Framework of a scoping review

The procedure of performing a scoping review follows similar steps as those used in systematic review approaches without limiting for study design of included studies and without a quantitative synthesis. The framework of Arksey and O’Malley for scoping reviews was implemented in this study [9]. The framework consists of five essential stages and one additional stage; 1) identifying the research question, 2) identifying relevant studies, 3) study selection, 4) charting the data, 5) collating, summarizing and reporting the results, and additionally 6) consultation of experts (optional). All stages can be performed in an iterative manner allowing refining of search parameters.

Identifying relevant studies

A comprehensive literature search was conducted in August 2015 and updated to 30^th September 2016 in seven electronic academic databases (PubMed, MEDLINE, Cinahl, Embase, PEDro, Web of Science, WorldCat). The search strategy was documented by title of the database searched, date of the search, the complete search string that was used and the number of articles found (Table 1). The development of each search string was an iterative process and familiarisation with the literature revealed additional search terms for aquatic cycling such as “immersed cycling” or “underwater pedalling”. These terms were combined with more general terms for aquatic therapy (e.g. hydrotherapy) the search included the following key terms: ergometer, immersion, hydrotherapy, aqua(tic), cycling, underwater (bi)cycle ergometer, immersed ergocycle.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Search strategy and results.

https://doi.org/10.1371/journal.pone.0177704.t001

Additionally, ResearchGate, GoogleScholar and relevant aquatic therapy websites (http://www.wcpt.org/apti, http://www.atri.org, https://www.aeawave.com) were examined. Moreover, the table of contents of the accessible key journals ‘International Journal of Aquatic Research and Education’ and ‘Journal of Aquatic Physical Therapy’ of the American Physical Therapy Association were checked for additional literature. Finally, reference lists of all included articles were hand-searched for new articles and the authors of this paper, all experts in the field of aquatic therapy and aquatic fitness, checked their own libraries for additional literature. The table of contents and reference lists were screened for the key words related to cycling and (immersion) exercise (testing) on land and in water. Throughout the search process it was noticed that no consistent terminology exists with regard to aquatic cycling. To ensure that the search terms used were correct and complete, the terminology used in included articles was re-evaluated. This post-hoc analysis (S1 File) addressing the terminology used to describe aquatic cycling confirmed our choice of search terms.

Study selection

The inclusion and exclusion criteria were developed in two stages. In phase one, the authors agreed to include all formats of full-text reports that focused on the effects of head-out aquatic cycling exercise on the human body (Table 2, stage one). After familiarisation with the literature the selection criteria were further specified (Table 2, stage two). In each step of the selection procedure two or more reviewers were involved and inclusion discrepancies were solved by discussion. Screening of titles and abstracts was performed by two reviewers (BW and SR) with the online programme “Covidence” (Covidence systematic review software, Veritas Health Innovation, Melbourne, Australia, available at: www.covidence.org). Next, all authors were involved with the full-text screening and all results were independently imported into a Microsoft Excel file and compared after completion of the review process. Information on the two-stage development of the inclusion criteria is available in a supporting file (S2 File).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Two stage expert consensus on inclusion and exclusion criteria.

https://doi.org/10.1371/journal.pone.0177704.t002

Charting the data

Descriptive data were extracted into Microsoft Excel tables including name of the first author, year of publication, primary research question, sample size, age, gender, health status of participants, exercise parameters, main results reported in the abstract, water temperatures, aquatic bike used and level of body immersion. Information on effects of resting immersion was not discussed for this review, but might have been part of the experimental set-up of the included studies. The tables were organised by the body position on the aquatic bike (upright and semi-recumbent), because physiological responses might vary with immersion level related to the body position on the ergometer [2]. All tables include information on interventions with healthy participants and patients. If patients were involved, information on the disease characteristics is reported in the tables. Articles that originated from the same data set, but focusing on different outcomes, were summarized and represented as one study in the tables, but references from all studies are included to aid identification of the separate articles.

Land-based cycling compared to aquatic cycling

Thirty-one studies compared aquatic cycling with land cycling (Table 3). Half of the studies (n = 15) used a maximal incremental exercise test to investigate the physiological responses during immersion versus on land exercise testing [11, 13–26]. Submaximal incremental exercise tests were conducted in six studies [27–32]. Increments were mostly achieved by an increase in pedalling frequency. Seven studies of the aforementioned studies controlled exercise intensity by electronically regulated pedalling resistance [10, 21, 24–27, 31]. An additional six studies compared submaximal continuous aquatic cycling with land cycling [33–38]. Three other studies evaluated aquatic cycling as a mean for active recovery after an extensive exercise bout on land [39–41]. Furthermore, one study compared the effect of moderate intense dryland cycling with high-intensity interval training (HIIT) on land and in water [42]. Two-third of the aquatic cycling sessions (n = 22) were conducted in an upright body position. Nine studies [11, 24–26, 30, 31, 34–36] compared semi-recumbent cycling on land and in water. Four semi-recumbent bikes also had arm pedals [11, 24, 30, 36]. The level of body immersion of the participants varied from chest level to chin level. The water temperature during the exercise sessions ranged from 18°C to 35°C.

All but three studies used a cross-over design to compare both environments. Additional study designs were a randomized controlled trial [41, 42] and a quasi-experimental study [27]. In 19 out of 31 studies participants were young, healthy males. Five studies included healthy participants of both sexes [13–16, 26, 36] and three studies included pregnant women [27, 37, 38]. In four other studies participants were middle-aged men [29], males with cardiovascular diseases [28, 32] and men and women with hypertension [42].

Studies (n = 21) investigating the difference in cardiovascular responses between aquatic versus land cycling compared oxygen consumption (VO₂), heart rate (HR), stroke volume, cardiac output and blood pressure [15–23, 26–33, 37, 38, 42]. In total eight studies investigated the maximum VO₂ response during land and aquatic cycling, with all but one study [15] reporting equivalent VO2_max values achieved by the participants on land and in water [17–23, 26]. Maximal HR was found to be lower during aquatic cycling at intensities higher than approximately 80% of the VO_2max in seven from ten studies [20–23, 26, 29, 30]. The remaining three studies reported similar maximal HR for the land and water conditions [16–18]. In men, following recovery from a myocardial infarction, no difference in submaximal HR on land and in water was found [28]. McMurray et al. reported a trend toward a lower HR at submaximal intensities in water in men with coronary heart disease [32]. In pregnant women moderate aquatic cycling resulted in lower maternal and foetal HR compared to land-based cycling [38]. Four studies reported higher stroke volume and cardiac output in the aquatic cycling group consiting of healthy participants [15, 21, 29, 30]. Systolic blood pressure was similar in healthy males during an incremental exercise test when using aquatic versus land-based cycling [18, 21]. In pregnant women and in men with coronary artery disease the systolic blood pressure was reported to be lower during submaximal aquatic cycling [27, 32, 37, 38]. Sosner et al. reported a similar post-exercise reduction in blood pressure in patients with hypertension after a high-intensity cycling session on land and in water [42].

Other key outcomes were ventilation parameters [23, 31, 33], lipid mobilisation and oxidation [18, 34], sympathoadrenal response [18, 20, 34], lactate accumulation and removal [17, 18, 20, 39, 40]. and thermoregulatory responses [35–37]. Further outcomes were the development of prediction equations to estimate oxygen consumption from pedalling rate during aquatic cycling [13, 14] and to calculate external power output of aquatic cycling [13]. Fenzl et al. compared the gas exchange measurements with the heart rate variability to estimate the ventilator threshold on an arm-leg aquatic bike [11].

Aquatic cycling under different exercise conditions

Twenty-five studies investigated the effect of several different exercise conditions during aquatic cycling (Table 4). The comparisons are based on cross-over studies with healthy young males with the exception that healthy (non-pregnant) females were included in three studies [43–45] and one study used a quasi-experimental design to compare age-matched healthy controls with heart disease patients [46]. Common core outcomes were cardiovascular [12, 23, 44, 45, 47–50], metabolic [36, 51–55] and thermal response [43, 49, 50, 52, 56–61] to different exercise conditions. Furthermore, approaches to estimate and regulate exercise intensity during aquatic cycling were evaluated [62–64].

Different exercise conditions were created mostly by changes in water temperature [12, 23, 43, 48–52, 57, 60, 61] and different exercise intensities (high versus low) [23, 43, 44, 50, 58–61, 63–65]. With regard to the exercise parameters intensity and duration, studies (n = 11) utilised continuous, submaximal exercise (40 and 60% of VO_2max) with a duration of 30 to 60 minutes [12, 23, 43, 49, 51, 52, 57, 60, 61]. Exercise intensities were either based on graded exercise testing on land [12, 23, 44, 45, 47, 50, 51, 53–56, 58] or in water [23, 43, 46, 48, 52, 57, 62–64]. The water temperatures that were compared ranged from cold (18–20°C) and cool (25°C) to thermoneutral (30–35°C). Other studies compared different levels of body immersion [46], different types of exercise (interval versus continuous cycling, arm versus arm-leg versus leg exercise) [45, 49, 61] and different aquatic bikes with each other [44]. Furthermore, the maternal and foetal response to submaximal (60% of VO_2max) aquatic cycling during different stages of pregnancy was studied [47, 53–56].

Fifteen studies used upright aquatic bikes [23, 44–47, 51–57, 62–64]. In all these studies pedalling frequency regulated exercise intensity while two studies focused on the influence of pedalling resistance provided by additional fins to the flywheel [44, 64]. Sogabe et al. used the additional fins to increase pedalling resistance in semi-recumbent cycling [65]. In all other semi-recumbent bikes intensity was set with electronically controlled pedalling resistance mechanisms [43, 48–50, 58–61].

Aquatic cycling intervention programmes

In total eight intervention studies, investigating the effects of a multiple sessions aquatic cycling exercise programme, were found [66–73]. The exercise programmes (Table 5) lasted between three and 36 weeks with an exercise frequency between two and five times per week. The duration of one session varied between 30 and 90 minutes. Exercise intensities were based on land-based maximal graded exercise-tests and the training intensities were set between 60 and 80% of the VO_2max in all but one study [66]. In a one-group test-retest study, Sheldahl et al. assessed weight loss in obese women after a low intense (30 to 40% of VO_2max) aquatic cycling programme [66]. Boidin et al. also evaluated the effects of aquatic cycling on cardiometabolic parameters in obese people [71]. In this retrospective study the participants underwent an extensive lifestyle programme including high-intensity aquatic cycling or land cycling. Furthermore, two randomised studies evaluated the cardiovascular effect of aquatic cycling compared to land cycling in young healthy males [68] and patients with multiple sclerosis [72, 73]. Two quasi-experimental studies investigated the influence of water temperature on heat tolerance and aerobic capacity [67, 69, 70].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 5. Aquatic cycling intervention programmes.

https://doi.org/10.1371/journal.pone.0177704.t005

Four studies reported a significant improvement of cardiorespiratory parameters compared to baseline in healthy (obese) people and multiple sclerosis patients [68, 71–73]. Aquatic and land cycling evoked similar improvements in cardiorespiratory parameters. Further, moderate land and aquatic cycling achieved similar improvements in health-related quality of life and self-reported physical fatigue in patients with multiple sclerosis [72, 73]. Boidin et al. reported comparable results in weight loss and reduction in fasting glycaemia and triglyceride levels in obese people [71]. In obese women, an eight week aquatic cycling programme in cold water did not lead to weight loss [66].

In young, healthy males, there was no superior effect of cold or warm water on the improvements in cardiovascular parameters [67, 69, 70], lactate accumulation lactate accumulation [69], dryland heat tolerance [67] and muscle glycogen utilization [69]

Land-based cycling versus aquatic cycling

The main body of the current research on aquatic cycling focuses on cardiovascular outcomes and the core findings for the comparison between land-based and water-based cycling showed similar trends. These latter studies [17–23, 26] reported comparable VO_2max values of aquatic and land-based cycling and therefore, the cardiac demand of aquatic cycling seems similar to land-based cycling. The results for HR were less consistent with a tendency for a lower HR during aquatic cycling compared to land-based cycling [20–23, 26, 29, 30]. Further, cardiac output and stroke volume was reported to be higher during aquatic cycling [15, 21, 29, 30]. These results are in line with the general understanding concerning the effects of water immersion on the human body. Hydrostatic pressure exerts external pressure on the immersed body, which increases with increased depth [2, 74]. Due to the hydrostatic pressure exerted there is a shift of blood from the extremities to the chest cavity, increasing arterial filling, and thus cardiac output and stroke volume are increased [2, 74]. Because cardiovascular parameters are modified by immersion, this could explain why the literature is inconclusive on the optimal recommendations for exercise prescription during aquatic cycling. Another explanation maybe as most aquatic bikes are not equipped with an electronically controlled pedalling resistance mechanism and approaches to estimate VO₂ from aquatic cycling are often based on pedalling frequency, with or without additional resistance. However, these equations cannot be used for all aquatic bikes, as the design and drag resistance created by pedals and resistance fins vary considerably across the aquatic bikes.

Aquatic cycling under different conditions

Due to the heterogeneous nature of aquatic cycling, many variables are involved when studying its impact on individuals, for example device-specific factors [44, 63–65] or environmental parameters as water temperature [12, 23, 43, 48, 49, 51, 52, 56–61]. Thus explaining why the cardiovascular response to different exercise conditions was frequently investigated. For example, it seems that VO_2max is comparable across different water temperatures and that participants perceived exercising in warm water as more exhaustive [23, 48, 49]. Further, included studies concluded that exercise intensities up to maximal limits are achieved by an increase in pedalling frequency and that VO_2peak does not differ between the different types of aquatic bikes [44, 64]. However, high-pedal frequencies are difficult to maintain during longer exercise sessions with a continuous character [44, 64]. To avoid discomfort with maintaining high pedal frequencies, exercise intensity can be modified by an increase in pedalling resistance or by utilising an interval training [45]. The latter was perceived less exhaustive than a continuous protocol [45].

Aquatic cycling as an intervention

Only six studies investigated the effect of multiple aquatic cycling sessions [66–73]. In four studies aquatic cycling was used in a clinical context for patients with multiple sclerosis and as exercise training for older adults and obese individuals. Research showed that aquatic cycling was equally effective than land-based cycling for improving cardiovascular fitness [66, 68, 71–73]. Furthermore, none of the included studies reported adverse events related to the training, suggesting that aquatic cycling is a safe exercise modality.

Most of the exercise protocols of the aquatic cycling intervention programmes consisted of steady cycling in a seated position with moderate intensity. Only Boidin et al. used an interval protocol for the training of obese individuals [71]. It seems that the full potential of aquatic cycling including out-of-saddle positions and arm and trunk exercises is not published yet in peer-reviewed journals [7]. Addition of these elements might prevent monotony especially in multiple session programmes [75] and results from supportive literature suggest that a full spectrum aquatic cycling programme is effective in patients with musculoskeletal disorders [76].

This scoping review has identified a number of areas for further research. Most of the included studies have a cross-over design with few cycling sessions and investigated the exercise response in young healthy males, because gender, body mass and morphology are known to affect the response to aquatic cycling [59, 77, 78]. Further, only six studies investigated the effect of an aquatic cycling intervention programme. To improve the use of aquatic cycling in healthcare, future studies, preferably RCTs, should investigate the effects of aquatic cycling interventions in different populations and on outcomes such as (joint) pain, muscle strength or physical functioning, which are yet to be investigated. Of specific interest may also be the biomechanics of aquatic cycling and differences of seated and out-of saddle cycling. Furthermore, the identified literature seems suitable for more systematic reviews. For example it seems worthy to synthesize the available evidence on cardiovascular responses to aquatic cycling.

To further improve the understanding of acute and long-term physiological adaptions to aquatic cycling training and facilitate between study comparisons, consistent reporting of the following parameters is recommended. Studies should describe the type of aquatic bike, body position, level of immersion, water temperature, methods used to control and assess exercise intensity i.e. training frequency, duration, rpm and pedalling resistance. Furthermore, it should be stated whether or not adverse events occurred. In addition to an accurate description of the aquatic cycling intervention, an agreement of experts on uniform keywords to describe the exercise activity is also strongly advised since this would improve the search in scientific databases. In this review the terms “aquatic cycling” and “aquatic bike” were used, as these expressions nowadays are commonly associated with this type of exercise.

This review has strengths and weaknesses. The extensive search procedure in this review resulted in more than sixty publications on aquatic cycling only, which were summarized and displayed. However, the presented studies should be interpreted with caution, because no quality assessment of the internal validity of the included studies was made in order to cover a broad spectrum of literature. Furthermore, this review provides a very general overview of the research on aquatic cycling without focusing on certain details of the included studies. For example, only the main outcomes reported in the abstract of the included studies were reported in this review. Yet, this comprehensive outline of available literature in this scoping review could serve as a starting point for systematic reviews or clinical studies on the effects of aquatic cycling on the cardiovascular responses.