Phase angle and trunk muscle mass as distinguishing factors between regular and non-regular players in amateur soccer teams

Abstract

[Purpose] This study aimed to identify the body composition characteristics that differentiate regular and non-regular players in amateur soccer teams, thereby providing objective metrics for player selection and enhancing team performance. [Participants and Methods] A total of 147 male amateur soccer players (mean age 17.2 ± 4.9 years) were included and categorized as regular (n=68) or non-regular (n=79) players. Parameters of body composition including muscle mass and phase angle were measured using an InBody S10 analyzer. Groups were compared using unpaired t-tests and analysis of covariance adjusted for age, height, and weight. [Results] The regular players demonstrated significantly higher lean body mass, muscle mass, and phase angle. The trunk muscle mass and phase angle emerged as the key distinguishing variables. Logistic regression analysis revealed a significant association between phase angle and player regularity. [Conclusion] Trunk muscle mass and phase angle are effective indicators of regular player status. Phase angle, in particular, showed utility in objective player evaluations. These findings support the integration of trunk muscle mass and phase angle into the selection criteria and conditioning protocols for amateur soccer teams.

INTRODUCTION

Soccer is one of the most widely played sports globally and requires a scientific approach to player selection. A range of performance tests has been developed to evaluate player capabilities, which are known to depend on physical morphology and body composition¹⁾. Regular assessments of body composition allow teams to individualize strength and conditioning protocols and nutritional strategies aimed at maintaining and enhancing athletic performance¹⁾.

Excess body fat adversely affects endurance²⁾, whereas greater muscle mass is associated with enhanced physical output. As such, body composition analysis is essential for optimizing sports performance, and elite athletes routinely undergo such evaluations³⁾. While these performance tests provide valuable insights, there remains a need for objective and quantifiable criteria that can be used consistently across teams for player selection.

Phase angle (PhA) is a key metric derived from bioelectrical impedance analysis (BIA) that reflects cellular membrane integrity and body cell mass. It is calculated as [arc tangent (Xc/R) × (180/π)], where resistance (R), reactance (Xc), and impedance (Z) are measured⁴⁾. Typically, higher PhA values are observed in healthy individuals and athletes⁵⁾. PhA is influenced by both race and sex, with lower values reported in Asian populations compared to Western populations, and higher values in men than in women⁵⁾. This parameter has been widely studied as a marker of nutritional status and as a prognostic indicator.

Notably, Sato et al. proposed a PhA cut-off value of 2.95° for predicting 1 year mortality in residents of elderly care facilities using the same BIA method applied in the present study⁶⁾. Although their study focused on an older population, the methodology highlights how PhA cut-off values may assist in classifying physiological states. Expanding on this application, recent research has explored PhA in athletes, examining seasonal changes in soccer players and associations between PhA and injury risk^{7, 8)}. However, research on PhA in Japanese athletes remains limited.

A soccer match consists of 11 primary players on the field, with several substitutes allowed during gameplay. In Japan, up to seven player substitutions are permitted per match⁹⁾, and it is standard practice to classify players as follows: “regular (R)” if they play or substitute in more than 50% of their total match appearances during a season, and “non-regular (NR)” if they participate in fewer than 50%¹⁰⁾.

Previous studies have primarily examined performance differences between R and NR players. R players tend to cover greater total distance, engage more frequently in sprinting, and perform more high-speed runs during matches than NR players¹¹⁾. In contrast, although NR players accumulate a lower overall workload because of reduced playing time, they often demonstrate a higher frequency of high-intensity efforts during their time on the field and are expected to influence match dynamics^{12, 13)}. Building on these findings, our study seeks to explore how PhA, a similar marker, might be applied specifically to soccer players and whether it can serve as a reliable indicator for distinguishing between regular and NR players.

In professional teams, player selection is typically informed by data from fitness assessments and input from the coaching staff. However, in community-based teams and school clubs, less experienced coaches may rely more on subjective judgment. Such subjectivity can compromise fairness and transparency, potentially diminishing player motivation and fostering a sense of inequity within the team. Therefore, scientific evaluation of player capabilities and selection based on objective indicators, are essential for player development.

Body composition assessment, which is noninvasive, quick, and requires minimal expertise, serves as a valuable objective standard in the player selection process. We hypothesized that phase angle would be significantly higher in regular players than in NR players, and that this parameter could serve as an objective index for player classification. This study aims to establish objective selection criteria that support performance enhancement and informed tactical decision-making.

PARTICIPANTS AND METHODS

This study included 147 male participants with a mean age of 17.2 ± 4.9 years, a mean height of 168.2 ± 7.8 cm, and a mean weight of 58.7 ± 10.2 kg. The cohort comprised 29 players from adult teams, 64 from high school club teams, and 54 from junior high school club teams. All participants were members of amateur soccer clubs competing at the prefectural level, and several had been selected for prefectural representative teams. The exclusion criterion was a traumatic injury sustained within the previous 3 months. All participants engaged in soccer training at least 5 times per week. Body composition assessments were conducted on non-training days, with participants instructed to consume meals at least 2 hours prior to measurement. Although fluid intake was not strictly restricted, participants were advised to avoid excessive fluid consumption before testing. Previous studies using multi-frequency BIA devices, including the Tanita MC-780PMA (Tanita Corporation, Tokyo, Japan), have demonstrated good to excellent reliability even following food and fluid intake, indicating that minor daily variations in hydration status are unlikely to significantly affect measurement accuracy¹⁴⁾. Participants were also instructed to empty their bowels before the assessment. For ethical compliance, informed consent was obtained from adult participants and from both guardians and coaches for minors before conducting body composition assessments using the InBody S10 (InBody Japan, Tokyo, Japan). Muscle mass (total body, upper limbs, trunk, and lower limbs), fat mass, lean body mass, skeletal muscle index, and PhA were measured. This study was approved by the Research Ethics Committee of the International University of Health and Welfare (approval number: 21-Io-34-2) and was conducted in accordance with the ethical standards of the 1964 Declaration of Helsinki and its subsequent revisions.

Participants were instructed to prepare for the assessment by removing their shoes and socks and wearing light clothing. Height and weight were recorded using a stadiometer and digital scale, respectively. Body composition parameters, including muscle mass, water content, and PhA, were evaluated using the InBody S10 body composition analyzer. Measurements were conducted with participants seated, backs not touching the chair’s backrest, arms relaxed and positioned approximately 15° from the trunk, and feet placed shoulder-width apart.

Statistical analyses were performed by dividing participants into two groups: regular (R) and NR. Initially, an unpaired t-test was used to compare body composition parameters between groups. To control for the potential effects of age, height, and weight on body composition, analysis of covariance (ANCOVA) was subsequently applied.

Additionally, a binomial logistic regression analysis was conducted to identify independent predictors of R/NR classification, using trunk muscle mass and phase angle as explanatory variables. The predictive performance of the model was evaluated using receiver operating characteristic (ROC) curve analysis and the corresponding area under the curve (AUC).

All statistical procedures were performed using SPSS version 29 (IBM, Chicago, IL, USA), with statistical significance set at p<0.05.

RESULTS

Data from the R and NR groups were analyzed using an unpaired t-test and analysis of covariance (ANCOVA), with age, height, and weight included as covariates (Table 1). The unpaired t-test revealed significant group differences in lean body mass (kg), total body muscle mass (kg), upper limb muscle mass (kg), trunk muscle mass (kg), lower limb muscle mass (kg), and PhA (p<0.05). ANCOVA further identified significant differences in trunk muscle mass (p<0.05) and PhA (p<0.001).

Table 1. Comparison of basic attributes between R and NR groups.

	R group	NR group	ALL	p-value
	n=68	n=79	n=147
Age (years)	17.8 ± 4.5	16.8 ± 5.2	17.2 ± 4.9
Height (cm)	169.5 ± 7.1	167.1 ± 8.2	168.2 ± 7.8
Weight (kg)	60.0 ± 8.9	57.4 ± 11.1	58.6 ± 10.2
Body Fat (kg)	7.0 ± 3.3	7.4 ± 3.8	7.2 ± 3.6
Lean body mass (kg)	53.0 ± 7.1	49.9 ± 8.9	51.3 ± 8.2	*
Muscle mass (kg)	49.9 ± 6.6	46.9 ± 8.4	48.3 ± 7.8	*
Upper limb muscle mass (kg)	5.3 ± 0.9	4.8 ± 1.3	5.0 ± 1.1	*
Trunk muscle mass (kg)	22.2 ± 2.9	20.6 ± 4.0	21.3 ± 3.6	*, #
Lower limb muscle mass (kg)	17.0 ± 2.5	16.0 ± 3.1	16.4 ± 2.9	*
PhA (°)	6.5 ± 0.7	6.0 ± 0.9	6.2 ± 0.9	*, #

Unpaired t-test: *p<0.05. Analysis of covariance: ^#p<0.05.

Covariates: Propensity scores were calculated from age, height, and weight.

PhA: phase angle; R: regular; NR: non-regular.

Binomial logistic regression analysis was conducted for trunk muscle mass and PhA parameters that showed significant group differences in the ANCOVA. This analysis identified PhA as a significant predictor of regular player classification (Table 2).

Table 2. Binomial logistic regression analysis for predicting R players: relationship with trunk muscle mass and PhA.

	OR	95% CI	p-value
Trunk muscle mass (kg)	2.33	0.85–1.13
PhA (°)	0.98	1.28–4.23	*

*p<0.05. Omnibus Test of Model Coefficients: χ² value was significant (p=0.000).

R: regular; OR: odds ratio; 95% CI: 95% confidence interval; PhA: phase angle.

ROC analysis for the identification of R players using PhA is presented in Table 3. PhA demonstrated predictive ability with an AUC of 0.685. The optimal cut-off value for distinguishing between R and NR players was 5.65° (AUC=0.685, sensitivity=86.8%, specificity=44.3%).

Table 3. Receiver operating characteristic curve analysis for R player identification by PhA.

	AUC**	95% CI	Cut-off	Sensitivity	Specificity
PhA	0.69	0.60–0.77	5.65	86.8%	44.3%

**p<0.01. R: regular; AUC: area under the curve; 95% CI: 95 confidence interval; PhA: phase angle.

DISCUSSION

In this study, we compared body composition components between R and NR players. Simple group comparisons revealed significant differences in lean body mass, total body muscle mass, upper limb muscle mass, trunk muscle mass, lower limb muscle mass, and PhA (Table 1). However, after adjusting for age and body size through ANCOVA, only trunk muscle mass and PhA remained significantly different between groups. These findings suggest that trunk muscle mass and PhA are key discriminators between R and NR players.

Trunk muscle mass plays a critical role in core stability by supporting the spine and pelvis and functioning as a central axis that integrates upper and lower body movement. Previous studies have demonstrated a correlation between trunk muscle thickness and 100-meter sprint performance in athletes, suggesting its contribution to explosive motor output¹⁵⁾. Moreover, a stable trunk enhances the efficiency of movements such as running, throwing, and kicking, while also reducing the risk of lumbar and lower limb injuries¹⁶⁾. The greater trunk muscle mass observed in R players in our study likely supported more effective movement mechanics, injury prevention, and overall performance.

Although binomial logistic regression analysis identified a significant group difference in trunk muscle mass following adjustment, trunk muscle mass did not emerge as a statistically significant independent predictor of regularity classification (p>0.05) (Table 2). In contrast, only PhA was significantly associated with regular player classification (p<0.05), reinforcing its utility as an effective indicator for distinguishing between R and NR players. Previous studies have also linked PhA to athletic performance levels, and our findings showing higher PhA values in R players are consistent with those prior observations¹⁷⁾.

We generated a ROC curve to evaluate the utility of PhA in identifying R players and calculated an optimal cut-off value of 5.65°. PhA increases with age and typically peaks during the third and fourth decades of life¹⁸⁾. In a previous study by the present authors, the mean PhA values for Japanese amateur soccer players were 5.6 ± 0.6° for ages 12–15 years, 6.5 ± 0.6° for 16–18 years, and 7.2 ± 0.6° for those aged 19 years and older¹⁹⁾. Given the mean ages of 17.8 ± 4.5 years in the R group and 16.8 ± 5.2 years in the NR group, the current results appear applicable to Japanese male soccer players at the junior and senior high school levels. However, variations by sex, age, race, and competition level may affect outcomes and should be explored through longitudinal analysis in future research.

Nevertheless, this study has some limitations that should be acknowledged. First, lifestyle factors such as training regimen, diet, hydration status, and sleep were not strictly controlled, and potential confounding variables such as playing position, athletic experience, and specific training content may also have influenced the results. While multi-frequency BIA devices have demonstrated robustness to minor daily fluctuations in hydration status, these variables remain potential sources of measurement variability. Future research incorporating these factors are necessary to clarify their effects. Second, the relatively low specificity of our findings indicates a limited ability to accurately classify NR players. This limitation may reduce the practical applicability of our proposed indicators in certain performance evaluation contexts.

Despite these limitations, our findings offer novel insights into the physiological characteristics of amateur soccer players and may serve as a foundation for future research aimed at supporting player development and establishing objective selection criteria in soccer.

Conflicts of interest

The authors declare no conflicts of interest.