Aquatic exercise training and stable heart failure

The following research the aquatic exercise training and stable heart failure is from Julie Adsett, Norman Morris, Jennifer Paratz, Heart Failure Service and Physiotherapy Department, Royal Brisbane and Women’s Hospital, Brisbane, Australia; School of Allied Health Sciences, Griffith University, Gold Coast, Australia; Alison Mudge, Department of Internal Medicine and Aged Care, Royal Brisbane and Women’s Hospital, Brisbane, Australia and Suzanne Kuys, School of Physiotherapy, Australian Catholic University, Brisbane, Australia; and was published in 2015 by International Journal of Cardiology.


Aim: A meta-analysis and review of the evidence was conducted to determine the efficacy of aquatic exercise training for individuals with heart failure compared to traditional land-based programmes.

Methods: A systematic search was conducted for studies published prior to March 2014, using MEDLINE, PUBMED, Cochrane Library, CINAHL and PEDro databases.

Key words and synonyms relating to aquatic exercise and heart failure comprised the search strategy. Interventions included aquatic exercise or a combination of aquatic plus land-based training, whilst comparator protocols included usual care, no exercise or land-based training alone.

The primary outcome of interest was exercise performance.

Studies reporting on muscle strength, quality of life and a range of hemodynamic and physiological parameters were also reviewed.

Results: Eight studies met criteria, accounting for 156 participants. Meta-analysis identified studies including aquatic exercise to be superior to comparator protocols for 6-minute walk test (p b 0.004) and peak power (p b 0.044).

Compared to land-based training programmes, aquatic exercise training provided similar benefits for VO2peak, muscle strength and quality of life, though was not superior.

Cardiac dimensions, left ventricular ejection fraction, cardiac output and BNP were not influenced by aquatic exercise training.

Conclusions: For those with stable heart failure, aquatic exercise training can improve exercise capacity, muscle strength and quality of life similar to land-based training programmes.

This form of exercise may provide a safe and effective alternative for those unable to participate in traditional exercise programmes.

Woman with stable heart failure performs aquatic training.

1. Introduction

Exercise training is a recommended component of the comprehensive management of patients with heart failure (HF).

Exercise based rehabilitation programmes have consistently shown positive improvements in patient symptoms, exercise capacity and quality of life, and a possible impact upon hospital readmissions and mortality.

Traditional land-based training programmes however, may not be suitable for all patients.

The frail elderly and those with co-morbid conditions including chronic pain, orthopedic or balance disturbances for example, may find these programmes difficult, contributing to lower levels of physical activity participation.

Aquatic exercise (exercise conducted in thermoneutral [32–34 °C] water) has been proposed as a possible alternative for these patients.

The warm water and low weight bearing environment reduce pain, and using the principles of hydrodynamics, allows exercise to be undertaken which may improve postural stability, exercise capacity and walking endurance.

Historically, aquatic exercise has not been recommended for individuals with heart failure.

Immersion in warm water leads to an increase in venous return as a consequence of hydrostatic pressure.

Clinicians have long been concerned that this increase in central blood volume and cardiac preload may not be tolerated by those with heart failure, leading to worsening of symptoms and a reduction in exercise capacity.

This potential risk has hampered trials and there are currently no clear recommendations for clinical practice.

Recent small studies however, have demonstrated that patients with stable heart failure not only tolerate immersion and exercise in this environment, but also benefit from a number of physiological sequelae.

The purpose of this systematic review and meta-analysis was to determine the effect of aquatic exercise training on a variety of functional and physiological outcome measures in this population.

Specifically, the study sought to determine functional benefits of aquatic exercise compared to usual activity and/or land-based exercise in people with heart failure.

2. Methods

2.1. Search strategy and study selection:

MEDLINE, PUBMED, Cochrane Library, CINAHL and PEDro databases were searched using the key words “aquatic exercise” OR “hydrotherapy” OR “water exercise” AND “heart failure” OR “cardiomyopathy”, OR “ventricular dysfunction”.

The search was limited to studies published prior to March 2014 in any language.

Identified titles and abstracts were independently scrutinized by two reviewers (JA and JP) and reference lists were assessed for additional relevant articles that met criteria.

Full text articles were extracted and independently reviewed by both reviewers when required, and any further disagreement was resolved

by discussion between reviewers.

Randomized controlled trials, pseudorandomized controlled trials, trials with historical controls and single group studies were included in the review.

Case studies, didactic articles and narrative reviews were excluded.

Studies were accepted if they included an aquatic exercise intervention of at least two weeks’ duration, conducted in a heated pool.

Only trials that recruited adult patients with left ventricular dysfunction (reduced or preserved ejection fraction) were accepted.

For intervention groups, training included either aquatic exercise alone or aquatic + land-based training.

Comparator groups included land based training, usual activity or no exercise.

As single group studies were included in the review, it was also possible for there to be no comparator group.

Exercise activities conducted in spas, sauna baths or other non-hydrotherapy pools were also excluded.

Included studies were independently scored for quality by the two reviewers using the validated PEDro scale.

Original authors were contacted for clarification of material or to provide missing data when required.

2.2. Outcome measures:

The primary outcome measure was change in exercise performance measured using either peak oxygen consumption (VO2peak), six-minute walk test (6MWT) or peak power.

Secondary outcomes included change in muscle strength, cardiac dimensions, hemodynamic parameters (including cardiac output, systemic vascular resistance and blood pressure), brain natriuretic peptide (BNP) and quality of life (QoL).

2.3. Data synthesis:

Comprehensive Meta-analysis software™ was used to compare results between studies.

For continuous variables, effect size for each individual study was determined using the Hedges g model, by calculating the difference between changes in the intervention group and comparator group by the pooled standard deviation.

The data were pooled using the fixed effect model, however when heterogeneity was statistically significant, (Q statistic p b 0.01), the data were reanalyzed using the random effects model.

Meta-analyses were conducted for VO2peak, peak power and 6MWT.

Due to insufficient studies reporting on specific outcomes, heterogeneity of methodology and lack of availability of some raw data, meta-analysis was not possible on other parameters.

3. Results

3.1. Studies included in the review:

Of the 73 papers identified, 37 were immediately excluded on the basis of duplication.

An additional 12 abstracts were excluded based upon criteria, leading to a review of 24 full text articles.

Of these, 16 were excluded for the purposes of not being clinical trials, investigating alternative outcome measures or for not including a water based exercise intervention of greater than two weeks’ duration.

As demonstrated in the flow diagram in Fig. 1, this systematic review and meta-analysis was conducted on the eight remaining studies, which included five randomized controlled trials (RCTs), two pre–post test design studies without a control group and one prospective cohort intervention with follow-up.

Figure 1: Flow diagram of included studies heart failure.

Six of the eight studies originated from two research Centre’s, thus decreasing generalizability.

All studies were published in English.

3.2. Study quality and participants:

Quality scores for included studies are listed in Table 1.

Table 1: PEDro scores for included-studies heart failure.

Main concerns included lack of assessor blinding, allocation to groups not being concealed, no documented evidence of outcome measures being obtained from N85% of participants and no documented evidence of “intention to treat” analysis.

Baseline characteristics for the eight included studies are depicted in Table 2.

Table 2: Baseline characteristics of included studies heart failure.

Studies were relatively small (n=12–25), and the overall total number of participants was 156.

The study groups were well matched for participant numbers, age, gender, left ventricular ejection fraction (LVEF) and heart failure etiology.

Participants were predominantly male (n = 139) with mean age 52–70 years.

All participants had heart failure with reduced ejection fraction (HFREF) which in most studies this was defined as LVEF b40–45%, whilst one study recruited only those with left ventricular ejection fraction b35%.

Etiology of heart failure was defined in all but one study, and included ischemic (n = 95) and dilated (n = 39) cardiomyopathies.

Studies mostly recruited participants with NYHA II–III symptoms.

Participants with co-morbid disease, including orthopedic or neurological conditions, were excluded from three of the eight studies.

3.3. Study design:

Training parameters for each study are reported in Table 3.

Table 3: Exercise training protocols for included studies heart failure.

All protocols were undertaken by experienced exercise professionals at local health facilities with water temperature maintained between 30 °C and 34 °C.

Depth of immersion varied and was either fixed (1.3 m), or individualized to the level of the xiphisternum or the neck. Immersion depth was not defined in one study and humidity was not reported in any of the included studies.

Training protocols varied between studies.

The aquatic intervention was confined to a water programme in only half of the studies with the exercise duration for these being 45min.

In the remaining four studies, the intervention included a combination of both aquatic and land-based training, the latter being 30 min of cycle ergometry.

Heterogeneity also existed for the length of the programme, varying from three weeks to 24 weeks.

In each of the short duration programmes, participants trained for either four or five days per week, compared to three times per week in longer programmes.

Protocols predominantly consisted of endurance training prescribed at an intensity of 40–70% maximum heart rate reserve (HRR) or 50–70% V̇O2peak. Resistance training was included in two studies.

3.4. Comparator protocols:

Comparator protocols also varied.

For the four studies that employed an isolated aquatic intervention, two were compared to usual activity, one used participants as self-controls following a period of no exercise and there was no comparator in one study.

For the combined interventions (aquatic + land training), three of these studies compared outcomes to a land-based intervention alone, whilst the remaining study used self-controls as the comparator.

In each of the combined interventions (aquatic + land), exercise intensity was matched across the two environments and total duration of exercise was equal for participant groups.

3.5. Summary of findings:

Meta-analysis was conducted for the primary outcomes of VO2peak, peak power and 6MWT (Fig. 2).

Figure 2: Change in VO2 peak and 6MWT for aquatic exercise training versus comparator protocols.

Secondary outcomes were unsuitable for meta-analysis but relevant study findings are summarized in Table 4 and Supplementary Table 1.

3.6. VO2peak, 6MWT and peak power:

Six studies measured V̇O2peak, including four randomized controlled trials.

Meta-analysis was conducted on the three studies that provided raw data, and revealed no significant difference between the aquatic and comparator groups (p b 0.87) (Fig. 2).

Figure 2: Change in Peak power for aquatic exercise training versus comparator protocols.

Review of individual studies suggested that when compared to controls who undertook usual activity, aquatic exercise led to a significant improvement in V̇O2peak, (between group difference p b 0.001 and p = 0.02 respectively).

When compared to land- based training alone, programmes that included a combined aquatic + land programme, elicited similar changes in V̇O2peak but were not superior.

In these latter two trials, a combined aquatic + land intervention induced an improvement in V̇O2peak by 6–11%.

The land only protocols in these same two trials led to an improvement of 10.5–11%.

Four studies reported 6MWT but only two were suitable for meta-analysis.

One further study did not provide raw data and one had no comparator group.

Overall, the analysis favored the aquatic intervention (p b 0.004) (Fig. 2).

Of these studies, one compared an aquatic intervention to usual activity, however it was the study that compared a combined protocol to a land only training that elicited the greatest effect size.

The study which did not provide raw data also reported a significant improvement in 6MWT compared to usual care.

For peak power data, meta-analysis conducted on three randomized controlled trials that included this parameter favored the aquatic intervention (p b 0.044) (Fig. 2).

One study compared an aquatic intervention to usual activity and the remaining two compared a combined training protocol (aquatic + land) to land based training alone.

Compared to usual activity, aquatic exercise significantly improved peak power (p b 0.05).

Compared to land-based training protocols however, the impact of aquatic exercise was less clear.

Teffaha et al., 2011 demonstrated greater benefit with aquatic compared to land-based training alone (between group difference p b 0.05), whilst Mourot et al., 2009, found a significant within group difference for both groups, but no between group difference.

3.7. Muscle strength:

Muscle strength was measured in only three of the eight included studies (Table 4).

Table 4: Muscle strength outcomes in aquatic exercise on heart failure.

Two of these studies (conducted by the same group), included a variety of lower limb isokinetic and isometric parameters, as well as measures of muscle endurance and grip strength.

When compared to usual activity, aquatic exercise significantly improved upper limb and lower limb endurance measures.

Whilst there was a trend favoring the aquatic intervention for measures of isokinetic knee extension peak torque, results did not consistently reach significance.

Grip strength and isometric lower limb peak torque measures did not significantly improve.

The third study investigating muscle strength, found a significant improvement in maximum voluntary contraction (MVC) of the quadriceps for both combined (aquatic + land) and land only exercise groups following an endurance training intervention (p b 0.05), however no between group difference.

Similarly, there was no significant difference between the aquatic and land interventions for measures of quadriceps peak torque. No resistance training was employed in this study.

3.8. LVEF and cardiac dimensions:

Aquatic exercise training did not significantly influence LVEF in three of the four studies that reported this parameter by comparison, Teffaha et al., did show a significant improvement in LVEF following a three-week training period (p b 0.05), with a slightly greater effect size elicited in the combined aquatic and land group compared to land only intervention (0.44 vs 0.21).

The effect of aquatic exercise training on cardiac dimensions is difficult to discern due to variation in study methodology and reporting.

Transthoracic echocardiography was used to record baseline and post programme cardiac dimensions in four studies. Results were reported according to left ventricular end diastolic and systolic diameters (LVEDD and LVESD) by two authors, whilst volume measures were reported by the remaining two authors.

Aquatic exercise did not significantly influence any of these parameters following the exercise training periods.

Whilst the exercise interventions did not significantly influence cardiac dimensions over the long term, immersion itself did have an acute impact upon these measures.

This ensues due to the increase in venous return that occurs as a consequence of hydrostatic pressure and was demonstrated by Svealv et al., 2009 who recorded cardiac volumes during immersion, both before and after aquatic exercise training.

Left ventricular end diastolic and systolic volumes (LVEDV and LVESV) both significantly increased during immersion compared to measures recorded on land at these same time points.

3.9. Hemodynamic parameters:

Interpreting the effect of aquatic exercise training on hemodynamic parameters is difficult, as results are influenced not only by variations in study design, but also the environment in which parameters are measured.

The majority of studies reported measures conducted on land, whilst others also included immersion measures.

To enable more suitable comparison of hemodynamic measures, studies listed below are summarized in Supplementary Table 1.

3.9.1. Land measures:

Following a period of aquatic exercise training, resting systolic blood pressure (SBP) does not appear to change when measured on land, but there may be an impact upon resting diastolic blood pressure (DBP).

Caminiti et al, reported a significant reduction in DBP following a 24-week combined aquatic + land intervention, which was not observed in those undertaking the land only training programme (between group difference p b 0.05). Shorter duration studies however, have not elicited similar results.

With regard to resting heart rate (HR), aquatic exercise training does not appear to provide additional benefit to land-based training.

Of the five studies that measured this parameter, only two demonstrated a reduction in HR following the intervention, and one demonstrated a significantly greater effect in the aquatic group (p b 0.05).

These studies also showed an increase in resting stroke volume (SV).

Of the four studies that measured systemic vascular resistance (SVR), three yielded no change following the intervention period. The study by Caminiti et al., however, reported a significant reduction following the 24-week intervention, favoring the aquatic group (between group difference p b 0.05).

In this study, a combined training approach (aquatic + land training) was compared to land based training alone.

Cardiac output was not observed to significantly change following any exercise intervention for studies included in this review.

3.9.2. Immersion measures:

In addition to the decrease in resting HR observed with acute immersion, further reductions are observed following a period of aquatic exercise training, when this parameter is measured during immersion.

Compared to land-based training however, this reduction in heart rate is no greater than that observed following land-based training, measured in this same environment.

Similar results are observed for increases in resting SV measured during immersion.

Measured during immersion, resting SBP, diastolic blood pressure and cardiac output do not appear to be influenced by aquatic exercise training.

3.10. Natriuretic peptide and plasma nitrate:

N-Terminal pro-brain natriuretic peptide (NT-proBNP) and BNP were measured prior to and following the exercise training protocol in three studies.

No significant difference was recorded in this parameter following the training period for any of these studies.

Plasma nitrate was reported by one author, who described a significant increase following the aquatic intervention (p b 0.05), whilst no significant change was observed in the land only training group.

Comparison of change between the two groups was not reported.

3.11. Quality of life:

Quality of life was measured in three studies using the Minnesota living with heart failure questionnaire (MLWHFQ).

Whilst a significant improvement (p= 0.01) was recorded in the aquatic intervention group in two of these studies, aquatic exercise was not superior to usual activity with regards to quality of life.

3.12. Safety:

No serious adverse events were reported in any of the studies included in this review.

Only one study documented negative consequences of the aquatic intervention.

These included a new episode of arrhythmia in an individual with paroxysmal atrial fibrillation and episodes of mild fatigue in three individuals undertaking the first two weeks of their training protocol.

4. Discussion

The findings from this review indicate that aquatic exercise training is beneficial for individuals with stable HF, as it improves exercise capacity, muscle strength and quality of life to a similar extent as land-based exercise training.

Some hemodynamic measures such as resting diastolic blood pressure, heart rate and systemic vascular resistance may also improve following this intervention.

A combined approach including aquatic exercise in addition to land-based training, may even be superior to land-based training alone for 6MWT and peak power.

The primary outcome in this review was exercise performance.

Of the small number of studies included, meta-analysis of aquatic exercise training versus alternative protocols, favored the aquatic intervention for 6MWT and peak power and demonstrated no significant impact upon V̇O2peak.

As 6MWT is a commonly used measure of performance in these patients, this outcome seems very encouraging.

However, these results are based upon only two studies and a limited number of participants, and as such should be interpreted with caution and should not be extrapolated to all individuals.

Land-based training programmes have been extensively studied in HF participants and improvements in V̇O2peak post intervention are commonly cited as being 15–25%.

Interestingly, smaller changes were reported in the studies included in this review for both the aquatic and land interventions (6–11% and 10.5–11% respectively).

This may suggest conservative or insufficient exercise volume or intensity in these studies.

In each study, exercise intensity was prescribed according to heart rate reserve and these same targets were used for both the aquatic and comparator protocols.

It is possible that heart rate may not be the best indicator of exercise intensity given the acute hemodynamic effects of immersion on this parameter.

Determining exercise intensity during aquatic exercise remains a challenge for the clinician.

Whilst aquatic exercise appears to evoke greater strength gains than usual care, its benefit compared to land-based training protocols is less certain.

In the one study that compared these two training environments, no resistance training was included in the protocol.

One might assume that during aquatic exercise, the water itself may provide a degree of resistance and consequent strength gains; however, experts in this field stress the importance of a strong working knowledge of the physiological principles of hydrodynamics to achieve optimal results.

Consistent with previous studies which have demonstrated that resistance training is the best modality for improving muscle strength in this population, merely transferring land-based exercises into the aquatic environment is not sufficient.

Best results would likely ensue when specific resistance exercises are employed.

From the studies, currently available, aquatic exercise does not appear to significantly influence cardiac dimensions or LVEF.

This is not an unexpected finding, given the relatively short duration of most interventions. In the one study of long duration included in this review, a number of positive hemodynamic benefits were reported that favored the aquatic intervention.

These included a reduction in resting diastolic blood pressure and heart rate and an increase in resting stroke volume post programme.

In this study, Caminiti et al., proposed that aquatic exercise may provide a greater stimulus for central adaptive changes, particularly with improved hemodynamics at rest.

When combined with the peripheral physiological adaptations that are well documented to occur with land-based training, a combined training approach may elicit even greater physiological and functional benefit.

This is yet to be confirmed however, and has not been observed in shorter duration studies.

No serious adverse events were reported for any of the 156 participants included in this review and adherence rates were listed between 92 and 95%, indicating a high acceptance in the selected trial participants.

Quality of life improved in the few studies that included this parameter though was not superior to land-based training. There are several limitations to this review.

Firstly, trials were small with carefully selected participants and results cannot be extrapolated to all individuals with HF.

Svealv et al., 2012 highlights potential concerns for those with biventricular failure and pulmonary hypertension.

This case report suggests that potential subgroups may exist for whom aquatic exercise training may not be recommended and at this time clinicians should use good clinical judgement when recommending this form of exercise for patients with HF.

Decompensated HF remains an absolute contra-indication for aquatic exercise training.

In addition to the small number of trials included in this review, three of the eight studies were not RCTs, and one did not include controls.

Raw data were not available for all measures.

Participants were predominantly males, had heart failure with reduced ejection fraction, and etiology was primarily ischemic.

Individuals with orthopedic or neurological co-morbidities were often excluded, yet these are the very patients who might benefit most from this intervention.

Despite these limitations, this review indicates that aquatic exercise training improves exercise performance in individuals with stable heart failure and may provide a safe and acceptable alternative for those unable to access or participate in traditional land-based programmes.

Further research should target those who are typically seen in clinical practice such as those with HF with preserved ejection fraction, the frail elderly, and those with significant co-morbidities.

Trials should investigate the impact of this form of exercise on other parameters pertinent to heart failure such as diuresis, thermoregulation, fatigue and respiratory muscle function, all of which may be influenced by exercise training in this environment.

Training intensity should be measured to inform prescription.

Finally, acceptability of aquatic exercise training for these individuals should also be explored.

5. Conclusion

For individuals with stable HF, aquatic exercise training improves exercise capacity, muscle strength and quality of life similar to land based training protocols.

Aquatic exercise training may provide a safe and effective alternative to traditional land-based programmes for appropriately selected patients.

Results cannot be extrapolated to all individuals with heart failure.

Future research should enroll more representative samples to identify feasible training protocols to inform translation of this research into everyday practice.

Woman with stable heart failure, doing Kaenz aquatic exercise in swimming pool.

Kaenz invitation

If you finished reading this post, most likely you will love, as we do, everything related to aquatic therapy, hydrotherapy and pool exercises. For this reason, we invite you to increase your income, becoming a Kaenz Representative in your city.

Notify of
Inline Feedbacks
View all comments