Summary of findings
Description of the condition
Chronic obstructive pulmonary disease (COPD) is a progressive disease state characterised by airflow limitation that is not fully reversible (McKenzie 2007). COPD is a major cause of morbidity, mortality and healthcare costs worldwide (Chapman 2006). People with COPD frequently experience breathlessness both at rest and on exertion, which limits their physical functioning and quality of life.
One of the most effective strategies for the management of COPD is land-based exercise training as part of integrated pulmonary rehabilitation. Land-based exercise training improves exercise capacity and quality of life (Lacasse 2006), and reduces admissions to hospital (Ries 2007) and length of stay (Golmohammadi 2004). However, land-based training is not always possible. COPD is prevalent in the older population (Cockram 2006), which includes a high proportion of physical co-morbidities (Fabbri 2008) that may preclude the elderly from participating in land-based training. This, combined with a high rate of non-completion of pulmonary rehabilitation programmes (Garrod 2006), means that it is important to explore alternative exercise options to enable people with COPD to complete some form of exercise training.
Description of the intervention
In the past, water-based exercise had been thought unsafe for people with COPD because of potential increases in cardiac and respiratory work as a consequence of increased venous return and increased chest wall pressure resulting from water immersion (Arborelius 1972). However, recent data have shown that a single head out of water exercise session in water can be performed safely without adverse events and with maintenance of oxygen saturation even in those with severe disease (Perk 1996). It is important to note that swimming is not considered in this review because of associated submersion of the head in water.
How the intervention might work
As a result of evidence that head out of water immersion and exercise in water are safe, water-based exercise can be considered as an alternative means of exercise training for people with COPD. It is hypothesised that when individuals complete a water-based exercise training programme of similar intensity and duration as land-based exercise training programmes that have previously been shown to be effective, exercise capacity and quality of life may improve to a similar degree. The unique properties of water, including buoyancy to support body weight and reduce mechanical impact on the body, water turbulence and resistance to increased muscle work when moving the body and limbs through the water, and warm water temperature, which may improve blood flow to muscles, may enable a higher intensity and duration of exercise, especially in people who have difficulty completing a land-based exercise training programme. These features of the water environment may mean that water-based exercise training is more suitable for people with comorbid conditions such as musculoskeletal or orthopaedic conditions.
Why it is important to do this review
Thus far, a systematic review of studies of water-based exercise training for people with COPD has not been conducted.
It is important to perform this review to evaluate the safety of water-based exercise training in people with COPD and to determine the effect of water-based training on exercise capacity and quality of life as described in the available literature.
To assess the effects of water-based exercise training in people with COPD.
Criteria for considering studies for this review
Types of studies
Randomised or quasi-randomised controlled trials in which water-based exercise training of at least four weeks' duration was compared with no exercise training or any other form of exercise training in people with COPD.
Types of participants
Adults with a clinical diagnosis of COPD based on the investigators' definition. The COPD should be stable (i.e. optimal and stable respiratory medications with no exacerbation or hospital admission within the previous month), and supplemental oxygen may be used.
Types of interventions
Studies examining water-based exercise training, supervised or unsupervised, were included if they included a comparison with no exercise training, a land-based exercise training programme or a sham water-based treatment. Trials in which water-based training was combined with another training intervention (e.g. land-based exercise training) were eligible for inclusion provided more than 50% of the training incorporated water-based exercise training. Swimming interventions were excluded.
Types of outcome measures
- Exercise capacity (functional or maximal), measured during formal exercise tests or field exercise tests.
- Quality of life, measured by generic or respiratory-specific quality of life questionnaires.
- Pulmonary function.
- Respiratory muscle strength.
- Upper and lower limb strength.
- Oxygen saturation.
- Level of activity.
- Psychological status.
- Self management/self efficacy.
- Healthcare utilisation.
- Adverse events.
- Body composition.
- Exercise training mode preference.
- Arterial blood gases.
Search methods for identification of studies
We identified trials from the Cochrane Airways Group Specialised Register of trials, which is derived from systematic searches of bibliographic databases including the Cochrane Central Register of Controlled Trials (CENTRAL), MEDLINE, EMBASE, CINAHL, AMED and PsycINFO, and we performed handsearching of respiratory journals and meeting abstracts (please see Appendix 1 for further details).
All records added to the Specialised Register and coded as ‘COPD’ were searched using the following terms: water* or aqua* or bath or pool or hydrotherap* or hydro-therap* or immers*.
We also conducted a search of ClinicalTrials.gov using the same terms. All databases were searched from their inception to August 2013, and no restriction regarding language of publication was imposed.
Searching other resources
We handsearched reference lists of all relevant studies to look for additional qualifying studies. Authors of identified and potentially eligible trials were contacted and were asked to identify further published and unpublished studies.
Data collection and analysis
Selection of studies
Studies identified through the literature searches were independently coded by two review authors (RJM and ZJM) for inclusion upon examination of titles and abstracts. Studies were categorised as follows.
- Include: Study categorically meets all review criteria.
- Unclear: Study appears to meet some review criteria, but insufficient information is available to categorically determine relevance.
- Exclude: Study does not categorically meet all review criteria.
The two review authors then used a full-text copy of each study in the first two categories to determine study inclusion. Disagreements were resolved by consensus, and when any disagreement could not be resolved, we consulted a third review author (JAA). A full record of decisions was kept, and simple agreement and kappa statistics were calculated.
Data extraction and management
Two review authors (RJM and ZJM) independently extracted data using a standard checklist before the primary review author (RJM) entered them into Review Manager (RevMan), and the second review author (ZJM) conducted random checks on accuracy. Disagreements were resolved by consensus. When two or more detailed reports described findings of the same study, data were extracted separately and then collated. Data collected included characteristics of included studies (methods, participants, interventions, outcomes) and results of the included studies. We recorded the specific details of exercise training (intensity, frequency, duration, type). Authors of included studies were contacted and were asked to provide missing data.
Assessment of risk of bias in included studies
Two review authors (RJM and ZJM) independently assessed the internal validity of included studies using The Cochrane Collaboration's 'Risk of bias assessment' tool (Higgins 2011) (including randomisation sequence generation; allocation concealment; blinding of participants, assessors and outcome assessments; completeness of outcome data; selective outcome reporting and any other possible sources of bias). Each item for each study was judged as having high, low or unclear risk of bias. Disagreements were resolved by consensus. Study authors were contacted to seek clarification in cases where quality was unclear.
Measures of treatment effect
The mean change from baseline with standard deviation (SD) for each group was recorded or calculated from available data for continuous variables. The mean difference (MD) for outcomes measured using the same measurement tool and the standardised mean difference (SMD) for outcomes measured with different measuring tools, as well as 95% confidence intervals (95% CIs), were calculated using RevMan 5.2. A pooled quantitative analysis was performed when trials were clinically homogeneous.
Unit of analysis issues
One study used a randomised cross-over trial design. Only the data from the first arm of this trial were incorporated in this review, given that a significant period of treatment interaction was found in the study. Another study used a semi-randomised design methodology whereby participants were randomly assigned to the two intervention groups, but the control (no intervention group) was not randomly assigned. Data from this non-randomised comparison control group were not used in this review.
Dealing with missing data
The original study investigators were contacted for further information when data were missing or could not be interpreted in the presented form.
Assessment of heterogeneity
Heterogeneity was assessed using the I
Assessment of reporting biases
We planned to create a funnel plot to test for publication bias and small-study effects if we had been able to pool ten or more trials.
A fixed-effect model was used in the analysis.
Subgroup analysis and investigation of heterogeneity
One subgroup analysis was specified a priori to explore possible sources of heterogeneity.
- Severity of lung disease: forced expiratory volume in one second (FEV
1) less than 40% predicted.
The small number of studies precluded this subgroup analysis. If in future updates, more studies are included, subgroup analysis will be performed.
A sensitivity analysis was not conducted because of the small number of studies. If more studies are included in future review updates, a sensitivity analysis will be performed to analyse the effects of allocation concealment and intention-to-treat analysis on results.
Description of studies
Results of the search
A total of 88 citations were identified by searching the databases. Six additional studies were identified upon handsearching of reference lists and completion of further searches by the review authors. From study titles and abstracts of references in this list, we identified and retrieved 26 papers for closer inspection. Study evaluation revealed that five studies (represented by a total of 12 citations) met the inclusion criteria for this review. A flow diagram of the search results is provided in Figure 1. Of the 26 studies analysed, the review authors agreed on 25 articles (96%) with kappa = 0.90, indicating excellent agreement. Disagreement was resolved by consultation with the third review author (JAA).
|Figure 1. Study flow diagram.|
Five studies (a total of 176 participants) met the inclusion criteria for this review. Full details can be found in the Characteristics of included studies table. One study was published in abstract form only; however, we were able to obtain unpublished data from the trialists (O'Brien 2004). Three studies were randomised controlled trials (de Souto Araujo 2012; McNamara 2013; Ozdemir 2010), one study was a semi-randomised controlled trial (Wadell 2004) and the remaining study was a randomised cross-over trial (O'Brien 2004). Only data from the first arm of the cross-over trial and only data from the randomised intervention groups in the semi-randomised controlled trial were used in analyses. All studies compared supervised water-based exercise training versus land-based exercise training and/or no exercise training in people with COPD (average FEV
Fourteen studies were excluded from this review upon evaluation against the inclusion and exclusion criteria. The most common reason for exclusion of studies was that they were not randomised controlled trials (n = 11). Full details can be found in the Characteristics of excluded studies table.
Risk of bias in included studies
|Figure 2. Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.|
|Figure 3. Risk of bias summary: review authors' judgements about each risk of bias item for each included study.|
All studies reported random allocation to groups; however, the amount of information provided was highly variable. Randomisation for sequence generation was reported by two studies (McNamara 2013; Ozdemir 2010). The remaining three studies provided insufficient information (de Souto Araujo 2012; O'Brien 2004) or did not conduct random sequence generation for the control (no exercise) group (Wadell 2004). Allocation sequence concealment with the use of sealed envelopes was reported by two studies (McNamara 2013; O'Brien 2004). Information regarding concealment of allocation sequence was insufficient in the remaining studies (de Souto Araujo 2012; Ozdemir 2010; Wadell 2004).
Participant and personnel blinding was not possible in any of the studies because of the physical nature of the intervention. Use of a blinded assessor to measure outcomes was reported in two studies (McNamara 2013; O'Brien 2004). Adequate assessment of assessor blinding in the three remaining studies could not be determined because sufficient information was lacking.
Incomplete outcome data
Four studies reported dropouts and losses to follow-up (de Souto Araujo 2012; McNamara 2013; O'Brien 2004; Wadell 2004). The single remaining study did not report dropouts or losses to follow-up (Ozdemir 2010).
All studies documented outcome measures, which were reported in the prespecified methods (see Characteristics of included studies).
Effects of interventions
The Data and analyses tables summarise the results of the meta-analyses for the two comparison pairs analysed in this review: (1) water-based exercise versus no exercise; and (2) water-based exercise versus land-based exercise. Meta-analyses are presented for all of the primary outcomes and for the two comparison pairs (regardless of the number of studies providing data), and for secondary outcomes only when sufficient data were available from at least two studies. Summary of findings for the main comparison and Summary of findings 2 summarise the quality of the evidence. For exercise capacity tests and tests of pulmonary function, positive values reflect an improvement. For quality of life measurement tools, negative values reflect benefit.
Four studies used the six-minute walk test to measure functional exercise capacity (de Souto Araujo 2012; McNamara 2013; O'Brien 2004; Ozdemir 2010). Results from three trials could be combined in a meta-analysis comparing water-based exercise (n = 48) versus no exercise (n = 51) (de Souto Araujo 2012; McNamara 2013; Ozdemir 2010), and results from three trials could be combined in a meta-analysis comparing water-based exercise (n = 28) versus land-based exercise (n = 34) (de Souto Araujo 2012; McNamara 2013; O'Brien 2004). When compared with no exercise, water-based exercise resulted in a mean difference change in distance walked of 62 metres (95% CI 44 to 80 metres; Figure 4). When compared with land-based exercise, water-based exercise resulted in a non-significant mean difference change in distance walked of 11 metres (95% CI -11 to 33 metres; Figure 5).
|Figure 4. Forest plot of comparison: 1 Water-based exercise versus no exercise, outcome: 1.1 Exercise capacity (mean change in metres).|
|Figure 5. Forest plot of comparison: 2 Water-based exercise versus land-based exercise, outcome: 2.1 Exercise capacity (mean change in metres).|
Two studies used the incremental shuttle walking test (McNamara 2013; Wadell 2004). The mean difference for change in distance walked of 50 metres (95% CI 20 to 80 metres) favoured water-based exercise (one study; n = 15) over no exercise (one study; n = 15) (Figure 4). No significant difference was found between water-based exercise (two studies; n = 30) and land-based exercise (two studies; n = 29) (MD 9 metres; 95% CI -15 to 34 metres) and high statistical heterogeneity was noted (I
Two studies used the endurance shuttle walk test (McNamara 2013; Wadell 2004). The mean difference for change in distance walked was significant at 371 metres (95% CI 121 to 621 metres) in favour of water-based exercise (one study; n = 15) over no exercise (one study; n = 15) (Figure 4) and 313 metres (95% CI 232 to 394 metres) in favour of water-based exercise (two studies; n = 30) over land-based exercise (two studies; n = 29) (Figure 5).
Quality of life
Health-related quality of life was measured in all studies; however, insufficient data were available for all studies to enable pooling of results in a meta-analysis. The potential impact of the fact that one study (Ozdemir 2010) did not contribute data to the synthesis is most likely small. In the water-based exercise versus no exercise comparison, data could be pooled from two studies (n = 23 water-based exercise; n = 26 no exercise). The Chronic Respiratory Disease Questionnaire (CRDQ) was used by McNamara 2013, and the St George's Respiratory Questionnare (SGRQ) was used by de Souto Araujo 2012. A standardised mean difference of -0.97 (95% CI -0.37 to -1.57) favoured water-based exercise (Figure 6). In the water-based exercise versus land-based exercise comparison, data were available from three studies that used the SGRQ (de Souto Araujo 2012; O'Brien 2004; Wadell 2004) and from one study that used the CRDQ (McNamara 2013) (n = 42 water-based exercise and n = 47 land-based exercise). A standardised mean change in total scores of -0.14 (95% CI -0.57 to 0.28; Analysis 2.2) indicated no significant difference in total quality of life. Moderate to substantial heterogeneity was noted (I
|Figure 6. Forest plot of comparison: 1 Water-based exercise versus no exercise, outcome: 1.2 Quality of life (mean change in total scores).|
In the comparison of water-based exercise versus no exercise, a mean difference change of 6.3% (95% CI 3.4 to 9.2) for FEV
Respiratory muscle strength
Two studies reported measures of respiratory muscle strength (de Souto Araujo 2012; McNamara 2013). All measures of respiratory muscle strength were seen to improve significantly when water-based exercise (n = 23) was compared with no exercise (n = 26); however significant heterogeneity was noted between the studies. The mean difference change for maximal inspiratory pressure was 14 cm H
Two studies provided information regarding adverse events (McNamara 2013; O'Brien 2004) upon completion of water-based exercise (n = 20). No adverse events were reported in the study by O'Brien 2004. One minor adverse event (an accidental skin tear from a fingernail scratch during the session in a water-based exercise group participant) was reported in the study by McNamara 2013.
Measurement of body weight (kg) was reported by two studies (McNamara 2013; Wadell 2004). Upon completion of water-based exercise (n = 30), participants had a mean weight loss of 1.29 kg (95% CI -2.65 to 0.07) compared with land-based exercise (n = 29). This mean weight loss did not reach statistical significance.
Exercise session attendance was reported by two studies (McNamara 2013; O'Brien 2004). No significant difference was observed in the number of exercise sessions attended between water-based exercise training and land-based exercise training (SMD 0.44, 95% CI -0.18 to 1.07).
Two studies examined participants' preferences for the exercise training environment upon completion of the training study period (McNamara 2013; O'Brien 2004). Of 41 participants who completed exercise training (i.e. water-based or land-based training), 49% (n = 20) reported that they would prefer exercise training in water, and 37% (n = 15) reported their preference for exercise training on land. Fifteen percent (n = 6) reported no preference for either environment for exercise training.
One study reported long-term effects of water-based exercise training six months after completion of a 12-week training programme (Wadell 2004). For outcome measures for which data were reported (quality of life and body composition), no significant change was observed between baseline results and results obtained six months post water-based exercise training.
Summary of main results
Five studies comparing water-based exercise training versus no exercise or land-based exercise training in people with COPD were identified for this review. Water exercise resulted in significant improvement in functional exercise capacity, peak exercise capacity, endurance exercise capacity and health-related quality of life when compared with no exercise. When compared with land exercise, water exercise elicited significantly greater improvement in endurance exercise capacity. To date, data are insufficient for conclusions to be drawn regarding the long-term effects of water-based exercise training in COPD.
Overall completeness and applicability of evidence
Significant mean improvements in functional, peak and endurance exercise capacity following water exercise compared with no exercise and in endurance exercise capacity following water-based exercise compared with land-based exercise surpassed the minimum clinically important differences (MCIDs) that have been reported in the literature. The mean improvement in the six-minute walk test of 62 metres compared with no exercise is greater than any of the previously recorded MCIDs of 25 metres (Holland 2010), 35 metres (Puhan 2008) and 54 metres (Redelmeier 1997) in people with COPD. The 50 metres mean improvement in incremental shuttle walk distance when water-based exercise training was compared with no exercise is greater than the 47.5 metres MCID reported by Singh 2008. Finally, improvement of 371 metres and 313 metres in the endurance shuttle walk test compared with no exercise and land exercise, respectively, clearly exceeded the MCID of 60 to 115 metres (Pepin 2011). The reason for the greater response in endurance walking capacity following water exercise is unclear; however, it is likely that a greater training stimulus is delivered in the water environment, where every motion in every direction of movement is resisted by the hydrostatic pressure of the water and the effect of water turbulence (Becker 2009). This greater endurance capacity may be better reflected in an endurance exercise test such as the endurance shuttle walk test rather than the six-minute walk test or the incremental shuttle walk test, in which the participant would have to walk faster rather than longer to demonstrate improvement.
Several components of health-related quality of life significantly improved upon completion of water-based exercise training compared with no exercise training. Significant changes were recorded by one study using the CRDQ (for total score, dyspnoea and fatigue subscores) (McNamara 2013) and by another study using the SGRQ (including total score and impact subscore) (de Souto Araujo 2012). However, no significant change in CRDQ emotional function and mastery subscores or in SGRQ symptom and activity subscores were found. When water-based exercise training was compared with land-based exercise training, the only significant change in quality of life favouring water-based exercise was seen in the CRDQ fatigue subscore. But this conclusion is based on the findings of only one study (McNamara 2013). It is important to note that most of the studies included in this review did not state their primary outcome measure, and only two studies calculated appropriate sample sizes, which were based on an exercise capacity outcome - not on quality of life (McNamara 2013; Wadell 2004). Therefore, based on the information provided, none of the studies in this review were adequately powered to determine a significant change in quality of life outcomes. Future studies will require sufficient sample sizes so it can be adequately determined whether water-based exercise training changes quality of life.
The studies in this review included COPD participants with an average FEV
Based on data from two studies (n = 20; GOLD stage II), water-based exercise was found to be safe for people with COPD, with only one minor adverse event reported. Future studies should report adverse events to further confirm the safety of water-based exercise training.
All studies in this review prescribed for both exercise training groups endurance exercise training or a combination of endurance and strength exercise training. The four studies that included comparisons of water-based exercise versus land-based exercise prescribed the water-based exercise training programme to match as closely as possible the land-based exercise training programme by using similar intensity, frequency and muscle groups trained (de Souto Araujo 2012; McNamara 2013; O'Brien 2004; Wadell 2004). This methodological feature gives assurance that the water exercise programmes adhered as closely as possible to current guidelines regarding exercise prescription in pulmonary rehabilitation (Nici 2006), especially given the difference in exercise training mediums.
Quality of the evidence
The quality of evidence for the primary outcomes in this review ranged from low to moderate. Interpretation of the findings of this review should be considered carefully because of potential sources of bias, especially the small sample size of each study. Lack of adequate randomisation in one study and unreported or poor random sequence generation, allocation concealment and blinding of outcome assessors are significant limitations.
Potential biases in the review process
Although the authors of all included studies were contacted to provide additional information and results, some information was not supplied, and this limited the data that could be included in some meta-analyses.
Agreements and disagreements with other studies or reviews
One previous review has examined the effect of water-based exercise training in people with COPD (McNamara 2011); however only one study was identified and reviewed. The inclusion of five studies in this current review (including the study previously examined in McNamara 2011) is a major strength.
Implications for practice
There is limited quality evidence that water-based exercise training is safe for people with COPD (GOLD stage II) and improves exercise capacity and health-related quality of life immediately following training for people with COPD GOLD stage II and III when compared with no training. There is limited quality evidence that water-based exercise training offers advantages over land-based exercise training in improving endurance exercise capacity, but we remain uncertain as to whether it leads to better quality of life. It is, therefore, appropriate to offer people with GOLD stage II and III COPD a water-based exercise training programme, especially when they have concurrent physical co-morbidities, or when the alternative is no exercise training.
Implications for research
More high-quality randomised controlled trials are required to further confirm exercise capacity and health-related quality of life outcomes following water-based exercise training in people with COPD. In particular, larger studies need to be conducted to determine whether disease severity affects the benefits of water-based exercise training and whether water exercise results in long-term effects. Study methodology in future trials needs to be rigorous with adequate sample sizes, randomisation, allocation concealment and blinding of outcome assessors, and this information must be reported in publications.
We thank the staff of the Cochrane Airways Group and the Australian Cochrane Airways Group Network for their guidance in completing this review. Thank you also to the Australian Cochrane Airways Group Network for providing a scholarship to enable review completion.
Anne Holland was the Editor for this review and commented critically on the review.
Data and analyses
- Top of page
- Summary of findings [Explanations]
- Authors' conclusions
- Data and analyses
- Contributions of authors
- Declarations of interest
- Sources of support
- Differences between protocol and review
- Index terms
Appendix 1. Sources and search methods for the Cochrane Airways Group Specialised Register (CAGR)
Electronic searches: core databases
Handsearches: core respiratory conference abstracts
MEDLINE search strategy used to identify trials for the CAGR
1. Lung Diseases, Obstructive/
2. exp Pulmonary Disease, Chronic Obstructive/
4. (chronic$ adj3 bronchiti$).mp.
5. (obstruct$ adj3 (pulmonary or lung$ or airway$ or airflow$ or bronch$ or respirat$)).mp.
Filter to identify RCTs
1. exp "clinical trial [publication type]"/
2. (randomised or randomised).ab,ti.
11. 9 not (9 and 10)
12. 8 not 11
The MEDLINE strategy and the RCT filter are adapted to identify trials in other electronic databases.
Protocol first published: Issue 1, 2010
Review first published: Issue 12, 2013
Contributions of authors
RJM: protocol initiation, development and writing; search for and retrieval of studies; study screening, quality appraisal and data extraction; author contact; data entry and analysis; manuscript writing.
ZJM: protocol development; study screening, quality appraisal and data extraction; data entry review; manuscript review.
DKM: protocol development; manuscript review.
JAA: protocol development; study quality appraisal; manuscript review.
Declarations of interest
The review authors (RJM, ZJM, DKM and JAA) conducted one of the included studies before commencing this review (McNamara 2013).
Sources of support
- No sources of support supplied
- Australian Cochrane Airways Group Network Scholarship, Australia.
Differences between protocol and review
Types of interventions: The criterion 'trials where water-based training was combined with another training intervention were included provided 50% or more of the training was water-based' was changed to a criterion whereby 'trials where water-based training was combined with another training intervention were included provided the water-based exercise training accounted for greater than 50% of the total training period'.
Types of outcome measures—secondary outcomes: Additional outcomes reported in trials (but not prespecified for this review) were included in the secondary outcome measures list in this review for future updates. These include body composition, attendance, preference for exercise training mode and arterial blood gases.
Subgroup analysis: This analysis could not be performed for disease severity, as we did not identify any trials with data presented according to disease severity.
Sensitivity analysis: This was not performed and funnel plots were not constructed because of the small number of included studies. If in future updates more studies are included, these analyses will be performed.
Medical Subject Headings (MeSH)
Breathing Exercises [methods]; Exercise Therapy [*methods]; Exercise Tolerance; Hydrotherapy [adverse effects; *methods]; Pulmonary Disease, Chronic Obstructive [*rehabilitation]; Quality of Life; Randomized Controlled Trials as Topic
MeSH check words
* Indicates the major publication for the study