*Article* **Single- and Multi-Joint Maximum Weight Lifting Relationship to Free-Fat Mass in Different Exercises for Upper- and Lower-Limbs in Well-Trained Male Young Adults**

**Danilo A. Massini <sup>1</sup> , Anderson G. Macedo 1,2 , Tiago A. F. Almeida <sup>2</sup> , Mário C. Espada 3,4,\* , Fernando J. Santos 3,4,5 , Eliane A. Castro 1,2,6 , Daniel C. P. Ferreira <sup>1</sup> , Cassiano M. Neiva 1,2,7 and Dalton M. Pessôa Filho 1,2**

	- merussi.neiva@unesp.br (C.M.N.); dalton.pessoa-filho@unesp.br (D.M.P.F.)
	- <sup>7</sup> MEFE—Metabolism and Exercise Physiology Laboratory, Faculty of Science, São Paulo State University (UNESP), Bauru 17033-360, Brazil
	- **\*** Correspondence: mario.espada@ese.ips.pt; Tel.: +351-265-710-800

**Abstract:** This study aimed to analyze whether the relationship between regional and whole-body fatfree mass (FFM) and strength is related to FFM distribution and area according to limb involvement. Thirty well-trained male young adults underwent one-repetition maximum test (1RM) to assess the strength in arm curl (AC), bench press (BP), seated row (SR), leg press 45◦ (LP45), knee extension (KE), and leg curl (LC). Dual-energy X-ray absorptiometry was used to evaluate FFM. The values for 1RM in AC, BP, and R correlated to FFM in upper limb (R<sup>2</sup> = 0.69, 0.84 and 0.75), without an effect of appendicular mass index (API) or area. For 1RM in KE, the correlation with FFM in lower limb increased with thigh area (R<sup>2</sup> = 0.56), whereas 1RM in LC and LP45 correlation to whole-body FFM increased with API (R<sup>2</sup> = 0.64 and 0.49). The upper limb's FFM may be reliable for indexing the arms and upper trunk strengths, whereas the relationships between FFM and strength in lower limb improve as muscle mass and thigh area increases between subjects.

**Keywords:** muscle strength; resistance exercise; body composition; early adulthood

### **1. Introduction**

Resistance exercise promotes muscular fitness (i.e., an increase in muscle strength and work economy, and improvement in power and speed during daily living or sporting tasks), which is undoubtedly accompanied by physiological and morphological muscle adaptations [1–3]. Nonetheless, muscle adaptation to resistance training requires that variables are planned (choice of exercise, order of exercise, load, volume, rest, frequency, and repetition velocity) to match a specific goal [2,4,5]. Indeed, when dealing with advanced practitioners (i.e., many years of training), further improvements in strength and muscle hypertrophy require the adequate management of training variables (e.g., load, repetition, sets, rest, and motor task) during a single session or throughout planning [1].

The loading in resistance training is operationally defined as the percentage of onerepetition maximum weight lifted (%1RM) in a single- or multi-joint exercise [4,5]. There

**Citation:** Massini, D.A.; Macedo, A.G.; Almeida, T.A.F.; Espada, M.C.; Santos, F.J.; Castro, E.A.; Ferreira, D.C.P.; Neiva, C.M.; Pessôa Filho, D.M. Single- and Multi-Joint Maximum Weight Lifting Relationship to Free-Fat Mass in Different Exercises for Upper- and Lower-Limbs in Well-Trained Male Young Adults. *Int. J. Environ. Res. Public Health* **2022**, *19*, 4020. https:// doi.org/10.3390/ijerph19074020

Academic Editors: Rafael Oliveira and João Paulo Brito

Received: 19 February 2022 Accepted: 24 March 2022 Published: 28 March 2022

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

**Copyright:** © 2022 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/).

are existing protocols for the measurement of the 1RM value [6]; however, these procedures are unreasonable when considering the training routine and planning for advanced practitioners, which include higher %1RM, high training volume (multiple sets, and a high number of repetitions), and high frequency to encompass a variety of single- and multi-joint exercises [1,5]. Alternatively, the monitoring of 1RM in terms of body composition and anthropometry is supported by the assumption that muscle strength increases in association with the modifications of fat-free mass (FFM) and, therefore, also influencing lift performance.

It was previously shown that segmental body area (arm circumference, arm muscle cross-sectional area, and thigh circumference) also makes a significant contribution to strength in highly resistance-trained athletes [7,8]. Moreover, the fewer joints and muscle groups involved in a weight lifting session, the greater the predictive accuracy from variables of body dimensions. However, the power of this relationship is controversial among studies [9,10]. Hortobágyi et al. [9] concluded that individual differences in muscular strength are poorly related to various measures of body size and segmental body dimensions, since correlations between strength vs. body mass, FFM, thigh and arm volume, cross-sectional area, and skinfolds ranged from −0.52 to 0.56 for trained and non-trained subject groups. Conversely, Hetzler et al. [10] evidenced improvements in the estimate of 1RM bench press using the repetitions to failure test with the addition of the arm circumference and arm length.

In the earliest studies reporting the relationship between 1RM values and anthropometric information, the coefficients widely ranged, but were not above 0.9 [8,11–15]. Therefore, when collectively analyzed, most of these previous studies have related sectional and muscle areas, circumference, and body mass to 1RM performance in multi-joint exercises (i.e., bench press and squat), resulting in predictive equations without the same robustness of the estimate as the models considering the submaximal level of muscle strength (i.e., repetition to failure based on a given weight, percentage of body mass, or fixed number of lifts) [16]. However, an improvement in correlation coefficient has been reported when FFM is considered as an independent variable to be related with the strength for exercises engaging single joints and small muscle groups [8,14] regardless of the level of training (i.e., moderate or advanced) of the participants [7,9,11,17].

Information is surprisingly lacking regarding the power of regional composition to monitor the 1RM value, despite findings indicating the influence of physical performance, FFM, and muscle fiber hypertrophy on the ability to lift heavier weight [18,19]. Indeed, if regional body tissue adaptations are considered to be meaningful information, combined with whole-body changes, and with practical (re)considerations for training control and planning across sexes and ages [20], it would be interesting to analyze how the regional composition information may be useful to evaluate the variations in 1RM in exercises regarding muscle mass participation in resistance exercises.

Thus, the objective of this study was to analyze whether regional and whole-body FFM, which are expected to correlate with 1RM in upper- and lower-limb exercises, follow a specific tendency concerning the limb engaged in exercise. In addition, we wondered whether regional FFM influences the change in 1RM values according to the differences in anthropometric and other composition variables between participants. We hypothesized that regional FFM correlation with 1RM values follows a specific trend regarding the limb engaged in the lift movement, therefore presenting a stronger coefficient compared to anthropometric and whole-body FFM variables. In others words, confirmation that strength and FFM are more strongly related at the body region level will demonstrate that muscle force and mass are both parameters of limb enhancement or a decreased ability in lifting exercises. This would support training and rehabilitation plans regarding body region requirements for strength improvements.

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

### *2.1. Participants*

Thirty well-trained male adult volunteers (23.7 ± 5.8 years, 178.7 ± 5.3 cm in height, 78.7 ± 11.3 kg in body weight, and 17.0 ± 5.4% in body fat), with resistance training experience of at least two years and no injury episode during the last six months, provided their written informed consent to participate in this study. Only male young adults participated to avoid the interference of maturation, sex, and aging process on muscle strength, fat-free tissue mass, and bone mineral content among subjects [17,21]. This research was approved by the Local Ethics Committee of the University (CAEE: 19824719.3.0000.5398).

### *2.2. Body Composition*

The dual-energy X-ray absorptiometry (DXA) method (Hologic® model, QDR Discovery Wi®, Beldford, MA, USA) was used to obtain the regional and whole-body composition. The software (Hologic APEX®, Beldford, MA, USA) provided values of FFM (fat-free mass and bone mineral content, in grams) for upper and lower limbs (UL-FFM and LL-FFM), and the submaximal whole-body FFM (WB-FFM, discarding values for the head). Other regional and whole-body composition variables were fat mass (FM), area, and appendicular fat-free mass index (API). The equipment was calibrated following the manufacturer's recommendations by a laboratory technician with experience in these procedures. According to Nana et al. [22], the standardized conditions for DXA scanning are: (i) participants be presented fasted, rested (no exercise), and with no fluid ingestion for at least three hours before the analysis, and (ii) should arrive wearing light clothing, without shoes or carrying any metallic object or body-worn accessories. During the DXA scanning, the participants remained lying in the supine position on the table until the end of the scan, with feet kept together (~15 cm apart) and arms arranged along the side of the trunk (in a midprone position with ~3 cm between the palms and trunk). The same technician adjusted the anatomical points following the manufacturer recommendations. The participants underwent DXA scanning during the first visit.

### *2.3. Strength Measurements*

Tests of 1RM were performed on the following exercises: (1) arm curl (AC), (2) horizontal bench press (BP), (3) seated row (SR), (4) knee extension (KE), (5) leg curl (LC), and (6) leg press 45◦ (LP45). All tests were performed after a non-specific warm-up of 15 min (static stretching, cycling, or running at exercise intensity ≤60% age-predicted maximal heart rate (i.e., HRmax = 220 − age, with age in years). The protocol of the 1RM test followed the recommendations of Mayhew et al. [23]: (1) a specific warm-up preceded the first attempt of the test and was performed with light weights to avoid concentric failure, and up to 8–10 non-maximal repetitions; (2) initial test weight was chosen based on the average rates for the strength of upper- and lower-limbs, according to age, sex, and body mass [6]; and (3) participants performed at least three attempts of one repetition each, with three minutes of rest between each attempt. The weight was increased or decreased from the initial weight by 1.1 to 4.5 kg based on the difficulty of the first lift. The weight that could not be lifted twice (i.e., self-reported inability, or failure in attempt, to perform the second lift) represented the 1RM reference [6,7]. The load value was reported in kilograms (kg). The participants were instructed to perform the movements with the proper technique, following recommendations from Baechle and Earle [24]. Moreover, two visits, separated by 24 h, were scheduled for the completion of all 1RM testing, following the order of small to large muscle groups, intercalating upper- and lower-limb exercises. Thus, AC, KE, and BP were tested in the first visit, and LC, SR, and LP45 in the second visit. Participants were instructed to avoid high-intensity resistance training 48 h before the testing, and to present themselves rested, fasten, and well-hydrated two hours prior to testing.

### *2.4. Statistical Analysis*

The data are reported as mean ± standard deviation, confidence interval (CI95%), and standard error of measurement (SEM). Normality was checked for the muscle strength variables by the Shapiro–Wilk test. The Pearson coefficient (r) was used to test the linear relationship (2-tailed) between maximum observed strength and body composition variables. The stepwise method was used to model the linear relationship between values of 1RM (as the dependent factor) and regional and whole-body composition variables (as independent factors). The input data for muscle strength in UL exercises considered regional and whole-body composition variables, except those for LL, and vice versa when the procedures were applied to analyze the relationship between LL strength exercises and body composition. To ensure that the correlations were not inflated for the differences in muscle area and musculature distribution, the analysis was controlled to segment area (i.e., arm or thigh, according to the body region involved in the exercise) and API (independently of the body region involved in the exercise). The Pearson coefficient was interpreted as <0.2 (trivial), 0.20–0.49 (small), 0.5–0.8 (medium), and >0.8 (strong) [24]. Scatterplots was used to analyze the explained variance (R<sup>2</sup> and R 2 adj) and standard error of the estimate (SEE) of the FFM-predicted 1RM distribution to the observed 1RM distribution across subjects, considering both coefficients as <0.04 (trivial), 0.04–0.24 (small), 0.25–0.63 (medium), and >0.64 (strong) [25]. All statistical procedures were performed in SPSS 26 (Statistical Package for Social Sciences, IBM, Armonk, NY, USA), with a significance level of *p* ≤ 0.05.

The sample power for the associations between the observed and predicted 1RM values were determined for each exercise, and the mean value was considered for analysis of the sample size (*n* = 30). Input parameters were: (a) the corresponding value of "*r*" from the coefficient for explained variance (R<sup>2</sup> ) given in scatterplots; (b) *Zα* = 1.96 for a security index of *α* = 0.05, following Díaz and Fernandéz [26]:

$$Z\_{1-\beta} = \sqrt{n-3}\frac{1}{2}\ln\left(\frac{1+r}{1-r}\right) - Z\_{1-\frac{\kappa}{2}}\tag{1}$$

To avoid anon-realistic statistical power by using information from the actual sample, the cross-validation process was performed using the predicted residual error sum of squares (*PRESS*) method [27,28]. From the *PRESS* statistic, a modified form of R<sup>2</sup> adjusted (R<sup>2</sup> <sup>p</sup>) and standard error of the estimate (SEEp) were recalculated, R<sup>2</sup> <sup>p</sup> = 1 – (*PRESS*/SSTotal) and SEE<sup>p</sup> = (*PRESS*/*n*) 1/2, in which *PRESS* is the sum of the squares of eliminated residuals:

$$PRESS = \sum\_{i=1}^{n} (y\_i - \hat{y}\_{i,-i})^2 \tag{2}$$

### **3. Results**

Table 1 presents regional and whole-body composition characteristics, anthropometric area, and 1RM values of the participants.


**Table 1.** Values of regional and whole-body composition, and muscle strength.

API: appendicular fat-free mass index; FFM: fat-free mass; WB: whole-body; UL upper limbs; LL: lower limbs; AC: arm curl; BP: horizontal bench press; SR: seated row; KE: knee extension; LC: leg curl; LP45: leg press 45◦ ; 1RM: one-repetition maximum; SD: standard deviation; CI95%: confidence interval; SEM: standard error of measurement.

The correlation coefficients between the regional and whole-body composition variables with 1RM values are shown in Table 2. All correlation coefficients for UL- and LL-FFM were observed to be at a higher level than those for API, WB-FFM, and arm and thigh areas, with the exceptions of KE, LC, and LP45, for which the correlations with WB-FFM and LL-FFM were quite similar.


**Table 2.** Coefficients for Pearson's correlation analysis between 1RM values and regional and wholebody composition variables.

API: appendicular fat-free mass index; WB: whole-body; FFM: fat-free mass; UL: upper limbs; LL: lower limbs; AC: arm curl; BP: horizontal bench press; SR: seated row; KE: knee extension; LC: leg curl; LP45: leg press 45◦ . \*\* *p* < 0.001, na: not analyzed.

Figure 1 depicts the scatterplots between values for the 1RM tests. For AC, BP, and SR, the explained variances from UL-FFM (Figure 1A–C) were higher when controlled by API (R<sup>2</sup> = 0.69, 0.84, and 0.75, respectively, (strong), *p* < 0.01). A similar result was observed for the KE variance explained by LL-FFM (Figure 1D), which increased when controlled by the thigh area (R<sup>2</sup> = 0.54 (medium), *p* < 0.01), and for the LC and LP45 variances explained by WB-FFM (Figure 1E and F), which also increased when controlled by API (R<sup>2</sup> = 0.62 (strong) and 0.46 (medium), respectively, *p* < 0.01).

The *PRESS* analysis is presented in Table 3. The stability of the correlations by shrinkage analysis from R<sup>2</sup> adj to R<sup>2</sup> <sup>p</sup> was ensured for all observed correlations between RE and

FFM variables, since the values for R<sup>2</sup> <sup>p</sup> were at a <sup>≤</sup>0.1 ratio from the previous R<sup>2</sup> adj values. Cross-validation was therefore acceptable from R<sup>2</sup> <sup>p</sup> for regression analysis in all resistance exercises. Calculated unbiased estimates of SEE<sup>p</sup> reduced when compared to those SEE shown in Figure 1 for all resistance exercises.


**Table 3.** Cross-validation values from *PRESS* analysis.

AC: arm curl; BP: horizontal bench press; SR: seated row; KE: knee extension; LC: leg curl; LP45: leg press 45◦ ; SEE: standard error of estimate. SEEDif: difference between SEE<sup>p</sup> and SEE. *Int. J. Environ. Res. Public Health* **2022**, *19*, 4020 6 of 10

**Figure 1.** Scatterplots between observed and predicted values for the 1RM tests in bench press (BP) (**A**); arm curl (AC) (**B**); seated row (SR) (**C**); knee extension (KE) (**D**); leg curl (LC) (**E**); and leg press 45° (LP45) (**F**). **Figure 1.** Scatterplots between observed and predicted values for the 1RM tests in bench press (BP) (**A**); arm curl (AC) (**B**); seated row (SR) (**C**); knee extension (KE) (**D**); leg curl (LC) (**E**); and leg press 45◦ (LP45) (**F**).

The *PRESS* analysis is presented in Table 3. The stability of the correlations by shrink-

exercises. Calculated unbiased estimates of SEE<sup>p</sup> reduced when compared to those SEE

AC 0.66 0.63 0.03 5.54 +2.21 BP 0.83 0.82 0.01 8.92 +2.41 SR 0.73 0.70 0.03 12.44 +2.89 KE 0.53 0.43 0.01 21.56 +7.85 LC 0.61 0.55 0.06 13.71 +5.79 LP45 0.44 0.40 0.04 46.93 +1.89 AC: arm curl; BP: horizontal bench press; SR: seated row; KE: knee extension; LC: leg curl; LP45:

FFM variables, since the values for R2p were at a ≤ 0.1 ratio from the previous R<sup>2</sup>

leg press 45°; SEE: standard error of estimate. SEEDif: difference between SEE<sup>p</sup> and SEE.

adj to R2p was ensured for all observed correlations between RE and

**adj R2p Shrinkage SEE<sup>p</sup> (kg) SEEDif (%)**

adj values.

shown in Figure 1 for all resistance exercises.

**R<sup>2</sup>**

**Table 3.** Cross-validation values from *PRESS* analysis.

**Exercise Model Cross-Validation**

age analysis from R<sup>2</sup>

### **4. Discussion**

The aim of this study was to analyze whether regional and whole-body FFM follows a specific tendency concerning the limb engaged in exercise. In addition, we wondered whether regional FFM influences the change in 1RM values according to the differences in anthropometric and other composition variables between participants. The findings from the present study showed that both UL- and LL-FFM are powerful indexes that are related to 1RM measurements for single- or multi-joint resistance exercises engaging upper- and lower-limb actions. Therefore, our findings are aligned with the assumption that resistance training can improve muscle strength, weight lifting capacity, and fat-free body mass [2,18]. However, information on the propensity of regional body composition to analyze muscle strength variance in different weight lifting exercises is still lacking in the literature. Thus, the current study evidenced that 1RM correlations with FFM in upper and lower limbs in exercises involving single- and multi-joint actions increased according to the content of FFM, regardless of the peripheral FFM distribution and thigh area between subjects when considering resistance exercises involving upper and lower limbs (respectively).

In this sense, the way that FFM variables related with 1RM values for UL singleor multi-joint exercises evidenced a higher power for regional than whole-body FFM, regardless of the arm sectional area between subjects. Moreover, the LL-FFM is a relevant variable for 1RM values when considering LL lifting weight capacity. However, the LL-FFM variable did not achieve a higher power than the whole-body FFM for the correlation with all resistance exercises. The FFM peripheral distribution (i.e., API) accounted for the increase in the correlation coefficients for LC and LP45; therefore, the results suggest that the greater the engagement of muscle mass for the execution of the exercise, the less the regional influence of FFM seems to be.

Undoubtedly, monitoring 1RM values based on regional FFM is an alternative way to control the muscle strength variation [8,29,30]. Moreover, a successful maximum lifted weight during a standard 1RM test protocol presumes: (i) movement expertise and engagement, (ii) soreness and injury possibilities, and (iii) changes in the weight lifted with the difference in mechanical demand of similar exercises. These are the greatest constraints for the testing protocol frequency and application to every exercise planned for training [29–31]. Therefore, the power of the interactions between maximum weight lifting capacity with body composition parameters (i.e., body mass, fat-free body mass, regional body area and volume, girth, and width) would provide confident references for 1RM measurements, controlling muscle strength improvements, and organizing or revising the overload during the training in accordance with the previous target weight and exercise volume [7,9,10,20,32–34].

However, the literature has shown conflicting results for assessing 1RM using anthropometric and body composition variables, mainly when it is carried out with subjects with differences in muscle strength. On the one hand, results showing that among trained subjects, anthropometric variables (arm circumference and length) improved the reliability (R<sup>2</sup> changed from 0.87 to 0.90) of 1RM estimation in the bench press [10]. Additionally, the predictive power (multiple regression coefficient, R<sup>2</sup> ) of the anthropometric dimension variables for 1RM estimates ranged from 0.52 to 0.87 for trained subjects [16,30]. Body composition and anthropometry have been related to variations in muscle strength among untrained subjects, but evidence of associations with 1RM were small to medium (Pearson's coefficient ranging from 0.42 to 0.67), mainly for LL and UL multi-joint resistance exercises [9,13,14,16].

Furthermore, 75.7% of the strength assessed in the bench press by trained men can be explained by the variations in the cross-sectional area of the arm, BMI, and fat percentage, with a standard error of 12.1 kg in the prediction [11]. In addition, the strength in the bench press exercise, in populations of both sexes and varying strength levels, showed a high correlation with the variable lean mass (0.77), and moderate correlations with height (0.59), body weight (0.56), arm circumference (0.66), and chest circumference (0.60), although only

the lean mass and submaximal load for 10RM estimated the bench press strength with 97.6% explanatory capacity [8].

Thus, the statement that highly trained athletes exhibit closer relationships between anthropometric dimensions and weight lifted, and probably, the fewer joints and muscle groups involved in a lift, the greater the predictive accuracy of maximum performance by structural proportions [8], remain theories about the association between training development and the responses in the body's dimensions. The results from the current study agree with this statement. Furthermore, we extend this assertion to exercises involving UL, considering that the association was independent of arm area size, but increased with FFM distribution in the upper limb. Moreover, for LL exercises, the control of 1RM values should consider changes in whole-body FFM and its peripheral distribution between subjects.

The lack of research relating regional body composition to 1RM for single-joint resistance exercises, contrast to those analyzing whole-body composition, anthropometry, and sub-maximal lifted weight relationships to 1RM for multi-joint resistance exercises. For example, the estimate of 1RM from a sub-maximal performance at 5RM or 10RM, with R<sup>2</sup> ranging from 0.96 to 0.99, and SEE lower than 6 and 24 kg, respectively, for bench press and leg press, has been widely accepted as the alternative reference to predicted maximal muscle strength [30]. However, even when relying on submaximal muscle strength scores to estimate 1RM, it is well recognized that the same intrinsic determinant, such as sex and training status, can alter the maximum number of repetitions performed at certain fractions of 1RM [20]. Moreover, each type of exercise prescribed in resistance training requires its specific 1RM reference, and submaximal equations were not available to predict 1RM in different single- or multi-joint resistance exercises. Indeed, athletes should not agree to participate in time-consuming test procedures, or non-specific weight lifting, as these may disrupt their training planning.

However, the lack of a comparable sample of subjects to perform cross-validation of the present relationships hindered a better emphasis of the power of regional and whole-body FFM to predict lifting abilities in single- and multi-joint exercises because reproducibility and sensitivity were not evaluated. Nevertheless, the sample power for correlation analysis was above 80%, which is satisfactory to prevent type II errors. Moreover, cross-validation by applying the *PRESS* approach yielded values for R<sup>2</sup> <sup>p</sup> and SEE<sup>p</sup> that were appropriate to strengthen the demonstrated correlations. In addition, the standardized 1RM protocol used in the current study may be a source of underestimation of the maximal strength during the attempt to attained the heaviest load in a single lifting [4]. Despite the possible underestimation of the actual maximal strength, this does not necessarily mean that a heavy load was not attained during the last lifting attempt, and the attained load was therefore ensured to be very close to the maximal one (i.e., >95% 1RM). Nonetheless, the results should, strictly, be applied to the management of 1RM values in subjects who met the following conditions: (a) expertise in the resistance exercise performance mode; (b) engagement in resistance training for at least two years; and (c) UL-FFM, WB-FFM, and arm cross-sectional area as adjustments to the observed correlation values.

### **5. Conclusions**

The current findings evidenced the role of regional fat-free tissue for monitoring the muscle strength development in specific body regions. This demonstrated that regional FFM may be applied to parametrize muscle strength in different resistance exercises for upper and lower limbs, and would explain rates of 81% and 75% for single-joint exercises, respectively. As a suggestion to improve the reliance in these or other indices of regional and whole-body composition, future analysis should focus on how maximal weight lifting relates to fat-free tissue across randomized trials for both sexes, before and after intervention with resistance exercises planned for muscle strength improvements in single- and multijoint exercises separately.

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

**Funding:** The authors would like to thank São Paulo Research Foundation—FAPESP (PROCESS 2016/04544-3) and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior—Brazil (CAPES— Finance Code 001) for the partial financial support. The collaboration of T.A.F.A. and E.A.C was possible thanks to the scholarships granted by the CAPES, in the scope of the Program CAPES-PrInt, process number 88887.310463/2018-00, Mobility number 88887.580265/2020-00 and International Cooperation Project number 88887.572557/2020-00. This research was also funded by Foundation for Science and Technology, I.P., Grant/Award Number UIDB/04748/2020.

**Institutional Review Board Statement:** The study considered the guidelines of the Declaration of Helsinki and was submitted to the Local Ethics Committee of the University (CAEE: 19824719.3.0000.5398).

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

**Data Availability Statement:** The data that support the findings of this study are available from the corresponding and last author (mario.espada@ese.ips.pt and dalton.pessoa-filho@unesp.br), upon reasonable request.

**Acknowledgments:** The authors would like to thank the team of Laboreh (human performance optimization laboratory) for the helpful participation in data sampling, as well as all participants in the University Social Program: Sport Square Gymnasium (PROEX–UNESP-2020).

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

### **References**


### *Article* **Body Composition and Bioelectrical-Impedance-Analysis-Derived Raw Variables in Pole Dancers**

**Giada Ballarin 1,2, Luca Scalfi <sup>2</sup> , Fabiana Monfrecola <sup>2</sup> , Paola Alicante <sup>2</sup> , Alessandro Bianco <sup>2</sup> , Maurizio Marra <sup>3</sup> and Anna Maria Sacco 2,\***


**Abstract:** Few data are available on the body composition of pole dancers. Bioelectrical impedance analysis (BIA) is a method that is used to estimate fat-free mass (FFM) and fat mass (FM), while raw BIA variables, such as the impedance ratio (IR) and phase angle (PhA), are markers of body cell mass and the ratio between extracellular and total body water. The aim of this study was to evaluate the body composition of pole dancers compared to controls, in particular, those raw BIA variables that are considered as markers of muscle composition. Forty female pole dancers and 59 controls participated in the study. BIA was performed on the whole body and upper and lower limbs, separately, at 5, 50, 100 and 250 kHz. The FFM, FFM index, FM and body fat percentage (BF%) were predicted. The bioelectrical impedance indexes IR and PhA were also considered. Pole dancers exhibited higher FFMI and BI indexes and lower BF%. PhA was greater and IRs were smaller in pole dancers than in controls for the whole body and upper limbs. Considering the training level, FFM, whole-body IR and PhA were higher in the professionals than non-professionals. Raw BIA variables significantly differed between the pole dancers and controls, suggesting a higher BCM; furthermore, practicing pole dancing was associated with a greater FFM and lower FM.

**Keywords:** bioelectrical impedance analysis; muscle composition; phase angle; impedance ratio; pole dance

### **1. Introduction**

The evaluation of body composition is crucial not only for assessing nutritional status in the general population but also for athletes for the monitoring of training and performance.

Anthropometry and bioelectrical impedance analysis (BIA) are both field methods that are widely used to assess the body composition of athletes [1]. In particular, BIA is a simple, non-invasive technique that measures the electrical characteristics of the human body, i.e., impedance (Z) and phase angle (PhA) (from those, resistance (R) and reactance (Xc) can also be derived). Total body water (TBW), fat-free mass (FFM) and fat mass (FM) can be estimated by means of predictive equations that include BIA variables and very often other variables, such as age, height and body mass; some equations were specifically developed for athletes. Since these specific equations [2–4] have not been definitively validated, the BIA-derived estimation of body composition should be considered with caution. In particular, the BIA method has an error of 4–8% compared to criterion methods, which could be even more evident in athletes [3]. On the other hand, the BIA estimates of body composition might give some interesting evidence on body composition on a

**Citation:** Ballarin, G.; Scalfi, L.; Monfrecola, F.; Alicante, P.; Bianco, A.; Marra, M.; Sacco, A.M. Body Composition and Bioelectrical-Impedance-Analysis-Derived Raw Variables in Pole Dancers. *Int. J. Environ. Res. Public Health* **2021**, *18*, 12638. https://doi.org/10.3390/ ijerph182312638

Academic Editors: Rafael Oliveira and João Paulo Brito

Received: 15 October 2021 Accepted: 25 November 2021 Published: 30 November 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/).

groupwise basis. Of note, the bioimpedance index at 50 kHz (BI index = height2/Z at 50 kHz) is commonly considered as a logical predictor of FFM and TBW [5,6]. Finally, it should be noted that BIA may be performed on the whole body but also separately for upper limbs and lower limbs (segmental BIA), giving, at least in theory, the chance for evaluating appendicular muscle mass [7–9].

Raw BIA variables, such as the impedance ratio (IR), which is the ratio between Z at high frequencies and Z at low frequencies, and PhA at 50 kHz, are those that are directly measured by an analyzer. Their assessment in sportspeople is motivated by the fact that IR and PhA may be considered as potential markers of both body cell mass (BCM) and the ratio between extracellular water and total body water (ECW/TBW ratio) [10–13]; in other words, both these variables may give information on the electrical properties, as well as the FFM composition and/or muscle composition. IR and PhA were related to muscle strength and physical activity [14,15] as well, while in the first decades of life and elderly people PhA was associated with muscle performance [16,17], isolated or grouped physical fitness indicators [18,19], and cardiorespiratory fitness [20]. As reported in a recent systematic review [21], it is still to be determined to what extent PhA differs between different sports and due to training/untraining; some studies showed that mean whole-body PhA is higher in athletes vs. controls [21,22], while, to the best of our knowledge, so far no data are available on IRs in sportspeople and only limited data exists on segmental BIA [7–9,21,22].

With regard to sports activities, pole dancing is a type of functional training that involves the use of a vertical pole to perform exercises and figures. A training session, called a pole class, lasts between 60 and 90 min (possibly depending on training level) and can be subdivided into three parts: warm-up and strengthening exercises are performed first; then the specific tool figures are studied, with increasing difficulty of execution, while cooldown exercises close the session. Pole dancing may be considered a moderate-intensity cardiorespiratory endurance exercise that, if practiced regularly, leads to a significant increase in aerobic capacity, resistance, flexibility, and motor coordination [23,24].

To the best of our knowledge, only a single study has evaluated the body composition of female pole dancers, attributing an increase in postural strength and stability to the more experienced athletes, but no changes in body composition [25]. Looking at similar sports, rhythmic gymnasts exhibited lower body mass, body mass index (BMI) and skinfold thickness compared to other athletes [26], while gymnasts had a reduced body fat percentage (BF%) compared to controls with the same BMI [27,28]. Dancers had similar BF% but higher levels of FFM and muscle mass than controls, whereas low values of FFM and fat mass (FM) were observed in the case of underweight athletes [29]. Finally, in sedentary women, a choreographed fitness group workout contributed to reducing FM and increasing muscle mass [30].

Against this background, the aim of this cross-sectional study was to evaluate the body composition of pole dancers (non-professional and professional athletes) compared to controls, with a particular interest in the raw BIA variables that are thought to be markers of FFM composition and/or muscle composition. In addition, a segmental BIA evaluation was performed to explore the electrical characteristics of upper or lower limbs.

### **2. Methods**

Forty female pole dancers and fifty-nine control young women participated in the study. Pole dancers were recruited from among those going to two gyms in Naples (a participation rate of 89%) and were non-professional performers (hereafter defined as amateurs) (*n* = 33), who trained 2–4 h a week in two sessions (18–36 months of specific training), and professionals (*n* = 7) who were pole dance trainers (at least 60 months and more than 6 h a week of specific training). Controls (*n* = 59) were sedentary women (at most 1 h of physical training twice a week) and were recruited from among the female students attending the "Federico II" University of Naples. All subjects were healthy. The Ethics Committee of the "Federico II" University of Naples approved the research protocol and subjects gave their informed consent to participate in the study.

The participants avoided physical exercise for 24 h before the measurement session and were studied by the same operator following standard procedures. Data were collected between March and April 2019 in four sessions for pole dancers and six sessions for controls (data on ≥10 women were collected in each session). The general schedule was similar in the two groups of pole dancers, with different intensities of training programs based on their training level.

Body mass was measured to the nearest 0.1 kg using a platform beam scale and height was measured to the nearest 0.5 cm using a stadiometer (Seca, Hamburg, Germany). Participants were asked to remove shoes and heavy clothes prior to weighing. BMI was then calculated as body mass (kg)/height<sup>2</sup> (m<sup>2</sup> ).

Height was measured according to standard procedures. The participants were asked to stand up straight against the backboard with their body weight evenly distributed and both feet flat on the stadiometer platform, while the head was in the Frankfort horizontal plane [31].

Mid-arm circumference and triceps skinfold thickness (Holtain skinfold caliper) were measured on both body sides and, subsequently, the arm muscle area (AMA), corrected for the bone area, and arm fat area (AFA) were calculated as follows [32]:

AMA = [(Mid-arm Circumference − π × TSF) × 2/4π] − 6.5

### AFA = Arm total area − AMA

BIA was performed using a HUMAN IM TOUCH multi-frequency analyzer (DS MEDICA, Milan, Italy) in standardized conditions: ambient temperature between 23–25 ◦C, fast for >3 h, empty bladder and supine position for 10 min. Data on Z at four different frequencies (5, 50, 100 and 250 kHz) and PhA at 50 kHz were considered for the statistical analysis. Precision resistors and capacitors (reference electronic circuits) were routinely used for calibration. The reproducibility of the BIA was previously assessed in ten healthy volunteers on subsequent days with a mean coefficient of variation of 1.5% for Z (at each of the different frequencies considered) and 2% for the phase angle at 50 kHz.

The 250 kHz/5 kHz IR may be used as a proxy marker of fluid distribution and was recently related by our group to mortality in patients with chronic obstructive pulmonary disease [10,14]. Subjects were asked to lie down with their legs and arms slightly abducted (~30◦ ) to ensure no contact between body segments. The measuring electrodes were placed on the anterior surface of the wrist and ankle, and the injecting electrodes were placed on the dorsal surface of the hand and the foot, respectively [13]. Segmental BIA was performed using a six-electrode technique according to Organ [33].

Whole-body BI indexes were calculated as height<sup>2</sup> divided by Z as markers of ECW (Z at a low frequency of 5 kHz) and FFM (Z at high frequencies of 50, 100 or 250 kHz). In addition, two other raw variables were measured for the whole body and upper or lower limbs separately: (1) IR is commonly calculated as the ratio between Z at 200, 250 or 300 kHz and Z at 5 kHz [10]. In the present study, data were obtained for three ratios: Z 50 kHz/Z 5 kHz (IR50/5), Z 100 kHz/Z 5 kHz (IR100/5), and Z 250 kHz/Z 5 kHz (IR250/5). (2) PhA was measured at 50 kHz, as described in the literature. To the best of our knowledge, there has been little interest in applied physiology and human nutrition for evaluating the phase angle at frequencies other than 50 kHz. In all cases, mean values for the dominant (D) and non-dominant (ND) body sides were considered for statistical analysis to give more consistent results for the entire body. FFM was estimated using the Sun equation [34], which is a well-known equation that was proposed for the general population aged 12–94 years and which is also expected to perform well in young women with a higher physical activity level but no very major changes in body composition.

Whole-body FFM was calculated as follows:

FFM = <sup>−</sup>9.53 + 0.69 <sup>×</sup> height2/resistance + 0.17 <sup>×</sup> body mass + 0.02 <sup>×</sup> resistance

where the resistance at 50 kHz was derived by multiplying Z by the cosine of PhA.

Finally, FM was obtained from the difference between body mass and FFM, while the fat-free mass index (FFMI) was calculated as FFM (kg)/height<sup>2</sup> (m<sup>2</sup> ).

### *Statistical Analysis*

Data obtained during the routine examination of athletes or control subjects were retrospectively retrieved. With a type I error rate of 0.05 and a type II error rate of 0.20, a sample size of 85 subjects is required to determine whether a correlation coefficient of 0.3 differs from zero.

Results are expressed as mean ± standard deviation (with some exceptions, see below). Statistical significance was pre-determined as *p* < 0.05. Effect size was calculated according to Cohen [35].

All statistical analyses were carried out using the Statistical Package for Social Sciences (SPSS Inc., Chicago, IL, USA) version 26. One-way analysis of variance was performed to assess the differences between two groups (pole dancers vs. controls or amateurs vs. professionals). Partial correlation was used to assess the relationships between the variables. The general linear model (GLM) was used to assess how several variables affected the continuous variables. From a practical point of view, it was used to compare the body composition between groups after controlling for body mass; adjusted means ± standard errors were provided by this statistical procedure.

### **3. Results**

The general characteristics of the study groups are summarized in Table 1. Despite no difference in body mass and BMI, the pole dancers exhibited lower BF% compared to the controls (−14%). Correspondingly, the AMA was significantly greater and the AFA was smaller in the pole dance than in the control group (Table 1).


Data are expressed as mean ± standard deviation. \* *p* < 0.05. BMI—body mass index. FFM and FM were estimated from the BIA; AMA was corrected for bone area. D—dominant side and ND—non-dominant side of the body. Effect size: Cohen's d ≤ 0.2 = small, 0.2 < d ≤ 0.5 = small to medium, 0.5 < d ≤ 0.8 = medium to large, d > 0.8 = large.

As for the raw BIA variables, the whole-body and upper limb Z values were lower in the pole dancers than in the controls; for instance, Z at 250 kHz was 485 ± 50 vs. 519 ± 38 kHz and 240 ± 28 vs. 271 ± 20 kHz, respectively (d = 0.39 and d = 0.72; *p* < 0.001), with small differences (<2%) between the D and ND body side. Furthermore, Table 2 indicates that the BI indexes at 5, 50, 100 and 250 kHz were higher in the pole dancers than in the controls (+4.3, +4.9, +5.3 and +5.3%, respectively). These differences in the mean values of different Z and BI indexes persisted after adjusting for age and mass (data not shown). After controlling for groups, a partial correlation indicated that whole-body BI indexes

were associated with AMA (r > 0.450 for 50, 100 and 250 kHz vs. r = 0.416 for 5 kHz) but not with AFA.

**Table 2.** Bioimpedance indexes, impedance ratios and phase angles that were measured for the whole body and upper and lower limbs in female pole dancers and controls.


Data are expressed as mean <sup>±</sup> standard deviation. \* *<sup>p</sup>* < 0.05. BI index—bioimpedance index (height2/Z), IR—impedance ratio, PhA—phase angle. Cohen's d ≤ 0.2—small, 0.2 < d ≤ 0.5—small to medium, 0.5 < d ≤ 0.8—medium to large, d > 0.8—large.

As shown in Table 2, PhA was greater in pole dancers than in controls by 3.8% for the whole body (d = 0.39 and *p* = 0.063) and by 10.7% for upper limbs (d = 0.89 and *p* < 0.001), whereas there was no difference for lower limbs. IRs were lower in the pole dance group than in the control group, again more markedly for upper limbs (Table 2). The differences for upper limbs were still found in both cases even after controlling for age and body mass. In particular, multiple regression analysis indicated age and body mass as predictors of IR250/5 (for the whole model: R<sup>2</sup> = 0.117, F(2,87) = 6.83, *p* = 0.002) and PhA (R<sup>2</sup> = 0.053, F(2,87) = 5.90, *p* = 0.017). Of note, no relationships were detected between IRs or PhA and body composition.

There was no significant association of PhA or IRs with height, mass, BMI, FFM, FM, AMA or BI indexes (*p* > 0.20, data not shown). On the other hand, after adjusting for groups, a partial correlation indicated a moderate association between the upper limb and lower limb values of PhA (r = 0.463), IR50/5 (r = 0.538), IR100/5 (r = 0.531) and IR250/5 (r = 0.514).

With respect to the training level, professional and amateur pole dancers did not differ in terms of body mass (55.6 <sup>±</sup> 4.2 vs. 57.3 <sup>±</sup> 7.3 kg) and BMI (22.0 <sup>±</sup> 2.3 vs. 22.2 <sup>±</sup> 2.3 kg/m<sup>2</sup> ). The GLM indicated that, after adjusting for body mass, FFM (mean ± SEM, 45.3 ± 0.6 vs. 43.7 ± 0.3 kg, *p* = 0.024) was greater in the more trained than in the less trained athletes, while BF% was smaller (21.4 ± 11.1 vs. 24.2 ± 0.5%, *p* = 0.023, respectively). In particular, multiple regression analysis was used to test whether training level and body

mass significantly predicted participants' FFM and BF%. The results indicated that the two predictors explained 75% of the total variance for FFM (R<sup>2</sup> = 0.75, F(2,86) = 130.9, *p* < 0.001) and 64% of the total variance for BF% (R<sup>2</sup> = 0.64, F(2,86) = 78.0, *p* < 0.001).

Turning to raw BIA variables, whole-body PhA and IRs were higher, but not significantly (d between 0.5 and 0.8; *p* between 0.05 and 0.10), in the professional athletes than in the amateur athletes (Table 3). More evident differences (Figure 1) emerged for the upper limbs: the professional pole dancers had significantly smaller IRs and greater PhA than the amateur athletes and controls, and the same was true when amateurs were compared to the controls (d = 0.99 and *p* < 0.05). After taking into consideration the training level as a predictor, no significant relationships were found between IRs or PhA and body mass or body composition.


**Table 3.** Bioimpedance index, impedance ratio and phase angle measured for the whole body and upper and lower limbs in amateur and professional pole dancers.

Data are expressed as mean <sup>±</sup> standard deviation. BI index—bioimpedance index calculated as height2/Z. \* *p* < 0.05. Cohen's d ≤ 0.2—small, 0.2 < d ≤ 0.5—small to medium, 0.5 < d ≤ 0.8—medium to large, d > 0.8—large.

**Figure 1.** Impedance ratio Z 250 kHz/Z 5 kHz and phase angle at 50 kHz in amateur or professional pole dancers compared to control women. *\* p* < 0.05 vs. controls *\*\* p* < 0.05 vs. amateurs and **Figure 1.** Impedance ratio Z 250 kHz/Z 5 kHz and phase angle at 50 kHz in amateur or professional pole dancers compared to control women. \* *p* < 0.05 vs. controls \*\* *p* < 0.05 vs. amateurs and controls.

In the present study, raw BIA variables that may be considered as markers of FFM composition and/or muscle composition significantly varied between female pole dancers and controls, showing different electrical characteristics of the body and suggesting

We performed a cross-sectional study on a relatively large group of pole dancers compared to sedentary controls, bearing in mind that the effects of this type of training on

higher BCM; in addition, pole dancers exhibited lower BIA-derived FM and BF%.

controls.

**4. Discussion** 

### **4. Discussion**

In the present study, raw BIA variables that may be considered as markers of FFM composition and/or muscle composition significantly varied between female pole dancers and controls, showing different electrical characteristics of the body and suggesting higher BCM; in addition, pole dancers exhibited lower BIA-derived FM and BF%.

We performed a cross-sectional study on a relatively large group of pole dancers compared to sedentary controls, bearing in mind that the effects of this type of training on body composition have so far been poorly explored [25]. Unfortunately, there was no information regarding participants' body composition before starting the training. Indeed, in light of the difficulties in carrying out long-term intervention studies, the present crosssectional study is expected to provide some preliminary insights regarding the effect of pole dancing on body composition.

Body composition was assessed using BIA, which is a technique that is widely used in athletes [1]. Since the specific equations developed for athletes [2–4] have not been definitively validated [3,13], BIA-derived estimation of body composition should be considered with caution. In particular, the BIA method has an error of 4–8% compared to criterion methods, which could be even more evident in athletes [3]. On the other hand, the BIA estimates of body composition might give some evidence on body composition on a groupwise basis. In the present study, the Sun equation was chosen to predict FFM [34]; this formula was developed in a large sample of healthy subjects using a multicomponent model, it is widely used, and it is expected to also perform well in young women with a higher physical activity level but no major changes in body composition.

Thus, we looked first at BIA-derived estimates of body compartments. Despite having similar body mass and BMI, pole dancers had lower FM and BF% compared to the controls. These findings are in agreement with those reported in previous cross-sectional studies that showed higher FFM and smaller FM in female gymnasts and dancers [26–28]. Of note, the study by Nawrocka et al. [25] on the body composition of pole dancers did not include a control group. Overall, our results suggest a significant, but small effect of pole dance training on body composition, with a moderate to high effect size for BF% (d = 0.74 and *p* = 0.001).

As an alternative approach, IRs and PhA (for the whole body and upper and lower limbs, separately) were directly (no predictive equations used) determined in pole dancers and controls as a qualitative approach to body composition analysis [13]. Both those raw BIA variables may be effective in exploring FFM composition and muscle composition in terms of the electrical characteristics of tissues, as well as BCM and the ECW/TBW ratio [10–13]. Interestingly, IRs and PhA have also been associated with muscle strength and physical activity [14,15,19]. A few cross-sectional studies showed that mean wholebody PhA is higher in athletes vs. controls, while, to the best of our knowledge, no data so far are available on IRs [21]; of note, a recent paper showed, as expected, a high correlation between IRs and PhA [19]. In addition, it is still to be determined to what extent IR and/or PhA may vary between different sports and with training/untraining [13,21]. Facing this background, although in our experience data on IR or PhA are very reproducible, the use of these BIA variables in longitudinal studies or single athletes should be better defined and considered with caution.

IR is commonly calculated as the ratio between Z at high frequency and Z at low frequency [10]. The ratio between Z at 200 kHz and Z at 5 kHz (IR200/5) is widely used but still not formally indicated as the only one to be taken into consideration. Results on three different IRs are reported here, with IR250/5 being very close to IR 200/5. The three IRs were all slightly smaller in the pole dance group compared to the control group. At first glance, these differences in IRs were small in percentage terms, but relevant when compared to the corresponding standard deviations. For instance, the difference in IR250/5 was 0.007, while the pooled standard deviation was 0.019 (d = 0.39 and *p* = 0.058). Regarding another raw BIA variable, whole-body PhA, which was measured at 50 kHz, as commonly described in the literature [10,21], was only slightly higher in pole dancers compared to controls (low size effect). Overall, only minor changes were observed for the whole body.

It is clear that segmental BIA, as performed on upper limbs and lower limbs separately, may give, at least in theory, the chance to evaluate appendicular muscle mass more directly [7–9]. Few previous papers have performed this type of measurement in athletes; they, for instance, showed greater PhA for both lower and upper limbs in female volleyball players compared to controls [8]. Our study yielded some results of interest: lower limb IRs and PhA did not differ between the groups, while a marked difference emerged for upper limbs (d = 0.72 and *p* < 0.001 and d = 0.89 and *p* < 0.001, respectively), suggesting some effects of pole dancing on different muscle groups. Of note, those differences persisted after adjusting for age plus body mass or plus body composition. Thus, segmental measurement seemed to be effective in detecting differences in raw BIA variables, which should be examined in detail by further studies that consider various types of training and use different criterion methods for assessing body composition.

Even if the interpretation of data on professional dancers (Table 3) should be discussed with caution due to the limited sample size, some stimulating findings emerged: compared to amateurs, they had lower IRs and higher PhA for the upper limbs, suggesting a possible relationship between workout volume and the electrical characteristics of muscle. In addition, smaller IRs and greater PhA for upper limbs were still observed in amateur athletes compared to the controls (Figure 1).

Athletes and controls were studied in standardized conditions by a single experienced operator, while BIA was performed on both body sides to ensure a more reliable assessment of the electrical characteristics of the body. A large proportion of the pole dancers going to two different gyms participated in the study, while control women were selected among those who were enrolled in a study on university students who did low amounts of physical activity.

Indeed, there are limitations to the study that should be considered. It was a singlecenter cross-sectional study in which body composition was evaluated by means of a field method. Furthermore, we specifically focused on the assessment of raw BIA variables, such as IR and PhA, that are markers of FFM composition or muscle composition and cannot easily be compared with a proper criterion technique. In addition, there was no information regarding participants' body composition before starting the training, and it was not possible to carry out a very accurate evaluation of the strengthening or conditioning workouts.

### **5. Conclusions**

In conclusion, care must be taken not to overinterpret the results of the present study. The main findings were that raw BIA variables that may be considered as markers of FFM composition or muscle composition significantly differed between female pole dancers and controls, suggesting higher BCM, as well as a lower ECW/TBW ratio; in addition, practicing pole dancing is associated with lower FM and BF%.

Differences in PhA and IRs may suggest modifications in the electrical characteristics of the body that seem to be more marked for the upper limbs and possibly in professional than amateur athletes and that was similar for the three IRs considered. These findings are in line with the literature describing changes in raw BIA variables and body composition due to regular physical exercise [8,9,21,22]. Further studies, especially intervention studies, are needed to define the best approach to use BIA in order to measure raw BIA variables and possibly track changes in the body composition of athletes with time.

**Author Contributions:** G.B., F.M. and A.B. collected the data. G.B., P.A. and M.M. analyzed the data. G.B., L.S., A.M.S. and M.M. designed the study and wrote the manuscript. A.M.S. supervised the project. All authors discussed the results and commented on the manuscript. 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 Ethics Committee of the "Federico II" University of Naples (N: 42/17) approved the research protocol and the study was conducted in accordance with the Declaration of Helsinki.

**Informed Consent Statement:** All participants gave written informed consent prior to being enrolled in the study.

**Data Availability Statement:** The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

**Acknowledgments:** The authors would like to thank Giuseppe Abate for his technical assistance.

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

### **References**


MDPI St. Alban-Anlage 66 4052 Basel Switzerland www.mdpi.com

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

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

Academic Open Access Publishing

mdpi.com ISBN 978-3-0365-9394-4