*Article* **Comparisons of the Prevalence, Severity, and Risk Factors of Dysmenorrhea between Japanese Female Athletes and Non-Athletes in Universities**

**Reiko Momma <sup>1</sup> , Yoshio Nakata <sup>2</sup> , Akemi Sawai <sup>3</sup> , Maho Takeda <sup>1</sup> , Hiroaki Natsui <sup>4</sup> , Naoki Mukai <sup>2</sup> and Koichi Watanabe 2,\***


**Citation:** Momma, R.; Nakata, Y.; Sawai, A.; Takeda, M.; Natsui, H.; Mukai, N.; Watanabe, K. Comparisons of the Prevalence, Severity, and Risk Factors of Dysmenorrhea between Japanese Female Athletes and Non-Athletes in Universities. *Int. J. Environ. Res. Public Health* **2022**, *19*, 52. https://doi.org/10.3390/ ijerph19010052

Academic Editors: Filipe Manuel Clemente and Ana Filipa Silva

Received: 31 October 2021 Accepted: 17 December 2021 Published: 21 December 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/).

**Abstract:** This study aimed to investigate the difference in the prevalence, severity, and risk factors of dysmenorrhea between Japanese female athletes and non-athletes in universities. The participants were 18 to 30 years old with no history of a previous pregnancy and/or childbirth. After application of the exclusion criteria, the cohort comprised 605 athletes and 295 non-athletes. An anonymous questionnaire, which included self-reported information on age, height, weight, age at menarche, menstrual cycle days, menstrual duration, dysmenorrhea severity, sleeping hours, dietary habits, exercise habits, training hours, and competition level was administered. Compared with athletes, nonathletes had a higher prevalence of dysmenorrhea (85.6% in athletes, 90.5% in non-athletes, *p* < 0.05); non-athletes also demonstrated increased severity (none/mild 27.8%, moderate 19.3%, and severe 52.9% in athletes; none/mild 21.2%, moderate 17.2%, and severe 61.6% in non-athletes; *p* < 0.05). Factors related to severe dysmenorrhea in athletes included long training hours, early menarche, and prolonged menstrual periods. In non-athletes, short menstrual cycle days and extended menstrual periods were related to severe dysmenorrhea. The prevalence and severity of dysmenorrhea were higher among non-athletes than among athletes; different factors were related to severe dysmenorrhea in these two groups. Thus, different strategies are necessary to manage dysmenorrhea for athletes and non-athletes in universities.

**Keywords:** menstruation disturbances; menstrual cycle; athletes; women's health; exercise

#### **1. Introduction**

Dysmenorrhea is an important women's health problem. It is experienced during menstruation and is associated with pain and discomfort such as headaches, abdominal pain, and back pain [1]. There are two types of dysmenorrhea: primary dysmenorrhea, which is caused by excessive prostaglandin secretion without an organic uterine disease, and secondary dysmenorrhea, which is caused by an organic disease of the uterus [2]. Previous studies have demonstrated that the prevalence of dysmenorrhea is approximately 80% among young women; 77.6% among working women (aged 25–55 years) [3]; 83.6% among college students [4]; and 89% among adolescent girls [5]. Moreover, dysmenorrhea is a severe problem in young women because it negatively impacts their lives; for example, it is a cause of absenteeism from school and work and decreased health-related quality of life [6,7].

Bad lifestyle habits may potentially be important risk factors of dysmenorrhea. Short sleeping hours and not having breakfast regularly were associated with moderate-to-severe

dysmenorrhea in a previous study [8]. In addition, caffeine consumption [9], alcohol consumption, and smoking [10] were also associated with dysmenorrhea. Moreover, mental stress [11–13] and a lack of exercise [14] were related to the severity of dysmenorrhea. Therefore, lifestyle changes may be a potential strategy to manage dysmenorrhea.

Armor et al. reported that dysmenorrhea lowered athletic performance during training and competitions [15]. Another study showed that the dysmenorrhea pain score was higher in athletes than in sedentary students [16]. Additionally, an interview-based study reported that menstruation-related symptoms reduce athletic performance in athletes [17]. In a previous study, athletes had a lower prevalence of dysmenorrhea than non-athletes (39.44% in athletes and 43.88% in non-athletes), although the difference was not significant [18]. An additional study showed that exercise can reduce dysmenorrhea [14]; however, the participants in this study were women with no exercise habits. Research on dysmenorrhea in athletes and non-athletes has therefore not yielded consistent results.

To address this issue, the present study aimed to investigate the difference in the prevalence, severity, and risk factors of dysmenorrhea between Japanese female athletes and non-athletes in universities. The present study hypothesized that athletes show an increased prevalence of severe dysmenorrhea relative to non-athletes and that different factors are associated with severe dysmenorrhea between these two groups of women.

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

#### *2.1. Study Design*

We conducted a cross-sectional, anonymous questionnaire survey administered from October 2019 to March 2020. The participants were recruited using a snowball sampling method, and all individuals consented to participating in this study. The Ethics Review Board of the Faculty of Health and Sport Sciences at the University of Tsukuba approved the study protocol (approval number: 19–85) on 19 September 2019.

#### *2.2. Participants*

Our cohort of participants included 961 athletes and 423 non-athletes who were recruited with the help of faculty members from six Japanese universities located in Tokyo (three universities), Ibaraki (two universities), Chiba (one university), and Okayama (one university). The athlete group consisted of university students who majored in physical education or sports science and/or who belonged to athletic clubs. The non-athlete group consisted of university students who majored in subjects other than sports science, such as nutrition and nursing, and/or those who did not participate in athletic competitions, such as managers of athletic clubs. This study included women who were aged 18 to 30 years, those who had never been pregnant and/or given birth, those who did not take oral contraceptives, and those who did not have irregular menstruation or secondary amenorrhea. University athletes were defined as those who belonged to an athletic club (not a recreational club), participated in competitions on a regular basis, and trained at least 3 days per week. As shown in Figure 1, 356 women in the athlete group and 126 women in the non-athlete group were excluded owing to the following exclusion criteria: those taking oral contraceptives (n = 22 and n = 15 in the athlete and non-athlete groups, respectively), those with irregular menstruation or secondary amenorrhea (n = 49 and n = 20, respectively), those who trained less than 3 days a week (n = 62 in the athlete group), and those with incomplete data (n = 223 and n = 93, respectively). The final analysis dataset comprised data from 605 athletes and 295 non-athletes. The athletes played basketball (n = 98), track and field (n = 88), lacrosse (n = 62), handball (n = 62), volleyball (n = 44), soccer (n = 33), rhythmic gymnastics (n = 32), dance (n = 32), softball (n = 27), kendo (n = 23), judo (n = 23), swimming (n = 19), badminton (n = 14), baseball (n = 13), tennis (n = 12), cheerleading (n = 9), gymnastics (n = 6), wheel gymnastics (n = 5), and wrestling (n = 3).

(n = 23), swimming (n = 19), badminton (n = 14), baseball (n = 13), tennis (n = 12), cheerleading (n = 9), gymnastics (n = 6), wheel gymnastics (n = 5), and wrestling (n = 3).

**Figure 1.** Participant flow diagram.

#### **Figure 1.** Participant flow diagram. *2.3. Questionnaire*

*2.3. Questionnaire*  A questionnaire that included questions related to age, height, weight, age at menarche, menstrual cycle days, menstrual duration, dysmenorrhea severity (none: 0 to heavy pain: 10), sleeping hours, dietary habits (skipping meals), exercise habits (in non-athletes), training hours (per week), and competition level (1: international, 2: national, 3: regional, 4: prefectural, 5: other, in athletes) was prepared. Body mass index (BMI) was calculated using the following formula: weight (kg) divided by the square of height (m2). The following question was asked about the prevalence and severity of dysmenorrhea. "What is the degree of pain you experience during menstruation? Please circle the number between 0 and 10 that is reflective of the pain you experience." Those with a severity score of ≥1 were defined as having dysmenorrhea. With reference to a previous study [19], the severity of dysmenorrhea was classified into three categories, namely, none/mild (0 to 3), moderate (4 to 6), and severe (7 to 10). Additionally, gynecological age was calculated by subtracting the age at menarche from the calendar age [20]. The severity of dysmenorrhea was com-A questionnaire that included questions related to age, height, weight, age at menarche, menstrual cycle days, menstrual duration, dysmenorrhea severity (none: 0 to heavy pain: 10), sleeping hours, dietary habits (skipping meals), exercise habits (in non-athletes), training hours (per week), and competition level (1: international, 2: national, 3: regional, 4: prefectural, 5: other, in athletes) was prepared. Body mass index (BMI) was calculated using the following formula: weight (kg) divided by the square of height (m<sup>2</sup> ). The following question was asked about the prevalence and severity of dysmenorrhea. "What is the degree of pain you experience during menstruation? Please circle the number between 0 and 10 that is reflective of the pain you experience." Those with a severity score of ≥1 were defined as having dysmenorrhea. With reference to a previous study [19], the severity of dysmenorrhea was classified into three categories, namely, none/mild (0 to 3), moderate (4 to 6), and severe (7 to 10). Additionally, gynecological age was calculated by subtracting the age at menarche from the calendar age [20]. The severity of dysmenorrhea was compared between competition levels or sport types in university athletes.

#### pared between competition levels or sport types in university athletes. *2.4. Statistical Analysis*

*2.4. Statistical Analysis*  Data were analyzed using SPSS version 26 (SPSS Inc., Chicago, IL, USA). The Kolmogorov–Smirnov normality test was used to examine normality. Because all variables were not normally distributed, the Mann–Whitney test was conducted to compare the characteristics of the participants, and the chi-square test was conducted to compare the prevalence and severity of dysmenorrhea between athletes and non-athletes, between sport types, and between competition levels. The Kruskal–Wallis test was used to compare the characteristics of the participants according to the severity of dysmenorrhea in athletes and non-athletes, followed by a Bonferroni post hoc test. The effect sizes were calculated and expressed as ES [21]. A logistic regression model was used to identify risk factors for severe dysmenorrhea (severe or not); the severity of dysmenorrhea was the dependent variable in the athlete and non-athlete groups. Independent variables were age, BMI, sleeping hours, skipping meals, age at menarche, menstrual cycle, menstrual period, training hours (in athletes), competition level (in athletes), and exercise hours (in non-athletes). Data were analyzed using SPSS version 26 (SPSS Inc., Chicago, IL, USA). The Kolmogorov– Smirnov normality test was used to examine normality. Because all variables were not normally distributed, the Mann–Whitney test was conducted to compare the characteristics of the participants, and the chi-square test was conducted to compare the prevalence and severity of dysmenorrhea between athletes and non-athletes, between sport types, and between competition levels. The Kruskal–Wallis test was used to compare the characteristics of the participants according to the severity of dysmenorrhea in athletes and non-athletes, followed by a Bonferroni post hoc test. The effect sizes were calculated and expressed as ES [21]. A logistic regression model was used to identify risk factors for severe dysmenorrhea (severe or not); the severity of dysmenorrhea was the dependent variable in the athlete and non-athlete groups. Independent variables were age, BMI, sleeping hours, skipping meals, age at menarche, menstrual cycle, menstrual period, training hours (in athletes), competition level (in athletes), and exercise hours (in non-athletes). The odds ratio and 95% confidence interval (95% CI) were calculated for each variable. Data are expressed as median (interquartile range) or frequency (%).

#### **3. Results**

#### *3.1. Participant Characteristics*

As shown in Table 1, some characteristics differed significantly between athletes and non-athletes. Athletes were taller, heavier, and higher in BMI and had shorter menstrual periods, a younger gynecological age, and longer sleeping hours than non-athletes. The prevalence of dysmenorrhea was significantly higher in non-athletes than in athletes (*p* = 0.04, ES = 0.07).

**Table 1.** The characteristics of the study participants.


Data are expressed as median [interquartile range] or frequency (%). BMI, body mass index; ES, effect size.

#### *3.2. Dysmenorrhea Severity*

As shown in Figure 2, the severity of dysmenorrhea differed significantly between athletes and non-athletes (*p* = 0.04, ES = 0.09). Dysmenorrhea was shown to be more severe in non-athletes (none/mild 21.2%, moderate 17.2%, and severe 61.6%) than in athletes (none/mild 27.8%, moderate 19.3%, and severe 52.9%).

Tables 2 and 3 present characteristics by the severity of dysmenorrhea in athletes and non-athletes, respectively. In athletes, there were significant differences in age, age at menarche, menstrual period, and gynecological age between the dysmenorrhea severity groups. However, there were no differences in these variables between competition levels or between sport types. In non-athletes, there were no significant differences in these variables between the dysmenorrhea severity groups.

**Figure 2.** Severity of dysmenorrhea in athletes and non-athletes. **Figure 2.** Severity of dysmenorrhea in athletes and non-athletes.


**Table 2.** Characteristics by severity of dysmenorrhea in athletes. **Table 2.** Characteristics by severity of dysmenorrhea in athletes.

Other 5 (3.0%) 4 (3.4%) 7 (2.2%) Data are expressed as median [interquartile range] or frequency (%). \*: no/mild vs medium (*p* < Data are expressed as median [interquartile range] or frequency (%). \*: no/mild vs. medium (*p* < 0.01), #: no/mild vs. severe (*p* < 0.01), \$: medium vs. severe (*p* < 0.05). BMI, body mass index; ES, effect size.

0.01), #: no/mild vs severe (*p* < 0.01), \$: medium vs severe (*p* < 0.05). BMI, body mass index; ES,

**Severe** 

**(n = 183) ES** 

**Table 3.** Characteristics by severity of dysmenorrhea in non-athletes.

Age (years) 20.0 [19.0–21.0] 20.0 [19.0–22.0] 20.0 [19.0–21.0] 0.01 Height (cm) 158.0 [153.6–161.1] 158.7 [155.4–161.0] 158.0 [155.0–162.0] <0.01

**Medium (n = 51)** 

effect size.

**No/Mild (n = 61)** 


**Table 3.** Characteristics by severity of dysmenorrhea in non-athletes.

Data are expressed as median [interquartile range] or frequency (%). BMI, body mass index; ES, effect size.

*3.3. Factors Related to Severe Dysmenorrhea*

Tables 4 and 5 illustrate the coefficients at 95% CIs generated from the logistic regression model using dysmenorrhea severity (severe or not) as the dependent variable in athletes and non-athletes, respectively. In athletes, long training hours, early menarche, and a long menstrual period were significantly related to severe dysmenorrhea (Table 4). In non-athletes, a short menstrual cycle and a long menstrual period were significantly related to severe dysmenorrhea (Table 5).

**Table 4.** Factors related to the severity of dysmenorrhea in athletes.


BMI, body mass index; CI, confidence interval; Exp(B), odds ratio.


BMI, body mass index; CI, confidence interval; Exp(B), odds ratio.

#### **4. Discussion**

The present study investigated the difference in the prevalence, severity, and risk factors of dysmenorrhea between Japanese female athletes and non-athletes in universities. The prevalence of dysmenorrhea was higher in non-athletes (90.5%) than in athletes (85.6%) (*p* = 0.04, ES = 0.07). Furthermore, the severity of dysmenorrhea was higher in non-athletes

than in athletes (*p* = 0.04, ES = 0.09). Although the effect sizes were small, significances were observed. The factors associated with severe dysmenorrhea were different between athletes and non-athletes. As mentioned earlier, long training hours, early menarche, and long menstrual periods were significant risk factors among athletes, while short menstrual cycles and long menstrual periods were shown to be significant risk factors among non-athletes. Therefore, different strategies may be necessary to address severe dysmenorrhea in athletes and non-athletes in universities.

Most previous studies have reported the prevalence of dysmenorrhea among the general population. Polat et al. reported that the prevalence of dysmenorrhea among adult university students in Turkey was 87.8% [22]. In contrast, Ortiz et al. reported a prevalence of 48.4% among Mexican high school students [23]. This inconsistency is partly due to the different definitions of dysmenorrhea. The former study defined dysmenorrhea as having pain during menstruation; the latter defined dysmenorrhea as having painful menstruation for the past 3 months. It is necessary to focus on this point when interpreting the prevalence of dysmenorrhea. The definition from the former study was used in the present study, and the prevalence of dysmenorrhea was similar in both [22].

Few previous studies have reported the prevalence of dysmenorrhea among athletes. Homai et al. compared the prevalence of dysmenorrhea between athletes and non-athletes (39.44% in athletes and 43.88% in non-athletes) [16]. However, they used a different definition of dysmenorrhea. The present study showed that the prevalence of dysmenorrhea was higher in non-athletes (90.5%) than in athletes (85.6%). While the prevalence of dysmenorrhea was much higher in the present study than in the previous study, the rank relationship noted in the present study was comparable to that reported in the previous study [16].

Many previous studies have reported the severity of dysmenorrhea among the general population of women; however, few previous studies compared differences in severity between athletes and non-athletes. Some observational studies reported that women without exercise habits had a high prevalence and severity of dysmenorrhea [8,14,24]. Some intervention studies demonstrated that an exercise intervention improved the severity of dysmenorrhea in sedentary women [25–28]. Therefore, exercise might be a potential strategy to manage dysmenorrhea in the general population of women.

However, very frequent training may be a risk factor for severe dysmenorrhea in athletes. Czajkowska et al. reported that premenstrual syndrome (PMS) might worsen in athletes due to high-intensity training and an extended competition history [29]. A previous study showed a correlation between the severity of PMS and the severity of dysmenorrhea [30]. Therefore, the present study hypothesized that the prevalence of severe dysmenorrhea is higher in athletes owing to consistent, high-intensity training. However, in the present study, the prevalence of severe dysmenorrhea was shown to be higher in nonathletes (61.6%) than in athletes (52.9%), which is not in line with the initial hypothesis. In addition, there was no difference in the severity of dysmenorrhea between competition levels or sport types in the present study. Further studies are necessary to be conducted in different populations.

The present study also examined the factors related to severe dysmenorrhea in athletes and non-athletes. In athletes, long training duration was a risk factor for severe dysmenorrhea, and this finding is similar to that of a previous study that reported that long training hours are associated with PMS [29]. Although the prevalence of severe dysmenorrhea was lower in athletes than in non-athletes, frequent training may be a risk factor for severe dysmenorrhea. Low-intensity exercises, such as yoga and Pilates, are thought to be beneficial for improving dysmenorrhea because it lowers the levels of cortisol, which in turn inhibits prostaglandin synthesis [31,32]. However, prolonged high-intensity exercise, which is performed by athletes, may increase the levels of inflammatory cytokines, which in turn may increase prostaglandin synthesis and increase the severity of dysmenorrhea [33]. Therefore, the management of training hours might be a crucial factor in controlling dysmenorrhea in athletes.

Long menstrual periods were a common risk factor for dysmenorrhea in university athletes and non-athletes. This result was consistent with those of previous studies conducted on the general population [23,34,35]. In non-athletes, short menstrual cycles were shown to be an important risk factor for dysmenorrhea, while, in contrast to previous studies, exercise habits were not [8,14,24]. Although the risk factors in athletes and nonathletes were examined separately, the previous studies may have included both athletes and non-athletes in the study population. Therefore, the difference in study designs and study populations might have resulted in different findings in these studies.

There were some limitations in this study. First, the study participants were not a representative sample. Athlete and non-athlete participants were recruited separately. Therefore, the overall sample included in this study could not be analyzed. Second, the participants were recruited from a limited number of universities, and the participants were pursuing studies in physical education, nursing, or nutrition. In addition, we enrolled athletes from many sports in this study. Third, self-reported data were collected using a questionnaire, which contained questions that required participants to recollect events that had happened in the past; this might have led to recall bias. Fourth, primary dysmenorrhea was not differentiated from secondary dysmenorrhea. The causes are different: primary dysmenorrhea is caused by prostaglandins and secondary dysmenorrhea is caused by an organic disease. As the causative mechanisms of primary and secondary dysmenorrhea are different, future studies involving the collection of the history of gynecological consultations and previous medical history are needed. In addition, the diseases that may cause dysmenorrhea were not investigated. Fifth, the validity and reliability of the questionnaires were not tested. Sixth, a detailed survey on the nutritional status of the participants was not conducted. Thus, caution is necessary when generalizing the results of this study.

#### **5. Conclusions**

The present study compared the prevalence and severity of dysmenorrhea between female university athletes and non-athletes in Japanese universities and investigated the risk factors. The prevalence and severity of dysmenorrhea were higher in non-athletes than in athletes. The risk factors for severe dysmenorrhea were long training hours, early menarche, and long menstrual periods in athletes. In contrast, short menstrual cycles and long menstrual periods were shown to be significant risk factors in non-athletes. Therefore, different strategies may be necessary to address dysmenorrhea in athletes and non-athletes in universities.

**Author Contributions:** Study concept and design: R.M., A.S., H.N., M.T., N.M. and K.W. Acquisition of data: R.M., A.S., M.T. and H.N. Analysis and interpretation: R.M., Y.N. and K.W. Writing the first draft: R.M. and Y.N. All authors have critically reviewed the article and agreed on the journal to which the article will be submitted. All authors have reviewed and agreed on all versions of the article before submission, during revision, the final version accepted for publication, and any significant changes introduced at the proofing stage. All authors have agreed to take responsibility and be accountable for the contents of the article. All authors have read and agreed to the published version of the manuscript.

**Funding:** The Total Conditioning Research Project of Japan Sport Council financially supported this research.

**Institutional Review Board Statement:** The Ethics Review Board of the Faculty of Health and Sport Sciences at the University of Tsukuba approved the study protocol (approval number: 19–85) on 19 September 2019.

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

**Data Availability Statement:** The data presented in this study are not publicly available in compliance with the investigation confidentiality and are available from the corresponding author on reasonable request.

**Acknowledgments:** We want to thank the participants for their cooperation and the collaborative faculty members for recruiting the participants.

**Conflicts of Interest:** The authors report no conflict of interest in this work.

#### **References**


### *Article* **On-Match Impact and Outcomes of Scoring First in Professional European Female Football**

**Patricia Sánchez-Murillo <sup>1</sup> , Antonio Antúnez <sup>1</sup> , Daniel Rojas-Valverde 2,3,\* and Sergio J. Ibáñez 1,\***


**Abstract:** Background: Scoring first seems to be a determinant in professional football playing; several factors could influence the development of the match and the outcome. This study aimed to identify which factors could influence scoring first and impact match outcomes in professional European female football. Methods: There were 504 official matches held on 74 match days during the 2018–2019 professional female European football seasons (*Primera Iberdrola*, *D1 Féminine*, and *Frauen-Bundesliga*), analysed using a notational and inferential assessment. Results: There was a direct positive relationship (*p* < 0.05) between scoring first and winning the match; 75.9% of the winning teams scored first. Moreover, those teams that usually scored first had a better final league classification (*p* < 0.05). These relationships were not influenced by home or away conditions. Conclusions: Scoring first is a determinant in the outcomes of professional European female football matches. Physical and tactical training and programming should focus on those variables, leading female teams to score first.

**Keywords:** women; football; final score; winning; match result; situational variables

#### **1. Introduction**

Football is well-known as one of the most played and popular sports worldwide. This sport attracts millions of fans, and the economic and social interest in related events, tournaments, and matches continues to grow. Curiously, despite its attractiveness, football match outcomes are often determined by relatively few critical actions, leading to small final scores. Usually, football matches have an average of 2.7 goals per game [1]. This reduced number of goals, which determines the final match result is the primary rationale of a study on the influence of scoring first in professional football [2].

Several studies analyse the impact of scoring first in male football [3–5]; but little studies focused on female teams [3]. All of the scientific evidence suggested that scoring first in female and male football is critical; this advantage increases the probability of winning the matches. In male football, 65–75% of the matches are won by the first-scoring teams, whereas female football lacks the evidence to summarise the real impact and benefits of scoring first.

Some factors may be determinant to scoring first, such as the home advantage, team league classification, and if the goal was scored in the first or second half of the match. The home advantage is understood as the relative advantage of being the match host. Previous studies in male football suggested that, when playing at home, the home team has winning odds of 74% [4]. Contrarily, if the visiting team scores first, the winning odds are 50% [5] to 63% [4]. These odds also seem to apply to women's football [6]; in an analysis

**Citation:** Sánchez-Murillo, P.; Antúnez, A.; Rojas-Valverde, D.; Ibáñez, S.J. On-Match Impact and Outcomes of Scoring First in Professional European Female Football. *Int. J. Environ. Res. Public Health* **2021**, *18*, 12009. https:// doi.org/10.3390/ijerph182212009

Academic Editors: Paul B. Tchounwou, Filipe Manuel Clemente and Ana Filipa Silva

Received: 12 October 2021 Accepted: 12 November 2021 Published: 16 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/).

of the top Spanish league, those female teams that scored first in the match won in more than 80 to 90% of cases, depending on the team's classification (top three vs. top ten) [7]. Additionally, in top European male football leagues, if the home team scores first, the winning odds are 62% [5].

The scoring-first effect on match outcome was studied in other sports, such as baseball and hockey. Scoring first (66.3%) and being the home team (61.7%) are determinants of the final score in baseball [8]. In hockey, the psychological momentum of scoring first causes a marked increase in the likelihood of winning the matches [9].

The abovementioned studies highlighted the second determinant of scoring first, proposing that the best-classified teams usually scored first during female matches [7]. This was also the cases in male football, where the best-ranked teams usually scored first and, as expected, won their matches [10]. Some evidence in male football suggested that higher-budget teams are more likely to win (14%) [11]. Additionally, the competitive balance, understood as the balance in the sport capabilities of teams, could influence the match outcome [12]. Additionally, other team characteristics, such as overall quality and overall ranking, are determinants in team sports [13,14].

Additionally, analyses of the top European championships (UEFA Champions League and European League) show that the first goal scored between 16 and 45 min of match time is more determinant in the outcome than those scored during the first 15 min of the match [15]. This evidence was confirmed by recent studies, suggesting that the team that scored its first goal during the final minutes of a game is usually the winner [1].

All of these situational factors could define the outcome of a match, and they are usually explained by a series of tactical and psychological reasons. Tactical passivity, decreased motivation and confidence, and reduced crew tactical structure and cohesion are common issues for losing teams after conceding a goal [16]. Additionally, with regard to the basic physiological differences [17], female football, as opposed to male football, has particular characteristics that could influence the match outcome and first goal, as budget disparities (competitive balance), quality, and technical and tactical skills differ between teams. To better understand female professional players' psychological, tactical, and physical behaviour and acknowledge the lack of studies on women's football, this study aimed to identify which factors could influence scoring first and how they could impact match outcomes in professional European female football.

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

#### *2.1. Sample*

This study was defined as observational since the authors did not influence the natural behaviour of the matches, using an ex post facto analysis [18]. All official matches were recorded and explored using a systematic notational analysis.

A total of 504 official matches held on 74 match days during the 2018–2019 top female European football seasons (*Primera Iberdrola*, *D1 Féminine*, and *Frauen-Bundesliga*) were assessed. The distribution of the matches by league was as follows: Primera Iberdrola = 240, D1 Féminine = 132 and Frauen-Bundesliga = 132. The French and German leagues included 12 teams and a total of 22 match days, while the Spanish league included 16 teams and a total of 30 match days. Consequently, a total of 1008 cases were analysed. The format of all leagues defined the champion as the team with the most points (best-ranked) following both home and away matches.

All of the data analysed were extracted from a digital database, accessible to the public from the leagues' official websites (e.g., www.laliga.es). This task was performed by two experts and was contrasted. If there were inconsistencies, both experts agreed after a consensus and final review of the databases.

#### *2.2. Variables*

Based on the previous literature, the selected variables were chosen as independent: first-scorer team (first scorer vs. the second scorer,) and the following as dependent

variables: match time when the first goal was scored (e.g., 0–15 min, 16–30 min, 30–45 min, 45–60 min, 60–75 min, 75–90 min, 90+ min), league (*Primera Iberdrola* vs. *D1 Féminine* vs. *Frauen-Bundesliga*), local conditions (home vs. away). Other quantitative variables were used, including final ranking (top 1–4 vs. positions 5–12 and 13–16 for Primera Iberdrola; 5–8 and 9–12 for Frauen-Bundesliga and D1 Féminine), number of goals scored (0–11 per team), and number of yellow (0–6 per team) and red cards (0–1 per team).

#### *2.3. Statistical Analysis*

The observational data collection was performed using a data sheet (Excel, Microsoft Office 365, Mountain View, CA, United States). Analyses were made using the Statistical Package for Social Sciences (SPSS v.21.0, Chicago, IL, USA).

The normality of the data was explored by Kolmogorov–Smirnov test. Categorical variables were treated as non-parametrical data. The data were presented using descriptive analysis and frequency distribution for qualitative variables; for quantitative variables, the data were presented as mean, minimum, maximum and typic deviations [19]. Inferential analyses were used to explore the potential influence of independent variables on the dependent variable. Association between variables was explored using chi-squared (*X* 2 ) and Cramer's V (*V*). The relationship between variables using *X* <sup>2</sup> was understood as *p* < 0.05. Cramer´s V was interpreted using previous criteria [20], as follows: trivial (<0.10), small (0.10–0.29), moderate (0.30–0.49) and large (>0.50).

The relationship level between variables was established using the *adjusted standardised residuals (ASR)* and different contingency tables (Field, 2009). Those residuals greater than 1.96 confirmed the association between variables, the interpretation made based on previous criteria [21].

Differences in goals, yellow cards, and red cards between first and non-first scorers were explored using a one-way analysis of variance (*F* value) (1 × 3).

#### **3. Results**

The descriptive results of match outcome, local conditions and the match time when the first goal was scored are presented by league (see Table 1). In the Primera Iberdrola League, the winning teams scored first in 77.3% of cases (*X* <sup>2</sup> = 393.5, *p* < 0.01; *V* = 0.6 *(large)*, *p* = 0 < 0.01, *ASR* = 15.2), D1 Féminine teams in 94.0% of cases (*X* <sup>2</sup> = 192.8, *p* < 0.01; *V* = 0.6 *(large)*, *p* < 0.01, *ASR* = 10.9) and Frauen-Bundesliga in 95.0% of the matches (*X* <sup>2</sup> = 174.0, *p* < 0.01; *V* = 0.6 *(large)*, *p*< 0.01, *ASR* = 11.1).

Moreover, must of those teams that scored first did so in the first 15 min (36.9% for Primera Iberdrola, 36% for D1 Féminine and 43.8% for Frauen-Bundesliga). Additionally, the home teams and first scorers won in 52.9%, 55.2% and 53.9% of cases, respectively. There was no statistical evidence suggesting a clear probability of winning when scoring at a specific match time for any league (Primera Iberdrola: *X* <sup>2</sup> = 1.5, *p* = 0.4; *V* = 0.1 *(small)*, *p* = 0.5, *ASR* = 1.2; D1 Féminine: *X* <sup>2</sup> = 12.7, *p* = 0.3; *V* = 0.1 *(small)*, *p* = 0.3, *ASR* = 1.6; Frauen-Bundesliga: *X* <sup>2</sup> = 1.6, *p* = 0.5; *V* = 0.1 *(small)*, *p* = 0.5, *ASR* = 1.2).

*Int. J. Environ. Res. Public Health* **2021**, *18*, 12009


**Table 1.** Descriptive analysis of the result, time point when the first goal was scored, and local status by the league and first scorer.

In Table 2, the descriptive data are shown relative to the number of goals and yellow /red cards, considering if the team scored first or not. First scorers scored a mean of 2.7 goals per match, significantly higher than non-first scorers. There were statistical differences in goals where first scorers scored more goals (*F* = 239.6, *p* < 0.01). No differences were found in number of yellow (*F* = 1.4, *p* = 0.2) or red (*F* = 0.1, *p* = 0.9) cards.

**Table 2.** Descriptive data of final number of goals and red/yellow cards by the first scorer.


Finally, the better-ranked teams usually scored first in all leagues. The association between variables for the best-ranked teams was as follows: (Primera Iberdrola: *X* <sup>2</sup> = 43.5, *p* < 0.01; *V* = 0.2 *(small)*, *p* = 0 < 0.01, *ASR* = 5.9; D1 Féminine: *X* <sup>2</sup> = 30.3, *p* < 0.01; *V* = 0.2 *(small)*, *p* < 0.01, *ASR* = 5.1; Frauen-Bundesliga: *X* <sup>2</sup> = 43.9, *p* < 0.01; *V* = 0.3 (*moderate)*, *p* < 0.01, *ASR* = 5.3).

#### **4. Discussion**

This study aimed to identify which factors could influence scoring first and match outcomes in professional European female football. The results of the analyses suggested that scoring first in female professional football was critical to winning. Female football teams in top European leagues who scored first won in 77–95% of the matches and were better ranked, significantly different results from non-first scorers. More than a third of the first scorers' teams scored in the first 15 min (36–43% of cases) but with no statistical differences compared to other match time points. First scorers scored more goals, but no differences were found in the number of yellow or red cards. Local status also did not influence the final match outcome.

The results of the study confirmed the critical role of scoring first in female professional football. Scoring the first goal may create an advantage, psychologically, tactically, and physically. The evidence suggests that when male football teams are winning, the team usually creates a positive psychological momentum and mindset that makes winning more probable [9]. When scoring first, the conceding team tend to have a higher ball possession than when drawing or wining [22,23]. Additionally, losing teams tend to make more mistakes (e.g., ball interceptions by a rival, fewer clearances) [24] and show more high-intensity actions [14,25], positioning the losing team in a technically and tactically disadvantageous position. Moreover, female losing teams make more tackles, lose the ball more often and accumulate more yellow and red cards than the winning teams [26].

Compared to other sports (50–65%) [8], in female football (77–95%) the odds of winning after scoring first are higher. Compared to male football (65–75%) [3–5], the probability of scoring first in female football is slightly higher. Additionally, these odds depend on the league and could be explained by the greater difference in the teams' quality and the competitive imbalance of female football compared to male football. This could be explained by the higher heterogeneity in the quality of female football compared to male football, with professional and semi-professional female players competing in the same league.

Furthermore, the match time point in which the first goal is scored is also a variable that was studied [4]. Critical moments of the match were identified as essential and are in accordance with the results of this study. The first 15 min of both halves and the last 15 min of the match are critical periods regarding scoring first [27]. These critical moments in the female matches usually see an increase in match workload; depending on the game situation, this increase can be between 20 and 25% [28]. This increase in some periods of the match should be considered when addressing periodisation and strategies to achieve the first goal.

Finally, the evidence suggests that home advantage depends more on the quality of the home team and its rival than on a home effect per se [29]. Indeed, the results of this study indicate that being the first scorer is a determinant regardless of whether the scoring team is the home or the away team. It is also known that teams playing at home tend to show a higher ball possession than teams playing away [22]. Nonetheless, when playing against a stronger opponent, the opposing team must perform better [30], which may cause physical exhaustion and alterations to tactics. In this sense, the best quality team have a more stable pattern of play [13], improving their ability to perform consistent high-intensity actions (e.g., sprints, high-speed running, accelerations, changes of direction), which are essential for physical improvements in female football [31,32].

In female football, the home advantage effect was not as high as in male football. It seems that, in the European leagues, the home advantage in female football is reflected in winning odds of 51–59%; in male football the odds are almost 60% [6]. Some factors could explain these differences, such as the crowd effect (size, intensity and density) on players and referees (referee bias) and gender perceptions of territorial protection and competitive balance [33], usually greater in male football [6]. The aggressivity of the sport and intensity of the match also influence the home advantage. These plausible reasons are supported by other sports studies of the differences in home advantage by gender (e.g., water polo, handball) [34,35]. In Western European football, the evidence suggests a decrease in the home advantage, compared to that of recent decades, due to changes in rules, sport structure and diffusion since the 1980s, leading to a competitive balance [33,36].

Finally, recent studies proposed some tactical, technical and physical factors that may increase the likelihood of winning in female football and could influence scoring first. For example, one-quarter of goals are scored from crosses [37], and so free kicks should be made by a direct free kick or a direct shot on goal [38], pass accuracy should be increased, and a better performance in offensive and defensive duels should be demonstrated [39]. These are technical and tactically significant contributors to victory. Therefore, these actions and situations should be incorporated into training and improved upon to increase the odds of scoring the first goal and winning the match. Additionally, regarding tactics, the winning teams often intercept and recover the ball in more advanced regions of the field than the losing teams [40]; this could suggest the need for some deep pressure strategies during the match, resulting in a higher number of goal attempts. Additionally, some key indicators, such as the high-intensity actions of sprinting, running distances, high ball possession and optimal attacking organisation could influence the match outcome [41]. A home disadvantage is supported by a hypothesis underlying the pressure of winning in front of a supportive audience (expectation of winning). There is a diffusion of responsibility among team members in football compared to other sports, such as basketball or baseball [42]. In other sports, such as hockey and rugby, there is a hypothesis regarding the inhibition of anxiety, which reduces pressure on the home team due to high physical contact [42].

Due to the common playing dynamics and considering that football is a low-scoring sport, with an average of three or fewer goals [1], being the first scoring team seems determinant in female professional football leagues [27]. Scoring first has a strong positive effect that influences the match outcome. These findings can be fundamental for football coaches when developing strategic and tactical planning to enhance the performance of their players, with regard to the different situational variables that their teams may face during the matches.

#### *Limitations*

The main limitation of this study is due to the insufficient number of published scientific articles, which analyse the situational and contextual variables that influence female professional football, specifically those focusing on official match analyses and those factors that impact the final match outcomes. Several studies explored the situational and conditional factors that could influence performance in male football. Based on the basic physiological, genetic, social and cultural differences between women and men [17], such information must be collected and analysed concerning performance and the factors that influence performance in the female sphere. This gender gap reveals the need to better explore female performance in professional football in future studies.

Additional limitations are based on the limited data availability of the different official web pages, with regard to the specification of players' characteristics, weather conditions, and other situational information that could affect the interpretation of the data. While this study analysed the 2018–2019 tournaments of the Spanish, German and French female professional football leagues, recent data are not available or conclusive due to the season cancellations cause by the COVID-19 pandemic.

#### **5. Conclusions**

Scoring first determines the outcomes of official matches from top female European leagues and impacts the final number of goals in a match and the final ranking of the league team. Additionally, it is well known that scoring first provides an advantage during matches and could condition game dynamics, tactics, and programming during a tournament. The top-ranked teams usually have better preparation processes and resources available to develop physical, psychological and tactical skills; these advantages allow them to score the first goal in most cases.

#### *Practical Applications*

Considering the impact of being the first scorer in a female football match, coaches may plan training sessions, bearing in mind that there are physical and psychological aspects to focus on that may boost team performance during the opening minutes of a match, and design tasks that may help achieve the first goal. Different offensive strategies may also help when taking into account physical conditioning, the quality of the opposing team, the available players and their abilities.

Additionally, some training strategies may be devised to overcome a situation where the team is not the first scorer, thus avoiding negativity and encouraging confidence in solving this issue. Moreover, those teams at the bottom of the table may focus their training towards achieving the first goal as a condition that may change the outcome of the matches.

The teams from minor divisions should be aware of these results and consider the critical opening minutes of a match as a determinant for winning the match, and thus the fundamental role of scoring first. Finally, coaches may analyse these tactical and physical situations where the team scores the first goal to understand how it was achieved to boost future performance, considering the variables that may help accomplish the first goal.

The internal and external variables mediating the scoring of the first goal or the offensive tactics that may lead to it should be monitored and considered when planning physical and conditioning training (e.g., high-intensity actions, accelerations, changes of direction).

Future studies could focus on the teams who score first, considering psychological, tactical and physical conditions. Additionally, studies could explore how female footballers, whose teams are losing during a match, can overturn the match outcome.

**Author Contributions:** Conceptualisation, P.S.-M., A.A. and S.J.I.; methodology, P.S.-M. and S.J.I.; software, P.S.-M. and D.R.-V.; validation, A.A., D.R.-V. and S.J.I.; formal analysis, P.S.-M. and D.R.-V.; investigation, P.S.-M.; resources, P.S.-M., A.A. and S.J.I.; data curation, P.S.-M., A.A. and D.R.-V.; writing—original draft preparation, P.S.-M. and D.R.-V.; writing—review and editing, P.S.-M., D.R.-V. and S.J.I.; visualisation, P.S.-M., A.A. and D.R.-V.; supervision, S.J.I.; project administration, S.J.I.; funding acquisition, S.J.I. All authors have read and agreed to the published version of the manuscript.

**Funding:** This study was partially subsidised by the Aid for Research Groups (GR18170) from the Regional Government of Extremadura (Department of Employment, Companies and Innovation), with a contribution from the European Funds for Regional Development of the European Union.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

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

#### **References**


### *Article* **No Relationship between Lean Mass and Functional Asymmetry in High-Level Female Tennis Players**

**Laurent Chapelle 1,\*, Chris Bishop <sup>2</sup> , Peter Clarys <sup>1</sup> and Eva D'Hondt 1,3**


**Abstract:** The relationship between lean mass and functional asymmetry in terms of their magnitude and direction was examined in 22 high-level female tennis players (20.9 ± 3.6 years). Lean mass of both upper and lower extremities was examined using Dual X-ray Absorptiometry. Functional asymmetry was assessed using a battery of field tests (handgrip strength, seated shot-put throw, plate tapping, single leg countermovement jump, single leg forward hop test, 6 m single leg hop test, and 505 change of direction (time and deficit)). Paired sample *t*-tests compared the dominant (overall highest/best (performance) value) against the non-dominant value (highest/best (performance) value of the opposing extremity). Linear regressions were used to explore the relationship between lean mass and functional asymmetry magnitudes. Kappa coefficients were used to examine the consistency in direction between the extremity displaying the highest lean mass value and the extremity performing dominantly across tests. Significant asymmetry magnitudes (*p* < 0.05) were found for all upper and lower extremity lean mass and functional values. No relationship was apparent between lean mass and functional asymmetry magnitudes (*p*-value range = 0.131–0.889). Despite finding perfect consistency in asymmetry direction (k-value = 1.00) for the upper extremity, poor to fair consistency (k-value range = −0.00–0.21) was found for the lower extremity. In conclusion, lean mass and functional asymmetries should be examined independently.

**Keywords:** women; performance; unilateral; racket sport

#### **1. Introduction**

As one of the most popular sports globally, tennis is characterised by short highintensity efforts which are alternated by bouts of recovery [1,2]. During these high-intensity efforts, tennis strokes are performed during which the preferred upper extremity of the player (i.e., the upper extremity holding the racket) is exposed to greater mechanical loading compared to the opposing upper extremity (i.e., the non-preferred upper extremity) [3]. Consequently, this predominantly unilateral sport is ideally suited to examine the occurrence of lean mass asymmetries (i.e., side-to-side differences in lean mass, expressed as a percentage) [4,5]. For instance, using Dual X-ray Absorptiometry (DXA), significant asymmetries between the preferred and non-preferred upper extremity in terms of lean mass (i.e., which includes muscle mass and body water) have previously been reported in both male (i.e., 9.7%) and female (i.e., 6.8%) tennis players [6,7].

In addition to the upper limbs, the lower extremities of tennis players are also subjected to asymmetrical loading due to their specific role in the kinetic chain when performing the various tennis strokes [8,9]. Several previous studies have examined lower extremity lean mass asymmetries by means of DXA in male youth [10], professional male adult [11] and high-level female adult tennis players [6], but reported varying results. For instance, the

**Citation:** Chapelle, L.; Bishop, C.; Clarys, P.; D'Hondt, E. No Relationship between Lean Mass and Functional Asymmetry in High-Level Female Tennis Players. *Int. J. Environ. Res. Public Health* **2021**, *18*, 11928. https://doi.org/10.3390/ ijerph182211928

Academic Editors: Filipe Manuel Clemente and Ana Filipa Silva

Received: 23 September 2021 Accepted: 11 November 2021 Published: 13 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/).

two beforementioned studies examining (youth) male tennis players indicated no significant lower extremity lean mass asymmetries (i.e., 0.6–0.8%), whilst the study examining female tennis players demonstrated significant lower extremity lean mass asymmetries (i.e., 4.8%). An important consideration, however, is that the latter study did not relate these significant side-to-side differences in lower extremity lean mass to players' tennisspecific physical performance (which may increase our knowledge regarding the impact of lean mass asymmetries). Hence, and in addition to the reported contradictory results, more research into (lower extremity) lean mass asymmetries in female tennis players is warranted.

Along with the occurrence of lean mass asymmetry, the presence of functional asymmetry (i.e., side-to-side differences in physical performance (e.g., strength or power), again expressed as a percentage) has also been established. Consequently, significant magnitudes of upper (i.e., 8.9–15.2%) and lower extremity (i.e., 1.8–9.4%) functional asymmetries have previously been reported in high-level female tennis players [12]. When examining functional asymmetries, it is essential to use a composite test battery (as opposed to isolated testing) given the direction specificity of asymmetries (i.e., which extremity displays higher values and/or is dominant in performance) between different sporting tasks [13]. For instance, the beforementioned study in high-level female tennis players reported that the preferred upper extremity consistently demonstrates superior performances compared to the opposing upper extremity. In contrast, the lower extremity was found to display poor levels of agreement as to which leg performed better across tests (i.e., the kappa coefficients ranged from −0.07 to 0.17), illustrating the direction specificity of lower extremity functional asymmetries [12].

It is important to note that both lean mass and functional asymmetries have, albeit separately, been associated with a decreased sport-specific performance, in addition to an increased injury risk [13–15]. However, no study has simultaneously examined both types of asymmetry. As a result, research regarding the relationship between lean mass asymmetry and functional asymmetry, both at the upper and lower extremity level, is currently lacking. More specifically, it is unknown whether a high(er) magnitude of lean mass asymmetry implies a high(er) magnitude of functional asymmetry (i.e., which could be the case since muscle mass (which entails lean mass) is reported to be a key determinant of functional strength and power) [16]. Similarly, regarding the agreement in direction between lean mass and functional asymmetry, it is unknown whether the extremity that displays the highest lean mass value also displays the best performance across body sides. As a result, the mutual relationship and the agreement in direction between both lean mass asymmetry and functional asymmetry remains to be investigated. Due to the lack of previous research in this respect (especially in female tennis players), this study aimed to examine the relationship between lean mass and functional asymmetry in terms of their magnitude and direction in high-level female tennis players.

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

#### *2.1. Participants*

Twenty-two high-level Belgian female tennis players aged between 17 and 27 years participated in this observational cross-sectional study. To be eligible for participation, these female tennis players had to be injury-free at the time of measurement and either have an international tennis ranking (i.e., Women's Tennis Association or International Tennis Federation) or a high national tennis ranking (i.e., being in the top 200 of the Belgian circuit ranking). Our study protocol was approved by the local university's medical ethics committee prior to data collection (B.U.N. 143201836107). The female tennis players, together with their legal guardians if they were still minor, were informed about the purposes of this study and signed an informed consent upon participation.

#### *2.2. Procedures and Experimental Design*

Data collection took place in the local university's biometry and biomechanics laboratory facilities between March 2019 and September 2020. Firstly, the female tennis players were asked to fill in a questionnaire to provide basic demographic and sport-specific information (i.e., date of birth, dominant upper extremity, starting age of tennis play and average weekly training volume over the last year). Next, after voiding their bladder and whilst being barefoot in light sports clothing, participants' body height and weight were measured to the nearest 0.1 cm and 0.002 kg using a stadiometer (SECA 217, Hamburg, Germany) and precision scale (RADWAG WLT 60/120/X/L3, All scales Europe, Veen, The Netherlands), respectively. Table 1 presents the demographic, sport-specific and anthropometric information of the 22 female tennis players included in our study sample.

**Table 1.** Demographic, sport-specific and anthropometric information of the high-level female tennis players (N = 22).


Note: Data are presented as n or mean ± standard deviation.

#### *2.3. Lean Mass*

DXA research scans (Norland Elite, Swissray, Fort Atkinson, WI, USA) of both the preferred and non-preferred upper extremity as well as the right and left lower extremity were conducted by the same researcher, who was intensively trained by the DXA scan manufacturer upon data collection, in order to determine participants' regional lean mass to the nearest 0.1 g. The DXA scanner was calibrated in accordance with the manufacturer's guidelines before each test session. Participants were instructed to lie as straight and still as possible in a supine position on the DXA scan table after the removal of all metal objects (e.g., earrings). The scan width was set to 6 × 6 mm, whilst a scan speed of 130 mm/s was applied. The upper extremity region included the upper arm, lower arm and hand, and was separated from the trunk by an inclined line passing through the scapula-humeral joint. The lower extremity region included the upper leg, lower leg and foot, and was separated from the trunk by an inclined line passing just below the pelvis [11]. The DXA research scans were analysed with the Norland Illuminatus software (Swissray, Fort Atkinson, WI, USA).

#### *2.4. Functional Test Battery*

A physical performance field-based test battery was used to examine the magnitude of functional asymmetry. Participants were instructed to wear their normal tennis outfit and sports shoes whilst performing the test battery, consisting of 8 different unilateral tests. A standardised 10-min warm-up, involving light running exercises and dynamic stretches, was implemented before completing the test battery. The different tests were always completed in the same order, ensuring alternation in testing the upper and lower extremities. The participants were guided through the test battery by the same well-trained researcher. Each participant was given three attempts per body side for every test. The first attempt of a test was always performed with the right body side, whereas the second attempt was always performed with the left body side, ensuring alteration between both sides of the body during testing. Participants were given 60 s of rest between attempts and 3 min of rest between tests to ensure adequate recovery.

Handgrip strength: Participants were instructed to squeeze as hard as possible (for three seconds) in a digital handheld dynamometer with an accuracy of 0.1 kg (Jamar Plus, Patterson Medical, Nottinghamshire, UK), while being seated in a chair without armrests. The elbow of the participants had to remain 90 degrees flexed throughout every attempt [17].

Seated shot-put throw: Participants were seated on the ground with their back against a wall and their hips, knees and ankles parallel to the ground. The non-throwing arm was placed on the opposite (i.e., throwing) shoulder. From this position, participants had to throw a 3-kg medicine ball as far as possible in a forward direction. The distance where the medicine ball landed on to the ground was measured to the nearest 1 cm using a tape measure [18].

Plate tapping: Two discs (with a diameter of 20 cm) were placed with their centres 60 cm apart on a table together with a 10 × 20 cm rectangle (which was placed in between the two discs). Participants started the test with one hand placed on one of the two discs, whilst the other hand was placed on the rectangle in the middle. The aim of the plate tapping test was to move one hand back and forth between both discs over the other hand (which was on the rectangle) as fast as possible. This action was repeated for 25 full cycles (i.e., 50 taps on the discs) and the time needed to complete this test was recorded to the nearest 0.01 s using a hand-held stopwatch [19].

Single leg countermovement jump: Participants were instructed to jump up as high as possible on one leg. Throughout the jump, they were instructed to hold their hands on their hips. Swinging of the non-jumping leg was not allowed and the jumping leg had to remain completely extended throughout the flight phase. Participants needed to keep their balance on one leg after landing, otherwise an extra attempt was provided. Jumping height was determined to the nearest 0.1 cm using the Optojump Next system (Microgate Bolzano, Italy) [20].

Single leg forward hop test: Participants stood on one leg behind a tape line whilst holding their hands on the hips. They had to jump as far as possible in a forward direction landing on the same foot without losing their balance (e.g., moving their foot on which they land or planting the other foot on to the ground). If participants were not able to maintain their balance on one leg after landing, an extra attempt was provided. The covered distance from the starting line to the heel of the participants' landing foot was measured to the nearest 1 cm using a tape measure [21].

6 m single leg hop test: Participants were instructed to cover 6 m as fast as possible whilst hopping on one leg. The time needed to cover these 6 m was measured to the nearest 0.001 s using electronic timing gates (Witty Wireless Training Timer, Microgate, Bolzano, Italy). These timing gates were placed at hip height and participants had to start behind a tape line which was located 30 cm from the first timing gate.

505 change of direction time (505 COD time) and deficit (505 COD deficit): First, participants' 10 m sprint time was measured to the nearest 0.001 s using electronic timing gates (Witty Wireless Training Timer, Microgate, Bolzano, Italy). Next, their 505 COD time was determined to the nearest 0.001 s based on performing the 505 COD test, which consisted of a 5 m sprint, followed by a 180◦ turn to either the left or the right side, and a 5 m sprint back to the starting line. Participants' 505 COD deficit was then calculated by deducting their 10 m sprint time from their 505 COD time [22].

#### *2.5. Asymmetry Calculations*

The dominant value was defined as the highest lean mass value or the best (i.e., highest or fastest) value for a test of the functional test battery. The non-dominant value was defined as the highest or best result of the same outcome measure for the opposing upper or lower extremity [23]. The magnitude of lean mass and functional asymmetry was calculated for every outcome measure and expressed as a percentage by using the percentage difference method (PDM): (dominant value − non-dominant value)/dominant value) × 100 [24].

#### *2.6. Statistical Analyses*

Data analysis was conducted using SPSS version 27.0 (IBM, Chicago, IL, USA). Normality of distribution was examined for every outcome measure using the Shapiro–Wilk test. Variability and reliability of every outcome measure was verified by calculating the coefficient of variation (CV) and a two-way random intraclass correlation coefficient (ICC) with 95% confidence intervals. CV values of less than 10% were considered acceptable and ICC values were classified as poor (<0.50), moderate (0.50–0.74), good (0.75–0.89) and excellent (>0.90) [25,26]. Paired sample *t*-tests were used for within-subject comparisons of the dominant against the non-dominant values for every outcome measure. Effect size analyses using Hedges' *g* were conducted of the side-to-side difference between the dominant and non-dominant values and classified as trivial (<0.20), small (0.20–0.49), medium (0.50–0.79) or large (>0.80) [27]. A linear regression analysis, adjusting for the participants' age, was used to examine the relationship between the magnitude of lean mass asymmetry and the magnitude of functional asymmetry [28]. Lastly, the consistency in direction as to which extremity displayed the dominant lean mass value and which extremity performed dominantly across the different field tests of the functional test battery was examined using Kappa coefficients. These Kappa coefficients were classified as poor (≤0), slight (0.01–0.20), fair (0.21–0.40), moderate (0.41–0.60), substantial (0.61–0.80), almost perfect (0.81–0.99) and perfect (1.00) [29]. All data are presented as means ± standard deviations and *p*-values <0.05 were considered statistically significant.

#### **3. Results**

Every outcome measure showed acceptable reliability (i.e., all CVs were below 10%) and excellent reliability (i.e., all ICCs were above 0.90) as presented in Table 2. The lean mass and functional asymmetry values for our study sample of high-level female tennis players are displayed in Table 3. Significant magnitudes of lean mass and functional asymmetry for all outcome measures were found (t-value range = 4.027–8.638; *p* < 0.001). Effect sizes between the side-to-side differences of the dominant and non-dominant values ranged from small to large.


**Table 2.** Variability and reliability of the DXA research scans and the unilateral tests of the functional test battery.

Note: DXA = Dual X-ray Absorptiometry; CV = coefficient of variation; ICC = intraclass correlation coefficient; 95% CI = 95% confidence interval.

(N = 22).

505 change of direction


**Table 3.** Upper and lower extremity lean mass and functional asymmetry values of the high-level female tennis players (N = 22).  **Dominant Value Non-Dominant Value ES (95% CI) PDM (%)**  Lean mass

*Int. J. Environ. Res. Public Health* **2021**, *18*, x 6 of 13

**Table 3.** Upper and lower extremity lean mass and functional asymmetry values of the high-level female tennis players

Note: Data are presented as mean ± standard deviation; ES = effect size; 95% CI = confidence interval; PDM = percentage difference method; \* Significant (*p* < 0.05) magnitude of asymmetry between body sides. 6 m single leg hop test (sec) 1.938 ± 0.168 2.010 ± 0.172 0.42 (−0.09, 0.84) 3.6 ± 3.2 \*

For every field-based test, the corresponding individual lean mass asymmetry magnitudes alongside functional asymmetry magnitudes are displayed in Figure 1 for the upper extremity and in Figure 2 for the lower extremity. No significant relationship between the magnitude of lean mass asymmetry and the magnitude of functional asymmetry (F-value range = 0.021–3.461; r-value range = −0.232–0.254; *p*-value range = 0.131–0.889) was found as lean mass asymmetry magnitude could only explain 0.1 to 15.9% of the functional asymmetry magnitude. Time (sec) 3.249 ± 0.174 3.311 ± 0.181 0.34 (−0.16, 0.84) 1.9 ± 1.7 \* Deficit (sec) 1.144 ± 0.109 1.207 ± 0.120 0.54 (−0.03, 1.04) 5.0 ± 4.3 \* Note: Data are presented as mean ± standard deviation; ES = effect size; 95% CI = confidence interval; PDM = percentage difference method; \* Significant (p < 0.05) magnitude of asymmetry between body sides. For every field-based test, the corresponding individual lean mass asymmetry magnitudes alongside functional asymmetry magnitudes are displayed in Figure 1 for the up-

The consistency in direction between the upper extremity displaying the dominant lean mass value and the upper extremity performing dominantly on the tests of the functional test battery was classified as perfect. For the lower extremity, the consistency between the lower extremity displaying the dominant lean mass value and the lower extremity performing dominantly across tests were classified from poor to fair (Table 4). per extremity and in Figure 2 for the lower extremity. No significant relationship between the magnitude of lean mass asymmetry and the magnitude of functional asymmetry (Fvalue range = 0.021–3.461; r-value range = −0.232–0.254; *p*-value range = 0.131–0.889) was found as lean mass asymmetry magnitude could only explain 0.1 to 15.9% of the functional asymmetry magnitude.

**Figure 1.** *Cont.*

**Figure 1.** Scatter plot illustrating the relationship between the magnitude of upper extremity lean mass asymmetry (x-axis) and the magnitude of upper extremity functional asymmetry (y-axis) for the high-level female tennis players (N = 22). Note: The dotted line represents the linear trend line; PDM = percentage difference method; r = correlation coefficient; R² = R squared value. **Figure 1.** Scatter plot illustrating the relationship between the magnitude of upper extremity lean mass asymmetry (x-axis) and the magnitude of upper extremity functional asymmetry (y-axis) for the high-level female tennis players (N = 22). Note: The dotted line represents the linear trend line; The consistency in direction between the upper extremity displaying the dominant lean mass value and the upper extremity performing dominantly on the tests of the functional test battery was classified as perfect. For the lower extremity, the consistency between the lower extremity displaying the dominant lean mass value and the lower extremity performing dominantly across tests were classified from poor to fair (Table 4).

r = −0.232; R² = 0.087

**Figure 2.** *Cont.*

*Int. J. Environ. Res. Public Health* **2021**, *18*, x 8 of 13

**Figure 2.** *Cont.*

**Figure 2.** Scatter plot illustrating the relationship between the magnitude of lower extremity lean mass asymmetry (x-axis) and the magnitude of lower extremity functional asymmetry (y-axis) for the high-level female tennis players (N = 22). Note: The dotted line represents the linear trend line; PDM = percentage difference method; r = correlation coefficient; R² = R squared value. **Figure 2.** Scatter plot illustrating the relationship between the magnitude of lower extremity lean mass asymmetry (x-axis) and the magnitude of lower extremity functional asymmetry (y-axis) for the high-level female tennis players (N = 22). Note: The dotted line represents the linear trend line; PDM = percentage difference method; r = correlation coefficient; R<sup>2</sup> = R squared value.

lean mass value and the upper extremity performing dominantly on the tests of the functional test battery was classified as perfect. For the lower extremity, the consistency between the lower extremity displaying the dominant lean mass value and the lower ex-**Table 4.** Kappa coefficients indicating the consistency in direction between the dominant lean mass value and the dominant performance value across unilateral tests for the high-level female tennis players (N = 22).

The consistency in direction between the upper extremity displaying the dominant


Plate tapping 1.00 Perfect Lower extremity lean mass Note: Kappa coefficients are classified as poor (≤0), slight (0.01–0.20), fair (0.21–0.40), moderate (0.41–0.60), substantial (0.61–0.80), almost perfect (0.81–0.99) and perfect (1.00).

Single leg forward hop test 0.00 Poor 6 m Single leg hop test 0.18 Slight 505 Change of direction time/deficit 0.21 Fair Note: Kappa coefficients are classified as poor (≤0), slight (0.01–0.20), fair (0.21–0.40), moderate (0.41–0.60), substantial (0.61–0.80), almost perfect (0.81–0.99) and perfect (1.00). The consistency in direction between the upper extremity displaying the dominant lean mass value and the upper extremity performing dominantly on the tests of the functional test battery was classified as perfect. For the lower extremity, the consistency between the lower extremity displaying the dominant lean mass value and the lower extremity performing dominantly across tests was classified from poor to fair, depending on the field test at hand (Table 4).

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

This observational cross-sectional study aimed to examine the relationship between lean mass and functional asymmetry in terms of their magnitude and direction in highlevel female tennis players. The results of our study indicated no meaningful relationships between the magnitude of lean mass asymmetry and functional asymmetry in either the upper or the lower extremities. Additionally, consistency in the direction of asymmetry between the extremity displaying the highest lean mass value and the extremity displaying the dominant performance value for the functional tests across body sides was perfect for the upper extremity, whereas this consistency in dominance for both types of asymmetry ranged from poor to fair as regards to the lower extremity. This observational cross-sectional study aimed to examine the relationship between lean mass and functional asymmetry in terms of their magnitude and direction in highlevel female tennis players. The results of our study indicated no meaningful relationships between the magnitude of lean mass asymmetry and functional asymmetry in either the upper or the lower extremities. Additionally, consistency in the direction of asymmetry between the extremity displaying the highest lean mass value and the extremity displaying the dominant performance value for the functional tests across body sides was perfect for the upper extremity, whereas this consistency in dominance for both types of asymmetry ranged from poor to fair as regards to the lower extremity.

Single leg countermovement jump 0.18 Slight

The significant magnitude of upper extremity lean mass asymmetry found in this study (i.e., 7.1%) can be largely attributed to the mechanical loading imposed to the preferred upper extremity associated with the repetitive performance of tennis strokes [4]. Interestingly, the preferred upper extremity of all high-level female tennis players included in the present always displayed the highest lean mass value. In agreement with the results of the upper extremity, significant lower extremity lean mass asymmetries (i.e., 4.8%) were found in our sample of Belgian high-level female tennis players. Even though most of them were right-handed (i.e., 21 out of 22 players), the majority displayed a higher lean mass of the left leg compared to the right leg (i.e., 18 out of 22 players). This could be explained by the previously reported occurrence of cross-asymmetry where the contralateral leg (i.e., the leg opposed to the preferred upper extremity) plays an important role in counterbalancing the torques of the upper extremity performing the various tennis strokes [6,8,9]. It is important to consider that the present study compared the dominant versus the nondominant value to examine and report lower extremity lean mass asymmetries as opposed to using the values of the self-reported preferred lower extremity by asking, for example, on which leg participants prefer to perform a single leg hop [24]. The latter could lead to an incorrect calculation of the asymmetry magnitude as a percentage should be calculated with respect to the highest value [24,30].

The magnitude of upper extremity functional asymmetry ranged from 9.5 to 13.2% in our study, which is indicative of significant inter-limb asymmetries. Again, these significant inter-limb asymmetries can be principally attributed to the predominantly unilateral nature of tennis [3]. It is important to note that the preferred upper extremity of the included high-level female tennis players always performed dominantly across all upper extremity tests. Although lower than the magnitude of upper extremity functional asymmetry, the overall magnitude of functional asymmetry at the lower extremity level ranged from 1.9 to 8.4%, indicating significant functional asymmetries for all lower extremity performance tests. However, due to the task specificity of lower extremity functional asymmetries, there was no occurrence of cross-asymmetry across the functional tests for the lower extremity, as also mentioned in earlier research [12]. The highest asymmetry magnitude was found for the single leg countermovement jump (i.e., 8.4%). This result is in agreement with previous studies that have reported jump height from the single leg countermovement jump as being a sensitive physical performance test to examine functional asymmetries, especially when compared to jumping in a forward direction [31,32]. Nevertheless, it can be argued that it is surprising to find significant lower extremity functional asymmetries in a study sample of high-level female tennis players because being equally physically skilled on both lower extremities could be advantageous from a performance perspective [13].

As indicated by the results of this study, lean mass asymmetry and functional asymmetry do not seem to be related in terms of their magnitude given that lean mass asymmetry magnitude could only explain between 0.1 and 15.9% of the functional asymmetry magnitude. This is surprising because lean mass (which also encompasses muscle mass) has been reported to be a key determinant of functional strength and power [16], although it has been reported that other factors such as neuromuscular control and joint coordination also contribute to strength and power development [15,33]. Therefore, it is recommended that practitioners examine lean mass and functional asymmetries independently from one another. Additionally, the non-existent relationship between lean mass and functional asymmetry may have implications when designing targeted training programmes to counteract the reported negative influences of asymmetry (as it is unclear whether practitioners should focus on lean mass and/or functional parameters) [13–15]. Regarding the direction of asymmetry, there was a poor to slight consistency between the lower extremity displaying the dominant lean mass value and the lower extremity performing dominantly across the functional tests. This result was in contrast to the upper extremity, which displayed perfect levels of agreement. Consequently, the reported lower extremity results in this respect highlight the task and direction specific nature of asymmetry during the execution of different tasks, with Kappa values of the present study being comparable to

those in previous research [20,32]. Because the extremity displaying the highest lean mass value does not consistently perform dominantly at lower limb level, it is recommended that practitioners examine and interpret both lean mass and functional asymmetry in an independent manner. Additionally, the assessment of asymmetries should be performed regularly and on an individual player basis, so that an asymmetry profile can be made to closely monitor each tennis player [20,34].

This is the first study to examine and report both lean and functional asymmetry of the upper and the lower extremity in high-level female tennis players using individual data. It can be argued that high-level tennis players are well suited to examine asymmetries because reaching such a level requires a high training volume and given the reported association between a high training volume and the occurrence of asymmetry [10]. Additionally, all players included in our study sample started to play tennis before the onset of puberty, which has been reported to result in greater asymmetry magnitudes [4]. Furthermore, functional asymmetry was examined using a valid, reliable and elaborated field-based test battery, as opposed to isolated testing, which is important given that asymmetries are reported to be movement or task-specific [34], as clearly demonstrated by our findings at the level of the lower limb. However, some limitations to our research are apparent. The present study implemented a cross-sectional design, which included a small sample size (although a post hoc power analysis revealed that the statistical power of this study was 91%). However, a control group was not included and the association between lean mass and functional asymmetry with decreased sport-specific performance, and injury incidence, was not examined. Therefore, future research is needed to examine the influence of lean mass and functional asymmetry on sports-specific performance and injury incidence using a longitudinal design. Additionally, more precise tools (e.g., force plates or isokinetic dynamometry) and outcome measures (e.g., leg stiffness, ground contact time or force) could be used when examining functional asymmetries [35].

#### **5. Conclusions**

To conclude, the significant lean mass and functional asymmetries of both the upper and lower extremity were not related in terms of their magnitude among high-level female tennis players. Additionally, the consistency between the extremity displaying the dominant lean mass value and the extremity displaying the dominant performance value across the functional tests was perfect for the upper extremity, whereas this consistency ranged from poor to fair for the lower extremity. When examining asymmetries in tennis players, it is recommended that both the magnitude and direction thereof should be considered and interpreted independently of one another in view of asymmetry profiling because no mutual relationship between both constructs could be demonstrated. It is also essential to examine and monitor both upper and lower extremity asymmetries on an individual player basis and to examine functional asymmetries using an elaborated field-based test battery. Future more in-depth research is also needed to investigate the impact of lean mass and functional asymmetries on female players' sports-specific performance and injury incidence using longitudinal (and/or experimental) study designs.

**Author Contributions:** L.C. recruited the participants, collected all data and also took responsibility for data analysis and drafting the manuscript. C.B., P.C. and E.D. substantially contributed to the conception of the study design, the interpretation of the data and critically revised the manuscript's drafting. All authors have read and agreed to the published version of the manuscript.

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

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Ethics Committee of the Vrije Universiteit Brussel (B.U.N. 143201836107).

**Informed Consent Statement:** Informed consent was obtained from all participants involved in the study before data collection took place.

**Data Availability Statement:** The data that support the findings of this study are available from the corresponding author upon request.

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

#### **References**


### *Article* **The Effects of Running Compared with Functional High-Intensity Interval Training on Body Composition and Aerobic Fitness in Female University Students**

**Yining Lu <sup>1</sup> , Huw D. Wiltshire <sup>1</sup> , Julien S. Baker <sup>2</sup> and Qiaojun Wang 3,\***



**Citation:** Lu, Y.; Wiltshire, H.D.; Baker, J.S.; Wang, Q. The Effects of Running Compared with Functional High-Intensity Interval Training on Body Composition and Aerobic Fitness in Female University Students. *Int. J. Environ. Res. Public Health* **2021**, *18*, 11312. https://doi.org/10.3390/ ijerph182111312

Academic Editors: Filipe Manuel Clemente and Ana Filipa Silva

Received: 13 October 2021 Accepted: 25 October 2021 Published: 28 October 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/).

**Abstract:** High-intensity interval running (HIIT-R) and high-intensity functional training (HIFT) are two forms of HIIT exercise that are commonly used. The purpose of this study was to determine the effects of HIFT on aerobic capacity and body composition when compared to HIIT-R in females. Twenty healthy, untrained female university students (age 20.5 ± 0.7 year) were randomly assigned to a 12-week HIIT-R or HIFT intervention. The HIIT-R group involved a 30 s maximal shuttle run with a 30 s recovery period, whereas the HIFT involved multiple functional exercises with a 2:1 work-active recovery ratio. Body composition, VO2max, and muscle performance were measured before and post intervention. As a result, HIIT-R and HIIT-F stimulated similar improvements in VO2max (17.1% ± 5.6% and 12.7% ± 6.7%, respectively, *p* > 0.05). Only the HIIT-F group revealed significant improvements in muscle performance (sit-ups, 16.5% ± 3.1%, standing broad jump 5.1% ± 2.2%, *p* < 0.05). Body fat percentage decreased (17.1% ± 7.4% and 12.6% ± 5.1%, respectively, *p* < 0.05) in both HIIT-R and HIIT-F with no between-group differences. We concluded that HIFT was equally effective in promoting body composition and aerobic fitness compared to HIIT-R. HIFT resulted in improved muscle performance, whereas the HIIT-R protocol demonstrated no gains.

**Keywords:** high-intensity interval training; high-intensity functional training; body composition; aerobic fitness; muscle performance

#### **1. Introduction**

Regular physical activity (PA) is beneficial for health [1–3]. Despite the well documented benefits of moderate- to vigorous-intensity PA, 31% of adults worldwide do not engage in sufficient PA for health benefits as recommended by the World Health Organization (WHO) and the American College of Sports Medicine (ACSM) [4–6]. Frequently reported barriers to physical activity are physical exertion, time, and financial expenditure [7,8]. Thus, compared to traditional continuous training, which is characterized by long-duration, continuous aerobic exercises, and moderate-intensities, high-intensity interval training (HIIT) appears to be an efficient pathway to enhance PA and improve health [9].

HIIT involves repeated bouts of high-intensity exercises separated by a recovery using low-intensity activities or inactivity [10]. Recent studies had indicated that HIIT has a similar, or even greater positive, effect on physical fitness, especially on body composition and cardiorespiratory health [11–14]. From a time/benefit perspective, HIIT appears to help physically inactive individuals overcome a major time and participation barrier to maintaining a healthier lifestyle [15].

Originally, HIIT was used to improve the performance of endurance athletes [16]. Cycling, running, and rowing are traditional exercise modalities that adopted the use of HIIT protocols, while for individuals who perform exercise for health and recreation, these traditional modalities seem boring and do not engage individuals because of the repetitive nature of the exercise combined with repetition. This is considered as a negative impact for maintaining regular exercise and has been cited as "lack of enjoyment" when investigating barriers to exercise [17].

The intrinsic factors of participants are also important when considering exercise adherence [18,19]. Several studies have revealed that adherence is affected by exercise intensity, especially among inactive individuals [20,21].

High-intensity functional training (HIFT) has become a relatively popular training modality in recent years and is an alternative to traditional aerobic activities. The HIFT protocol consists of a variety of functional movements that are executed at a high intensity [22,23]). Recently, several investigators have studied the effects of HIFT on physical fitness promotion. After engaging in HIFT protocols, participants show significant improvements in cardiorespiratory fitness [24,25] and body composition [25,26]. Providing similar or greater health promotions compared to moderate-intensity continuous training, HIFT demonstrates further improvements in muscle fitness [27,28]. Additionally, participants perceive this type of activity to be more enjoyable when engaging in HIFT compared to those individuals performing traditional HIIT [29,30]. Moreover, most HIFT protocols are executed using the participant's own body weight, allowing the participant to control the exercise intensity. This helps to improve exercise adherence [7,19,20].

Although studies have shown that HIFT has similar or superior benefits for physical fitness compared to moderate-intensity continuous training and have indicated more enjoyment compared to HIIT, the question remains as to whether HIFT is as efficient as HIIT for improving health-related fitness.

While HIFT is not synonymous with HIIT, they share an important conceptual commonality in the modality of both being of a high intensity. The current study was undertaken to clarify how a functional exercise based on HIIT would improve fitness parameters such as fat mass, blood pressure, VO2max, and muscle endurance following a 12-week intervention compared to changes achieved using a running-based HIIT. The purpose of this study was to investigate the effects of different kinds of training on fitness parameters in untrained female university students. It was hypothesized that (a) aerobic fitness would be increased in both the HIIT-F and HIIT-R groups; (b) that fat mass would be decreased in both the HIIT-F and HIIT-R groups; and (c) that muscular strength and endurance would be improved in the HIIT-F group.

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

#### *2.1. Participants*

Twenty untrained healthy females who were physical inactive volunteered to participate the study. Participants who did not exercise for more than 2 h weekly for at least 12 months were considered as physically inactive [31]. All of the participants were in their second year of a non-physical education-related degree at Ningbo University. Similar self-reported menstrual cycles were required, ensuring the simultaneity of testing and training. Interventions were suspended for 1 week during menstruation, and the normal menstruation period lasted for 3 to 10 days [32,33]. A randomized controlled research design was utilized, and participants were randomly assigned into a running-based HIIT (HIIT-R) (*n* = 10) or a functional training-based HIIT (HIIT-F) (*n* = 10). The participants were nonsmokers and were instructed to maintain their normal dietary intake and lifestyle habits (sleep, sit, and physical activity) throughout the intervention. Nutritional supplements and intense exercise beyond their usual exercise habits were forbidden during the intervention period [31]. All of the participants were fully familiarized with the test procedures and data collection methods prior to the intervention. Written informed consent was provided

by all participants. The study was approved by the Ningbo University ethics committee. The characteristics of the participants at baseline are detailed in Table 1.


**Table 1.** Baseline characteristics of HIIT-R and HIIT-F group.

Notes: BMI, body mass index; bpm, beats per minute, HIIT-F, functional exercise-based high-intensity interval training; HIIT-R, running-based high-intensity interval training; HR resting, resting heart rate; HR max, maximal heart rate; VO2max, maximal oxygen uptake; WHR, waist to hip ratio; yrs, years old.

#### *2.2. Procedures*

A randomized controlled trial was used in this study. Each participant completed twelve weeks of 36 sessions of HIIT-R or HIIT-F intervention (three sessions per week) comprising a total of 19 min per session (10 min warm-up, 4 min work-out, and 5 min cool-down). All sessions were conducted and monitored at the same indoor stadium and at the same time of day between 9:00–10:00 a.m. Heart rates (HR) were collected with an activity wristband (Mi Smart Band 5, Xiaomi, Beijing, China) during each session to ensure that the required high intensity was achieved. The reliability and validity of the heart rate index and distance index were reported in a previous study [34]. The activity wristband was required to be worn tightly on the participant's wrist. The HR index was measured based on changes in the light transmittance caused by blood flow density using optical sensing technology, and the distance index was measured by a triaxial acceleration sensor. Two measurement time points (pre- and post-intervention) were included. The participants were instructed to abstain from drugs, alcohol, and intense exercise two days prior to the baseline and post-intervention measurements. On the first measurement day, the participants presented themselves at 8:00 a.m. and underwent a body composition analysis, physical, and physiology measures as well as resting heart rate (HRresting) and blood pressure (BP) measurements under standardized conditions. The aerobic fitness assessment was conducted using a 12 min running test, which was completed on two days, with 24 h observed between each test. The first running test was scheduled on the first measurement day following the completion of all of the other tests, and the second trial was 24 h later. The average of the two data sets was used to assess aerobic fitness. After resting for a week [35], both groups began the training intervention. Post-intervention measurements were performed using the same methodologies as at baseline and were undertaken two days following all of the training sessions [11]. During the intervention period, additional exercises including habitual training were suspended.

#### *2.3. Physical, Physiological and Body Composition Assessment*

Participants were instructed to arrive at the laboratory 9:00 a.m. after a normal breakfast. Before the measurements were taken, participants were asked to empty their bladder to minimize measurement errors caused by "electrically silent" [36]. Under the guidance of two skilled operators and while wearing normal PE clothing, the participants stood on a bioelectrical impedance analysis device (BIA) (MC-180, TANITA CO., Dongguan, China) and data were presented from the device's associated software and included height, weight, waist and hip circumference, lean muscle mass, and body fat percentage. Body mass index

(BMI) was obtained by dividing weight (kg) by height (m) squared. Waist-to-hip ratio (WHR) was obtained by dividing waist (cm) by hip (cm). Blood pressure and resting HR were measured using an automatic upper arm blood pressure monitor (HEM-1000, Omron, Dalian, China). The average of the two data sets was used for analysis.

#### *2.4. Aerobic Fitness Test*

The most reliable and effective way to measure aerobic capacity is to record each individual subject's VO2max [37]. Although maximal-effort tests are commonly used to measure VO2max, for untrained participants, submaximal exercises can be used as a reliable measure to estimate this value. Cooper's 12 min running test was used to assess aerobic fitness in this study. All of the participants completed two trials of the running test separated by 24 h of rest. After a 5 min warm up, the participants were required to wear an activity wristband (Mi Smart Band 5, Xiaomi, China) and commenced running on a standard 400-metre running track. Subjects were instructed to run as many laps as possible on a standard outdoor track during the 12 min test period. All of the participants were encouraged verbally and were instructed to focus on their own pace throughout the test. The experimenter verbally provided the elapsed time at 3, 6, and 9 min. At the end of the 12 min period, the experimenter called "stop". All of the participants ceased running and stood still, until the distance achieved, and maximal heart rate (HRmax) were recorded. The HRmax displayed on the activity wristband was recorded immediately upon the cessation of exercise, and the higher value of the two trials was used for analysis. The total distance run was determined by measures obtained from the activity band. An estimated VO2max was calculated using Cooper's standardized equation [38]. The calculated VO2max was highly correlated with the laboratory-determined one and had acceptable reliability and validity (r = 0.897) [38]. The average of the two data sets was used to determine the VO2max.

#### *2.5. Muscle Performance Test*

Muscle performance was assessed using a field-based muscle fitness test battery. Timed sit-ups, push-ups with flexed knee (modified for females), and standing broad jump were recommended by previous studies to assess muscle performance [26,39–41]. All of the participants were instructed to perform the tests under supervision, and the data were recorded by the same experimenter. To assess abdominal muscular performance, the participants were asked to perform as many sit-ups as possible during a one-minute test period. The number of sit-ups that were completed correctly were recorded. A sit-up that met the following criteria was recorded: the participant lay supine on the mat with their hands crossed behind their head, elbows pointed straight forward, and knees bent at 90 degrees. The ankles were firmly held by the experimenter. During the execution of the test, the participants sat up with their heads clasped in their hands, and then their elbows touched or went over the knees, and the participant went back with their shoulders touching the mat [42]. To assess upper body strength and endurance, the flexed knee push-up option was used as a gender modification [39]. A correct flexed knee push-up met the following criteria: participants knelt on the mat with their knees bent to the mat with their arms propped on the mat slightly wider than the shoulders. When the test began, the participants were instructed to lower their body by bending their arms until their elbows were bent at a 90-degree angle and their chest was placed within 2 inches of the mat, subjects then pushed up to the starting position [43,44]. The number of correctly completed push-ups during a one-minute test period was recorded as upper body strength and endurance. Finally, the standing broad jump test was used to assess the muscle power of the lower limbs. The participants wore sneakers and stood behind the starting line with their feet placed naturally at a shoulder width apart. When testing began, the participants were instructed to bend the knees, swing the arms, and jump with both feet at the same time [45]. The jumping distance measured in centimeters was recorded, and the best of

three jumps was used to determine lower limb performance. All scores were compared for statistical analysis.

#### *2.6. Intervention*

Exercise interventions commenced one week after the last measurement day. Both the HIIT-R and HIIT-F interventions were conducted three days per week on Mondays, Wednesdays, and Saturdays for twelve weeks. If the participants were unable to attend a scheduled exercise day, the exercise was performed on the next day and was monitored by the same researcher.

Participants in the HIIT-R group were required to complete 144 repetitions of maximal shuttle running for a total exercise time of 72 min. Each bout included a 30 s maximal shuttle run between cones placed 20 m apart with a 30 s recovery period between runs. The validity and reliability of 40 m maximal shuttle run as a measure of anaerobic performance has been reported previously [46]. The participants completed 4 bouts per session over three sessions per week. Prior to the intervention, a familiarization trial was provided to acquaint the participants with the training procedure. Running and recovery times were recorded manually using a digital stopwatch by the same experimenter. Participants were encouraged to run at their individual maximal speed for each bout.

Participants in the HIIT-F group performed multiple functional exercises using their own body weight based on Tabata training [47]. According to a recent study [48], eight movements were implemented in each session (Table 2). Participants were motivated to complete as many repetitions of a given movement as possible over 20 s followed by a 10 s recovery in the form of low intensity stepping. There was no rest period between each movement. The total training time for each session was 4 min.


**Table 2.** Details of the functional high-intensity interval training intervention.

The training frequency was the same as the HIIT-R group. All training exercises were recorded by video, which was provided to the HIIT-F participants prior to intervention to ensure that they were familiar with the movements and procedures. This video was played on a screen during the training intervention to ensure that the participants kept up with the rhythm of each movement.

To ensure that the interventions were performed at adequate exercise intensity, participants' HRs were recorded throughout the session with an activity wristband. The peak heart rate (HR peak) of each session was considered to be 75% or more of the HRmax that had been recorded during Cooper's 12 min running test. All of the sessions began with a standardized 10 min low-to-moderate running and stretching followed by maximal shuttle run or functional training and ended with a 5 min cool-down and stretching.

#### *2.7. Statistical Analyses*

Statistical analyses were performed using SPSS, version 23.0 (Chicago, IL, USA). Data were presented as means x± SD. A two-factor analysis of variance with repeated measures was used to analyze differences in body composition, muscle performance, and aerobic capacity, with intervention (pretraining and post training) as a within-group factor and group (HIIT-R and HIIT-F) as a between-group factor. A significant intervention x group interaction was used to identify training-induced changes in body composition, muscle performance and aerobic capacity. Data were subsequently checked by Tukey's post hoc test if a significant interaction was revealed. Furthermore, paired t-tests were used to estimate within-group effects, and independent t-tests were conducted to examine differences between groups. The significance level was established as *p* < 0.05.

#### **3. Results**

All of the participants completed all of the sessions over the twelve-week period. There were no significant between-group differences in the variables measured at baseline (Table 1).

• Body Composition

Body composition data are presented in Table 3. There was a significant decrease (17.4% ± 7.4% for HIIT-R and 12.6% ± 5.1% for HIIT-F, *p* < 0.05) in the percent body fat for both groups (Figure 1a), with no interaction effect between HIIT-R and HIIT-F (*p* > 0.05). Body mass index (BMI) (Figure 1b) and waist hip ratio (WHR) (Figure 1c) did not change in either intervention (*p* > 0.05). Lean muscle mass increased in both groups (1.8% ± 1.4% for HIIT-R and 1.2% ± 1.2% for HIIT-F, *p* < 0.05).

### • Resting Heart Rate and Blood Pressure

Resting HR (*p* < 0.05) was improved compared to baseline in both intervention groups, while no interaction effect was observed. Resting systolic BP and diastolic BP remained unchanged (*p* > 0.05) after training in both the HIIT-R and HIIT-F groups.

• Aerobic Capacity

VO2max data was calculated from the following Cooper's equation: VO2max (mL/kg/min) = (distance(m)-506)/45. VO2max data for all participants are presented in Table 3. A significant increase (*p* < 0.05) in the VO2max was demonstrated in both training groups compared to baseline measures, while no significant intervention x group interaction was revealed between HIIT-R and HIIT-F after intervention compared to baseline (Figure 1c). (*p* > 0.05). The extent of the change in VO2max was 17.1% ± 5.6% and 12.7% ± 6.7% in the HIIT-R and HIIT-F groups, respectively.

• Muscle Performance

A significant intervention x group interaction displayed significant changes in the HIIT-R and HIIT-F groups in terms of measures of abdominal and lower limb strength (Figure 1d). In the HIIT-F group, repetitions completed during the one-minute sit-up test increased (*p* < 0.05) by 16.5% ± 3.1% and the distance obtained in the stand broad jumping test improved (*p* < 0.05) by 5.1% ± 2.2%, whereas these variables were unaltered (*p* > 0.05) in the HIIT-R group. Flexed push-ups were unaltered in both the HIIT-R and HIIT-F groups (Table 4).


**Table 3.** Body composition and aerobic capacity data from HIIT-R and HIIT-F groups.

Note: ∆ (post-baseline)/baseline; ns, no significance; partial *η* <sup>2</sup> value for effect size.

**Figure 1.** Changes in (**a**) BMI, (**b**) body fat%, (**c**) WHR, (**d**) VO2max, and (**e**) muscle performance change. Note: \*\* significantly different from baseline at *p* < 0.01. **Table 3.** Body composition and aerobic capacity data from HIIT-R and HIIT-F groups. **Figure 1.** Changes in (**a**) BMI, (**b**) body fat%, (**c**) WHR, (**d**) VO2max, and (**e**) muscle performance change. Note: \*\* significantly different from baseline at *p* < 0.01.

1.4% *<sup>p</sup>* < 0.05 36.0 ± 2.1 36.4 ± 2.1 1.2% ±

2.1% ns 22.4 ± 2.2 21.9 ± 2.1 <sup>−</sup>1.9% ±

0.9% ns 0.8 ± 0.0 0.8 ± 0.0 <sup>−</sup>0.3% ±

7.4% *<sup>p</sup>* < 0.01 32.3 ± 3.6 28.3 ± 3.9 <sup>−</sup>12.6% ±

Weight (kg) 56.6 ± 6.7 55.8 ± 6.5 <sup>−</sup>1.3% ±

mass (kg) 36.5 ± 1.7 37.2 ± 1.8 1.8% ±

BMI (kg/m2) 21.9 ± 3.1 21.6 ± 3.1 <sup>−</sup>1.3% ±

WHR 0.8 ± 0.0 0.8 ± 0.0 <sup>−</sup>0.6% ±

Body fat (%) 31.6 ± 4.1 26.3 ± 4.8 <sup>−</sup>17.1% ±

Lean muscle

**HIIT-R Group (***n* **= 10) HIIT-F Group (***n* **= 10) Interaction Effect**

3.0% ns ns 0.020

1.2% *<sup>p</sup>* < 0.05 ns 0.056

3.0% ns ns 0.018

0.5% ns ns 0.032

5.1% *<sup>p</sup>* < 0.01 ns 0.118


**Table 4.** Muscle performance data from HIIT-R and HIIT-F groups.

Note: ∆ (post-baseline)/baseline; ns, no significance; partial *η* <sup>2</sup> value for effect size.

#### **4. Discussion**

The present study aimed to investigate the effects of running and functional highintensity training on body composition, aerobic capacity, and muscle fitness. The primary finding was that high-intensity functional training was as effective as high-intensity interval running for aerobic capacity and body composition promotion in healthy inactive females, and moreover, it induced a significant improvement in muscle fitness. The validity of this finding is supported by the fact that the mean heart rate of all of the participants reached 75% VO2max or above throughout the intervention. Increases in resting heart rate were also detected after training in both groups.

#### *4.1. Body Composition*

Our findings that HIIT-R and HIIT-F had positive effects on body composition promotion regarding the reduction of the body fat percentage were consistent with other researchers. A previous study [49] showed improved body mass, BMI, and percent body fat among obese females after a total of 108 min HIIT-R. Similarly, previous research [50] found that HIIT-R was effective in reducing BMI and body fat percentage in overweight adults. Additionally, for individuals with normal BMI, body composition improved by decreasing fat mass and increasing lean mass after a 6 -week HIIT-R intervention [51].

Not surprisingly, body composition benefits were also found in other studies investigating HIFT. Improved body fat percentage was reported after a 5-week, thrice weekly HIFT intervention [25], and further studies have also indicated a beneficial influence of HIFT on body composition [52].

However, current research has indicated that body fat percentage was significantly improved after an eight-week HIFT, while body mass was unaltered [31]. Likewise, after 16 weeks of HIFT, a significant decrease in body fat percentage was observed with no changes in the body mass [26]. Previous HIIT-R studies have provided similar results [53,54]. These results are consistent with our findings that although body fat percentage was improved, body mass and BMI were not affected by the intervention. The improved body fat percentage may be explained by the significant increase in lean muscle mass (*p* = 0.001for HIIT-R and *p* = 0.006 HIIT-F) without significant changes in the body mass (*p* = 0.064 for HIIT-R and *p* = 0.051 for HIIT-F). The non-significant change in BMI may be due to the following reasons: the insufficient exercise duration per session (2 min vs. 6–10 min); the uncontrolled dietary intake during the intervention; and the characteristics of the participants regarding body weight. This suggestion has been highlighted in a recent systematic review [55] that indicated that for normal weight populations, low-volume HIIT is inefficient for body composition improvement. Furthermore, several studies have indicated that HIIT-R and HIFT have a more significant effect on weight loss or body fat loss among obese individuals [41,48–50].

Finally, no significant interaction effect was revealed for any body composition variables. This suggests that HIIT-R and HIIT-F were equally effective in the modulation of body fat percentage.

#### *4.2. Aerobic Fitness*

VO2max was assessed in the present study to estimate the effects of HIIT-R and HIIT-F protocols on aerobic fitness. Running-based HIIT has been shown to increase aerobic capacity in numerous previous investigations. Several studies have reported significant increases in VO2max after HIIT [11,12,55]. Furthermore, a systematic review also showed that HIIT was beneficial for aerobic fitness improvements among healthy young people [56]. Nevertheless, there has been no consensus on the effect of HIFT on aerobic capacity. Some studies investigating HIFT have shown an improvement in VO2max [35,52,57]. On the contrary, recent research has only found aerobic capacity improvement in underweight and overweight boys, with no changes being found among normal weight people [39]. Similarly, no significant changes in VO2max were found after a 6-week HIFT protocol [58,59].

In our study, participants from both the HIIT-R and HIIT-F groups experienced improvements in VO2max (17.1% ± 5.6% and 12.7% ± 6.7%, respectively). In line with the magnitude of our results, an increase of 8% in the VO2max was found after a low-volume HIFT [28]. It should be noted that in the current study, the enhanced VO2max observed in the HIFT group was significantly higher than values recorded in previous studies. VO2max has been reported to improve by 5% after a HIFT with no aerobic exercise [60]. Another study showed a moderate improvement in the VO2max of 6.3% [31]. In our study, the greater response of VO2max to HIFT could be explained by the following reasons: firstly, improvements in VO2max were related to the testing modality [61]. Cooper's 12 min run test demonstrated a systematic bias in favor of higher-scoring individuals [62]; secondly, this study used a longer duration (12 weeks vs. 6–8 weeks) for the implementation of functional exercises. Short or low-volume training reported no improvements in aerobic capacity, which was shown to require continuous training [14,63]. However, other investigations reported that the extent of improvement was not clearly related to training duration but to training intensity [56,64]. Therefore, further studies are required to investigate the effectiveness of the duration (work bouts/total work duration) and intensity on traininginduced aerobic capacity improvement; finally, the magnitude of the improvement in VO2max can be attributed to the fatigue index, which was not measured in our study [11].

Although high-intensity running and functional training were both beneficial for aerobic capacity promotion, few studies have compared the effectiveness of these two exercise modalities in terms of aerobic capacity enhancement. In the current study, we controlled for the same intervention intensity and duration and found that surprisingly, there was no significant difference in terms of the changes in VO2max between the HIIT-R and HIIT-F groups. It is worth noting that running showed higher oxygen consumption for the same intensity compared to other modalities [65]. Our findings were partially in line with a previous study [66] that indicated no significant differences in VO2max promotion between high-intensity cycling and HIFT. The results from the present study illustrate that functional training is as effective as running for aerobic fitness improvement when performed at high intensity with the same volume and intensity.

#### *4.3. Muscle Performance*

Importantly, the repetition of sit-ups and the distance of the standing broad jump were significantly increased after HIIT-F, whereas both parameters remained unaltered in the HIIT-R group. Moreover, significant interaction effects were observed in terms of the effects on abdominal and lower limb strength and duration. Our finding is consistent with other HIFT studies. A significant increase in muscle performance after 6 weeks of HIFT was reported, whereas no increase was found in HIIT group using rowing as the exercise modality [27]. Significant improvements in lower body strength and power among patients and Army personnel were also evident [24,25]. Likewise, a study with female participants compared the effects of HIFT and endurance treadmill training on muscle fitness and demonstrated that sit-ups, chest presses, and push-ups improved by 64%, 207% and 135%, respectively, in the HIFT group after 4 weeks of intervention [24].

It is worth noting that the number of flexed push-ups that was completed in the repetition exercise was unchanged in both groups. The unchanged results are in contrast with findings from other investigations. Findings from recent studies revealed increased upper body strength and endurance after functional training executed at a high intensity [24,26–28]. It was possible that the observed unvaried parameters were the consequence of insufficient movements during our functional training, which lacked upper body adaptations [26]. Additionally, the assessment methods used in the present study could have also induced unaltered results. Although the flexed push-ups had been modified for females and even though the participants were familiarized with testing procedures, the participants in the present study had no or little experience and were not familiar with this movement. Furthermore, they had no knowledge of specific strategies that could be used to maximize their performance.

The effects of HIFT on muscle performance varies across exercise design and test methods. HIIT significantly increases the proportion of type I fibers [67], while muscle adaptions are specific to the exercise modality. A previous study revealed that compared to high-intensity interval running, strength training with functional movements resulted in type I muscle fibers increasing in size and a higher percentage of type IIA muscle fibers [68]. In the present study, functional exercise was more effective in strengthening muscle power than running when both were performed at relatively the same high intensity and for the same duration. However, further studies are required to investigate the training-induced individual changes in the type and size of muscle fibers between participants. Additionally, the functional exercise design should consider the fitness of the participants to reduce muscle soreness, and a previous study reported no injuries using this methodology [31].

A general limitation in the HIFT investigation was the different types of functional exercises that were included. The results might be dissimilar if HIFT was performed with other combinations of movements. Furthermore, the results of our study came from a small sample size and a non-exercising control group was not used. Finally, dietary intake was not controlled during the intervention, and the total calories consumed were not calculated. In addition, the fatigue index was not measured during the aerobic test.

#### **5. Conclusions**

Twelve weeks of high-intensity training based on running or functional exercises were both effective in reducing body fat percentage and improving aerobic capacity among healthy inactive females. Relative to running-based high-intensity training, HIFT shows an equally effective alternative with more exercise enjoyment and much stronger adherence regarding body composition and aerobic fitness promotion. Additionally, HIFT resulted in greater muscle performance increases than running-based high-intensity training, after which no gains were observed in terms of muscle fitness.

HIFT with self-selected intensity represents an alternative to high-intensity interval running for eliminating exercise barriers for physical exertion. Furthermore, HIFT can be performed anywhere at any time, which limits the barriers of lacking time/money. Finally, HIFT reveals strong exercise adherence and more enjoyment among females. It may be helpful for individuals to promote physical activity and the associated benefits of a prolonged healthy lifestyle.

**Author Contributions:** Conceptualization, Y.L., H.D.W. and J.S.B.; methodology, H.D.W. and J.S.B.; software, Y.L.; validation, Y.L., H.D.W. and J.S.B.; formal analysis, Y.L.; investigation, Y.L.; resources, Y.L. and Q.W.; data curation, Y.L.; writing—original draft preparation, Y.L.; writing—review and editing, H.D.W. and J.S.B.; visualization, Y.L.; supervision, H.D.W. and J.S.B.; project administration, Y.L.; funding acquisition, Q.W. All authors have read and agreed to the published version of the manuscript.

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

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki and was approved by the Institutional Review Board (or Ethics Committee) of Ningbo University (RAGH202103150366.8; 15 March 2021).

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

**Data Availability Statement:** The data presented in this study are available upon request from the corresponding author. The data are not publicly available due to student privacy.

**Acknowledgments:** I would like to show my deepest gratitude to my supervisory team, Wiltshire and Professor Baker; they are respectable, responsible and resourceful scholars. They have provided me with valuable guidance in every stage of my research. Additionally, I would like to thank the researchers from Faculty of Sport Science, Ningbo University, for helping me to complete the assessment. Finally, many thanks to the participants who continued to exercise for their health.

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

#### **References**


3

### *Article* **Associations between Physical Status and Training Load in Women Soccer Players**

**Lillian Gonçalves 1,\* , Filipe Manuel Clemente 2,3,\* , Joel Ignacio Barrera 4,5 , Hugo Sarmento 4,5 , Gibson Moreira Praça <sup>6</sup> , André Gustavo Pereira de Andrade <sup>6</sup> , António José Figueiredo 4,5 , Rui Silva <sup>2</sup> , Ana Filipa Silva 2,7 and José María Cancela Carral <sup>1</sup>**


**Abstract:** This study aimed to analyze the variations of fitness status, as well as test the relationships between accumulated training load and fitness changes in women soccer players. This study followed an observational analytic cohort design. Observations were conducted over 23 consecutive weeks (from the preseason to the midseason). Twenty-two women soccer players from the same first Portuguese league team (22.7 ± 5.21 years old) took part in the study. The fitness assessment included anthropometry, hip adductor and abductor strength, vertical jump, change of direction, linear speed, repeated sprint ability, and the Yo-Yo intermittent recovery test. The training load was monitored daily using session rating of perceived exertion (s-RPE). A one-way repeated ANOVA revealed no significant differences for any of the variables analyzed across the three moments of fitness assessments (*p* > 0.05). The *t*-test also revealed no differences in the training load across the moments of the season (*t* = 1.216; *p* = 0.235). No significant correlations were found between fitness levels and accumulated training load (range: *r* = 0.023 to −0.447; *p* > 0.05). This study revealed no differences in the fitness status during the analyzed season, and the fitness status had no significant relationship with accumulated training load.

**Keywords:** football; athletic performance; training load; sports training; physical fitness

#### **1. Introduction**

Soccer is a high-intensity intermittent sport that recruits different energetic systems based on the intermittence of the match [1,2]. Among other factors, soccer performance requires technical skills, tactical awareness, and physical fitness [1,3]. In women's soccer, players may cover 9–12 km in total in a single match, with 1.5–2.5 km covered during high-intensity runs [4–6]. Moreover, throughout a women's soccer match, the average heart rate can reach up to 167 beats per minute (bpm), and the maximum heart rate (HRmax) can reach up to 193 bpm [7]. Therefore, to be successful, women soccer players should possess well-developed aerobic and anaerobic capacities, as well as good neuromuscular properties [2].

Well-developed physical fitness can help ensure overall success to the same extent as other important factors such as technical and tactical skills [1,8]. Accordingly, seeking

**Citation:** Gonçalves, L.; Clemente, F.M.; Barrera, J.I.; Sarmento, H.; Praça, G.M.; Andrade, A.G.P.d.; Figueiredo, A.J.; Silva, R.; Silva, A.F.; Carral, J.M.C. Associations between Physical Status and Training Load in Women Soccer Players. *Int. J. Environ. Res. Public Health* **2021**, *18*, 10015. https:// doi.org/10.3390/ijerph181910015

Academic Editors: Cristina Cortis and Paul B. Tchounwou

Received: 20 August 2021 Accepted: 22 September 2021 Published: 23 September 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/).

an improvement in fitness status, it is necessary to understand the status of players, thus making the assessment a determinant factor for individualization of the training and controlling the development of the players [1,9]. Regarding the control of evolution, it is also expectable that some fitness variations may occur across a season, specifically considering the three main periods of training and competition: (i) preseason, (ii) earlyseason, and (iii) end-season [1,10,11].

For example, body fat is usually lower after the preseason training period than at the start of the preseason [10,12]. Additionally, significant changes occur in the biomarkers of physiological stress [1,13]. Considering the physical fitness of female soccer players, it was found that countermovement jump scores seem to improve during the season [1]. Furthermore, the linear speed at 15 m improves during the preseason before stabilizing until the end of the season, whereas the linear speed at 25 m starts to decrease at the end of the season [11]. Naturally, considering seasonal variations, most of the fitness changes occur during the preseason because the training sessions during this phase are focused on establishing the players' fitness [3,14,15]. In contrast, during the season, more focus is placed on tactical and technical skills [16], with some efforts to stabilize players' fitness.

Even though no perfectly related variations were observed across the season, physical/physiological adaptations could be related to the training load and stimuli imposed on the players [17]. Therefore, a dose–response relationship is expected to arise between the training load and changes in fitness that may occur in soccer players [18]. However, such a relationship can vary on the basis of the training load measures and fitness parameters used; moreover, the relationship might not be as obvious or straightforward as expected [18,19]. As an example, in a study conducted on professional soccer players, relationships were found between accumulated perceived exertion and the speed achieved in the 30–15 Intermittent Fitness Test by professional players [20]. However, in another study (also on soccer players), such a relationship was not meaningful [21].

As mentioned above, the magnitude of the relationship between load and adaptations can vary as a function of the measures used. In the case of training load monitoring, one of the most commonly used measures is the rating of perceived exertion (RPE) [22,23]. This measure has been confirmed as valid and reliable, based on different scales (e.g., CR-10, CR-100), to estimate the intensity of a training session. According to the score provided by the player, RPE can be used to calculate the session RPE (s-RPE), which is the multiplication of the RPE score by the duration of the session (in minutes) [23,24]. Since this measure (s-RPE) has been highly correlated with internal load markers (e.g., heart rate measures) and external load markers (e.g., total distance, player load) [25,26], it seems to be a good measure to test relationships with fitness adaptations across the season.

Adaptions in soccer players take time and are influenced by multiple factors, such as age, gender, training history, psychological factors, and the duration, intensity, and frequency of training [18,27]. Therefore, it is difficult to understand which factors promote changes in players in women's soccer. In the particular case of women's soccer, dose– response relationships have not been explored' for that reason, there is a need to test whether such a relationship exists.

Testing the possibility of relationships between accumulated training load and the changes in fitness status would help to identify whether training load is a determinant of these changes or if there are other factors that coaches should be aware of. For that reason, the aims of this study were to analyze variations in the fitness status of women soccer players over time (repeated measures) and test the relationships between accumulated training load and fitness variations.

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

#### *2.1. Experimental Approach*

This study followed an observational analytic cohort design. Observations were made across 23 consecutive weeks (from the preseason to midseason). Fitness assessments of the players were performed three times: (i) at the beginning of the preseason, (ii) at the end of **2. Materials and Methods**  *2.1. Experimental Approach* 

**Figure 1.** Timeline of the study.

*2.2. Participants* 

of humans.

the preseason, and (iii) during the middle of the season. Internal loads were collected daily in all training sessions between August and January (Figure 1). end of the preseason, and (iii) during the middle of the season. Internal loads were collected daily in all training sessions between August and January (Figure 1).

made across 23 consecutive weeks (from the preseason to midseason). Fitness assessments of the players were performed three times: (i) at the beginning of the preseason, (ii) at the

**Figure 1.** Timeline of the study.

#### *2.2. Participants*

*Int. J. Environ. Res. Public Health* **2021**, *181*, 15 3 of 10

The cohort included 22 female soccer players (age: 22.7 ± 5.21 years; height: 162 ± 6.84 cm; weight: 57.6 ± 4.9 kg) competing in the first Portuguese League. The team had four weekly training sessions and one official match per week. The eligibility criteria for being considered in the analysis were as follows: (i) participation in at least 85% of the training sessions during the study, (ii) participants were present in all three assessments, (iii) absence of injuries or illness in the last four consecutive weeks, and (iv) players had at least 2 years of experience. Three players were excluded because they did not participate in all physical assessments. Before the assessments, all players were informed about the study procedures and signed an informed consent. The study was approved by the local The cohort included 22 female soccer players (age: 22.7 ± 5.21 years; height: 162 ± 6.84 cm; weight: 57.6 ± 4.9 kg) competing in the first Portuguese League. The team had four weekly training sessions and one official match per week. The eligibility criteria for being considered in the analysis were as follows: (i) participation in at least 85% of the training sessions during the study, (ii) participants were present in all three assessments, (iii) absence of injuries or illness in the last four consecutive weeks, and (iv) players had at least 2 years of experience. Three players were excluded because they did not participate in all physical assessments. Before the assessments, all players were informed about the study procedures and signed an informed consent. The study was approved by the local university and followed the ethical standards of the Declaration of Helsinki for the study of humans.

#### university and followed the ethical standards of the Declaration of Helsinki for the study *2.3. Fitness Assessment*

*2.3. Fitness Assessment*  Fitness assessments were conducted between August and January. All tests were performed during the same day of the week, following the same order, and at the same time of the day (7:30 p.m.) to limit data bias. During the three periods of assessments, all tests were distributed across three sessions, interspersed by 24 h of recovery. We acknowledge the fact that a testing battery can be carried out in a single day [28]. However, it can ideally be distributed over 2–3 days [29]. Regardless of the days, it is important that the sequence is designed with the aim of ensuring the most adequate conditions of absence of fatigue in tests with a greater need for neuromuscular recruitment, leaving the tests with greater metabolic stress to the end [29]. Bioenergetic and neuromuscular considerations resulted in the applied test sequencing in the present study. Regarding the warm-up protocol, it was out of the scope of the authors to intervene as it was always the team staff (physical trainer) conducting the warm-ups. The warm-Fitness assessments were conducted between August and January. All tests were performed during the same day of the week, following the same order, and at the same time of the day (7:30 p.m.) to limit data bias. During the three periods of assessments, all tests were distributed across three sessions, interspersed by 24 h of recovery. We acknowledge the fact that a testing battery can be carried out in a single day [28]. However, it can ideally be distributed over 2–3 days [29]. Regardless of the days, it is important that the sequence is designed with the aim of ensuring the most adequate conditions of absence of fatigue in tests with a greater need for neuromuscular recruitment, leaving the tests with greater metabolic stress to the end [29]. Bioenergetic and neuromuscular considerations resulted in the applied test sequencing in the present study. Regarding the warm-up protocol, it was out of the scope of the authors to intervene as it was always the team staff (physical trainer) conducting the warm-ups. The warm-ups consisted of low and self-paced running, followed by calisthenic exercises in which players performed two sets of 10 repetitions of walking lunges, single-leg deadlifts, and fontal and lateral high knee movements. These warm-ups were based on proposed strategies, highlighting the post-activation potentiation (PAP) exercises, as previously recommended and used [30,31].

ups consisted of low and self-paced running, followed by calisthenic exercises in which players performed two sets of 10 repetitions of walking lunges, single-leg deadlifts, and fontal and lateral high knee movements. These warm-ups were based on proposed strategies, highlighting the post-activation potentiation (PAP) exercises, as previously recommended and used [30,31]. The first assessments comprised anthropometry and hip adductor and abductor The first assessments comprised anthropometry and hip adductor and abductor strength tests. The second assessments comprised lower-body power, change-of-direction (COD), and linear speed tests. The third assessments comprised repeated sprint ability (RSA) and Yo-Yo intermittent recovery (YYIR) tests. All indoor tests were performed in a room with a stable temperature of 23 ◦C and relative humidity of 55%. All field tests were conducted on a synthetic turf with a mean temperature of 19.5 ± 3.4 ◦C and a relative humidity of 63% ± 4%.

strength tests. The second assessments comprised lower-body power, change-of-direction (COD), and linear speed tests. The third assessments comprised repeated sprint ability (RSA) and Yo-Yo intermittent recovery (YYIR) tests. All indoor tests were performed in a room with a stable temperature of 23 °C and relative humidity of 55%. All field tests were A measuring tape (SECA 206, Hamburg, Germany) and a digital scale (SECA 874, Hamburg, Germany) were used to measure the participants' height and body weight, measured to the nearest 0.1 kg. During both assessments, all participants were in a vertical position and had no shoes and unnecessary accessories. To measure hip strength, the

squeeze test was conducted using a dynamometer (Smart Groin Trainer, Neuro excellence, Portugal), as in a previously recommended protocol [32]. For lower-body power performance, the squat jump (SJ) and countermovement jump (CMJ) with both hands on hips were assessed, using the Optojump system (Optojump, Microgate, Bolzano, Italia [33]. The jump height was used for analysis. The 20 m zig-zag test was conducted to measure the participants' COD performance, using photocell timing gates (Photocells, Brower Timing System, USA) with a protocol described elsewhere [34]. The best time in seconds was used for further analysis. A 30 m linear sprint test was executed using three pairs of photocell timing gates (Photocells, Brower Timing System, UT, USA). Three maximal trials were performed, and the best time was used for analysis. Furthermore, an RSA protocol was conducted using two pairs of photocell timing gates (Photocells, Brower Timing System, UT, USA). The running anaerobic sprint test (RAST) test was conducted. This test consisted of six 35 m linear sprints, interspersed by 10 s of recovery. The best time to complete the test, peak power, and fatigue index measures were used for analysis [35]. The minimum and maximum peak power and the fatigue index were determined using the following equations [36]:

$$\text{Power} = \frac{\text{Weight} \times \text{Distance}^2}{\text{Time}^3} \text{ and } \text{Fatingule} \\ \text{Index} = \frac{\text{Max}\_{\text{Power}} - \text{Min}\_{\text{Power}}}{\text{Sum of 6 spints (s)}}.$$

Lastly, the participants completed the YYIR test to measure the VO2max. All player had to run 20 m from cone A to cone B and return to cone A (total: 40 m). After every 40 m covered, a 10 s recovery period was ensured. The speed started at 10 km/h, following progressive increases in velocity throughout the test. The YYIR ended when the player achieved total exhaustion or did not reach one of the 20 m cones at the beep timing. The number of completed shuttles and the total distance covered were recorded [37]. Additionally, during the YYIR test, all players used individual Bluetooth HRsensors for heart rate monitor (Polar H10, Polar-Electro, Kempele, Finland, recorded in 5 s intervals) to quantify each athlete's heart rate maximum (HRmax).

#### *2.4. Training Load Monitoring*

For measuring the internal load, 10 to 30 min after each training session, all players were asked about how hard the training session was, scored from 1–10, were 1 corresponds to "very light activity" and 10 corresponds to "maximal exertion" [38]. These scores were based on the CR-10 Borg scale [23]. All players were previously familiarized with this daily practice. The collected scores were then multiplied by the total duration in minutes of each training session, to obtain the session RPE [23]. The session RPE for each training session was used as the final outcome for further analysis.

#### *2.5. Statistical Analysis*

Subjects' characteristics are presented as means and standard deviations of the variables. For the variables of fitness assessment, a one-way repeated-measure analysis of variance (ANOVA) was performed to clarify the differences among the three assessments. If there was a significant effect, we used the Bonferroni multiple comparison test to determine significant differences among the three conditions for each variable. Eta squared (η 2 ) values were used as an indicator of effect size. An η <sup>2</sup> value of 0.00–0.19 was considered trivial, 0.20–0.49 was small, 0.50–0.79 was moderate, and ≥0.80 was large [39]. The strength of the relationship between the variables of the fitness assessment and accumulated training load was determined using a Pearson product moment linear correlation coefficient (*r*). A paired *t*-test was used to compare the training load between the periods (preseason and midseason). Statistical analyses were conducted using the Statistical Package for the Social Sciences (SPSS version 22.0; Chicago, IL, USA), with a significance level of 0.05.

#### **3. Results**

The one-way repeated ANOVA revealed no significant differences for any of the variables analyzed at the three moments of fitness assessment (Table 1). As there were no significant changes in the three moments observed, we chose to use the mean as a representative measure of physical status. The *t*-test revealed no differences in the training load between the periods of the season (*t* = 1.216; *p* = 0.235).

**Table 1.** Descriptions, F-statistics, and *p*-values of the fitness variables analyzed at the three moments of fitness assessment.


M1, M2, and M3: three measurement moments; *p*: *p*-value of F-statistic; η 2 : eta squared values; HRmax: heart rate maximum; VO2max: maximum oxygen volume; V10: 10 m sprint; V30: 30 m sprint; COD20: 20 m zig-zag test; p.max: maximum power; p.min; minimum power; FI: fatigue index; SJ: squat jump; CMJ: countermovement jump; YYIR: Yo-Yo intermittent recovery test; Addu: adductors; Abdu: abductors. *Int. J. Environ. Res. Public Health* **2021**, *181*, 15 6 of 10

> The time-course of the training load accumulated in the different microcycles is shown in Figure 2.

Correlations between fitness variables and average training load can be observed in

The current study aimed to analyze the variations of fitness status in women soccer players over time (repeated measures) and test the relationships between accumulated training load and fitness variations. To the best of our knowledge, this is the first study to simultaneously analyze variations in fitness status and training load from the beginning of preseason to the end of midseason, in the context of women's soccer. Concerning the first aim, there were no differences in fitness status during the analyzed period, contrary

**Variable** *r p***-Value** *r p***-Value**  HRmax −0.126 0.585 −0.447 0.048 VO2max −0.042 0.850 −0.157 0.486 V10 −0.187 0.417 0.056 0.816 V30 −0.123 0.596 0.023 0.922 COD20 −0.091 0.695 −0.225 0.341 p.max 0.249 0.276 0.058 0.808 p.min 0.351 0.119 0.256 0.276 FI 0.080 0.731 −0.104 0.662 SJ 0.314 0.166 0.330 0.156 CMJ 0.351 0.119 0.441 0.052 YYIR −0.059 0.811 −0.261 0.295 HRmax: heart rate maximum; VO2max: maximum oxygen volume; V10: 10 m sprint; V30: 30 m sprint; COD20: 20 m zig-zag test; p.max: maximum power; p.min; minimum power; FI: fatigue index; SJ: squat jump; CMJ: countermovement jump; YYIR: Yo-Yo intermittent recovery test.

**Figure 2.** Average training load in the 23 weeks of observation. **Figure 2.** Average training load in the 23 weeks of observation.

**Table 2.** Correlations between mean values fitness and average training load.

Table 2.

**4. Discussion** 

Correlations between fitness variables and average training load can be observed in Table 2.


**Table 2.** Correlations between mean values fitness and average training load.

HRmax: heart rate maximum; VO2max: maximum oxygen volume; V10: 10 m sprint; V30: 30 m sprint; COD20: 20 m zig-zag test; p.max: maximum power; p.min; minimum power; FI: fatigue index; SJ: squat jump; CMJ: countermovement jump; YYIR: Yo-Yo intermittent recovery test.

#### **4. Discussion**

The current study aimed to analyze the variations of fitness status in women soccer players over time (repeated measures) and test the relationships between accumulated training load and fitness variations. To the best of our knowledge, this is the first study to simultaneously analyze variations in fitness status and training load from the beginning of preseason to the end of midseason, in the context of women's soccer. Concerning the first aim, there were no differences in fitness status during the analyzed period, contrary to our original hypothesis. Furthermore, no significant relationships were observed between the fitness status and the accumulated training load, which is also contrary to our hypothesis.

The literature has shown that athletes generally change their fitness status over the season, although this is not so straightforward. For example, a study showed that players' aerobic capacity was higher in the midseason than in the pre- and postseason, indicating that the participants tend to reach a peak performance in this variable in the middle of the competitive schedule before it decreases over the subsequent weeks [17]. However, another study revealed that these changes can be very different from season to season [40]. Furthermore, physical performance changes throughout a soccer season can be dependent on the fitness status observed at the beginning of the preseason period [41]. Furthermore, the abovementioned study revealed a lack of positive changes after a preseason period [41].

Similar results were observed regarding VO2max, 15 m sprint, and agility tests in another study [10]. On the other hand, [1] found no differences in performance in the countermovement jump with arm-swing and sprint performance over a season, similar to the current results. In addition, previous studies showed that training loads can vary between the different periods of a season [14,15], which suggests that variations in fitness status could be related to the training loads that the players experience.

However, in the current study, there were no significant differences in the training load when comparing the preseason and the midseason, which might justify the absence of differences in players' fitness status. In fact, there is a need to respect the training principles, such as progressive overload, individualization, and variation for ensuring training adaptations [42]. Indeed, training variation assumes an important role for avoiding a monotonous training cycle and allowing supercompensation to occur [43,44]. For those reasons, the lack of differences in training load during the period observed in the present study may be related to poor management of training loads, as well as poor micro- and mesocycle planning [41]. Therefore, a dose–response relationship between training load variations and fitness status changes in the season could be suggested, although additional studies are required to confirm such an assumption. Specifically, studies in which the

training load is consciously manipulated to generate different magnitudes of changes are welcome.

Concerning the associations between physical status and training load, no significant values were reported. The literature suggests that a higher accumulated match time during the season is linked to better speed and CMJ performance, while a higher total exposure time is related to decreased power performances [19]. Therefore, it appears that both training and competitions define soccer players' physical status. However, another study showed significant negative associations between sRPE and physical fitness changes after a 9 week training period [21]. In fact, corroborating the abovementioned statement regarding the progressive overload training principle need for adaptations to occur, a study conducted on 34 junior male soccer players revealed that elite players with higher perceived training loads throughout a training period presented greater improvements in aerobic performance, when compared with their nonelite counterparts [45]. These facts reinforce the need for a better load management, respecting the training principles and biologic individuality for ensuring positive fitness changes after a training intervention.

At this point, sRPE is postulated as a consistent method of measuring internal training load among sessions for an entire season for youth soccer players [46]. However, sRPE is weakly related to independent high-intensity external load measures [47], indicating that this measure might not capture the complexity of training loads in soccer. Therefore, we suggest that future studies investigate the correlations between fitness status and external training loads in soccer, which could shed light on this topic.

Studies in women's soccer are still scarce, which justifies one of the strengths of the current research. Moreover, we monitored the players over half a soccer season, which indicates the relevance of the current results to understanding the complexity of the relationship between training loads and fitness status in women's soccer. However, there is a need for future studies to analyze an entire season, for a greater perception of such relationships.

Nevertheless, caution is required when interpreting the data. First, all athletes in this study belonged to the same team, which reduces the study's external validity. Additionally, no measures of external training load were taken, which limits the comprehension of the phenomenon. Lastly, we were unable to collect measures of fitness status during the midseason period (just after this stage). For this reason, changes might have occurred that were not captured by our measurements. For those reasons, we recommend future studies to expand the current findings by investigating athletes from a larger sample, collecting external load measures (i.e., high-intensity data), and including more intermediate assessments of fitness status.

#### **5. Conclusions**

This study revealed no differences in the fitness status of women soccer players during the analyzed season. Moreover, fitness status had no significant relationship with accumulated training load. Further studies should be conducted to identify other possible relationships and eventually determine how specific elements of fitness status are associated with specific efforts exerted during training drills.

**Author Contributions:** L.G., F.M.C. and J.M.C.C. led the project, established the protocol, and wrote and revised the original manuscript; J.I.B. and H.S. collected the data, and wrote and revised the original manuscript; G.M.P., A.G.P.d.A., A.F.S., A.J.F. and R.S. wrote and revised the original 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 study was conducted according to the guidelines of the Declaration of Helsinki and approved by the Institutional Review Board (or Ethics Committee) of the Polytechnic Institute of Viana do Castelo School of Sport and Leisure (code: CTC-ESDL-CE001-2021).

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

**Acknowledgments:** This work was funded by Fundação para a Ciência e Tecnologia/Ministério da Ciência, Tecnologia e Ensino Superior through national funds and, when applicable, co-funded EU funds under the project UIDB/50008/2020 to Filipe Manuel Clemente. Hugo Sarmento gratefully acknowledges the support of a Spanish government subproject 'Integration ways between qualitative and quantitative data, multiple case development, and synthesis review as the main axis for an innovative future in physical activity and sports research' [PGC2018-098742-B-C31] (Ministerio de Economía y Competitividad, Programa Estatal de Generación de Conocimiento y Fortalecimiento Científico y Tecnológico del Sistema I + D + i), which is part of the coordinated project 'New approach of research in physical activity and sport from a mixed methods perspective' (NARPAS\_MM) [SPGC201800x098742CV0].

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

#### **References**

