*2.7. Studies' Risk of Bias Assessment*

Risk of bias in individual studies was judged using Cochrane's Risk of Bias Assessment for Non-randomized studies (RoBANS; [48]), evaluating six domains: (i) the participant selection; (ii) confounding variables; (iii) the exposure measurement; (iv) the outcome assessments blinding; (v) incomplete outcome data; and (vi) selective outcome reporting.

#### *2.8. Effect Measures*

IVV mean ± SD or median ± IQR values were calculated, and, when needed, two authors (AF and BM) independently extracted data from graphs using the WebPlotDigitizer v4.5 (Pacifica, CA, USA) [49].

#### *2.9. Synthesis Methods*

A narrative synthesis of the main findings was performed and supplemented with an interactive evidence gap map (generated by EPPI-Mapper v.2.2.3, London, UK, powered by EPPI Reviewer and created by the Digital Solution Foundry team). This map can be accessed online, providing interactive ways to visualize the current review's included studies (including authors, abstracts and keywords) and the primary and secondary outcomes.

#### **3. Results**

The initial search identified 227 potentially relevant articles, with 126 being duplicates, which were consequently removed (Figure 1). Following the titles and abstract screening, 17 and 10 studies were excluded by eligibility criteria and article type (respectively). After the seventy-four full texts were screened, one was excluded by type [50], six by exposure [51–56], seven by outcomes [57–63] and one by participant [64] eligibility criteria. Reference list analysis revealed 31 studies on the topic as potentially meeting the inclusion criteria, with full-text analysis excluding 10 articles by type [65–74], 2 by exposure [75,76] and 8 by outcomes [18,77–83]. Seven additional studies from snowballing citation tracking process were deemed eligible for inclusion, and all were included [29,84–89]. Expert consultations did not yield any new studies, so the combined total sample was n = 76 corresponding to 68 trials. Studies from the same trial were grouped for the analysis [4,22,25–27,29–31,33,39,90–93].

**Figure 1.** Search and screening processes used in the current study displayed as a PRISMA 2020 flow diagram.

#### *3.1. Studies Risk of Bias Assessment*

Sixty-eight trials were considered for judging risk of bias, with 20 [6,20,28,36,37,40, 41,87–89,94–103] and 48 considered as having overall low and high risk (respectively). The selection of participants showed a low risk of bias for 79% of the trials due to the overall purpose of evaluating competitive swimmers (Figure 2). However, 19% of the trials presented high risk due to the unbalanced number of females versus males [7,104,105], heterogeneity of participants [86,92,106,107], lack of information [10,92,108], or the noncompetitive or inexperienced participation in the trials [84,109–111]. Two studies [26,27] were judged unclear because of the uncertainty of how swimmers were analysed. Fifty-one percent of the trials had a high risk of bias in the domain of confounding variables due to participant-related problems (lack of information [10,23,26,27,108–110,112], swimmers with different characteristics mixed in the same group [15,17,24,85,86,92,104,107,111,113–115], swimmers experience [3,84,116,117] and specialty [118]) and protocol-related problems

(snorkel use [4,21,25,30,34,35,42,90,91,105,119,120], possible fatigue effect [32,121] and different evaluation conditions [122]).

**Figure 2.** Percentage for each risk-of-bias domain regarding the included trials.

Considering that no data were provided concerning the validity and reliability of the software used or whether the process was fully automated in the different studies analysed, exposure measurement was judged unclear for 63% of the trials. High risk was evaluated for 6% of the trials with specific measurement issues; in particular, (i) the electrical resistance variation method had not been previously validated, with authors not providing proof of its reliability [104]; (ii) the preparation procedures and the evaluation protocol were performed for different swimming techniques [123]; (iii) various devices were used for different swimmers, and the evaluation frequency varied substantially in a retrospective study [124], raising questions concerning the actual measurement exposure consistency; and (iv) evaluations did not respect the same time period from the main competitions [125].

Many trials (74%) did not mention outcome assessment blinding and it was unclear if video analysis was fully automated (probably interfering with the measurements). High risk was attributed to 7% of the trials due to no blinding and to the inexistence of data concerning the reliability of the automated process [4,16,25,30,31,90,91,93,116,124]. Due to an absence of information on whether the selected swimmers were part of a larger sample, incomplete outcome data were judged unclear for 88% of the trials, except for a case study [125] and a trial that included an a priori sample-power analysis [109]. High risk was evaluated for 9% of the trials due to missing data, given that this could influence the study outcomes [4,19,25,29,30,39,90,91,111,126,127]. Eighty-eight percent of the trials had no preregistered protocol to compare to, with the selective outcome reporting unclear. High risk was judged for the trials belonging to the same study [4,22,25–27,29–31,33,39,90,91,93] and for those that did not fully report the pre-defined primary outcomes [19,104].

#### *3.2. Studies Characteristics*

The included trials' main characteristics are presented in Table 2. Across the 68 trials, 1440 swimmers were evaluated for IVV (55.2% male and 10.7% missing information), with n = 1–126 sample sizes and 11.7 ± 0.8–42.5 ± 9.5 years of age. Some trials did not present information regarding IVV [16,19,111,120], female swimmers' participation [10, 23,26,27,92,102,112], competitive level [104,110], age category [7,10,21,104], or protocol intensity [111]. Thirty-nine trials assessed IVV as the main study purpose, of which three analysed and described the swimming cycles curves [7,104,118]; nine related IVV with coordination [22,23,26,27,30–33,103], six with swimming economy [21,30,34,35,90,105,123], six with fatigue [26,27,84,107,108,112], six with technique [4,36,41,107,111,113] and five with velocity [3,6,37,41,124]; three analysed different swimming techniques variants [35,106,126]; two related to force [94,99]; six were methodological [10,17,19,20,86,87]; one was a dynamical systems approach [40]; and one was a training intervention [88].


*Bioengineering*

 **2023**, *10*, 308





119

*Bioengineering*

 **2023**, *10*, 308



IVV was not the primary outcome in 31 trials but was included in a larger analysis, being described [127] and analysed together with anthropometric, kinematic, energetic, coordinative neuromuscular activity and other biomechanical variables [25,30,39,91,97,110, 117,119]. Trials also related IVV with coordination [28,29,95], swimming economy [42,92], fatigue [28,29,95,121,128], technique [24,93,120] and velocity [125]. Thus, IVV was included in methodological approaches [15,16,89,96,100], dynamical systems approaches [85,102, 114,122] and training interventional trials [29,98,101,109,115]. No conflicting interests were declared or were not addressed by 34 and 66% of the trials. Funding information was not reported by 54% of the trials, while 40% had financial support. Trials dissemination was growing over time (records were published every year) and 2016 was the year with the most publications (seven records).

#### *3.3. Evidence Synthesis*

The evaluation of the evidence gap map and trials' risk of bias can be accessed through the Supplementary File S1. IVV was assessed in 46 front crawl, 10 backstroke, 24 breaststroke and 14 butterfly-related trials, most of them focusing on mixed and male-only groups regarding swimmers' sex (56 and 39%, respectively). High-level swimmers were the most studied, followed by elite and trained, recreational, world-class and sedentary swimmers (37, 25, 26, 9, 3 and 1%, respectively), from which senior, youth and junior, young and master swimmers participated (50, 18, 18, 10 and 5%, respectively). Regarding the protocol intensity, most trials focused on swimming at sprint and severe intensities (36 and 19%), and fewer implemented incremental protocols that include other intensities (extreme, heavy, moderate and low: 11, 11, 12 and 11%, respectively). Trials conducted in swimming pool conditions were used 99% of the time.

Image- (47 and 53% in two and three dimensions) and mechanical-based methods were used (56 and 41%, respectively), with speedometers being mostly selected (82%). Velocity was calculated using an anatomical fixed point as a reference, most of the time with the hip chosen (and only twice selecting the head/neck) rather than the centre of mass (71 and 29%, respectively). The coefficient of variation was preferred regarding IVV quantification versus the difference between the maximum and minimum instantaneous velocity (dv; 61 and 7%, respectively), the ratio maximum and minimum instantaneous velocity difference/intracycle mean velocity (dv/v; 7%), the ratio of the mean velocity/difference between the maximal and minimum instantaneous velocity (3%) and other methods (such as cycle characterization, curves acceleration and dynamic indexes; 23%). Twenty-five trials reported variables associated with swimming economy (such stroke length and stroke index) and only two reported hydrodynamic drag related variables.

Front-crawl-related trials almost covered all secondary outcomes, even though gaps were identified for the four conventional swimming techniques. No trials were conducted with world-class swimmers focused on extreme, heavy, moderate and low intensities; used accelerometers; or quantified IVV with overall methods. Young swimmers were not used as samples in trials that were conducted at extreme and low swimming intensities, accelerometers were employed, and, when characterizing these age group IVV, its quantification was performed using only three methods. Master swimmers were not called to participate in protocols with extreme intensity and were not evaluated using accelerometers, while IVV quantification in this population was conducted only through the coefficient of variation. Trials using youth/junior, world-class, elite, highly trained and trained swimmers did not have associated IVV and hydrodynamic drag.

#### *3.4. Study Results*

Higher-level swimmers presented superior mean velocities for the same swimming intensity, but IVV was not related to swimming competitive levels or to the mean velocities regarding the four swimming techniques (Table 3). Except for front crawl, studies were mostly interested in analysing IVV when swimmers were performing at maximal intensity. IVV was not related to mean velocity in front crawl or backstroke [37,40,41,100], even if

a non-linear relationship was also observed (with the velocity increase leading to a IVV decrease in young swimmers in the four swimming techniques [3] and in the swimmers with high-level front crawl [4]). Data gathered from so many swimmers and diverse samples should be cautiously analysed. Some outputs were obtained from a single trial performed at a specific swimming intensity, while others were gathered by averaging the data available. In addition, in some studies, swimmers from different competitive levels were pooled, and data were presented as a single group.

**Table 3.** Mean ± SD or median ± IQR mean velocity and IVV values obtained in the swimming trials included in the current study.


Legend: IVV quantified by dv/v is presented in the breaststroke row.

In breaststroke, IVV is usually quantified by dv/v (m/s), as presented in Equation (1), with vmax,LL as the maximum centre of mass's velocity achieved at the end of lower limb propulsion; vmin,LL as the first minimum peak of the centre of mass's velocity following upper and lower limbs recovery (corresponding to the beginning of lower limb propulsion); vmax,UL as the maximum centre of mass's velocity at the end of the upper limb propulsion; and vmin,T as the minimum centre of mass's velocity during the transition between upper and lower limb propulsion (corresponding to the centre of mass's velocity during gliding).

$$\text{IVV} = \frac{\text{vmax}, \text{LL} - \text{vmin}, \text{LL} + \text{vmax}, \text{UL} - \text{vmax}, \text{T}}{\text{vmean}} \tag{1}$$

Some trials showed periodic velocity fluctuations related to the upper limbs' actions and the rate and the number of peaks per cycle, with a higher IVV range in lower- than in higher-level swimmers [7,104,118,127]. Furthermore, successful swimmers were able

to more effectively combine intracycle peak velocity with relatively longer cycle periods [6]. When a front crawl technical training intervention period was conducted, IVV decreased [29,88,109] or did not change [98,115]. Although propulsive and drag forces were higher in swimmers of superior level, larger index of coordination values for front crawl were also presented even if IVV did not change across intensities [10,21,23,33,34,38,39], suggesting that better propulsive continuity allows a stable IVV [22,24–33]. Conversely, IVV increased throughout paces in less skilled swimmers [23]. IVV for highly trained swimmers was lower than for trained counterparts at all front crawl swimming velocities (in both senior and youth age groups) [23,39] but in backstroke, IVV did not differ between elite and highly trained swimmers [40].

IVV was directly related to swimming economy in the four swimming techniques [21, 34,35,105,123,126] even though, in one study, no association between these variables was reported [90]. However, front crawl and backstroke IVV did not differ; nonetheless, lower energy cost values for front crawl vs. backstroke were observed [42], and they showed a tendency to decrease in a maximal lactate steady-state test [92]. Similarly, swimmers maintained their IVV values when performing at submaximal intensity, but IVV rose at maximal intensity [84,107,108,112,123,128], even though others described no changes [26, 27,121]. This IVV increase with effort is probably justified by the progressive increase in fatigue, resulting in swimmers becoming less mechanically efficient. Swimmers with higher intracycle force variation also presented higher IVV values, leading to a progressive decrease in performance [94,99].

Methodological trials mainly assessed the relationship between the hip and the centreof-mass kinematics to provide simpler methods to quantify IVV in swimming. It seems consensual that the hip does not adequately represent the centre of mass in intracycle variation in butterfly, breaststroke and front crawl. Some authors clearly state that this anatomic point should not be used in this kind of assessment [16,19,20] because it greatly overestimates the swimmer's real variation in velocity [15,17]. Other trials aimed to validate methods to quantify and express IVV [10,86,87,89,96]. When applying dynamical system approaches to swimming, nonlinear properties can be observed [114], with their magnitude differing according to the swimming technique and the swimmer's level. The breaststroke and butterfly techniques displayed more complex (but predictable) patterns [85,114,122] and elite vs. non-elite swimmers' performances were more unstable and complex (even though their IVV did not differ) [40].

#### **4. Discussion**

The current systematic scoping review focused on the IVV assessment in swimming that is retrospectively available for almost a century. The IVV-related trials' main interest is in the interactions between the cyclical propulsive and drag forces, which help understand the cyclic effectiveness of the upper and lower limbs while swimming and, consequently, swimmers' technical efficiency. In the first studies on IVV, breaststroke was the most studied swimming technique due to the simultaneity between the movements of the upper and lower limbs (which allowed researchers to easily identify when these movements were occurring) [6,118]. Then, new methodologies were developed, with researchers focusing their attention on the four conventional techniques, but our results showed that front crawl aroused greater interest. It is now accepted that the techniques with simultaneous movements (butterfly and breaststroke) present higher IVV than those with alternated movements (front crawl and backstroke) due to the mechanical impulses applied to the swimmer's body [3,114,122]. Furthermore, the alternated techniques' IVVs are very similar due to the biomechanical similarities between the front and back crawl (an "old" term used to designate backstroke) [42].

From the analysed trials, we could observe that male swimmers were the most studied even though mixed groups were also used due to the interest in checking differences between female and male swimmers (particularly regarding anthropometric characteristics [39,117], mechanical power output [6,33], technical proficiency and hydrodynamic

profile [33,126]). Researchers focused their attention on trained, highly trained and elite swimmers, with the most elevated competitive levels being preferred for analysis. Most trials focused on senior swimmers, displaying strong confidence in results due to their experience. The same was not observed for trials conducted in master swimmers, with considerable gaps found, probably due to their heterogeneity of age and competitive level. Swimmers were mainly evaluated using maximal-intensity protocols to assess the kinematics directly related to the competitive events with the most participation (the 50 and 100 m distances). The 200 m distance was also often investigated, since its metabolic characteristics are important determinants of the kinematic variables' behaviour during these mixed aerobic–anaerobic events [4,37]. Few studies have focused on the backstroke, breaststroke and butterfly techniques at heavy, severe and extreme intensities.

The included trials used distinct evaluation protocols, with some analysing nonbreathing cycles [15,24,26,27,31,32,85,93,96,104,105,119–121,127] and other not reporting the breathing condition or the inclusion of a specific space in which the participants were not allowed to breathe [3,6,10,16,17,19,21–25,30,31,33,36,37,42,85,87,88,90,91,93,95,104,107, 108,111,112,115,116,122,124,126,127]. Even though breathing was shown to lead to coordination asymmetry [129], upper-limb-cycle kinematics with individual breathing patterns presented IVV similarities to those in apnoea [41]. Data from trials that used a snorkel for assessing oxygen consumption should be carefully analysed [4,25,30,34,35,42,86,89– 91,102,105,107,108,119,123]. Concerning the use of the hip vs. the centre of mass for assessing IVV, it was clear that the latter was the most reliable method to measure kinematical variables, although some authors still consider hip movements to provide a good IVV estimate [3,15]. These methods were previously compared with the hypothesis that the hip represented the centre of mass (and not the opposite), which was considered a priori the best methodology [15–17,19,20]. Future studies should clarify why the centre of mass is the gold standard considering the complexity of evaluation.

As a consequence of specific front crawl intervention protocols, IVV decreased or remained stable due to better swimming technique [6,102,106,111,113]. This also might have happened in other swimming techniques, with butterfly IVV decreasing when the hands' velocity at the end of the underwater path and the vertical velocity during the lower limbs' actions increased, and the velocity during the hands' entry decreased [111,113]. The hands, trunk and lower limbs role are also fundamental for lowering IVV [4,6,93,126]. Even though it is widely accepted that lower IVV should be achieved for enhanced performance, IVV has no standardized values and is highly variable according to the studied population and the methods used. Therefore, it would be very useful to implement more frequent intervention programs with strategies to upgrade swimmers' technique and overall performance.

Researchers have started to characterize swimming cycles' shape and number of peaks, developing quantification methods such as the absolute average velocity, root mean square [10], coefficient of variation and range of maximum and minimum velocities in a cycle [130]. Unfortunately, only one work compared these measurements [131], concluding that the coefficient of variation was the only approach sensitive to the mean swimming velocity and to the instantaneous velocity dispersion during the cycle. Mathematically, it is the more accurate method for IVV quantification but it may overestimate its value in breaststroke (due to this technique's complexity regarding mechanical impulses and coordination). Nevertheless, even this measure does not reflect the hydrodynamic drag characteristics, and it may be helpful to develop a new method of IVV determination.

Swimmers at a higher level present higher IVV values due to their capacity to generate and sustain the highest velocities (rather than being more economical), displaying larger amplitude of velocity [36,124]. However, breaststrokers eliminated in the preliminaries of a World Swimming Championships displayed higher IVV values than those that qualified for the semi-finals [127], probably as a result of a very low minimal instantaneous velocity (and not necessarily related to the maximal velocity value achieved within a cycle). In short distances, depending on the swimming technique, better swimmers find solutions to improve technical proficiency, producing high mechanical power to generate

superior propulsive forces, reducing hydrodynamic drag, and adopting greater propulsive continuity [33,34,38,41], which will cause different IVV.

The quality of the trials included in the current study can be questioned due to the lack of detailed information and uncertainty of the evidence provided (being indeterminate whether it would result in a high or low risk of bias). Disregarding the already mentioned factors that influenced a high risk of bias, most variables were unclear because it the validity and reliability of the exposure measurement were not mentioned, nor were the blinding of the outcome assessment or even the information about whether swimmers belonged to a larger sample. In the scope of swimming, experimental protocols aim to replicate swimmers' performance and are not usually registered in databases. Furthermore, the current scoping review included trials since 1971 that were not as concerned about the studies' quality as is dictated today.

#### **5. Study Limitations**

The number of included trials highlighted the importance and utility of performing a systematic scoping review in swimming IVV. We believe that including the Proceedings Books of the Biomechanics and Medicine in Swimming Symposia strengthened our work, since this book series contains several important documents that added relevant information to the current review. This research aimed to provide an overall representation of the IVV scope of competitive swimming, but we recognize that considering IVV calculations in conditions such as using snorkelling or swimming with/without breathing could affect its interpretation. For sake of the clarity, those studies were properly identified.

#### **6. Conclusions**

The current study compiles the studies available on the topic of the swimming IVV in the most respected and well-known literature databases. We have described the literature gaps and the most interesting IVV-related topics within almost the past century. IVV was often used in front-crawl-related studies, involving mixed samples and senior swimmers that performed at sprint intensity in swimming pools and were evaluated with cable speedometer using an anatomical fixed point as a reference and that quantified IVV using the coefficient of variation. There is a clear need for investigating backstroke, breaststroke and butterfly swimming techniques performed at heavy, severe and extreme intensities. Since these paces correspond to the characteristics of the official competitive events, it would be imperative to assess them more often. Young and youth swimmers were less studied, even though their performance development in swimming is important in their training process throughout their careers. It would be very helpful to evaluate world-class swimmers as well to acknowledge the top-level performers' behaviour. Although there is no proof that the coefficient of variation is the best measure to assess IVV, researchers generally agreed that it best reflects the velocity fluctuations in swimming.

#### **7. Future Directions**

Future investigations should cover the gaps found in the current study to allow for meaningful results and possible comparisons. IVV measurements should be revised, and a new approach that accounts for hydrodynamic characteristics is welcome to standardize results according to these factors. Future research should strive to reduce the risk of bias by (i) attending to a balance between female and male swimmers, looking for better sample homogeneity; (ii) providing important personal characteristics; (iii) controlling the evaluation conditions; (iv) providing the software validity and reliability; (v) blinding the outcome evaluators; (vi) providing data on the inter-evaluator reliability of outcome measurement or measures of error for the methodologies used (when applicable); (vii) providing information about whether swimmers are part of larger samples; and (viii) pre-registering the research protocols.

**Supplementary Materials:** The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/bioengineering10030308/s1, File S1: Evidence Gap Map.

**Author Contributions:** Conceptualization, A.F., J.A., F.N., B.M., J.P.V.-B. and R.J.F.; Methodology, A.F., J.A. and B.M.; Formal analysis and investigation A.F., J.A., F.N., B.M. and R.J.F.; Resources, J.P.V.-B. and R.J.F.; Writing—original draft preparation, A.F.; Writing—review and editing, A.F., J.A., F.N., B.M., J.P.V.-B. and R.J.F.; Supervision B.M., J.P.V.-B. and R.J.F.; Project administration, A.F., J.P.V.-B. and R.J.F.; Funding acquisition, J.P.V.-B. and R.J.F. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by Fundação para a Ciência e Tecnologia, grant number 2020.06799.BD.

**Institutional Review Board Statement:** The study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of Faculty of Sport of University of Porto (CEFADE 24 2020, 11 November 2020).

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** We would like to acknowledge the support of Daniel Daly and Flávio Castro for their role as external experts. They verified our eligibility criteria and our list of included studies, suggesting additional potentially relevant studies. This acknowledgement was consented by them.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


**Disclaimer/Publisher's Note:** The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
