There is a lack of meta-analysis and systematic reviews to describe the effects of water-based exercise on physical fitness parameters in elderly subjects. In fact, the existing high quality reviews are mainly focused on knee osteoarthritis , fibromyalgia  and the relief from back pain . Additionally, the evidence of an effective improvement of physical fitness parameters with a water-based exercise program in healthy older adults is lacking.
When considering the studies analyzing the benefits of water exercises programs in terms of aerobic capacity, strength, flexibility and body composition improvements, several limitations were identified. Firstly, examining cohort characteristics of the included studies, seven of them involved a sample of female subjects, while only two manuscripts [16, 17] presented mixed groups. This means that in the available literature, the overall number of healthy older male subjects studied was 12. Although some evidences revealed no gender differences in the magnitude of improvement in aerobic capacity with endurance training [24, 25], few older studies suggested that women were less trainable than men [26, 27]. Conversely, Peterson et al.  reported that the regression analysis performed in their systematic review failed to identify an association between gender and strength main effects, suggesting a significant potential adaptive response for both men and women. Further, Hakkinen et al.  indicated women as hypertrophycally responsive to a twice-a-week strength training, while the same program showed limited effects on muscle hypertrophy in men. Thus, similar to water-based strength exercise, whether muscular adaptations to different land-based activities between genders are comparable, is still unclear. Moreover, the normal decline in physical function occurring with aging could alter the analysis of gender effect on training of flexibility and body composition.
From a methodological point of view, a further limitation was the paucity of studies evaluating or reporting participants' level of physical activity at the start of the exercise protocol [13, 16, 18, 19]. From this perspective, elderly with higher or lower than normal functional capacity could have shown different outcomes at the end of the exercise interventions, especially in studies using higher volumes and intensities.
Six studies measured aerobic capacity (Table 3). All exercise protocols were similar for duration (8–12 weeks) and frequency (three times a week) but differed for modality of the aerobic exercise. Protocols included aerobic exercise , combined aerobic and resistance exercise [12, 13, 18, 19] and a sequence of specific water-based exercise . On average, the intensity was set between 60 and 85% of age-predicted maximal HR. Despite these small differences, the mean improvement in aerobic capacity (VO2 max) ranged between 10 and 15%. Only one study, a RUT , paradoxically the one with the longest duration of training (9 months), was unable to detect any significant change. However, this study was mainly focused on strength enhancement, while aerobic exercise was poorly developed, and its intensity was not quantified. On the other hand, the greater increase in VO2 max (+42%), observed by Bocalini et al. , can probably be explained by their use of an indirect approach for maximal aerobic power estimation (ACSM's equation) .
On the whole, it would seem that these protocols, although slightly different for intensity, produced similar improvements in terms of VO2 max. This suggests that water-based exercise, performed from moderate to high intensities, should also be considered a useful tool for the improvement of cardiovascular capacity in healthy elderly.
Most part of the studies evaluated strength using different testing modalities and different methods of strength training intensity monitoring; this variability yield a further limitation for data comparison and interpretation. Katsura et al.  described the intensity of strength exercises using Borg's rating of perceived exertion (RPE); Takeshima et al.  did not describe the intensity of the aquatic exercise although the authors imposed the full range of motion at the maximal velocity attainable in each set and repetition. Bocalini et al.  followed the exercise intervention methods of Takeshima et al.  with subtle adaptations, specific for older women. Finally, Taunton et al.  included strength and endurance (muscle endurance) exercises in their protocol without a suitable quantification of exercise intensity. A standardized method to evaluate intensity when considering water-based resistance exercises was proposed by Colado et al. : the authors proposed the use of the exercise rhythm and the RPE as a valid method for reproducing the intensity of effort among different sets of the same aquatic resistance exercise. Future researches should use a common system to evaluate and monitor resistance training intensity, possibly adopting the method recommended by Colado , which also appears more suitable for fitness leaders of water-based exercise programs.
Although our analysis of literature shows that water-based exercise is a valid tool for the development of strength and muscle endurance, the intensity and the progression of the stimulus required are not well defined. In fact, the wide variability in strength tests and muscular groups evaluated precludes an objective analysis of the magnitude of the adequate stimulus. Future studies should focus on exercise intensity and its progression in resistance training protocols, not disregarding the practical applicability for exercise professionals.
A lack of evidences about the effect of flexibility training on a range of motion (D category of evidence) is highlighted by the ACSM/AHA physical activity guidelines for older adults. In this systematic review, out of nine papers, six specifically investigated flexibility, reporting contradicting results. If we separately analyze each paper, we observe that Taunton et al.  and Takeshima at al.  did not observe any change in flexibility after their water-based exercise protocols. Conversely, Bocalini et al. , Cancela Carral et al. , Tsorlou et al.  and Katsura et al.  described remarkable improvements on lower body flexibility (up to +50%). Despite none of these manuscripts specifically described the type and the intensity of flexibility exercise performed, five studies [12, 15, 17–19] reported a time of 7 to 20 min (time per exercise session) as being dedicated to the flexibility/stretching activities. Only one study provided data about the upper body flexibility , showing a 40% improvement, while papers reporting positive results on lower body flexibility described a mean improvement ranging between 7 and 19%. In light of this analysis, we emphasize the need for future research through well-defined training and assessment protocols in water exercise activities.
Currently, applying the ACSM/AHA recommendations, in water-based activity programs, we should include at least a twice-a-week flexibility exercise at moderate intensity (5–6 on the Borg's Scale from 0 to 10). Further, as pointed by Barbosa et al. , these exercises should be adapted to the aquatic condition. In fact, during flexibility exercises, body temperature does not reach the same levels in respect to aerobic activities. Then, these exercises should be carried out in higher water temperature or alternated with other exercises in order to reduce the loss of body heat.
Only four studies investigated body composition. Tsourlou et al. , using bioimpedance analysis, found a 3.4% significant increase of fat free mass. This improvement was concomitant with an enhancement on both upper and lower limb strength. In addition, Takeshima  found a statistically significant reduction of 7.9% in the sum of skinfold. Although the method of skinfold is not exempt from errors of measure, this decrease in subcutaneous fat indicates a clinically relevant improvement in body composition. The other two papers [16, 19] did not find any significant change.
We believe that these different outcomes could have been affected by the rather short duration of the exercise protocols, the lack of control of diet and the different methods used to evaluate body composition. Further, differently from the other studies, in the two “effective” protocols [14, 15], none of subjects had ever participated in a weight training program.
Among all variables of physical fitness, the modification of body composition was the least studied. Future high qualities studies should consider the inclusion of more advanced body composition analysis techniques (e.g., dual energy x-ray absorptiometry), which are able to evaluate whole and segmental body composition as well as a standardization of other important variables such as dietary intake and previous weight training activities. Considering the relevance of sarcopenia in the aging process, these methods could add clinically important information.