Effectiveness of Deep Water Running on Improving Cardiorespiratory Fitness, Physical Function and Quality of Life: A Systematic Review

Deep Water Running (DWR) is a form of aquatic aerobic exercise simulating the running patterns adopted on dry land. Little is known on the effectiveness of DWR despite gaining popularity. The objective of this study is to systematically review the effects of DWR on cardiorespiratory fitness, physical function, and quality of life in healthy and clinical populations. A systematic search was completed using six databases, including SPORTDiscus, MEDLINE, CINAHL, AMED, Embase, and The Cochrane Library, up to February 2022. Eleven studies evaluating the effectiveness of DWR on cardiorespiratory fitness (CRF), physical function, or quality of life (QoL), compared with no interventions (or standard treatment) or land-based trainings were identified. Data relevant to the review questions were extracted by two independent reviewers when means and standard deviations were reported, and standardized mean differences were calculated. A quality assessment was conducted using selected items from the Downs and Black checklist. A total of 11 clinical trials (7 randomized controlled trials) with a total of 287 participants were included. Meta-analyses were not completed due to the high level of clinical and statistical heterogeneity between studies. Compared with land-based training, DWR showed similar effects on CRF with limited studies reporting outcomes of physical function and QoL compared with the no-exercise control group. DWR appears to be comparable to land-based training for improving CRF. The aquatic environment may provide some advantages for off-loaded exercise at high intensity in populations that are weak, injured or in pain, but more studies are required.


Introduction
Deep water running (DWR) has gained popularity among different patient populations in recent decades for cardiorespiratory conditioning. DWR can be defined as running with 70% of the body immersed, submerged at shoulder level, with or without the use of a floatation device [1]. It simulates land running movement patterns performed but is conducted in a non-weight-bearing aquatic environment. With the advantage of the unloading property of buoyancy, DWR eliminates vertical ground reaction forces, and hence reduces joint loadings and the potential risk for injury to the musculoskeletal system [2,3].
Changes in the properties of water, including temperature, as well as the instructions for training, can influence the outcomes of DWR. For instance, water temperature can directly influence the physiological mechanisms brought about by DWR. A temperature of 26 • C to 28 • C is recommended for DWR performed at maximal intensities [1]. This included in the review. If in disagreement, a third reviewer (SH) resolved any discrepancies between the two reviewers, and the article was removed from the review.

Search Strategy
This systematic review was guided by the Preferred Reporting Items for Systematic reviews and Meta-analysis (PRISMA) guidelines [15], and was registered in PROS-PERO database (CRD42020154988) prior to conducting the review. A complete search of six databases: SPORTDiscus, MEDLINE, CINAHL, AMED, Embase, and The Cochrane Library, were conducted up to February 2022 with a combination of text words listed in Table S1. In an effort to unveil complex evidence in obscure locations, manual searches of "reference tracking" and "citation tracking" of relevant articles were completed [16]. With the use of citation tracking databases (i.e., "Cited by" in Google Scholar), the forward tracking of selected studies was processed.

Inclusion Criteria
The articles included were based on the following criteria: (1) Studies must be interventional studies, either randomised or non-randomised trials; (2) studies must be published as a full paper in English; (3) study participants were not limited to only healthy individuals (i.e., included patients with at least one health condition); (4) the interventions being evaluated were predominantly DWR (≥50%)-based or combined with passive treatments (e.g., education or stretching); and (5) outcomes included cardiorespiratory measures (e.g., VO 2 max, maximal heart rate, blood pressure, ventilatory threshold), physical function related to running or walking (e.g., timed test, endurance test, or functional test), or health-related QoL (validated questionnaires or assessment tools) (Table S2).

Exclusion Criteria
The studies were excluded if: (1) The study designs were non-experimental (e.g., descriptive or exploratory studies) and/or cross-sectional studies; (2) participants were under 18 years of age; (3) they studied exercises other than DWR (e.g., general aquatic exercise, shallow water running, underwater treadmill) (Table S2).

Data Extraction
Data were extracted independently by 2 reviewers (MK and BS) and included study design, baseline demographics of study participants (age, sex, health status, body weight, height, body mass index), sample size, intervention details (frequency, intensity, duration, length of intervention, DWR technique, instructions, supervision of exercise intervention, equipment used, pool temperature, and room temperature), outcomes of interest (measurements on cardiorespiratory outcomes, physical functional outcomes and health-related QoL), and measures of variance of the outcomes of interest. If there were missing data or data potentially included in error, the study authors were contacted via email and asked to provide further information.

Quality and Risk of Bias Assessment
Each included study was critically appraised by two independent reviewers (MK and BS) using selected items from a checklist based on Downs and Black (Table 1) [17]. In the preliminary search of articles, relevant studies of multiple study designs were drawn. The Downs and Black quality assessment tool designed for randomized and non-randomized trials was chosen to assess the quality. Despite there being limitations, it has been shown to be an effective tool irrespective of the tool development and extensive domains covered [18]. Five subscales, including reporting, external validity, internal validity in bias, confounding, and power were assessed according to the characteristics of reviewed studies. Scores for this tool can be interpreted as follows: Excellent (26)(27)(28); good (20)(21)(22)(23)(24)(25); fair (15)(16)(17)(18)(19); and poor (≤14) [17].
Estimates of the random variability for the main outcomes provided

Data Analysis
The outcomes of interest and program parameters in specific populations were pooled for data analysis using Cochrane Software Review Manager (RevMan software, version 7.1; The Nordic Cochrane Centre, Copenhagen, Denmark). The effects of DWR on the three domains of interest were analysed with standardized mean difference (SMD) calculations if the studies reported means and standard deviations [29]. The SMD was significant only if the 95% CI did not include zero [30]. Effect-size thresholds were classified as small effect (0.2), medium effect (0.5), or large effect (0.8) [29]. The SMD and 95% confidence intervals (CI) were reported to demonstrate effects of DWR in comparison with a control group (without exercise or other interventions such as land-based training). Meta-analyses were not completed due to Downs and Black poor-quality score of (≤14) [10,19,20,[25][26][27][28]. Therefore, results are presented as effect sizes only.

Selection of Studies
Following an initial search identification of 1416 articles, 11 articles were included in the review ( Figure 1) for quantitative synthesis.

Data Analysis
The outcomes of interest and program parameters in specific populations were pooled for data analysis using Cochrane Software Review Manager (RevMan software, version 7.1; The Nordic Cochrane Centre, Copenhagen, Denmark). The effects of DWR on the three domains of interest were analysed with standardized mean difference (SMD) calculations if the studies reported means and standard deviations [29]. The SMD was significant only if the 95% CI did not include zero [30]. Effect-size thresholds were classified as small effect (0.2), medium effect (0.5), or large effect (0.8) [29]. The SMD and 95% confidence intervals (CI) were reported to demonstrate effects of DWR in comparison with a control group (without exercise or other interventions such as land-based training). Meta-analyses were not completed due to Downs and Black poor-quality score of (≤14) [10,19,20,[25][26][27][28]. Therefore, results are presented as effect sizes only.
All the studies instructed participants to perform running in the water; however, there were variations in body positions during DWR. Four studies required participants to maintain a vertical body position during DWR [20,22,26,28]. Davidson and McNaughton (2000) allowed participants to perform DWR in a vertical or slightly forward leaning position, while Broman et al. (2006) and Michaud et al. (1995) instructed participants to run in a slightly forward bending position [10,19,25]. Three studies did not specify the running positions adopted [21,23,24,27].

Cardiorespiratory Fitness
Peak Heart rate (HR peak) Three studies evaluated the changes in HR peak [10,25,27], but did not report means and SD; therefore, SMDs could not be calculated.

Physical Function
Deep Water Running Versus Control (no active exercise) One study, Alberti et al. (2017) evaluated physical function between the DWR group and control group with three walking tests (4MWT, 6MWT, 10MWST), and two functional mobility and dynamic balance tests: Five Times Sit to Stand Test (FTSST) and Timed Up and Go Test (TUGT) [24]. Large effect sizes of DWR were observed in the three walking tests (SMD = −2.39 to −1.33). Effect sizes revealed no significant effect of DWR compared to the no active exercise for FTSST and TUGT (SMD = −0.49 to −0.36), although there were mean improvements favouring the DWR group in this study (Figure 4).

Physical Function
Deep Water Running Versus Control (no active exercise) One study, Alberti et al. (2017) evaluated physical function between the DWR group and control group with three walking tests (4MWT, 6MWT, 10MWST), and two functional mobility and dynamic balance tests: Five Times Sit to Stand Test (FTSST) and Timed Up and Go Test (TUGT) [24]. Large effect sizes of DWR were observed in the three walking tests (SMD = −2.39 to −1.33). Effect sizes revealed no significant effect of DWR compared to the no active exercise for FTSST and TUGT (SMD = −0.49 to −0.36), although there were mean improvements favouring the DWR group in this study ( Figure 4).  Deep Water Running Versus Land Training Two studies measured running outcomes, reporting similar effects on physical function between DWR and land training groups ( Figure 5). Mckenzie et al. (1991) found a similar effect on the time to reach volitional exhaustion (i.e., time to fatigue) while running on a treadmill in DWR and land running groups (SMD = 0.82) [26]. Eyestone et al. (1993) found a similar effect of DWR to both land-running and cycling for two-mile run time (SMD = −0.41 to −0.27) [28]. Deep Water Running Versus Land Training Two studies measured running outcomes, reporting similar effects on physical function between DWR and land training groups ( Figure 5). Mckenzie et al. (1991) found a similar effect on the time to reach volitional exhaustion (i.e., time to fatigue) while running on a treadmill in DWR and land running groups (SMD = 0.82) [26]. Eyestone et al. (1993) found a similar effect of DWR to both land-running and cycling for two-mile run time (SMD = −0.41 to −0.27) [28].

Discussion
This systematic review found similar effects of DWR and land-training on improving VO2 max in active individuals, sedentary participants and people with chronic health conditions. Similar effects were also found for physical function and QoL outcome measures for water-based and land-based training; however, the conclusions for these domains are weaker due to the smaller number of studies. There were few studies reporting the effect of DWR to no-exercise across all three domains of interest with mixed results. Heterogeneity and reporting of interventions, as well as the varied types of participants, limits stronger conclusions.
The results of the review reflect that DWR can be an equally effective form of exercise to maintain or improve CRF when compared with land-based training. VO2 max is considered to be the highest value of VO2 attained upon an exercise stress test [31]. The test of VO2 max is designed to bring the subject to the limit of exhaustion, a gold-standard measure of CRF [32]. VO2 max can be influenced by a restriction in chest expansion when inspiratory muscle contractions are unable to equal or overcome the force of hydrostatic pressure under immersion [33]. It is suggested that central hypervolaemia, or the shift of blood into the chest cavity, reduces lung volume, as well as reducing lung compliance and promoting gas trapping, which narrows the airways [5,34]. Compression of both the abdomen (in turn pushing up the diaphragm) and the chest wall itself by water also in-

Discussion
This systematic review found similar effects of DWR and land-training on improving VO2 max in active individuals, sedentary participants and people with chronic health conditions. Similar effects were also found for physical function and QoL outcome measures for water-based and land-based training; however, the conclusions for these domains are weaker due to the smaller number of studies. There were few studies reporting the effect of DWR to no-exercise across all three domains of interest with mixed results. Heterogeneity and reporting of interventions, as well as the varied types of participants, limits stronger conclusions.
The results of the review reflect that DWR can be an equally effective form of exercise to maintain or improve CRF when compared with land-based training. VO2 max is considered to be the highest value of VO2 attained upon an exercise stress test [31]. The test of VO2 max is designed to bring the subject to the limit of exhaustion, a gold-standard measure of CRF [32]. VO2 max can be influenced by a restriction in chest expansion when inspiratory muscle contractions are unable to equal or overcome the force of hydrostatic pressure under immersion [33]. It is suggested that central hypervolaemia, or the shift of blood into the chest cavity, reduces lung volume, as well as reducing lung compliance and promoting gas trapping, which narrows the airways [5,34]. Compression of both the abdomen (in turn pushing up the diaphragm) and the chest wall itself by water also in-

Discussion
This systematic review found similar effects of DWR and land-training on improving VO 2 max in active individuals, sedentary participants and people with chronic health conditions. Similar effects were also found for physical function and QoL outcome measures for water-based and land-based training; however, the conclusions for these domains are weaker due to the smaller number of studies. There were few studies reporting the effect of DWR to no-exercise across all three domains of interest with mixed results. Heterogeneity and reporting of interventions, as well as the varied types of participants, limits stronger conclusions.
The results of the review reflect that DWR can be an equally effective form of exercise to maintain or improve CRF when compared with land-based training. VO 2 max is considered to be the highest value of VO 2 attained upon an exercise stress test [31]. The test of VO 2 max is designed to bring the subject to the limit of exhaustion, a gold-standard measure of CRF [32]. VO 2 max can be influenced by a restriction in chest expansion when inspiratory muscle contractions are unable to equal or overcome the force of hydrostatic pressure under immersion [33]. It is suggested that central hypervolaemia, or the shift of blood into the chest cavity, reduces lung volume, as well as reducing lung compliance and promoting gas trapping, which narrows the airways [5,34]. Compression of both the abdomen (in turn pushing up the diaphragm) and the chest wall itself by water also increases the work of breathing [5,35]. With such unique physiological adaptations during aquatic training, minute ventilation and breathing frequency could potentially be increased when compared with land-based training with an equivalent exercise intensity [36]. Additionally, it is suggested that by performing movements in DWR, the physiological effects of water immersion contribute to a reduction in joint loading, while the tactile, thermal stimulation and drag force may enhance joint proprioception, body balance and muscle strength [22]. Since aquatic-based maintenance of cardiorespiratory conditioning offers a number of additional advantages over land-based training, it may be ideal for populations unable to exercise on land, or those who exclusively train on land, and desire to cross-train in water for rehabilitative purposes.
This review found that DWR is more effective in untrained, sedentary healthy elderly populations [10,25] than trained or physically active subjects [19,20,[26][27][28]. One possible explanation of DWR improving aerobic capacity in sedentary individuals and elderly populations is their low initial fitness [1]. Of note, such a substantial improvement likely has more significant clinical relevance in sedentary healthy elderly populations than trained or physical active subjects, given that cardiorespiratory function decreases with primary ageing, and that CRF declines steadily in sedentary individuals at a rate of approximately 1% per year after the third decade of life [37]. Such findings are in agreement with a previous study by Reilly et al. (2003) [1], and other systematic reviews by Chu and Rhodes (2001) [12] and Jorgic et al. (2012) [38]. As such, CRF training, for instance DWR, promotes improvement in CRF and has shown important clinical implications for sedentary populations, particularly in individuals with advanced age.
A key component of aquatic-based training is carryover to land function as well as broader QoL improvements. One of the aims of this review was to analyse the effect of DWR compared to land-based walking and running. This review found a large effect for DWR in walking tests (4MWT, 6MWT and 10MWT) when compared with the no-active-exercise control groups, but only one study reported this [24]. Similarly, few data on the effect of DWR on quality were found. When DWR was compared with the no-active-exercise control group, DWR showed a significant effect upon physical health and mental health in SF-12, but only in one study [23]. There is a potential for QoL to be influenced via changes in physical function or to improve well-being by promoting relaxation, vasodilation, and a reduction in joint loading, and to produce an analgesic effect [22]. DWR may also release cortisol and adrenaline into the bloodstream, thus increasing the pain threshold for those subjects who suffered from pain [23], along with the removal of metabolic waste and reduction in nociceptor activation [39]. More evidence is required to understand if DWR has enough of an effect to improve both land-based function and QoL.

Study Limitations
Although the synthesized evidence in this review is encouraging, this study has several limitations. Firstly, the majority of the included studies had small sample sizes, and the recruitment of participants was from convenience sampling. This may have increased the chance of committing a type II error, affecting the results of efficacy regarding subjects' representativeness and generalisability. Secondly, heterogeneity of participants' characteristics also made it difficult to draw definitive conclusions about clinical outcomes. Furthermore, a meta-analysis was not completed due to the clinical diversity in study interventions, variations in methodology, and risk of bias in individual studies. For instance, there were a wide range of DWR protocols and methods of measurements for the main outcomes; that is, the measurements used or methods in measuring VO 2 max were distinct across some studies. This limits the ability of this review to conclude a recommended dosage of DWR for participants. As a result, a limited number of studies on homogeneous groups of subjects have been conducted in this review. Lastly, the lack of a meta-analysis hinders a precise estimation of effect and a statistical analysis of DWR. More clinical trials in specific populations with larger sample sizes could help yield more solid conclusions and consistency on the effectiveness of DWR.

Future Directions
Conducting correlational studies could be useful with the application of biopsychosocial approaches, to investigate the relationships between measured outcomes and DWR with a larger and more homogenous group of participants. This could further justify the clinical significance of DWR in specific groups of populations. Additionally, the studies measured short-term effects immediately after DWR sessions. However, long-term effects could be considered after completion of intervention to evaluate the carryover effects. Currently, there is a lack of evidence of participants' experience and perceptions towards DWR programs. Qualitative research is therefore suggested to explore the attitudes, behaviours, beliefs, and satisfaction towards effects of DWR to further consider their exercise adherence in addition to quantitative data on both land-based function and QoL. It is also difficult to conclude optimal dosages of a training programme; therefore, future investigations on the programme intensity are warranted.

Conclusions
DWR appears to be comparable to land-based training, with mixed results of effects compared to no exercise for improving CRF, physical functions and QoL. The small number of studies and quality of the evidence limits further conclusions. To further understanding of the potential benefits of DWR, future research is needed to develop an effective prescription for targeted populations.