control (Falk 1993; Jarad 2010; Mortensen 1991; van der Schans 1991). The study by Jarad, however, included one group which did not receive a recognised airway clearance technique (hydro acoustic therapy (HAT)) as one of the interventions and therefore this treatment arm of the study was excluded from the current review (Jarad 2010). 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 undertook single treatments per day (Elkins 2005; Falk 1993; Mortensen 1991; Pfleger 1992; Rossman 1982; van der Schans 1991). One study administered therapy twice a day for two days (Braggion 1995), and one study repeated each intervention on two successive days (Jarad 2010).

Participants

The number of participants in the studies ranged from six (Rossman 1982) to 19 (Jarad 2010). The age of patients included in the studies ranged from 9.8 years (Pfleger 1992) to 34 years (Elkins 2005). One study included only males (Rossman 1982) and three studies did not report the gender of participants (Elkins 2005; Falk 1993; van der Schans 1991). The remaining studies all included both male and female participants: eight male, eight female (Braggion 1995); 11 male, eight female (Jarad 2010); six male, four female (Mortensen 1991); five male, nine female (Pfleger 1992). The studies included participants with a wide range of disease severity. The mean (SD) FEV₁% predicted was reported in three studies: 61.7% (17%) (Braggion 1995); 53% (21%) (Pfleger 1992) and 70% (24%) (van der Schans 1991). Elkins included patients with a mean (range) FEV₁% predicted 53% (16% to 88%) (Elkins 2005). Mortenson included participants with a median (range) FEV₁% predicted 38.5% (26% to 101%); and Rossman included participants with FEV₁% predicted ranging from 12% to 77.7% (Rossman 1982). Jarad stated they included patients with FEV₁% predicted less than 80%, however only absolute values were reported in the 'Results' so the disease severity of those entered into this study is unclear (Jarad 2010). Falk did not report details on patients' severity of disease (Falk 1993).

Interventions

The active interventions studied varied greatly between the included studies. The most often used active intervention was PEP breathing combined with the FET (Braggion 1995; Falk 1993; Mortensen 1991); two studies combined postural drainage with FET (Falk 1993; Mortensen 1991) and another study combined HFCC with FET and also with postural drainage (Braggion 1995). Braggion also used postural drainage combined with vibrations, deep breathing, percussion and coughing after each of the active therapies as well as the control sessions (Braggion 1995). Elkins compared postural drainage with percussion and vibrations, PEP, oscillating PEP and matched cough (where participants coughed the maximum numbers of coughs during any of the preceding intervention groups,including control) (Elkins 2005). One study compared PEP breathing, AD, PEP breathing followed by AD or AD followed by PEP breathing with a control period (Pfleger 1992). Van der Schans also used PEP in comparison with control, but varied the water pressure between the two active treatment arms (5 cm water pressure and 15 cm water pressure respectively) and followed this with directed vigorous coughing (van der Schans 1991). Jarad investigated the effect of the flutter and HAT compared to control (although the HAT group has been excluded from this review as it is not a recognised form of airway clearance) (Jarad 2010). The remaining study compared directed vigorous cough, postural drainage, postural drainage with mechanical percussion and conventional physiotherapy with control (Rossman 1982).

The control intervention in three studies was directed coughing (Pfleger 1992; Rossman 1982; van der Schans 1991) and in four studies it was spontaneous coughing (Braggion 1995; Elkins 2005, Falk 1993; Mortensen 1991). The control intervention in the study by Jarad was a placebo form of HAT which involved sitting in a bath with sounds audible, but without delivery of the acoustic waves thought to provide external thoracic oscillation therapy (Jarad 2010).

Outcome measures

The most common outcomes measured by the studies were: sputum weight, assessed by five of the included studies (Braggion 1995; Jarad 2010; Mortensen 1991; Pfleger 1992; Rossman 1982), two of which specified both wet and dry sputum weight (Braggion 1995; Jarad 2010); and radioactive tracer clearance also assessed by five studies (Elkins 2005; Falk 1993; Mortensen 1991; Rossman 1982; van der Schans 1991). Additionally, four studies reported pulmonary function tests (Braggion 1995; Jarad 2010; Pfleger 1992; van der Schans 1991). Three of these reported FEV₁ and FVC (Braggion 1995; Jarad 2010; Pfleger 1992); and two studies additionally reported FEF_25‐75% (Braggion 1995; Jarad 2010). Other pulmonary function tests reported were: forced expiratory flow at 75% (FEF₇₅) (Jarad 2010); residual volume as a fraction of total lung capacity (RV/TLC) and airway resistance (Raw) (Pfleger 1992); TLC and FRC (van der Schans 1991). Finally, two studies reported on patients' subjective assessment of the interventions (Braggion 1995; Jarad 2010).

Excluded studies

A total of 149 studies were excluded; 136 were excluded as they lacked a 'no treatment' or 'spontaneous coughing' control group. The remaining 13 studies were excluded for other reasons as follows: two studies were not clinical trials (Langenderfer 1998; Thomas 1995); four included diagnoses other than CF (Cochrane 1977; Parker 1984; Sutton 1985; van Hengstum 1988); four did not evaluate chest physiotherapy (Delk 1994; Fauroux 1999; Gayer 1988; Wordsworth 1996); one did not use any of the outcome measures defined for this review (Murphy 1988); one because the intervention was not thought to improve mucus clearance (Stites 2006); and one was in participants during an intra‐operative period under anaesthesia (Tannenbaum 2001).

Risk of bias in included studies

The methodological quality scored using the Jadad score for the original review is as follows:

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

Open in a new tab

The maximal score according to Jadad is five, however, two items are related to blinding of the investigator. Since blinding of the investigator is impossible in case of chest physiotherapy the maximal possible score for these studies is only three.

Details of the risk of bias assessment undertaken at the 2013 update follow.

Allocation

All eight studies were described as randomised; however, only one study gave any details on the method of randomisation (Latin square design) (Braggion 1995). We therefore judged the studies to have an unclear risk of bias for the generation of the randomisation sequence, except for the Braggion study which had a low risk of bias.

None of the eight studies discussed the concealment of the allocation and so all were deemed to have an unclear risk of bias.

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. Only one study was described as single‐blind, which we assume refers to the outcome assessors being blinded (Mortensen 1991). Since, except for the outcomes well‐being and therapy compliance, all other outcome measurements are physiological data, we do not consider the fact that participants, caregivers or clinicians were not blinded as an important source of bias.

Incomplete outcome data

Six of the included studies did not mention any dropouts. One study reported that one participant withdrew from the study due to time constraints (Jarad 2010); and another study stated that one participant was excluded from the study due to respiratory infection (Pfleger 1992). Since all included studies were short term and the effect was measured immediately after the intervention we do not consider dropouts an important risk of bias.

Selective reporting

In order to assess for selective reporting, we planned to compare the study protocols with the final publication. However, it was only possible to obtain a study protocol for one study (Jarad 2010). The protocol in this instance stated secondary outcome data would be collected on oxygen saturation (SpO₂₎, respiratory rate (RR), heart rate (HR) and blood pressure (BP) 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 and giving a high risk of bias.

Two of the included studies were only available in abstract form; therefore there was insufficient information available to assess for selective reporting. These studies were judged to be of unclear risk (Falk 1993; Elkins 2005).

For the remaining five studies which were published in full, there was no discrepancy identified between the methodology and results sections, thus raising no suspicion of selective reporting. Therefore these studies were judged to be at low risk (Braggion 1995; Mortensen 1991; Pfleger 1992; Rossman 1982; van der Schans 1991).

Other potential sources of bias

Mortenson was the only study in the current review to report a source of funding (Mortensen 1991). However, because the interventions being studied in the remaining studies either do not require equipment or use equipment widely available it was not considered by the authors that potential funding sources for the other studies in the review represented a significant risk of bias.

The study by Jarad included sputum wet and dry weight as an outcome, but did not report the unit of measurement in either the protocol or published study (Jarad 2010). If it is assumed that 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 patient with that technique. Therefore, naivety of patients to some but not all interventions being studied could introduce a potential source of bias. No reference is made to how experienced patients were with each intervention in six of the studies reviewed (Braggion 1995; Elkins 2005; Falk 1993; Mortensen 1991; Rossman 1982; van der Schans 1991). In the study by Jarad, there was the potential for this practised effect as four of the eighteen participants' usual physiotherapy technique was the flutter with the remainder of participants naive to this treatment intervention (Jarad 2010). Only Pfleger stated that all participants were trained in the techniques being studied during the six month period preceding commencement of the study (Pfleger 1992).

Effects of interventions

Eight studies were included; all are cross‐over studies and no meta‐analysis was possible. All were short‐term studies (less then seven days). Six studies were single treatment studies (Elkins 2005; Falk 1993; Mortensen 1991; Pfleger 1992; Rossman 1982; van der Schans 1991); in one study each intervention was performed twice on successive days (Jarad 2010); and in one study each physiotherapy treatment was given four times over two days (Braggion 1995).

Primary outcomes

1. Expectorated secretions

Expectorated secretions were reported in five studies (Braggion 1995; Jarad 2010; Mortensen 1991; Pfleger 1992; Rossman 1982), but sputum weight was not reported in the remaining three studies (Elkins 2005; Falk 1993; van der Schans 1991).

Four studies, involving 28 participants, found a higher amount of expectorated secretions during chest physiotherapy compared to the control period (Braggion 1995; Mortensen 1991; Pfleger 1992; Rossman 1982). Braggion found a mean wet weight of expectorated secretions during the control day of 6 g and during the chest physiotherapy sessions 23 g to 30 g (Braggion 1995). Mortenson reported medians and ranges, making comparisons between studies difficult but did report a significantly larger (P < 0.01) sputum weight in grams with FET + PEP median (range) 8.6 (3.5 to 19.9) and PD + FET 8.0 (2.3 to 13.9) compared to control 0.0 (0.0 to 2.1) during the treatment period; but this increase in sputum weight was not sustained in the post‐intervention follow up (Mortensen 1991). In the paper by Pfleger, the mean (SD) weight of expectorated mucus was presented for each treatment arm as a figure, but no specific data is given 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 chest physiotherapy (range 34 g to 45 g) (Pfleger 1992). Pfleger 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 (Pfleger 1992). Rossman found a statistically significant higher volume of expectorated secretions during the different forms of chest physiotherapy compared to the control session (Rossman 1982). Jarad, however, found no significant differences in wet or dry weight of expectorated sputum between the placebo or flutter groups; the P values and absolute value of measurement for sputum weight were not stated in the results making it difficult to draw comparisons between the studies (Jarad 2010).

2. Mucus transport rate as assessed by radioactive tracer clearance

In five studies radioactive tracer clearance was used as an outcome variable (Elkins 2005; Falk 1993; Mortensen 1991; Rossman 1982; van der Schans 1991). In four of these (40 participants), it was found that chest physiotherapy, including coughing, increased radioactive tracer clearance as compared to the control period (Elkins 2005; Falk 1993; Mortensen 1991; Rossman 1982). Elkins 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 also reported greater % C30 with PEP, oscillating PEP and matched cough compared to control but these did not reach significance (Elkins 2005). Falk found approximately 6% clearance during the control measurement and 9% during chest physiotherapy (Falk 1993). In the study by Mortensen, median clearance after 30 minutes during control was 7% and during two different chest physiotherapy sessions was 33% and 34% (Mortensen 1991). Rossman found 32% radioactive tracer clearance during the control measurement and 40% to 46% during the different forms of chest physiotherapy (Rossman 1982). One study (eight participants) reported no significant difference between two different chest physiotherapy sessions of PEP breathing (clearance 10% and 6%), without coughing, compared to a control period (clearance 8%) (van der Schans 1991). Different outcomes between the study by van der Schans and the other studies can be explained by the fact that in the van der Schans study participants were requested not to cough, but coughing was encouraged in the other studies as a part of the treatment (van der Schans 1991).

3. Pulmonary function tests

Pulmonary function measures were used as an outcome in four of the available studies (Braggion 1995; Jarad 2010; Pfleger 1992; van der Schans 1991).

Braggion reported no significant difference between any of the three treatments or control in FEV₁, FEF_25‐75, or FVC measured 30 minutes post intervention (Braggion 1995). Jarad reported a statistically significant reduction in values for FEV₁ (P = 0.028) and FEF_25‐75 (P = 0.03) following flutter but these were not observed in the placebo group (Jarad 2010). These changes were short‐lived and values had returned to baseline on the second study day (each intervention was repeated on two successive days). No statistically significant changes in FVC were reported in either the flutter or control groups (Jarad 2010). Pfleger measured FEV₁ and FVC at five different time points during each intervention session and reported means and SDs at the end of the first and fifth intervention session where the intervention varied in a random order from session two to five, hence we are unable to enter the data in a meta‐analysis (Pfleger 1992). The study found significant improvement in FEV₁ comparing the first and last time‐point analysis in the PEP (P < 0.01), AD (P < 0.05) and AD‐PEP (P < 0.01) groups but not in the control or AD‐PEP groups. Significant improvement in FVC at the same time‐points was also observed, but only in the PEP (P < 0.05) and PEP‐AD (P < 0.02) groups (Pfleger 1992). In the study by van der Schans, no differences were reported in FEV₁ measures for the treatment or control, although data are not fully reported for this variable (van der Schans 1991).