Intragroup Significant difference between Pre-PAPE and Pre -Match with the rest of post-intervention tests.

In Table 3, intragroup differences in CMJ can be appreciated for experimental and control groups. The control group presents lower values for CMJ in the Post-PAPE, Pre-Match, and Sets one to five tests, whereas the experimental group increases the jump height in comparison with Pre-PAPE (baseline) values in all tests. On the other hand, another baseline at the beginning of the match (Pre-Match) allows for the analysis of jump performance in the five sets of the match, which showed different behaviors between both groups. The experimental group showed significant differences with the second set, while the control group did in the third and fifth sets.

Therefore, the improvement percentage values are lower in the control group than the experimental group as observed in Figure 2. There are significant differences between groups until Set 2. Also, it could be an increase in the values of improvement percentage in the experimental group up to Set 2, where there is a drop in performance until the end of the intervention, Set 5. On the other hand, the control group does not show significant improvement percentages until Set 3.

improvement percentages until Set 3.

**Figure 2.** Comparison of the improvement values from Pre-PAPE performance expressed as percentage. \* Significant difference between experimental and control groups at the time points: pre-PAPE, Pre-match and Set 1, Set 2, Set 3, Set 4, and Set 5. Bars, whiskers, and dots represent mean, standard deviation and individual values, respectively. **Figure 2.** Comparison of the improvement values from Pre-PAPE performance expressed as percentage. \* Significant difference between experimental and control groups at the time points: pre-PAPE, Pre-match and Set 1, Set 2, Set 3, Set 4, and Set 5. Bars, whiskers, and dots represent mean, standard deviation and individual values, respectively.

groups until Set 2. Also, it could be an increase in the values of improvement percentage in the experimental group up to Set 2, where there is a drop in performance until the end

#### **4. Discussion 4. Discussion**

The aim of this study was to analyze the effects of PAPE for professional female volleyball players during a match. In general, the results highlighted that squat-based preactivation stimulates higher levels of PAPE shown as improvements in VJ height after activation, these improvements in VJ remained for several minutes during the match. To our knowledge, this is the first study of this kind with elite volleyball players. The aim of this study was to analyze the effects of PAPE for professional female volleyball players during a match. In general, the results highlighted that squat-based pre-activation stimulates higher levels of PAPE shown as improvements in VJ height after activation, these improvements in VJ remained for several minutes during the match. To our knowledge, this is the first study of this kind with elite volleyball players.

PAP is an electrically evoked mechanism that produces an increase of muscle strength and twitch forces, as a result of its contractile history [12,47]. According to this, PAP produces improvements in the rate of force development (RFD) and maximal voluntary contractions (MVC) for a specified level of neural activation [12]. PAP is induced by the rise in the phosphorylation of regulatory light chains, which renders actin-myosin more sensitive to submaximal Ca2+ concentrations [48]. This activation occurs with more intensity in the fibers with the isoform II, which are involved in high intensity and short duration actions such as the VJ [19]. Other factors, such as the reduction in the pennation angle after a maximal voluntary contraction, are also suggested as possible mechanisms PAP is an electrically evoked mechanism that produces an increase of muscle strength and twitch forces, as a result of its contractile history [12,47]. According to this, PAP produces improvements in the rate of force development (RFD) and maximal voluntary contractions (MVC) for a specified level of neural activation [12]. PAP is induced by the rise in the phosphorylation of regulatory light chains, which renders actin-myosin more sensitive to submaximal Ca2+ concentrations [48]. This activation occurs with more intensity in the fibers with the isoform II, which are involved in high intensity and short duration actions such as the VJ [19]. Other factors, such as the reduction in the pennation angle after a maximal voluntary contraction, are also suggested as possible mechanisms of PAP.

of PAP. On the other hand, PAPE is associated whit an intensification in force production induced by previous muscle activity (i.e., voluntary contraction), and its presence is confirmed by performance outcomes [10,12]. Mechanisms proposed for PAPE are different from PAP, nevertheless, there are not well defined yet, but PAPE may be associated whit more lasted processes such as an increase in muscle temperature [49,50]. Also, the Muscle flow or/and water content and muscle activation (Partly through motivation) are mechanisms proposed for PAPE [12]. Finally, the increase in plasma catecholamines induced by exercise [51], and intensification in excitability of high order motor units [48,52,53] are proposed as mechanisms of PAPE, their effects may be observed until 20 min after Pre-On the other hand, PAPE is associated whit an intensification in force production induced by previous muscle activity (i.e., voluntary contraction), and its presence is confirmed by performance outcomes [10,12]. Mechanisms proposed for PAPE are different from PAP, nevertheless, there are not well defined yet, but PAPE may be associated whit more lasted processes such as an increase in muscle temperature [49,50]. Also, the Muscle flow or/and water content and muscle activation (Partly through motivation) are mechanisms proposed for PAPE [12]. Finally, the increase in plasma catecholamines induced by exercise [51], and intensification in excitability of high order motor units [48,52,53] are proposed as mechanisms of PAPE, their effects may be observed until 20 min after Pre-activation at least. However, more investigation is needed in order to confirm those effects.

Most studies showed that PAPE protocol increased the performance in VJ in volleyball players [35,52–55]. Similar results can be found in our study with elite players as the experimental group showed improvements in VJ performance, while the control group has an opposite trend. Nevertheless, these differences between Pre-PAP and Post-PAPE tests are not statistically significant (*p* > 0.05). These results are in concordance with the study by [26,36] in which improvements in VJ were found, although not statistically significant. Due to the difficulty of access to elite athletes, the low sample size in our study may limit the statistical power to show differences between measures taken before and after activation. However, a moderate effect size of the VJ performance was observed, confirming the jump improvement tendency observed.

Volleyball players usually perform a typical explosive strength workout [56]. Those workouts include intensity loads ranging from 40 to 70% of 1RM, which are far from 90% of 1RM, and therefore, the classification as trained subjects [26,57] must be questioned. intensity loads of 90% of 1RM could produce an excess of fatigue in volleyball players and, as a result, the subjects could become non-responders, according to the criteria of [21,58]. Under these circumstances, the load may not fully adjust to the characteristics of the group, and therefore the response obtained is a smaller quantity than expected. Therefore, it is necessary to individualize the PAPE very carefully in order to adjust the activation intensity and volume loads to the individual characteristics of female volleyball players. Previous studies comparing routines based on peak strength and hypertrophy find that explosive-based workouts generate less fatigue [59].

On the other hand, if the improvement percentage values of both groups are compared, as shown in Figure 2, there are significant differences between control and experimental groups in the improvement reached in the post-PAPE tests. Positive improvement percentage values were observed for the experimental group (4.12%) while the control group showed an opposite trend (−5.37%). In addition, these statistically significant results are consistent with their moderate effect sizes, as depicted in Table 3, showing practical significance for the improvement percentage and the jump height reached in the CMJ in the Post-test. Hence, our study suggests that a conditioning activity would generate a positive effect on VJ performance, i.e., PAPE, as a result of an increase in muscle strength obtained 8 min after activation protocol [29,31,47,52,56,59].

The jumps distribution profile during the match was clearly different for control and experimental groups, which suggests that the effect of the activation could be one of the causes of this difference. After the peak of PAPE had occurred in the Post-PAPE test for the experimental group, the CMJ heights still progress, as shown in Table 2, peaking at Set 2 and reaching the end of the match with values similar to those at the start. These results agree with studies in which CMJ is used to evaluate fatigue after using loads and intensities higher than those used in our study (3 sets of 3 repetitions, 90% 1 RM) [59], in the analyzed study a decrease of 6% is observed immediately after the load, but an increase of 2% is observed in the CMJ 24 h after workout. Significant differences and large effect sizes were observed for all test occasions compared to the Pre-PAP test. However, the control group showed a different trend: all VJ heights were lower than the experimental, only Set 3 and Set 5 showed significant differences in regards to the Pre-PAPE test, and peaking at Set 3, later than experimental. This trend can also be analyzed through improvement percentage, shown in Figure 2. The experimental group achieved larger values and, again, showed a peak in Set 2, followed by a decrease in improvement percentage. The control group did not show improvement until Set 3.

The difference between groups could be explained by the presence of PAPE in the experimental group, which effect would extend beyond the time window of 7–12 min [25,26], increasing the jump performance, and also making the athlete more sensitive to future stimulus. Therefore, if PAP is combined with more explosive actions in warm-up routines, a summative effect may occur and therefore a performance improvement. As a result, the effect of PAPE combined with a standard volleyball warm-up, in which numerous jumps are executed [60], may be effective for elicited VJ enhancements.

The effects of PAPE last longer than those of PAP [12], but as in PAP, these effects will depend on the relationship with accumulated fatigue [11]. For the experimental group, the effects of PAPE could be largest than fatigue until Set 2 and consequently, a better improvement in jump performance than in the control group is observed The possible effects of PAPE were evaluated from activation at times ranging from 2 to 20 min maximum

in other protocols [30,36]. In our study, CMJ tests were taken in longer time spans: after activation (8 min), at the beginning of the match (23 min), and in Sets 1 to 5 (46, 68, 95, 120, and 123 min, respectively). The experimental group peaked at Set 2 that occurred at 45 min from the beginning of the match and 68 min from activation, while the control group peaked at 90 min after activation. From Set 3 onwards, both groups appear to have similar conditions, and the values for improvement percentage are similar. The possible effect of accumulated fatigue, in addition to the dissipation of activation, causes the behavior of both groups to be more similar, which could be understood as the effect of PAPE is no longer present in the experimental group from Set 3 to the end of the match (i.e., from 90 to 130 min). Performance improvements are probably not mostly due to PAPE, but it is intuited that PAPE helps to generate a summation effect that produces an increase in the performance in VJ of volleyball players.

However, attributing the improvement percentage exclusively to the effect of PAPE generated by an initial conditioning activity and warm-up would not be entirely correct. Numerous physical and cognitive factors can affect the final performance, which is very difficult to control in a real game situation. For these reasons, the individualization of stimuli is very important. Despite this, the two groups in this study were in similar situations and only the group that performed a previous potentiation obtained a better improvement percentage in all sets with a greater effect size magnitude, and therefore, sports performance will be greater in this group.

The main limitation of this study is the sample size due to limited access to elite players in match conditions. The restricted statistical power because of the sample size in this study may have influenced the significance of some of the statistical comparisons conducted. A post hoc power analysis revealed that, for the lowest effect size of interest observed in the present study (*d* = 0.7), the number of players would have been approximately 25 for each group to obtain statistical power at the recommended 80% level. The results of this study serve as a basis that can be generalized to larger populations. Thus, more investigation with larger samples is needed to determine the effects of PAPE in volleyball female players and related sports.

#### **5. Conclusions**

The use of conditioning activity consisting of three repetitions of 90% of 1RM in the back half-squat exercise generates differences in the increase in CMJ heights between control and experimental group in Post-PAPE tests and elicited larger PAPE effects that remain until the second set of a volleyball match. The results of this study suggest that, if the activation is fitted individually on the correct form, and combined with an optimal warm-up, PAPE may be used to improve vertical jump performance, a key feature in volleyball and other related sports. Therefore, the inclusion of such protocols in volleyball warm-ups should be considered by coaches and physical trainers of volleyball teams. However, further investigation should be carried out, following different warm-up strategies with a wider sample in order to generalize the results achieved in the present study.

**Author Contributions:** Conceptualization, L.V.-G., B.P. and J.M.J.-O.; Data curation, A.P.-T.; Formal analysis, S.S.-A. and B.P.; Funding acquisition, A.P.-T.; Investigation, L.V.-G., A.P.-T., S.S.-A. and J.M.J.-O.; Methodology, A.P.-T., S.S.-A., B.P. and J.M.J.-O.; Supervision, J.M.J.-O.; Writing-original draft, L.V.-G. and B.P.; Writing-review & editing, L.V.-G., A.P.-T., S.S.-A., B.P. and J.M.J.-O. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by Generalitat Valenciana, grant number GV/2021/098.

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Ethics Committee of University of Alicante (UA-17 November 2018, date of approval 20 December 2018).

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** Data can be obtained through the corresponding author on reasonable request.

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

#### **References**


## *Article* **Effects of Acute Microcurrent Electrical Stimulation on Muscle Function and Subsequent Recovery Strategy**

**Alessandro Piras <sup>1</sup> , Lorenzo Zini <sup>2</sup> , Aurelio Trofè 2 , Francesco Campa 2,\* and Milena Raffi <sup>2</sup>**


**\*** Correspondence: francesco.campa3@unibo.it

**Abstract:** Microcurrent electrical neuromuscular stimulation (MENS) is believed to alter blood flow, increasing cutaneous blood perfusion, with vasodilation and hyperemia. According to these physiological mechanisms, we investigated the short-term effects of MENS on constant-load exercise and the subsequent recovery process. Ten healthy subjects performed, on separate days, constant-load cycling, which was preceded and followed by active or inactive stimulation to the right quadricep. Blood lactate, pulmonary oxygen, and muscle deoxyhemoglobin on-transition kinetics were recorded. Hemodynamic parameters, heart rate variability, and baroreflex sensitivity were collected and used as a tool to investigate the recovery process. Microcurrent stimulation caused a faster deoxyhemoglobin (4.43 ± 0.5 vs. 5.80 ± 0.5 s) and a slower VO<sup>2</sup> (25.19 ± 2.1 vs. 21.94 ± 1.3 s) on-kinetics during cycling, with higher lactate levels immediately after treatments executed before exercise (1.55 ± 0.1 vs. 1.40 ± 0.1 mmol/L) and after exercise (2.15 ± 0.1 vs. 1.79 ± 0.1 mmol/L). In conclusion, MENS applied before exercise produced an increase in oxygen extraction at muscle microvasculature. In contrast, MENS applied after exercise improved recovery, with the sympathovagal balance shifted toward a state of parasympathetic predominance. MENS also caused higher lactate values, which may be due to the magnitude of the muscular stress by both manual treatment and electrical stimulation than control condition in which the muscle received only a manual treatment.

**Keywords:** MENS; oxygen consumption; deoxyhemoglobin kinetics; near-infrared spectroscopy; lactate; cycling

#### **1. Introduction**

Microcurrent electrical neuromuscular stimulation (MENS) involves series of stimuli delivered superficially, in the microampere range, through special transducer gloves that allow managing microcurrent signals through manipulation techniques. It is a key component for many medical and sport applications, and it is largely used for rehabilitation, training, and recovery purposes [1].

Nowadays, interest in the use of low-intensity current such as MENS is increasing, as its effects take place at the cell level (protein synthesizing activity; increased ATP generation), with sub-sensory application (i.e., painless), besides the absence of collateral effect, low cost, and easy utilization [2]. The utilization of electric field and currents comparable to different cells results in the stimulation of growth and tissue restoration [3] and diminution of edema [4]. Electric stimulations ranging from 10 to 1000 µA increase ATP levels and protein synthesis of rat skin, without having an impact on DNA metabolism [2]. The consequences on ATP production are described by proton actions [5], whereas the amino acids transport through the cell are facilitated by the alterations of the electrical gradients across the membranes [2]. Throughout stimulation of damaged muscles, MENS manages the modified membrane function by various processes, such as the preservation of intracellular Ca2+ homeostasis and with the augmented production of ATP levels [6].

**Citation:** Piras, A.; Zini, L.; Trofè, A.; Campa, F.; Raffi, M. Effects of Acute Microcurrent Electrical Stimulation on Muscle Function and Subsequent Recovery Strategy. *Int. J. Environ. Res. Public Health* **2021**, *18*, 4597. https:// doi.org/10.3390/ijerph18094597

Academic Editor: Olga Scudiero

Received: 10 March 2021 Accepted: 24 April 2021 Published: 26 April 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

Prior studies demonstrated that muscle damage treatment through microcurrent at low amperage (<500 µA) can decrease the severity of muscle symptoms [7] and a quicker regrowth of atrophied animals leg muscles [8]. In addition, microcurrents tone up the smooth muscles of blood vessels, improve skin turgor and tissue temperature, with an increase in blood flow through area treated [9]. All these characteristics are associated with vasodilation, then stimulating the metabolism of waste and toxins from the blood, therefore increasing healing and decreasing pain [2].

Based on these health-related cellular effects, a combination of stimulation plus exercise might improve exercise performance, and it might also be valuable to accelerate the subsequent recovery, thanks to the increased muscle blood flow that accelerates muscle metabolites removal [10]. One of the most physiological variables used to evaluate the recovery process is the heart rate variability (HRV). At the end of exercise, HRV returns exponentially to control value, and its increment is functionally related to athletes' training status and to the exercise intensity previously executed. HRV is the tool used to analyze the cardiac autonomic responses in combination with the baroreflex sensitivity (BRS), which is a reflex that adapts the heart period in response to variations in systolic blood pressure. These parameters have been used to evaluate the different adaptations to exercise and the recovery times after exercise [11–15]. Regardless of several research and medical applications, few studies have investigated the MENS effect before or after endurance exercise [7,16]. To date, only one study has investigated MENS effects in combination with aerobic exercise in reducing abdominal fat [16]. Authors found that microcurrent application with a frequency range of 25–50 Hz, combined with aerobic exercise, led to a significant decrease in subcutaneous abdominal fat thickness through the lipolysis stimulation [16]. Furthermore, the majority of studies performed with MENS reported a significant reduction of delayed onset muscle soreness after strength exercise [7,17,18]. In elderly people, Kwon et al. [19] found that MENS, after 40 min of short-term application, has an effect on muscle function, enhancing handgrip strength and single leg heel-rise.

Considering the influence of MENS on microcirculation, vascularization, and cellular energy production described above, and that endurance exercise stimulates the microvascular oxygenation following the onset of contractions [20], it could be interesting to investigate the effect of MENS stimulation on muscle tissue oxygenation and its influence on pulmonary oxygen kinetics during cycling. The rapid increase of the pulmonary oxygen kinetics at the transition between rest and exercise is a determinant of aerobic performance and an indicator of a well-done state of oxidative energetic system activity [21]. Additionally, the faster rise in VO<sup>2</sup> after the onset of exercise indicates a higher muscle oxygen utilization, which is a characteristic of elite athletes and trained subjects [22].

Until now, to our knowledge, no studies have investigated the effects of MENS at the human muscle tissue level, and more precisely, on factors of endurance capacity that are related to performance, such as faster oxygen kinetics, higher muscle oxygen release, or reduced blood lactate level at higher aerobic intensity. Thus, it is possible to hypothesize that the instantaneous and short-term effects of MENS might enhance the individual's capabilities on exercise at submaximal intensities and to accelerate the subsequent recovery process. Therefore, the aim of our study was to evaluate the acute effects of MENS on the muscle endurance capacity and subsequent recovery in sport. The results could be of great importance for elucidation of the O<sup>2</sup> release to acute, localized MENS exposure and the development of efficacious performance and recovery modalities.

#### **2. Materials and Methods**

#### *2.1. Participants and Inclusion Criteria*

Experiments were performed in 10 healthy subjects (2 females, 8 males; mean ± SD: age 27.2 <sup>±</sup> 3.6 years; body mass index (BMI) 23.4 <sup>±</sup> 2.5 Kg/m<sup>2</sup> ; VO2peak 49.9 ± 7.9 mL/kg/min). The subjects were recreationally active but not highly trained. All subjects were volunteers, healthy, non-smokers, and none of them were taking medications or supplements. None of the subjects reported physical deficit or injuries during the study. All participants received a verbal explanation of experimental procedures, and informed consent was obtained before the beginning of recordings. In agreement with the Declaration of Helsinki, the experimental protocol was approved by our University Institutional Ethic Committee. sent was obtained before the beginning of recordings. In agreement with the Declaration of Helsinki, the experimental protocol was approved by our University Institutional Ethic Committee.

were volunteers, healthy, non-smokers, and none of them were taking medications or supplements. None of the subjects reported physical deficit or injuries during the study. All participants received a verbal explanation of experimental procedures, and informed con-

*Int. J. Environ. Res. Public Health* **2021**, *18*, x FOR PEER REVIEW 3 of 15

#### *2.2. Study Design and Test Protocol 2.2. Study Design and Test Protocol*

The study design was a cross-sectional, single-blind, randomized controlled trial. For the realization of this study, the participants visited our laboratory five times, with at least three days between each visit, in which we performed different recordings. In the first visit, the participants performed an incremental test on a cycle-ergometer (H-300-R Lode), to determine ventilatory threshold (VT), respiratory compensation point (RCP), and peak oxygen consumption (VO2peak) to identify their individual workload for the succeeding four recording sessions. Expirated gases were analyzed using a Quark b<sup>2</sup> breath-by-breath metabolic system (Cosmed, Rome, Italy). After the incremental session, the subjects came to our laboratory and performed two repetitions of each conditions of ON (MENS stimulation) and OFF (sham stimulation). The cycling exercise protocol consisted of 1 min of unloaded exercise followed by 5-min of heavy-intensity exercise. On the following days, the four recordings were performed in random order, and the participants were never informed about the status of stimulation, because it was not perceived by the subjects at the cutaneous level (single-blind). The study design was a cross-sectional, single-blind, randomized controlled trial. For the realization of this study, the participants visited our laboratory five times, with at least three days between each visit, in which we performed different recordings. In the first visit, the participants performed an incremental test on a cycle-ergometer (H-300-R Lode), to determine ventilatory threshold (VT), respiratory compensation point (RCP), and peak oxygen consumption (VO2peak) to identify their individual workload for the succeeding four recording sessions. Expirated gases were analyzed using a Quark b<sup>2</sup> breath-by-breath metabolic system (Cosmed, Rome, Italy). After the incremental session, the subjects came to our laboratory and performed two repetitions of each conditions of ON (MENS stimulation) and OFF (sham stimulation). The cycling exercise protocol consisted of 1 min of unloaded exercise followed by 5-min of heavy-intensity exercise. On the following days, the four recordings were performed in random order, and the participants were never informed about the status of stimulation, because it was not perceived by the subjects at the cutaneous level (single-blind).

The incremental test consisted of one minute of unloaded pedaling, followed by a warm-up of 5 min at 50 W. Then, at a constant cycling frequency of 75 rpm, the power output started at 80 W and was increased of 20 W/min until volitional exhaustion was reached or the required pedal rate could not be maintained [23]. Ventilatory and gas exchange variables were measured continuously breath by breath throughout the test. The highest VO<sup>2</sup> averaged over a 20 s interval was taken as VO2peak. The LT and RCP were estimated from gas exchange measurements using the V-slope method, ventilatory equivalents, and end-tidal gas tensions [24]. Briefly, VT was determined from a different measurement, such as (i) the first unbalanced increase in CO<sup>2</sup> production (VCO2) with respect to VO2; (ii) an increase in expired ventilation (VE/VO2) with no increase in VE/VCO2; and (iii) an increase in end-tidal oxygen tension with no fall in end-tidal carbon dioxide tension. RCP was determined from a number of measurements including (i) an increase in VE/VCO2; and (ii) an increase in end-tidal CO<sup>2</sup> tension. The incremental test consisted of one minute of unloaded pedaling, followed by a warm-up of 5 min at 50 W. Then, at a constant cycling frequency of 75 rpm, the power output started at 80 W and was increased of 20 W/min until volitional exhaustion was reached or the required pedal rate could not be maintained [23]. Ventilatory and gas exchange variables were measured continuously breath by breath throughout the test. The highest VO<sup>2</sup> averaged over a 20 s interval was taken as VO2peak. The LT and RCP were estimated from gas exchange measurements using the V-slope method, ventilatory equivalents, and end-tidal gas tensions [24]. Briefly, VT was determined from a different measurement, such as (i) the first unbalanced increase in CO<sup>2</sup> production (VCO2) with respect to VO2; (ii) an increase in expired ventilation (VE/VO2) with no increase in VE/VCO2; and (iii) an increase in end-tidal oxygen tension with no fall in end-tidal carbon dioxide tension. RCP was determined from a number of measurements including (i) an increase in VE/VCO2; and (ii) an increase in end-tidal CO<sup>2</sup> tension.

Then, participants come to our laboratory for the exercise sessions. The procedure followed 4 stages: baseline; MENS-pre-exercise; exercise; and MENS-post-exercise, as illustrated below (Figure 1). Then, participants come to our laboratory for the exercise sessions. The procedure followed 4 stages: baseline; MENS-pre-exercise; exercise; and MENS-post-exercise, as illustrated below (Figure 1).

**Figure 1.** Graphical overview of the experimental protocol. HRV = heart rate variability, BRS = baroreflex sensitivity, Hp = hemodynamic parameters, MENS = microcurrent electrical neuromuscular stimulation, RPE = rate of perceived exertion, VO<sup>2</sup> = oxygen consumption, HHb = deoxyhemoglobin value. Gloves black and white represent active (ON) and inactive (OFF) stimulation, respectively. **Figure 1.** Graphical overview of the experimental protocol. HRV = heart rate variability, BRS = baroreflex sensitivity, Hp = hemodynamic parameters, MENS = microcurrent electrical neuromuscular stimulation, RPE = rate of perceived exertion, VO<sup>2</sup> = oxygen consumption, HHb = deoxyhemoglobin value. Gloves black and white represent active (ON) and inactive (OFF) stimulation, respectively.

*Baseline.* Participants stayed in a supine position in a quiet room, with a comfortable temperature (22–25 ◦C), for 10 min. They underwent noninvasive continuous blood pressure monitoring using servo-controlled infrared finger plethysmography (Portapres device; TNO/BMI, Amsterdam, the Netherlands) for analysis of heart rate variability (HRV) and baroreflex sensitivity (BRS). HRV is the amount of heart rate fluctuations around the mean heart rate, and it reflects the cardiorespiratory control system. The BRS is an established tool for the assessment of the sympathetic and parasympathetic role of the autonomic nervous system [25]. Tests were performed under a standardized procedure at the same time of the day (9:00–12:00) to prevent circadian effects. Then, we took a blood sample at ear lobe for lactate measurement (Lactate Scout, SensLab, Leipzig, Germany). The reliability of the portable blood lactate analyzer was <0.5 mM for concentrations in the range of ≈1.0–10 mM [26].

*MENS pre-exercise*. Participants, in the same supine position, after the baseline procedure and before exercise, were manipulated with MENS (Electra Microlab, LED, Via Selciatella, Italy). The operator, one of the authors, applied electric pulses through the machine using special transducer gloves to manage microcurrent signals over the most common manipulation techniques, which is necessary to massage and stimulate the quadriceps of the right leg. During sham stimulation (MENS OFF), the operation followed the identical procedure used during the real stimulation, except for the fact that the instrument was turned off. It was possible because the electric pulse was not perceived by the subjects at the cutaneous level (single-blind). The entire massage and stimulation had a duration of 20 min with a Faradic current with rectangular waveform (1 s of impulse duration; frequency of 256 Hz; amplitude of 400 µA; Positive/Negative polarity with change direction), with the intention to stimulate hyperemic vasodilation. After stimulation, we took a second blood sample for lactate measurement; then, participants were ready for the exercise test.

*Exercise.* The data collected during the incremental test were used to calculate the work rates used during the subsequent constant-load exercise tests. Specifically, the individualized workload for each athlete (mean ± SD: 311.4 ± 70.1 watt) corresponded to ~ ≈50% of the difference between power (watt) reached at VT and at the RCP (≈50% ∆ RCP-VT). Pedaling frequency was kept at about 70–80 revolutions/min. On-transitions were from unloaded pedaling to the imposed load, which was attained in about 3 s. Pulmonary ventilation (VE), oxygen consumption (VO2), and carbon dioxide output (VCO2) were determined with a Quark b<sup>2</sup> breath-by-breath metabolic system (Cosmed, Rome, Italy) previously calibrated according to the manufacturer's guidelines (included room air calibration, reference gas calibration, and turbine calibration with a 3-L syringe).

The changes in the vastus lateralis muscle oxygenation were evaluated by nearinfrared spectroscopy (NIRS). A portable NIRS single-distance continuous-wave photometer (NIMO, Nirox Srl, Brescia, Italy) was utilized for the present study. In brief, the procedure is based on the changes of oxygen absorption with near infrared light, and it includes an emission probe that emits 3 wavelengths (685, 850, and 980 nm) and a photon detector. The transmitted light was recorded continuously at 40 Hz and utilized to quantity deoxygenated myoglobin and hemoglobin levels [27]. The deoxygenated value is less conditioned by the variations of the blood flow, and it is considered as a measure of fractional oxygen extraction inside the microvascular tissue [28]. After that, we had carefully shaven the skin, we attached the probe on it, covering, about 10–12 cm above the knee joint, the lower extremity of the right leg vastus lateralis muscle [29]. Then, the probe and the skin were wrapped with black cloth to prevent corruption from ambient light.

At the third minute of effort, we took a third blood sample for lactate measurement. At the end of the cycling, rate of perceived exertion was recorded with 6–20 Borg scale. Participants were asked how hard they felt the exercise [30]. The constant load exercise had a duration of about 10 min.

*MENS-post-exercise.* Immediately after exercise, participants were back positioned in a supine position, in the same room used before the exercise. To assess the effect of MENS on recovery, we stimulated the quadriceps of the right leg for 20 min with the same protocol described for MENS-pre-exercise. After that, in order to quantify the recovery level, we took a fourth blood lactate and a second continuous blood pressure monitoring using the same plethysmography used before exercise (Portapres device) for HRV and BRS analysis.

#### *2.3. Data Analysis*

*VO<sup>2</sup> and HHb Kinetics*. Breath by breath VO<sup>2</sup> and muscle oxygenation (from this point forward identified as HHb, expressed in µM) data obtained in the different repetitions of the exercise protocol (ON; OFF) were time aligned, interpolated on a second-by-second basis, and then superimposed for every athlete. Average values (every 1 s) were calculated and utilized for kinetics analysis. Data equivalent to the "cardiodynamic phase", recorded during the first 20 s of the on-transition were not included from the analysis [31]. To evaluate mathematically the VO<sup>2</sup> and HHb on-transition kinetics, data were fitted with two-exponential terms (primary and slow component of Equations (1) and (2)):

$$\begin{array}{l} \text{VO}\_{2}(\text{t}) = \text{VO}\_{2}(\text{b}) + \text{AP} \, \* \, (1 - \text{e} - \text{(t-TDp/\tau p)} \, \text{(phase 2)} \, \text{(primary component)} + \\ \text{As } \, ^\ast \text{(1 - e - (t-TDs/\tau s) (phase 3) (slow component)} \end{array} \tag{1}$$

and

$$\begin{array}{l} \text{HHb(t)} = \text{HHb(b)} + \text{AP} \,\text{\*}\,(1 - \text{e} - \text{(t-TDp/\tau p)} \,\text{(phase 2)} \,\text{(primary component)} + \\ \text{As\*} \,(1 - \text{e} - \text{(t-TDs/\tau s)} \,\text{(phase 3)} \,\text{(slow component)}. \end{array} \tag{2}$$

In Equations (1) and (2), Ap, As, TDp, TDs, and τp and τs denote the amplitude, time delay, and time constant, respectively, of the primary and slow component phases. Equations (1) and (2) were used on the basis of which equations yielded the lowest sum of squared residuals. We calculated also the percent contribution of the slow component with respect to the total amplitude of the response. Moreover, the gain of VO2, as the increase in VO<sup>2</sup> above baseline to the reached steady state, and corrected for individualized workload (WL), was also calculated according to this equation (Equation (3)):

$$\text{Gain} = (\text{VO}\_2 \, [150 \,\text{s} - 180 \,\text{s}] - \text{VO}\_2 \,\text{bas})/\text{WL}.\tag{3}$$

*Heart rate variability.* Time and frequency domain parameters were calculated regarding the HRV task force guidelines [25]. For the time domain, the square root of the mean squared differences of successive R-R intervals (RMSSD), and the standard deviation of successive R-R intervals (SDRR) were examined. Spectral analysis provides two main frequency parts: low frequency (LF) ranging between 0.04 and 0.15 Hz and high frequency (HF) positioned at the breathing frequency of 12 breath/minute. It has been revealed that HF is an index of the vagal tone, whereas LF reflects both sympathetic and vagal activities. Both indices (variables with skewed distributions) were log transformed (Ln). The LF/HF ratio provide quantitative markers of the cardiac sympathetic and the vagal modulation [25].

*Baroreflex sensitivity.* It was evaluated with Beatscope version 1.1 a (TNO/BMI, the Netherlands) with a BRS add-on module based on cross correlation analysis [13,14]. The slope of the regression line between SBP (systolic blood pressure) and R-R interval (all intervals between adjacent QRS complexes resulting from sinus node depolarizations) variations are considered as an index of BRS modulation of HR.

*Hemodynamic parameters.* The pulse contour method of Wesseling (the Modelflow method) was used to evaluate cardiac output (CO), stroke volume (SV), ejection time (EJT), and total peripheral vascular resistance (TPR) from the blood pressure waveform [32].

#### *2.4. Statistical Analysis*

All data are shown as means ± SD. All dependent parameters (VO2; HHb; HRV; BRS; hemodynamic and lactate values) were compared between conditions (ON; OFF) with the paired sample t-test, in which means were considered significantly different at *p* < 0.05. To determine the magnitude of the stimulation effects, effect sizes (Cohen's *d*) were calculated as the mean difference standardized by the between-subject standard deviation and interpreted according to the thresholds: <0.20; small, >0.20–0.60; moderate, >0.60–1.20; large, >1.20–2.00; very large, >2.00–4.00; extremely large, >4.0 [33]. Data were analyzed with SPSS v22.0 (IBM, New York, NY, USA). Regression analysis was done by the least squared residuals technique. Even though several powerful and dedicated software have been commercialized for regression analysis, we used the Solver add-in bundled of Microsoft Excel [34].

#### **3. Results**

All participants completed the protocol. Individualized oxygen uptake at maximal and submaximal level, with personalized workload obtained at exhaustion during the incremental exercise are shown in Table 1. VT occurred at 63% of the VO2peak and at 56% of the maximum workload; consequently, during constant-load exercise, participants pedaled at a mean workload of ~65% (±5) of their maximum.


**Table 1.** Oxygen uptake and workload characteristics.

Abbreviations: VO2peak, peak oxygen consumption; RCP, respiratory compensation point; VT; ventilatory threshold.

*Oxygen uptake kinetics*. Figure 2 shows VO<sup>2</sup> on-kinetic analysis, in both conditions, for a typical subject. A slow component was observed in both experimental conditions, with non-significant slightly higher value during OFF than the ON condition (4.3% vs. 3.7%). Mean oxygen kinetic parameters for the exponential curve fitting are shown in Table 2. Analysis showed a slower primary component (t(9) = −3.38; *p* = 0.004; mean diff. = 3.25; d = 0.60), with a slower mean response time (t(9) = −2.57; *p* = 0.015; mean diff. = 4.88; d = 0.66) and a shorter time delay of the slow component (t(9) = 1.90; *p* = 0.045; mean diff. = −29.83; d = −0.70) during ON in comparison to the OFF condition. Moreover, VO<sup>2</sup> at the steady-state level was higher during ON than it was in the OFF condition (t(9) = 2.90; *p* = 0.010; mean diff. = 0.14; d = 0.26). We did not find any significant differences for the increase in VO<sup>2</sup> per unit increase in work rate (the gain of the primary phase), with values corresponding to 9.14 ± 0.5 and 9.05 ± 0.5 mL/min/W during the OFF and ON condition, respectively.

**Figure 2.** Characteristics of the two-component exponential model describing oxygen uptake (VO2) during the on-transient of heavy intensity. Data refer to transition (at time 0, vertical dashed line) from unloaded pedaling to constant-load exercise during OFF (■) and ON (▲) experimental condition. Data points are average values calculated over 1 s. Horizontal dashed line represents baseline. Data obtained during the first 20 s of the transition were excluded from analysis. **Figure 2.** Characteristics of the two-component exponential model describing oxygen uptake (VO<sup>2</sup> ) during the on-transient of heavy intensity. Data refer to transition (at time 0, vertical dashed line) from unloaded pedaling to constant-load exercise during OFF () and ON (N) experimental condition. Data points are average values calculated over 1 s. Horizontal dashed line represents baseline. Data obtained during the first 20 s of the transition were excluded from analysis.

**Table 2.** Pulmonary VO<sup>2</sup> on-kinetics parameters from unloaded pedaling to constant-load exercise across conditions (OFF; **Table 2.** Pulmonary VO<sup>2</sup> on-kinetics parameters from unloaded pedaling to constant-load exercise across conditions (OFF; ON).


component; As, amplitude of response for slow component; TDs, time delay for slow component; τs, time constant for slow component; MRTp and MRTs, mean reaction time for primary and slow component; Sc, slow component. Bold values with asterisk indicate significant differences between conditions at *p* < 0.05. Values are mean ± SD. VO2(b), oxygen consumption at baseline level; VO2(ss), oxygen consumption at steady-state level; Ap, amplitude of response for primary component; TDp, time delay for primary component; τp, time constant for primary component; As, amplitude of response for slow component; TDs, time delay for slow component; τs, time constant for slow component; MRTp and MRTs, mean reaction time for primary and slow component; Sc, slow component. Bold values with asterisk indicate significant differences between conditions at *p* < 0.05.

> resentative subject are shown in Figure 3. Time values of HHb were significantly lower during ON than OFF conditions both at the primary, with faster τp (t(9) = 2.96; *p* = 0.008; mean diff. = −1.37; d = −0.88) and mean response time (t(9) = 2.65; *p* = 0.013; mean diff. = −1.39; d = −0.82), and at the secondary component with faster mean response time (t(9) = 2.35; *p* = 0.022; mean diff. = −32.75; d = −0.63) from rest-to-exercise transition (see Table 3). *Muscle oxygenation parameters.* HHb on-kinetics analysis, in both conditions, for a representative subject are shown in Figure 3. Time values of HHb were significantly lower during ON than OFF conditions both at the primary, with faster τp (t(9) = 2.96; *p* = 0.008; mean diff. = −1.37; d = −0.88) and mean response time (t(9) = 2.65; *p* = 0.013; mean diff. = −1.39; d = −0.82), and at the secondary component with faster mean response time (t(9) = 2.35; *p* = 0.022; mean diff. = −32.75; d = −0.63) from rest-to-exercise transition (see Table 3).

*Muscle oxygenation parameters.* HHb on-kinetics analysis, in both conditions, for a rep-

**Figure 3.** Characteristics of the two-component exponential model describing deoxyhemoglobin (HHb) during the on-transient of heavy intensity. Data refer to transition (at time 0, vertical dashed line) from unloaded pedaling to constant-load exercise during OFF (■) and ON (▲) experimental condition. Data points are average values calculated over 1 s. Data obtained during the first 20 s of the transition were excluded from analysis. **Figure 3.** Characteristics of the two-component exponential model describing deoxyhemoglobin (HHb) during the on-transient of heavy intensity. Data refer to transition (at time 0, vertical dashed line) from unloaded pedaling to constant-load exercise during OFF () and ON (N) experimental condition. Data points are average values calculated over 1 s. Data obtained during the first 20 s of the transition were excluded from analysis.

**Table 3.** Deoxygenated hemoglobin on-kinetics parameters from unloaded pedaling to constant-load exercise across conditions (OFF; ON). **Table 3.** Deoxygenated hemoglobin on-kinetics parameters from unloaded pedaling to constant-load exercise across conditions (OFF; ON).


constant for slow component; MRTp and MRTs, mean reaction time for primary and slow component; Sc, slow component. Bold values with asterisk indicate significant differences between conditions at *p* < 0.05. *Hemodynamic, cardiac autonomic variables.* Table 4 shows all hemodynamic and auto-Values are mean ± SD. HHb(b), deoxyhemoglobin at baseline level; HHb(ss), deoxyhemoglobin at steady-state level; Ap, amplitude of response for primary component; TDp, time delay for primary component; τp, time constant for primary component; As, amplitude of response for slow component; TDs, time delay for slow component; τs, time constant for slow component; MRTp and MRTs, mean reaction time for primary and slow component; Sc, slow component. Bold values with asterisk indicate significant differences between conditions at *p* < 0.05.

> nomic variables investigated. Significant differences were found for systolic (t(9) = 2.67; *p* = 0.013; mean diff. = 7.48; d = 0.70) and mean arterial pressure (t(9) = 2.43; *p* = 0.024; mean diff. = 3.23; d = 0.51), with greater values during ON than OFF conditions. Time and frequency domain analysis showed significant differences for RMSSD (t(9) = 1.75; *p* = 0.047; mean diff. = 4.86; d = 0.28), HF (t(9) = 2.56; *p* = 0.015; mean diff. = 0.30; d = 0.37), and LF/HF ratio (t(9) = 1.95; *p* = 0.044; mean diff. = 0.42; d = 0.62) between ON and OFF conditions, respectively. It appeared that during stimulation (ON), participants recovered faster than during placebo condition (OFF). *Hemodynamic, cardiac autonomic variables*. Table 4 shows all hemodynamic and autonomic variables investigated. Significant differences were found for systolic (t(9) = 2.67; *p* = 0.013; mean diff. = 7.48; d = 0.70) and mean arterial pressure (t(9) = 2.43; *p* = 0.024; mean diff. = 3.23; d = 0.51), with greater values during ON than OFF conditions. Time and frequency domain analysis showed significant differences for RMSSD (t(9) = 1.75; *p* = 0.047; mean diff. = 4.86; d = 0.28), HF (t(9) = 2.56; *p* = 0.015; mean diff. = 0.30; d = 0.37), and LF/HF ratio (t(9) = 1.95; *p* = 0.044; mean diff. = 0.42; d = 0.62) between ON and OFF conditions, respectively. It appeared that during stimulation (ON), participants recovered faster than during placebo condition (OFF).

> **Table 4.** Hemodynamic and autonomic variables. *OFF ON* **ΔSAP (mmHg) −4.98 ±2.10 2.50 ±2.00 \* ΔDAP (mmHg)** 0.01 ± 1.50 2.72 ± 2.10 **ΔMAP (mmHg) −1.35 ±1.50 1.88 ±2.00 \* ΔCO (L/min)** 0.16 ± 0.20 0.31 ± 0.30 *Lactate and the rate of perceived exertion*. Figure 4 shows lactate trends, analyzed at baseline; after the first MENS treatment that preceded the exercise; during the third minute of exercise; and after the second MENS treatment that has followed the exercise. A T-test revealed significant differences for comparison between ON and OFF conditions done before (t(9) = 1.65; *p* = 0.048; mean diff. = 0.16; d = 0.51), and after (t(9) = 1.89; *p* = 0.046; mean diff. = 0.36; d = 0.70) cycling. Rate of perceived exertion showed not significant difference between conditions (12.45 ± 2 vs. 11.25 ± 1.8 for ON and OFF, respectively).

> > **ΔSV (mL/min)** −8.35 ± 1.60 −5.41 ± 3.40


**Table 4.** Hemodynamic and autonomic variables.

Delta values (mean ± SD) are obtained subtracting the recovery values from the baseline values. SAP, systolic arterial pressure; DAP, diastolic arterial pressure; MAP, mean arterial pressure; CO, cardiac output; SV, stroke volume; HR, heart rate; EJT, ejection time; TPR, total peripheral resistance; HRV, heart rate variability; SDRR standard deviation of the R-R intervals; RMSSD, root mean square of the successive differences; LF, low frequency; HF, high frequency; BRS, baroreflex sensitivity; Ln, logarithm. Bold values with asterisk represent significant differences at *p* < 0.05. line; after the first MENS treatment that preceded the exercise; during the third minute of exercise; and after the second MENS treatment that has followed the exercise. A T-test revealed significant differences for comparison between ON and OFF conditions done before (t(9) = 1.65; *p* = 0.048; mean diff. = 0.16; d = 0.51), and after (t(9) = 1.89; *p* = 0.046; mean diff. = 0.36; d = 0.70) cycling. Rate of perceived exertion showed not significant difference between conditions (12.45 ± 2 vs. 11.25 ± 1.8 for ON and OFF, respectively).

**Figure 4.** Mean (±SD) lactate values recorded at the baseline, after the first MENS treatment (MENS-pre-Ex) before exercise, at the third minute of the constant-load exercise (Exercise), and after the second MENS treatment (MENS-post-Ex), subsequently to exercise performance. Black dashed line represents OFF, the gray dashed line represents ON condition. Asterisks showed mean significant differences at *p* < 0.05. **Figure 4.** Mean (±SD) lactate values recorded at the baseline, after the first MENS treatment (MENSpre-Ex) before exercise, at the third minute of the constant-load exercise (Exercise), and after the second MENS treatment (MENS-post-Ex), subsequently to exercise performance. Black dashed line represents OFF, the gray dashed line represents ON condition. Asterisks showed mean significant differences at *p* < 0.05.

#### **4. Discussion**

Different studies have discovered that the application of electric fields through the human body can significantly enhance cell metabolism [35] and injury restoration [36] when applied following exercise. The rationale behind the application of MENS is based on its efficacy to generate ATP at the cellular level and other health-related benefits, such as the increase in mitochondrial numbers [16], protein synthesis [7], and the activation of hormone-sensitive lipase, which increases the lipolysis process [16]. According to these different physiological mechanisms found in MENS treatment, our intention was to investigate the short-term effects of MENS on constant-load exercise at submaximal intensities and to the subsequent recovery process. The key results of our study are that MENS stimulation applied before exercise produced an increase in oxygen extraction at muscle microvasculature, while when applied after exercise, improved recovery through faster parasympathetic reactivation with respect to control condition. Moreover, electrical stimulation caused higher lactate levels, which may be due to the magnitude of the muscular stress by both manual treatment and electrical stimulation with respect to the control condition in which the right quadricep received only a manual treatment.

*VO<sup>2</sup> and HHb kinetics.* The main finding of the present study was the faster HHb on-transition kinetics during exercise executed after MENS stimulation, and surprisingly, by slower VO<sup>2</sup> on-transition kinetics. After the onset of exercise, a delay has been reported before an increase in muscle O<sup>2</sup> consumption [37], suggesting that the activation of mitochondrial respiration does not increase immediately, but rather, it has been delayed relative to the start of exercise. It could be argued that combining MENS with exercise might have been increased vasodilation and stimulated hyperemia, which could have, consequently, released nitric oxide, with the effect of accelerating O<sup>2</sup> availability at the muscle level. Nitric oxide represents an important component of the metabolic inertia to the VO<sup>2</sup> kinetics during supra-maximal exercise [38]. The precise mechanism by which nitric oxide contributes to the metabolic inertia at exercise onset is unclear but, in vitro, it has been demonstrated its role in inhibiting several mitochondrial enzymes, as it is a competitive inhibitor of oxygen consumption in the mitochondrial respiratory chain [39]. MENS treatment has received more widespread attention in the last years, as it not only relieves pain but also has a positive effect on reparative processes in the skin [1]. Microcurrents penetrate in the body's cells, normalize the biochemical processes, such as improving metabolism, increasing enzyme activity, ATP synthesis, proteins, lipids, and other vital substances [2]. In addition, microcurrents tone up the smooth muscles of blood vessels as well as improve skin turgor and tissue temperature, with an increase in blood flow through area treated [9]. They are associated with vasodilation, then stimulating the metabolism of waste and toxins from the blood, therefore increasing healing and decreasing pain [2]. Vasodilation and hyperemic processes might have stimulated NO release, even if, to the best of our knowledge, studies are lacking to support this hypothesis.

As shown in previous studies [40,41], the NIRS-derived HHb signal provides a continuous, noninvasive measurement of changes in muscle deoxygenation and reflects the balance between local muscle O<sup>2</sup> delivery and utilization. Our results have shown an immediate increment in muscle fractional O<sup>2</sup> extraction after a few seconds of delay (≈10 s) following the onset of contraction. The rate of adaptation of muscle deoxygenation was faster than the adaptation of the primary phase of the VO2, reflecting an accelerated O<sup>2</sup> extraction in the active muscle microvasculature as a consequence of microstimulation. Our results are in agreement with other studies that investigate increasing the availability of muscle O2; through hyperoxia, adenosine, or drug administration to the O2–hemoglobin dissociation curve, which facilitated O<sup>2</sup> release at the working muscle, the primary component of pulmonary VO<sup>2</sup> does not accelerates, even during high-intensity exercise [42,43]. This is in accordance with the hypothesis that VO<sup>2</sup> during the transition from rest-toexercise is not managed by the rate of adjustment of convective oxygen delivery to the exercising muscles [42]. After the time delay, during the ON condition, HHb increased more rapidly toward a "steady-state" level, suggesting that oxygen delivery in the on-

transition was more adequate to meet the metabolic demand of the muscle, thus requiring a rapid increase in O<sup>2</sup> extraction [41]. The slightly lower but not significantly different muscle HHb value exhibited by MENS stimulation at steady-state level, in concomitant with the significantly higher value of VO<sup>2</sup> consumption at the same working rate, suggests that our procedure may have improved oxygen availability/distribution within the muscle microvasculature. The cause of the slower phase II of VO<sup>2</sup> kinetics is unclear, although it is known that this parameter is sensitive to a number of factors, including the high percentage of type II fiber distribution in the working muscles [44]. However, it is difficult to see how MENS stimulation could alter muscle fiber recruitment patterns, although this should not, of course, be excluded yet.

*Autonomic nervous system parameters.* A second purpose of the present investigation was to examine the different physiological recovery responses to MENS exposure after exercise. The common physiological variable used to evaluate recovery time is the heart rate variability. At the end of exercise, HRV returns exponentially to control value, and its increment is functionally related to the athlete's training status and the exercise intensity previously executed. HRV is the tool used to investigate the cardiac autonomic responses in combination with the baroreflex sensitivity, which is a reflex that adapts the heart period in response to variations in systolic blood pressure. These parameters have been used to evaluate the different adaptations to exercise and the recovery times after exercise [11–14]. With a transition from exercise to passive recovery, there is a loss of central command and activation of the arterial baroreflex, resulting in a decrease in heart rate toward its preexercise level [45]. The vagal system plays a main role in reducing heart rate immediately after the cessation of exercise, and its further decrease is mediated by both the vagal and sympathetic system [13]. In the present study, we found significantly different effects on the autonomous nervous system parameters, with higher increase in vagal reactivation (RMSSD and HF band of the HRV frequency spectrum) after MENS compared to shamexposure. Moreover, sympathovagal balance, assessed by LF/HF ratio, was shifted toward a state of parasympathetic predominance, revealing a faster recovery after stimulation treatment than in the control condition. A possible explanation of the microcurrent effects on faster recovery after exercise could be related to its effect on muscle metaboreflex. Until now, no study has investigated the effect of MENS on metaboreflex activity. One study found that the transcutaneous electric nerve stimulation, a technique similar to MENS (both are accepted mode of electrotherapy) [1], augments peripheral blood flow by reduction of the muscle metaboreflex, increasing oxygen supply to stimulated muscles, with a decrease in sympathetic activity evaluated with the heart rate variability [46]. These findings support the idea that the acute application of electrotherapy improves sympathovagal balance, which could be linked to an intense peripheral vasodilatation response, contributing to a faster recovery process.

*Lactate and rate of perceived exertion.* Hyperlactatemia is observed during exercise and severe inflammation [47], as well as in muscle cells subjected in vitro to electrical pulse stimulation [48]. In the present study, lactate levels were significantly higher after MENS treatments, both before and after exercise, whereas during constant-load cycling, participants produced the same lactate values in both experimental conditions. We can speculate that higher lactate values could have caused vasodilation at muscle level, through the changes in osmolarity and acidity, which are necessary to speed-up HHb on-transition kinetics but not higher enough to accelerate VO<sup>2</sup> on-kinetics. The finding that the lactate values increased with respect to inactive stimulation is somewhat surprising. We could assume that MENS increased the magnitude of muscular activity, which may be due to both manual and electrical stimulation with respect to a sham condition in which the right quadricep received only a manual treatment. Moreover, the lactate is crucial for muscle to make cytosolic NAD+, which is necessary to ATP regeneration from glycolysis, protecting muscles from acidosis. Lactate utilizes two protons, which is necessary to promoting proton elimination from muscles. Moreover, MENS efficacy on blood lactate values could be influenced by parameters used (pulse duration, frequency, amplitude, and muscles

stimulated) [1,9], target population [49], and the type of fatiguing exercise or duration of recovery [9,10]. Finally, the RPE was not significantly different between conditions, and this result is similar to that of Barcala-Furelos et al. [10], in which electrical stimulation did not alter the RPE values when compared with the passive recovery in lifeguards following a water rescue.

*Limitations of the study.* Some limitations to the current investigation warrant discussion. Although we know that NIRS has several limitations, most of them have been prevailed by recent technological developments. For example, when the probe is applied on the skin overlying the muscle that we want to investigate, NIRS can measure only a relatively small and superficial volume of skeletal muscle tissue. However, the method has also important strengths and can give valuable and noninvasive useful insights into skeletal muscle oxidative metabolism in vivo during exercise [50].

Furthermore, the main limitation of the whole-body pulmonary oxygen consumption measurements is the difficulty to differentiate between the exercising muscles and the rest of the body, or between different muscles involved in the exercise. Moreover, the presence of O<sup>2</sup> stores between the location of measurement (the mouth) and the sites of gas exchange at the skeletal muscle level complicate data interpretation during metabolic transitions [50].

The sample size was in line with the most important articles published in this area, in which the number of participants is under 10 units and unbalanced between male and female. Although we have applied it on athletes, it could be important to direct future studies on patients and aged, in which their agility and life quality are limited for impairment in oxygen delivery and utilization.

We have also highlighted the fact that exercise duration, rate of increased in work rate, blood sampling location, instrument utilized, and measurement error are all potential sources of variability in measuring lactate values. However, the reliability of our portable blood lactate analyzer was <0.5 mM for concentrations in the range of ≈1.0–10 mM, with a measurement error of ≤3% [26]. Further investigations are needed, in terms of stimulation parameters (e.g., time, frequency, amplitude, duration) and with different exercise protocols.

#### **5. Conclusions**

In summary, we found that MENS stimulation causes a faster HHb and a slower VO<sup>2</sup> on-transition kinetics during exercise, with higher lactate levels immediately after the treatments. Sympathovagal balance was shifted toward a state of parasympathetic predominance, revealing a faster recovery after stimulation executed following cycling. These results could be due to the increased vasodilation and hyperemia, which are a consequence of stimulation. It seems plausible to consider MENS as an electrotherapy useful for improving recovery through faster parasympathetic reactivation following exercise. The absence of any positive effect on VO<sup>2</sup> on-transition kinetics could be partly imputed to methodological procedures such as the arbitrary choice of stimulation intensity and duration. Nevertheless, additional studies are needed to approve or discard this hypothesis and to shed light on the correlation of these consequences with a short period of training program with concurrent MENS stimulation.

**Author Contributions:** A.P. conceived and designed research; A.P. and L.Z. performed experiments; A.P. analyzed data; A.P., L.Z., A.T., F.C., and M.R. interpreted results of experiments; A.P. prepared figures; A.P. drafted manuscript; A.P., L.Z., A.T., F.C., and M.R. edited and revised manuscript; A.P. approved final version of manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** Supported by University of Bologna, RFO 2018.

**Institutional Review Board Statement:** This study was approved by the Bioethics committee of the University of Bologna in accordance with the Declaration of Helsinki. The protocol n◦ 0021246 was approved on 03-02-2020.

**Informed Consent Statement:** Written informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** Data sharing is not applicable to this article because of the consent provided by participants on the use of confidential data.

**Acknowledgments:** The authors would like to thanks all subjects involved in this study.

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

#### **References**


## *Article* **Effect of an Endurance and Strength Mixed Circuit Training on Regional Fat Thickness: The Quest for the "Spot Reduction"**

**Antonio Paoli <sup>1</sup> , Andrea Casolo <sup>1</sup> , Matteo Saoncella <sup>1</sup> , Carlo Bertaggia <sup>1</sup> , Marco Fantin <sup>1</sup> , Antonino Bianco <sup>2</sup> , Giuseppe Marcolin <sup>1</sup> and Tatiana Moro 2,\***


**Abstract:** Accumulation of adipose tissue in specific body areas is related to many physiological and hormonal variables. Spot reduction (SR) is a training protocol aimed to stimulate lipolysis locally, even though this training protocol has not been extensively studied in recent years. Thus, the present study sought to investigate the effect of a circuit-training SR on subcutaneous adipose tissue in healthy adults. Methods: Fourteen volunteers were randomly assigned to spot reduction (SR) or to a traditional resistance training (RT) protocol. Body composition via bioimpedance analysis (BIA) and subcutaneous adipose tissue via skinfold and ultrasound were measured before and after eight weeks of training. Results: SR significantly reduced body mass (*p* < 0.05) and subcutaneous abdominal adipose tissue (*p* < 0.05). Conclusions: circuit-training SR may be an efficient strategy to reduce in a localized manner abdominal subcutaneous fat tissue depot.

**Keywords:** spot reduction; body composition; resistance training; adipose tissue

#### **1. Introduction**

Regular physical activity can impact body composition, reducing fat mass and therefore positively improving health status. Accumulation of adipose tissue (AT) in specific areas of the body can be influenced by lifestyle behavior, such as working for most of the time in a sitting position or using only the upper body. Adipose tissue does not develop regularly but normally spread in distinctive anatomical depots [1]. Approximately 10–20% of total fat mass is contained in the visceral adipose tissue (VAT), located centrally and surrounds the internal organs [2]. The majority of total body fat is represented by the subcutaneous AT (SAT) positioned immediately below the skin: SAT normally accumulates in the gluteal, femoral and abdominal region and its distribution are regulated by different physiological and/or hormonal variables [3]. A small portion of AT consists in the ectopic AT and is localized around vital organs, such as the liver, heart and kidney.

The total amount of fat mass is considered a risk factor for several cardiometabolic diseases [4,5]; however, the location of lipids storage seems also to be critical for cardiometabolic consequences [6–8]. If central obesity is associated with metabolic dysfunction and hypertension [9,10], lower-body fat accumulation appears to have a protective effect and seems to be negatively correlated with cardiovascular disease and type 2 diabetes mellitus development [11,12]. The reduction of total body fat can be achieved through diet and/or exercise intervention [13,14]. While it has been widely demonstrated that an adequate amount of physical activity can have a favorable impact on the weight loss process [15], the existence of a "localized fat loss" is still on debate. As a matter of fact, for more than 60 years, the possibility of a localized removal of AT has raised interest in the scientific and social community. Even the "father" of the Mediterranean diet, Ancel Keys, admitted

**Citation:** Paoli, A.; Casolo, A.; Saoncella, M.; Bertaggia, C.; Fantin, M.; Bianco, A.; Marcolin, G.; Moro, T. Effect of an Endurance and Strength Mixed Circuit Training on Regional Fat Thickness: The Quest for the "Spot Reduction". *Int. J. Environ. Res. Public Health* **2021**, *18*, 3845. https:// doi.org/10.3390/ijerph18073845

Academic Editors: António Carlos Sousa and Pantelis T. Nikolaidis

Received: 11 March 2021 Accepted: 5 April 2021 Published: 6 April 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

the possibility of a localized fat reduction, although not in a scientific journal, but rather on Vogue in 1956 [13]. Later, in the '50s, some researchers reported that certain sports, such as gymnast [16], basketball [17] or running [18], promoted greater loss of fat mass in those parts of the bodies that were vigorously exercised. Since then, different strategies have been developed to advance the localized loss of SAT with exercise, and all those protocols have been termed "spot reduction". More recently, Stallknecht and coll. [19], hypothesized that exercise on specific muscles may induce "spot lipolysis" via an increased blood flow and release of fatty acids in the SAT nearby the contracting muscle regardless of exercise intensity. However, most of the studies found conflicting conclusions: some authors found a positive effect of spot reduction on localized lipolysis [20–22], while others were inconclusive [23–26]. The discrepancy between results can be found on the several exercise modalities employed, on the different body areas examined and, on the technique used for measuring SAT [27]. On the latter, most of the studies used skinfolds to evaluate changes after training [20–25]; recent comparative studies indicated that skinfold measurements do not permit accurate evaluation of SAT thickness because it is operator-dependent and influenced by anatomical site and skin thickness [28,29]. Regarding training modalities, it is well known that combining in the same training session endurance and strength exercises may exert a greater effect on total body fat loss [30] and, as recently demonstrated, and it might also have some positive effects on regional fat loss [31]. However, the effects of an alternation of strength and endurance training (mixed circuit training: MCT) has, until now, not been investigated. As a matter of fact, we demonstrated that a MCT induces greater total body fat loss and an improvement of metabolic variables compared to endurance training [32,33] but, except for our pilot trial in the 900 s [34] no one analyzed the effects of a MCT on regional fat loss.

In the light of the above, the purpose of the present study was to reconsider the spot reduction approach using a modified MCT protocol. MCT protocols usually alternate various total body strength exercises with short bouts of aerobic training. In the present study, we aimed to emphasize the positive effect of MCT by streamlining the order of proposed exercise to focus the major metabolic stress on the target body areas. To our best knowledge, this strategy has not been explored yet. We hypothesized that an MCT protocol concentrated on specific muscle would have exerted a great local lipolytic effect compare to a non-circuit mixed training. We tested our hypothesis on a group of healthy adults, using skinfolds and a modern ultrasound technique to measure SAT before and after eight weeks of training targeting abdominal and triceps.

#### **2. Materials and Methods**

This study is a randomized controlled parallel study. The study was approved by the ethical committee of the Department of Biomedical Sciences, University of Padova (HEC-DSB 05/17, 22 March 2017), according to the current Declaration of Helsinki. All participants read and signed a written informed consent form before enrollment.

Subjects were evaluated in a single visit before and after eight weeks of intervention. During the visit, height and body mass were measured, body composition was assessed via bioimpedance analysis (BIA) and skinfolds, and ultrasound was used to quantify the thickness of the adipose panicle. All measurements were taken by the same operator before and after the study. After the first screening visit, participants were randomized into two different groups: spot reduction (SR) or traditional resistance training (RT) and started the supervised training program.

#### *2.1. Subjects*

Eighteen volunteers (9 female and 9 male) aged between 20 and 46 years took part in the study. To be included in the protocol, subjects had to pass a medical interview and be aged 18–50 years old. Exclusion criteria were more than 1 year of training experience, chronic use of medication, metabolic disorders or any other clinical problems that could be aggravated by the study procedures or engagement in weight loss dietetic regimen.

During the intervention, participants were allowed to continue their recreational physical activity, but were instructed not to perform any structured, high-impact training. Four subjects (two from each group) were excluded from the final analysis due to noncompliance with the training schedule. To be considered for the analysis, subjects had to complete all sessions and maintain a frequency of 3 training per week. Table 1 shows the anthropometric characteristics at baseline.


**Table 1.** Baseline characteristics of spot reduction (SR) and resistance training (RT) groups.

All values are means ± SD.

#### *2.2. Measurements*

Body mass index (BMI) was calculated in kg/m<sup>2</sup> , obtained from body mass and height measurement using a Wunder stadiometer (Holtain Ltd., Crymych, UK) with a precision of 0.1 kg and 1 cm, respectively.

Before proceeding with the measurements of the adipose panniculus by skinfolds and ultrasound, specific detection points were traced on the right portion of the body. For skinfolds, the 4 points described by the Durnin protocol [35] were used: bicipital, triceps, suprailiac and subscapularis. With regard to ultrasound scans, to standardize the procedure and detection points between subjects, we employed the protocol described by Muller et al. [36], and 8 regions were selected for the analysis: upper abdomen, lower abdomen, spinal erectors, distal triceps, brachioradialis, front thigh, medial calf, lateral thigh.

A mechanical caliper (GIMA, Gessate MI, Italy) was used to determined skinfolds to the nearest 1 mm. Each skinfold was measured 3 times, and the arithmetic means were used as the final value. Test–retest reliability for skinfold analysis was ICC = 0.96. Test–retest intra-observer reliability for fat adipose tissue thickness in our lab was ICC = 0.96, similar to previous findings [28,37].

Using the points previously marked, the skin fold was "pinched" between thumb and forefinger one centimeter above the measurement site, perpendicularly for the triceps and bicipital folds, and at a 45◦ angle to the longitudinal axis for the suprailiac and subscapular fold. All measurements were taken with the subject in an upright position and with the arms relaxed at the sides, while for the suprailiac point, the subject's right arm was placed over the operator's shoulder. Successively, body density was determined with Durnin– Womersley method [38], according to the methods Equation (1) was used to estimate body dansiy in male, whilst Equation (2) was used in female volounteers:

$$\text{Male BD} = 1.1765 - \text{(0.744} \times \log\_{10} \Sigma \text{ skínfolds)};\tag{1}$$

$$\text{Female BD} = 1.1567 - (0.0717 \times \log\_{10} \Sigma \text{ skinídols)}.\tag{2}$$

Ultimately, body density from Equation (1) and (2) was used in Equation (3) to estimate body fat percentage using the Siri formula [39]:

$$\text{FAT\%} = ((4.95/\text{BD}) - 4.5) \times 100\tag{3}$$

Ultrasound measurements were performed using Xario 100 ultrasound (TOSHIBA, Tustin, CA, US) with a surface probe set on MSK 1. The probe was equipped with a spirit level to maintain the same inclination in all readings. During acquisition, no pressure was ever exerted on the probe placed on the subject's skin, except for the natural one resulting

from the weight of the probe itself, by keeping the probe by the cable. For each detection point 3 snapshots were taken, following Muller's procedures:


Body composition, total body water (TBW), extracellular (ECW) and intracellular (ICW) water, body cell mass (BCM) and phase angle (PA) were measured via bioelectrical impendence analysis (Akern, Body Pro, Pontassieve, Italy). Subjects were asked to empty their bladder and rest for ~3–5 min in a supine position, while four electrodes were placed on their hands and feet to start the analysis. Test–retest reliability for body composition analysis using bioelectrical impendence was ICC = 0.99.

#### *2.3. Training Protocols*

Subjects were required to perform the prescribed exercise protocol 3 times per week for a total of 8 weeks. The training was supervised by a certified trainer whose task was to check adherence to the study and the correct execution of training protocols.

Training protocols were comparable for the type of exercise, intensity, and volume but differed for the order of execution. The SR protocol was an alternation between endurance strength exercise (MCT) in which, specifically, abdominal, triceps and aerobic exercise were performed in a circuit (as described in Table 2) while the other muscular area (back, shoulders, arms, and lower limbs) were trained through RT exercises at the end of the circuit. RT group completed first all the aerobic exercises, and in the second part, the resistance exercises. Aerobic exercise intensity was settled at 65% of max HR using Cooper formula, while resistance load was assessed based on the previous training schedule and during a preliminary familiarization session.

#### *2.4. Statistical Analysis*

Results are presented as mean ± SD. The sample size was calculated based on preliminary data from our laboratory, assuming within-subject variability of 25% and a fixed power of 0.8 and an alpha risk of 0.05 for the main variables (skinfolds). Initially, the analysis revealed that 9 subjects per group were needed to achieve the above parameters. However, only 7 participants were included in the final evaluation; we thus perform a post hoc analysis and the achieved power with the real sample size was 0.75. An independent *t*-test was performed on baseline characteristics to ensure no difference between groups. After checking for normal distribution via the Shapiro–Wilk W test, a two-way ANOVA for repeated measures was performed to compare the effect of training modalities through a "time x training" analysis. The post hoc Bonferroni test was used to identify specific intragroup differences when suitable. For each group, Cohen's d effect size was assessed by dividing the difference between mean values by the pooled SD. The *p*-value was set at 0.05. Data analysis was performed using GraphPad Prism software version 8.4.3 (GraphPad Software, San Diego, CA, USA).


**Table 2.** Training protocols.

SR, Spot Reduction group; RT, Resostance Training group.

#### **3. Results**

Body mass significantly decreased (F(1,12) = 14.304; *p* = 0.003) in the SR group (from 69.24 ± 6.90 kg to 67.74 ± 6.34 kg; *p* = 0.01, d = −0.32), but not in the RT group (from 75.93 ± 12.47 kg to 74.96 ± 12.08 kg, *p* > 0.05, d = −0.11). As a consequence, also the BMI was significantly reduced (F(1,12) = 14.605; *p* = 0.002) only in the SR group (from 23.78 <sup>±</sup> 2.11 kg/m<sup>2</sup> to 23.27 <sup>±</sup> 1.93 kg/m<sup>2</sup> , *p* = 0.01, d = −0.36) compare to RT group (from 24.67 <sup>±</sup> 3.54 kg/m<sup>2</sup> to 24.36 <sup>±</sup> 3.44 kg/m<sup>2</sup> , *p* > 0.05, d = −0.13) as shown in Figure 1.

**Figure 1.** (**A**) body mass and (**B**) body mass index (BMI). RT resistance training group; SR, spot reduction group. \* significantly different from pre-value (*p* < 0.05); \*\* significantly different from pre-value (*p* < 0.01).

> No differences in skinfolds were detected after RT protocol; while SR resulted particularly effective on suprailiac (−13.29%; *p* = 0.02, d = −0.56) and subscapularis (−7.59%, *p* = 0.04, d = −0.44) skinfold. In both sites, the two-way ANOVA analysis revealed a significant main Time effect (suprailiac: F(1,12) = 6.993; *p* = 0.01; subscapularis: F(1,12) = 5.822; *p* = 0.01). Furthermore, body fat percentage estimated with Siri equation presented a significant main effect of time (F(1,12) = 7.776; *p* = 0.02), with a significant decrease observed only in the SR group (*p* = 0.01, d = −0.21). Data on skinfolds results are shown in Table 3.


**Table 3.** Skinfold results of spot reduction (SR) and resistance training (RT) groups.

All values are means ± SD. \* significantly different from pre-value (*p* < 0.05); § significant time effect (*p* < 0.05).

We observed a significant main time effect on ultrasound measurements of adipose panicle for upper abdomen (F(1,12) = 6.888; *p* = 0.02), spinal erectors (F(1,12) = 10.209; *p* = 0.01) and front thigh (F(1,12) = 5.855; *p* = 0.03) (Table 4). Post hoc test revealed a significant reduction only in the SR group for upper abdomen (−18.89%, *p* = 0.05, d = −0.66), spinal erector site (−19.45%, *p* = 0.04, d = −0.55).

**Table 4.** Ultrasound results of spot reduction (SR) and resistance training (RT) groups.


All values are means ± SD. \* significantly different from pre-value (*p* < 0.05). § significant time effect (*p* < 0.05).

Results from the body composition analysis via BIA are shown in Table 5. A significant time effect (F(1,12) = 5.776; *p* = 0.03) was observed only in the Intracellular body water, where RT resulted in a significant reduction (−7.89%, *p* = 0.04, d = −0.76). No other detectable differences were found.

**Table 5.** Bioimpedance analysis (BIA) results of spot reduction (SR) and resistance training (RT) groups.


All values are means ± SD. \* significantly different from pre-value (*p* < 0.05); § significant time effect (*p* < 0.05).

#### **4. Discussion**

The study aimed to revisit the spot reduction training in the light of new methods to analyze SAT. We observed a significant general reduction of body mass and abdominal SAT, measured both with skinfold and ultrasound after 12 weeks of spot reduction training. Skinfold measurements also showed a reduction on the subscapularis site, while ultrasound revealed a decrease in the spinal erectors SAT.

Compared to ultrasounds, the skinfolds' technique measures SAT within a compressed double layer of skin. Skin thickness may vary substantially among body area, for example, is lower in the upper arm compared to the abdomen [28,36], reducing the accuracy between measurements. A recent analysis revealed that ultrasound might overcome the problem linked to the compressibility and viscoelasticity of adipose tissue and thus represents a better tool to estimate SAT [36,40]. Despite the limitation mentioned, we decided to include skinfolds measurements to be consistent with most of the studies that have analyzed spot reduction protocols. It is also worth mentioning that the ultrasound technique relies on the operator performing the measurements as much as skinfolds. Despite trying to comply with all the standard procedures to reduce variability during measurements, we observed relatively large standard deviations of up to 5–6 times the observed difference. Overall, the pre-post analysis revealed a relative medium effect size (as suggested by the observed Cohen's d > 0.5 for most measures) which may slightly weaken the validity of our findings.

Spot reduction is a training protocol aimed to reduce subcutaneous fat on a particular part of the body. The first protocols of spot reduction were created based on the assumption that the accumulation of fat in a specific body area is related to the activity of the adjacent muscles [20,22,41]. However, a better understanding of the mechanism of adipose accumulation/oxidation has revealed that this assumption might not be completely correct. Fatty acids taken from the diet are deposited in the adipose tissue based on hormonal and receptor action [42,43], like energy storage. During physical activity, muscle contraction demands energy; if the energy request is not solved with glycogen store, fats are mobilized from adipose tissue, released into the bloodstream, and carried to target cells to be oxidized. Lipolysis is mediated by hormonal fluxes (catecholamines, insulin and autocrine/paracrine factors), which reach adipose tissue passing through the circulatory network [42,43]. Circulating fatty acids can be provided from any body district, which does not necessarily must be involved with muscular effort; therefore, performing countless series only of a specific exercise may not be sufficient to promote lipolysis in that specific site. However, it was recently observed that lipolytic activity is associated with an increase of blood flow in the adipose tissue and, thus, to the oxygenation of the adipocyte, suggesting that "blood flow and lipolysis are generally higher in subcutaneous adipose tissue adjacent to contracting than adjacent to resting muscle irrespective of exercise intensity. Thus, specific exercises can induce "spot lipolysis" in adipose tissue" [19]. Based on these premises, the goal of spot reduction training should be to increase blood perfusion in the areas where it is most needed, which are where the adipose tissue is located; and sequentially promote fat oxidation. For this reason, the SR protocol we have employed in the present study was composed by a circuit training, in which the localized SAT mobilization was stimulated by target exercises (crunches for the abdomen and dumbbell overhead extension for the triceps), while fat oxidation was induced by the aerobic phases. Apart from our previous pilot study [34], this is the first attempt to use an MCT in an SR protocol.

We observed a significant reduction in the suprailiac skinfold and in the upper abdomen measure via ultrasound. These data are in accordance with others [20,21] and support the idea that spot reduction protocol can improve local lipolysis in the abdomen. We also observed a significant SAT reduction in the spinal erector site, which was adjacent to the subscapularis skinfold site, while we did not observe any direct effect on the triceps measurements. This was an unexpected result, as our hypothesis was that the specific triceps exercise included in the circuit training would have reduced the local SAT. We included tricipital exercise into our protocol because this is one of the areas in which subcutaneous adipose tissue can concentrate and also because triceps brachii can be exercised with several specific exercises. It is possible that, due to their inexperience, participants had involved more the shoulder and scapula-stabilizing muscles than the triceps brachialis during the execution of the dumbbell overhead extension exercise, reducing the effect on the specific site. It is also worth noting that, although not significantly, the front thigh SAT increased in the SR group. This result may be explained with a greater effort expressed from the participant during the first MCT part, which might have tired them out before

facing the second part of the workout. It is, therefore, possible that lower limbs were not successfully trained. Although this is only speculation, and these results raise interesting future questions on the SR approach.

Overall, we found that SR reduced total body mass, while any significant difference was obtained after a traditional resistance training protocol. However, body compositional analysis via BIA was unable to detect any significant changes in total fat mass or lean body mass. We observed a significant decrease in the intracellular water after RT, which normally indicates alteration of the number and size of muscle cells; however, this did not reflect on lean body mass value. Using the Siri equation to estimate body fat percentage, we found a significant decrease in the SR group compared to the RT group. However, the Siri formula is dependent on the precision of skin-fold measurements, and generally, the error of this method is approximately 5% [44].

The two protocols contained the same exercises and were comparable for duration and volume. This implicates that to reduce body mass, training intensity is a more important variable than the type of exercise. Training intensity could be manipulated in several ways: by increasing loads or oxygen consumption level, but also reducing rest intervals or altering movement velocity. In the initial MCT part of the spot reduction protocol, subjects did not rest between exercises, which increased the overall training metabolic demand. This is in contrast with other studies comparing the general and localized type of exercise training [21,45], which found similar effects on body composition. However, in the mentioned studies, the two protocols did not match the intensity of training modalities; for example, Noland and coll. Compared a general aerobic training with a localized calisthenic-type exercise [21]; while Schade concentrated one protocol only in the hip and abdominal areas and expanded the generalized training adding exercises on the upper and lower body [45]. Finally, we hypothesized that the alternation of endurance and strength exercises, or put another way, the insertion of a strength training exercise for specific muscles inside an endurance training might enhance the reduction of the fat tissue adjacent to the exercising muscles.

A limitation of the present study was the reduced number of participants. Due to the variability between subjects, it would be important to increase the sample size in future studies to determine the effectiveness of spot reduction in a larger population.

#### **5. Conclusions**

Spot reduction training, conducted in a mixed circuit-training format (triceps and abdomen inside an endurance training), seems to be efficient in promoting adipose tissue reduction in the subcutaneous abdominal region, but was not efficient on the triceps site.

**Author Contributions:** Conceptualization, A.P.; methodology, A.P., G.M. and T.M.; data curation: M.S., C.B., G.M. and M.F.; formal analysis: A.P., A.C., M.S., C.B., M.F., A.B., G.M. and T.M.; supervision: A.P., and T.M.; writing—original draft preparation, T.M.; writing—review and editing, A.P., and A.B.; All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Ethics Committee of the Department of Biomedical Sciences, University of Padova (HEC-DSB 05/17, 22 March 2017).

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** Original data are available upon request to the corresponding author.

**Acknowledgments:** We would like to thank all the subjects involved in the experiment for their patience and availability.

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

#### **References**


## *Article* **The Activation of Gluteal, Thigh, and Lower Back Muscles in Different Squat Variations Performed by Competitive Bodybuilders: Implications for Resistance Training**

**Giuseppe Coratella 1,\* , Gianpaolo Tornatore <sup>1</sup> , Francesca Caccavale <sup>1</sup> , Stefano Longo <sup>1</sup> , Fabio Esposito 1,2 and Emiliano Cè 1,2**


**Abstract:** The present study investigated the activation of gluteal, thigh, and lower back muscles in different squat variations. Ten male competitive bodybuilders perform back-squat at full (full-BS) or parallel (parallel-BS) depth, using large feet-stance (sumo-BS), and enhancing the feet external rotation (external-rotated-sumo-BS) and front-squat (FS) at 80% 1-RM. The normalized surface electromyographic root-mean-square (sEMG RMS) amplitude of *gluteus maximus, gluteus medius, rectus femoris, vastus lateralis, vastus medialis, adductor longus, longissimus,* and *iliocostalis* was recorded during both the ascending and descending phase of each exercise. During the descending phase, greater sEMG RMS amplitude of *gluteus maximus* and *gluteus medius* was found in FS vs. all other exercises (*p* < 0.05). Additionally, FS elicited *iliocostalis* more than all other exercises. During the ascending phase, both sumo-BS and external-rotated-sumo-BS showed greater *vastus lateralis* and *adductor longus* activation compared to all other exercises (*p* < 0.05). Moreover, *rectus femoris* activation was greater in FS compared to full-BS (*p* < 0.05). No between-exercise difference was found in *vastus medialis* and *longissimus* showed no between-exercise difference. FS needs more backward stabilization during the descending phase. Larger feet-stance increases thigh muscles activity, possibly because of their longer length. These findings show how bodybuilders uniquely recruit muscles when performing different squat variations.

**Keywords:** EMG; quadriceps; gluteus maximus; adductor longus; weight training; strength training; front squat; back squat; feet stance

#### **1. Introduction**

The squat is one of the most popular exercises used to elicit lower-limb strength, hypertrophy, and power [1–3]. It consists of a simultaneous flexion-extension of the hip, knee, and ankle joints, with the important role of the lower back muscles that stabilize the upper body, and consequently the whole movement [4,5]. Particularly, the gluteal, thigh, and lower back muscles are strongly activated during both the ascending and descending phase [6–9].

The squat can be performed in a multitude of variations, depending for example on the place where the barbell is located, the squatting depth, the feet stance, and/or feet rotation. Consequently, we may have back-squat (BS) or front-squat (FS) when the barbell is placed over the shoulders or in front of the clavicular line, respectively [10]. Alternatively, the squatting depth may lead to parallel or full squat, where the descending phase ends when the thighs are parallel to the ground or below this line, respectively [7]. Furthermore, the feet stance may be regular or wide, leading in the latter case to a so-called sumo-squat, where the direction of the feet can be parallel or rotated externally [11]. Obviously, all

**Citation:** Coratella, G.; Tornatore, G.; Caccavale, F.; Longo, S.; Esposito, F.; Cè, E. The Activation of Gluteal, Thigh, and Lower Back Muscles in Different Squat Variations Performed by Competitive Bodybuilders: Implications for Resistance Training. *Int. J. Environ. Res. Public Health* **2021**, *18*, 772. https://doi.org/10.3390/ ijerph18020772

Received: 3 December 2020 Accepted: 14 January 2021 Published: 18 January 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

these independent parameters can be miscellaneously used to create many combinations of squatting techniques and exercises. Among these, full-BS, parallel-BS, FS, sumo-BS, and external-rotated-sumo-BS are widely performed in practice.

A number of studies have investigated the difference in muscle activation when different squat variations are performed. Overall, squatting depth was shown to affect *gluteus maximus* activation with inconsistent results, with greater activation recorded in partial vs. full squat performed by young resistance-trained men [12], greater activation in full vs. partial performed by experienced lifters [7], or no difference when performed by resistance-trained women [13]. Additionally, quadriceps activation was overall greater in full vs. partial squat [14]. Interestingly, no difference in muscle activation was found comparing BS vs. FS performed with 70% 1-RM by healthy men [10], while larger stance specifically activates medial thigh muscles in experienced lifters [15], although no difference was found in gluteal muscles activation [11].

Bodybuilders have a unique capacity to perform exercises with a profound consistency of their technique, and were recently used to investigate the differences in muscle activation when bench press [16] or shoulder raise variations [17] are performed. Additionally, examining the muscle activation during both the concentric and the descending phase may help practitioners to characterize the strength and the hypertrophic stimuli, given both the short-term [18,19] and long-term unique responses following traditional or eccentric-based exercise training [2,20–22]. Therefore, the present study investigated the differences in the gluteal, thigh, and lower back muscles' activation in bodybuilders when varying the squatting technique. Particularly, the exercises selected were full-BS, parallel-BS, sumo-BS, external-rotated-sumo-BS, and FS, and the gluteal, thigh, and lower back muscles' activation were recorded during both the ascending and descending phase.

#### **2. Materials and Methods**

#### *2.1. Study Design*

The present investigation was designed as a cross-over, repeated-measures, withinsubject study. The participants were involved in seven different sessions. In the first five sessions, the 1-RM was measured in full-BS, parallel-BS, sumo-BS, external-rotation-sumo-BS, and FS in random order. In the sixth session, the participants were familiarized with the selected loads and the electrodes placement. In the seventh session, the muscles' maximum activation was first measured, i.e., the activation during a maximum voluntary contraction. Then, after a minimum of 30 min of passive recovery, the participant performed a nonexhausting set for each exercise performed in a random order, with an inter-set pause of 10 min. Each session was separated by at least three days, and the participants were instructed to avoid any further form of resistance training for the entire duration of the investigation.

#### *2.2. Participants*

The present investigation was advertised by the investigators during some regional and national competitions, and to be included in the study, the participants had to compete in regional competitions for a minimum of 5 years. Additionally, they had to be clinically healthy, without any reported history of upper-limb and lower back muscle injury and neurological or cardiovascular disease in the previous 12 months. To avoid possible confounding factors, the participants competed in the same weight category (Men's Classic Bodybuilding <80 kg, <1.70 m), according to the International Federation of Body Building Pro-League. The use of drugs or steroids was continuously monitored by a dedicated authority under its regulations, although we could have not checked for it. Thereafter, 10 male competitive bodybuilders (age 29.8 ± 3.0 years; body mass 77.9 ± 1.0 kg; stature 1.68 ± 0.01 m; training seniority 10.6 ± 1.8 years) were recruited for the present procedures. The participants were asked to abstain from alcohol, caffeine, or similar beverages in the 24 h preceding the test. After a full explanation of the aims of the study and the experimental procedures, the participants signed a written informed consent. They were

also free to withdraw at any time. The current design was approved by the Ethical Committee of the Università degli Studi di Milano (CE 27/17) and performed following the Declaration of Helsinki (1975) for studies involving human subjects.

#### *2.3. Maximum Voluntary Isometric Activation*

The maximum voluntary isometric activation of gluteus maximus, gluteus medius, rectus femoris, vastus lateralis, vastus medialis, erector spinae longissimus, and erector spinae iliocostal was assessed in random order. The electrodes were placed on the dominant limb, defined as the one preferred to kick a ball [2]. The participants were required to exert their maximum force against manual resistance. Each attempt lasted 5 s and three attempts were completed for each movement separated by 3 min of passive recovery [16,17]. The operators provided strong standardized verbal encouragements to push as hard as possible against the resistance exerted. The surface electromyography (sEMG) electrodes were placed following the SENIAM recommendations [23]. To check for appropriate electrodes placement previous procedures were followed [17]. For example, if the electrode shifted over the innervation zone during part of the movement, the EMG amplitude was underestimated. Therefore, to check for any consequence due to a possible shift of the surface electrode over the innervation zone, a Fast-Fourier Transform approach was used, as suggested in a previous investigation [24]. Briefly, the electrode placement on each muscle was checked during the warm-up phase of each exercise, analyzing the power spectrum profile of the sEMG signal recorded at the starting-, middle-, and endpoint of each exercise in all muscles. The correct electrode placement results in a typical bellyshaped power spectrum profile of the EMG signal, while noise, motion artifacts, power lines, and electrodes placed on the innervation zone or myotendinous junction generate a different power spectrum profile [24]. If the power spectrum did not match with the typical belly-shaped power spectrum profile in any of the temporal points, the electrodes were repositioned, and the procedures repeated so to have a clear EMG signal from all the muscles throughout the movement. The same experienced operator placed the electrodes and checked the power-spectrum profile. This approach was shown to provide very high reliability in sEMG data [16,17].

For *gluteus maximus*, the participants laid prone with the flexed knee and the electrode was placed below the line between the posterior-superior iliac spine and the trochanter major [13]. The participants were then asked to extend the hip against a manual resistance on the distal thigh [13]. For *gluteus medius*, the participant laid on a side and the electrodes were placed at 50% on the line from the crista iliaca to the trochanter. The participant was then asked to abduct the limb against manual resistance [23]. For *rectus femoris*, *vastus lateralis,* and *vastus medialis,* the participants sat on a table with the knees in slight flexion and the trunk slightly bent backward. The electrode were respectively placed at 50% and 2/3 on the line between the anterior-superior iliac spine and the lateral side of the patella and at 90% on the line between the anterior-superior iliac spine and the joint space in front of the anterior border of the medial ligament [23]. The participant were then asked to extend the knee against manual resistance [23]. The *adductor longus* belly was found midway between the origin at the pubic tubercle and the insertion at the medial linea aspera of the femur [25]. To ensure electrode placement, the test leg was passively abducted and the adductor longus muscle belly was palpated just distal to the muscle's tendon, traced from the pubic tubercle on the medial side of the leg, and the participant was then asked to actively adduct the leg against resistance [25]. For *erector spinae longissimus* and *iliocostalis*, the participant laid prone and the electrodes were respectively placed at 2-finger width lateral from the processus spinalis of L1 and 1-finger width medial from the line from the posterior-superior iliac spine to the lowest point of the lower rib, at the level of L2 [23]. The participant was then asked to extend the trunk against manual resistance [23].

The electrodes were equipped with a probe (probe mass: 8.5 g, BTS Inc., Milano, Italy) that permitted the detection and the transfer of the sEMG signal by wireless modality. sEMG signal was acquired at 1000 Hz, amplified (gain: 2000, impedance and the common rejection mode ratio of the equipment are >1015 Ω//0.2 pF and 60/10 Hz 92 dB, respectively), and driven to a wireless electromyographic system (FREEEMG 300, BTS Inc., Milano, Italy) that digitized (1000 Hz) and filtered (filter type: IV-order Butterworth filter; bandwidth: 10–500 Hz) the raw sEMG signals.

#### *2.4. 1-RM Protocol*

The squat 1-RM was assessed following previous procedures [26] using an Olympic bar (Vulcan Standard 20 kg, Vulcan Strength Training System, Charlotte, NC, USA). Briefly, after a standardized warm-up consisting of 30 weight-free squats, the 1-RM attempts started from 80% of the self-declared 1-RM and additional 5% or less was added until failure [27]. Each attempt was separated by at least 3 min of passive recovery. A standard time under tension (2 s for the ascending and descending phase, 0.5 for the isometric phase) was used and the participants had to lower the bar until the thighs were parallel to the ground. A metronome was used to pace the intended duty cycle and a camera was used to provide a feedback about the squatting technique and depth. Strong standardized encouragements were provided to the participants to maximally perform each trial.

#### *2.5. Exercises' Technique Description*

The selected exercises are shown in Figure 1, and described here from left to right, first the upper and then the lower row. In parallel-BS, the bar was placed over the shoulder and the participants were required to descent until the thighs were parallel to the ground, with a regular feet stance. In full-BS, the bar was placed over the shoulder and the participants were required to descent below the parallel thighs, with a regular feet stance. In FS, the bar was placed in front of the clavicular line and sternum, and the participants were required to descent until the thighs were parallel to the ground, with a regular feet stance. In sumo-BS, the bar was placed over the shoulder, and the participants were required to descent until the thighs were parallel to the ground, with a two-fold feet stance compared to the previous exercises. In external-rotated-sumo-BS, the participants received the same instructions as for sumo-BS, with the exception of the feet that were rotated externally. Six non-exhaustive repetitions were performed for each exercise.

**Figure 1.** The squat variations are shown. From the left to the right, in the upper row: full-back squat (BS), parallel-BS, and front squat (FS). In the lower row: sumo-BS and external-rotated-sumo-BS.

#### *2.6. Data Analysis*

The sEMG signals from both the peak value recorded during the maximum voluntary isometric activation and from the ascending and descending phases of each exercise were analyzed in time-domain, using a 25-ms mobile window for the computation of the root mean square (RMS). For the maximum voluntary isometric activation, the average of the RMS corresponding to the central 2 s was considered. During each exercise, the RMS was calculated and averaged over the 2 s of the ascending and descending phase. To identify the ascending and the descending phase, the sEMG was synchronized with an integrated camera (VixtaCam 30 Hz, BTS Inc., Milano, Italy) that provided the duration of each phase. Such a duration was used to mark the start and the end of each phase while analyzing the sEMG signal. The sEMG data were averaged excluding the first repetition of each set, to possibly have more consistent technique during the following repetitions. After, the sEMG RMS of each muscle during each exercise was normalized for its respective maximum voluntary isometric activation [16,17,27] and inserted into the data analysis.

#### *2.7. Statistical Analysis*

The statistical analysis was performed using a statistical software (SPSS 22.0, IBM, Armonk, NY, USA). The normality of data was checked using the Shapiro–Wilk test and all distributions were normal. Descriptive statistics are reported as mean (SD). The differences in the normalized EMG RMS were separately calculated for each exercise (5 levels) and phase (2 levels) using a two-way repeated-measures ANOVA. Multiple comparisons were adjusted using the Bonferroni's correction. Significance was set at *p* < 0.05. The differences are reported as mean with 95% of confidence interval (95%CI). Cohen's *d* effect size (ES) with 95% confidence interval (CI) was reported and interpreted according to the Hopkins' recommendations: 0.00–0.19: trivial; 0.20–0.59: small: 0.60–1.19: moderate; 1.20–1.99: large; ≥2.00: very large [28].

#### **3. Results**

The 1-RM were as follows: 215(28) kg for full-BS, 238(31) kg for parallel-BS, 255(36) kg for sumo-BS, 258(41) kg for external-rotated-sumo-BS, and 176(33) kg for FS.

The results for *gluteus maximus* are shown in Figure 2. No phase x exercise interaction (*p* = 0.197) was found for the normalized RMS of *gluteus maximus*. A main effect was found for factor phase (*p* < 0.001), but not exercise (*p* = 0.097). With the exception of FS (11.1%, −6.5% to 28.8%, *p* = 0.11; ES: 0.48, −0.43 to 1.48), greater normalized RMS was found during the ascending vs. descending phase in all exercises (16.0% to 41.1%, *p* < 0.05; ES: 1.55 to 3.99). During the ascending phase, no between-exercise difference was observed. During the descending phase, greater normalized RMS was found in FS vs. full-BS (46.6%, 8.4% to 84.8%, *p* = 0.017; ES: 2.94, 1.58 to 4.05), parallel-BS (40.9%, 14.0% to 67.9%, *p* = 0.005; ES: 2.58, 1.31 to 3.63), sumo-BS (40.1%, 7.1% to 73.0%, *p* = 0.017; ES: 2.38, 1.16 to 3.41), and external-rotated-sumo-BS (44.9%, 10.3% to 79.5%, *p* = 0.012; ES: 2.83, 1.49 to 3.92).

The results for *gluteus medius* are shown in Figure 2. No phase x exercise interaction (*p* = 0.157) was found for the normalized RMS of *gluteus medius*. A main effect was found for factor phase (*p* = 0.002), but not exercise (*p* = 0.125). Greater normalized RMS was found during the ascending phase in full-BS (12.0%, 9.1% to 15%, *p* < 0.001; ES: 2.92, 1.56 to 4.02) and external-rotated-sumo-BS (12.9%, 4.2% to 21.7%, *p* = 0.010; ES: 1.57, 0.51 to 2.49). During the ascending phase, no between-exercise difference was observed. During the descending phase, greater normalized RMS was found in FS vs. full-BS (19.0%, 4.9% to 33.1%, *p* = 0.010; ES: 2.16, 0.98 to 3.16), parallel-BS (13.6%, 0.4% to 26.8%, *p* = 0.016; ES: 1.35, 0.10 to 2.70), sumo-BS (17.5%, 5.4% to 29.6%, *p* = 0.006; ES: 1.90, 0.78 to 2.86), and external-rotated-sumo-BS (19.4%, 7.3% to 31.5%, *p* = 0.003; ES: 2.10, 0.93 to 3.08).

**Figure 2.** The surface electromyographic root-mean-square (sEMG) RMS amplitude of *gluteus maximus* and *gluteus medius* is shown. BS: back squat. \*: *p* < 0.05 ascending vs. descending phase. a: *p* < 0.05 vs. full-BS. b: *p* < 0.05 vs. parallel-BS. c: *p* < 0.05 vs. sumo-BS. d: *p* < 0.05 vs. external-rotated-sumo-BS.

The results for *rectus femoris* are shown in Figure 3. Phase x exercise interaction (*p* = 0.038) was found for the normalized RMS, and no main effect was found for factor phase (*p* = 0.417) and exercise (*p* = 0.231). Greater normalized RMS was found during the ascending compared to the descending phase in FS (30.1%, 7.8% to 52.3%, *p* = 0.015; ES: 1.35, 0.33 to 2.25). During the ascending phase FS showed greater normalized RMS than full-BS (24.0%, 1.9% to 46.0%, *p* = 0.032; ES: 1.21, 0.21 to 2.11). No between-exercise difference was found during the descending phase.

The results for *vastus lateralis* are shown in Figure 3. Phase x exercise interaction (*p* = 0.026) was found for the normalized RMS, and a main effect was found for factor phase (*p* = 0.011), but not exercise (*p* = 0.457). Compared to the descending phase, greater normalized RMS was found during the ascending phase in full-BS (22.1%, 6.1% to 38.1%, *p* = 0.013; ES: 1.60, 0.54 to 2.53), sumo-BS (28.8%, 8.4% to 49.1%, *p* = 0.012; ES: 1.64, 0.57 to 2.58), and external-rotated-sumo-BS (30.0%, 14.6% to 45.5%, *p* = 0.002; ES: 1.26, 0.25 to 2.16). During the ascending phase, both sumo-BS (19.8%, 0.8% to 38.8%, *p* = 0.040; ES: 0.97, 0.01 to 1.85) and external-rotated-sumo-BS (23.0%, 3.8% to 42.1%, *p* = 0.019; ES: 0.88, −0.07 to 1.76) had greater normalized RMS than FS. No between-exercise difference was found during the descending phase.

The results for *vastus medialis* are shown in Figure 3. No phase x exercise interaction (*p* = 0.133) was found for the normalized RMS, and a main effect was found for factor phase (*p* < 0.001), but not exercise (*p* = 0.102). Compared to the descending phase, greater normalized RMS was found during the ascending phase in full-BS (25.2%, 11.8% to 38.5%, *p* = 0.003; ES: 1.06, 0.08 to 1.94) and sumo-BS (25.9%, 10.6% to 41.2%, *p* = 0.005; ES: 1.27, 0.26 to 2.17). No between-exercise difference was observed during both the ascending and descending phase.

The results for *adductor longus* are shown in Figure 3. Phase x exercise interaction (*p* = 0.032) was found for the normalized RMS, and a main effect was found for factor phase (*p* < 0.001) and exercise (*p* = 0.021). Compared to the descending phase, greater normalized RMS was observed during the ascending phase in all exercises (ES: 2.25 to 5.39). During the ascending phase, greater normalized RMS was found in external-rotated-sumo-BS than full-BS (17.9%, 1.7% to 34.0%, *p* = 0.029; ES: 2.01, 0.86 to 2.98), parallel-BS (16.7%, 3.0% to 30.3%, *p* = 0.017; ES: 1.47, 0.43 to 2.39), and FS [26.9%, 7.3% to 46.5%, *p* = 0.009; ES: 2.64, 1.35 to 3.70). Greater normalized RMS was also found for sumo-BS than FS (19.7%, 5.6% to 33.9%, *p* = 0.008; ES: 2.15, 0.98 to 3.14).

**Figure 3.** The surface electromyographic root-mean-square (sEMG) RMS amplitude of *rectus femoris, vastus lateralis, vastus medialis* and *adductor longus* is shown. BS: back squat. \*: *p* < 0.05 ascending vs. descending phase. a: *p* < 0.05 vs. full-BS. b: *p* < 0.05 vs. parallel-BS. e: *p* < 0.05 vs. parallel front squat.

The results for *erector spinae longissimus* are shown in Figure 4. Phase x exercise interaction (*p* = 0.004) was found for the normalized RMS, and a main effect was found for factor phase (*p* = 0.015), but not exercise (*p* = 0.477). Compared to the descending phase, greater normalized RMS was found during the ascending phase in full-BS (39.6%, 16.0% to 63.1%, *p* = 0.005; ES: 1.76, 0.67 to 2.71). No between-exercise difference was observed during both ascending and descending phase.

**Figure 4.** The surface electromyographic root-mean-square (sEMG) RMS amplitude of *erector spinae longissimus* and *erector spinae iliocostalis* is shown. BS: back squat. \*: *p* < 0.05 ascending vs. descending phase. a: *p* < 0.05 vs. full-BS. b: *p* < 0.05 vs. parallel-BS. c: *p* < 0.05 vs. sumo-BS. d: *p* < 0.05 vs. external-rotated-sumo-BS.

The results for *erector spinae iliocostalis* are shown in Figure 4. Phase x exercise interaction (*p* = 0.020) was found for the normalized RMS, and a main effect was found for factor exercise (*p* = 0.040), but not phase (*p* = 0.431). Compared to the descending phase, the normalized RMS was greater during the ascending phase in full-BS (9.4%, 4.5% to 14.3%, *p* = 0.003; ES: 1.91, 0.78 to 2.87) and lower in FS (−10.4%, −18.6 to −2.3, *p* = 0.019; ES: −1.14, −2.03 to −0.15). During the descending phase, FS showed greater normalized RMS than full-BS (22.9%, 14.6% to 31.2%, *p* < 0.001; ES: 3.29, 1.84 to 4.46), parallel-BS (18.1%, 0.5% to 35.7%, *p* = 0.043; ES: 2.37, 1.14 to 3.39), sumo-BS (18.2%, 12.1% to 24.3%, *p* < 0.001; ES: 2.14, 0.97 to 3.13), and external-rotated-sumo-BS (19.2%, 7.1% to 31.2%, *p* = 0.004; ES: 2.43, 1.19 to 3.46). No between-exercise difference was found during the ascending phase.

#### **4. Discussion**

The current study examined how different squat variations influence the activation of the main muscles involved in these exercises. Both *gluteus maximus* and *gluteus medius* were more active during the descending phase of FS compared to all other exercises. *Rectus femoris* was more active during the ascending phase of FS compared to full-BS compared to all other exercises, while no between-exercise difference was visible for *vastus medialis*. *Vastus lateralis* and *adductor longus* were more active during the ascending phase of sumo-BS and external-rotation-sumo-BS compared to all other exercises. Lastly, while no betweenexercise difference was observed for *erector spinae longissimus*, *erector spinae iliocostalis* was more active during the descending phase of FS elicited compared to all other exercises. As such, varying the squatting technique seems to affect selectively the muscle activation.

#### *4.1. Gluteal Muscles*

FS showed very large increases in the *gluteus maximus* activation compared with all other exercises during the descending phase, with no between-exercise difference recorded during the ascending phase. A direct comparison with the literature is challenging, since few previous studies used similar design. When recording the sEMG RMS amplitude of *gluteus maximus* and distinguishing the ascending from the descending phase, no difference in FS vs. BS was found [29]. However, the load was maximal and performed by healthy men that limits the inference towards the present population. Additionally, we found that the *gluteus maximus* activation recorded here is much greater compared to the aforementioned study (e.g., 70% vs. 30% of the maximum activation during the descending phase of FS), which underlines the capacity of bodybuilders to increase muscle activation while training [30]. Moreover, no difference in gluteus maximus activation was found comparing FS, full-BS, and parallel-BS in trained women [13]. However, the authors did not specifically state which phase (ascending or descending or both) was examined, since it leads to argue that these findings are consistent with the no between-exercise difference recorded here during the ascending phase. Additionally, FS vs. BS was previously investigated, but no gluteal muscle was examined [10]. Lastly, the effect of stance does not seem to play a key role in *gluteus maximus* activation, which contrasts with the greater activation reported at greater stance [11,15]. Again, it is possible that the present bodybuilders population may have cancelled such a difference, since they were able to recruit the gluteus maximus more than just experienced lifters irrespectively of the stance. Similarly, *gluteus medius* resulted in greater activation during the descending phase of FS compared to all other exercise, with no between-exercise difference during the ascending phase. In a previous study, no difference in *gluteus medius* activation was observed when increasing the feet stance, confirming the present findings [11]. Taking all together, gluteal muscles seem to be particularly involved during the descending phase of FS. This may derive from the need to maintain an adequate trunk extension to avoid the barbell slipping forward (i.e., *gluteus maximus*), and to avoid a medial collapsing of the knees (i.e., *gluteus medius*), particularly when controlling the descending phase. As such, a frontal barbell placement seems to be a good option to increase the stimuli towards gluteal muscles while squatting.

#### *4.2. Thigh Muscles*

*Rectus femoris* showed greater activation in FS compared to full-BS during the ascending phase, with no other between-exercise differences. The lack of differences between full-BS and parallel-BS agrees with the no-difference found previously in powerlifters or weightlifters [31] or in healthy resistance-trained men [12]. Similarly with previous results, no difference in rectus femoris activation was reported when varying the squatting stance [11]. The reduced activation in full-BS vs. FS can be possibly explained by the greater *rectus femoris* length forced by the more vertical trunk in FS, which agrees with the greater work performed by the aforementioned gluteal muscles. Indeed, since *rectus femoris* acts as hip flexor, a more extended trunk corresponds to a longer length throughout the whole movement, thus increasing its activation as previously shown for deltoids [17] and triceps [32]. Both the sumo squats showed greater activation in *vastus lateralis* vs. FS. As suggested previously, larger stance makes hip and knee joint to exert more force to lift the load due to the non-favorable less vertical lever, thus increasing their recruitment [15]. Indeed, larger stance was shown to increase the *vastus lateralis* activation [33], rather than an external feet rotation alone, as previously reported [34]. Moreover, *vastus medialis* showed no between-exercise difference, with all exercises highly recruiting it. This may depend by the role of profound stabilizer of the patella across all movements, that enhances its activation when high loads have to be lifted. Lastly, larger stance and feet external rotation increased the *adductor longus* activation. This may depend on the need to stabilize the thigh position and keep the trajectory as vertical as possible in conjunction with the thigh external rotators, and on the longer muscle length at which adductor longus act at larger squat stance [11,15,35]. Taking together, larger feet stance may be used as an effective stimulus to increase the thigh muscles activity and could be implemented in the training practice accordingly.

#### *4.3. Lower Back Muscles*

*Erector spinae longissimus* showed no between-exercise difference, displaying a great activation across all exercises and during both the ascending and descending phase. In line with our results, no difference was found between BS and FS in experienced lifters [10], not even at different squatting depth in resistance trained men [12]. The study that investigated the effects of feet stance did not examine any lower back muscle [11,15,36], so a direct comparison cannot be made. However, given the high load and the consistent squatting technique, it is possible that the feet stance does not play a role in the *erector spinae longissimus* activation. Intriguingly, the activation of *erector spinae iliocostalis* was greater in FS compared with all other exercises during the descending phase. This may imply that FS needs additional balance control by mean of the trunk extensors to avoid any possible forward unbalancing. However, it should be noted that the net activation was much lesser than what observed in *longissimus,* meaning that the whole trunk and not only the lower back is involved in stabilizing the body. Lastly, both erectors' activation was greater during the ascending vs. descending phase in full-BS. This may be accounted for the very closed joint angles that could need an additional backward action to start the movement from a non-favorable body position. In practice, in conjunction with the greater stimulus for the gluteal muscles, FS might be recommended to enhance the work of the lower back muscles.

#### *4.4. Limitations*

A number of limitations should be acknowledged. First, there is no information of any rear thigh muscle (e.g., *biceps femoris*) that could have deepened the between-exercise differences. Second, similarly, the stabilizer role of any anterior trunk muscle (e.g., *rectus abdominis*) was not examined. Third, we selected a group of squat variations among several possible different combinations, that cannot be examined in a single study, so further research is needed to widen these aspects. Fourth, adding kinematic data would deepen the knowledge and should be considered in future research. Last, it is acknowledged that the present results are specific for the present populations, and different sport background may result in different muscle activation.

#### **5. Conclusions**

In conclusion, the present study showed different muscle activation depending on the squat variation in competitive bodybuilders. A front vs. back bar position led to greater gluteal and lower back muscles activation compared to all other exercises. Additionally, larger feet stance increases the thigh muscles activation, particularly *rectus femoris*, *vastus lateralis,* and *adductor longus*. Lastly, squatting depth does not seem to promote any specific difference in muscle activation, with the exception of the greater *rectus femoris* activation in FS vs. full-BS. These findings could be used in resistance training practice to vary the training stimuli when performing the squat exercises depending on the muscle group needed to be highlighted. Additionally, the specific differences observed during the ascending or descending phase may increase the specificity of the training-induced effects.

**Author Contributions:** Conceptualization, G.C., F.E., and E.C.; methodology, G.T., F.C., and S.L.; formal analysis, G.C., G.T., F.C., and S.L.; investigation, G.T.; data curation, G.T.; writing—original draft preparation, G.C. and F.C.; writing—review and editing, G.C., G.T., F.C., S.L. F.E., and E.C.; All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Ethics Committee of the Università degli Studi di Milano (protocol code CE 27/17, October 2017).

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Acknowledgments:** The Authors are grateful to the participants that volunteered for the present investigation.

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

#### **References**


International Journal of *Environmental Research and Public Health*

## *Article* **The Impact of Wrist Percooling on Physiological and Perceptual Responses during a Running Time Trial Performance in the Heat**

#### **Kelsey Denby, Ronald Caruso, Emily Schlicht and Stephen J. Ives \***

Department of Health and Human Physiological Sciences, Skidmore College, Saratoga Springs, NY 12866, USA; kdenby@skidmore.edu (K.D.); rcaruso@skidmore.edu (R.C.); eschlich@skidmore.edu (E.S.)

**\*** Correspondence: sives@skidmore.edu; Tel.: +1-518-580-8366

Received: 23 September 2020; Accepted: 15 October 2020; Published: 17 October 2020

**Abstract:** Environmental heat stress poses significant physiological challenge and impairs exercise performance. We investigated the impact of wrist percooling on running performance and physiological and perceptual responses in the heat. In a counterbalanced design, 13 trained males (33 ± 9 years, 15 ± 7% body fat, and maximal oxygen consumption, VO2max 59 ± 5 mL/kg/min) completed three 10 km running time trials (27 ◦C, 60% relative humidity) while wearing two cooling bands: (1) both bands were off (off/off), (2) one band on (off/on), (3) both bands on (on/on). Heart rate (HR), HR variability (HRV), mean arterial pressure (MAP), core temperature (TCO), thermal sensation (TS), and fatigue (VAS) were recorded at baseline and recovery, while running speed (RS) and rating of perceived exertion (RPE) were collected during the 10 km. Wrist cooling had no effect (*p* > 0.05) at rest, except modestly increased HR (3–5 ∆beats/min, *p* < 0.05). Wrist percooling increased (*p* < 0.05) RS (0.25 <sup>∆</sup>mi/h) and HR (5 <sup>∆</sup>beats/min), but not TCO (<sup>∆</sup> 0.3 ◦C), RPE, or TS. Given incomplete trials, the distance achieved at 16 min was not different between conditions (off/off 1.96 ± 0.16 vs. off/on 1.98 ± 0.19 vs. on/on 1.99 ± 0.24 miles, *p* = 0.490). During recovery HRV, MAP, or fatigue were unaffected (*p* > 0.05). We demonstrate that wrist percooling elicited a faster running speed, though this coincides with increased HR; although, interestingly, sensations of effort and thermal comfort were unaffected, despite the faster speed and higher HR.

**Keywords:** exercise; cooling; recovery; fatigue; thermal; environment; endurance

#### **1. Introduction**

Environmental stress, specifically heat stress, increases demand placed on the cardiovascular system [1,2]. Exercise also induces stress on the cardiovascular system, and the combination of heat stress with exercise can lead to a physiological challenge where demands for blood flow begin to challenge the maximal output of the heart, eventually leading to fatigue, exhaustion, and/or a decline in performance [1–7].

Accordingly, researchers have been developing strategies to prevent heat stress associated declines in exercise performance. One such approach has been the use of precooling, or reducing body temperature prior to exercise in the heat [4,8,9]. A review suggested that precooling via cold water immersion likely benefits performance, where ingestion of crushed ice/water ice slurry does not likely benefit performance [4]. Although the benefits of precooling, such as with cold water immersion, are not to be ignored, the issue of practicality raises concern over feasibility of implementation, and thus alternative methods ought to be explored. Strategies of attempting to cool during exercise, termed percooling, have demonstrated a positive effect on exercise performance, on par with precooling [8,9], though studies of percooling are far less abundant.

Recently, a company has developed a wearable, active cooling method (dhamaSPORTtm, DhamaUSA, Scotts Valley, CA, USA) that is light weight (115 g, 6 cm wide) and can be worn on the wrist during activity while posing minimal disruption or burden to the athlete (e.g., ice vest). While we have demonstrated that this wrist cooling device improved physiological recovery and reduced fatigue from an occupationally relevant model of exercise-induced heat stress [10], it has yet to be determined whether wrist percooling is capable of improving endurance performance in the heat, and if this might impact post-exercise recovery. Aside from the obvious potential to provide cooling, mitigating exercise-induced elevations in core temperature, surface percooling might activate the transient receptor potential melastatin 8 (TRPM8) "cold receptor", which might alter thermal sensation and/or exercise performance [11]. Further, recent work by Phillips et al. [7] suggests that cooling can modulate prefrontal cortex activation, perceptions of muscle fatigue or effort, and partially mitigate declines in local muscle performance. However, the impact of wrist percooling on perceptual responses during exercise in hypothermia is unknown. DhamaUSA, Scotts Valley, CA, USA) that is light weight (115 g, 6 cm wide) and can be worn on the wrist during activity while posing minimal disruption or burden to the athlete (e.g., ice vest). While we have demonstrated that this wrist cooling device improved physiological recovery and reduced fatigue from an occupationally relevant model of exercise-induced heat stress [10], it has yet to be determined whether wrist percooling is capable of improving endurance performance in the heat, and if this might impact post-exercise recovery. Aside from the obvious potential to provide cooling, mitigating exercise-induced elevations in core temperature, surface percooling might activate the transient receptor potential melastatin 8 (TRPM8) "cold receptor", which might alter thermal sensation and/or exercise performance [11]. Further, recent work by Phillips et al. [7] suggests that cooling can modulate prefrontal cortex activation, perceptions of muscle fatigue or effort, and partially mitigate declines in local muscle performance. However, the impact of wrist percooling on perceptual responses during exercise in hypothermia is unknown. Therefore, the purpose of this study was to investigate whether the wrist cooling improves

Recently, a company has developed a wearable, active cooling method (dhamaSPORTtm,

Therefore, the purpose of this study was to investigate whether the wrist cooling improves exercise performance in the heat or lessen physiological strain, and if this effect is "dose-dependent." We hypothesized that wrist percooling would reduce perceptions of effort and thermal stress, reduce heart rate, and/or improve performance on a 10 km running time trial, and these effects would be greater with the active cooling of both wrists. Second, use of the cooling bands would improve recovery as assessed by heart rate, heart rate variability, core temperature, and reduce fatigue and thermal sensations, all of which would also be greater with the active cooling of both wrists. exercise performance in the heat or lessen physiological strain, and if this effect is "dose-dependent." We hypothesized that wrist percooling would reduce perceptions of effort and thermal stress, reduce heart rate, and/or improve performance on a 10 km running time trial, and these effects would be greater with the active cooling of both wrists. Second, use of the cooling bands would improve recovery as assessed by heart rate, heart rate variability, core temperature, and reduce fatigue and thermal sensations, all of which would also be greater with the active cooling of both wrists.

#### **2. Materials and Methods 2. Materials and Methods**

#### *2.1. Subjects 2.1. Subjects*

Fourteen exercise-trained healthy male volunteers between the ages of 18 and 54 years were recruited for this study (Figure 1). To participate in this study, all participants must have been regularly exercise training for more than one hour at least three times a week for the past four months, have a maximal oxygen consumption (VO2max) of >45 mL/kg/min, and >1 year of experience in competing in running events (e.g., 5 km, 10 km, half-, full-, ultra-marathon, half-, full-ironman, etc.). Participants were screened, via health history, and those with cardiovascular, pulmonary, musculoskeletal, or metabolic disease, those taking regular medication, or presenting contraindications to the ingestible telemetry pill (*n* = 1) were excluded. Methodologically, women were excluded to avoid the long periods of time that would be necessary (up to 3 months) to ensure adequate recovery with parallel desire for testing to occur in a singular phase of the menstrual cycle, thus reducing the impact of hormonal fluctuations. All participants provided written informed consent prior to any testing. The protocol was approved by the Institutional Review Board of Skidmore College (IRB # 1612-568) and was conducted in accordance with the most recent revisions to the Declaration of Helsinki. Fourteen exercise-trained healthy male volunteers between the ages of 18 and 54 years were recruited for this study (Figure 1). To participate in this study, all participants must have been regularly exercise training for more than one hour at least three times a week for the past four months, have a maximal oxygen consumption (VO2max) of >45 mL/kg/min, and >1 year of experience in competing in running events (e.g., 5 km, 10 km, half-, full-, ultra-marathon, half-, full-ironman, etc.). Participants were screened, via health history, and those with cardiovascular, pulmonary, musculoskeletal, or metabolic disease, those taking regular medication, or presenting contraindications to the ingestible telemetry pill (*n* = 1) were excluded. Methodologically, women were excluded to avoid the long periods of time that would be necessary (up to 3 months) to ensure adequate recovery with parallel desire for testing to occur in a singular phase of the menstrual cycle, thus reducing the impact of hormonal fluctuations. All participants provided written informed consent prior to any testing. The protocol was approved by the Institutional Review Board of Skidmore College (IRB # 1612-568) and was conducted in accordance with the most recent revisions to the Declaration of Helsinki.

**Figure 1.** Experimental overview: RMR, resting metabolic rate; IET, incremental exercise test. **Figure 1.** Experimental overview: RMR, resting metabolic rate; IET, incremental exercise test.

#### *2.2. Study Overview*

The current study was conducted in a single blind, counterbalanced, crossover design to investigate the potential impact of wrist cooling on performance in, and recovery from, exercise in the heat (Figure 1). As the number of participants did not equal or equally multiply by the number of possible sequences of three trials, we used a Latin square approach to counterbalance. All testing was conducted in the Environmental Physiology Laboratory at Skidmore College. For each visit, participants were asked to avoid strenuous exercise for 24 h prior and alcohol/caffeine use 12 h prior to each study visit. Participants were instructed to maintain a similar diet and sleep regimen throughout the duration of the study. Participants were asked to wear shorts and t-shirt and to dress similarly across trials. Finally, participants were instructed to arrive each day fueled and hydrated as if preparing for a race, which included drinking the proper amount of fluids prior to each experimental visit (e.g., ~500 mL 2–3 h prior and 250 mL within 15 min of the visit). All participants reported to the laboratory on four separate occasions: a screening day and three experimental trials. While wearing two cooling bands, the three trials were conducted as follows: (1) both bands were off (off/off), (2) one band on (off/on), (3) both bands on (on/on). In the off/on condition, the right wrist was always activated. All experimental trials for a subject were completed at the same time of day to reduce impact of diurnal variation. In a thermoneutral (21 ± 1 ◦C, 29 ± 12% relative humidity) and normobaric (~750 mmHg) environment, the first screening visit assessed participant characteristics, which included anthropometrics (height, weight), body composition using air displacement plethysmography (BodPod, Cosmed, Chicago, IL, USA), and aerobic fitness via graded exercise testing on a treadmill (PPS Med, Woodway, Waukesha, WI, USA). A running protocol (modified McConnell) [12,13] was used to determine maximal oxygen consumption (VO2max) using open circuit spirometry and gas analysis (TrueOne 2400, Parvomedics, Sandy, UT, USA) [14]. Prior to each experimental trial, participants were given an FDA approved core temperature telemetry pill (HQ Inc, Palmetto, FL, USA), which was taken 8–12 h prior to the study visit [15,16].

#### *2.3. Procedures*

Upon arrival, a urine sample was collected and hydration status was confirmed via urine specific gravity (USG < 1.020) as described previously [17]. If USG was >1.020, participants were given 500 mL of water and USG was retested thereafter (though as the participants were familiar with race preparations, this only occurred once out of 39 total visits). Participants were then instrumented with a heart rate monitor (H7, Polar USA, Lake Success, NY, USA), and the presence of the core temperature telemetry pill was confirmed (CorTemp Recorder, HQ Inc, Palmetto, FL, USA). Participants were then seated and were outfitted with two wrist cooling bands (Dhama Sport Pro, Dhama USA, Scott's Valley, USA) (Figure 2). In the "on" condition, the bands were activated and set to the coolest setting (7.2 ◦C). While we attempted to avoid investigator cues and reduce possible anticipatory responses by single blinding and not making the participants' aware of which condition they were receiving, when the band was active, participants' were able to detect the cooling, but when the band was off (off/off condition) they were unsure. The device elicits cooling through a one-inch square ceramic cooling plate placed over the anterior vascular portion of the wrist, which dissipates heat via Peltier effect over a larger heat sink area on the exterior portion of the device. The heat transfer rate for this device ranges from 0.2 to 200 watts, with typical values of 0.5–50 watts, depending upon ambient conditions. After 10 min of quiet rest, a one minute [18], breathing frequency paced [19], recording of heart rate (HR) and R-R intervals were obtained via HR monitor, sent to a mobile device (IPad Pro, Apple, Cupertino, CA, USA) via Bluetooth™ and analyzed by a mobile device application (Elite HRV, Gloucester, MA, USA). The Elite HRV application performs artifact correction and has been shown to be valid [20] and has been used in previous studies [10,20–22]. Specifically, along with mean HR, R-R intervals were analyzed for the standard deviation of R-R intervals, SDNN; root mean square of successive differences, RMSSD; and the log transformed RMSSD, LnRMSSD. HRV was measured to assess potential impacts of wrist cooling on recovery as it is an increasingly recognized method to assess

or monitor athlete acute and chronic physiological response to training, or recovery and readiness to train [22–26]. After HR and HRV were obtained, to further characterize potential impacts of wrist cooling on recovery, blood pressure (BP) was measured via oscillometric cuff method (Mobilograph, GmbH, Stolberg, Germany) [27–29], after which thermal sensation/comfort via thermal sensation (TS) scale (0 "unbearably cold" to 8 "unbearably hot"), and fatigue via a visual analog scale (0 "no fatigue" to 10 "severe fatigue") were recorded. HR, heart rate; MAP, mean arterial pressure; SBP, systolic blood pressure; DBP, diastolic blood pressure; LnRMSSD, natural log transformed root mean squared of successive differences; A.U., arbitrary units; SDNN, standard deviation of normal R-R intervals; Msec, milliseconds; TS, thermal sensation; RPE, rating of perceived exertion; VAS, visual analog scale. \* main effect of time; † main effect of condition, # vs. off/off, *p* < 0.05. Note: all time effects were *p* < 0.05 pre vs. post. Means ± SD. *3.3. Ten km Time Trial Performance* 

Post 6.2 ± 1.8 6.4 ± 1.6 5.3 ± 2.5 \*

**Figure 2.** Raw performance data for all trials (*n* = 13): **(A)** average run time, including incomplete trials across condition; **(B)** number of incomplete trials across condition; and **(C)** average time to incompletion across conditions. Data are means ± Standard Error. **Figure 2.** Raw performance data for all trials (*n* = 13): (**A**) average run time, including incomplete trials across condition; (**B**) number of incomplete trials across condition; and (**C**) average time to incompletion across conditions. Data are means ± Standard Error.

Raw running time, including any incomplete trials, tended to decrease with the use of the bands (Figure 2A, off/on *p* = 0.14 and on/on *p* = 0.08 vs. off/off). However, due to participants reaching our institutionally mandated core temperature safety cut off, and importantly not of volitional means, this aforementioned time is tainted by a number of incomplete trials that tended to increase with the use of the bands (Figure 2B), and trended to an earlier incompletion time (Figure 2C). Chi squared analysis found that the proportion of incomplete trials did not significantly differ by condition (*p* = 0.16), nor were times to incompletion also not different between trials (pairwise comparisons, *p* = 0.51–0.81). Additionally, Kaplan–Meier survival curve analysis using a log rank test revealed no Participants were then allowed to warm up for a maximum of 5 min outside the chamber in the thermoneutral laboratory, typically followed by use of the restroom to void their bladder. Subsequently, participants entered the heated environmental chamber (26.7 ◦C, 60% relative humidity, heat index of 28 ◦C, "caution") and were instructed to complete the 10 km time trial (~6.2 miles, since the treadmill was in English units) as fast as possible at 0% grade. Thus, participants were able to see their speed and allowed direct control of the treadmill speed. Verbal encouragement was provided to all participants in a consistent manner between subjects and across trials. Participants were allowed to drink water, ad libitum, during all trials, but were asked to consume fluid in a similar volume and manner across

trials, matched for their first trial completed. During exercise, participants were asked to report their thermal sensation, rating perceived exertion using standardized visual scales every 5 min, while HR and core temperature (TCO) were monitored continuously and recorded every minute. Due to safety concerns, and institutional restrictions, in effort to avoid heat related injury, if core temperature reached two consecutive measures of 39.1 ◦C, or a single measure of 39.2 ◦C or higher, the trial was ended and the participant was immediately removed from the chamber and into a cool-down period in the thermoneutral laboratory. In such case, post-measures were obtained in an identical manner as if they had completed the trial.

Once the 10 km trial was completed, participants were escorted from the chamber and completed a 5 min cool-down, walking on a treadmill in the thermoneutral laboratory. HR and TCO were continuously monitored for safety reasons. Fifteen minutes after the cessation of the exercise, a post-exercise assessment of the baseline measures, except USG, was performed, namely: VAS, thermal sensation, core temperature, HR, HRV, and BP. The timing of the post-exercise measurements was maintained for all trials, including those that were ended due to core temperature reaching our institutional safety threshold. Once post-exercise measures were obtained, the wrist cooling units were turned off and removed. Participants reported back to the laboratory to complete the other two trials in a randomized counterbalanced order as described above. Visits were completed with a minimum of 48 h in between (average time between visits ~72 h).

#### *2.4. Statistical Analysis*

Data were analyzed using commercially available software (SPSS v26, IBM, Armonk, NY, USA) As the number of athletes who reached our institutional safety cutoff turned out to be larger than anticipated, additional analyses were conducted to compare the number of incomplete trials between wrist percooling conditions using a chi square test, and pairwise comparisons were used to determine if the time to incompletion differed between conditions. Further, a Kaplan–Meier survival curve analysis was conducted to compare the rate and time of incompletion using a log rank test. Again, due to athletes' core temperatures reaching our institutional safety cutoff, to allow direct time-matched pairwise comparisons between trials, all trial data were analyzed to the point at which we had complete data for all participants (16 min), as well as the final data point for each participant. The final data point was either the final data at incompletion due to reaching the temperature cutoff or the final value at the end of the 10 km time trial. Thus, data were analyzed and plotted to the longest common time, plus each athlete's final data point, and only for one athlete in one condition were 16 min and final the same. Further, to estimate effects of wrist percooling on 10 km time trial performance, if the trial was incomplete for reaching core temperature cutoff (see Figure 2), trial performance was estimated using average running speed for the trial.

Prior to analysis, any anomalous individual data points presenting as an outlier (>2 SD) were removed from the data set, and where appropriate, interpolated using a linear function. Accordingly, heart rate, core temperature, and speed were analyzed using a 3 (condition) by 17 (time, 16 min + final) repeated measures ANOVA. For RPE and TS, a 3 (condition) by 4 (time) repeated measures ANOVA was completed. To compare the pre- and post-measurements, a 3 (condition) by 2 (time) repeated measures ANOVA was completed for HR, core temperature, RMSSD, MAP, SDNN, LnRMSSD, diastolic blood pressure (DBP), systolic blood pressure (SBP), VAS, RPE, and TS Significance was established at *p* < 0.05. Data are presented as means ± standard deviation (SD), unless indicated otherwise.

#### **3. Results**

#### *3.1. Participant Characteristics*

The participant characteristics are presented in Table 1. Most participants (*n* = 10 of 13) were active triathletes, having competed in half or full distance Ironman events as well as running events, but all had road and/or trail running racing experience. Participants were fit, with an average VO<sup>2</sup> max of 59 mL/kg/min (range 50–71), particularly considering their average age of 33 years.


**Table 1.** Subject characteristics (*n* = 13).

#### *3.2. E*ff*ects of Wrist Cooling on Baseline Parameters*

No significant differences were found in resting core temperature, indicators of heart rate variability, blood pressure, thermal sensation, rating of perceived exertion, or in reported fatigue using a visual analog scale between conditions (Table 2). However, use of the bands tended to affect heart rate (*p* = 0.05), where HR was elevated by ~5 beats/min in the off/on condition (Table 2).

**Table 2.** Pre- and post-measurements for all three conditions (*n* = 13).


HR, heart rate; MAP, mean arterial pressure; SBP, systolic blood pressure; DBP, diastolic blood pressure; LnRMSSD, natural log transformed root mean squared of successive differences; A.U., arbitrary units; SDNN, standard deviation of normal R-R intervals; Msec, milliseconds; TS, thermal sensation; RPE, rating of perceived exertion; VAS, visual analog scale. \* main effect of time; † main effect of condition, # vs. off/off, *p* < 0.05. Note: all time effects were *p* < 0.05 pre vs. post. Means ± SD.

#### *3.3. Ten km Time Trial Performance*

Raw running time, including any incomplete trials, tended to decrease with the use of the bands (Figure 2A, off/on *p* = 0.14 and on/on *p* = 0.08 vs. off/off). However, due to participants reaching our 3).

institutionally mandated core temperature safety cut off, and importantly not of volitional means, this aforementioned time is tainted by a number of incomplete trials that tended to increase with the use of the bands (Figure 2B), and trended to an earlier incompletion time (Figure 2C). Chi squared analysis found that the proportion of incomplete trials did not significantly differ by condition (*p* = 0.16), nor were times to incompletion also not different between trials (pairwise comparisons, *p* = 0.51–0.81). Additionally, Kaplan–Meier survival curve analysis using a log rank test revealed no significant differences in survival distribution between wrist percooling conditions (*p* = 0.14, Figure 3). *Int. J. Environ. Res. Public Health* **2020**, *17*, x 7 of 17 significant differences in survival distribution between wrist percooling conditions (*p* = 0.14, Figure

**Figure 3.** Kaplan–Meier survival curve across time (seconds) between wrist cooling conditions during the 10 km time trial**. Figure 3.** Kaplan–Meier survival curve across time (seconds) between wrist cooling conditions during the 10 km time trial.

There was a significant interaction of wrist percooling condition and time on running speed, where RS tended to increase more over time with wrist percooling (Figure 4A), though there was no main effect for condition (*p* = 0.13). Focusing on the time to which all participants had completed, the distance achieved at 16 min was not different between conditions (off/off 1.96 ± 0.16 vs. off/on 1.98 ± 0.19 vs. on/on 1.99 ± 0.24 miles, *p* = 0.490, Figure 4B). Using both actual or projected 10 km times, there was no statistically significant effect of the bands (off/on *p* = 0.49 on/on *p* = 0.77 vs. off/off), despite a tendency for an approximate 30 s improvement in 10 km time for the off/on condition and a 10 s improvement in 10 km time in the on/on condition (off/off: 50:14.6, off/on: 49:45.9, on/on: 50:04.2 min:s). Using only those with complete trials, within a condition, the trend is less clear (off/off: 49:17.4, off/on: 50:02.3, on/on: 48:51.3 min:s). There was a significant interaction of wrist percooling condition and time on running speed, where RS tended to increase more over time with wrist percooling (Figure 4A), though there was no main effect for condition (*p* = 0.13). Focusing on the time to which all participants had completed, the distance achieved at 16 min was not different between conditions (off/off 1.96 ± 0.16 vs. off/on 1.98 ± 0.19 vs. on/on 1.99 ± 0.24 miles, *p* = 0.490, Figure 4B). Using both actual or projected 10 km times, there was no statistically significant effect of the bands (off/on *p* = 0.49 on/on *p* = 0.77 vs. off/off), despite a tendency for an approximate 30 s improvement in 10 km time for the off/on condition and a 10 s improvement in 10 km time in the on/on condition (off/off: 50:14.6, off/on: 49:45.9, on/on: 50:04.2 min:s). Using only those with complete trials, within a condition, the trend is less clear (off/off: 49:17.4, off/on: 50:02.3, on/on: 48:51.3 min:s).

*Int. J. Environ. Res. Public Health* **2020**, *17*, x 8 of 17

**Figure 4.** Running performance across wrist percooling condition. **(A**) Self-selected running speed during the 10 km time trial in the heat. Results of two-way ANOVA are presented (inset). Note: due to safety tolerance in core temperature trials were ended early and plotted to the shortest time, plus each athlete's final data point (with SE for time). **(B)** Distance to 16 minutes across wrist percooling condition (*n* = 13). This time was chosen as it was the longest point to which all participants completed at least 16 min for all 3 trials. Data are means ± Standard Error. **Figure 4.** Running performance across wrist percooling condition. (**A**) Self-selected running speed during the 10 km time trial in the heat. Results of two-way ANOVA are presented (inset). Note: due to safety tolerance in core temperature trials were ended early and plotted to the shortest time, plus each athlete's final data point (with SE for time). (**B**) Distance to 16 min across wrist percooling condition (*n* = 13). This time was chosen as it was the longest point to which all participants completed at least 16 min for all 3 trials. Data are means ± Standard Error.

#### *3.4. Physiological Response to 10 km Time Trial in the Heat 3.4. Physiological Response to 10 km Time Trial in the Heat*

A significant interaction between condition and time was found for heart rate (*p* = 0.00, Figure 5A). Expectedly, a main effect for time was found for heart rate throughout the trial (*p* = 0.00). No A significant interaction between condition and time was found for heart rate (*p* = 0.00, Figure 5A). Expectedly, a main effect for time was found for heart rate throughout the trial (*p* = 0.00). No significant differences were found for heart rate between conditions (*p* = 0.39).

significant differences were found for heart rate between conditions (*p* = 0.39). There was no significant interaction of condition by time for core temperature during the 10 km time trial (TT) (*p* = 0.15, Figure 5B). No main effect of condition was found during the 10 km time trial for core temperature (*p* = 0.88). Expectedly, however, there was a significant effect of time on core temperature during the trial (*p* < 0.001, Figure 5B). There was no significant interaction of condition by time for core temperature during the 10 km time trial (TT) (*p* = 0.15, Figure 5B). No main effect of condition was found during the 10 km time trial for core temperature (*p* = 0.88). Expectedly, however, there was a significant effect of time on core temperature during the trial (*p* < 0.001, Figure 5B).

#### *3.5. Perceptual Measures during 10 Km Time Trial in the Heat*

There was no significant interaction between condition and time for thermal sensation during the 10 km (*p* = 0.96, Figure 6A). A significant main effect was found for time, where the participants' TS increased over time (*p* = 0.00), though no significant differences were observed between conditions for TS (*p* = 0.47).

**Figure 5.** Physiological responses to 10 km time trial in the heat across wrist percooling condition. **A**) Heart rate and **B**) core temperature during 10 km time trial. Note: due to safety tolerance in core temperature, trials were ended early and plotted to the shortest time, plus each athlete's final data point. Data are means ± Standard Error (*n* = 13). **Figure 5.** Physiological responses to 10 km time trial in the heat across wrist percooling condition. (**A**) Heart rate and (**B**) core temperature during 10 km time trial. Note: due to safety tolerance in core temperature, trials were ended early and plotted to the shortest time, plus each athlete's final data point. Data are means ± Standard Error (*n* = 13).

*3.5. Perceptual Measures during 10 Km Time Trial in the Heat*  There was no significant interaction between condition and time for thermal sensation during No significant interaction of condition and time was found for RPE (*p* = 0.38, Figure 6B). Naturally, a main effect for time was found (*p* = 0.000). However, RPE did not significantly differ between conditions (*p* = 0.93).

#### the 10 km (*p* = 0.96, Figure 6A). A significant main effect was found for time, where the participants' TS increased over time (*p* = 0.00), though no significant differences were observed between conditions *3.6. Impact of Wrist Cooling on Recovery*

for TS (*p* = 0.47). No significant interaction of condition and time was found for RPE (*p* = 0.38, Figure 6B). Naturally, a main effect for time was found (*p* = 0.000). However, RPE did not significantly differ between conditions (*p* = 0.93). Pre- and post-measurements are shown in Table 2. No significant interaction (*p* = 0.58) was found between condition and time for core temperature, though core temperature was on average 0.2 to 0.7 ◦C cooler in recovery with use of the bands. Although core temperature approached significance, there was not a significant effect of time (*p* = 0.05); core temperature was not different from baseline and had recovered. However, there was no effect of condition on core temperature (*p* = 0.64).

There was no significant interaction between condition and time for heart rate (*p* = 0.36, Table 2). There was a significant main effect for time (*p* = 0.00) on heart rate with elevations post-exercise. Heart rate significantly differed between conditions, where heart rate tended to increase with the use of the bands (condition effect *p* = 0.03). To measure heart rate variability, the root mean square of the successive differences (RMSSD) was measured. No significant interaction effect for condition by time (*p* = 0.23) and no main effect of condition was found (*p* = 0.97). A significant main time effect was found for RMSSD (*p* = 0.000, Table 2) with a significantly reduced RMSSD post-exercise. In addition,

= 13).

*3.6. Impact of Wrist Cooling on Recovery* 

there was no significant interaction effect for condition by time (*p* = 0.43) and no difference between conditions (*p* = 0.96) for log transformed RMSSD (LnRMSSD). Expectedly, similar to RMSSD, there was a main time effect for LnRMSSD (*p* = 0.00, Table 2) with lower HR variability post-exercise. Lastly, SDNN had no significant condition effect (*p* = 0.41) or condition by time effect (*p* = 0.15). Corresponding to the other heart rate variability variables, there was a main time effect for SDNN (*p* = 0.00) with lower HR variability post-exercise. *Int. J. Environ. Res. Public Health* **2020**, *17*, x 10 of 17

**Figure 6.** Perceptual measures during 10 km TT**. (A)** Thermal sensation (TS) and **(B)** Rating of perceived exertion. Note: due to safety tolerance in core temperature, trials were ended early and plotted to the shortest time, plus each athlete's final data point. Data are means ± Standard Error, (*n* **Figure 6.** Perceptual measures during 10 km TT. (**A**) Thermal sensation (TS) and (**B**) Rating of perceived exertion. Note: due to safety tolerance in core temperature, trials were ended early and plotted to the shortest time, plus each athlete's final data point. Data are means ± Standard Error, (*n* = 13).

and had recovered. However, there was no effect of condition on core temperature (*p* = 0.64).

°C cooler in recovery with use of the bands. Although core temperature approached significance, there was not a significant effect of time (*p* = 0.05); core temperature was not different from baseline

No significant interaction for condition by time was found for mean arterial pressure, MAP (*p* = 0.90, Table 2). During the recovery process, a main effect of time was found for MAP (*p* = 0.01) with lower MAP in recovery, though no significant differences between conditions were found (*p* = 0.08). Systolic blood pressure, SBP, showed no significance for main time effect, condition by time, or condition (all *p* > 0.05, Table 2). Diastolic blood pressure, DBP, had no significant interaction of condition by time (*p* = 0.28) or effect of condition (*p* = 0.11). In contrast to SBP, DBP had a significant main time effect (*p* = 0.02) with a reduction in diastolic pressure post-exercise.

During recovery, there was no significant interaction between condition and time for RPE (*p* = 0.08) and TS (*p* = 0.72). There was a significant main effect of time for RPE and TS (both *p* = 0.000, Table 2). No significant differences between conditions were present for RPE (*p* = 0.43) and TS (*p* = 0.33). There was no significant interaction between condition and time for the fatigue visual analog scale (*p* = 0.47, Table 2). Expectedly, a main effect of time was present during the recovery process for VAS (*p* = 0.00), showing higher reported levels of fatigue, but no significant difference was found between conditions in the recovery process (*p* = 0.10).

#### **4. Discussion**

The intent of this study was to ascertain whether percooling via wrist cooling bands would improve 10 km running time trial performance in the heat, or lessen the physiological strain, and enhance recovery. The main finding of this study was that the use of the bands seemed to promote the participants to run at a faster speed over time. Thus, when using the bands, there was a corresponding increase in heart rate over time as a result of this increased speed and energy demand. On average, core temperature, thermal sensation, and rating of perceived exertion were not different when using the bands. Use of the bands did not appear to alter baseline or enhance physiological or perceptual indicators of recovery from the 10 km running bout. There was also no clear evidence that two bands were more advantageous than one, in terms of the physiological and perceptual responses to exercise, performance, or recovery. Thus, athletes considering use of wrist percooling, wearing one band is likely sufficient and possibly optimal. In conclusion, the cooling bands elicited a faster running speed over time; however, this comes at a physiological cost, but surprisingly not a perceptual one. Further work in the field or in unrestricted settings are needed to ultimately demonstrate efficacy of wrist percooling.

#### *4.1. Ten km Time Trial Performance*

In the present study, we observed that wrist percooling through the use of wrist cooling bands seemed to elicit a faster running speed in the participants over time (interaction effect, Figure 4). The faster self-selected running speed over time could be interpreted as an increase in performance because, all else held constant, would be expected to lead to a faster 10 km time trial. Focusing on distance covered at a common time, the distance covered by the athletes was not statistically higher at 16 min, a critical time point in our study. Using completed and estimated 10 km times, wrist percooling via use of the band's lead to times that were on average ~20 s faster, but were not significantly different from control. The faster running speed, as an indicator of performance, agrees with prior research using precooling, which also observed an increase in performance [9,30–34].

Previous reviews indicated that the magnitude of the effect of precooling likely depends upon the modality of precooling (e.g., water immersion, and depth, ice slurry, cooling garment, etc.) and how performance is assessed (i.e., time to exhaustion, and prescribed intensity, time trial, distance trial, etc.) [4,9]. Specifically, cold water immersion elicits an effect size (Cohens d) of 0.4 to 2 on performance vs. control, whereas precooling garments elicit an effect size of 0.1 to 0.5 (small to medium) [4], the latter more reflecting the magnitude effect in the current study (Cohens d effect size of 0.2, small effect, in off/off vs. off/on). Although the trend for a positive effect on actual or estimated 10 km time was not statistically significant, it should not be interpreted necessarily as useless. A 10–30 s effect on 10 km performance could have practical implications. To put this modest effect into perspective, when looking at the professional men's results of the 2016 AJC Peachtree road race, one of the nation's largest

10 km events, run in Atlanta, GA, during July (similar environmental conditions at race start to those in our study), a 20 s boost in performance could mean being on the podium or not. For the Bolder Boulder 10 km race, one of the largest 10 km races in the world, first place through sixth place in the pro division was separated by 20 s. Again, further field testing is needed to support this notion, though as climatic temperatures rise and running events are held in hot environments (e.g., Tokyo 2021 Olympic games), developing viable methods of supporting athlete's performance is increasingly paramount.

#### *4.2. Impact of Wrist Percooling on Physiological Responses to 10 km TT in the Heat*

In the current study, core temperature during the 10 km time trial was unaffected by percooling via wrist cooling bands (Figure 5). The present findings are in contrast to previous work using precooling, which demonstrated reduced core temperature particularly during initial stages of exercise [31,32,35,36]. For example, in a study by Lee and Haymes [32], 30 min of precooling reduced core temperature at rest and during exercise, though final core temperatures were not different between precooled and control. In agreement, when using a mix-method of ice bags and a cooling vest, Duffield et al. [31] saw a decrease in core temperature during warm up and the first sprinting periods; however, there was no significant difference in the final core temperature. Precooling is thought to reduce core temperature, creating a larger reservoir or tolerance for core temperature increases, postponing increases in temperature to the latter stage of exercise. In the current study, it was hypothesized that the use of the bands would blunt the rise of core temperature and lower it during the 10 km time trial, but this was not observed. However, given the running speed was increased over time, further challenging the cooling capacity of the unit due to the additional metabolic heat load, in already loaded condition, observing such an effect may not be possible. The cooling power of the device, maximally 200 watts, is simply not capable of ablating human heat production, estimated at over 1000 watts [37], but whether it might attenuate the rise in core temperature in individuals warrants further study.

Concordant to the increased running speed over time seen with wrist percooling, heart rate also was elevated over time during the 10 km time trial (Figure 5). These results somewhat support one previous study done by Duffield et al. [31], where they found no significant difference in heart rate during the exercise. However, previous research found with various models of precooling that HR was suppressed [30,32,33,35,38], at least during the initial stages of exercise, as final HR often was not different between conditions. However, some of these protocols used steady state exercise models [35], others incremental [33,38], or time to exhaustion [30,32], where speed or work load were matched. Thus, the higher heart rate over the trial with the use of the bands, while perhaps in disagreement with previous studies and not supporting the hypothesis, makes physiological sense in the context of the increased running speed.

#### *4.3. Impact of Wrist Percooling on Perceptual Measures during 10 km TT in the Heat*

The current study found no significant difference between conditions and no interaction effect for RPE (Figure 6). In support of the present data, Duffield et al. [31,36] also found no significant difference between the precooling and control conditions for RPE during the performance trial. However, other studies showed that RPE decreased with the use of precooling methods during the exercise performance [35,39]. However, in both of these studies walking/running speed was fixed.

Similar to RPE, in the present study, thermal sensation or perception during the 10 km time trial was not different between conditions and did not see an interaction effect with the use of the bands (Figure 6). Only one previous study supports the present data that showed no significant difference between conditions during the performance [31]. Other research has proven that with the use of cooling, the TS decreases during exercise [35,39,40].

Thus, while the present data does not support the hypothesis that the use of the cooling bands would have lowered RPE and TS, the present data are to be considered in the context of altered running speed over time. Recent meta-analysis suggests that topical or ingestion of menthol, a known agonist of the "cold receptor," TRPM8, can alter thermal sensation and/or exercise performance [11], perhaps independent of core temperature [41]. Relatedly, work by Phillips et al. [7] suggests that precooling might modulate the prefrontal cortex and/or its processing of afferent feedback regarding perceptions of effort, fatigue, or skin temperature, which might support greater exercise tolerance, and ultimately an attenuation of muscle fatigue. Thus, while it is tempting to speculate that RPE and TS might have been expected to increase in response to the increased running speed, and that the cooling bands mitigated the expected increase in perception of effort and thermal strain, further work is needed to confirm this hypothesis and explore the potential psychophysiological effects of wrist percooling.

#### *4.4. Impact of Wrist Percooling on Physiological Recovery*

In contrast to our initial hypothesis, core temperature did not significantly differ between conditions during recovery (Table 2). A prior meta-analytical review demonstrated that more aggressive cooling methods, such as whole body cooling likely help to recover performance [42], though few studies have focused on recovery of core temperature. Much of this work has been done in occupational models, such as firefighting, using multiple interventions, some invasive, to induce cooling and recovery of heart rate after firefighting [43–47]. Accordingly, our previous work using this wrist cooling device to induce cooling after an occupational model of exercise-induced heat stress via exercise in encapsulating clothing, revealed a significant positive impact of wrist cooling on recovery of temperature and heart rate [10]. However, in that study the exercise was necessarily more modest (walking) and shorter, thus the rise in core temperature was lower, all only increasing by 1 ◦C or less, potentially creating multiple differentials between the present and the aforementioned study. Interestingly, heart rate was found to be significantly impacted by wrist cooling at rest, and agrees with previous work that demonstrated skin cooling to 7 ◦C resulted in a modest 5 beat/min increase, likely the result of activating sensory afferent neurons [48]. This resting difference, we believe, contributed to a main effect of condition when exploring rest and recovery, as post-exercise HR values were not actually different between conditions. In support of no difference in heart rate during recovery between conditions, Edmonds et al. [40] using the wrist cooling device also found no significant difference in heart rate after high intensity physical activity. Thus, wrist cooling may be insufficient to hasten recovery of HR after high intensity activity in the heat.

#### *4.5. Impact of Wrist Percooling on Recovery of Perceptual Measurements*

During the recovery period, there was no significant difference in TS. However, TS during the recovery period was recorded highest in the off/off condition (4.6) and the lowest in the on/on condition (4.2). In support of this, Edmonds et al. [40] found a significant decrease during the recovery period, at the 10 min marker. The present data do show a lower value with the use of the bands; however, due to the time to incompletion decreasing with the bands, and the number of incomplete trials increasing, perceptual cues could be altered due to the decrease in performance. In recovery, RPE, or perhaps more appropriately the fatigue visual analog scale values after the exercise did not significantly differ, which does disagree with our prior work [10] using wrist cooling. Although for the reasons mentioned above, this may be expected. Future studies should explore the potential impact of wrist cooling on recovery in the field or athletic setting where immediate cooling applications may not be readily available and thus wrist cooling could perhaps be a bridge from or to more powerful cooling methods, such as cold-water immersion.

#### *4.6. Experimental Considerations*

The 10 km time trial took place in a controlled environment during the winter months and therefore the athletes were not acclimated to running in hot and humid environments. Institutional safety mandated cessation of exercise just above 39 ◦C (Figure 2), impairing our ability to determine whether performance would have truly been affected; indeed, previous work has suggested that high-level athletes are capable of tolerating such core temperatures or higher, perhaps even to 41.5 ◦C [49]. Anecdotally, none of the participants exhibited any heat-related illness signs or symptoms, suggestive

of a greater tolerance, and this may have underestimated potential performance effects as the athletes were unable to fully execute their individual race plan (e.g., negative splits or sprint at the end). Nonetheless due to this cutoff, in trials where the athlete reached this threshold we estimated or projected their performance. Future studies are needed to determine possible effects in a relatively unrestricted or field environment to observe fully the potential effects on performance. Although ingestible temperature telemetry pills have been demonstrated valid and able to track changes over time with heating or cooling [15,16], there was some variability in core temperature measurements with the telemetry pill and future studies might consider using more invasive measures such as esophageal or rectal thermistors. It was impossible to conduct the study in a fully blinded or placebo-controlled manner, though the research team sought to minimize eliciting any anticipatory responses, and participants wore both bands during all three trials. Measures of skin temperatures and/or localized thermal sensations would have enhanced the study, and future studies using this cooling method should include these measures, as well as pulmonary measures (e.g., VO2, ventilation, respiratory exchange ratio).

#### **5. Conclusions**

In the present study, wrist percooling during a 10 km TT in the heat resulted in a faster self-selected running speed and higher heart rates, though thermal sensation or perceptions of effort were unaffected. The increased running speed over time with wrist percooling might be practically meaningful, but further work is needed to determine the potential impacts of wrist cooling on performance, particularly in the field.

**Author Contributions:** Conceptualization, S.J.I.; methodology, K.D., R.C., E.S. and S.J.I.; formal analysis, K.D., R.C., E.S. and S.J.I.; investigation, K.D., R.C., E.S. and S.J.I.; resources, S.J.I.; data curation, K.D., R.C., E.S. and S.J.I.; writing—original draft preparation, K.D. and S.J.I.; writing—review and editing, K.D., R.C., E.S. and S.J.I.; visualization, K.D. and S.J.I.; supervision, S.J.I.; project administration, S.J.I.; funding acquisition, S.J.I. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was supported, in part, by DhamaUSA.

**Acknowledgments:** The authors would like to thank the Empire Endurance triathlon team and those who graciously volunteered for the study. We would also like to thank Michael Lopez in the Department of Mathematics and Statistics at Skidmore for his assistance. We would like to thank DhamaUSA for providing cooling units and financial support.

**Conflicts of Interest:** The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

#### **References**


**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

International Journal of *Environmental Research and Public Health*

## *Article* **Sarcopenia as a Mediator of the E**ff**ect of a Gerontogymnastics Program on Cardiorespiratory Fitness of Overweight and Obese Older Women: A Randomized Controlled Trial**

**Pablo Jorge Marcos-Pardo 1,2 , Noelia González-Gálvez 1,2,\* , Gemma María Gea-García 1,2 , Abraham López-Vivancos 1,2, Alejandro Espeso-García <sup>1</sup> and Rodrigo Gomes de Souza Vale 2,3**


Received: 4 August 2020; Accepted: 24 September 2020; Published: 27 September 2020

**Abstract:** The objectives were to analyze the effect of a gerontogymnastics program on functional ability and fitness on overweight and obese older woman and to understand if sarcopenia mediates its effect. This randomized controlled trial involved 216 overweight and obese women. The experimental group (EG) carried out 12 weeks of a gerontogymnastics program. The assessment was of gait speed, cardiorespiratory fitness, functional capacity, and muscle strength. EG showed significant improvements in almost every test. When the effect of training was adjusted by gait speed, the improvement of the 6 min walk test (MWT) for the trained group was no longer significant (*p* = 0.127). The improvement of the 6 MWT was significantly and positively associated with the 10 m test (β = −10.087). After including the 10-m test in the equations, the association between the 6MWT and carrying out the training program decreased but remained significant (β = −19.904). The mediation analysis showed a significant, direct and indirect effect with a significant Sobel test value (*z* = 6.606 ± 7.733; *p* = 0.000). These results indicate that a gerontogymnastics program improves functional capacity and fitness; and the effect of a gerontogymnastics program on CRF is mediated by sarcopenia in older women who are overweight and obese.

**Keywords:** Sarcopenia; gait speed; cardiorespiratory responses; walking; physical fitness; older people

#### **1. Introduction**

Sarcopenia is a progressive disease that involves the loss of muscle mass and strength [1]. It is associated with aging and causes a decrease in functional capacity, increasing the risk of falls and negatively affecting the quality of life, and in many cases may require hospitalization or rehabilitation [2,3]. Sarcopenia affects about 5 to 13% of individuals in their 60s and 70s, and 11 to 50% of octogenarians [4]. From the age of 40, there is a decline of muscle mass of approximately 8% per decade, which around the age of 70 can reach up to 15% per decade [5]. Besides, older women are more susceptible to present sarcopenia, as opposed to young women and men [6].

Age is associated with loss of mass, strength, and muscle power [7–9]. The loss of muscle mass and strength increases the risk of falls and potential fractures and contributes to the loss of functionality, impedes the older individual's ability to live independently [10], and results in a worse quality of life of a person [11–13]. The ability of the leg muscles to produce strength is a key factor in maintaining the older person's balance and walking speed [14]. According to the definition by the European Working Group on Sarcopenia in Older People (EWGSOP) [3,15], the diagnosis of sarcopenia can be made by assessing the low muscle mass, plus low muscle strength or low physical performance. The most commonly used parameters to measure muscle mass loss are dual energy x-ray absorptiometry (DEXA) and bioelectrical impedance analysis (BIA), to measure muscle strength it's handgrip strength, and to measure physical performance it's short physical performance (SPPB) and gait speed (GS) (Table 1).


**Table 1.** Sarcopenia: measurable variables and cut-off points [15].

The EWGSOP suggests the gait speed test as an easy and valid method for assessing physical performance [3,15,16]. Gait speed has been performed to evaluate various health-related factors such as physical functions, health status [17–19], and quality of life [20]. A well-known meta-analysis of 2888 older people set the minimum health threshold for gait speed at ≥ 0.8 m/s [21]. The promotion of physical exercise programs can prevent or even reverse the loss of functional capacity associated with sarcopenia if they emphasize improving the gait speed of the older population [22].

The increased fat mass associated with sarcopenia has been linked with a higher incidence of chronic diseases in older people [5,23,24], and developing sarcopenia contributes to the development of cardiovascular and metabolic diseases [25]. Being sedentary or not very physically active contributes to sarcopenia, as shown by research result, which have also linked age to the development of sarcopenia. The ICD-10 code for sarcopenia in 2016 was established by the World Health Organization to promote effective therapeutic strategies that include physical exercise to prevent loss of muscle mass and function with aging [26].

Recent research suggests effective intervention strategies to combat sarcopenia that include physical exercise, and more specifically, strength training [27–31]. Also, maximum oxygen consumption (VO2max) is a measurement of cardiorespiratory fitness (CRF), which can predict longevity in older adults [32]. People who have a higher risk of cardiovascular disease (CVD) and mortality are those who have a lower CRF [33]. For older individuals, regular aerobic exercise helps them to attain better VO<sup>2</sup> values [34]. With age, exercise that includes long-term aerobic exercise can help combat the effects of sarcopenia [35].

Being overweight and obese, coupled with a poor physical condition, are related to aging and are also associated with the risk of death from chronic diseases. Therefore, strategies are needed to encourage changes in body composition and physical condition [36,37]. Exercise programs, such as gerontogymnastics, which include resistance and aerobic training, are an optimal strategy for maintaining muscle mass and its protective effects against a variety of chronic diseases [38–42]. However, older adults with low functional capacity may not be able to develop resistance programs leading to improved CRF, due to their low fitness [43]. Besides, the improvement of CRF could also be influenced by improved strength [44].

Therefore, the objectives of the study were the following: a) to analyze the effect of a gerontogymnastics program with overweight and obese older women (≥65 years old) on functional ability and fitness, and b) to understand if sarcopenia mediates the effect of a gerontogymnastics program on cardiovascular fitness. We hypothesized that the older women who participate in the trained group will show improvements in all the tests, whereas the control group will not show changes, and that sarcopenia will mediate the effect of the program on cardiorespiratory capacity.

#### **2. Materials and Methods**

#### *2.1. Design*

The study was conducted from September to December 2019, with a total of 12 weeks of training. This is a randomized controlled trial study that investigates the effect of gerontogymnastics program with overweight and obese older women on functional ability and physical fitness. This study trial followed the Consolidated Standards of Reporting Trials (CONSORT) guidelines. Older women were informed of the study and signed an informed consent to participate, according to the process approved by the local ethics committee (CE111908) of the San Antonio Catholic University of Murcia (Spain) and in accordance with the Declaration of Helsinki. This study was conducted at a women's association and sport science laboratory from Region de Murcia (Spain).

#### *2.2. Participants*

Participants were recruited through advertisements in women's associations, senior centers, and presentations in the local community. Recruitment was done before registration for convenience and accessibility to the sample, but the intervention did not begin until the registration was completed. Inclusion criteria were: (a) at least 1 year not engaged in a structured exercise program, (b) having a body mass index between 25 and 29.9 (overweight) or between 30 and 34.9 (obese), (c) women aged between 65 and 90 years old, and (d) being physically independent. The exclusion criteria were: (a) having musculoskeletal injuries or limitations that could affect the person's health and physical performance; (b) being under a doctor's prescription for taking medication that could influence physical performance; (c) no regular attendance at the proposed sessions.

Sample size and power were established in connection with the 10-m walk test in a previous study [45]. An estimated error of 0.045 s and a significance level of α = 0.05 were utilized. A valid sample size for a confidence interval of 95% was 207.84. Based on previous research, a dropout rate of 10% was assumed; therefore, 230 participants were recruited.

Participants were divided to a trained group (TG) and a control group (CG). Thus, the experimental sample of the study consisted of 230 women, being electronically randomized [46] into the TG (115 subjects) and the CG (115 subjects). A researcher who was not involved in participant recruitment performed the randomization.

Two-hundred-and-sixteen older women aged between 65 and 88 years old volunteered (mean ± standard deviation [SD]; age = 68.26 ± 4.19 years, body mass = 71.11 ± 10.66 kg, height = 1.55 ± 0.07 m; BMI = 29.97 ± 3.86) and completed the study (TG = 114; CG = 102). The CONSORT 2010 flow diagram is shown in Figure 1. *Int. J. Environ. Res. Public Health* **2020**, *17*, x 4 of 16

**Figure 1.** Flow diagram of the sample. **Figure 1.** Flow diagram of the sample.

Anthropometric measurements were recorded. Weight (kg) was evaluated in light clothing without footwear to the nearest 0.1 kg by using an electronic scale, and height (cm) was measured using a stadiometer to the nearest millimeter (Seca 763 digital scale, Birmingham, UK). Body mass index (BMI) was calculated by dividing their weight in kilograms by their height in square meters (kg/m2). All anthropometric measurements were completed by experienced and well-trained persons (ISAK level 1 and 2 certificate). The same researchers performed all the measurements in a single Anthropometric measurements were recorded. Weight (kg) was evaluated in light clothing without footwear to the nearest 0.1 kg by using an electronic scale, and height (cm) was measured using a stadiometer to the nearest millimeter (Seca 763 digital scale, Birmingham, UK). Body mass index (BMI) was calculated by dividing their weight in kilograms by their height in square meters (kg/m<sup>2</sup> ). All anthropometric measurements were completed by experienced and well-trained persons (ISAK level 1 and 2 certificate). The same researchers performed all the measurements in a single session between the hours of 9:00 and 13:00 without warming up and allowing for a 5-min break between tests.

session between the hours of 9:00 and 13:00 without warming up and allowing for a 5-min break

#### between tests. *2.3. Procedure*

#### *2.3. Procedure*  2.3.1. Trained Group

2.3.1. Trained Group The gerontogymnastics program was implemented following the recommendations of the Otago Exercise Program (OEP) [47], as it is a renowned exercise program with widespread use at the international level that aims to improve strength and mobility to help in the prevention of falls. The gerontogymnastics program was planned principally to help prevent the risk of falls with exercise training that improves muscle strength, power, and balance in the lower extremities and The gerontogymnastics program was implemented following the recommendations of the Otago Exercise Program (OEP) [47], as it is a renowned exercise program with widespread use at the international level that aims to improve strength and mobility to help in the prevention of falls. The gerontogymnastics program was planned principally to help prevent the risk of falls with exercise training that improves muscle strength, power, and balance in the lower extremities and cardiorespiratory endurance. A professional Sports Physical Educator directed the training. The total training group was divided into subgroups of a maximum of 20 subjects for security

joint mobility and stretching of the main muscle groups involved; (b) 30 min of an exercise circuit (three sets of 15 repetitions in each exercise, with 2 min rest between sets). The exercise circuit

cardiorespiratory endurance. A professional Sports Physical Educator directed the training. The total

and the correct direction by the trainer. The participants in the TG trained for 1 h, three times a week for the 12 weeks of intervention (36 sessions). The training session consisted of: (a) 10 min of warm-up, consisting of joint mobility and stretching of the main muscle groups involved; (b) 30 min of an exercise circuit (three sets of 15 repetitions in each exercise, with 2 min rest between sets). The exercise circuit consisted of 12 exercises, of which, seven focused on strength: knee extension, squat, knee curl, leg press, elbow curl, chest press, and shoulder overhead press using the OMNI resistance exercise scale [48]; five focused on balance: walking on marked lines on the floor, walking on tiptoes, walking sideways, walking on heels, and walking from heel to toe; and (c) 20 min of cardiovascular exercises. The cardiovascular exercise consisted of walking at maximum speed without running to maintain a moderate to hard level of perception of exertion [49]. The training sessions were held on non-consecutive days to facilitate recovery.

#### 2.3.2. Control Group

The CG participants were asked to carry out their normal life and not to alter their habits during the study period, and they did not practice any physical activity or exercise program.

#### 2.3.3. Assessment of gait speed

Gait speed was assessed with a 10-m test. The time in seconds that the person took to walk 10 m was analyzed, with the person walking at their usual pace. This test has been widely used in large epidemiological studies, showing high concurrent and predictive validity [45,50–53]. The results of previous studies [45] indicate an excellent relationship between the 4-m test and the 10-m test (ICC = 0.959 and 0.976, respectively), with a good average between both tests (ICC = 0.867). The 10-m test proved to be somewhat better. This test is a valid method for predicting sarcopenia [54]. The reference value for gait speed in the 10-m test is 0.8 m/s [55]. In the present study, in order to obtain reliable measurements, two photocells were placed at the beginning and the end of a 10-m lane, and through a connection to a computer, recorded the time spent in carrying out the test (MuscleLab, Ergotest, Langesund, Norway). The older women were asked to stand at the starting line mark and walk at their usual pace at the sound signal. Two attempts were made and the average value between the two repetitions was recorded.

#### 2.3.4. Assessment of Cardiorespiratory Fitness Level

Aerobic endurance was assessed using the 6-min walk test (6 MWT). The 6 MWT has been shown to be a valid, reliable, objective, inexpensive, and easy test used to evaluate cardiorespiratory capacity [56–60]. It is a simple test to perform and is better related to the person's daily life activities than other tests [58,59]. It is used to measure an individual's sub-maximum aerobic capacity while walking for 6 min.

It is suggested that this test should be performed on a flat surface that allows walking for 20 to 30 m. The subject should be relaxed and wear comfortable clothing and shoes and the heart rate of each subject was recorded with a POLAR 400 heart rate monitor just before the start of the test and just after the end. The route was marked every 5 m and cones were placed at the turning area. The subject was walking at a pace appropriate to his/her condition, being able to stop or slow down if he/she is fatigued and resume as soon as possible. The trainer can motivate the subjects with phrases such as "You are doing well", and the total meters walked is recorded [61]. This test has good reliability (ranging from 0.95 to 0.97) [62].

#### 2.3.5. Assessment of Functional Capacity

The Latin American Group for Maturity (GDLAM) protocol is used to evaluate the functional capacity in older adults [27,63,64]. The battery consists of five tests: walking 10 m; rising from a sitting position; standing up from a prone position on the floor; getting up from a chair and moving around; and the putting on and taking off a T-shirt test. These tests were to calculate the GDLAM functionality index (GI) using a mathematical formula. The material needed for carrying out the tests consisted of a standard chair with a height of 48 cm from the seat to the floor, a digital chronometer, four cones, a sports mat, and a metal measuring tape. The magnitude of the statistical significance demonstrated high reliability (*r* = 0.9; *p* < 0.001) and validity [63].

#### 2.3.6. Assessment of Muscle Strength

Two tests from the "Senior Fitness Tests" (SFT) battery [59,65] were used to assess strength variables: extension flexion elbow test and lift chair 30 s test. The extension flexion elbow test measures the muscle strength of the upper extremity. The subject, while sitting on a chair, was asked to perform the maximum number of repetitions for 30 s with a dumbbell (2.3 kg for women). The lift chair 30 s test reflects lower body strength. The participant was asked to sit on a chair with his arms across his chest and perform the most sitting and standing repetitions for 30 s. Reliability and validity indicators for the standards ranged between 0.79 and 0.97 [66].

#### *2.4. Data Analysis*

The normality of the data was evaluated using the Kolmogorov–Smirnov test, and Mauchly's W-test was used to analyze the normality and the sphericity of the data. The inter- and intra-groups differences and the interaction between groups and time were analyzed with a two-way ANOVA with repeated measurements of one factor (time). Also, an ANCOVA (adjusted for gait speed) with repeated measurement of one factor (time) was used. To check intra-groups change, the post-hoc Bonferroni test and the Wilcoxon signed-rank test were used to evaluate the statistical significance of parametric and non-parametric variables, respectively. The Mann–Whitney test was used to check for inter-group differences for non-parametric variables. The partial eta-squared (η2p) for variance analysis was used to calculate the size effect, and this was defined as small: ES ≥ 0.10; moderate: ES ≥ 0.30, large: ES ≥ 1.2; or very large: ES ≥ 2.0, with an error of *p* ≤ 0.05 utilized [67].

To determine if the effect on the 6MWT test was mediated by the change in the 10-m test, the analysis of the mediation variables was performed using the Process macro for SPSS (SPSS Inc, Chicago, Illinois). A resample procedure of 10,000 bootstrap samples for non-parametric variables was utilized, [68] and the classical Baron and Kenny step regression method was used for parametric ones. [69]. In order to analyze the statistical significance of the mediation effect, the Sobel test was used [70]. If after the mediation, the independent variable was no longer associated with the dependent variable, it was considered complete mediation. However, if after the mediation, the independent variable was reduced but was still significant, it was considered partial mediation. The statistical analysis was performed using IBM SPSS Statistics (version 24.0), and an error of *p* ≤ 0.05 was set for the analysis.

#### **3. Results**

The characteristics of the participants are shown in Table 2. The TG showed significant improvements in the 10-m test (*p* < 0.000), the 6 MWT (*p* = 0.001), stand from siting test (*p* < 0.000), the rising from sitting test (*p* < 0.000), the rise from the floor test (*p* < 0.000), the t-shirt test (*p* < 0.000), the GDLAM index (*p* < 0.000), the extension and flexion elbow test (*p* < 0.000), and the lift chair 30 s test (*p* < 0.000). TG did not show changes in the stand-up and go test (*p* = 0.150) and showed an increase of BMI (*p* = 0.021) but with a very low effect size (ES = 0.03).

The CG experienced a significant decrease in the 10-m test (*p* < 0.000), 6 MWT (*p* = 0.011), rise from the floor test (*p* = 0.032), stand-up and go test (*p* < 0.000), and extension flexion elbow test (*p* < 0.000), and showed a significant improvement in the rise from the floor test (*p* = 0.032), although they did not show changes in the rest of the tests. Although both groups showed a significant improvement in the rise from the floor test, the effect size was small for the CG (ES = 0.12), whereas the effect size for the TG was large (ES = 0.72) (Table 3).


**Table 2.** Characteristics of the participants.

Legend: s = seconds; m = meters; rep = repetitions; M = Mean; SD = Standard Deviation; 6 MWT = 6 min walk test.

**Table 3.** Differences pre- to post-test (intra-groups) for functional and fitness test.


Legend: TG = trained group; CG = control group; M = Mean; SD = Standard Deviation; ES = Effect Size; s = seconds; m = meters; rep = repetitions.

Table 4 shows the differences between groups in the changes pre- and post-test. The results show a difference between groups for all the functional and fitness tests in favor of TG.


**Table 4.** Differences pre to post-test (intergroups) for functional and fitness test.

Legend: s = seconds; m = meters; rep = repetitions; M = Mean; SD = Standard Deviation; 6 MWT = 6 min walk test; ES = effect size.

When the effect of training was adjusted according to gait speed, the improvement of the 6 MWT for TG was no longer significant (TG = difference post-pre (M ± SD): −9.476 ± 6.178; *p* = 0.127; CI 95% (Mpost–Mpre): −21.653;2.702; CG = difference post-pre (M ± SD): 5.601 ± 6.633; *p* = 0.399; CI 95% (Mpost–Mpre): −7.473;18.675).

The improvements in the 6 MWT (β = −32.129) and 10-m test (β = 1.689) were significantly associated with carrying out the training program (TG). The improvement in the 6 MWT was significantly and positively associated with the 10 m test (β = −10.087). After including the 10 m test in the equations, the association between the 6MWT and carrying out the training program (TG) decreased, although it remained significant (β = −19.904). The mediation analysis showed significant, direct and indirect effects with a significant Sobel test value (*z* = 6.606 ± 7.733; *p* < 0.000). These results indicate that gait speed (10 m test) acts as a mediator on the effect of the exercise program on the 6 MWT (Figure 2).

**Figure 2.** Mediation of exercise intervention and 6MWT by 10-metres test. \*\* *p* ˂ 0.001; \* ˂ 0.05. **Figure 2.** Mediation of exercise intervention and 6MWT by 10-metres test. \*\* *p* < 0.001; \* < 0.05.

#### **4. Discussion 4. Discussion**

The first objective of this randomized controlled trial was to analyze the effect of a gerontogymnastics program for overweight and obese older women on functional ability and fitness. Significant improvements in functional capacity (10-m test, rise from sitting test, rise from the floor test, t-shirt test, and GDLAM index), CRF (6MWT) and muscle strength and endurance (extension and flexion elbow and lift chair 30 s test) were reported by the group that carried out the intervention program. The CG showed a significant decrease in the 10-m test, 6MWT, rise from the floor test, stand-up and go test, and extension flexion elbow test; and did not show changes in the rest of the tests. In connection with the stand-up and go test, the TG did not show any changes; however, the CG experiment showed a significant decrease. This could be interpreted as the intervention program preventing the physical decline due to age. Although both groups showed a significant improvement in the rise from the floor test, the effect size was small for the CG (ES = 0.12), whereas the effect size for the TG was large (ES = 0.72); and there was also an inter-groups difference that indicated that the The first objective of this randomized controlled trial was to analyze the effect of a gerontogymnastics program for overweight and obese older women on functional ability and fitness. Significant improvements in functional capacity (10-m test, rise from sitting test, rise from the floor test, t-shirt test, and GDLAM index), CRF (6MWT) and muscle strength and endurance (extension and flexion elbow and lift chair 30 s test) were reported by the group that carried out the intervention program. The CG showed a significant decrease in the 10-m test, 6MWT, rise from the floor test, stand-up and go test, and extension flexion elbow test; and did not show changes in the rest of the tests. In connection with the stand-up and go test, the TG did not show any changes; however, the CG experiment showed a significant decrease. This could be interpreted as the intervention program preventing the physical decline due to age. Although both groups showed a significant improvement in the rise from the floor test, the effect size was small for the CG (ES = 0.12), whereas the effect size for the TG was large (ES = 0.72); and there was also an inter-groups difference that indicated that the TG significantly improved more than the CG.

TG significantly improved more than the CG. Other studies that implemented a similar exercise program also reported improvements in functional capacity and fitness [42,47,71–73]. These studies implemented their programs from 8 to 18 weeks, with a frequency of three times per week and a session duration ranging from 50 min to 60 min. Related to this, our study included different sets of exercises for strength training, balance, and cardiovascular endurance. This exercise program is adapted to older women who are overweight Other studies that implemented a similar exercise program also reported improvements in functional capacity and fitness [42,47,71–73]. These studies implemented their programs from 8 to 18 weeks, with a frequency of three times per week and a session duration ranging from 50 min to 60 min. Related to this, our study included different sets of exercises for strength training, balance, and cardiovascular endurance. This exercise program is adapted to older women who are overweight and obese.

and obese. A 12-week, low-to-moderate-intensity at maximal fat oxidation intensity (FATmax; 37% –54% VO2max) exercise program for overweight and obese older women resulted in favorable changes in body composition and functional capacity in the exercise (training) group, compared with the outcomes of the control group [74]. Another study revealed that 12 weeks of elastic resistance training exerted positive effects on functional mobility outcomes of older women with sarcopenic obesity [75]. No prevalence of obesity, a higher level of physical activity, and baseline grip strength were associated with better mobility performance among the older population [76]. Physical activity mitigated the deleterious effects of the loss of functional capacity and muscle strength in obese A 12-week, low-to-moderate-intensity at maximal fat oxidation intensity (FATmax; 37–54% VO2max) exercise program for overweight and obese older women resulted in favorable changes in body composition and functional capacity in the exercise (training) group, compared with the outcomes of the control group [74]. Another study revealed that 12 weeks of elastic resistance training exerted positive effects on functional mobility outcomes of older women with sarcopenic obesity [75]. No prevalence of obesity, a higher level of physical activity, and baseline grip strength were associated with better mobility performance among the older population [76]. Physical activity mitigated the deleterious effects of the loss of functional capacity and muscle strength in obese individuals,

individuals, highlighting its importance in the creation of strategies for the preservation of physical

highlighting its importance in the creation of strategies for the preservation of physical function with age [77]. These results support the evidence that a 12-week gerontogymnastics program that included endurance and strength training exercises improves functional capacity, CRF, and strength and endurance of musculature of overweight and obese older women; and could thus delay the harmful effects of aging.

The second objective of this study was to understand if sarcopenia mediated the effect of a gerontogymnastics program on cardiovascular fitness. The major finding of our study was that an improvement in CRF was associated with an improvement in gait speed, in consonance with the decrease in sarcopenia. Our results are in agreement with a previous study, showing a connection between CRF and gait speed and sarcopenia [78]. In our study, sarcopenia acted as a partial mediator on the association between carrying out a gerontogymnastics program and improved CRF. To the best of our knowledge, this is the first randomized controlled trial with an analysis of the mediation that assesses how sarcopenia influences the effect of an exercise program on CRF.

A recent study [79] assessed 527 women aged 75 years and older (79.7 ± 3.5) in a cross-sectional study. The objective of this study was to investigate if the connection between physical activity and gait speed was mediated by strength and weight. These authors reported that the association between physical activity and gait speed was partially mediated by the absolute and relative strength of the lower limbs and that muscle mass partially mediated the relationship between physical activity and muscle strength.

On the other hand, it has been demonstrated that there is a connection between walking balance and strength [80] and that sarcopenia influences walking balance [81]. A study with older adults with mild to moderate frailty improved their CRF but at a modest level [82]. This suggests that it will be necessary to increase leg strength to further increase walking speed, in order to improve CRF. In this sense, a study was performed to determine the mechanisms responsible for the effect of exercise training on CRF in older adults, utilizing a strength training program before an endurance training program, with a sample of 22 older adults, to improve their functional capacity [44].

In agreement with another study [43], older adults with declined functionality were not able to participate in endurance training until they improved their neuromuscular capacity. Therefore, endurance training for older women should be performed with previous strength and resistance training to achieve the highest CRF adaptations.

It has also been reported that gait speed, muscle mass, and sarcopenia are strongly associated with functional capacity [17–19]. However, our study expands this finding by showing that sarcopenia is not just a predictor, but also an important mediator of the effect of an exercise program on another important factor for the health such as CRF.

Strong research methodologies, such as a randomized clinical trial with a blinded examiner, is one of the strengths of the present study. Also, to minimize the risk of bias, a large sample size was utilized. However, our study is not without limits. This research was developed with older women who were overweight and obese, and thus, we are not able to generalize the result to other populations of interest.

#### **5. Conclusions**

A gerontogymnastics program improves the functional capacity and fitness of older women who are overweight and obese. Sarcopenia acts as a mediator of the effect of a gerontogymnastics program on CRF in overweight and obese older women.

In this sense, the results support the new interest in changing the type of intervention and could be used to suggest that the improvements in strength, gait speed, and reduction of sarcopenia at the start of the exercise program could be needed to secure or improve the effects of the program on CRF and help improve the health of overweight and obese older people.

**Author Contributions:** Conceptualization, P.J.M.-P., N.G.-G. and R.G.d.S.V; Formal analysis, P.J.M.-P. and N.G.-G.; Methodology, P.J.M.-P., N.G.-G., A.L.-V. and A.E.-G.; Resources, P.J.M.-P., N.G.-G., G.M.G.-G., A.L.-V., A.E.-G. and R.G.d.S.V.; Supervision, P.J.M.-P. Validation, P.J.M.-P., N.G.-G. and R.G.d.S.V; Writing—original draft, P.J.M.-P.,

N.G.-G., A.L.-V., A.E.-G. and R.G.d.S.V; Writing—review & editing, P.J.M.-P., N.G.-G., G.M.G.-G., A.L.-V., A.E.-G. and R.G.d.S.V. All authors have read and agreed to the published version of the manuscript.

**Funding:** The present research on active aging of members (GISAFFCOM) of HEALTHY-AGE Network (reference 08/UPR/20) is supported by a grant from the Spanish Ministry of Culture and Sport- Sports Sciences Networks and GISAFFCOM research group is supported by a grant from the Spanish Ministry of Science, Innovation and Universities- RETOS I+d+i 2018 (RTC-2017-6145-1).

**Acknowledgments:** The research team would like to thank the heads of the social and women's centers and all the older women for their participation in this research, and the San Antonio Catholic University of Murcia (UCAM) for its support to the line of research on healthy and active aging.

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

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

International Journal of *Environmental Research and Public Health*

## *Article* **Relationships between Exercise Modality and Activity Restriction, Quality of Life, and Hematopoietic Profile in Korean Breast Cancer Survivors**

#### **MunHee Kim <sup>1</sup> , Wi-Young So <sup>2</sup> and Jiyoun Kim 3,\***


Received: 21 August 2020; Accepted: 17 September 2020; Published: 21 September 2020

**Abstract:** This study aimed to examine the relationships between activity restriction, quality of life (QoL), and hematopoietic profile in breast cancer survivors according to exercise modality. The subjects in this study were 187 female breast cancer survivors among a total of 32,631 participants in the Korea National Health and Nutrition Examination Survey, which was conducted from 2016 to 2018. The selected subjects participated in a questionnaire survey and blood analysis. A cross-analysis was conducted to determine the relationship between participation in various modality of exercise (e.g., aerobic exercise, resistance exercise, walking exercise). The phi coefficients or Cramer's V value for activity restriction and QoL were calculated; an independent *t*-test was conducted to evaluate the differences between hematopoietic profiles based on the modality of exercise. Statistically significant correlations were seen between obesity and aerobic exercise and walking frequency, as well as between diabetes and aerobic exercise and activity restriction. With respect to QoL, there was a statistically significant correlation between participation in aerobic exercise and exercise ability, participation in aerobic exercise and anxiety/depression, participation in resistance exercise and subjective health status, participation in resistance exercise and exercise ability, and participation in weekly walking exercise and self-care ability. Regarding hemodynamic changes, red blood cells increased significantly in breast cancer survivors who participated in weekly resistance exercise compared to in those who did not. In conclusion, exercise participation had a positive effect on activity restriction, QoL, and hematopoietic profile in breast cancer survivors; in particular, some modalities of aerobic exercise were more effective.

**Keywords:** aerobic exercise; obesity; resistance exercise; subjective health status; walking

#### **1. Introduction**

Breast cancer is the most common cancer among Korean women. According to a report from the Korea Central Cancer Registry under the Ministry of Health and Welfare, the age-adjusted cancer incidence rate in 2016 was 62.6 out of 100,000 women, which is significantly higher than 54.7 in 2014 and 56.1 in 2015. When noted according to age group, breast cancer is most common among women in their 40s (44.3%), followed by those in their 50s (30.2%) and 60s (16.1%) [1]. The Korea Ministry of Health and Welfare reported that, while the number of patients with cancer who survived for more than 5 years after cancer diagnosis exceeded 1 million in 2017, and the cancer survival rate reached 70%, 40% of patients with breast cancer developed depression [1,2]. Breast cancer survivors experience many activity restrictions because of sexual problems, infertility, fatigue, appearance, separation or divorce from their spouses, fear of cancer recurrence and death, etc. [2–6].

Because of these psychological stresses, those with breast cancer tend not to actively participate in many activities, which leads to the deterioration of cardiovascular health, muscle strength, and bone health, thus increasing the risk of osteoporosis and cardiovascular disease [7]. These diseases eventually cause activity restrictions for breast cancer survivors, including discomfort in daily life and absenteeism due to concurrent diseases [8]. In fact, 48.4% of breast cancer survivors in Korea are obese, which increases the risk of breast cancer recurrence and mortality to 35–40%, and may cause insulin resistance, metabolic syndrome, and type 2 diabetes [9]. Since obesity and diabetes also increase the risk of other cancers, more care is required [10]. In recent studies, the expression and activity of iron-related proteins (ferritin, hepcidin, and ferroportin) in breast cancer cells affected the prognosis of breast cancer [11]. In particular, poor iron metabolism (anemia) in patients with breast cancer is a common phenomenon based on tumor stage and anticancer treatment used, and about 43–47% of patients with breast cancer develop anemia [12,13]. In addition, patients with cancer experience inflammatory reactions in their bodies due to obesity, as their level of activity decreases because of fatigue [14,15].

The prevention of cancer is of primary importance; however, women who already have cancer need proper physical and emotional care to maintain their quality of life (QoL). For cancer survivors, the ongoing management of lifestyle (nutrition, physical activity, sleep, stress) is important. Among lifestyle factors, physical activity is widely recognized as an effective non-pharmacological treatment for patients with cancer [16–18]. In order to manage or prevent breast cancer, various modalities of exercise are used. Therefore, this study, utilizing the Korea National Health and Nutrition Examination Survey (KNHANES) conducted from 2016 to 2018, aimed to examine the relationships between participation in various modalities of exercise and activity restriction, QoL, and hemodynamic changes in breast cancer survivors, and to determine which modality of exercise is more effective for breast cancer survivors.

#### **2. Materials and Methods**

#### *2.1. Study Design and Participants*

The KNHANES is a national survey conducted by trained experts every year under the supervision of the Korea Centers for Disease Control and Prevention. This study was conducted with 187 live female patients with breast cancer of 32,631 participants who participated in the KNHANES for 3 years. The mean age of the subjects and mean age at diagnosis were post-menopausal. For more information about the physical characteristics of the subjects, please see Table 1.



#### *2.2. Physical Activity Assessment*

The KNHANES physical activity levels were measured using 3 exercise categories: (1) aerobic exercise only (medium intensity aerobic activity of 150 min per week or high-intensity aerobic activity for 75 min per week); (2) resistance exercise only (>1 time per week); (3) walking exercise only (1–2/times per week, 3–5/times per week, 6–7/times per week). Each participant could belong to multiple exercise categories.

#### *2.3. Activity Restriction*

The questions asked during the health interview survey regarding activity restrictions consisted of: the presence of activity restriction, causes of activity restriction, diseases, and experience of absenteeism. To ascertain activity restriction, 5 questions (discomfort in the last 2 weeks, disease in the past month, disease in the last year, absenteeism in the last month, absenteeism in the last year) were asked, and were designed to be answered with a "yes" or "no." In addition, the causes of activity restriction, the presence of obesity, diabetes, and anemia (dizziness), which are closely associated with breast cancer, were examined.

#### *2.4. Subjective Health Status and QoL*

For subjective health status, the question "How do you think your health is in normal times?" was asked, and answers were scored from 1 point for "Very Bad" to 5 points for "Very Good" using a 5-point Likert scale. The higher the score, the better the perceived health status by the subject. The EuroQoL-5 dimension (EQ-5D) developed by the EuroQoL Group was used to measure QoL related to overall health. It consists of 5 multiple-choice questions concerning exercise ability, self-care, daily activities, pain/discomfort, and anxiety/depression. Each of the 5 questions can be answered with 1 of 3 responses: "No problem at all", "There are some problems", and "There are serious problems". In this study, Cronbach's alpha for the instrument was 0.78.

#### *2.5. Blood Analysis*

Fasting blood samples from all participants in the Korea National Health and Nutrition Examination Survey were collected; white blood cells, red blood cells (RBCs), hemoglobin, platelets, hematocrit, and hs-C-reactive protein (CRP) levels were analyzed. For diabetes, stage 3 diabetes (normal, impaired glucose tolerance (IGT) and diabetes) was based on blood glucose after fasting for more than 8 h. Anemia was determined based on a hemoglobin level <12 g/dL.

#### *2.6. Ethics Statement*

The KNHANES was conducted as an interview survey in which the investigators interviewed the subjects and collect responses to the questions. In this study, the raw data from the seventh survey (2016–2018) that met the criteria of the study were downloaded from the KNHANES website (http://knhanes.cdc.go.kr/). In order to use the data, protocols for using the raw data from the KNHANES website were followed. Since the KNHANES is considered a public welfare study conducted by the Korean government, this study was conducted without the prior approval of the Research Ethics Review Committee.

#### *2.7. Statistical Analysis*

Phi coefficients or Cramer's V value were calculated using cross-analysis to determine whether there was a relationship between exercise participation, activity restriction, subjective health status, and QoL by exercise modality. An independent *t*-test was conducted to examine the differences between exercise modality participation by hematopoietic profile, and alpha (α) was set to 0.05. The reasons for the different case numbers for each variable is due to missing data from those who did not respond to the questionnaire. All analyses were conducted using SPSS version 18.0 (IBM Corp., Armonk, NY, USA).

#### **3. Results**

This study examined the relationships between exercise modality, activity restriction, subjective health status, QoL, and hematopoietic profile in breast cancer survivors who participated in the 2016–2018 KNHANES. The results of the cross-analysis, conducted to determine the correlation between exercise participation and activity restriction-related variants by exercise modality (aerobic exercise, resistance exercise, walking exercise) in the breast cancer survivors, are presented in Table 2. There were no statistically significant correlations between participation in various modalities of exercise and activity restriction (discomfort in the past 2 weeks, disease in the last month, disease in the last year, absenteeism in the last month, absenteeism in the last year) in the breast cancer survivors.

Among activity restriction due to disease, there was a statistically significant correlation between obesity and aerobic exercise participation (*p* < 0.046) and walking exercise frequency (*p* < 0.029). However, there was an exception; one subject who participated in aerobic and resistance exercises had a higher obesity rate than those who did not participate. There was also a significant correlation between diabetes and aerobic exercise participation at the level of *p* < 0.038. The subjects who participated in aerobic exercise showed a lower prevalence of diabetes compared to those who did not participate in aerobic exercise.

The results of the cross analysis, conducted to examine the correlation between subjective health status and QoL by exercise modality (aerobic exercise, resistance exercise, walking exercise) in breast cancer survivors, are shown in Table 3. There was a statistically significant correlation between subjective health status and resistance exercise participation at the level of *p* < 0.180. There was a statistically significant correlation between mobility, aerobic exercise participation, and resistance exercise participation at the levels of *p* < 0.028 and *p* < 0.026. There were also a statistically significant correlation between self-care and walking exercise frequency, and anxiety/depression and aerobic exercise participation at the levels of *p* < 0.037 and *p* < 0.017.

The independent *t*-test conducted to examine the effect of exercise participation on the hematopoietic profile by exercise modality in breast cancer survivors is presented in Table 4. The RBC was significantly higher at the level of *p* < 0.028 for those who participated in resistance exercise compared to those who did not.


**Table 2.** Correlation between exercise modality and activity restriction-related variants in breast cancer survivors.


**Table 3.** Correlation between exercise modality, subjective health status and QoL in breast cancer survivors.


**Table 4.** Differences between exercise modality participation in relation to hemodynamic variables in breast cancer survivors.

*Int. J. Environ. Res. Public Health* **2020**, *17*, 6899

*p* < 0.05; tested by an independent *t*-test.

#### **4. Discussion**

Breast cancer is affected by genetic and environmental factors, such as menarche, menopause, childbirth, and lactation experience; it is reported that a Western diet and inactive lifestyle increase the incidence [19]. Women who have undergone surgery because of the development of breast cancer, ovarian cancer, and uterine cancer may develop depression because they feel deprived of femininity, which may lead to family or social problems. In addition, it has been reported that the risk of myopathy, osteoporosis, and cardiovascular disease increases in breast cancer survivors [7].

Cancer survivors are recommended to participate in various modalities of exercise to prevent daily fatigue and cancer recurrence. In this study, there was no significant correlation between exercise participation and activity restriction-related discomfort or disease, and absenteeism for the last 2 weeks; however, there was a correlation between obesity and diabetes and activity restriction. Specifically, obesity and diabetes were significantly correlated with aerobic exercise participation and walking exercise frequency in breast cancer survivors. In this study, aerobic exercise and walking exercise showed a significantly positive correlation. It is suggested that aerobic exercise and walking (6–8 repetitions per week) are good solutions for obesity in breast cancer survivors. In addition, diabetes showed a correlation with aerobic exercise, showing that participation in aerobic exercise has a lower prevalence of diabetes compared with no participation in aerobic exercise. A meta-analysis conducted by Protani et al. [20], reported that the risk of cancer recurrence or death was 30% higher in breast cancer survivors who were obese than in breast cancer survivors of normal weight. It has also been reported that excess fat tissue caused by obesity increases the recurrence rate of breast cancer [21], and aerobic exercise (walking exercise) reduces the size of fat cells [22] and improves immune function [23]. However, increased fatigue due to a rapid increase in the level of activity may lower the immunity in patients with cancer, so care must be taken during exercise [24]. In addition, patients with breast cancer tend to lose muscle strength because of changes in body composition during anticancer treatment; resistance exercise has a positive effect on maintaining body composition and strength [25], indicating that breast cancer survivors need to participate in various modalities of exercise to further reduce cancer-related risk factors and prevent concurrent diseases.

In this study, there was a significant correlation between resistance exercise participation and subjective health status in breast cancer survivors. With respect to QoL, mobility and anxiety/depression were significantly correlated with aerobic exercise participation self-care, and walking exercise frequency. These results are similar to those of a study that reported a significant increase in QoL, fatigue, and depression symptoms after cancer survivors participated in exercise [26].

Regarding the examination of the relationship between exercise modality and the hematopoietic profile of the breast cancer survivors in this study, RBC significantly increased depending upon weekly resistance exercise participation; thus, those who participated in resistance exercise had higher RBC counts than those who did not. An increase in RBC count is closely associated with the prevalence of anemia. The blood cells of patients with cancer do not pass through blood vessels because of the deformation of red blood cells, which forms congestion and causes anemia. This phenomenon has been reported in more than 40–64% of patients with cancer [27,28]. Anemia can cause dizziness, weakness, and fatigue in everyday life, which may, in turn, deteriorate the QoL and restrict the activities of breast cancer survivors [29]. Although not statistically significant, it was found that the prevalence of anemia was higher in those who participated in all modality of exercise than in those who did not, as shown in Table 1, which is considered to be closely associated with increased RBCs, even though they were within the normal range. Mohamady et al. [30] and Drouin et al. [31] reported that participation in a 7-week exercise program prevented the increase in RBC and hemoglobin in patients with breast cancer who were undergoing radiation therapy. It has also been reported that exercise improves systemic inflammation in cancer survivors [32–35].

However, in this study, the inflammatory index hs-CRP was within the normal range for all exercise modality, and there was no difference. The results of this study found that participation in physical activities (aerobic exercise, resistance exercise, walking exercise) lowered the prevalence of obesity and diabetes affecting patients with breast cancer. Physical activity participation improved subjective health status and exercise ability, and reduced depression and anxiety, thus improving the quality of life of breast cancer survivors in Korea. Among the modalities of exercise assessed, aerobic exercise had a greater positive correlation, indicating that it may be more effective.

There are several limitations to this study. First, the amount of exercise participation was not directly measured by objective observance, but surveyed indirectly by using a questionnaire. Second, there was a lack of a methodological approach for measuring the proper amount of exercise according to the grade of breast cancer and cancer therapy method, suggesting the necessity of a follow-up study. Third, this study was conducted only with patients with breast cancer; thus, the findings cannot be generalized to other cancer patients. Fourth, additional physical activity evaluations, such as activities of daily living or instrumental activities of daily living, were not conducted, suggesting the necessity for a further study with additional variables. However, combining resistance exercise and aerobic exercise to lessen muscle weakening is recommended. The habit of performing exercise on a regular basis is considered most important for breast cancer survivors for the prevention of cancer recurrence and for cancer recovery.

#### **5. Conclusions**

This study found that, for breast cancer survivors, participation in physical activity, such as aerobic exercise, resistance exercise, and walking exercise may lower the prevalence of diseases such as obesity and diabetes. Furthermore, physical activity can reduce depression and anxiety and improve subjective health status, exercise ability, and quality of life. In particular, aerobic exercise was shown to be effective in positively affecting a number of variables, but resistance training is also recommended to prevent muscle loss. The effort to establish regular exercise habits, regardless of modality, seems to be important for the mental and physiological health of breast cancer survivors.

**Author Contributions:** Study design: M.K., J.K., and W.-Y.S., Study conduct: M.K., J.K., and W.-Y.S., Data collection: M.K., J.K., and W.-Y.S., Data analysis: M.K., J.K., and W.-Y.S., Data interpretation: M.K., J.K., and W.-Y.S., Drafting manuscript: M.K., J.K., and W.-Y.S., Revising the manuscript content: M.K., J.K., and W.-Y.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by the Ministry of Education of the Republic of Korea and the National Research Foundation of Korea (NRF-2019S1A5B6102784).

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

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

*Int. J. Environ. Res. Public Health* **2020**, *17*, 6604 International Journal of *Environmental Research and Public Health*

## **Body Water Content and Morphological Characteristics Modify Bioimpedance Vector Patterns in Volleyball, Soccer, and Rugby Players**

**Francesco Campa 1,\* , Analiza M. Silva <sup>2</sup> , Catarina N. Matias <sup>2</sup> , Cristina P. Monteiro <sup>2</sup> , Antonio Paoli <sup>3</sup> , João Pedro Nunes <sup>4</sup> , Jacopo Talluri <sup>5</sup> , Henry Lukaski <sup>6</sup> and Stefania Toselli <sup>7</sup>**


Received: 3 August 2020; Accepted: 9 September 2020; Published: 10 September 2020

**Abstract: Background:** Bioimpedance vector analysis (BIVA) is a widely used method based on the interpretation of raw bioimpedance parameters to evaluate body composition and cellular health in athletes. However, several variables contribute to influencing BIVA patterns by militating against an optimal interpretation of the data. This study aims to explore the association of morphological characteristics with bioelectrical properties in volleyball, soccer, and rugby players. **Methods:** 164 athletes belonging to professional teams (age 26.2 ± 4.4 yrs; body mass index (BMI) 25.4 ± 2.4 kg/m<sup>2</sup> ) underwent bioimpedance and anthropometric measurements. Bioelectric resistance (R) and reactance (Xc) were standardized for the athlete's height and used to plot the vector in the R-Xc graph according to the BIVA approach. Total body water (TBW), phase angle (PhA), and somatotype were determined from bioelectrical and anthropometric data. **Results:** No significant difference (*p* > 0.05) for age and for age at the start of competition among the athletes was found. Athletes divided into groups of TBW limited by quartiles showed significant differences in the mean vector position in the R-Xc graph (*p* < 0.001), where a higher content of body fluids resulted in a shorter vector and lower positioning in the graph. Furthermore, six categories of somatotypes were identified, and the results of bivariate and partial correlation analysis highlighted a direct association between PhA and mesomorphy (*r* = 0.401, *p* < 0.001) while showing an inverse correlation with ectomorphy (*r* = −0.416, *p* < 0.001), even adjusted for age. On the contrary, no association was observed between PhA and endomorphy (*r* = 0.100, *p* = 0.471). **Conclusions:** Body fluid content affects the vector length in the R-Xc graph. In addition, the lateral displacement of the vector, which determines the PhA, can be modified by the morphological characteristics of the athlete. In particular, higher PhA values are observed in subjects with a high mesomorphic component, whereas lower values are found when ectomorphy is dominant.

**Keywords:** body composition; BIVA; phase angle; R-Xc graph; somatotype; total body water; vector length

#### **1. Introduction**

In recent years, the bioimpedance vector analysis (BIVA) has been widely used in the sports field for the assessment of body composition and cellular health in athletes [1]. This is because BIVA is not subject to errors related to prediction equations since it interprets the raw bioimpedance values [resistance (R) and reactance (Xc)], and it is an easy to use and non-invasive method. Its application is based on the bivariate interpretation of R and Xc standardized for height on a graph. Vector displacements identify increases or losses in total body water (TBW) or in the ratio between intra (ICW) and extracellular (ECW) fluids for which increases or decreases correspond to shifts to the left or to the right of the R-Xc graph, respectively [2]. In addition, it is possible to obtain an immediate analysis of the subject's body composition by comparing the vector position with tolerance ellipses built from the data of the reference population [3–6].

Recent studies have focused on evaluating factors influencing the vector position of athletes, including maturity status, [7–10] dehydration [11–14], and fitness level [15]. Since the bioelectric properties of body tissues depend on body fluids and cells' membrane integrity [16], the main determinant of the vector position in the R-Xc graph is the TBW and the distribution of the fluids among the two compartments (ICW and ECW). In fact, the vector length is inversely proportional to TBW, while the lateral displacements of the vector are directly correlated with the ICW/ECW ratio [17–20]. In addition to these body composition variables, Campa et al. [21] have recently suggested that the somatotype also influences vector position in the R-Xc graph, where athletes with higher mesomorphic and endomorphic components are positioned more to the lower-left than athletes with a dominant ectomorphy. However, as the content of total body fluids is greatly associated with the vector, correct discrimination in the R-Xc graph based on somatotype categories may be compromised. Indeed, athletes with high body weight can still be located at the bottom of the graph regardless of their morphology.

The bioelectric values and the vector position reflect body composition; in particular, the phase angle (PhA), obtained as the arctangent of Xc/R, correctly mirrors the ICW/ECW ratio [17–20]. In fact, athletes with a high PhA are positioned to the left portion in the R-Xc graph [21,22], and increases in muscle mass, body cell mass, and, therefore, ICW lead to vector shifts further to the left over time [23,24]. A high muscularity is observed in subjects with a high PhA or in those whose somatotype shows a dominant mesomorphic component [21,22], as mesomorphy characterizes skeletal muscle features [25]; moreover, both PhA and somatotype can be modified with nutrition and exercise [23,26]. However, to the best of our knowledge, no study has explored the associations between somatotype and PhA, while analyzing the influence of morphology on the vector position for similar TBW values in athletes.

Therefore, this study aimed to analyze the associations of morphological characteristics with bioelectrical properties using the BIVA approach, according to different levels of body water content in male volleyball, soccer, and rugby players. Our hypothesis was that PhA was associated with morphological characteristics in the athletes.

#### **2. Methods**

#### *2.1. Subjects*

This was a cross-sectional observational study conducted in 164 athletes engaged in 7 professional Italian teams participating in Series A2, Series B, and Series A divisions of volleyball, soccer, and rugby, respectively (age 26.2 ± 4.4 yrs; body mass index (BMI) 25.4 ± 2.4 kg/m<sup>2</sup> ; age at start competition

14.2 ± 1.3 yrs). The following inclusion criteria were used: (i) A minimum of 10 h of training per week; (ii) tested negative for performance-enhancing drugs; and (iii) not taking any medication. The athletes were tested in the morning (9:00 AM) in the facilities of the teams. All measurements were performed under resting conditions in the second off-season period. All participants gave informed consent after receiving a detailed description of the study procedures. The project was conducted in accordance with the guidelines of the Declaration of Helsinki and was approved by the local Bioethics Committee of the University of Bologna. (Ethical Approval Code: 25027).

#### *2.2. Procedures*

All athletes were tested to ensure a well-hydrated state using the urine specific gravity test (refractometer Urisys 1100; Roche Diagnostics), according to Armstrong et al. [27]. A urine specific gravity value < 1.022 for the first urine was used to identify an euhydration state.

The anthropometric traits were body mass, height, humerus and femur breadths, contracted arm and calf girths, and 4-skinfold thicknesses (triceps, subscapular, supraspinal, and medial calf). All anthropometric measurements were taken by a certified anthropometrist according to standard methods in the literature [28], whose technical error was 5% and 1.5% for skinfolds and all other measurements, respectively. Height was recorded to the nearest 0.1 cm using a stadiometer (Raven Equipment Ltd., Great Donmow, UK) and body mass was measured to the nearest 0.1 kg using a high-precision mechanical scale (Seca, Basel, Switzerland). BMI was calculated as the ratio of body weight to height squared (kg/m<sup>2</sup> ). Girths were taken to the nearest 0.1 cm using a tape measure (GMP, Zürich, Switzerland). Breadths were measured to the nearest 0.1 cm using a sliding caliper (GMP, Zürich, Switzerland). Skinfold thicknesses were measured to the nearest 0.1 mm using a Lange skinfold caliper (Beta technology Inc., Cambridge, MD, USA).

Bioimpedance analysis (BIA) was performed by a phase-sensitive single-frequency bioimpedance analyzer (101 Anniversary, Akern, Florence, Italy), which applied an alternating current of 400 microamperes at 50 kHz. Vector length (VL) was calculated as (adjusted R<sup>2</sup> + adjusted Xc<sup>2</sup> ) 0.5 and PhA as the arctangent of Xc/R × 180/π. BIVA was applied to normalize V, R, and Xc for height (H) in meters [29]. TBW was calculated from bioimpedance values, according to specific equations developed for athletes using a 4-compartment model as a criterion method [30] then the athletes were divided into quartiles.

Somatotype components were calculated according to the Heath and Carter method [25] as follow:

*Endomorphy* = − 0.7182 + 0.1451 (X) − 0.00068 (X 2) + 0.0000014 (X 3), where X = (sum of triceps, subscapular and supraspinal skinfolds) multiplied by (170.18/H in cm);

*Mesomorphy* = 0.858 × humerus breadth + 0.601 × femur breadth + 0.188 × corrected arm girth + 0.161 × corrected calf girth − H 0.131 + 4.5;

*Ectomorphy* = 0.732 HWR − 28.58, where HWR = (height divided by the cube root of weight).

From the 13 initial proposed categories by Heath and Carter [25], the athletes were grouped in 6 somatotype categories:


#### *2.3. Statistical Analysis*

To verify the normality of the data, the Shapiro-Wilk test was applied. The athletes were divided into groups limited by quartiles of TBW and the one-way ANOVA was performed to evaluate the difference in BIVA patterns (PhA and VL/H). When a significant F ratio was obtained, the Bonferroni post hoc test was used to assess the differences between the 4 groups, setting the significance at *p* < 0.008. The two-sample Hotelling's T<sup>2</sup> test was used to compare the mean impedance vectors among the athletes grouped according to quartiles of TBW. Bivariate and partial (controlling for age) correlations were performed to evaluate the associations between PhA and the somatotype components. The mean standard deviation was calculated for each variable. Data were analyzed with IBM SPSS Statistics, version 24.0 (IBM Corp., Armonk, NY, USA). *Int. J. Environ. Res. Public Health* **2020**, *17*, x 4 of 11 post hoc test was used to assess the differences between the 4 groups, setting the significance at *p* <

#### **3. Results**

No significant difference (*p* > 0.05) for age and for age at start of competition among the athletes was found. The soccer, volleyball, and rubgy players showed an average EcM (endomorphy: 1.6 ± 0.3; mesomorphy: 4.7 ± 0.9; ectomorphy: 2.9 ± 0.8), EcM (endomorphy: 2.0 ± 0.7; mesomorphy: 4.0 ± 1.3; ectomorphy: 3.2 ± 1.1), and EnM (endomorphy: 2.1 ± 0.7; mesomorphy: 6.0 ± 1.1; ectomorphy: 0.9 ± 0.3) somatotype, respectively (Figure 1). components. The mean standard deviation was calculated for each variable. Data were analyzed with IBM SPSS Statistics, version 24.0 (IBM Corp., Armonk, NY, USA). **3. Results** No significant difference (*p* > 0.05) for age and for age at start of competition among the athletes was found. The soccer, volleyball, and rubgy players showed an average EcM (endomorphy: 1.6 ± 0.3; mesomorphy: 4.7 ± 0.9; ectomorphy: 2.9 ± 0.8), EcM (endomorphy: 2.0 ± 0.7; mesomorphy: 4.0 ± 1.3; ectomorphy: 3.2 ± 1.1), and EnM (endomorphy: 2.1 ± 0.7; mesomorphy: 6.0 ± 1.1; ectomorphy: 0.9 ± 0.3) somatotype, respectively (Figure 1).

0.008. The two-sample Hotelling's T<sup>2</sup> test was used to compare the mean impedance vectors among the athletes grouped according to quartiles of TBW. Bivariate and partial (controlling for age) correlations were performed to evaluate the associations between PhA and the somatotype

**Figure 1.** Representation of the athletes' somatotype. **Figure 1.** Representation of the athletes' somatotype.

Descriptive body fluids and bioelectrical characteristics are presented in Table 1, while the mean impedance vectors of the athletes divided according to quartiles of TBW are shown in Figure 1. Fortytwo athletes were included in the first group (Q1) (endomorphy: 1.8 ± 0.6, mesomorphy: 4.4 ± 1.1, ectomorphy: 2.9 ± 1.0), 40 in the second group (Q2) (endomorphy: 2.2 ± 0.7 , mesomorphy: 4.6 ± 1.5, ectomorphy 2.8 ± 1.3), 41 in the third group (Q3) (endomorphy: 2.2 ± 0.7, mesomorphy: 4.3 ± 1.4, ectomorphy 2.8 ± 1.3) and 41 in the fourth group (Q4) (endomorphy: 2.4 ± 0.7, mesomorphy: 5.2 ± 1.6, ectomorphy 2.0 ± 1.5). Six somatotype categories were identified, and their absolute frequencies for each group are presented in Figure 2. The results of the two-sample Hotelling t<sup>2</sup> test showed significant differences between all the groups (Q1 vs. Q2, *t* = 21.1, *p* < 0.001; Q1 vs. Q3, *t* = 105.8, *p* < 0.001; Q1 vs. Q4, *p* < 0.001; *t* = 201.6, *p* < 0.001; Q2 vs. Q3, *t* = 39.4, *p* < 0.001; Q2 vs. Q4, *t* = 98.1, *p* < 0.001; Q3 vs. Q4, *t* = 22.7, *p* < 0.001) indicating that the athletes with higher TBW were positioned to the lower left in the R-Xc graph than those with a lower TBW, as displayed in Figure 2. In addition, Descriptive body fluids and bioelectrical characteristics are presented in Table 1, while the mean impedance vectors of the athletes divided according to quartiles of TBW are shown in Figure 1. Forty-two athletes were included in the first group (Q1) (endomorphy: 1.8 ± 0.6, mesomorphy: 4.4 ± 1.1, ectomorphy: 2.9 ± 1.0), 40 in the second group (Q2) (endomorphy: 2.2 ± 0.7, mesomorphy: 4.6 ± 1.5, ectomorphy 2.8 ± 1.3), 41 in the third group (Q3) (endomorphy: 2.2 ± 0.7, mesomorphy: 4.3 ± 1.4, ectomorphy 2.8 ± 1.3) and 41 in the fourth group (Q4) (endomorphy: 2.4 ± 0.7, mesomorphy: 5.2 ± 1.6, ectomorphy 2.0 ± 1.5). Six somatotype categories were identified, and their absolute frequencies for each group are presented in Figure 2. The results of the two-sample Hotelling t<sup>2</sup> test showed significant differences between all the groups (Q1 vs. Q2, *t* = 21.1, *p* < 0.001; Q1 vs. Q3, *t* = 105.8, *p* < 0.001; Q1 vs. Q4, *p* < 0.001; *t* = 201.6, *p* < 0.001; Q2 vs. Q3, *t* = 39.4, *p* < 0.001; Q2 vs. Q4, *t* = 98.1, *p* < 0.001; Q3 vs. Q4, *t* = 22.7, *p* < 0.001) indicating that the athletes with higher TBW were positioned to the lower left in the R-Xc graph than those with a lower TBW, as displayed in Figure 2. In addition, significant differences (*p* < 0.008) were found between the 4 groups for VL/H but not for PhA, as reported in Table 2.

Figure 3 illustrates the mean vectors of the athletes subdivided by somatotype in each TBW group. For each TBW group, somatotype categories with a dominant mesomorphy (EnM, BM, and EcM) showed a vector tending to be positioned more to the left than those with a greater ectomorphy

reported in Table 2.

TBW (L)

R/H (Ohm/m)

Xc/H (Ohm/m)

(MEc and BEc). Moreover, as displayed in Figure 4, PhA was directly correlated with the mesomorphic component (*r* = 0.401, *p* < 0.001; Panel A) and inversely with the ectomorphic component (*r* = −0.416, *p* < 0.001; Panel B), even when corrected for age (*p* < 0.001). On the contrary, no association was observed between PhA and endomorphy (*r* = 0.100, *p* = 0.471; Panel C). BEc 36.1 ± 3.3 34.2 ± 1.5 30.4 ± 3.1 29.0 ± 3.8 Whole sample 35.9 ± 3.0 34.3 ± 3.1 30.5 ± 2.7 28.4 ± 3.5 Note: Data are presented as mean ± SD. TBW = total body water, R/H = resistance standardized for height, Xc/H = reactance standardized for height, EnM = Endomorphic Mesomorph, BM = Balanced Mesomorph, EcM = Ectomorphic Mesomorph, M-Ec = Mesomorph Ectomorph, MEc = Mesomorphic Ectomorph, BEc = Balanced Ectomorph.

*Int. J. Environ. Res. Public Health* **2020**, *17*, x 5 of 11 significant differences (*p* < 0.008) were found between the 4 groups for VL/H but not for PhA, as

**Table 1.** Descriptive statistics for the athletes divided by total body water quartile [first (Q1), second

**Q2 (n = 40)**

EnM 49.2 ± 0.3 52.4 ± 1.5 56.6 ± 1.4 67.5 ± 3.8 BM 45.9 ± 1.6 51.6 ± 1.6 58.1 ± 1.1 65.1 ± 5.2 EcM 45.9 ± 3.1 52.5 ± 0.9 56.7 ± 1.3 63.9 ± 4.7 M-Ec - 53.7 ± 0.8 56.7 ± 0.9 62.4 ± 1.2 MEc - 52.4 ± 0.7 57.7 ± 1.5 61.2 ± 1.2 BEc 44.2 ± 3.6 51.7 ± 0.9 56.9 ± 1.3 61.6 ± 0.7 Whole sample 46.1 ± 2.9 52.3 ± 1.3 56.9 ± 1.3 65.2 ± 4.3

EnM 247.0 ± 8.8 238.5 ± 14.1 214.9 ± 18.8 203.9 ± 18.1 BM 266.8 ± 17.3 238.2 ± 16.9 223.5 ± 6.0 211.6 ± 17.6 EcM 262.2 ± 15.9 245.5 ± 18.4 223.5 ± 10.8 203.1 ± 12.5 M-Ec - 249.2 ± 16.1 237.7 ± 12.8 234.6 ± 29.0 MEc - 268.0 ± 9.5 234.5 ± 12.5 224.2 ± 4.9 BEc 287.8 ± 17.6 253.9 ± 9.5 243.1 ± 17.2 219.6 ± 14.0 Whole sample 262.9 ± 17.1 245.5 ± 17.3 227.5 ± 17.2 209.0 ± 17.9

EnM 35.9 ± 1.4 34.7 ± 3.3 30.7 ± 2.8 28.3 ± 4.0 BM 36.0 ± 2.7 33.4 ± 2.6 31.9 ± 3.6 27.6 ± 3.9

M-Ec - 32.2 ± 2.3 31.5 ± 2.6 31.4 ± 3.1 MEc - 34.7 ± 2.3 31.1 ± 2.6 30.2 ± 2.2

**Q3 (n = 41)**

**Q4 (n = 41)**

**(n = 42)**

(Q2), third (Q3), and forth (Q4) quartile] and somatotype. **Variable Somatotype Q1**

**Figure 2.** Scattergram of the mean impedance vectors of the athletes divided by the total body water groups plotted on the 50%, 75%, and 95% tolerance ellipses of the healthy male Italian population [31]; on the right side, the absolute frequencies of the somatotype categories for each quartile is shown. **Figure 2.** Scattergram of the mean impedance vectors of the athletes divided by the total body water groups plotted on the 50%, 75%, and 95% tolerance ellipses of the healthy male Italian population [31]; on the right side, the absolute frequencies of the somatotype categories for each quartile is shown. EnM = Endomorphic Mesomorph, BM = Balanced Mesomorph, EcM = Ectomorphic Mesomorph, M-Ec = Mesomorph Ectomorph, MEc = Mesomorphic Ectomorph, BEc = Balanced Ectomorph. *Int. J. Environ. Res. Public Health* **2020**, *17*, x 7 of 11

**Figure 3.** Scattergram of the mean impedance vectors of the athletes are categorized by somatotype and divided according to groups of TBW. EnM = Endomorphic Mesomorph, BM = Balanced Mesomorph, EcM = Ectomorphic Mesomorph, M-Ec = Mesomorph Ectomorph, MEc = Mesomorphic Ectomorph, BEc = Balanced Ectomorph. **Figure 3.** Scattergram of the mean impedance vectors of the athletes are categorized by somatotype and divided according to groups of TBW. EnM = Endomorphic Mesomorph, BM = Balanced Mesomorph, EcM = Ectomorphic Mesomorph, M-Ec = Mesomorph Ectomorph, MEc = Mesomorphic Ectomorph, BEc = Balanced Ectomorph.

**Figure 4.** Correlation between phase angle and mesomorphy (**A**), ectomorphy (**B**), and endomorphy

= subjects from Q4.

= subjects from Q2. □ = subjects from Q3. △

(**C**). ◇

**4. Discussion**

= subjects from Q1. ○




Note: Data are presented as mean ± SD. TBW = total body water, R/H = resistance standardized for height, Xc/H = reactance standardized for height, EnM = Endomorphic Mesomorph, BM = Balanced Mesomorph, EcM = Ectomorphic Mesomorph, M-Ec = Mesomorph Ectomorph, MEc = Mesomorphic Ectomorph, BEc = Balanced Ectomorph.



Note: Data are presented as mean ± SD. PhA = phase angle, VL/H = vector length standardized for height, EnM = Endomorphic Mesomorph, BM = Balanced Mesomorph, EcM = Ectomorphic Mesomorph, M-Ec = Mesomorph Ectomorph, MEc = Mesomorphic Ectomorph, BEc = Balanced Ectomorph. \* Results of the one-way ANOVA considering the athletes as a whole sample; 1 Different (*p* < 0.008) from Q1; 2 Different (*p* < 0.008) from Q2; 3 Different (*p* < 0.008) from Q3; 4 Different (*p* < 0.008) from Q4.

**Figure 4.** Correlation between phase angle and mesomorphy (**A**), ectomorphy (**B**), and endomorphy (**C**). = subjects from Q1. # = subjects from Q2. = subjects from Q3. = subjects from Q4.

#### **4. Discussion**

The aim of this study was to analyze the associations of morphological characteristics with bioelectrical properties in volleyball, soccer, and rugby players. An important finding has emerged from our results regarding the role of body fluids on vector length. In addition, this study has shown, for the first time, the associations between the somatotype components and PhA. As hypothesized, when considering subjects with a similar TBW, the differences in vector position may reflect morphological peculiarities; this was possible to observe due to the data analysis carried out in this study, in which the athletes were divided into separate groups limited by quartiles of TBW.

We observed that body fluids content was a determining factor for vector length, extending the findings of previous research studies [17,20]. In particular, athletes with a higher TBW (Q4) showed a mean vector positioned lower than the other athletes (Figure 2). This is in line with previous studies that observed vector length and its changes to accurately reflect changes in TBW using dilution techniques to assess water and its compartments [17,20,21]. In addition, when the mean vectors for each somatotype category were plotted on the R-Xc graph, it was possible to observe how athletes with a higher mesomorphic component showed a vector tending to be more left in the graph. Conversely, athletes with a dominant ectomorphy presented a vector displacement to the right. In this regard, our results have shown that PhA correlates directly with the mesomorphic component, while an inverse association was observed with the ectomorphic component of the somatotype. These findings are in line with previous investigations that observed in athletes with a higher muscle mass, including bodybuilders, a vector position at the limits of the 95th percentile to the left of the reference ellipses of the normal population [3,22].

The athletes belonging to the six identified somatotype categories were distributed among the four groups of TBW except for the first group where mesomorph ectomorph, and mesomorphic ectomorph athletes, were not present. As a result, it was possible to explore the association between PhA and morphological features. The endomorphic mesomorph, balanced mesomorph, and ectomorphic mesomorph somatotypes are characterized by a dominant mesomorphy, due to a muscular related body shape. On the contrary, mesomorphic ectomorph and balanced ectomorph categories imply a dominant ectomorphy, and therefore, athletes tend to be taller with a lower muscle mass than the other somatotype categories [25]. Rakovi'c et al. [32] showed how mesomorphic features are linked to individual sports that require higher muscle strength, while ectomorphy is predominant in runners [33], especially those involved in long distance. In previous research [21], it was highlighted how R/H and Xc/H were able to discriminate somatotypes. However, if we consider this new and more individual approach, considering the TBW values, probably some of those athletes needed to be revised, since body fluids have a great influence on the vector position. For this reason, the athletes' somatotypes were analyzed according to groups of body fluids to reduce the differences attributed to TBW, and consequently, to better understand how somatotype is associated with the vector position. Due to this approach, it was possible to observe how the vector position changes based on the morphological features. Indeed, when the athletes were divided into TBW groups, significant differences were found in vector length, but not for PhA, which instead represents the lateral displacement of the vector. This suggests that athletes with a similar PhA could have a greater TBW and, therefore, a different vector position and body composition characteristics. In fact, when comparing the four groups, athletes with a shorter mean vector were those with a higher TBW. A recent literature review on PhA in sports [34] concluded that it was not clear whether PhA differs among athletes engaged in different sports. On the contrary, studies on athletes practicing the same sports, but at different competitive levels, have shown that elite athletes show a higher PhA than those engaged in lower-level categories [4–6]. In particular, Micheli et al. [6] suggested that this is due to a lower R/H in relation to Xc/H and reflects a condition of greater muscularity and body cell mass content in the athletes that compete in the higher levels. In this study, it was shown how the interpretation of the vector position in the R-Xc graph overcomes the limits linked to the interpretation of the PhA alone. In this regard, Reis et al. [35] recently showed that the bioimpedance vector position varied in response to changes in the macrocycle training load in

swimmers of both sexes. The authors identified an accumulation of fluids and, therefore, a shortening of the vector following a first phase characterized by a high training load, and then a subsequent shift of the vector position to the left side as a result of muscle adaptations that occurred after a recovery period. Similarly, Mascherini et al., in 2014 [24], for the first time studied vector changes over the competitive season in soccer players and showed that PhA decreased during the preparatory phase and then increased near the beginning of the competition. In line with this, Nabuco et al. [36] highlighted a moderate and inverse association between PhA and the values obtained from a fatigue assessment test.

The results of the present study add important and useful information for a correct interpretation of BIVA. The careful evaluation and monitoring of body composition allow the athlete to be predisposed to achieving high peak performance [37]. Through BIVA, it is possible to evaluate the progress in the athlete's physical condition during the season [38] in response to a training program or a nutritional intervention [37], avoiding incurring decreases in physical performance. This method, in addition to monitoring body fluids, also allows information about other body composition variables at a whole-body level [39]. In fact, it is not always possible to collect the various anthropometric measures that allow the evaluation of the somatotype; in this regard BIVA, if correctly interpreted, can provide important information, minimizing the need for skinfolds and girths collection by a certified anthropometrist. While this method requires further study, especially concerning monitoring the bioimpedance parameters in the short term, the innovation of this study lies in the fact that it provides useful information for the correct interpretation of the vector position in the R-Xc graph, specifically the role of body water as a mediator of the morphological associations with bioelectrical parameters.

Some limitations of this study need to be addressed. First of all, our results are only generalized for male athletes. Secondly, the findings of this study are only applicable to single-frequency BIA equipment. In fact, different results in measuring raw BIA parameters are obtained using devices that work on single- or multi-frequency [40]. On the other hand, the fact that it was ensured that the athletes were in a state of euhydration is a strength of the present work. Future research should study the potential of BIVA patterns as a biomarker of physical condition during the training process, as in specific microcycles. Indeed, as bioimpedance analysis (BIA) equipment provides an easy, simple, and low-cost application, it allows for a frequent assessment to optimize monitoring of the athlete's physical condition.

#### **5. Conclusions**

BIVA provides meaningful information on body composition assessment in athletes. This study showed that body fluid content affects the vector length in the R-Xc graph, while the lateral displacement of the vector can be modified by the morphological characteristics of the athlete. In particular, the mesomorphic component of the somatotype is related to higher PhA, an important marker of cellular integrity and overall physical function.

**Author Contributions:** Conceptualization, A.M.S., C.N.M., J.T., H.L. and S.T.; data curation, F.C. and C.P.M.; formal analysis, F.C., C.P.M.and J.P.N.; investigation, A.P., J.T., and S.T.; methodology, A.M.S., C.N.M., J.P.N., and H.L.; supervision, H.L. and S.T.; visualization, S.T.; writing—original draft, F.C., A.M.S., C.N.M., and J.P.N.; writing—review and editing, F.C., A.M.S, C.N.M., C.P.M., A.P., and S.T. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Acknowledgments:** The authors are grateful to all the athletes who took part in this study.

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

#### **References**



© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

International Journal of *Environmental Research and Public Health*

## *Article* **E**ff**ects of Aerobic and Anaerobic Fatigue Exercises on Postural Control and Recovery Time in Female Soccer Players**

**Özkan Güler <sup>1</sup> , Dicle Aras <sup>1</sup> , Fırat Akça <sup>1</sup> , Antonino Bianco 2,\* , Gioacchino Lavanco <sup>2</sup> , Antonio Paoli <sup>3</sup> and Fatma Ne¸se ¸Sahin <sup>1</sup>**


Received: 4 August 2020; Accepted: 24 August 2020; Published: 28 August 2020

**Abstract:** Sixteen female soccer players (age = 20.19 ± 1.52 years; body mass = 56.52 ± 4.95 kg; body height = 164.81 ± 4.21 cm) with no history of lower extremity injury participated in the study. The Biodex SD Balance system was used to determine the non-dominant single-leg stability. In anaerobic exercise, each subject performed four maximal cycling efforts against a resistance equivalent to 0.075 kg/body mass for 30 s with three-minute rest intervals. In aerobic exercise, subjects performed the Bruce protocol on a motorized treadmill. After each exercise, subjects subsequently performed a single-leg stability test and then repeated the same test for four times with five-minute passive rest periods. In accordance with the results, it was found that the impairment observed right after the aerobic loading was higher (*p* < 0.001) compared to the anaerobic one. However, the time-related deterioration in both aerobic and anaerobic loadings was similar. The B-pre value was lower than Bpost and B<sup>5</sup> (*p* < 0.01) and B<sup>10</sup> (*p* < 0.05) in both conditions. Subjects could reach the initial balance level at B<sup>15</sup> after aerobic and anaerobic loadings. The lactate level did not reach resting value even after 20 min of both fatigue protocols. Although the fatigue after aerobic and aerobic exercise negatively affects a single-leg dynamic balance level, single leg balance ability returns to the baseline status after 10 min of passive recovery duration.

**Keywords:** balance; fatigue; female; support leg; recovery

#### **1. Introduction**

The popularity of soccer among females is increasing each passing day. It is estimated that around 30 million females are actively playing licensed soccer in more than 100 countries around the world. Studies indicate that as the participation of females in soccer increases, the incidence of injury increases at a high rate [1–3]. In a soccer match, soccer players engage in many moves such as high-intensity acceleration, deceleration, sudden change of direction, bounce, and other soccer-oriented movements. Along with these moves, soccer players often experience various injuries when using one leg for stopping and cutting during pressure, while using the other leg to tackle the ball [4]. In addition to these, injuries in soccer are caused by sudden acceleration and deceleration without impact, rapid disorientation, and exposure to high loads while maintaining the stability of the knee joint in unpredictable movements [5–8]. When the injuries experienced in soccer were evaluated according to sex differences, It was reported that female athletes had a higher incidence of anterior cruciate

ligament (ACL) experience in lower extremity injuries due to biomechanical and neuromuscular differences [5,6] than males [9].

Furthermore, previous studies indicate that female soccer player has the risk of ACL injury nine times greater than males [10]. Several risk factors cause these injuries in female soccer players. These risk factors in female soccer players include a previous history of injury [11], as well as a decline in hip strength [12] due to accumulated fatigue [11] and deterioration of lower extremity dynamic balance [13]. Epidemiological studies have pointed out that 50% of the injuries occur at the end of competitions or sports activities, and 58% of these injuries are due to non-impact conditions. That fatigue is an essential element of sensory-motor changes associated with injury [14,15]. Ekstrand et al. [9] report in their study that traumatic injuries occur more often in the last minutes of both halves of a soccer match [9]. In addition to these, it was reported in another study that lower extremity injuries were commonly seen at the last minutes of competition in sports such as soccer, which includes high-intensity moves and multi-directional sprints [9,16]. Therefore, it can be stated that non-contact injuries caused by fatigue occurred in the last fifteen minutes of play in both the first and second half of games. In non-contact injuries, neuromuscular fatigue is seen as a risk factor [17–20]. Neuromuscular fatigue is divided into two, according to the intensity and duration of exercise, like peripheral and central nervous system fatigue. Long-term activities affect the central nervous system, while short-term high-intensity activities cause peripheral fatigue [21,22]. Peripheral fatigue arises when there is not adequate energy provided to the muscles, despite the increasing energy need [22].

Moreover, it has been reported that muscle fatigue affects both peripheral and central proprioceptive processes [23–25]. Balance is defined as being able to hold the body center of gravity within the center of support [26]. In order to maintain balance visual, vestibular, and proprioceptive systems play a crucial role, and these systems are affected by many factors [27–32]. The proprioceptive system consists of the Golgi tendon organ, the muscle spindle, the Pacini corpuscle, free nerve endings, and the receptors in the joint capsules and skin [32–34]. It ensures maintenance of the balance with the information collected from these structures [32,34]. The proprioceptive system is affected by fatigue, aging, sarcopenia, neurological disease fibromyalgia, cancer, and rheumatological diseases and may result in impaired balance [35–38]. Many researchers have shown that fatigue negatively affects dynamic postural control [39–42]. Fatigue is an essential factor that acutely affects balance ability. In a study in which the center of pressure (COP) was measured before, in halftime and immediately after a soccer match, it was determined that the balance skill of the support leg was impaired in the post-match measurement [41]. The return of balance ability to average values after fatigue depends on many factors. The return of post-fatigue balance ability to initial level depends on the duration, intensity, and type of intensity of the fatigue protocol performed [43]. Deficits in dynamic postural control is a risk factor in experiencing falls and lower extremity injuries [13,42,44–47]. The deterioration of dynamic balance is associated with reactive and compensating movements, and it is stated that this impairment has been linked to falling risk and lower extremity injuries [18,48,49]. Since fatigue increases the rate of injury in athletes [9,16], and the lack of postural control is a lower extremity injury risk factor [13,50,51], it can be expected that the rate of injury as a result of fatigue-induced postural control (fatigue-induced balance deficits) may increase.

Soccer is classified as both an aerobic and an anaerobic sport. In the game, players may experience fatigue from time to time as a result of aerobic and anaerobic activities. There are studies in the literatüre on the effects of anaerobic fatigue on balance performance in soccer players. However, up to date, no previous studies have examined the effects of both types of fatigue in soccer players. Besides, many activities such as passing, kicking, and jumping in soccer are carried out on the support leg. This research will be the first to examine the effects of fatigue on support leg balance performance. Therefore, this study aims to determine the effects of different fatigue protocols on balance performance of the support-leg in female soccer players and to understand the time required for the balance to recover after loading.

#### **2. Materials and Methods**

Sixteen sub-elite female soccer players (with mean age of 20.19 ± 1.52 years, body mass 56.52 ± 4.95 kg, body height 164.81 ± 4.21 cm, percent body fat 22.63 ± 2.42%, and maxVO<sup>2</sup> 52.33 ± 5.74 mL.kg.−1min−<sup>1</sup> ) participated in the study voluntarily. Players who suffered lower extremity injuries for the last six months were not included in the study. Participants were instructed not to perform exercises that may cause exhaustion 48 h before the tests and not to use stimulants such as alcohol, caffeine, or drugs in the last 24 h before the study. The study was conducted according to the Declaration of Helsinki and was approved by the Ethics Committee of Ankara University, Approval code 21-1300-17, released in December 2017.

#### *2.1. Balance Test*

The participants were invited to participate in the Biodex SD Balance System (Biodex, Shirley, NY, USA) athletic single-leg testing protocol [52]. As noted, the Biodex Balance System (BBS) uses a circular platform that is free to move in the anterior-posterior and medial-lateral axes simultaneously. The BBS measures, in degrees, the tilt about each axis during dynamic conditions and calculates an overall stability index (OSI). A high score in the OSI indicates poor balance. The platform stability ranges from 1–12, with 1 representing the most significant instability.

The familiarization protocol was implemented before the experiments. All participants were instructed to perform the balance test on five different days of the targeted week.

The athletic single leg test protocol consisted of 3 trials of 20 s of upright stance on support-leg with 10 s of rest intervals between trials. Participants were asked to place their feet with the malleolar axis aligned with the midpoint of the platform over the center dot of the platform in a comfortable position. An athletic single-leg test was conducted on BBS with the platform set at level 4. Balance tests were carried out on the non-dominant leg. The same test protocol was performed before (Bpre) and right after (Bpost) both aerobic and anaerobic fatigue protocols, and repeated at the 5th (B5), 10th (B10), 15th (B15), and 20th (B20) minutes. The participants were allowed to rest passively during the 5 min of recovery periods. There was a 2-day period between aerobic and anaerobic fatigue protocols (Figure 1).

#### *2.2. Aerobic Fatigue Protocol*

The Bruce protocol was performed on a motorized treadmill (Cosmed, Rome, Italy) in order to create aerobic fatigue in soccer players [53]. The participants continued the test until they were exhausted. At the end of the test, the maxVO<sup>2</sup> consumption values of the participants were calculated and recorded. Rating of perceived exertion (RPE) was obtained using the 6–12 point Borg scale at the end of every load [54]. MaxVO<sup>2</sup> was defined as the highest 30 s average in oxygen uptake and maximal heart rate (HRmax) as the highest every 10 s average during the Bruce protocol. A test was considered maximal when four of the following criteria were completed: VO2 plateau at peak exercise, respiratory exchange ratio ≥ 1.10 greater age-predicted maximal heart rate (220-age), and an indication of 18–20 rating on the Borg RPE scale [55].

#### *2.3. Anaerobic Loading Protocol*

Anaerobic fatigue protocol was performed using a bicycle ergometer (Monark Erogomedic 894 E Peak Bike Vansbro Sweden). The Wingate test protocol was used for anaerobic fatigue. In the Wingate protocol, participants were asked to pedal at maximal speed for 30 s. As the intensity of the training, a weight equivalent to 7.5% of the participants' body mass was placed on the load scale. Once the participants started pedaling, the scale dropped when the bike's wheel revolution went up to 150 rpm and the maximal pedaling for 30 s. Soccer players were verbally motivated during training. Participants repeated the Wingate test protocol a total of 6 times with intervals of 4 min.

Basel, Switzerland).

Participants' blood lactate values were determined immediately after the fatigue protocol and during the recovery period (Lpre) just before the balance tests at 0th (Lpost), 5th (L5), 10th (L10), 15th (L15), and 20th (L20) minutes. During the lactate test, participants' fingertips were wiped with alcoholbased tissue paper, and their capillary blood samples were taken with a lancet pen. Blood lactate

**Figure 1.** One subject during the data collection phase. Female soccer player. **Figure 1.** One subject during the data collection phase. Female soccer player.

#### *2.5. Statistical Analysis 2.4. Lactate Testing*

In all statistical analyses, SPSS version 20 was used (SPSS Inc., Chicago, IL, USA). First, because the number of participants was below 50, the normality of the data was analyzed with the Shapiro-Wilk test. Depending on the distribution, lactate and balance values obtained at different times following aerobic and anaerobic fatigue protocols were compared by the Paired Sample t-Test or Wilcoxon Test. For the intergroup analyzes, either the Repeated Measurements Analysis of Variance (Aerobic Lpre, L5, L10, L15, L20; Bpre, B5, B15, B20; Anaerobic Lpost, L5, L10, L15, L20; Bpre, B5, B15, Participants' blood lactate values were determined immediately after the fatigue protocol and during the recovery period (Lpre) just before the balance tests at 0th (Lpost), 5th (L5), 10th (L10), 15th (L15), and 20th (L20) minutes. During the lactate test, participants' fingertips were wiped with alcohol-based tissue paper, and their capillary blood samples were taken with a lancet pen. Blood lactate levels of the participants were determined by an Accutrend Plus lactate device (Roche Diagnostics, Basel, Switzerland).

#### B20) or the Friedman test (Aerobic Lpost, Bpost, B10; Anaerobic Lpre, Bpost, B10) was used again depending on the distribution. In the case of the dataset exhibited both non-normally distributed and *2.5. Statistical Analysis*

norma distributed data, we proceeded with the non-parametric analysis. In all statistical analyses, the alpha value was considered to be 0.05. **3. Results**  The following variables were shown to be not normally distributed (Aerobic Lpost, Bpost, B10; Anaerobic Lpre, Bpost, B10). The lactate values obtained from the participants before, during, and after aerobic and anaerobic fatigue protocols are shown in Table 1. In all statistical analyses, SPSS version 20 was used (SPSS Inc., Chicago, IL, USA). First, because the number of participants was below 50, the normality of the data was analyzed with the Shapiro-Wilk test. Depending on the distribution, lactate and balance values obtained at different times following aerobic and anaerobic fatigue protocols were compared by the Paired Sample t-Test or Wilcoxon Test. For the intergroup analyzes, either the Repeated Measurements Analysis of Variance (Aerobic Lpre, L5, L10, L15, L20; Bpre, B5, B15, B20; Anaerobic Lpost, L5, L10, L15, L20; Bpre, B5, B15, B20) or the Friedman test (Aerobic Lpost, Bpost, B10; Anaerobic Lpre, Bpost, B10) was used again depending on the distribution. In the case of the dataset exhibited both non-normally distributed and norma distributed data, we proceeded with the non-parametric analysis. In all statistical analyses, the alpha value was considered to be 0.05.

#### **3. Results**

The following variables were shown to be not normally distributed (Aerobic Lpost, Bpost, B10; Anaerobic Lpre, Bpost, B10).

The lactate values obtained from the participants before, during, and after aerobic and anaerobic fatigue protocols are shown in Table 1.

**Table 1.** Mean comparisons of participants with lactate and balance, which vary depending on aerobic and anaerobic loading. In horizontal, the repeated measures *p* values, in vertical the *p* values for aerobic vs. anaerobic comparisons.


L: Lactate; B: Balance; \* *p* < 0.05; \*\* *p* < 0.01.

According to the results, there was no significant difference between participants' resting lactate concentration values obtained before aerobic and anaerobic loading (*p* > 0.05). However, the lactate values obtained immediately after, 5th, 10th, 15th, and 20th min were statistically significant (*p* < 0.01) according to the fatigue conditions. After anaerobic loading, lactate values were found to be higher. Besides, it was understood that 20 min was not sufficient for the recovery, regardless of the type of loading.

Lactate values obtained at six different phases of the aerobic fatigue protocol were significantly different (*p* < 0.01). The resting lactate value was determined to be lower than all others reached after loading. Also, a significant difference was found between Lpost and L10, L<sup>15</sup> and L20; L<sup>5</sup> and L10, Lpost L15, and L20; L<sup>10</sup> and L<sup>15</sup> and L20 (*p* < 0.01); and with L<sup>5</sup> and L<sup>20</sup> (*p* < 0.05).

In lactate values measured after anaerobic loading, the resting value was determined to be significantly lower than in all other measurements. The difference between Lpost and L<sup>5</sup> was not significant, similar to the aerobic one. However, a significant difference at the level of *p* < 0.01 was found between Lpost and L10, L15, and L20, between L<sup>5</sup> and L10, L15, and L<sup>20</sup> and between L<sup>15</sup> and L20.

When the results related to the balance were examined, a significant difference was found at the level of *p* < 0.01 between the balance values obtained immediately after aerobic and anaerobic loading and at the 5th minute. It was understood that there is more deterioration in the balance after aerobic loading. No significant difference was observed in the values obtained after 10th, 15th, and 20th min, depending on the type of loading (Figure 2).

When the values obtained due to aerobic loading are taken into account, it was understood that the differences between Bpre and Bpost, B<sup>5</sup> (*p* < 0.01), and between Bpre and B<sup>10</sup> were (*p* < 0.05) statistically significant. However, there was no significant difference with the values reached between B15, B20, and Bpre. Accordingly, it can be stated that recovery takes place after aerobic load in the state of balance between 10th and 15th minutes. In the other results, on the other hand, a significant difference was found between Bpost and B5, B10, B15, and B<sup>20</sup> (*p* < 0.01). The measurement of B<sup>5</sup> was found to be significantly higher than the measurements of B10, B<sup>15</sup> and B<sup>20</sup> (*p* < 0.01).

*Diagnostics* **2020**, *10*, x FOR PEER REVIEW 6 of 11

**Figure 2.** Overall stability index across the experimentation. Aerobic condition vs. anaerobic **Figure 2.** Overall stability index across the experimentation. Aerobic condition vs. anaerobic condition.

#### condition. **4. Discussion**

When the values obtained due to aerobic loading are taken into account, it was understood that the differences between Bpre and Bpost, B5 (*p* < 0.01), and between Bpre and B10 were (*p* < 0.05) statistically significant. However, there was no significant difference with the values reached between B15, B20, and Bpre. Accordingly, it can be stated that recovery takes place after aerobic load in the state of balance between 10th and 15th minutes. In the other results, on the other hand, a significant difference was found between Bpost and B5, B10, B15, and B20 (*p* < 0.01). The measurement of B5 was found to be significantly higher than the measurements of B10, B15 and B20 (*p* < 0.01). **4. Discussion**  This study aimed to investigate the acute effects of aerobic and anaerobic exercises on dynamic balance skill and recovery time in female soccer players. In order to control the fatigue level, lactate concentrations of the subjects were also collected. According to the lactate test results, both fatigue protocols were found to be successful in creating fatigue. Although it was higher after the anaerobic exercise, the lactate level did not return to the initial level within 20 min after both fatigue conditions. Besides, as a result of strenuous aerobic or anaerobic exercise, female soccer players' ability to balance the support leg was affected negatively. After both aerobic and anaerobic loading, the recovery time of balance skill lasted about 10 min. The deterioration in the athletic single-leg stability test was observed to be higher after the aerobic fatigue protocols in all of the measurements. Furthermore, the difference was statically significant in Bpost (*p* < 0.01) and B5 (*p* < 0.05) values. In the literature, many studies suggest that aerobic and anaerobic fatigue negatively affect balance ability [41–43,56–61]. There are similar studies on this subject in the literature. In a study in which both aerobic and anaerobic fatigue protocol was implemented, and the balance level was determined with the Balance Error Scoring System (BESS), no difference was found between the balance performance and its time of recovery after both fatigue protocols. However, athletes returned to their initial balance performance values within 8–13 min after both fatigue protocols [58]. Steinberg et al. investigated the This study aimed to investigate the acute effects of aerobic and anaerobic exercises on dynamic balance skill and recovery time in female soccer players. In order to control the fatigue level, lactate concentrations of the subjects were also collected. According to the lactate test results, both fatigue protocols were found to be successful in creating fatigue. Although it was higher after the anaerobic exercise, the lactate level did not return to the initial level within 20 min after both fatigue conditions. Besides, as a result of strenuous aerobic or anaerobic exercise, female soccer players' ability to balance the support leg was affected negatively. After both aerobic and anaerobic loading, the recovery time of balance skill lasted about 10 min. The deterioration in the athletic single-leg stability test was observed to be higher after the aerobic fatigue protocols in all of the measurements. Furthermore, the difference was statically significant in Bpost (*p* < 0.01) and B<sup>5</sup> (*p* < 0.05) values. In the literature, many studies suggest that aerobic and anaerobic fatigue negatively affect balance ability [41–43,56–61]. There are similar studies on this subject in the literature. In a study in which both aerobic and anaerobic fatigue protocol was implemented, and the balance level was determined with the Balance Error Scoring System (BESS), no difference was found between the balance performance and its time of recovery after both fatigue protocols. However, athletes returned to their initial balance performance values within 8–13 min after both fatigue protocols [58]. Steinberg et al. investigated the balance level after a Yo-Yo test with the Interactive Balance System (Tetrax) device and reported that balance skill returned to the initial level within 10 min after fatigue [62]. In the current study, balance skills returned to the initial level after approximately 10–15 min after both aerobic and anaerobic fatigue. Moreover, the present study found no difference between balance levels during the recovery time after aerobic or anaerobic fatigue. In a study conducted on a bicycle ergometer, participants performed two maximal Wingate tests lasting 30 s with a rest interval of 2 min. At the end of high-intensity activation, it was determined that balance skill was affected negatively and returned to the baseline level within 10 min [63]. Ishizuka et al. applied the functional fatigue protocol to 14 male and 9 female college-level soccer players and determined the balance with the Biodex Limit of Stability

balance level after a Yo-Yo test with the Interactive Balance System (Tetrax) device and reported that balance skill returned to the initial level within 10 min after fatigue [62]. In the current study, balance Test. As a result, the subjects were found to have returned their initial level within 10 min after the fatigue [59]. In a study by Matsuda et al., a functional fatigue protocol was applied to 100 recreationally active college students. After the functional fatigue protocol, it was reported in the measurements made with the balance error score system that the balance performance returned to its initial level within approximately 20 min [64]. Although a similar fatigue protocol was implemented, different balance performance recovery times were observed between the study mentioned above and our study. This difference is thought to be since soccer players have a better balance skill than other athletes and sedentary people. Contrary to the result of the current study, it is reported in some studies that the recovery time of balance ability lasts longer than 10 min, while in some studies, it is reported that balance ability is not affected after fatigue. In a study where the fatigue protocol involving sports activities was applied, it was reported that the participants' balance values returned to their initial values after 20 min as a result of the balance test conducted with the balance error score system [56]. In another study, in which balance measurements were made after a 25-min treadmill run, it took approximately 15 min for the athletes participating in the study to return their balance performance to its initial level [65]. In another study, where balance measurement was performed with the biodex balance system before and during a soccer match, it was determined that the dominant leg balance performance of the players was impaired while no change was observed in the support leg balance performance [66]. In contrast to this, soccer players' support leg balance skills were found to be negatively affected following the fatigue protocol in the present study. The reason for the difference is thought to be due to the degree of difficulty differences between the protocols of balance tests. In the literature, some studies did not observe any changes experienced in balance performance after fatigue. After soccer-specific fatigue [67] and after soccer training [68], balance measurements made with the Biodex Balance System have reported that the balance performance of soccer players is not affected by fatigue. Paillard reported that the return of post-fatigue balance ability to initial values depends on the duration, density, and intensity of the fatigue protocol performed [43]. In these studies, it is thought that the reason why there is no difference after fatigue is due to insufficient density and intensity of the fatigue protocol for impaired balance ability. In addition, longer times to return to initial values of balance performance were reported in these studies compared to the current study. The reason for these varied results may have been the difference between the branches of the athletes participating in the studies or the difference in balance measurement methods.

#### **5. Conclusions**

As a result of this study, it was clearly observed that balance performance is impaired in soccer players after both aerobic and anaerobic fatigue. The impairment of fatigue and balance performance are seen as significant risk factors. Although there is not enough data on the effects of fatigue on balance ability in soccer players, it is stated in several studies that fatigue increases the incidence of injury experienced [69] and that deterioration in balance performance may increase ankle injuries [70]. Many researchers also suggest that balance training should be performed to prevent injuries [1,71,72]. Therefore, trainers should give importance to balance training in order to prevent non-contact injuries caused by loss of balance. In future studies, it is suggested to investigate the effects of fatigue on the balance ability in athletes performing balance training.

The study was performed exclusively on healthy young adult female soccer players (who suffer from a higher prevalence of non-contact ACL injuries). A limitation of the study could be seen as a lack of control of the menstrual cycle. In addition, in this study, anaerobic fatigue protocol was performed with a bicycle ergometer and aerobic fatigue with a treadmill. The measurement of balance performance after a real soccer match is thought to provide a clearer picture of the effects of soccer-specific fatigue mechanisms on balance performance.

**Author Contributions:** Conceptualization, Ö.G.; Data curation, D.A. and F.A.; Formal analysis, Ö.G. and F.N.¸S.; Funding acquisition, G.L.; Investigation, D.A.; Methodology, Ö.G., F.A., and F.N.¸S.; Project administration, A.B. and F.N.¸S.; Supervision, G.L. and A.P.; Visualization, A.P.; Writing—original draft, D.A., F.A., and F.N.¸S.;

Writing—review & editing, A.B. and A.P. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

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

#### **References**


© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*International Journal of Environmental Research and Public Health* Editorial Office E-mail: ijerph@mdpi.com www.mdpi.com/journal/ijerph

MDPI St. Alban-Anlage 66 4052 Basel Switzerland

Tel: +41 61 683 77 34 Fax: +41 61 302 89 18

www.mdpi.com ISBN 978-3-0365-4251-5