*Article* **Quantifying Coordination between Agonist and Antagonist Elbow Muscles during Backhand Crosscourt Shots in Adult Female Squash Players**

**Abdel-Rahman Akl <sup>1</sup> , Amr Hassan <sup>2</sup> , Helal Elgizawy <sup>1</sup> and Markus Tilp 3,\***


**Abstract:** The purpose of this study was to quantify the coordination between agonist and antagonist elbow muscles during squash backhand crosscourt shots in adult female players. Ten right-handed, international-level, female squash players participated in the study. The electrical muscle activity of two right elbow agonist/antagonist muscles, the biceps brachii and triceps brachii, were recorded using a surface EMG system, and processed using the integrated EMG to calculate a co-activation index (CoI) for the preparation phase, the execution phase, and the follow-through phase. A significant effect of the phases on the CoI was observed. Co-activation was significantly different between the follow-through and the execution phase (45.93 ± 6.00% and 30.14 ± 4.11%, *p* < 0.001), and also between the preparation and the execution phase (44.74 ± 9.88% and 30.14 ± 4.11%, *p* < 0.01). No significant difference was found between the preparation and the follow-through phase (*p* = 0.953). In conclusion, the co-activation of the elbow muscles varies within the squash backhand crosscourt shots. The highest level of co-activation was observed in the preparation phase and the lowest level of co-activation was observed during the execution. The co-activation index could be a useful method for the interpretation of elbow muscle co-activity during a squash backhand crosscourt shot.

**Keywords:** racket sport; injury; elbow; electromyography; co-activation

#### **1. Introduction**

The popularity of squash is increasing and now it is one of the racket sports that is played in most countries in the world. Similarly, the number of squash studies is growing, together with the interest of scientists who have analyzed various aspects of the game [1–4].

Modern squash is a fast-performing sport including complex and multidirectional movement patterns with a high density and intermittent rhythm. Therefore, it is multifaceted in its motor skills and its physiological, kinetic, and cognitive requirements. Performance success depends to a large extent on the interaction and complementarity of these factors [3,5–7].

Previous studies analyzed the effect of upper extremity movement and racquet speed during skill performance in squash [8], examined the electromyographical activity during strokes [9], and performed three-dimensional kinematic analyses of the forehand [10] and backhand strokes [11].

In their study, Hong, Chang, and Chan [10] reported interesting differences in the types of skills used during a squash game, among them the fact that the backhand was played more frequently (63.1%) than the forehand (36.9%), which underlines the importance of the backhand stroke in squash. Others investigated the association between the rotating motion of the upper extremities and racket speed when playing squash., The rate of performance of the forehand and backhand stroke was similar in other studies [8,9,12].

**Citation:** Akl, A.-R.; Hassan, A.; Elgizawy, H.; Tilp, M. Quantifying Coordination between Agonist and Antagonist Elbow Muscles during Backhand Crosscourt Shots in Adult Female Squash Players. *Int. J. Environ. Res. Public Health* **2021**, *18*, 9825. https://doi.org/10.3390/ ijerph18189825

Academic Editors: Filipe Manuel Clemente and Ana Filipa Silva

Received: 23 August 2021 Accepted: 16 September 2021 Published: 17 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/).

The backhand in squash is different from other strokes because the player must control the racket to fully control the angle of the hit without losing control of the swing speed while controlling the angular displacements of the elbow, torso, and shoulder joints [12].

Seoung Eun, Seung Nam, and Murali [12] used a 3D motion analysis system to analyze the backhand stroke performance of both elite and novice players. The study aimed to compare the displacement and velocity of the trunk and racquet, and the angular displacements and the velocities of the elbow and shoulder joints. The significant differences observed between novice and expert players underline the importance of studying the muscular activity during the backhand stroke. Vukovic et al. [13] measured the trajectory and velocity of movement using a tracking system to determine whether there were significant differences between winners and losers. They analysed the used skills, the time patterns, and the position of the squash players during their performance. Subsequently, they compared the dynamic movements of players of different technical abilities and related them to the tactics adopted by different players in 24 competitive matches with elite male squash players [14]. McGarry [15] examined the space–time patterns of squash players as they move around the squash court in the context of a dynamical system using movements analysis of forty-eight squash rallies—twelve from each quarter-final match in a high-level knock-out competition.

Besides the analysis of performance, squash injuries have also been a focus for researchers. Finch and Eime [7] conducted a review on retrospective studies of squash injuries that analyzed records of hospitalized, injured, or emergency patients, and surveys from squash players. The studies included data from 2232 domestic league players and university teams from the USA, UK, New Zealand, Germany, and the Netherlands and concluded that better-controlled studies are needed, particularly to determine the risk of injuries associated with squash.

In recent years, there have been steady increases in the duration of the game, perhaps due to the improvement in the physical and technical abilities of squash players. In addition, it should be noted that in 2009 squash underwent changes in the rules (e.g., changing the scoring system in squash to 'Point-A-Rally' (PAR) to 11 points per game) according to which players have less time to perform shots, which increases the workload of the players [14]. Both developments may cause high loads leading to injury.

Therefore, an important research aim in sports medicine has been to understand the relationship between the intensity or volume of training and the type and grade of injuries mainly due to overuse, particularly following stroke training [16–18]. During the last few years, some researchers have focused on injuries in squash [18]. Horsley et al. [19] studied the diversity of injuries suffered by professional squash players in both training and competition through a survey of injury records between 2004–2015. However, this study only looked at injuries of the lower limbs because of the mechanical loading of the players during strokes in racket games. Habitual participation in racquet games over years often results in specific strength and flexibility imbalances [20]. Previous studies have reported that the ratio of upper-limb injuries is around 36% of all squash injuries, making the elbow the most commonly injured body region [3,5,7]. Despite this, little is known about injury mechanisms, exhaustion, recovery, and performance in training and competition [21–23].

Previous studies have indicated the relationship between muscular co-activation and injury. Hirokawa, et al. [24] reported that increased quadriceps–hamstring muscle co-activation at the knee may reduce the risk of anterior cruciate ligament (ACL) injury.

Lehman [25] reported that there is a relationship between muscle extensor endurance with decreases in injury risk. The aberrant flexor/extensor endurance ratios have also been correlated with a history of injury. From this perspective, adequate joint stability is related to the amount of muscle co-activation [26,27]. Elbow stability is not provided by one specific muscle but rather via the coordinated efforts of agonist and antagonists muscles. These muscles are active throughout the whole backhand crosscourt movement in squash. Due to the muscular demands of the backhand crosscourt shot and the prevalence of injury

in the elbow, training the agonist and antagonist musculature may improve performance and decrease injury risk.

Surprisingly, so far, no study has investigated the electrical activity of the muscles during the backhand stroke, which has been identified as the skill with the highest frequency and also the highest cause of injury rate [3,7].

Given the importance of the coordination between the muscles working on the elbow joint with regards to performance and injury prevention, especially among female athletes, this study aims to investigate the coordination between agonist and antagonist elbow muscles during the backhand crosscourt stroke in adult female squash players.

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

#### *2.1. Participants*

Female, right-handed elite squash players (*n* = 10) participated in the present study (age: 18.4 ± 0.8 years; body mass: 60.8 ± 1.8 kg; height: 165.2 ± 1.6 cm; training age: 9.1 ± 0.9 years). The subjects were officially ranked between 4 and 20 in the Egyptian squash federation and were currently competing in professional squash tournaments (national and international). Written informed consent of the players was obtained, and the study was approved by the institutional ethics committee of studies and research.

#### *2.2. Experiment Protocol*

After a 15 min warm-up including general, elbow-, and shoulder-specific mobility exercises, as well as stretching and familiarization with the protocol, participants performed Squash backhand crosscourt strokes. A total of three successful attempts were recorded for each player, with a one-minute rest between attempts. The Squash backhand crosscourt skill was broken into three phases: the preparation phase, the execution phase, and the follow-through phase.

#### *2.3. Data Recording*

The electrical muscle activity of two right elbow agonist/antagonist muscles, the biceps brachii (BB) and triceps brachii (TB) were recorded using surface EMG system (Myon m320RX; Myon, Switzerland). The skin over the muscles of the dominant arm was shaved and cleaned with alcohol and bipolar, circular 10 mm diameter silver chloride surface electrodes (SKINTACT FS-RG1/10, Leonhard Lang GmbH, Archenweg 56, 6020 Innsbruck, Austria) were secured on the selected muscles. Electrodes were attached over each muscle following the SENIAM guidelines maintaining a 2 cm center to center inter-electrode spacing [28]. The EMG signals were stored at a sampling frequency of 1000 Hz and digitized using a 16-bit analog to digital (A/D) converter. EMG data was processed using Visual 3D software (C-Motion, Germantown, MD, USA). Raw EMG data were band-pass filtered (20 Hz–450 Hz) applying a Butterworth filter. The signals were preprocessed using full wave rectifier and a linear envelope obtained using the root mean square (RMS) approach with a window size of 100 ms. Data were normalized to an isometric maximum voluntary contraction (MVC), which was recorded after each subject finished the experimental tasks. To obtain the MVC values, subjects performed three repetitions for 5 s, with 60 s rest in between while sitting in a stable chair with forearm resistance. Peak muscle activity over the three repetitions for each muscle was taken as the MVC value.

#### *2.4. Co-Activation Index*

Muscle co-activation was estimated by the calculation of a co-activation index (CoI) using the following equation adapted from Kellis et al. [29]

$$CoI = \frac{iEMG\_{anta}}{\left(iEMG\_{anta} + iEMG\_{agg}\right)} \times 100$$

where *iEMGanta* and *iEMGago,* respectively, refer to *iEMG* of antagonist and agonist muscle in different movement phases. The preparation phase was defined from the beginning of the movement to the end of the elbow flexion, the execution phase was defined from the beginning of the elbow extension until the shot, and the follow-through phase was defined from the instant of the shot until the end of the movement, see Figure 1. The phases were defined by video analysis using 3D simi motion capture, which was synchronized with EMG. cle in different movement phases. The preparation phase was defined from the beginning of the movement to the end of the elbow flexion, the execution phase was defined from the beginning of the elbow extension until the shot, and the follow-through phase was defined from the instant of the shot until the end of the movement, see Figure 1. The phases were defined by video analysis using 3D simi motion capture, which was synchronized with EMG.

(ாெீೌೌାாெீೌ)

where *iEMGanta* and *iEMGago,* respectively, refer to *iEMG* of antagonist and agonist mus-

Muscle co-activation was estimated by the calculation of a co-activation index (CoI)

× 100

*Int. J. Environ. Res. Public Health* **2021**, *18*, 9825 4 of 9

using the following equation adapted from Kellis et al. [29]

ೌீೌாெ = ܫܥ

*2.4. Co-Activation Index* 

**Figure 1.** Backhand crosscourt phases (Preparation phase, Execution phase, Follow-Through phase) of the Biceps brachii (BB) and Triceps brachii (TB). (**A**) EMG raw data, (**B**) EMG rectified data, and (**C**) EMG RMS. Means (solid lines) and standard deviation (shaded areas) of three stroke attempts from a representative subject. **Figure 1.** Backhand crosscourt phases (Preparation phase, Execution phase, Follow-Through phase) of the Biceps brachii (BB) and Triceps brachii (TB). (**A**) EMG raw data, (**B**) EMG rectified data, and (**C**) EMG RMS. Means (solid lines) and standard deviation (shaded areas) of three stroke attempts from a representative subject.

#### *2.5. Statistical Analysis*

Descriptive statistics were reported as means and standard deviations (mean ± SD). The normality of the data was analyzed using the Shapiro–Wilk test and all data were found to be suitable for parametric analysis. Repeated Measures Analysis of Variance (ANOVA) with Sidak post hoc tests were used to detect significant differences and compare the mean of each variable during the three phases (preparation, execution, and follow-through). Partial eta squared (η <sup>2</sup>p) was calculated to assess the effect size. The statistical analysis was performed using IBM SPSS software Statistics v21 (IBM® Corporation, Armonk, NY, USA).

#### **3. Results** *3.1. Muscular Activity*

**3. Results** 

#### *3.1. Muscular Activity* Average values and standard deviations for the normalized RMS for the BB are pre-

*2.5. Statistical Analysis* 

Average values and standard deviations for the normalized RMS for the BB are presented in Figure 2, and the TB in Figure 3, during the three analyzed phases (the preparation, the execution, and the follow-through phase). The highest activities of the BB were observed during the follow-through phase, followed by the execution phase, and the preparation phase, with values of 13.80 ± 2.97%, 11.57 ± 1.45%, and 8.32 ± 3.47%, respectively. There was a significant difference between the BB activity during the preparation compared to the follow-through phases (*p* < 0.05; η <sup>2</sup>p = 0.50, Figure 2). For TB, the highest activities were observed during the execution phase, followed by the follow-through phase and the preparation phase, with values of 27.02 ± 3.43%, 16.14 ± 2.32%, and 6.38 ± 1.86%, respectively, while high significant differences for the TB were observed among the three phases (*p* < 0.001; η <sup>2</sup>p = 0.97, Figure 3). sented in Figure 2, and the TB in Figure 3, during the three analyzed phases (the preparation, the execution, and the follow-through phase). The highest activities of the BB were observed during the follow-through phase, followed by the execution phase, and the preparation phase, with values of 13.80 ± 2.97%, 11.57 ± 1.45%, and 8.32 ± 3.47%, respectively. There was a significant difference between the BB activity during the preparation compared to the follow-through phases (*p* < 0.05; η2p = 0.50, Figure 2). For TB, the highest activities were observed during the execution phase, followed by the follow-through phase and the preparation phase, with values of 27.02 ± 3.43%, 16.14 ± 2.32%, and 6.38 ± 1.86%, respectively, while high significant differences for the TB were observed among the three phases (*p* < 0.001; η2p = 0.97, Figure 3).

Descriptive statistics were reported as means and standard deviations (mean ± SD). The normality of the data was analyzed using the Shapiro–Wilk test and all data were found to be suitable for parametric analysis. Repeated Measures Analysis of Variance (ANOVA) with Sidak post hoc tests were used to detect significant differences and compare the mean of each variable during the three phases (preparation, execution, and follow-through). Partial eta squared (η2p) was calculated to assess the effect size. The statis-

*Int. J. Environ. Res. Public Health* **2021**, *18*, 9825 5 of 9

**Figure 2.** Average values and standard deviations for the normalized EMG (%MVC) per phase of the biceps brachii (η2p = 0.50). **Figure 2.** Average values and standard deviations for the normalized EMG (%MVC) per phase of the biceps brachii (η <sup>2</sup>p = 0.50). *Int. J. Environ. Res. Public Health* **2021**, *18*, 9825 6 of 9

**Figure 3.** Average values and standard deviations for the normalized EMG (%MVC) per phase of the Triceps brachii (η2p = 0.97). **Figure 3.** Average values and standard deviations for the normalized EMG (%MVC) per phase of the Triceps brachii (η <sup>2</sup>p = 0.97).

#### *3.2. Co-Activation Index 3.2. Co-Activation Index*

tively.

**4. Discussion** 

and follow-through phases (*p* = 0.95).

A significant effect (*p* < 0.01, η2p = 0.73) of the phases on the CoI was observed (Figure 4). Post hoc analyses showed that the co-activation was significantly higher in the follow-A significant effect (*p* < 0.01, η <sup>2</sup>p = 0.73) of the phases on the CoI was observed (Figure 4). Post hoc analyses showed that the co-activation was significantly higher in the

0.001), and also between the preparation phase and the execution phase (44.74 ± 9.88% and 30.14 ± 4.11%, *p* < 0.01). No significant difference was found between the preparation

**Figure 4.** Average values and standard deviations for the co-activation index (CoI) (%) per phase (η2p = 0.73). The agonist/antagonist muscle of each phase was BB/TB, TB/BB, and TB/BB, respec-

The main aim of this study was to determine muscle co-activation of elbow muscles as an indicator of coordination between agonist and antagonist muscle activity during three phases of the squash backhand crosscourt shots in adult female players. While we observed similar co-activation in the preparation and follow-through phase, the co-acti-

vation was significantly decreased during the execution phase.

follow-through phase compared to the execution phase (45.93 ± 6.00% and 30.14 ± 4.11%, *p* < 0.001), and also between the preparation phase and the execution phase (44.74 ± 9.88% and 30.14 ± 4.11%, *p* < 0.01). No significant difference was found between the preparation and follow-through phases (*p* = 0.95). through phase compared to the execution phase (45.93 ± 6.00% and 30.14 ± 4.11%, *p* < 0.001), and also between the preparation phase and the execution phase (44.74 ± 9.88% and 30.14 ± 4.11%, *p* < 0.01). No significant difference was found between the preparation and follow-through phases (*p* = 0.95).

A significant effect (*p* < 0.01, η2p = 0.73) of the phases on the CoI was observed (Figure 4). Post hoc analyses showed that the co-activation was significantly higher in the follow-

**Figure 3.** Average values and standard deviations for the normalized EMG (%MVC) per phase of

*Int. J. Environ. Res. Public Health* **2021**, *18*, 9825 6 of 9

**Figure 4.** Average values and standard deviations for the co-activation index (CoI) (%) per phase (η2p = 0.73). The agonist/antagonist muscle of each phase was BB/TB, TB/BB, and TB/BB, respectively. **Figure 4.** Average values and standard deviations for the co-activation index (CoI) (%) per phase (η <sup>2</sup>p = 0.73). The agonist/antagonist muscle of each phase was BB/TB, TB/BB, and TB/BB, respectively.

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

the Triceps brachii (η2p = 0.97).

The main aim of this study was to determine muscle co-activation of elbow muscles as an indicator of coordination between agonist and antagonist muscle activity during three phases of the squash backhand crosscourt shots in adult female players. While we observed similar co-activation in the preparation and follow-through phase, the co-acti-The main aim of this study was to determine muscle co-activation of elbow muscles as an indicator of coordination between agonist and antagonist muscle activity during three phases of the squash backhand crosscourt shots in adult female players. While we observed similar co-activation in the preparation and follow-through phase, the co-activation was significantly decreased during the execution phase.

vation was significantly decreased during the execution phase. The main activities for the BB and the TB were observed during three phases in which they acted as a prime mover (agonist), BB during the preparation phase and TB during the execution and follow-through phases.

Low values of muscle activity were observed in the preparation phase where both BB and TB showed activations less than 10% of MVC (BB: 8.3% MVC and TB 6.4% MVC). This can be explained by the fact that the elbow flexion includes muscle synergies (e.g., brachioradialis and anterior deltoid muscle) [30,31]. However, the relative muscle activity of BB was greater than that of TB muscle activity because the BB is the prime mover during elbow flexion.

Greater TB activation was observed during the execution phase with 27.02% MVC (BB: 11.57% MVC). Despite substantial activation differences between TB and BB during the execution phase, the observed BB activation was still greater when compared with the preparation phase. This is somewhat surprising, since the BB is an agonist muscle in the preparation phase but an antagonist muscle in the execution phase. The reason for this result may be the variation in movement muscle activity amplitude. While the agonist muscle activity increased during the execution phase to accelerate the movement, the antagonist initiated an increase in muscle activity, possibly to prevent an elbow joint injury due to overextension of the elbow joint in the follow-through phase [32].

Previous studies [33–38] indicate that experienced athletes could have a distinct muscle activation pattern with less antagonist muscle activation, implying that antagonistic muscle coupling might be altered by specialized activity. As a result, top athletes may have lower muscle co-activation than non-athletes, particularly during fast movements [39].

According to the observed results, muscle co-activation was greater in the preparation and the follow-through phases compared to the execution phase. Both Bazzucchi et al. [40] and Rouard and Clarys [31] reported greater co-activation values of the arm muscles during fast compared to slow movements, increasing at the preparation phase, decreasing at execution to allow faster acceleration, and increasing at the end of the movement to provide dynamic braking, which is similar to the elbow extension in the backhand crosscourt shot. In addition, Darainy and Ostry [41] and Bazzucchi, Sbriccoli, Marzattinocci, and Felici [40] underlined that greater antagonist activity could make the motor task controllable and also increases the stiffness and stability of the joint.

Wagner et al. [42] reported that overarm movements are essential skills in different types of sports. Hence, strong elbow extension might be considered as one of the determinants of efficiency in the squash backhand crosscourt shot. Bazzucchi, Riccio, and Felici [39] observed that muscle co-activation decreased during the execution phase for generating higher forces to increase performance. This is in accordance with our findings where the TB as the agonist muscle showed a strong activity during the execution phase, with low values of the BB as the antagonist muscle. This led to a low value of co-activation during the execution phase with values around 30%.

Muscle co-activation increased again in the follow-through phase to inhibit end range elbow extensions. The high value of co-activation in the follow-through phase with values of 45.93% increased joint stiffness and, therefore, stability [40]. This result was expected as increased co-activation is considered a determinant for preventing injury of the elbow joint.

Thus, an initial decreased co-activation allows a faster acceleration in the execution phase, while an increase at the end of the movement range provides dynamic braking of the movement. Furthermore, the high co-activation at the end of the movement allows players to better prepare the arm for the next response during the game [30,40].

Whereas most previous research has concentrated on antagonist muscle co-activation during maximal isometric efforts or as a function of isokinetic velocity, our research focused on sports movements including both concentric and eccentric contractions [43].

#### **5. Conclusions**

In summary, the co-activation of the elbow muscles varies within the squash backhand crosscourt stroke. The highest level of co-activation was observed in the preparation phase to control the forearm velocity before the execution phase, and in the follow-through phase to stabilize the elbow joint to prevent injuries and slow down the arm at the end of the movement. The lowest level of co-activation was observed in the execution phase for generating the appropriate force from the prime mover muscle to increase the efficiency of the backhand crosscourt shot.

**Author Contributions:** Conceptualization, A.-R.A. and H.E.; methodology, A.-R.A. and H.E.; software, A.-R.A.; validation, A.-R.A., A.H. and M.T.; formal analysis, A.-R.A., A.H. and M.T.; investigation, A.-R.A.; resources, A.-R.A.; data curation, A.-R.A., A.H. and M.T.; writing—original draft preparation, A.-R.A., H.E., A.H. and M.T.; writing—review and editing, A.-R.A., A.H. and M.T.; visualization, A.-R.A., A.H. and M.T.; supervision, A.-R.A.; project administration, A.-R.A. 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 Alexandria University.

**Informed Consent Statement:** Informed consent was freely obtained, and the study was approved by the institutional ethics committee of studies and research.

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

**Acknowledgments:** Open Access Funding by the University of Graz.

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

#### **References**


### *Article* **The Impact of Physical Performance on Functional Movement Screen Scores and Asymmetries in Female University Physical Education Students**

**Dawid Ko´zlenia \* and Jarosław Domaradzki**

Department of Biostructure, Faculty of Physical Education & Sport, University School of Physical Education in Wroclaw, al. I.J. Paderewskiego 35, 51-612 Wroclaw, Poland; jaroslaw.domaradzki@awf.wroc.pl **\*** Correspondence: dawid.kozlenia@awf.wroc.pl

**Abstract:** Association between physical performance and movement quality remains ambiguous. However, both affect injury risk. Furthermore, existing research rarely regards women. Therefore, this study aimed to assess the impact of physical performance components on FMS scores and asymmetries among young women—University Physical Education Students. The study sample was 101 women, 21.72 <sup>±</sup> 1.57 years, body mass index 21.52 <sup>±</sup> 2.49 [kg/m<sup>2</sup> ]. The FMS test was conducted to assess the movement patterns quality. Physical performance tests were done to evaluate strength, power, flexibility. Flexibility has the strongest correlation with FMS overall (r = 0.25, *p* = 0.0130) and single tasks scores. A higher level of flexibility and strength of abdominal muscles are associated with fewer asymmetries (r = −0.31, *p* = 0.0018; r = −0.27, *p* = 0.0057, respectively). However, the main findings determine that flexibility has the strongest and statistically significant impact on FMS overall (ß = 0.25, *p* = 0.0106) and asymmetries (ß = −0.30, *p* = 0.0014). Additionally, a significant effect of abdominal muscles strength on FMS asymmetries were observed (ß = −0.29, *p* = 0.0027). Flexibility and abdominal muscles strength have the most decisive impact on movement patterns quality. These results suggest possibilities for shaping FMS scores in young women.

**Keywords:** movement quality; physical performance; strength; power; flexibility; women; physical activity

#### **1. Introduction**

The physical performance and movement patterns quality affect injury risk [1,2]. Physical performance is defined as a body function that an appropriate test can objectively measure. It is a multidimensional concept which involves musculoskeletal system function, cardiorespiratory and nervous system. Physical performance is expressed by a level of single components such as strength, flexibility, speed, or endurance. [3,4]. Movement patterns quality is mainly examined by the FMS test, which detects dysfunctional movement patterns [5]. The relationships between the quality of movement patterns and physical performance components have been investigated. However, these associations have not been clearly defined. The studies published so far have focused mainly on men and mixed groups. Therefore, it is needed to establish these associations among women.

The tool for assessing the quality of the movement patterns is the Functional Movement Screen (FMS), which allows for a comprehensive evaluation of the functional state of the movement apparatus and to identify dysfunctional movement and differences in paired tests to identify asymmetries [5]. Numerous studies have indicated associations of low FMS scores with more injuries among men and women as well [6–9] and the possibility of injury prediction based on FMS score [10]. Mokha et al. [9] and Chalmers et al. [11] also demonstrated the relationship between asymmetries in the FMS test and injuries, studying young athletes. However, these results should be treated with some caution. In a replication study by Chalmers et al. [12], no similar links were observed.

**Citation:** Ko´zlenia, D.; Domaradzki, J. The Impact of Physical Performance on Functional Movement Screen Scores and Asymmetries in Female University Physical Education Students. *Int. J. Environ. Res. Public Health* **2021**, *18*, 8872. https://doi.org/ 10.3390/ijerph18168872

Academic Editor: Corneel Vandelanotte

Received: 11 July 2021 Accepted: 20 August 2021 Published: 23 August 2021

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

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

The attempts to determine the relationship between the quality of movement patterns and physical performance components indicate their existence [13]. However, the direction and strength of these relationships are unclear. Especially possible associations remain ambiguous among women due to a low number of studies regarding females. Parchmann and McBridge [14], in a mixed group, and Lockie et al. [15] among female team sports athletes, did not show any links between the quality of the movement patterns, speed, and agility. These reports are opposed to Ko´zlenia et al. [16], who showed that better quality of the movement patterns is associated with better speed and agility tests outcomes among men. Support for these results can be found in the studies by Campa et al. [17]. Chang et al. [18] indicated the relationship between a trunk stability push-up with agility t-test result. Sannicandro et al. [19] showed a connection between the FMS score and the power of the lower limbs among professional footballers, showing that better quality of movement patterns was associated with greater power of the lower limbs. Chimera et al. [20] established strong relationships between flexibility and the trunk muscles' strength with movement patterns quality. Similar observations also provide studies by Silva et al. [21,22] which showed the strength of the trunk muscles as a factor determining the quality of the movement patterns. However, the studies mentioned above [14,16–22] regard men and mixed groups, not focusing only on women and possible sex differences in associations between FMS score and asymmetries with physical performance components. However, Kibler et al. [23] proved that women are characterized by greater flexibility compared to men, who, in turn, have greater strength than women. The above observations could translate into relationships between the results of physical performance tests and the movement patterns quality and cause the sex differences in the single motor tasks scores in the FMS test described by Schneiders et al. [24]. They showed that men performed better than women in the trunk stability push-up (TSPU) and rotary stability (RS). Miller and Susa [25] noted a similar observation.

In the light of this observation, there is a need to keep in mind that physical performance and movement quality have an influence on injury risk [1,2]. Therefore, their interconnectedness should be explored. However, published studies mostly regard men and mixed groups in the literature, not only on women. Thus, this study aims to assess the impact of physical performance components on FMS scores and FMS asymmetries among young women—University Physical Education Students. Specifically, it was also examined the simple association between physical performance tests and FMS scores. These observations let to described which physical performance components are crucial to improving the quality of movement patterns, what can positively affect physical fitness and reduce injury risk.

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

#### *2.1. Study Sample*

The study sample consisted of 101 young adult women whose average age was 21.72 ± 1.57 years. All subjects were volunteers recruited from students at the University School of Physical Education in Wroclaw, Faculty of Physical Education and Sport. The average body weight was 60.54 ± 9.05 [kg], body height 1.68 ± 0.07 [m], and body mass index 21.52 <sup>±</sup> 2.49 [kg/m<sup>2</sup> ]. Subjects averagely declared 5.04 ± 3.56 h per week of physical activity. The inclusion criteria were no injuries before six weeks of the start of the study and no other medical contradiction for physical activity. All participants were fully informed about the purpose, type, research methodology, and participation conditions. They could withdraw from the research at any time without giving any reason.

#### *2.2. Measurements*

We followed the methods described in the study by Ko´zlenia and Domaradzki [2]. Participants were informed to avoid any physical activity directly before the measurements and tests. The measurements were performed in the Biokinetics Research Laboratory of

the University School of Physical Education in Wroclaw (Quality Management System Certificate PN-EN ISO 9001: 2009; No. PW-48606-10E).

A SECA model 764 height and weight measuring device (SECA manufactured, Hamburg, Germany. Quality control number C-2070) was used to measure body height and weight. Based on the obtained values, the index of relative body mass BMI (kg/m<sup>2</sup> ) was calculated according to the formula: BMI = body weight [kg]/body height [m<sup>2</sup> ].

The quality of the movement patterns was assessed using the Functional Movement Screen (FMS). The FMS test is a battery of seven movement tasks that make up the entire test: Deep Squat (DS), Hurdle Step (HS), (In-Line Lunge (IN-L), Shoulder Mobility (SM), Active Straight Leg Raise (ASLR), Trunk Stability Push-up (TSPU), Rotary Stability (RS). The tests were performed with a standard FMS kit (Functional Movement Systems, Inc, Chatham, MA, USA). According to Cook et al. [5], no warm-up directly before the test was performed. Single motor tasks are rated on a scale of 0 to 3 according to clear guidelines described for each test [5]. The maximum score is 21 points. From 14 points and below, the risk of injury increases significantly [26].

Handgrip strength of the upper limbs was examined using a hydraulic dynamometer with an adjustable handle SAEHAN SH5001 (Saehan Corporation, Masan, South Korea). The measurements were done with an accuracy of 1 kg. The subject keeps his arm lowered so that the upper limb does not touch the body. Holding the dynamometer tightly their hand, hand clenching was performed with maximum force for about 2 s. Two attempts were made for each limb. The best result on both limbs was considered.

A long jump test was used to assess lower limbs power. From two made attempts, the better result was considered. Jump length was measured from the back of the heels. The measurement was performed with an accuracy of 0.5 cm. The subject performed the test from the established line, made a jump from both lower limbs with a swing of the upper limbs landing on both legs.

The strength of the abdominal muscles was tested with the sit-ups test. The test consists of making as many sit-ups as possible within 30 s. One attempt was made. The subject was laid down with the lower limbs bent at the knee joints at an angle of 90◦ . The feet were blocked. The subject began the test by lying down with her hands clasped behind the nape of her neck, performing torso bends and touching knees with elbows.

Flexibility was measured using the sit-and-reach test. The measurement was performed with an accuracy of 0.5 cm. The examined person sat down with the lower limbs straightened in the knee joints by placing feet against the sidewall of the table. While maintaining the extension in the knee joints, the subject bent forward and tried to move the ruler on the table as far as possible along the scale. The tests scores were measured from the 0 cm point. The measurement was performed with an accuracy of 0.5 cm. Of the two trials, the better result was considered.

#### *2.3. Statistics*

The means, standard deviations, and confidence intervals were calculated for the data meeting the assumptions of normality of the distribution or the median, and standard errors for the data that did not meet the assumptions of the normal distribution. Spearman's rank correlation was calculated to investigate the strength and direction of relationships between the quality of the movement patterns (FMS scores) and physical performance tests. Multiple regression was used to determine the impact of the physical performance components on FMS overall and asymmetries. The significance level α = 0.05. Statistica v13.3 from Statsoft Polska (Cracow, Poland) was used for statistical analyses.

#### **3. Results**

Table 1 includes descriptive statistics for physical performance tests results.


**Table 1.** Physical performance tests results.

Table 2 shows the FMS overall score, single tasks score, and asymmetries numbers in the bilateral test. The mean FMS overall score is 14.96 ± 2.21, and a median of 15 points indicates the study sample has high-quality movement patterns.

**Table 2.** Characteristics of the FMS scores.


Abbreviations: FMS—overall score; DS—deep squat; HS—hurdle step; HS A—hurdle step-asymmetry; IN-L—inline lunge; IN-L A—in-line lunge-asymmetry; SM—shoulder mobility; SM A—shoulder mobility-asymmetry; ASLR—active straight leg raise; ASLR—active straight leg raise-asymmetry; TSPU—trunk stability push-up; RS—rotary stability—overall; RS A—rotary stability-asymmetry.

Spearman's correlation for FMS overall, single tasks score, and asymmetries number revealed the higher sit-and-reach test result is associated with the better FMS overall r = 0.25, *p* = 0.2130 and lower number of FMS asymmetries r = −0.31, *p* = 0.0018, better hurdle step score (HS) r = 0.21, *p* = 0.0357, shoulder mobility score (SM) r = 0.34, *p* = 0.0133, and asymmetries r = −0.23, *p* = 0.2010, active straight leg raise score r = 0.50, *p* > 0.0001 and asymmetries r = −0.27, *p* = 0.0055. Additionally, the higher sit-ups test results were associated with the lower number of FMS asymmetries r = −0.27, *p* = 0.0057, and inline lunge asymmetries r−0.25, *p* = 0.0126. No other statistically significant correlation was observed.

The multiple regression model for FMS overall is presented in Figure 1. The model is statistically significant, *p* < 0.0406.

Multiple regression results for FMS overall score are included in Table 3. Flexibility had the strongest statistically significant impact on FMS overall ß = 0.25, *p* = 0.0206. An increase in the sit-and-reach test score by 1 cm is associated with improving FMS overall score by 0.08 points.

**Figure 1.** Multiple regression model for FMS overall score. **Figure 1.** Multiple regression model for FMS overall score.


Multiple regression results for FMS overall score are included in Table 3. Flexibility **Table 3.** Multiple regression results for FMS overall score.

FMS overall Hand grip (N/kg) 0.11 0.10 0.04 0.04 1.08 0.2811 Long jump (cm) 0.09 0.10 0.01 0.01 0.89 0.3781 Sit-ups (reps/30 s) 0.15 0.10 0.07 0.05 1.49 0.1395 The multiple regression model for FMS asymmetries is presented in Figure 2. The model is statistically significant, *p* < 0.0005. *Int. J. Environ. Res. Public Health* **2021**, *18*, x 6 of 11

**Table 4.** Multiple regression results for FMS asymmetries number.

women due to the low number of studies among females.

**Independent Varia-**

Flexibility had the strongest statistically significant impact on FMS asymmetries ß = −0.30, *p* = 0.0014. Additionally, FMS asymmetries depend on abdominal muscles strength ß = −0.29, *p* = 0.0027. An increase in the sit-and-reach test score by 1 cm is associated with reducing asymmetries by 0.04. An increase in the sit-ups test by one rep reduces FMS

**bles ß ß SE B b SE t** *<sup>p</sup>*

Hand grip (N/kg) 0.04 0.09 0.01 0.01 0.42 0.6739 Long jump (cm) 0.01 0.09 0.00 0.00 0.06 0.9553 Sit-ups (reps/30 s) −0.29 0.09 −0.05 0.02 −3.08 0.0027 Sit and reach (cm) −0.30 0.09 −0.04 0.01 −3.29 0.0014

The quality of movement patterns and the level of physical performance affect the risk of injury, which raises further questions about their relationship [1,2]. There have been attempts in the literature to answer this type of ambiguity. However, some differences in observation do not allow clear conclusions on this matter, especially considering

Our results showed that flexibility and abdominal muscle strength have an influence on movement patterns quality among young women. The strength of abdominal muscles is crucial in trunk stability, whereas a good level of flexibility aids the optimal range of motion in joints [27,28]. Both mentioned physical performance components allow movement without restrictions in joints with simultaneous stabilization in various body positions, thus avoiding compensations that disturb the movement patterns. [29]. Our results

**Figure 2.** Multiple regression model for FMS asymmetries number. **Figure 2.** Multiple regression model for FMS asymmetries number.

asymmetries number by 0.05.

**Dependent Variable** 

FMS asymmetries

4. **Discussion** 

Multiple regression results for FMS asymmetries numbers are included in Table 4. Flexibility had the strongest statistically significant impact on FMS asymmetries ß = −0.30, *p* = 0.0014. Additionally, FMS asymmetries depend on abdominal muscles strength ß = −0.29, *p* = 0.0027. An increase in the sit-and-reach test score by 1 cm is associated with reducing asymmetries by 0.04. An increase in the sit-ups test by one rep reduces FMS asymmetries number by 0.05.


**Table 4.** Multiple regression results for FMS asymmetries number.

#### **4. Discussion**

The quality of movement patterns and the level of physical performance affect the risk of injury, which raises further questions about their relationship [1,2]. There have been attempts in the literature to answer this type of ambiguity. However, some differences in observation do not allow clear conclusions on this matter, especially considering women due to the low number of studies among females.

Our results showed that flexibility and abdominal muscle strength have an influence on movement patterns quality among young women. The strength of abdominal muscles is crucial in trunk stability, whereas a good level of flexibility aids the optimal range of motion in joints [27,28]. Both mentioned physical performance components allow movement without restrictions in joints with simultaneous stabilization in various body positions, thus avoiding compensations that disturb the movement patterns. [29]. Our results confirm this approach that is supported by the literature which provides related observations [20–22].

Most researchers agree that sex is not a factor that differentiates the overall score of the FMS test within one research group. Schneiders et al. [24] and Miller and Susa [25] did not show a statistically significant difference in the mean FMS score between physically active men and women. On the other hand, diversity can be seen in comparisons between groups from different research studies. The type of physical activity may explain the differences in FMS scores between groups from other studies. This can be seen when comparing the FMS scores results among various sports groups (e.g., footballers [30], volleyball players [31], or runners [8]). The quality of movement patterns can be shaped by appropriate training [32,33].

In the case of assessments of the FMS single motor tasks, sex differences are observed. It was shown that men performed better in a test that required stability and strength (TSPU and RS), whereas women achieved higher scores in a mobility test (SM and ASLR) [24]. A study by Miller and Susa [25] confirms this observation, which indicated that women achieved better results in shoulder mobility (SM) and active straight leg raises (ASLR), while men had higher scores in the trunk stability push-up (TSPU) and in-line lunge (IL-L). Anderson et al. [34] also indicated that women performed a weaker trunk stability push-up (TSPU). Similar observations were made in the study by Chimera et al. [20]. Women achieved better results in tasks requiring flexibility, active straight leg raise (ASLR), and shoulder mobility (SM), whereas men achieved better results than women in trunk stability push-up (TSPU) and rotary stability (RS). These results indicate that, despite averagely comparable results in FMS overall score between sexes, its structure could be differentiated by scores in single tasks. It could be associated with sex differences in physical performance [23]. Therefore, it suggests a need to consider sex differences in targeting the development of physical performance to improve movement quality.

The median value of the FMS overall score of our study sample was 15, indicating highquality movement patterns associated with low injury risk [26]. Literature shows that highquality movement patterns characterize young, physically active women. Chimera et al. [6] revealed that female athletes' average FMS overall score was above 14 points. A similar observation was added by Anderson et al. [34], which showed the mean FMS score among females was above 15. Our study sample performed better in the standing long jump than nonathletes females [35]. In the case of the sit-and-reach test, our study sample achieved comparable results to the female athletes from the study by Lopez-Minaro et al. [36]. In the handgrip test and 30 s sit-ups test, our study sample achieved better results than average populations [37,38].

Analyzing the relationship between the quality of movement patterns and physical performance components indicates a high, statically significant correlation between the sit-and-reach test score (flexibility), the FMS overall, and the FMS asymmetries number. The association in single tasks score was observed as an active straight leg raise (ASLR), shoulder mobility (SM), and asymmetries in these tests. A better result in the sit-and-reach test score was also associated with a better hurdle step (HS) score. Additionally, a higher level of flexibility related to a lower number of asymmetries was also observed in better results in the sit-ups test, suggesting that strong abdominal muscles are essential in maintaining functional symmetry. Other physical performance tests showed low and statistically insignificant correlations, suggesting no linear correlation between physical performance tests, such as handgrip or long jump, and the FMS scores. Similar observations were noted in other studies, especially regarding the relationship between flexibility and the quality of movement patterns. However, literature has limited observations considering the association between movement quality and FMS scores with physical performance components such as flexibility or strength. Therefore, it is hard to refer our results to only female groups. Some references must be made to mixed and male groups. Grygorowicz et al. [39] observed that female soccer players have better quality of movement patterns with a higher level of flexibility, which confirms our observations. Glass et al. [40] showed associations between the higher level of strength, balance, and flexibility with the quality of the movement patterns. In mixed groups, Jenkins et al. [41] observed that a better range of hip joint motion was associated with higher quality of movement patterns. Similarly, Yildiz et al. [42] indicated improved FMS scores and flexibility among children during tennis training after using a training intervention. Song et al. [43] noted that flexibility development was associated with improved FMS scores in males. In a study of male baseball players, Liang et al. [44] showed associations of FMS scores with the flexibility of the rectus femoris and speed abilities. Chimera et al. [20] showed that worse ranges of motion in the hip and knee joints negatively affect the movement patterns quality in a mixed study sample. The same authors [18] also observed that stronger trunk muscles positively influenced the FMS scores. Surprisingly, in our research, a higher score in trunk stability push-up (TSPU) was not directly associated with a better result of the sit-ups, whereas the study by Silva et al. [20,21] found links between the trunk stability push-up (TSPU) and the physical fitness of the study group. However, handgrip strength and the FMS scores do not seem to have clear connections [20]. Sannicandro et al. [19] indicated that footballers presented better quality movement patterns that generated greater lower limb power. Willigenburg and Hewett [45] observed correlations between the length of the long jump and the overall FMS score among American football players. In our study, a correlation between the long jump test score and the FMS overall and the deep squat (DS) score was not observed in women. The conclusions drawn from the above observations indicate a clear relationship between the quality of movement patterns and physical performance components. The single motor tasks of the FMS test are specifically related to the level of selected physical performance components. In this type of analysis, there is a need to consider the possible differences between sexes. However, our study shows that flexibility has the most decisive impact on the FMS overall score among female students.

Our analysis also showed that the number of asymmetries observed during the FMS test is related to flexibility and abdominal muscle strength, indicating these factors' significant role in shaping symmetrical movement patterns. However, the literature so far does not pay much attention to the number of FMS asymmetries concerning the level of physical fitness. Athletes with a higher number of asymmetries are 1.8 times more likely to be injured [46]. Similar observations are made by Mokha et al. [9], which also indicated a significant risk of injury due to asymmetries. Considering the relationship between flexibility and muscle strength and the risk of injury [47,48], the importance of these abilities as a key to developing high-quality movement patterns is growing.

#### **5. Conclusions**

In young, healthy women, flexibility and abdominal muscles strength are significantly associated with the quality of movement patterns, expressed as an FMS overall score and FMS asymmetries. Furthermore, flexibility is a component of physical performance with the most decisive impact on movement patterns quality in FMS overall score and asymmetries number, whereas abdominal muscle strength only influences on asymmetries in FMS. Our results indicate the importance of flexibility and abdominal muscle strength for movement patterns quality among young women. The appropriate range of motion in joints with abdominal muscle strength that provides trunk stability helps to avoid compensation in movement. It potentially suggests that FMS scores can be shaped throughout the development of flexibility and abdominal muscle strength. However, further studies need to verify if developing these abilities improves movement patterns quality in young women.

We are aware that our study has some limitations. The analysis could be augmented with more physical performance tests measuring other abilities, such as speed and endurance. It is also worth analyzing how the type of physical activity undertaken affects the relationship between the quality of movement patterns and physical performance. These aspects should be considered in further studies.

**Author Contributions:** Conceptualization, D.K.; methodology, D.K., J.D.; software, J.D.; validation, D.K., J.D.; formal analysis, J.D.; investigation, D.K.; resources, D.K., J.D.; data curation, J.D.; writing original draft preparation, D.K., J.D.; writing—review and editing, D.K., J.D.; visualization, D.K.; supervision, J.D.; project administration, D.K.; funding acquisition, D.K. 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 Senate Research Ethics Committee at the University School of Physical Education in Wrocław (ECUPE No. 16/2018; approved on 1 March 2018).

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

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

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

#### **References**


### *Article* **Impact of Rowing Training on Quality of Life and Physical Activity Levels in Female Breast Cancer Survivors**

**Juan Gavala-González 1,2 , Amanda Torres-Pérez 2,3,\* and José Carlos Fernández-García 2,3**


**Abstract:** The aim of this longitudinal study was to determine whether a rowing training program improved the quality of life and the physical activity levels in female breast cancer survivors (*n* = 28) (stage 1–4.54%; stage 2–36.36%; stage 3–54.54%; and stage 4–4.54%), diagnosed 4.68 ± 3.00 years previously, who had undergone a subsequent intervention (preservation 56.53% and total mastectomy 43.47%) and had a current mean age of 52.30 ± 3.78 years. The participants (*n* = 28) engaged in a 12-week training program, each week comprising three sessions and each session lasting 60– 90 min. The short form of the International Physical Activity Questionnaire (IPAQ-SF) and the Short Form 36 Health Survey (SF-36) were also administered. The results showed statistically significant improvements in levels of physical activity and in the dimensions of quality of life. We can conclude that a 12-week rowing training program tailored to women who have had breast cancer increases physical activity levels, leading to improved health status and quality of life.

**Keywords:** breast cancer; rowing; exercise; quality of life; perceived health; IPAQ-SF; SF-36

#### **1. Introduction**

Cancer is the second leading cause of death worldwide, representing about 9.6 million deaths in 2018, which means that one in six deaths globally is due to this disease [1,2]. In women, breast cancer is the most common cancer, affecting around 2.1 million women in 2018, i.e., one in four cancers diagnosed is breast cancer [2–4].

The rise in the number of cancer cases diagnosed in recent years has been associated with population growth, closely linked to increased life expectancy, and therefore with aging, considering age as a fundamental risk factor for developing cancer. It has also been related to the increase in early detection as well as improvements in primary care and in early diagnosis programs, which, although they lead to higher numbers of cases, are in turn related to a decrease in mortality [4–6].

One-third of diagnosed cancer cases could be prevented if exposures to various lifestyle-related risk factors were eliminated or reduced, such as smoking; consumption of harmful substances such as alcohol; an unhealthy, high-calorie diet with a high intake of saturated animal fats and sugars; and/or a sedentary lifestyle [1,4,5,7]. A large body of scientific evidence shows that physical activity has positive effects on the general population, improving health status, mood, body composition, quality of life [7–11], and preventing the onset of numerous diseases, including various types of cancer, such as breast cancer [12–14].

Physical activity has been associated with a lower risk of developing breast cancer [12,13,15], with a decrease in the probability of relapse and with a higher survival rate [13–15]. In individuals with cancer, physical activity has benefits for their health:

**Citation:** Gavala-González, J.; Torres-Pérez, A.; Fernández-García, J.C. Impact of Rowing Training on Quality of Life and Physical Activity Levels in Female Breast Cancer Survivors. *Int. J. Environ. Res. Public Health* **2021**, *18*, 7188. https:// doi.org/10.3390/ijerph18137188

Academic Editors: Filipe Manuel Batista Clemente and Ana Filipa Silva

Received: 7 May 2021 Accepted: 26 June 2021 Published: 5 July 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/).

reduced fatigue, improved strength levels, and improved quality of life and physical function [15–19].

However, despite the evidence supporting physical activity, two out of three cancer patients do not perform the minimum levels of exercise recommended by the American College of Sports Medicine (ACSM), which considers it essential to perform 150 min of moderate aerobic activity or 75 min of vigorous aerobic activity per week and at least 2 days of resistance training [13,17,20].

The relationship between physical activity and breast cancer has been demonstrated in several studies that analyzed and compared the effects of different exercise programs in breast cancer survivors, finding significant improvements in quality of life [21,22], physical function, and muscle strength [21,23]. More specifically, the study by Wiskemann et al. (2016) based on a 12-week resistance training program showed gains in muscle strength [24]. In several studies where the training programs combined endurance with aerobic exercise for 12 weeks, the improvements were significant in muscle strength, level of physical activity, and quality of life [21,25–29]. In addition, studies on mixed programs combining aerobic and strength exercises [30,31] showed important improvements in aerobic capacity, maximum oxygen consumption, muscle strength, reduction in the percentage of fat mass, and, above all, improved quality of life. Improvements were also found in physical, psychological, social, and quality of life parameters from training programs based on dragon boat rowing in women with breast cancer [18,32–34].

Women breast cancer survivors have found rowing to be an activity that improves the sequelae of the disease [35], such as reducing pain, increasing the range of movement in the upper limbs, improving muscle activation, and increasing strength and muscle function [36,37].

In this sense, rowing is considered one of the most complete water sports, involving the work of the musculature of both the upper and lower limbs [38] and almost all the body's musculature [39]. It is a sport in which symmetrical movements are performed that do not require forced position and that combine the work of strength and aerobic endurance [18]. Several studies have shown that this type of activity improves the quality of life of cancer patients, including psychological, physical, social, and emotional aspects, favoring their rehabilitation, self-esteem, and normalizing their daily life [37,40,41].

In addition, the practice of rowing has psychosocial benefits for cancer survivors [37,42] because it is a team sport that promotes the development of social relationships, and they find the support they need in other women who have gone through or are going through the same situation. Additionally, rowing is an outdoor activity, which provides them with extra motivation to adhere to physical activity and improves their quality of life [41,43,44].

Finally, in studies using the short form of the International Physical Activity Questionnaire (IPAQ-SF) to examine changes in physical activity [28,29] and the Short Form 36 Health Survey (SF-36) for changes in quality of life [27,30] following physical activity interventions in breast cancer patients, it was found that activity levels increased, and the different domains of quality of life improved after completion of the training programs.

The purpose of the present longitudinal study is to determine the influence of a 12-week rowing training program on quality of life and physical activity levels in women who have survived breast cancer.

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

#### *2.1. Design and Participants*

The study, according to Hernández, Fernández, and Baptista (2014), is a non-experimental longitudinal panel design, given that the same participants have been measured or observed at all times or points in time [45].

Participants (*n* = 28) aged 52.3 ± 3.8 years were recruited under the condition of having overcome breast cancer. The women were diagnosed 4.7 ± 3.0 years earlier, had different stages of disease, and had undergone surgery, as shown in Table 1.

Breast cancer survivor (BCS) is the name given to women who have been diagnosed with breast cancer and have had to undergo surgery and chemotherapy and/or radiotherapy treatments.

To take part in this research, we searched various breast cancer associations in Malaga for women who wanted to do sports (rowing) and who met the following characteristics: having overcome breast cancer, having completed chemotherapy and radiotherapy treatments, and having the oncologist's approval to do physical activity. No more than 10 years had passed since the cancer diagnosis. All of them were still taking tamoxifen.

The sample was selected based on compliance with these inclusion criteria.


**Table 1.** Characteristics of the breast cancer survivor sample.

The study was carried out at the RC Mediterráneo in Málaga involving women from the Sport Association Málaga D.B. Forty-eight participants were invited to take part: 10 initially withdrew due to compatibility/work/family and transport problems, and 10 were excluded because they did not attend 90% of the sessions.

After the initial selection, the nature of the study was explained to the participants, indicating that their anonymity would be maintained at all times, following the ethical considerations of the Sport and Exercise Science Research [46], as well as the principles included in the Declaration of Helsinki [47], which define the ethical guidelines for research in human subjects. The University of Malaga assigned the identification number 65-2020-H, which is registered with the Ethics Committee. The participants provided written informed consent, and throughout the intervention and afterwards, we acted under the provisions of the Organic Law 3/2018, of December 5, on the Protection of Personal Data and Guarantee of Digital Rights, regarding the protection of personal data under Spanish legislation. After signing the informed consent, the physical activity (IPAQ-SF) and the health-related quality of life (SF-36) questionnaires were administered.

The intervention lasted 12 uninterrupted weeks in which the women carried out two weekly sessions as described above. Both at the beginning and at the end of the program the participants were asked to complete the questionnaires.

#### *2.2. Instruments*

The participants also completed the short version of the International Physical Activity (IPAQ-SF) questionnaire to assess physical activity levels over the last 7 days. This questionnaire consists of seven questions that have acceptable measurement properties to monitor physical activity levels for adults aged 18 to 65 years in various settings, and it also reports the number of metabolic equivalents (METS) over the last 7 days [48]. Several studies have demonstrated the reliability of the IPAQ-SF for measuring the level of physical activity or the number of METS achieved during the last 7 days, obtaining similar results to other types of tests such as accelerometry or podometry [49–51].

The SF-36 Health Survey was used to assess health-related quality of life. This questionnaire consists of 36 items that report both positive and negative health status covering eight dimensions: physical function, social function, physical role, emotional role, mental health, vitality, bodily pain, and general health.

#### *2.3. Intervention*

Before starting the training program, participants were asked to complete the IPAQ-SF and SF-36 Health Survey questionnaires; in addition, they filled in the informed consent document to participate in the study.

The 12-week rowing training program was carried out at the RC Mediterráneo in Málaga, and it was divided into three parts of 4 weeks each. These stages progressively increased in intensity and were regulated through the participants' subjective perception of effort using the Börg scale [52].


Throughout the program, a weekly schedule was established consisting of three training days lasting 60–90 min per session. These sessions were supervised by a trainer who ensured attendance, correct execution of the tasks, and intensity of the sessions, in addition to excluding from the study those subjects who did not comply with at least 90% participation. All the exercises in these sessions were performed in a group. Exercises were generic and adjusted for people who have never rowed before. Each of the training sessions had the same structure:


Both at the beginning and at the end of the program sessions were held to discuss issues, and the participants were asked to complete the questionnaires.

#### *2.4. Data Analysis*

All analyses were performed with IBM SPSS, version 25 (IBM Corp, Armonk, NY, USA). The significance level was defined as *p* < 0.05. The fit of the different variables to the normal distribution was assessed using graphic procedures and the Shapiro–Wilk test.

To examine the differences resulting from the rowing training performed by the participants, the medians of each variable pre- and post-intervention were analyzed using the Wilcoxon test for related samples (paired data). In addition, graphic analysis of the different variables was carried out using box-and-whisker plots. In addition, the effect size for the Wilcoxon test (r) was calculated by the Z-score [54]. In the Cohen's guidelines for r, a large effect is defined as 0.5, a medium effect as 0.3, and a small effect as 0.1.

#### **3. Results**

Descriptive analyses of the different study variables (Table 2), level of physical activity, and quality of life are shown below, differentiating between pre-intervention and postintervention values.

*Int. J. Environ. Res. Public Health* **2021**, *18*, 7188


**Table 2.** Descriptive analysis of the variables pretest and posttest.

For the variables associated with engaging in physical activity obtained from the IPAQ-SF and the quality of life through the SF-36 questionnaire, the Wilcoxon test was carried out to determine whether significant differences exist between the pretest and posttest data.

Figure 1 depicts all physical activity variables, showing improvements after the intervention in levels of walking (Dif Mdn = 495 < 1287, z = −4.201, *p* = 0.000, r = 0.79), moderate (Dif Mdn = 0.00 < 720.00, z = −4.314, *p* = 0.000, r = 0.81), vigorous (Dif Mdn = 80.00 < 1440.00, z = −4.043, *p* = 0.000, r = 0.76), and total physical activity (Dif Mdn = 1075.50 < 3483.00, z = −4.286, *p* = 0.000, r = 0.81), all of which were significant. The lower value obtained for the IPAQ sitting variable (Dif Mdn = 240.00 > 150.00, z = −3.075, *p* = 0.002, r = 0.58) after the intervention compared to before the intervention indicates the participants were more active and spent less time sitting throughout the week. Regarding the effect size, the differences had a large effect and were statistically significant, as they presented values greater than 0.5. *Int. J. Environ. Res. Public Health* **2021**, *18*, x FOR PEER REVIEW 6 of 12

**Figure 1.** Physical activity levels (IPAQ-SF); PRE = pretest; POST = posttest; METS = metabolic equivalents. **Figure 1.** Physical activity levels (IPAQ-SF); PRE = pretest; POST = posttest; METS = metabolic equivalents.

Regarding the variables associated with quality of life, Figure 2 displays the results for the mental dimensions, showing improvements after the intervention in vitality (Dif Mdn = 50.00 < 60.00, z = −2.879, *p* = 0.004, r = 0.54), social function (Dif Mdn = 75.00 < 100.00, z = −3.247, *p* = 0.001, r = 0.61), emotional role (Dif Mdn = 75.00 < 100.00, z = −3.268, *p* = 0.001, r = 0.61), and mental health (Dif Mdn = 58.00 < 68.00, z = −2.836, *p* = 0.005, r = 0.54), all of which were significant. As for the effect size, its values were relevant as they were above 0.5. Regarding the variables associated with quality of life, Figure 2 displays the results for the mental dimensions, showing improvements after the intervention in vitality (Dif Mdn = 50.00 < 60.00, z = −2.879, *p* = 0.004, r = 0.54), social function (Dif Mdn = 75.00 < 100.00, z = −3.247, *p* = 0.001, r = 0.61), emotional role (Dif Mdn = 75.00 < 100.00, z = −3.268, *p* = 0.001, r = 0.61), and mental health (Dif Mdn = 58.00 < 68.00, z = −2.836, *p* = 0.005, r = 0.54), all of which were significant. As for the effect size, its values were relevant as they were above 0.5.

**Figure 2.** Quality of Life (SF-36)—Mental Dimensions; PRE = pretest; POST = posttest.

Focusing on the physical dimensions, Figure 3 shows that all variables improved significantly after the intervention, including general health (Dif Mdn = 61.00 < 72.00, z = −2.006, *p* = 0.045, r = 0.37). In the case of the variation in bodily pain (Dif Mdn = 51.50 < equivalents.

0.5.

**Figure 1.** Physical activity levels (IPAQ-SF); PRE = pretest; POST = posttest; METS = metabolic

Regarding the variables associated with quality of life, Figure 2 displays the results for the mental dimensions, showing improvements after the intervention in vitality (Dif Mdn = 50.00 < 60.00, z = −2.879, *p* = 0.004, r = 0.54), social function (Dif Mdn = 75.00 < 100.00, z = −3.247, *p* = 0.001, r = 0.61), emotional role (Dif Mdn = 75.00 < 100.00, z = −3.268, *p* = 0.001,

which were significant. As for the effect size, its values were relevant as they were above

**Figure 2.** Quality of Life (SF-36)—Mental Dimensions; PRE = pretest; POST = posttest. **Figure 2.** Quality of Life (SF-36)—Mental Dimensions; PRE = pretest; POST = posttest.

Focusing on the physical dimensions, Figure 3 shows that all variables improved significantly after the intervention, including general health (Dif Mdn = 61.00 < 72.00, z = −2.006, *p* = 0.045, r = 0.37). In the case of the variation in bodily pain (Dif Mdn = 51.50 < Focusing on the physical dimensions, Figure 3 shows that all variables improved significantly after the intervention, including general health (Dif Mdn = 61.00 < 72.00, z = −2.006, *p* = 0.045, r = 0.37). In the case of the variation in bodily pain (Dif Mdn = 51.50 < 61.00, z = −2.472, *p* = 0.013, r = 0.47), this may indicate that the participants reported less pain overall after the intervention. In other words, engaging in rowing decreases bodily pain in women who are breast cancer survivors or increases the pain threshold in the women studied. In addition, improvements after the intervention in physical role (Dif Mdn = 71.88 < 84.38, z = −3.866, *p* = 0.000, r = 0.73) and physical function (Dif Mdn = 82.50 < 90.00, z = −3.256, *p* = 0.001, r = 0.62) were shown, which indicates that the participants had fewer limitations when performing any physical activity compared to prior to the intervention. In terms of effect size, the data show that the differences were relevant, as they presented values between 0.3 and 0.5. This suggests that a 12-week rowing training program, adapted to women who have had breast cancer, helps to improve the perceived ability to perform other activities. *Int. J. Environ. Res. Public Health* **2021**, *18*, x FOR PEER REVIEW 7 of 12 61.00, z = −2.472, *p* = 0.013, r = 0.47), this may indicate that the participants reported less pain overall after the intervention. In other words, engaging in rowing decreases bodily pain in women who are breast cancer survivors or increases the pain threshold in the women studied. In addition, improvements after the intervention in physical role (Dif Mdn = 71.88 < 84.38, z = −3.866, *p* = 0.000, r = 0.73) and physical function (Dif Mdn = 82.50 < 90.00, z = −3.256, *p* = 0.001, r = 0.62) were shown, which indicates that the participants had fewer limitations when performing any physical activity compared to prior to the intervention. In terms of effect size, the data show that the differences were relevant, as they presented values between 0.3 and 0.5. This suggests that a 12-week rowing training program, adapted to women who have had breast cancer, helps to improve the perceived ability to perform other activities.

**Figure 3.** Quality of Life (SF-36)—Physical Dimensions; PRE = pretest; POST = posttest. PRE-TEST POST-TEST **Figure 3.** Quality of Life (SF-36)—Physical Dimensions; PRE = pretest; POST = posttest.

rowing program, a tendency towards an improved perception of health emerged: the number of women who reported having poor or fair health decreased, and those who claimed having "very good health" increased, with 75% of the participants reporting their

Perceived Health

Bad Regular Good Very good

health status to be good or very good.

Pts

Finally, in terms of overall perceived health of the participants, a significant improvement was detected after the physical activity intervention (Figure 4). After completing the rowing program, a tendency towards an improved perception of health emerged: the number of women who reported having poor or fair health decreased, and those who claimed having "very good health" increased, with 75% of the participants reporting their health status to be good or very good. Finally, in terms of overall perceived health of the participants, a significant improvement was detected after the physical activity intervention (Figure 4). After completing the rowing program, a tendency towards an improved perception of health emerged: the number of women who reported having poor or fair health decreased, and those who claimed having "very good health" increased, with 75% of the participants reporting their health status to be good or very good.

**Figure 3.** Quality of Life (SF-36)—Physical Dimensions; PRE = pretest; POST = posttest.

*Int. J. Environ. Res. Public Health* **2021**, *18*, x FOR PEER REVIEW 7 of 12

ability to perform other activities.

61.00, z = −2.472, *p* = 0.013, r = 0.47), this may indicate that the participants reported less pain overall after the intervention. In other words, engaging in rowing decreases bodily pain in women who are breast cancer survivors or increases the pain threshold in the women studied. In addition, improvements after the intervention in physical role (Dif Mdn = 71.88 < 84.38, z = −3.866, *p* = 0.000, r = 0.73) and physical function (Dif Mdn = 82.50 < 90.00, z = −3.256, *p* = 0.001, r = 0.62) were shown, which indicates that the participants had fewer limitations when performing any physical activity compared to prior to the intervention. In terms of effect size, the data show that the differences were relevant, as they presented values between 0.3 and 0.5. This suggests that a 12-week rowing training program, adapted to women who have had breast cancer, helps to improve the perceived

**Figure 4.** Variation in perceived health; Pts = points.

In summary, after an intervention based on a 12-week rowing training program tailored to women who have had breast cancer, all variables showed significant improvements, including those concerning level of physical activity and quality of life. That is, the participants were more physically active and less sedentary throughout the week. In addition, their perceived physical, emotional, and mental health status improved, which indicates fewer limitations in terms of physical activity, social life, and vitality.

#### **4. Discussion**

To date, studies relating rowing to improvements in quality of life or in the level of physical activity in breast cancer survivors are practically non-existent. In this sense, some research can be found in which other types of sports, such as dragon boating [32,33], yoga, Pilates [55,56], or endurance, strength, or aerobic exercise training programs [22,24,31], report improvements at physical, psychological, and social levels in women breast cancer survivors.

More specifically, upon examining studies that used different types of physical activity programs, the 12-week resistance training program proposed by Wiskemann et al. (2016) reported improvements only in muscle strength [24]. Other programs that combined aerobic and strength exercises for 24 weeks led to improvements in muscle strength, aerobic capacity, and only some quality of life dimensions [31], while those that combined resistance and aerobic exercises for 16 weeks showed improvements in the quality of life and physical fitness of the participants [26]. Harris (2012) and McDonough et al. (2018) reported that dragon boat rowing [32,33] led to physical, psychological, and social improvements in women breast cancer survivors.

In our study, based on a rowing training program lasting only 12 weeks, improvements were observed in all levels of physical activity, as measured by the IPAQ-SF, including increased levels of walking, moderate and vigorous physical activity, and total physical activity as well as decreased sedentary activity of the participants; improvements were shown in all dimensions of quality of life. The results obtained in the present research

are superior to those reported in previous studies; in addition, a shorter intervention time was required.

These results may be due to the fact that rowing is a sport that involves the muscles of both the lower and upper limbs and almost every muscle in the body [38]. In addition, rowing is a cyclic and symmetrical sport that combines overall strength with aerobic endurance [18]; it is based on cyclic and alternating movements of flexion/extension of the limbs and stabilization of the trunk and back muscles [57]. In this sense, the involvement of the whole body in physical activity, such as rowing, leads to improvements in quality of life and physical function as well as reduction in body fat in women breast cancer survivors [44,58,59].

In the study by Park et al. (2019), through a training program combining aerobic and resistance exercises, physical activity levels increased after the 12-week intervention, rising in 63,4% of the participants. At baseline, 33% of participants were inactive, 49.6% minimally active, and 17.4% health-enhancing physical activity (HEPA) active; after intervention the percentages improved, being 15.3%, 50.4%, and 34.2%, respectively [28]. It is of interest to compare these results with those of our study, which used an intervention of the same duration. In our study, overall physical activity levels increased in all participants after the rowing program with respect to the initial measurement. It should also be noted that prior to the intervention, about 50% of the participants reported no moderate or vigorous physical activity, whereas afterwards, all the participants reported engaging in moderate physical activity, and less than 10% reported no vigorous activity.

Regarding quality of life parameters, the study by Mascherini et al. (2020) shows improvements in physical function, social function, general health, and mental health after 6 months of training [60]. Di Blasio et al. (2017) after 12 weeks of intervention only found improvements in physical function, physical role, bodily pain, and general health [23]. Dolan et al. (2018) also showed that after a 22-week training program all SF-36 quality of life dimensions improved, with the exception of bodily pain [61].

Some of the limitations of this study reside in the scarcity of the sample. Although homogeneity was presented for the variables evaluated in this article, it would be interesting to know what the effects of this rowing program are from different times since the cancer was diagnosed or whether mastectomy is present or not.

The results of the present study demonstrate that rowing training has a greater influence on quality of life in breast cancer survivors than previously reported, as significant improvements were found in all quality of life parameters after an intervention of shorter duration. This implies that perceived physical, emotional, and mental health status and the perceived ability to perform other activities improved to a greater extent after our rowing-based training program. Indeed, 75% of the participants reported their health status as good or very good after the intervention, with a considerable increase compared to the initial data.

#### **5. Conclusions**

Previous studies have reported improved quality of life and reduced physical inactivity in female breast cancer survivors through training programs such as aerobic, strength, resistance, and dragon boat exercises. In the present study, we have shown that an intervention of just twelve weeks in length using rowing training tailored to women who have had breast cancer produced improvements in all levels of physical activity and reduced physical inactivity, with respect to the initial measurement. In addition, after the intervention, the physical, emotional, and mental health status of the participants improved, leading to fewer limitations in their daily routines, physical activity, social life, and vitality. Our rowing program showed greater benefits in health and quality of life than other studies of various durations. We can therefore conclude that rowing training contributes to increasing daily physical activity as well as improving health status and the different dimensions of quality of life in women who have survived breast cancer.

**Author Contributions:** Conceptualization J.G.-G. and A.T.-P.; methodology J.C.F.-G. and J.G.-G.; software, A.T.-P.; validation, J.C.F.-G. and J.G.-G.; formal analysis, J.C.F.-G. and A.T.-P.; investigation J.C.F.-G. and J.G.-G.; resources, J.C.F.-G. and J.G.-G.; data curation, A.T.-P.; writing—original draft preparation, J.G.-G. and A.T.-P.; writing—review and editing, J.G.-G.; J.G.-G. and J.C.F.-G.; visualization, J.G.-G.; A.T.-P. and J.C.F.-G.; supervision, J.C.F.-G.; project administration, J.C.F.-G. and J.G.-G.; funding acquisition, J.C.F.-G. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the "Researching in Sport Sciences" research group (CTS-563) of the Andalusian Research Plan.

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki and approved by Ethics Committee of University of Málaga (protocol code 65-2020-H and date of approval 30 September 2020).

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

**Data Availability Statement:** Data consent was obtained from all subjects involved in the study.

**Acknowledgments:** We would like to thank the Real Club Mediterráneo of Malaga, its rowing captain Juan Carlos Marfil Rodríguez, and the team of the Malaga Dragon Boat BCS for opening their doors to us, letting us use their facilities, and for giving us the most important thing: their time and inspiring us with their impetus and energy, thanks to which we can present the first results of this study. To Maria Repice, for her help with the English version of this manuscript. All authors have read and agreed to the published version of the manuscript.

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

#### **References**


### *Article* **Contribution of Solid Food to Achieve Individual Nutritional Requirement during a Continuous 438 km Mountain Ultramarathon in Female Athlete**

**Kengo Ishihara 1,\* , Naho Inamura <sup>1</sup> , Asuka Tani <sup>1</sup> , Daisuke Shima <sup>1</sup> , Ai Kuramochi <sup>1</sup> , Tsutomu Nonaka <sup>2</sup> , Hiroshi Oneda <sup>3</sup> and Yasuyuki Nakamura 1,4**


**Citation:** Ishihara, K.; Inamura, N.; Tani, A.; Shima, D.; Kuramochi, A.; Nonaka, T.; Oneda, H.; Nakamura, Y. Contribution of Solid Food to Achieve Individual Nutritional Requirement during a Continuous 438 km Mountain Ultramarathon in Female Athlete. *Int. J. Environ. Res. Public Health* **2021**, *18*, 5153. https://doi.org/10.3390/ijerph 18105153

Academic Editor: Filipe Manuel Clemente

Received: 21 March 2021 Accepted: 10 May 2021 Published: 13 May 2021

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

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

**Abstract:** Background: Races and competitions over 100 miles have recently increased. Limited information exists about the effect of multiday continuous endurance exercise on blood glucose control and appropriate intake of food and drink in a female athlete. The present study aimed to examine the variation of blood glucose control and its relationship with nutritional intake and running performance in a professional female athlete during a 155.7 h ultramarathon race with little sleep. Methods: We divided the mountain course of 438 km into 33 segments by timing gates and continuously monitored the participant's glucose profile throughout the ultramarathon. The running speed in each segment was standardized to the scheduled required time-based on three trial runs. Concurrently, the accompanying runners recorded the participant's food and drink intake. Nutrient, energy, and water intake were then calculated. Results: Throughout the ultramarathon of 155.7 h, including 16.0 h of rest and sleep, diurnal variation had almost disappeared with the overall increase in blood glucose levels (25–30 mg/dL) compared with that during resting (*p* < 0.0001). Plasma total protein and triglyceride levels were decreased after the ultramarathon. The intake of protein and fat directly or indirectly contributed to maintaining blood glucose levels and running speed as substrates for gluconeogenesis or as alternative sources of energy when the carbohydrate intake was at a lower recommended limit. The higher amounts of nutrient intakes from solid foods correlated with a higher running pace compared with those from liquids and gels to supply carbohydrates, protein, and fat. Conclusion: Carbohydrate, protein, and fat intake from solid foods contributed to maintaining a fast pace with a steady, mild rise in blood glucose levels compared with liquids and gels when female runner completed a multiday continuous ultramarathon with little sleep.

**Keywords:** sports nutrition; continuous glucose monitoring; carbohydrate; protein; hydration; trail running; Freestyle Libre

#### **1. Introduction**

Ultramarathon is a longer-distance marathon that has increasingly gained popularity in recent years [1]. The total energy expenditure of a 100-mile (160 km) ultramarathon reaches approximately 13,000 kcal in a 180 cm, 75 kg, middle-aged experienced male runner. The carbohydrate-derived energy in 140 min of the mountain marathon was reported to be 68%, which was lower than in 20 to 30 min of track (98%) or mountain running (86%). The percentage of lipid utilization would increase further as time and distance increased [2]. Thus, nutritional strategies are essential for ultramarathon runners wanting to improve their race results and also for those focusing primarily on finishing the event.

For endurance sports, the recommended carbohydrate intake is 30–60 g/h; however, for exercises lasting more than 3 h, the advocated recommendation is higher (i.e., ≤90 g/h and glucose:fructose ratio of 2:1) [3,4]. For a single-stage ultramarathon that generally lasts for more than 3 h, a carbohydrate level of 30–50 g/h is recommended because of numerous barriers to achieve 90 g/h consumption of a multiple-transportable carbohydrate blend [5]. Some of these barriers were described as follows. First, observational studies demonstrated that the actual carbohydrate intake during ultramarathons was less than 60 g/h in most runners [2,6,7], including slower runners consuming 37 g/h [8], and very few reached more than 60 g/h [9,10]. Second, the absolute exercise intensity of an ultramarathon was not as high as some other endurance activities because of its extremely long duration (6, 13, 24, 48, 72, or 10 days) [11]. Third, intestinal absorption might be affected by undertaking highly intensive and long-duration exercises because of the changes in the splanchnic blood flow. In addition, heat, endotoxin, or vertical shaking of their digestive system during rough terrain races could lose the appetite of ultramarathon runners [11,12]. Fourth, ultramarathon runners have to carry their food and fluids in their backpacks during long hours of racing; considering the additional weight being carried, the exercise intensity was increased [8]. Fifth, runners might find food intake difficult while maintaining their balance with both hands when running down steep mountains or climbing steep slopes.

In recent years, races and competitions of over 100 miles have increased, but the specific nutritional and hydration requirements during a continuous multiday ultra-endurance running are still insufficiently known. Reports on races longer than 100 miles [13–18] and especially on the nutritional intake are also limited [19–21].

A review on nutritional supplementation during ultramarathons mainly covers running events of 100–160 km, with a maximum of 217 km [2]. Likewise, a recently published position statement of the International Society of Sports Nutrition [5] and practical recommendations for ultramarathon events [8,14,22] are mainly based on 100–160 km studies.

Since energy intake in an ultramarathon usually exceeds the energy expenditure [5], the effects of energy deficiency would be apparent when the competition time becomes longer than several days and close to a week. The energy deficit could lead to hypoglycemia, degradation of organs, reduction of energy substrates in the blood, and decline in running performance. Firstly, hypoglycemia could reduce the running speed [23]. Hence, the minimum required amount of carbohydrates in each athlete must be identified to maintain their blood glucose levels using a continuous glucose monitoring system. Continuous blood glucose monitoring has widely been used in patients with diabetes and healthy people during exercise [24,25].

Secondly, degradation of tissues could become apparent [8,26,27]. Degradation of tissues could be attenuated by carbohydrate supplementation of 120 g/h, which is far above the recommended amount. However, it cannot enhance running performance and can even increase the incidence of gastrointestinal problems [28,29].

Third, the effects of sleep deprivation on metabolism and athletic performance might be more pronounced. The effects of sleep loss on physiological responses and exercise remain equivocal. An exhaustive review reported that sleep deprivation decreased exercise time to exhaustion, mean power, and increased heart rate [30]. In healthy non-athletes, sleep restriction induced both weight gain and diabetes risk by altering the glucose metabolism, upregulation of appetite, and decreased energy expenditure [31]. However, there are no reports on sleep deprivation and glucose control during ultra-distance events.

Optimal nutrition leads to a decreased risk of energy depletion, better performance [32], the prevention of acute cognitive decline, and improved athlete safety on ultramarathon courses with technical terrain or those requiring navigation [5]. However, the execution of the precise nutrition plan might be difficult for the runner [33] because the nutrient requirements for ultramarathon racing vary greatly, depending on the individual [5], such as age, sex, body mass, or exercise intensity.

A continuous glucose monitoring system encodes individual fluctuations of blood glucose levels in all life situations, including extreme endurance sports [24,25,34–36]. For endurance athletes, wearable devices enable them to compete while maintaining social distance and race across time and space by keeping the records in the cloud [37]. A combination of these devices could potentially improve our understanding of the complex interplay between nutrition and exercise performance.

Through analyzing the relationship between blood glucose fluctuation, running speed, and nutrient intake, one's optimal nutritional requirement could be determined. This approach might help acquire one's appropriate energy and nutrient intake, especially during long-distance events [24,34–36].

This study aimed to examine the variation of blood glucose control and its relationship with the nutritional intake and running performance in a professional female athlete during the 155.7 h of 438 km ultramarathon race with little sleep.

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

#### *2.1. Study Design*

This case study was designed to define the normal range of blood glucose during a continuous multiday ultramarathon with little sleep and observe the relationships among nutrient intakes, running performance, and blood glucose level. The Ryukoku University Human Research Ethics Review Board (No. 2019-35) approved all our procedures, which complied with the code of ethics of the World Medical Association (Declaration of Helsinki). Written informed consent was obtained from the participant before the study commenced.

#### *2.2. Study Population*

A professional female trail runner (age, 44 years; height, 1.52 m; body mass, 42 kg; body mass index, 18.2 kg/m<sup>2</sup> , body fat percentage 18%) voluntarily participated in the study. Her annual mileage, elevation gain, and training time were 4440 km, 212,000 m, and 720 h, respectively. She had completed various ultramarathon races, including 330 km certified by the International Trail Running Association, in the last 3 years. In the 35 international races she has competed in, she finished 29 times in the top 10, 16 times in the top 3, and won 4 times. Hence, she is well experienced in ultramarathons.

#### *2.3. Mountain Ultramarathon Course*

This study was conducted during a mountain ultramarathon challenge for the fastestknown time in Shiga Round Trail/Shigaichi (https://fastestknowntime.com/route/shigaround-trail-shiga-ichi-japan, accessed on 12 May 2021) held around the Lake Biwa, the largest lake in Japan (ambient temperature range: 18.3 ◦C–30.5 ◦C, relative humidity range: 42–75%), during the first week of June 2020. The distance of the course covered 438 km, and the total elevation gain was 28,300 m. Over 90% of its length unfolded along the unpaved forestry trail. The course was divided into 32 segments by 33 timing gates where investigators recorded each runner's passing time. The distances between each timing gate were 13.69 ± 5.3 SD km (range: 5.9–26.8 km). The running time and speed between each timing gate were obtained by investigators, and the global positioning system (GPS) was recorded by a wristwatch (SUUNTO 9, Suunto, Finland). Furthermore, the location and running speed of the runner were broadcasted live on the internet via a GPS-based tracking system (https://ibuki.run/, archived on 12 May 2021, IBUKI live, ONDO Inc., Kyoto, Japan). Specifically, the running time between each timing gate was 4:52 ± 2:16 h (range: 1:10–11:40 h). During the race, the runner carried her backpack containing necessities, such as food and fluid, which could be replenished at each timing gate. The total running time was 155.7 h, including 16.0 h of rest and sleep. The total hours of rest and sleep each day were 0:20, 1:55, 1:40, 4:00, 3:30, 2:45, 1:50 h:min plus 2 h of fragments of sleep during running.

#### *2.4. Running Pace Data Collection and Standardization*

The running performance during the ultramarathon was calculated using the following arbitrary formula:

$$\text{Pace (min/km)} = (\text{E} - \text{A}) / \text{(Distance of the segment (km))}$$

where

E: the estimated running time of the segment (min); A: the actual running time of the segment (min).

The pace value was positive when the runner ran faster than the estimated running time but negative when the runner ran slower. Before the ultramarathon, the runner ran 3 1/2 trial laps of this course to determine the estimated running time of the segment.

#### *2.5. Glucose Data Collection and Standardization*

All of the blood glucose profile was monitored by flash glucose monitoring (FGM), a minimally invasive method described in previous reports [38–42]. Briefly, the FGM system (FreeStyle Libre; Abbott Diabetes Care, Alameda, CA, USA) mechanically reads and continuously measures the glucose concentration in the interstitial fluid collected right below the skin and subsequently reveals the corresponding ambulatory glucose profile. From one day before the race to three days after the race, the FGM sensor was applied at the back of the runner's upper arm, and glucose concentrations were obtained every 15 min [39]. The highest and lowest glucose concentration levels in each segment and their difference were used as representative values in every segment.

The samples for the plasma clinical parameters were collected after an overnight fast one month before and one week of an off-training period after the ultramarathon and analyzed. The ultramarathon was planned to take place a month earlier, but due to the COVID-19 infection situation and the soft lockdown by the Japanese government, the start of the ultramarathon was delayed by a month. Blood was collected in a clotted vial, and the serum obtained was analyzed by clinical laboratory testing (Falco Biosystems, Inc., Kyoto, Japan). Aspartate aminotransferase (AST), alanine aminotransferase (ALT), and creatine kinase (CK) were estimated by the JSCC standard method. Alkaline phosphatase (ALP) and lactate dehydrogenase (LDH) were estimated by the IFCC standard method. Triglyceride and LDL-cholesterol were estimated by the enzyme colorimetric method. Sodium, potassium, and chloride were estimated by the ion-selective electrode method. Calcium was estimated by the arcenazo III colorimetric test.

#### *2.6. Diet Supply Data Collection*

Investigators followed the runner and recorded the entire food and drink intake in relays throughout the ultramarathon. They reported the timing and volume of consumed food and fluid products based on pictures taken during the race. In detail, one or two of the investigators always ran with the runner, taking turns in each segment. They checked the current location, based on GPS, when the runner consumed the refueling meal and recorded the consumption point on a map. Food and fluid products consumed more than 60 min before the start of the ultramarathon were excluded in the nutrient intake calculation with reference to previous studies [23,43]. The nutrition information indicated on the cover of the food and fluid products was our basis when calculating the energy and nutrient intake. If data were unavailable, the intake was calculated according to the standard tables of food composition in Japan 2015 (7th revised edition) [44]. The energy and nutrient intake was expressed relative to the pre-race body weight (kg) per running time (h). In reference to previous research [23,44], all foods were categorized as follows: sports drinks (isotonic and hypertonic formulas), cola, gels, milk product, tea, soup, other liquids (all other drinks consumed), fruits, sweets, bars, noodles, bread, rice products, wheat products, powder, and other solids (all other products consumed).

As shown in Table 1, the runner consumed energy and nutrients from liquids, gels, fruits, sweets, and other solids. The hourly intake of energy, protein, fat, carbohydrate, water, and salt was 170.8 kcal, 5.9 g, 3.1 g, 29.7 g, 263.0 g, and 1.1 g, respectively. The protein:fat:carbohydrate ratio of the ingested nutrients was 13.7:16.5:69.5. The runner consumed 30.1% and 58.3% of their energy from liquids and gels, and solids, respectively. Identically, the intake of carbohydrates from solids (44.4%) was similar to that from liquids or gels (41.4%). Meanwhile, proteins (89.5%) and fats (84.3%) were mostly consumed from solids. Other solids included smoked chicken, potatoes, risotto, lasagna, and protein powder. Other liquids included smoothies and coffee.


**Table 1.** Total energy and nutrients consumed during the ultramarathon.

The subtotal of each category is shown in bold.

#### *2.7. Statistics*

Herein, numerical data are presented as means and standard deviations unless otherwise specified. Data from a female ultrarunner were processed and analyzed in GraphPad Prism for Mac (version 9.0.1, GraphPad Inc., San Diego, CA, USA). The associations between the running performance, glucose level, and nutrient intake were investigated using Spearman's rank correlation coefficients. The differences among each situation of the blood glucose level were compared by a Mann–Whitney test by ANOVA followed by Dunn's multiple comparison test for the comparison among more than three groups. Results were considered significant when *p* < 0.05. Limitations of the single-subject research design are the generalizability of the study conclusions and were described in the discussion section.

#### **3. Results**

#### *3.1. General Results of Blood Glucose Fluctuation during the Ultramarathon*

During the 7-day ultramarathon, the regular circadian rhythms, including breakfast, lunch, and dinner, almost disappeared, as detected in the blood glucose levels (Figure 1A). Additionally, the mean blood glucose levels (25–30 mg/dL) were higher than those during the preliminary and post-ultramarathon periods (*p* < 0.0001, Figure 1B). The mean daytime and nighttime blood glucose levels during the ultramarathon were 130.0 ± 16.2 and 124.7 ± 17.3 mg/dL, respectively, with a slight difference (*p* < 0.001, Figure 1C). Moreover, the blood glucose levels during the ultramarathon were controlled within a narrower range than during the preliminary period (Figure 1D).

than during the preliminary period (Figure 1D).

Additionally, the mean blood glucose levels (25–30 mg/dL) were higher than those during the preliminary and post-ultramarathon periods (*p* < 0.0001, Figure 1B). The mean daytime and nighttime blood glucose levels during the ultramarathon were 130.0 ± 16.2 and 124.7 ± 17.3 mg/dL, respectively, with a slight difference (*p* < 0.001, Figure 1C). Moreover, the

**Figure 1.** Blood glucose fluctuation during the 7-day ultramarathon. Each solid line represents the daily glucose variation (**A**). Scatter plot of blood glucose during a preliminary, ultramarathon, and post-ultramarathon periods, respectively (**B**). Scatter plot of blood glucose levels during night (dark) and day (light) throughout the ultramarathon (**C**). Histogram of blood glucose fluctuation during preliminary, ultramarathon (night and day), and post-ultramarathon periods (**D**). \*\*\* *p* < 0.001, \*\*\*\* *p* < 0.0001 The differences between dark and light were compared by the Mann–Whitney test, and those among pre, ultramarathon post were compared by the Kruskal–Wallis nonpar-**Figure 1.** Blood glucose fluctuation during the 7-day ultramarathon. Each solid line represents the daily glucose variation (**A**). Scatter plot of blood glucose during a preliminary, ultramarathon, and post-ultramarathon periods, respectively (**B**). Scatter plot of blood glucose levels during night (dark) and day (light) throughout the ultramarathon (**C**). Histogram of blood glucose fluctuation during preliminary, ultramarathon (night and day), and post-ultramarathon periods (**D**). \*\*\* *p* < 0.001, \*\*\*\* *p* < 0.0001 The differences between dark and light were compared by the Mann–Whitney test, and those among pre, ultramarathon post were compared by the Kruskal–Wallis nonparametric ANOVA, followed by Dunn's multiple comparison test.

#### ametric ANOVA, followed by Dunn's multiple comparison test. *3.2. Relationship between the Amount of Nutrient Intake and Maintenance of Glucose Level during the Ultramarathon*

*3.2. Relationship between the Amount of Nutrient Intake and Maintenance of Glucose Level during the Ultramarathon*  Significant correlations between the runner's blood glucose levels and nutrient intake were not observed. The lowest glucose level between the segments tended to correlate with protein intake (*p* = 0.06). The highest blood glucose level in the segment was not significantly related to nutrient intake. The difference between the two lines was 50 mg/dL Significant correlations between the runner's blood glucose levels and nutrient intake were not observed. The lowest glucose level between the segments tended to correlate with protein intake (*p* = 0.06). The highest blood glucose level in the segment was not significantly related to nutrient intake. The difference between the two lines was 50 mg/dL approximately, which was the difference between the highest and lowest blood glucose levels in each segment. Among all nutrients, the difference did not vary significantly from about 50 mg/dL, regardless of the amount consumed (Figure 2).

about 50 mg/dL, regardless of the amount consumed (Figure 2).

approximately, which was the difference between the highest and lowest blood glucose levels in each segment. Among all nutrients, the difference did not vary significantly from

*Ultramarathon* 

(Figure 4).

**Figure 2.** Scatter plots showing the relationships between nutrient intake and blood glucose level. The intake of energy (**A**), carbohydrate (**B**), protein (**C**), fat (**D**), water (**E**), and salt (**F**) was calculated according to the consumed food and fluid products. Each plot indicates one segment. **Figure 2.** Scatter plots showing the relationships between nutrient intake and blood glucose level. The intake of energy (**A**), carbohydrate (**B**), protein (**C**), fat (**D**), water (**E**), and salt (**F**) was calculated according to the consumed food and fluid products. Each plot indicates one segment. *3.3. Relationship between Glucose Level and Running Pace*  During exercise, the runner was within the expected normoglycemic range (86–185

#### *3.3. Relationship between Glucose Level and Running Pace* mg/dL), with no extreme hyperglycemia or hypoglycemia (Figure 1). Therefore, the run-

*3.3. Relationship between Glucose Level and Running Pace*  During exercise, the runner was within the expected normoglycemic range (86–185 mg/dL), with no extreme hyperglycemia or hypoglycemia (Figure 1). Therefore, the running pace had no significant correlation with the highest blood glucose level (*p* = 0.79), lowest blood glucose level (*p* = 0.32), and delta (difference between the highest and lowest During exercise, the runner was within the expected normoglycemic range (86–185 mg/dL), with no extreme hyperglycemia or hypoglycemia (Figure 1). Therefore, the running pace had no significant correlation with the highest blood glucose level (*p* = 0.79), lowest blood glucose level (*p* = 0.32), and delta (difference between the highest and lowest blood glucose levels, *p* = 0.36) between segments (Figure 3). ning pace had no significant correlation with the highest blood glucose level (*p* = 0.79), lowest blood glucose level (*p* = 0.32), and delta (difference between the highest and lowest blood glucose levels, *p* = 0.36) between segments (Figure 3).

*Int. J. Environ. Res. Public Health* **2021**, *18*, x FOR PEER REVIEW 7 of 17

segments. The pace was calculated from the difference between the estimated running time and actual running time of the segment, as described in the Methods section. Briefly, the pace value was positive when the runner ran faster than the estimated running time and negative when the runner ran slower. *3.4. Relationship between the Amount of Nutrient Intake and Running Pace during the*  **Figure 3.** Relationship between the running pace and the highest blood glucose level, lowest blood glucose level, and delta (difference between the highest and lowest blood glucose levels) between segments. The pace was calculated from the difference between the estimated running time and actual running time of the segment, as described in the Methods section. Briefly, the pace value **Figure 3.** Relationship between the running pace and the highest blood glucose level, lowest blood glucose level, and delta (difference between the highest and lowest blood glucose levels) between segments. The pace was calculated from the difference between the estimated running time and actual running time of the segment, as described in the Methods section. Briefly, the pace value was positive when the runner ran faster than the estimated running time and negative when the runner ran slower.

*3.4. Relationship between the Amount of Nutrient Intake and Running Pace during the* 

The running pace significantly correlated with energy (*p* = 0.02) and carbohydrates

(*p* = 0.01). The running pace tended to correlate with protein (*p* = 0.10) and water intake (*p*

= 0.06). Other nutrient intake data did not show the correlation with the running pace

(*p* = 0.01). The running pace tended to correlate with protein (*p* = 0.10) and water intake (*p* = 0.06). Other nutrient intake data did not show the correlation with the running pace

runner ran slower.

*Ultramarathon* 

(Figure 4).

cates one segment.

#### *3.4. Relationship between the Amount of Nutrient Intake and Running Pace during the Ultramarathon*

The running pace significantly correlated with energy (*p* = 0.02) and carbohydrates (*p* = 0.01). The running pace tended to correlate with protein (*p* = 0.10) and water intake (*p* = 0.06). Other nutrient intake data did not show the correlation with the running pace (Figure 4). *Int. J. Environ. Res. Public Health* **2021**, *18*, x FOR PEER REVIEW 8 of 17

**Figure 4.** Scatter plots showing the relationships between nutrient intake and running pace. The intake of energy (**A**), carbohydrate (**B**), protein (**C**), fat (**D**), water (**E**), and salt (**F**) was calculated **Figure 4.** Scatter plots showing the relationships between nutrient intake and running pace. The intake of energy (**A**), carbohydrate (**B**), protein (**C**), fat (**D**), water (**E**), and salt (**F**) was calculated according to the nutrition information of the consumed food and fluid products. Each plot indicates one segment.

according to the nutrition information of the consumed food and fluid products. Each plot indi-

#### *3.5. Comparison of Nutrient Intake between Fast and Slow Running Paces*

*3.5. Comparison of Nutrient Intake between Fast and Slow Running Paces*  The energy and nutrient intake in the positive running pace (faster than planned) were compared with that in the negative running pace (slower than planned). The energy (*p* < 0.01), carbohydrate (*p* < 0.05), protein (*p* < 0.01), fat (*p* < 0.05), water (*p* < 0.01), and salt (*p* < 0.05) intake in the positive running pace was significantly higher than that in the neg-The energy and nutrient intake in the positive running pace (faster than planned) were compared with that in the negative running pace (slower than planned). The energy (*p* < 0.01), carbohydrate (*p* < 0.05), protein (*p* < 0.01), fat (*p* < 0.05), water (*p* < 0.01), and salt (*p* < 0.05) intake in the positive running pace was significantly higher than that in the negative running pace. In the positive running pace, the median hourly nutrient intake for carbohydrate, protein, fat, water, and salt was 38.0 g/h, 9.0 g/h, 5.0 g/h, 413.0 mL/h, and 1.6 g/h, respectively, and the median hourly energy intake was 225 kcal/h (Figure 5).

#### ative running pace. In the positive running pace, the median hourly nutrient intake for *3.6. Comparison of Food Type between Fast and Slow Running Paces*

carbohydrate, protein, fat, water, and salt was 38.0 g/h, 9.0 g/h, 5.0 g/h, 413.0 mL/h, and 1.6 g/h, respectively, and the median hourly energy intake was 225 kcal/h (Figure 5). Food product types used for energy and nutrient consumption were compared in terms of the positivity or negativity of the running pace. When the running pace was positive, the energy and nutrient intake from solids was approximately two times higher than that when it was negative (*p* < 0.05). The water intake was mainly derived from liquids or gels and the intake of liquid (Figure 6).

**Figure 5.** Comparison of energy (**A**), carbohydrate (**B**), protein (**C**), fat (**D**), water (**E**), and salt (**F**) intake among different running paces (fast, when the running pace was faster than planned; on time, when the running pace was the same as planned; slow, when the running pace was slower than planned). \* *p* < 0.05 and \*\* *p* < 0.01 between fast and slow, Mann–Whitney test. The horizontal bar represents the median in each group. **Figure 5.** Comparison of energy (**A**), carbohydrate (**B**), protein (**C**), fat (**D**), water (**E**), and salt (**F**) intake among different running paces (fast, when the running pace was faster than planned; on time, when the running pace was the same as planned; slow, when the running pace was slower than planned). \* *p* < 0.05 and \*\* *p* < 0.01 between fast and slow, Mann–Whitney test. The horizontal bar represents the median in each group. *Int. J. Environ. Res. Public Health* **2021**, *18*, x FOR PEER REVIEW 10 of 17

**Figure 6.** Comparison of product type for energy (**A**), carbohydrate (**B**), protein (**C**), fat (**D**), water (**E**), and salt (**F**) consumption among different running paces (fast, when the running pace was faster than planned; on time, when the running pace was the same as planned; slow, when the running pace was slower than planned). \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001 between fast and slow, **Figure 6.** Comparison of product type for energy (**A**), carbohydrate (**B**), protein (**C**), fat (**D**), water (**E**), and salt (**F**) consumption among different running paces (fast, when the running pace was faster than planned; on time, when the running pace was the same as planned; slow, when the running pace was slower than planned). \* *p* < 0.05, \*\* *p* < 0.01, \*\*\* *p* < 0.001 between fast and slow, Mann–Whitney test. Values are means ± SEM.

Triglycerides in the blood were greatly reduced, suggesting that lipid utilization contributed significantly to energy production. The total protein in the blood was also slightly decreased. The decrease in ALP could be due to insufficient zinc intake. On the other hand, there was only a slight increase in AST and ALT compared to the marked increase in LDH and CK, suggesting that although there was muscle damage, the damage on the

**Table 2.** Serum parameters pre- and post-ultramarathon compared with the Japanese population.

 **Normal Range Ultramarathon Mean (95% CI) Pre Post**  Total protein g/dL 7.4 ± 0.5 (6.5–8.3) 6.9 5.8 L Triglyceride mg/dL 89.5 ± 30.4 (30.0–149.0) 140 33 LDL-cholesterol mg/dL 104.5 ± 17.6 (70.0–139.0) 65 L 42 L AST U/L 23.0 ± 7.7 (8.0–38.0) 31 87 H ALT U/L 23.5 ± 9.9 (4.0–43.0) 35 84 H ALP U/L 232.0 ± 62.2 (110.0–354.0) 258 70 L LDH U/L 183.0 ± 31.6 (121.0–245.0) 184 428 H CK U/L 117.0 ± 40.3 (38.0–196.0) 102 1312 H Na mEq/L 142.5 ± 3.8 (135.0–150.0) 137 142 Cl mEq/L 104.0 ± 3.1 (98.0–110.0) 101 106 K mEq/L 4.4 ± 0.5 (3.5–5.3) 4.7 3.9 Ca mg/dL 9.3 ± 0.5 (8.4–10.2) 9 8.5 L, lower than normal range CI; H, higher than normal range; AST, aspartate aminotransferase; ALT, alanine aminotransferase; ALP, alkaline phosphatase; lactate dehydrogenase; LDH, lactate

*3.7. Change of Serum Parameters Pre- and Post-Ultramarathon* 

liver function was small (Table 2).

dehydrogenase; CK, creatine kinase.

#### *3.7. Change of Serum Parameters Pre- and Post-Ultramarathon*

Triglycerides in the blood were greatly reduced, suggesting that lipid utilization contributed significantly to energy production. The total protein in the blood was also slightly decreased. The decrease in ALP could be due to insufficient zinc intake. On the other hand, there was only a slight increase in AST and ALT compared to the marked increase in LDH and CK, suggesting that although there was muscle damage, the damage on the liver function was small (Table 2).


**Table 2.** Serum parameters pre- and post-ultramarathon compared with the Japanese population.

L , lower than normal range CI; <sup>H</sup>, higher than normal range; AST, aspartate aminotransferase; ALT, alanine aminotransferase; ALP, alkaline phosphatase; lactate dehydrogenase; LDH, lactate dehydrogenase; CK, creatine kinase.

#### **4. Discussion**

This study aimed to examine the variation of blood glucose control and its relationship with nutritional intake and running performance in a professional female athlete during the continuous over 400 km ultramarathon race with little sleep. Diurnal variation had almost disappeared with the overall average glucose increase of approximately 30 mg/dL compared to resting. A significantly faster running speed correlated with a higher energy and nutrient intake from solid foods than from gels and liquids. Interestingly, the median energy and carbohydrate intake in the fast-running pace were within the recommended energy and carbohydrate intake, mainly covered 100–160 km ultramarathon [5,6].

Protein intake contributed to the maintenance of blood glucose levels as carbohydrate intake was at the lower end of the recommended amount. Sufficient energy and nutrient intake prevented hypoglycemia, thereby maintaining the running speed during the ultramarathon. Consistent with the findings from a 100-mile race [23], the highest blood glucose concentration obtained was not associated with the running speed, indicating that instead of the rapid availability of carbohydrates, nutrient intake from solid foods for controlling glucose homeostasis was the key determinant of performance especially in an ultramarathon of over 400 km.

Recording the amount of energy and nutrient intake during prolonged exercise events had corresponding difficulties. A bicycle equipped with a camera cycled alongside the runner to accurately record the results [45]. In the present study, given that the course mostly covered a single track where bicycles could not pass, the ultramarathon runners were followed by other runners to record the food and drink intake; hence, the energy and nutrient intake recorded was precise and valuable. This ultramarathon gained considerable attention that several runners took turns to accompany the ultramarathon runner for approximately one week to keep track of her meals and drinks.

Gluconeogenesis played an essential role in maintaining blood glucose levels during an ultramarathon, considering that meeting carbohydrate consumption throughout the entire ultramarathon race was not feasible, not even in typical durations of an ultramarathon (6–48 h). Energy deficiency was common in ultramarathons [5–7,32,45–48]. Studies using

a doubly labeled water technique or respiratory gas analysis estimated that the energy expenditure during 160 km ultramarathons was approximately 13,000 kcal [2,49,50]. The previously reported rates of gluconeogenesis and hepatic glycogenolysis in a resting state in low-carbohydrate–fed subjects were 0.07 and 0.03 g/kg/h, respectively [51]. The sum of these two values (0.1 g/kg/h), otherwise known as endogenous glucose production, would be the minimum required amount of carbohydrates to maintain blood glucose levels during a resting state. The endogenous glucose production significantly increased to 0.36 g/kg/h during exercise at 55% of peak power output [51] or 0.48 g/kg/h during exercise at the lactate threshold level in fasted, well-trained subjects [52]. In accordance with this calculation, the runner in the present study required 29.7 and 5.9 g/h of carbohydrate and protein intake, respectively, to maintain her blood glucose concentrations during the ultramarathon.

Consuming mostly solid foods alongside other carbohydrate forms can be a practical option for the supplementation of adequate energy and nutrients with fewer gastrointestinal problems, especially for a race that lasts for several days. Previous reports using <sup>13</sup>C-labeled isotopes revealed that carbohydrates from solid foods (as well as from liquids) were effectively oxidized during exercise and could suppress gastric emptying compared with the liquid form [43]. Solid foods slightly elevated blood glucose levels and secreted less insulin [53] and glucose-dependent insulinotropic peptide compared with liquid foods [54]. Furthermore, gastric emptying of semisolid food was not affected by exercises at intensities of the 40% VO<sup>2</sup> peak [55]. Solid foods could maintain the same blood glucose levels as gel foods and perform the same intensity of cycle exercise and time trials [56]. According to recent reports, ingestion of a larger volume of carbohydrate solution at less frequent intervals during prolonged submaximal running spared endogenous carbohydrate oxidation rates. It did not cause increased markers of gastrointestinal discomfort compared with the smaller volumes at more frequent intervals [57].

A marked increase in CK, compared with the other biochemical markers, such as ALT, AST, LDH, were reported in previous studies on 130 to 160 km of ultramarathon [26,27]. In a longer-distance ultramarathon, a significant increase in CK was observed [8]. On the other hand, ALP was mildly elevated in this previous study but decreased in the present study. The reduction of ALP might be due to insufficient zinc intake in this study.

In the present study, there was a correlation between nutrient intake and speed, as the intake of energy and nutrients were insufficient compared with the energy expenditure during the ultramarathon. Insufficient intake was speculated by the decrease in blood triglycerides and total protein concentration in the present ultramarathon. Our previous study also revealed that the lowest blood glucose level in each section was the cause of the running speed reduction, though the highest blood glucose level in each section of the run was not related to the running speed [23]. Fatigue is caused by various factors, and excessive intake did not entirely enhance performance. The weak correlation between the blood glucose level and the running speed could be explained by the previously reported gender-specific differences in fuel utilization during exercise. Females showed higher lipid oxidation caused by higher plasma adiponectin levels [58], higher muscle triglyceride utilization [59], low plasma glucose levels [60], and higher fasting hepatic glucose uptake than males [61].

Levels of glucose increased by an average of approximately 20 percent during the early part of nocturnal sleep but returned to baseline levels in the morning because of reduced glucose utilization during sleep [62]. Similarly, glucose tolerance was optimal in the morning and reached its minimum in the middle of the night [63,64]. Another study revealed an association between sleep and glucose regulation during constant glucose infusion, which was a condition that inhibited endogenous glucose production and, therefore, revealed changes in glucose utilization [62].

The intravenous glucose tolerance test during the sleep restriction condition demonstrated that the rate of glucose clearance was approximately 40% lower and the acute insulin response to glucose was 30% lower compared to the sleep extension condition [65]. Interestingly, the diurnal rhythm of blood glucose levels in the present study was almost abolished compared to the resting state, and the difference between daytime and nighttime blood glucose levels was significant but not pronounced. This change was caused by the combined factors of running and sleep deprivation, but the mean and SD of blood glucose levels did not gradually increase along with the accumulation of sleep deprivation during the ultramarathon. These results suggested that running in a sleep-deprived state for at least up to a week did not cause extreme fluctuations in blood glucose levels.

Sleep deprivation of 30 to 72 h did not drastically affect cardiovascular and respiratory responses to exercise of varying intensity or the aerobic and anaerobic performance capability [66]. For example, during prolonged treadmill walking at about 80% of the VO2max, the reduction of work time to exhaustion was only 11% after 30 h of sleep deprivation [67]. Another study reported that the maximal isometric and isokinetic muscular strength and endurance of selected upper and lower body muscle groups, the performance of the Wingate Anaerobic Power Test, simple reaction time, the blood lactate response to cycle exercise at 70% VO2max, and most of the cardiovascular and respiratory responses to treadmill running at 70% and 80% VO2max, were not significantly altered as a result of sleep deprivation of 60 h [68]. Although the sleep deprivation in this study was 155.7 h, the runner was not in complete sleep deprivation, and its relationship to performance needed to be carefully examined.

Meanwhile, the main limitation of this study was the number of subjects. It was not feasible to have sufficient subjects, as few people even attempt to run through the entire route of the 438 km of a mountain ultramarathon for continuous several days. In the present ultramarathon, another runner started simultaneously but retired in the middle of the race. Though this research was a case study, our study athlete set the fastest time; thus, our record could be considered valuable. Our observational study supported the effectiveness of the position statement of the International Society of Sports Nutrition [5] and practical recommendation for ultramarathon participants to prevent hypoglycemia during exercise.

The limitations of the single-subject research design were the generalizability of the study conclusions and the methodological and statistical assumptions that were typically needed for inferential statistical tests. A single-subject design provided limited support for conclusions regarding populations of subjects [69]. Nonetheless, single-subject studies have been conducted in rehabilitation, disability [70,71], or psychological research [72]. These study designs employ a comparison in the AB design. In this design, A represents the baseline, and B represents the treatment. The subject is treated repeatedly as AB or ABA [73,74]. Similarly, ABAB or other extensions of AB designs are stronger designs than the simple phase change of an AB design [71]. In the present study, the statistical data was processed as an extended AB design consisting of segments of faster or slower than the planned pace. In the future, these findings need to be accumulated to reach a general conclusion.

Another methodological limitation was the large fluctuations in running speed during the ultramarathon. The running speed in the mountain ultramarathon usually varied by vast diversities of terrain [75] and physiological changes such as muscle fatigue and energy deficiency. We could not measure the heart rate as the runners did not tolerate the discomfort of wearing the belt for six days. We used a GPS tracking system to monitor this run, but it was difficult to calculate the intensity from the distance and slope because more than 90% of this route was a track in the forest, and the surface was diverse. Therefore, the running speed values of the runner were standardized using the preliminary planned running time, which was obtained by allowing the runner to run 3 1/2 trial laps of this course before the ultramarathon. The precise and objective power meters were already applicable in cycling studies [76].

Continuous glucose monitoring systems were less accurate than the gold standard for intermittent self-measurement of blood glucose [40]. The present research adopted the system, which was pronounced as superior performance during exercise compared with the other GM systems [41]. Although we rarely observed hypoglycemia in this study, caution should be exercised regarding the accuracy of values at low blood glucose levels as the median absolute relative difference between the reference values and those obtained by the sensor across the glycemic range overall was 22 (13.9–29.7)% and was 36.3 (24.2–45.2)% during hypoglycemia, 22.8 (14.6–30.6)% during euglycemia and 15.4 (9–21)% during hyperglycemia [77].

GPS monitoring throughout the entire running and sharing the runner's location was linked to the assured dietary support and safety of the runner in the present study. Sharing the blood glucose level as well was expected to ensure further safety. Accurate physical workload calculation based on GPS monitoring or running power meter would enable a more accurate analysis of the running performance and blood glucose fluctuations.

#### **5. Conclusions**

In conclusion, the diurnal variation of the plasma glucose level had almost disappeared with the overall slight glucose increase during a continuous multiday ultramarathon in a female athlete. The intake of protein and fat directly or indirectly contributed to maintaining the blood glucose levels and running speed as gluconeogenesis source or energy source when the intake of carbohydrates was at the lower limit of dietary recommendation. Carbohydrate, protein, and fat intake from solid foods contributed to maintaining a fast pace compared with liquids and gels.

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

**Funding:** This research was funded by the Ryukoku University and Nagatasangyo.co. (Hyogo, Japan).

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Institutional Review Board of Ryukoku University (Protocol no. 2019-35, Date of approval: 18 Februay 2020).

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study. Written informed consent has been obtained from the patient to publish this paper.

**Acknowledgments:** We express our gratitude and deep appreciation to Kaori Niwa, Hitomi Matsubara, and Noriyuki Niwa for their kind support. We also thank all the participants for their cooperation in the investigation.

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

#### **References**

