This study is a systematic review in which relevant clinical trials were identified, selected and evaluated, providing practical and theoretical support to the classification and analysis of the literature review. The methods used for the systematic review follow below.

In January 2013, computer searches for articles were performed in Scopus (www.scopus.com) and PubMed (www.pubmed.gov) databases, while master and doctoral theses were performed in LUME (on-line database of the Federal University of Rio Grande do Sul, Brazil, containing master and doctoral theses). Publication year was unlimited and the following key words were used: “physical activity”, “exercise”, “training”, “aquatic”, “water-based”, “hydrotherapy”, “hydrogymnastics”, “deep water running”, “deep water walking” and “head-out aquatic”, brought them together by means of the boolean operator “OR”. Besides this, references provided in the documents identified by these searches were screened with the purpose of finding additional studies.

Due to the higher availability and ease for searching and reading, only documents written in English, Portuguese and Spanish were used in this study.

Criteria for inclusion of studies in the present review were as follows:

• (1)

  Intervention studies (acute and chronic treatment).

• (2)

  Use of aquatic exercises and/or aquatic exercise training executed in an upright position (with the head out of the water).

• (3)

  Use of human subjects.

• (4)

  Assessment of lipids or glycaemic profile variables.

Regarding the exercise and training aims, protocols aiming at strength, aerobic or combined (strength and aerobic) improvements were included in the review, without restriction of training frequency and intensity, as well as duration of training sessions.

The PEDro (Physiotherapy Evidence Database) scale 18 was used for rating the documents included in the analysis. This scale was developed by the Centre for Evidence-Based Physiotherapy with the purpose of objectively evaluating the internal validity of studies. Each document was assessed through several criteria previously established, that overall, aim to evaluate four fundamental methodological aspects of a study such as the random process, the blinding technique, group comparison, and the data analysis process, being attributed a maximal value of 10 points. These 10 points were distributed among the seven criteria scale of possible control in intervention studies with physical training. Thus, studies were considered of excellent quality those who completed at least six criteria (8.57 points), not being associated with excellence the criterion on intention-to-treat analysis, which although very important clinically, it is still very unusual in studies aiming to analyse responses to physical training. Two raters evaluated the documents independently using the scale; in case of differences between the values attributed by them, the evaluation was discussed until a consensus was reached.

The average (± standard deviation) score of the analysed articles on the PEDro scale was 8.10 ± 0.83 points; only three studies were found and no study reached the highest score. The selected studies presented the specified eligibility criteria and the random distribution of treatments, but since they involved physical training, the scale criteria 3, 5 and 6, related to blinded interventions, were not considered, as it would not be possible to blind subjects and/or researchers in such interventions. By exclusion these three criteria of scale, the 10 points considered for assessment of methodological quality are the result of a redistribution of the total score of the seven criteria PEDro scale (2, 4, 7, 8, 9, 10 and 11), being assigned 1.42 for each criterion.

Table 1 shows the acute effects of aquatic training on lipids profile and glycaemia found in the three studies analysed, together with a summary of the development of the studies, including intervention duration, type, volume and intensity of training sessions, sample, main outcomes, percentage of improvement and score in the PEDro scale.

The three analysed studies 19–21 involved aerobic aquatic exercise, and the interventions were effective in improving lipids profile in two 20,21 of them. This effect was independent on the training status of the participants.

Although all training sessions of these studies were focused on aerobic exercise, they were executed through different manners (underwater cycle ergometer, 19 underwater treadmill 20 and water aerobics, 21 emphasizing the different types of exercise that can be used for aerobic training in the aquatic environment.

In spite of the benefits in lipids profile described by two these studies, 20,21 methodological differences between them, such as population and characteristics of exercise session impairs the formulation of definitive conclusions and practical applications regarding this subject.

The number of studies evaluating chronic outcomes (n = 9) was slightly higher than for the acute effects studies. Average (± standard deviation) score obtained by these studies in the PEDro scale was 8.58 ± 1.65 points, with three 22–24 studies obtaining the maximal score. Only one study 22 did not specify its eligibility criteria. Concerning random distribution of interventions, only one study 23 did not meet this criterion; nevertheless, this was not due to low methodological quality, but to the study design, since the authors analysed the effect of one aquatic exercise training protocol on two different populations. In this case, the score was redistributed within the other scale criteria. As in the studies about acute effects, scale criteria 3, 5 and 6 were not considered for the same reason.

Table 2 shows the chronic effects of aquatic exercise training on lipids profile and glycaemia in the nine studies analysed, as well as research development, intervention duration, type, volume and intensity of the training protocol, environment, sample, main results, percentage of improvement and score in the PEDro scale.

From the nine reviewed studies, three analysed aerobic exercise, 20–22 five analysed combined training (resistance and aerobic), 23–27 and only one analysed isolated resistance training. 28

Among the studies involving aerobic exercise, two presented interventions using water aerobics and one used walking and/or running on an underwater treadmill as exercise type. All studies demonstrated significant improvements in at least one of the lipids profile parameters, while only two studies measured glycaemia through glycated haemoglobin (HbA1c) 22,27 values, finding significant effects of exercise on this variable.

The majority of the studies investigating the chronic effects of aquatic exercise training used a combination of aerobic and strength training. 23–27 Of these studies, four 24-27 used water aerobics as exercise type, with three 24-26 of them showing positive results of exercise in lipids and lipoprotein profiles. Only one 27 of these studies evaluated HbA1c values, with authors reporting decreases in these values and, therefore, efficacy of the intervention in improving glycaemic control in DM-2 patients. Additionally, the study using deep-water running 23 as exercise type for healthy individuals and individuals presenting glucose intolerance also found diminished postprandial glycaemic values in the patients with glucose intolerance after the training intervention.

Only one study used strength training without combination with aerobic training. 28 In this study, no significant alterations in lipids profile and glycaemia were found in response to aquatic exercise training.

Two different designs were identified in studies regarding aerobic training, comparisons between exercise on dry land and in water; and comparisons between a training group and a control group. The populations examined were elderly men with history of heart ischaemic disease, dyslipidemic premenopausal women, overweight and obese subjects and patients diagnosed with DM-2; thus, responses to exercise and water immersion could be reviewed only for these populations. In the three 20-22 studies examining the effects of chronic exercise training, interventions had duration of 12 weeks; it was not possible to find studies with longer periods of intervention. While in one study 21 participants exercised twice a week during the intervention period, in the others studies individuals underwent three exercise sessions per week. In one study, 19 the duration of the exercise sessions was 15 min, considerably lower for aerobic exercise aiming at improvements in lipids metabolism, which is between 30 and 60 min. The minimal duration recommended (30 min) was reached by the other two studies 21,22 ; only one study 20 controlled exercise duration through energy expenditure instead of time in minutes. Three different manners of controlling intensity could be identified in the studies: rate of perceived exertion (RPE), percentage of the maximal heart rate, and velocity corresponding to a given percentage of maximal oxygen consumption (VO2max). Considering the three studies 20–22 with a 12-week intervention, two 20,21 presented intensity progression throughout the training period, one 20 with intensities varying between 60% and 85% of VO2max and the other 21 with RPE values varying between 11 and 15 points using the Borg Scale (minimum of 6 and maximum of 20 points).

No studies evaluating glycaemic measures were found. However, improvements in lipids profile were found both acutely and chronically, with exception of the study by Berminghan et al., 19 in which the lack of improvements most likely results from the intensities and short duration of the training sessions. Nuttamonwarakul et al. 22 controlled glycaemic values by measuring HbA1c and, despite the lack of intensity progression during the training period, considerable improvements in this parameter were found, exceeding the improvements found by previous studies involving on-land exercise in individuals with DM-2. 29,30

Future studies could examine the effects of longer interventions, higher training volumes and intensity control according to the anaerobic threshold or its markers, since this parameter is considered excellent for controlling intensity during aerobic exercise training. 31 Furthermore, comparing continuous and interval training, as well as analysing the mechanisms underlying the effects of immersion in water on lipids profile and glycaemic variables would be possible and highly relevant.

Among the studies analysed, only one 28 investigated the effects of aquatic strength training, comparing them to those of strength training executed with elastic bands on dry land. In this study, postmenopausal women underwent a 12-week intervention, with training frequency increasing from two to three sessions per week during the training period. Additionally, over the training period, volume increased from one set to two supersets and intensity increased from 5 to 7 points of the OMNI scale (RPE). No significant differences were found in fasting glycaemia and lipids profile in response to this training mode, which appears not to be the more adequate for populations in which these alterations are important aims. It is also important to consider that the magnitude of intensity progression was not very high; this could have been a reason for the lack of improvements in biochemical parameters found, since high-intensity strength training on dry land can enhance glycaemia and lipids profile in DM-2 patients. 32 Currently, research has been demonstrating increased strength in different populations when aquatic strength training is executed at maximal velocities. 33,34 The impact of this strategy on biochemical variables is, though, still unclear.

Definitive conclusions about the effects of this type of training on the parameters focused in this review are prevented by lack of studies with a higher level of evidence, such as randomized controlled trials. In addition, the only study involving aquatic strength training presented in its prescription moderate intensities, which are not characteristic of training programmes aiming at strength gains.

Studies analysing responses of lipids profile and glycaemia to combined training can be divided into three groups according to their methods: (a) comparisons between exercise in water and on dry land (one study 26 ); (b) comparisons between training and control groups (two studies 24,27 ); and (c) comparison of pre- and post-training values of one group (two studies 23,25 ). The populations examined were healthy premenopausal and elderly women, middle-aged overweight women (both glucose tolerant and intolerant), middle-aged men with coronary artery disease, and men and women with DM-2 and/or cardiac insufficiency. Intervention duration of the studies ranged between eight and 20 weeks, with an average training frequency of three sessions per week.

The study 26 with the highest training frequency – four sessions per week – presented the only protocol in which aerobic and strength-training sessions were separately executed in different days. This can be an interesting strategy when planning exercise training programmes, since it leads participants to engage in physical activity four days per week and diminishes the monotony of the sessions.

On average, total session duration of the analysed studies was approximately 60 min. In almost all studies, aerobic training sessions had a duration of about 30–40 min, with no changes along the training period. Control of session duration through energy expenditure was not used in any of these studies. In turn, strength-training volume varied between one and six sets per exercise and duration of sets was controlled through different manners, such as maximal repetition number or percentages of it, execution time or fixed number of repetitions at maximal velocity. As a consequence of the considerable differences in training volume between the studies, it remains difficult to explore and associate the results with training prescription.

With respect to aerobic training intensity, three 23,25,26 studies controlled this parameter during aerobic training through percentages of maximal heart rate, which ranged between 50% and 85%. Other manners of prescribing exercise intensity found were the use of heart rate reserve measures and the identification of heart rate values at the anaerobic threshold. None of the analysed studies presented intensity progression throughout the training period, which is a possible limiting factor for several physiological adjustments.

Regarding intensity prescription of strength training, exercise execution at maximal velocity was used in two studies. Another form of controlling this variable when exercising at submaximal intensities is the use of percentages of the maximal number of repetitions; this method, however, presents the disadvantage of requiring the execution of repetition maximum series in the water – in these cases, fatigue can lead to a diminution in velocity of execution, which in turn decreases the water resistance to the movement, altering the exercise load. Thus, it is possible to control intensity by determining a given number of repetitions to be executed within a determined time; however, it has to be considered that this measure is not related to repetition maximum values. The other methods used for controlling intensity were not sufficiently described in the studies, preventing their discussion. Only one study 25 explored an interesting intensity progression throughout the strength-training period, starting with two sets of 30 s, progressing until six sets of 10 s at maximal velocity.

Only two studies 23,27 examined glycaemic variables and four studies 23–27 assessed lipids profile. In all studies, combining aerobic and strength training elicited significant improvements in at least one of these variables when compared to the control group or pre-training values. The only study comparing aquatic exercise training to training on dry land demonstrated a lack of differences between these two exercise types. This emphasizes the advantages of aquatic exercise training, since this type of training could elicit the same biochemical improvements resulting from training on dry land, with the additional benefit of increased buoyancy and diminished load imposed to the musculoskeletal system. 3

Suggestions for further studies include the comparison between biochemical responses elicited by exercise training in water and on dry land. Another important point would be to prescribe exercise based on maximal and/or submaximal tests executed in the training environment. Prescribing aquatic exercise training based on results reached during tests on dry land can lead to considerable errors and imprecisions, since the physiological peculiarities of water immersion, discussed in this review, are neglected in such situations.

The reduced number of database searched is a limitation in our literature search. However, the broad reach of the searched databases (PubMed and Scopus), we believe that our findings reflect the current literature on the topic of study.