*Article* **Different Effects of the COVID-19 Pandemic on Exercise Indexes and Mood States Based on Sport Types, Exercise Dependency and Individual Characteristics**

**Alireza Aghababa <sup>1</sup> , Georgian Badicu <sup>2</sup> , Zahra Fathirezaie 3,\*, Hadi Rohani <sup>4</sup> , Maghsoud Nabilpour <sup>5</sup> , Seyed Hojjat Zamani Sani <sup>3</sup> and Elham Khodadadeh <sup>3</sup>**


**Abstract:** Exercise indexes have been affected by the coronavirus disease 2019 (COVID-19) pandemic and its related restrictions among athletes. In the present study, we investigated the exercise frequency and intensity before and during the COVID-19 pandemic, and also current exercise dependency and mood state among non-contact individual, contact individual, and team sports athletes. A total of 1353 athletes from non-contact individual sports athletes (NCISA), contact individual sports athletes (CISA) and team sport athletes (TSA) participated; 45.4% of them were females that completed a series of self-rating questionnaires covering sociodemographic information, former and current exercise patterns, exercise dependency and mood states. NCISA had less exercise frequency than CISA, both before and during the COVID-19 pandemic, and NCISA had less exercise frequency than TSA during the COVID-19 pandemic. Regarding exercise intensity, CISA had higher scores than NCISA and TSA before the COVID-19 pandemic, and CISA had more exercise intensity than TSA during the COVID-19 pandemic. Frequency and intensity were reduced from before to during the COVID-19 pandemic in the three groups, except for TSA intensity. In addition, positive and negative mood states were correlated with exercise dependency. CISA were more discouraged and vigorous than NCISA and TSA, respectively. For NCISA, CISA, and TSA, ordinal regressions separately showed that adherence to quarantine and exercise dependency were better predictors of exercise indexes. Finally, exercise dependency subscales were different among sports, but it was not in exercise dependency itself. Although the decrease in exercise indexes was noticeable, there was no consistent pattern of change in exercise behavior in all sports. Additionally, during the COVID-19 pandemic, negative moods were predominant among all athletes. The results discussed are based on exercise nonparticipating, sport type, and affect regulation hypothesis.

**Keywords:** exercise indexes; exercise dependency; COVID-19 pandemic; team sports; individual sports

#### **1. Introduction**

Psychological and social pressures such as economic problems and illness may always be around us and can lead to some changes in our lifestyle [1,2]. Of course, in addition to stress, the perception of eustress can also have better effects. In fact, what we perceive

**Citation:** Aghababa, A.; Badicu, G.; Fathirezaie, Z.; Rohani, H.; Nabilpour, M.; Zamani Sani, S.H.; Khodadadeh, E. Different Effects of the COVID-19 Pandemic on Exercise Indexes and Mood States Based on Sport Types, Exercise Dependency and Individual Characteristics. *Children* **2021**, *8*, 438. https://doi.org/10.3390/ children8060438

Academic Editor: Filipe Manuel Clemente

Received: 19 April 2021 Accepted: 21 May 2021 Published: 24 May 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/).

from external events affects our perception, behavior, and life. The emergence of the coronavirus disease 2019 (COVID-19) is one of these stressors that has greatly affected economic, social, and even personal life. Of course, different strategies have been taken to prevent infection, and some suggestions have been made. In this regard, although the importance of maintaining exercise in all its dimensions, such as physiological or psychological effects, is recommended, the closure of sports centers and the possibility of air pollution in these spaces may have reduced the amount of physical activity (PA) during lockdown [3], and may, in turn, induce numerous health problems such as stress, depression, and anxiety related to the confinement and prolonged periods of inactivity [4]. Recent findings have shown this in various countries [5]; however, a limitation of previous studies was that apparently no distinction was made between individual and team sports athletes. For the following reason, this is critical: compared to individually exercising athletes, it is conceivable that team sports athletes decreased their exercise levels more rigorously when compared to exercise levels before the lockdown, because of social distancing. Additionally, limited training group sizes might have impacted team sports athletes more severely. Thus, the first aim of the present study was to investigate the differences in exercise frequency and intensity of individual and team sports athletes before and during the lockdown.

Decreased exercise in addition to physiologically destructive effects can also have psychological effects, although recent studies have shown a decreased PA during the COVID-19 pandemic and mood swings [6], it seems that these effects may be more in people who did collaborative in team sports compare to individual sports. Research has shown that people who do individual sports have different self-regulatory skills [7], coping strategies [8], some individual characteristics [9,10], and personality characters [11] from people who do team sports. Additionally, it has been claimed that team sport athletes are at high genetic risk of severe COVID-19 [12]. Therefore, the second aim of this study was the investigation of individual and team sport athlete's mood states during the COVID-19 pandemic. For a deeper understanding of the issue, we investigated contact and non-contract individual and team sport athletes.

In addition to the study of exercise frequency and intensity, we also predicted them by other influential factors. It seems that craving for doing a sport, which is known as exercise dependency, along with individual factors such as age and gender, are some possible predictors of exercise indexes during the COVID-19 pandemic. Therefore, the third aim of this study was predicting exercise frequency and intensity by individual characteristics including exercise dependency, age, gender, and adherence to quarantine.

Although the positive relationship between higher expert-paced PA intensity levels and mood states have been shown [13–15], it seems that a lockdown-related change in PA levels is associated with mood, which, in turn, is influenced by the type of physical activity (team or individual sport) and exercise dependency in individuals. As previous studies have shown [16], there seemed to be a conceptual relationship between mood states during the COVID-19 pandemic and exercise dependency in this study; therefore, we investigated possible relationships between them among all groups.

Mental health conditions among the general and professional populations were reported by cross-sectional [17] and longitudinal [18] studies during the COVID-19 pandemic, but there currently is not clear evidence on the possible effect of the dependence of exercise on the other athlete's behavior which, in turn, could affect PA. Therefore, in this study, we investigated non-contact individual sport athletes (NCISA), contact individual sport athletes (CISA), and team sport athletes (TSA) in terms of exercise indexes, exercise dependency, and mood states.

To the best of our knowledge, no study to date has evaluated the differences in frequency and intensity of PA, exercise dependency and positive/negative mood states, and their possible differences among individual and team sport athletes during the COVID-19 pandemic. Therefore, we wanted to find out the difference in exercise characteristics and mood states among contact and non-contract individual and team sport athletes. Additionally, we investigated the possible relationship between exercise dependency and

positive/negative mood states. Finally, we looked to find out whether exercise frequency and intensity could be predicted by exercise dependency, age, gender, and adherence to quarantine during the COVID-19 pandemic.

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

#### *2.1. Procedure*

Individual and team sports athletes were approached via social network sites (SNS) to participate in the present online study, and they were asked to fill out a questionnaire package on exercise frequency and intensity before and during the COVID-19 pandemic, mood states, and exercise dependency from 5 March to 30 April 2020. Before starting, the objectives of the research, the anonymous data gathering techniques, the confidential data handling practices, and the ethical approval of the study were explained to the participants on the first pages of the study. Next, participants accepted informed consent by clicking a box of agreement. Additionally, the Human Research Ethics Board at the Sport Sciences Research Institute of Iran approved the study (approval ID: IR.SSRC.REC.1399.070), which was performed in accordance with the last revision of the Declaration of Helsinki [19].

#### *2.2. Participants*

A sample of 1353 Iranian athletes with 45.4% females participated in this study. They included non-contact individual sportspeople (skating, *n* = 95; fitness and body building, *n* = 165; swimming, *n* = 80; gymnastics, *n* = 70) with athletes whose mean age was 26.8 years (SD = 10.53 years); contact individual sport athletes (karate, *n* = 85; taekwondo, *n* = 87; judo, *n* = 94, wushu, *n* = 76; boxing, *n* = 91; wrestling, *n* = 68) whose mean age was 23.76 years (SD = 9.86 years); and team sport athletes whose mean age was 24.79 years (SD = 10.41 years) (football, *n* = 102; futsal, *n* = 85; volleyball, *n* = 98; handball, *n* = 115; and basketball, *n* = 110).

#### *2.3. Measures*

#### 2.3.1. Exercise Level

Exercise levels were measured by inquiring about the type, frequency, and intensity of exercise (from low to very high) before and during the COVID-19 pandemic, which was extracted from the 5-item PA questionnaire developed based on Cho's study [20]. The reliability and validity of this tool have been confirmed by Cho [21]. The first question was related to the type of activities in which the athletes participated before/during the COVID-19 pandemic. Individual and team sports were the main sports of athletes before the COVID-19 pandemic; an open-ended question was asked about physical activities during the COVID-19 pandemic. The second question was "before/during the COVID-19 pandemic, how often do you participate in the activity?" The choices were "every day, 6 days/week, 5 days/week, 4 days/week, 3 days/week, 2 days/week, 1 day/week and anytime". Additionally, the last question was "how intensely do you participate in the activity before/during the COVID-19 pandemic?" The choices for intensity were "light, moderate hard and very hard".

#### 2.3.2. Mood State

To evaluate positive and negative mood states, we used a shortened version of the Brunel Mood Scale (BRUMS) [22,23]. The questionnaire included items related to 16 mood states. In the mood test, the participants were asked to express their current feelings according to the instructions. Each response was scored on a five-point scale (ranging from 0 = no to 4 = extremely). The internal consistency values (Cronbach's alpha) of all dimensions and the total scale ranged from 0.82 to 0.96 [24], while in the present study the total scale was 0.90.

#### 2.3.3. Exercise Dependency

The Exercise Dependency measure was measured via 16 items on a seven-point Likert scale. It included the following five factors: expected positive consequences, interference with social life, health, withdrawal symptoms, and exercise as a possibility to compensate for psychological problems. This scale has already been used by previous researchers and validated with internal consistency (α = 0.643–0.808) and fitted the model [25].

Additionally, questions regarding some individual characteristics such as age and social measures concerning the COVID-19 pandemic were asked, such as adherence to quarantine, type and duration of applied confinements, social distancing, and lockdown of gyms, outdoor sports centers and parks.

#### *2.4. Statistical Analysis*

Differences in exercise indexes among athletes of different sports were analyzed by Kruskal–Wallis one-way analysis of variance for between-group effects, and Mann–Whitney U test for within-group effects. Additionally, multivariate analysis of variance was used to investigate negative and positive mood states among different sport groups during the COVID-19 pandemic. In addition, the relationship of exercise dependency with mood states was analyzed by Spearman's correlation coefficients. Additionally, ordinal regressions were used to predicting exercise frequency and intensity by exercise dependency, age, gender, and adherence to quarantine among different sport groups. Finally, one-way analysis of variance and MANOVA were used to analyze exercise dependency and its subscales among different sport groups. The level of significance was set at alpha < 0.05. All statistical analyses were computed utilizing IBM Corp. Released 2015. IBM SPSS Statistics for Windows, Version 23.0. Armonk, NY, USA: IBM Corp and Microsoft Excel (2013).

#### **3. Results**

#### *3.1. Descriptive Statistics of Studied Variables*

Table 1 shows the mean or median scores of exercise frequency and intensity before and during the COVID-19 pandemic, as well as mood states, exercise dependency, and their subscales.

**Table 1.** Descriptive statistics of studied variables.


#### *3.2. Exercise Indexes of Different Sport Athletes*

Kruskal–Wallis one-way analysis of variance showed that NCISA had less frequent exercise than CISA, both before and during the COVID-19 pandemic, and NCISA had less frequent exercise than TSA during the COVID-19 pandemic. Regarding exercise intensity, CISA had higher scores than NCISA and TSA before the COVID-19 pandemic, and CISA had higher exercise intensity than TSA during the COVID-19 pandemic (Table 2).


**Table 2.** Group differences of exercise frequency and intensity before and during COVID-19.

NCISA, non-contact individual sport athletes; CISA, contact individual sport athletes; TSA, team sport athletes; \* *p* ≤ 0.05.

In addition, Mann–Whitney U test for within-group effects showed that the frequency and intensity were reduced from before to during the COVID-19 pandemic in the three groups, except for TSA intensity (Tables 1 and 3). However, this latter variable was not significant.

**Table 3.** Intergroup changes in exercise frequency and intensity from before to during COVID-19.


NCISA, non-contact individual sport athletes; CISA, contact individual sport athletes; TSA, team sport athletes; \* *p* ≤ 0.05.

#### *3.3. Mood States of Athletes during the COVID-19 Pandemic*

Descriptive analyses of mood states (12 items for negative mood states and 4 items for positive mood states) among NCISA, CISA, and TSA are shown in Figure 1.

MANOVA showed that there were significant differences in negative mood states among different athletes (F = 1.66, *p* = 0.022, Wilks' Lambda = 0.968, Partial Eta Squared = 0.016). Thus, the pairwise comparisons showed that CISA were more discouraged than NCISA (mean differences = 0.253, *p* = 0.010). Additionally, positive mood state analysis showed that there were significant differences among different sports (F = 2.10, *p* = 0.032, Wilks' Lambda = 0.986, Partial Eta Squared = 0.007). The results showed that CISA were more vigorous than TSA (mean differences = 0.243, *p* = 0.005). However, there were no significant differences among other negative or positive mood states (all *p*-values > 0.05).

**Figure 1.** Mood states of different groups of athletes during the COVID-19 pandemic.

#### *3.4. Correlation Coefficients of Exercise Dependency, Negative and Positive Mood States*

The Pearson's correlation coefficients also showed that there was a positive significant relationship between exercise dependency with positive mood states (r = 0.198, *p* = 0.0001) and a negative significant correlation with negative mood states (r = −0.077, *p* = 0.008), but their effect size was very small (4% and 0.5%, respectively).

#### *3.5. Predicting of Exercise Indexes by Exercise Dependency, Age, Gender, and Adherence to Quarantine in Different Sport Groups*

Additionally, we investigated whether exercise frequency and intensity could be predicted by exercise dependency, age, gender, and adherence to quarantine among different sports. For NCISA, CISA, and TSA, ordinal regressions showed that adherence to quarantine and exercise dependency were the best predictors. (Table 4).

−


**Table 4.** Ordinal regression parameters of NCISA, CISA, and TSA during the COVID-19 pandemic.

NCISA, non-contact individual sport athletes; CISA, contact individual sport athletes; TSA, team sport athletes; \* *p* ≤ 0.05.

**−** Parameter estimates showed that exercise frequency could be predicted by exercise dependency (in the NCISA and CISA), adherence to quarantine (in the CISA and TSA), and age (in the CISA). In addition, exercise intensity could be predicted by exercise dependency (CISA), adherence to quarantine (in the NCISA and CISA), and gender (in the CISA).

#### *3.6. Perceived Exercise Dependency and Its Subscales among Sports Groups during COVID-19*

≤ Finally, one-way analysis of variance showed that there was no significant difference in exercise dependency (total score) among different sport groups (F = 2.08, *p* = 0.132). However, MANOVA (Wilk's Lambda value = 0.978, F = 2.64, *p* = 0.003) showed that CISA had more withdrawal symptoms than NCISA and TSA (mean differences = 0.75, 1.17; *p* = 0.049, 0.002, respectively). Additionally, regarding health status, NCISA had higher scores than TSA (mean differences = 0.94, *p* = 0.006).

#### **4. Discussion**

This survey reports some data from online research among Iranian individual and team sport athletes during the COVID-19 pandemic. A total of 1353 athletes from individual and team sports participated in this study. Changes in exercise before and during the COVID-19 pandemic showed that the intensity and frequency of exercise were higher before the COVID-19 pandemic than during it among all groups, apart from the intensity metric of TSA. Interestingly, intensity of TSA was higher during the COVID-19 pandemic but was not significant. These results are in line with most studies among the general population [5,26–28]. The scientific community has highlighted the real benefits of PA during the pandemic [29,30]; however, our results showed a reduction in exercise indexes among NCISA and CISA and exercise frequency among TSA. It seems that most team athletes may have performed home exercise, fitness, stretching, walking, and individual exercises at different intensities during this pandemic. Although the frequency of exercise decreased among them, in-depth checking of the present study results showed that the type of exercises had changed, which complied with exercise and PA recommendations during the coronavirus outbreak [31]. This experience could be due to essential changes in sports training schedules. Sports training was discontinued at the original coronavirus outbreak and research data were collected at that time; therefore, the present findings can be justified. This issue was consistent with Lim [32], who cited that many athletes attempt to maintain active lifestyles by themselves. Therefore, changes in daily life activities are necessary. It seems TSA may push themselves to keep fit and stay in shape by doing some other types of PA and exercise. It should also be noted that at the time of data collection of the present study, all places of sports activities, both individual and team, were closed, except for unorganized individual activities outdoors. However, at the same time, there were no nationwide closures for parks, shopping malls, etc., which may have affected the obtained results.

To address this issue, Mutz and Gerke [33] reported a significant decline in sport and exercise activities among Germans. Additionally, about one-third of the studied population reduced their sport and exercise activities, and only 6% intensified sport and exercise levels. They cited that this last group increased home-based workouts and outdoor endurance sports, while others did not adapt their sporting routines to the present conditions.

Based on Figure 1, all the positive moods were less than the negative ones among all groups. However, our results showed that CISA were more discouraged and vigorous. It seems that the type of sport could affect negative and positive mood states; sports in which the participants necessarily come into bodily contact with one another seemed to be more affected. Results of negative moods are consistent with recent research which indicates that the increasing menace of the epidemic has resulted in depression due to disrupted travel plans, social isolation, and media information overload [34].

In addition, positive relationships of exercise dependency with positive mood states and negative relationships with negative mood states were in line with a previous study [16]. They showed small to moderate correlations between exercise dependence with mood states. Previous research suggested that mood states could play a critical role in the development or the maintenance of exercise dependency [16,35,36]. The "affect regulation" hypothesis could justify this issue [36]. Therefore, PA results in improvements in positive mood states and decreases in negative mood states. As the exercise cycle continues, increased amounts of exercise are needed to experience improvement in affect and mood. On the other hand, lack of exercise can lead to a weakening of positive moods and an increase in negative moods.

In addition, exercise indexes were predicted by age, gender, adherence to quarantine, and exercise dependency in the three groups. Adherence to quarantine and exercise dependency were the best predictors of exercise frequency and intensity. Additionally, age and gender were able to predict the frequency and intensity of exercise in the CISA group.

It was further shown that CISA had unpleasant feelings of leaving exercise, and TSA had an unhealthier status than NCISA. Therefore, it seems that the nature of how an athlete interacts with other athletes may be important in the context of athletes' feelings during the COVID-19 pandemic.

Despite the new findings, several limitations warn against overgeneralization of the results. Firstly, the cross-sectional design of the study precludes conclusions about the studied variables. Secondly, although we used open- and closed-ended questions in this study, data collection through self-reporting may be biased. Thirdly, the present study data were collected only a few months after the onset of the COVID-19 pandemic; however, longitudinal studies seem to better clarify the changing trends of exercise indexes.

#### **5. Conclusions**

Among a large sample of Iranian athletes of different ages, gender, and sports, changing exercise indexes were not similar among groups; there was a dominant reduction pattern among all sports, and a non-significant increasing trend was also observed in team sports. Unlike previous studies, the present project did not focus only on the overall scores of negative or positive mood; it also presented new findings related to negative and positive mood subscales. Additionally, in this study, we showed that exercise dependency has a significant relationship with both positive and negative mood states. Finally, it was shown that with increasing exercise dependency, exercise intensity and frequency increased. Additionally, adherence to quarantine and exercise dependency were the best predictors of exercise indexes during the COVID-19 pandemic.

**Author Contributions:** Conceptualization, A.A. and H.R.; methodology, M.N.; software, Z.F.; S.H.Z.S. and E.K.; validation, A.A. and H.R.; formal analysis, S.H.Z.S.; investigation, Z.F.; resources, Z.F. and E.K.; data curation, S.H.Z.S.; writing—original draft preparation, Z.F.; writing—review and editing, A.A., G.B., H.R. and S.H.Z.S.; visualization, A.A. and G.B.; supervision, A.A.; project administration, A.A. and H.R. 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 Human Research Ethics Board at the Sport Sciences Research Institute of Iran approved the study (approval ID: IR.SSRC.REC.1399.070), which was performed in accordance with the seventh and current revision of the Declaration of Helsinki.

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study. Individual and team sports athletes were approached via social network sites (SNS) to participate in the present online survey on past and current exercise patterns, mood states and exercise dependency. On the first page of the online survey, participants were informed about the aims of the study, the anonymous data gathering, the confidential data handling, and the ethical approval of the study. Next, to provide informed consent, participants clicked a box of agreement.

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

**Acknowledgments:** We thank the Sport Sciences Research Institute of Iran for supporting this project. Additionally, the authors are grateful to all athletes who took part in the present study.

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

#### **References**


## *Article* **Somatotype, Accumulated Workload, and Fitness Parameters in Elite Youth Players: Associations with Playing Position**

**Hadi Nobari 1,2,3,4,\* , Rafael Oliveira 5,6,7 , Filipe Manuel Clemente 8,9 , Jorge Pérez-Gómez <sup>2</sup> , Elena Pardos-Mainer <sup>10</sup> and Luca Paolo Ardigò 11**


**Abstract:** The purpose of this study was three-fold: (1) to describe anthropometric, maturation, and somatotype differences of players based on playing positions; (2) to analyze variations of accumulated load training (AcL) and fitness parameters between playing positions; and finally (3) to explain the variation of maximal oxygen uptake (VO2max) and peak power (PP) through the AcL, body fat (BF), maturity, somatotype and fitness levels. Twenty-seven male youth soccer players under-16 were divided by the following positions participated in this study: six central midfielders, four wingers (WG), five forwards, eight defenders, and four goalkeepers (GK). They were evaluated on two occasions: pre-season and after-season. Height, sitting height, body mass, BF, girths, percentage of BF (BF%), lean body mass, maturity, somatotype, sprint test, change of direction test, Yo-Yo intermittent recovery test level 1, Wingate, PP, VO2max and fatigue index were assessed. Then, AcL was monitored during training sessions. The main results revealed significant differences between player positions for maturity offset (*p* = 0.001), for BF (*p* = 0.006), BF% (*p* = 0.015), and lean body mass kg (*p* = 0.003). Also, there were significant differences for AcL and fatigue index in pre-season between player positions (*p* < 0.05). In addition, there were some significant differences in pre- and after-season for VO2max and PP between player positions (*p* < 0.05). In conclusion, GK showed higher values in anthropometric, body composition variables and maturity offset compared to the other positions, while WG presented lower levels of BF. In pre-season, there were more differences by player positions for the different variables analyzed than after-season that reinforces the tactical role of the positions, and the emphasis in increased load in the beginning of the season. This study could be used by coaches, staff, and researchers as a reference for athletes of the same sex, age, and competitive level.

**Keywords:** VO2max; anthropometric; body composition; maturation; peak power; training load

#### **1. Introduction**

Soccer has specific requirements in different competitive levels, playing positions and age categories [1]. Soccer is multifactorial and conditioned by multiple variables such as

**Citation:** Nobari, H.; Oliveira, R.; Clemente, F.M.; Pérez-Gómez, J.; Pardos-Mainer, E.; Ardigò, L.P. Somatotype, Accumulated Workload, and Fitness Parameters in Elite Youth Players: Associations with Playing Position. *Children* **2021**, *8*, 375. https://doi.org/10.3390/children 8050375

Academic Editor: Zoe Knowles

Received: 20 March 2021 Accepted: 6 May 2021 Published: 10 May 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/).

anthropometric, body composition, somatotype, physical, physiological, and soccer-specific skills [2,3]. In this sense, scientific research regarding these topics has been developed, but still provides inclusive information [4]. With special regard to young age categories, there is a need to identify the differences between young soccer players, as any category can include different chronological and biological ages [5]. Therefore, anthropometry, somatotype and some fitness parameters are necessary to know the actual state of the player.

The identification of the somatotype helps to individualize exercise training programs which can differ by positions. Furthermore, this identification facilitates an understanding of the differences in adiposity level, robustness and musculoskeletal linearity [6]. Along with somatotype, anthropometric, body composition and physiologic variables are considered main areas regarding athletes' performance [7]. Chamari et al. [8] reinforce the finding that technical and tactical developments are influenced by morphological characteristics. In addition, the positions of youth soccer players can differ from others on body composition [9].

Soccer can be characterized by a predominant low-to-moderate intensity activity, interspaced with periods of high-intensity actions [10]. Therefore, soccer players depend on well-developed aerobic and anaerobic metabolisms to sustain the different efforts exerted in a match. Despite a predominance of aerobic activity, the most decisive skills, such as to perform a high jump, sprint or score a goal, come from the anaerobic system [11]. Therefore, soccer players need to develop different physical qualities to ensure the best performance during a match.

One of those qualities is the maximum rate of oxygen consumption (VO2max). This is the physiological index most widely used for measuring aerobic fitness of soccer players and can be determined in both, laboratory and field tests. It helps to clarify the level of physical fitness and if it is high, it probably prevents or reduces the risk of injury [12–14].

A soccer game is composed of two parts of 45 min, where the movements, when playing on the field, are very complex and varied, so players require cruising capabilities throughout the game. For this reason, to achieve excellent physical fitness, a soccer player should have a high aerobic capacity [13,15].

In addition, coaches and sport science staff usually perform internal training load quantification to avoid high levels of fatigue and to reduce high risk of illness or injury [16,17]. Also, it allows a better individual and group training periodization [18,19]. Through the rating of perceived exertion (RPE) scale, it is possible to collect considerations regarding physiological characteristics applied during training sessions [20,21].

Furthermore, the knowledge of the aforementioned variables through the season is crucial and can impact training and performance during competition. Therefore, it is important to monitor, access and compare all variables during the different phases of the season. Thus, the literature is somewhat inconclusive about establishing differences in training load, anthropometric, body composition, somatotype, physical, physiological, and soccer-specific skills for player positions in youth players. Moreover, the majority of the studies split the different variables mentioned and do not use them simultaneously.

Therefore, the aims of this study were: (i) to describe baseline anthropometric, maturation, somatotype and fitness parameters differences of players based on playing positions; (ii) to analyze variations of accumulated load training (AcL) between playing positions for periods analyzed; and finally (iii) to explain the variation of VO2max and peak power (PP) through the AcL, body fat (BF), maturity, somatotype, and baseline fitness levels.

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

#### *2.1. The Experimental Approach to the Problem*

The present study consists of two parts: the first is a semi-experimental design with pre- and post-test; the second is a cohort with daily monitoring for 18 weeks in the competitive season. They practiced 5 sessions a week with one match. The team usually performed one resistance training session, one short-speed training session, agility, and small-sided games (SSG) with skill and tactical training per week. The season was divided into two

equal periods of eight weeks: early-season and end-season. The first assessment was in the week before starting the league (pre-season), and the second measurement was done after the league (post-season). Players were assessed on four consecutive days. Anthropometric and body composition were assessed in one day (e.g., height, sitting height, body mass, BF, and girths) then based on this information the percentage of body fat (BF%), lean body mass (LBM), the maturity and somatotype of the players were calculated. The next day, the sprint and change of direction (COD) tests were performed. In the following day, Wingate test was performed to obtain PP and fatigue index (FI), which was considered as a criterion for assessing the anaerobic capacity of players, and on the last day, the Yo-Yo intermittent recovery test level 1 (YYIRT1) was performed to estimate the aerobic power of the players along with the calculation of the VO2max. All tests were performed for each participant under similar environmental conditions and in the same order. Thirty minutes after each training session, all players reported load of training, then each 'training load' was used with training time for calculating AcL for the two periods.

#### *2.2. Participants*

Twenty-seven elite soccer players, belonging to the same national under-16 team competing in the national league, were evaluated for 18 weeks during a competitive season. In total, 76 training sessions and 16 competitive matches were held. To analyze the differences between player positions, they were organized by six central midfielders (CM) with maturity offset 1.72 ± 0.18 years (yrs), four wingers (WG) with maturity offset 1.55 ± 0.17 yrs, five forwards (FW) with maturity offset 2.08 ± 0.33 yrs, eight defenders (DF) with maturity offset 1.94 ± 0.19 yrs, and four goalkeepers (GK) with maturity offset 2.38 ± 0.32 yrs. Inclusion and exclusion criteria in this study were: (i) players who participated in at least 90% of training seasons; (ii) players that did not participate in another training plan along with this study; (iii) each player who was not participating in the match during a week was practicing a separate session, without the ball or through SSG. Before starting this study, explanations about the different phases of the research were given to all participants along with their parents. Also, they were informed of the potential risks and benefits of participating in the study. The study was conducted in accordance with the Declaration of Helsinki; players and their parents given and signed their informed consent to participate in this study, which was approved by the Ethics Committee of the Sport Sciences Research Institute (IR.SSRC.REC.1399.060).

We calculated an a-posteriori estimation of sample size, accepting an alpha risk of 0.05 and a beta risk of 0.2 in a two-sided test; 4 players are necessary for each group to be recognized as statistically significant, with a minimum difference of 10.06 units between any pair of groups, assuming that 5 groups exist. The common deviation is assumed to be 3.64. A drop-out rate of 0% was anticipated.

#### *2.3. Procedures*

#### 2.3.1. Anthropometric and Body Composition

To measure the standing height, participants stood in the stadiometer without shoes and socks. They kept heels, hips, shoulder blades and back of the head as close as possible to the stadiometer, and then feet were placed beside each other. For sitting height, participants sat on a 50 cm bench and brought their buttocks as close as possible to the stadiometer, holding their upper body straight and placing their hands on their feet, then their heights were assessed. The distance between the highest point of the head and the bench, which was at 50 cm, was calculated as sitting height. For this measurement, portable stadiometer SECA (Model 213, Germany) was used with an accuracy of 5 mm.

For measuring maturity offset and age at peak high velocity (PHV), we used the formula: Maturity offset = −9.236 + 0.0002708 (leg length × sitting height) − 0.001663 (age × leg length) + 0.007216 (age × sitting height) + 0.02292 (Weight by Height ratio), R = 0.94, R2 = 0.891, and SEE = 0.592) and for leg length = Standing Height (cm) − Sitting

height (cm) [22]. To measure weight, participants only wore one pair of sports shorts for body weight on the scale SECA (model 813, England), with an accuracy of ± 0.1 kg.

The subcutaneous fat thickness of the seven points of the body including the chest, abdomen, thigh, triceps, subscapular, suprailiac and midaxillary were calculated for body density (BD) by Jackson and Pollock method and for BD and BF% with Brozek's formula [23]. Skin thickness was obtained by calibrating Lafayette Instrument Company (Lafayette, IN, USA) with an accuracy of 0.1 mm. All measurements were performed twice on the right side of the body, the final score recorded with the mean of two measurements. The technical standard error of subcutaneous fat measurement was performed according to previous studies [24]. Other anthropometric measurements such as girths (cm), relaxed arm, flexed arm, chest, waist, hip, upper thigh, mid-thigh, calf and abdomen were measured using the techniques provided by the International Society for the Advancement of Kinanthropometry Advance also used in previous study [25]. The technical error of measurement, inter- and intra-observer, was lower than 3% for the other variables.

#### 2.3.2. Somatotype

Body somatotype, the three-dimensional distance from a profile to the mean of all profiles (endomorph, mesomorph and ectomorph) and height to weight ratio (HWR), according to Carter and Heath [5], were calculated from anthropometric measures including height, weight, four skinfold thickness (triceps, subscapular, supraspinal, and medial calf), two epicondylar breadths (humerus and femur) and two girths (upper arm flexed and tensed, and calf). The somatotypes were plotted in agreement with previous studies [25,26] on a two-dimensional grid system somatochart using the appropriate software https: //www.somatotype.org/ (accessed on 20 March 2021) (Somatotype 1.2 software). All measurements were performed by an expert with five years of background in this area. All anthropometric measurements were performed in the morning [25].

#### 2.3.3. Change of Direction Test

Players did the "modified 505 agility test" [27]. A photo-finish system recorded the time of a complete 5 m turn (2 × 5 m). All procedures were described in our previous study [28]. The best of the efforts performed was used for statistical analysis. The intra-class correlation coefficient (ICC) in this study was equal to two replicates of 0.90 in this test.

#### 2.3.4. Sprint Test

For sprint test a digital timer connected to two photocells was placed at hip height, and after 10-min specific warm-up subjects stood 70 cm before the start line. To calculate the sprint time [29], the test was performed at a distance of 30 m. The best value obtained from 3 trials was used for statistical analysis. Subjects had to rest for at least 3 min between each trial. All phases of testing were monitored by the coach. In this study, COD and sprint tests were performed with the Newtest Powertimer 300-series testing system (Tyrnävä, Finland). The ICC in this study was equal to two replicates of 0.87 in this test.

#### 2.3.5. Anaerobic Power

The Wingate test [30,31] was selected to measure anaerobic power (PP and FI). After giving a warm-up to subjects, the seat height was adjusted so that knee flexion degrees were 170–175, with leg extended fully. At first, to determine the repetition per minute (RPM), the subject began to pedal at their maximum speed for 5 s. RPM was recorded immediately from the ergometer monitor. According to the calculated value and body mass (75 g per kg of body mass), the resistance load of the test was set. The testing procedure consisted of the participants performing a 10-s countdown phase and a 30-s quick pedaling phase; all subjects were verbally encouraged to continue to pedal as fast as they could for the entire 30 s. Ultimately, the desired indicators were calculated using the Wingate power software program of the Monark model 894-E ergometer (Vansbro, Sweden). The ICC in this study was equal to two replicates of 0.94 in this test.

#### 2.3.6. Aerobic Power Test

To evaluate the aerobic power, the YYIRT1 was used and then, VO2max was calculated based on the following formula: VO2max (mL·kg−<sup>1</sup> ·min−<sup>1</sup> ) = YYIRT1 distance (m) × 0.0084 + 36.4 [32]. The ICC in this study was equal to two replicates of 0.86 in this test.

#### 2.3.7. Monitoring Accumulated Training Load

Players were monitored daily for their RPE using the CR-10 Borg's scale, a valid and reliable scale to estimate the intensity of a session [33]. To the question "How intense was your session?" players answered in the interval of number zero for the day without training, 1 for minimum effort and 10 for maximal effort. Players provided responses 30 min after the end of the training session [12,34]. Additionally, the duration of the training sessions (in minutes) was recorded for each player. As a measure of internal load, the s-RPE was calculated by multiplying the score in the CR-10 scale by the duration of the session in minutes [35,36]. Players were previously familiarized with the scale through spending two years at the club. In this study, the AcL (for training and competition) was used for 18 weeks. These weeks of the full competitive season were divided into two periods: early-season, from week (W) 1 to W8 (includes 8 competitions and 39 practice sessions); and end-season, from W9 to W16 (includes 8 competitions and 37 practice sessions).

#### 2.3.8. Statistical Analysis

Statistical analyses were performed using SPSS (version 23.0, IBM SPSS Inc., Chicago, IL, USA) and Graph-Pad Prism 8.0.1 (GraphPad Software Inc, San Diego, California, CA, USA). The significance level was set at *p* < 0.05. Data are presented as mean and standard deviation (SD). Then, inferential tests were executed. Changes between the two in-season periods were assessed using a repeated-measures analysis of variance (ANOVA), followed by Bonferroni post hoc test for pairwise comparisons. Partial eta squared (ηp 2 ) was calculated as effect size of the repeated-measures ANOVA. Besides this, a one-way ANOVA was applied to compare the different assessment variables, by playing position, in each season period. Hedge's g effect sizes with 95% confidence interval were also calculated to determine the magnitude of pairwise comparisons for between-period comparisons. The Hopkins' thresholds for effect size statistics were used, as follows: ≤0.2, trivial; >0.2, small; >0.6, moderate; >1.2, large; >2.0, very large; and >4.0, nearly perfect [37]. Then, multiple linear regression analysis between the percentage of change in fitness levels include VO2max and PP which were calculated by this formula ([POST − PRE]/PRE TEST) × 100). The independent variables considered for multiple linear regression were AcL, BF%, maturity, somatotype, and baseline fitness levels in the soccer players. The Akaike information criterion (AIC) for each model's regression was calculated. Multiple linear regression analysis and AIC were calculated with the R software version 4.0.2 (22 June 2020; R Foundation for Statistical Computing, Vienna, Austria). The test-retest reliability assessments, ICCs, were used. The ICC >0.7 was suitable [38]. G-Power software (University of Düsseldorf, Düsseldorf, Germany) was used for the sample size calculated with the design of the study.

#### **3. Results**

Table 1 shows comparisons between the different playing positions for anthropometric, maturity, body composition and somatotype variables. The most important results of oneway ANOVA showed significant differences between playing positions for maturity offset (*p* = 0.001). Hence, goalkeeper (GK) presented a significantly greater value than central midfielders (CM) (*p* = 0.007; CI95% = 0.14–1.18) and wingers (WG) (*p* = 0.002; CI95% = 0.25– 1.40). Also, defender (DF) presented a significantly greater value than WG (*p* = 0.032; CI95% = 0.03–1.02). For body composition variables, it shows significant differences in (body fat) BF kg (*p* = 0.006), BF% (*p* = 0.015), and lean body mass (LBM; *p* = 0.003). Those differences were found between playing positions for BF%, where WG presented a significantly smaller BF% than CM and GK, respectively (*p* = 0.022; CI95% = −11.02–−0.59

and *p* = 0.025; CI95% = −11.98–−0.55). For LBM, GK presented a significantly greater value than CM and WG (*p* = 0.002; CI95% = 3.94–22 and *p* = 0.022; CI95% = 1.09–20.87), respectively. Further results are shown in Table 1.

**Table 1.** Absolute size characteristic, body composition, somatotype and anthropometric of soccer player by playing positions.


CM, central midfielders; WG, winger; FW, forward; DF, Defender; GK, goalkeeper; PHV = Peak height velocity; BF = Body fat; LBM = lean body mass; HWR = Height to weight ratio; yrs, years; SD, standard deviation. € Represents a statistically significant difference between groups to one-way ANOVA (*p* < 0.05); \* Represents a statistically significant difference compared with goalkeepers (*p* < 0.05); # Represents a statistically significant difference compared with wingers (*p* < 0.05).

> Figure 1 shows the somatotype according to divided playing positions. The somatotype (*p* = 0.020) only showed significant differences between playing positions for endomorph where CM presented a significant greater (*p* = 0.011; CI95% = 0.29–3.15) than WG.

> Significant differences between season periods in accumulated load training (AcL) demonstrated main effects of time (*F* (1, 7.73) *p* = 0.011; η<sup>p</sup> <sup>2</sup> = 0.261) and group effect (*F* (4, 3.43) *p* = 0.025; η<sup>p</sup> <sup>2</sup> = 0.384). Post hoc tests using the Bonferroni correction revealed a significant increase in AcL. There was an only a significant difference between early-season and end-season in forwards (FW) (*p* = 0.043; CI95% = 125.17–4942.44). Tables 2 and 3 show comparisons between the different playing positions for fitness status in pre- and post-season, respectively. This variable was also the analysis of one-way ANOVA with a comparison between different playing position groups in each test time, and it was demonstrated that there was a difference in AcL compared to early-season (WG vs. FW: *p* < 0.021, *g* = 3.09; WG vs. GK: *p* < 0.001, *g* = 2.35; and DF vs. GK: *p* < 0.050, *g* = 1.32). However, no differences were found between playing position in end-season period for AcL.

> There was a significant main effect of time for VO2max (*F* (1, 10.37) *p* = 0.004; η<sup>p</sup> <sup>2</sup> = 0.320) and group effect (*F* (4, 9.45) *p* < 0.001 η<sup>p</sup> <sup>2</sup> = 0.632). Post hoc analysis revealed VO2max was significantly greater at post-season in CM (*p* = 0.042; CI95% = 0.08–3.18) and DF (*p* = 0.048; CI95% = 0.01–1.95). Also, between player positions, this variable demonstrated that there was a significant difference within the pre-season (CM vs. GK: *p* ≤ 0.001, *g* = 4.10; WG vs. GK: *p* ≤ 0.001, *g* = 6.75; FW vs. GK: *p* < 0.002, *g* = 3.11; and DF vs. GK: *p* < 0.001, *g* = 2.57) as well as in the post-season (CM vs. GK: *p* ≤ 0.001, *g* = 3.88; WG vs. GK: *p* < 0.003, *g* = 5.66; FW vs. GK: *p* < 0.003, *g* = 2.83; and DF vs. GK: *p* < 0.001, *g* = 2.57).

–

**Figure 1.** Individual somatotypes by the 2-D somatochart (**a**) Central midfielder, (**b**) Winger, (**c**) Defender, (**d**) Forward, (**e**) Goalkeeper. O = the mean somatotype.

η η – Peak power (PP) levels demonstrated main effects of time (*F* (1, 22.21) *p* ≤ 0.001; ηp <sup>2</sup> = 0.502) and group effect (*F* (4, 9.12) *p* ≤ 0.001; η<sup>p</sup> <sup>2</sup> = 0.624). Post hoc tests using the Bonferroni correction revealed a significant increase in PP between pre-season and postseason in CM (*p* = 0.024; CI95% = 19.93–183.40), DF (*p* = 0.047; CI95% = 1.11–132.39) and GK (*p* = 0.041; CI95% = 3.41–80.59). Also, between player positions, this variable demonstrated that there was a significant difference in the pre-season (WG vs. FW: *p* = 0.001, *g* = 3.93; FW vs. DF: *p* = 0.009, *g* = −1.75; and FW vs. GK: *p* = 0.001, *g* = −4.81) and ultimately, in the post-season (CM vs. FW: *p* = 0.008, *g* = 1.83; WG vs. FW: *p* = 0.002, *g* = 2.91; FW vs. DF: *p* = 0.004, *g* = −1.93; and FW vs. GK: *p* = 0.001, *g* = −3.33). Further results regarding FI are shown in Tables 2 and 3.


**Table 2.** Between-group comparisons for accumulated load and fitness parameters between playing positions in pre-season.


**Table 2.** *Cont.*

M, Mean; diff, difference; AcL, accumulated load training; COD = change of direction; VO2max, maximal oxygen consumption; CM, central midfielder; WG, winger; FW, forward; DF, Defender; GK, goalkeeper; SD, standard deviation; A.U., arbitrary units; CI, confidence interval, and *p*, *p*-value at alpha level 0.05; Hedge's g (95% CI), Hedge's g effect size magnitude with 95% confidence interval. \* The mean difference is significant at the 0.05 levels; # Indicates a significant difference between the groups with Bonferroni at the 0.05 levels.

**Table 3.** Between-group comparisons for accumulated load and fitness parameters between playing positions in after-season.



**Table 3.** *Cont.*

M, Mean; diff, difference; AcL, accumulated load training; COD = change of direction; VO2max, maximal oxygen consumption; CM, central midfielder; WG, winger; FW, forward; DF, Defender; GK, goalkeeper; SD, standard deviation; A.U., arbitrary units; CI, confidence interval, and *p*, *p*-value at alpha level 0.05; Hedge's g (95% CI), Hedge's g effect size magnitude with 95% confidence interval. \* The mean difference is significant at the 0.05 levels; # Indicates a significant difference.

> Multiple linear regression analysis was calculated to predict the percentage of change in fitness levels (i.e., VO2max (mL·kg−<sup>1</sup> ·min−<sup>1</sup> ) and peak power (PP, watts)) based on

AcL, BF%, maturity, somatotype, and baseline fitness levels in soccer player (Figure 2 and Table 4). The first analysis of VO2max showed that there were significant (*F* (8, 18) = 2.71, *p* = 0.038), with a R <sup>2</sup> of just 0.55. Participants showed good predictions for VO2max; (Y) is equal to Beta 0 + Beta1 (Acl) + Beta2 (BF%) + Beta3 (peak height velocity, PHV) + Beta4 (mesomorph) + Beta5 (sprint) + Beta6 (PP) + Beta7 (FI) + Beta8 (VO2max), where AcL was measured as A.U, PHV was measured as years, fitness status was measured as COD (change of direction, seconds), PP (watts), FI (fatigue index, %), and VO2max (mL·kg −1 ·min −1 ) in order based on the equation. − −

= −1.75; and FW vs. GK: = −4.81) and ultimately, in the post

−1.93; and FW vs. GK: = −3.33). Further results regarding FI are shown in Tables

− −

− − **Figure 2.** Multiple linear regression analysis was calculated to predict the percentage of change in fitness levels (**A**) VO2max and (**B**) PP based on accumulated training load, body fat %, maturity, somatotype, and baseline fitness levels in the soccer players. Also, residual plot was calculated to predict the percentage of change in fitness levels (**C**) VO2max and (**D**) PP; the difference between the actual value of the dependent variable and the value predicted by the residual provided. Note: VO2max = maximal oxygen consumption (mL·kg −1 ·min −1 ); PP = Peak power (watts).

There was significant statically found in PP (*F* (8, 18) = 3.80, *p* = 0.009), with an R <sup>2</sup> of 0.63. Participants showed good predictions for PP; (Y) is equal to Beta 0 + Beta1 (Acl) + Beta2 (BF%) + Beta3 (maturity offset) + Beta4 (mesomorph) + Beta5 (COD) + Beta6 (PP) + Beta7 (FI) + Beta8 (VO2max), where AcL was measured as A.U, maturity offset was measured as years, fitness status was measured as sprint (seconds), PP (watts), FI (%), and VO2max (mL·kg −1 ·min −1 ) in order based on the equation.


**Table 4.** Multiple linear regression analysis: percentage of change in VO2max and peak power with workload, body fat, maturity, somatotype, and baseline fitness levels.

Note: β0 = Y; CI = confidence interval; AIC = Akaike information criterion; AcL = accumulated load training; BF = body fat; PHV = Peak height velocity; COD = change of direction; VO2max = maximal oxygen consumption; % = The the percentage of change in between assessments from pre to post-test; A.U. = arbitrary units; yrs = years. \* The significant differences at the 0.05 levels.

#### **4. Discussion**

The purposes of this study were: (i) to describe anthropometric, maturation, and somatotype differences of players based on playing positions; (ii) to analyze variations of accumulated load training (AcL) and fitness parameters between playing positions; and (iii) to show a multiple linear regression analysis between the percentage of change in fitness levels and variables of AcL, body fat percentage (BF%), maturity, somatotype, and baseline fitness levels. In this context, the present study contributes to the existing literature, providing information about the variables mentioned in youth athletes of a professional soccer club.

Regarding the first aim of this study, it was found that goalkeepers (GK) presented higher height, weight, maturity offset, BF and lean body mass (LBM) than other positions. Then, the wingers (WG) showed significantly less BF than GK, and the others positions as well, but the central midfielders (CM) presented lower LBM than other positions. The results are similar to those found in under-16 Spanish soccer players [39] and under-17 Brazilian soccer players [40].

The somatotype results showed some differences between player positions. For instance, the CM and GK presented high endomorph values while WG, defenders (DF) and forwards (FW) presented high ectomorph values. These results are consistent with anthropometric and body composition variables from the present study. Also, it is possible to observe some differences by player positions when compared with the study of Fidelix et al. [40], where it was found that the morphological configuration of DF, FW and GK was classified as a balanced mesomorph, while midfield players were classified as ectomorphmesomorphs; in the present study the morphological configuration of the majority of GK and FW was classified as endomorph ectomorph, while DF and WG were considered balanced ectomorph and CM as endomorph-ectomorph. Some years ago, Rienzi et al. [41] found that GK possessed different somatotype characteristics from the other field positions. Even before, Casajús and Aragonés [42] found higher endomorph values for GK. Gil et al. [39] observed that under-16 soccer players presented higher mesomorph values (2.3-4.3-3.1), but the present study presented more balanced data. Possible explanations

for the present result could be associated with food habits from Iran and some genetic influence which was not controlled in the present study.

Despite some differences between player positions, there are some findings somewhat hard to explain. From the assessment in pre-season, it was shown that WG accumulated higher training loads than other positions and GK accumulated lower training loads than other positions. In agreement with these findings, there was a lower VO2max for GK, while other positions presented similar values and WG the highest VO2max. Also, higher fatigue index (FI) was shown for GK. However, sprint and change of direction (COD) was similar between positions, and peak power (PP) was higher for GK and WG, while FW presented the lowest value. Considering the anthropometric, somatotype, maturation and body composition assessment of the soccer players, the results are difficult to explain and it is not possible to identify a pattern for each player's position. Previously, it was suggested that players that accumulated higher training loads required higher levels of aerobic capacity [43]. This per se, is associated with higher fat-free mass [44]. On the one hand, in the scenario of the present study, GK presented higher LBM but lower VO2max. On the other hand, GK presented higher BF which is in opposition to the statements of Goran et al. [44]. It is important to highlight that the results for VO2max came from Yo-Yo intermittent recovery test level 1 (YYIRT1) and different results could occur with a continuous incremental and maximal test.

In the present study, the only players who showed characteristics different from other positions' somatotype were midfield players. The distance traveled by midfield players is significantly higher than that of backs and FW [45], suggesting that this playing position requires a higher level of aerobic capacity [35], which is strongly influenced by fat-free mass [45]. Also, GK presented lower training load accumulation and higher FI, but it is suggested that higher training load accumulation should lead to higher levels of fatigue [46]. In addition, it is suggested that players with a greater intermittent aerobic capacity have reduced fatigue [47] and vice-versa, which is supported by our study. Furthermore, we also speculate that the higher values for FI could be associated with the number of impacts with soil that GK suffered.

Previously, it was reported that the magnitude of the relationships between age, maturation, body dimensions and match running performance were position dependent. Within a single age-group in the present player sample, maturation had a substantial impact on match running performance, especially in attacking players. Coaches may need to consider players' maturity status when assessing their on-field playing performance [48]. The present study supports the mentioned findings; however, the GK presented the highest level of maturity offset and the FW the second highest level. Meanwhile, this study did not assess match running performance, but it measured AcL, which reflects internal training load perceived by the external load experienced. However, in this scenario, the GK presented the lowest values and the FW the second lowest values during pre-season, but after-season FW presented the highest values.

While Taylor [49] states that in soccer, the best players can reach VO2max levels of 65– 70 mL·kg−<sup>1</sup> ·min−<sup>1</sup> , depending on their age, level of individual performance and position on the pitch, Slimani et al. [3] consider a wide range between 48 and 62 mL·kg−<sup>1</sup> ·min−<sup>1</sup> and specifically, by position, they reported 48.4–57.5 mL·kg−<sup>1</sup> ·min−<sup>1</sup> for GK, 53.2– 62.8 mL·kg−<sup>1</sup> ·min−<sup>1</sup> for DF, 54.7–63 mL·kg−<sup>1</sup> ·min−<sup>1</sup> for CM, and 54.5–62.9 mL·kg−<sup>1</sup> ·min−<sup>1</sup> for FW. Regarding pre-season, our data seems to be in line for all FW, CM, WG and DF, which presented similar values between 48.6 and 50.1 mL·kg−<sup>1</sup> ·min−<sup>1</sup> with the exception of GK who presented the lowest value (41.5 mL·kg−<sup>1</sup> ·min−<sup>1</sup> ). In addition, it was shown that 55% of VO2max variability and 63% of PP variability is explained by independent variables, respectively (see Table 4 and Figure 2). After-season, FW, CM, WG and DF revealed a slight increase with a range between 49.6 and 51.6 mL·kg−<sup>1</sup> ·min−<sup>1</sup> , with GK presenting 41.9 mL·kg−<sup>1</sup> ·min−<sup>1</sup> .

In the present study of twenty-seven under-16 elite soccer players from the Iranian League, what have been found are significant relationships between VO2max and peak height velocity (PHV) and between peak power with BF and peak power with sprint (all, *p* < 0.05, Table 4);recently [28], it was suggested that the higher physical capacity allows soccer players to perform with stronger exertion, which could be expressed in the values of the chronic workload and the accumulated training monotony. The same study found a relationship between the PHV and the accumulated training monotony, and between chronic workload and physical abilities that can be expressed by VO2max [28].

Due to the limited sample of our study, we suggest larger samples with the same analysis to better interpret if the other variables, such as, AcL, maturity, somatotype and baseline fitness levels can predict VO2max and peak power.

Higher VO2max is associated with higher performance in matches such as distance traveled, intensity, number of sprints, and the amount of player involvement with the ball [50]. Also, soccer players will have more energy to move with few limitations and will have a fast recovery without increasing fatigue substantially. The statements presented regarding VO2max supported some results found in the present study during pre-season. For instance, in pre-season, the position with higher VO2max was WG, which was also revealed to be the position with higher AcL and more sprints. However, they did not present higher values of PP and COD tests. Possible explanations for these differences could be associated with the use of non-specific tests to assess PP for soccer players by using a bicycle ergometer; COD testing is a skill that could be developed with special care considering a specific position. After-season, the position with higher VO2max was CM, but this time it only displays higher results in COD tests while other tests analyzed present different player positions with higher values (see Table 3). Despite the possible physiological and positional adaptations that may occur during the season, a justification for the presented results could be associated with previous studies that stated that in-season training load variability is very limited, and that only minor decrements or a maintenance during the season might occur [51–54], which is in line with Malone et al. [13], who posit that it is the need to win matches that influences a possible specific peak for strength and conditioning.

In agreement with Fidelix et al. [40] study, there are some limitations regarding our participants, as they belong to the same team which per se can be associated with a specific somatotype profile of the club's intention, its geographical location and others. Also, the small sample size and their specific team and country do not allow generalizing the present finding.

This study provides useful information regarding the AcL, anthropometric, body composition, maturity, somatotype and fitness levels such as VO2max, PP, anaerobic power, aerobic power, COD, sprint and YYIRT1 of a youth soccer team during different in-season periods. It provides further evidence of the value of using a combination of different monitoring and assessment measures to fully evaluate the youth soccer player across a full competitive season. Moreover, it identifies differences between player positions which allow coaches, staff, and the scientific community to analyze youth soccer player with greater knowledge. Also, for coaches, this study could provide important information to be considered when planning training sessions and/or weekly periodization. For instance, coaches can use information from RPE to produce AcL and to better understand the load perceived by young soccer players. In addition, with the information from the present study, it is suggested to include the following fitness parameters when analyzing young soccer players: PHV, body composition variables such BF, somatotype, VO2max, sprint and COD tests.

In future studies, it would be interesting to replicate the present study with more teams in the same season, level of competition, or even with different age groups to better interpretation of the results. Furthermore, it would be pertinent to replicate the present study with female soccer players and different age categories in order to increase knowledge on the variables analyzed.

#### **5. Conclusions**

In general, GK showed higher values in anthropometric, body composition variables and maturity offset compared to the other positions. In the opposite direction, WG presented lower levels of BF. In addition, there was only one significant difference in somatotype, where DF presented a higher endomorph value than WG.

Furthermore, there were several differences in the beginning of the season and few after-season. In pre-season, AcL, VO2max, sprint was found to be higher for WG while COD was found to be higher for CM. PP and FI was found to be higher for GK. After-season, AcL, is similar for all positions except for the GK. VO2max was found to be higher for CM. Sprint was higher for WG. COD was found to be higher for CM and GK. Still, PP and FI was found to be higher for GK. These finding reinforce the tactical role of the positions as they produce different adaptations during the season. Our multiple linear regressions support these findings because they indicated that our model explains more than 50% of all the variability of the responses.

This information is useful for coaches and professionals involved in sports, as it can be used in the process of talent selection and the development of training programs because they serve as a reference for athletes of the same sex, age and competitive level.

**Author Contributions:** Conceptualization, H.N., R.O., J.P.-G. and L.P.A.; methodology, H.N., R.O., F.M.C., E.P.-M. and J.P.-G.; software, H.N., F.M.C., R.O. and J.P.-G.; formal analysis, H.N., F.M.C., and R.O.; investigation, J.P.-G., H.N. and R.O.; writing—original draft preparation, H.N. and R.O.; writing—review and editing, H.N., E.P.-M., R.O., J.P.-G. and L.P.A. All authors have read and agreed to the published version of the manuscript.

**Funding:** Portuguese Foundation for Science and Technology, I.P., Grant/Award Number UIDP/ 04748/2020.

**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 Ethics Committee of the Sport Sciences Research Institute (IR.SSRC.REC.1399.060).

**Informed Consent Statement:** Informed consent was obtained from all subjects and their parents involved 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 the team's coaches and players for their cooperation during all data collection procedures.

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

#### **References**


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

*Children* Editorial Office E-mail: children@mdpi.com www.mdpi.com/journal/children

MDPI St. Alban-Anlage 66 4052 Basel Switzerland

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

www.mdpi.com ISBN 978-3-0365-3128-1