The Combined Additive Effect of Inter-Limb Muscle Mass Asymmetries and Body Composition Indices on Lower Limb Injuries in Physically Active Young Adults

Abstract

Biological measurements that predict injury risk are crucial diagnostic tools. Yet, research on improving diagnostic accuracy in detecting accidents is insufficient. Combining multiple predictors and assessing them via ROC curves can enhance this accuracy. This study aimed to (1) evaluate the importance of lower limb muscle mass asymmetry and body composition (BMI and FMI) as predictors of injuries, (2) explore the role of the most effective body composition index in the relationship between muscle asymmetry and injury, and (3) assess the prognostic potential of combined predictors. Cross-sectional sampling was used to select students from a university. The sample included 237 physically active young adults (44% males). The independent variables were inter-limb muscle mass asymmetry (absolute asymmetry, AA), BMI, and FMI; the dependent variable was the number of injuries in the past year. Using zero-inflated Poisson regression, we examined the relationships, including a moderation analysis (moderated multiple ZIP regression). The mediation by body composition was tested using ZIP and logistic regression. The predictive power was assessed via ROC curves. The significance level was set at an α-value of 0.05. No significant difference in injury incidence between males and females was found (χ² = 2.12, p = 0.145), though the injury types varied. Males had more muscle strains, while females had more bone fractures (χ² = 6.02, p = 0.014). In males, the inter-limb asymmetry and FMI predicted injuries; in females, the BMI and FMI did, but not asymmetry. No moderating or mediating effects of body composition were found. In males, combined asymmetry and the FMI better predicted injuries (AUC = 0.686) than separate predictors (AA: AUC = 0.650, FMI: AUC = 0.458). For females, the FMI was the best predictor (AUC = 0.662). The most predictive factors for injuries in males were both muscle asymmetry and the FMI (as combined predictors), while in females, it was the single FMI. The hypothesis regarding the mediating role of body composition indicators was rejected, as no moderation or mediation by the FMI was detected in the relationship between absolute asymmetry (AA) and injuries. For clinical practice, the findings suggest that practitioners should incorporate assessments of both muscle asymmetry and body composition into routine screenings for physically active individuals. Identifying those with both high asymmetry and an elevated FMI can help target preventative interventions more effectively. Tailored strength training and conditioning programs aimed at reducing asymmetry and managing body composition may reduce the risk of injury, particularly in populations identified as high-risk.

1. Introduction

The relationship between physical activity (PA) and health, particularly regarding its effects on body composition, cardiovascular parameters, mental health, and general well-being, appears to be well-established [1,2]. The apparently unequivocally positive role of physical activity (PA) diminishes in significance when considering the injury risks associated with sporting activities. Increasing sport activities, even at the amateur level, leads to a higher incidence of injuries, as demonstrated in several relevant cohort studies [3,4,5].

The pro-healthy effects of physical activity (PA) are closely related to energy expenditure. Physical exercise significantly contributes to activity-induced energy expenditure, which is the most important component of daily energy expenditure [6]. Energy expenditure influences body composition, especially when the resulting energy imbalance is compensated for by the mobilization of body fat [7]. The effect is decreased body fat (BF) and body mass index (BMI). However, it has been proven that physical activity (PA) also alters body composition by increasing muscle mass [8]. Therefore, the BMI is recognized as a better index for muscle mass than for adiposity in active individuals [9]. People with excessive body weight and a BMI over 35 have a threefold higher risk of musculoskeletal (MSK) injuries [10]. Hence, studies and analyses have focused on examining the effect of the fat-to-height ratio, known as the fat mass index (FMI), as a better predictor of injuries in physically active individuals and athletes [11]. Several studies have shown, for example, that the resultant cut-off points for the fat mass index (FMI), derived from ROC analyses, can more effectively distinguish injured from non-injured individuals than cut-off points for the BMI [12].

The risk of musculoskeletal (MSK) injury, as well as the specific body parts exposed to damage, is related to the type or form of physical activity [13,14]. The lower limbs (LL) are particularly at risk, as they endure the most stress in many sports and activities [12]. Consequently, studies dedicated to injuries in these body parts are among the most commonly conducted, with knee and foot injuries being the most frequently studied. In Finch et al.’s study of 1512 non-professional players across various team sports, lower limb injuries were found to be twice as common as injuries to other body parts [15]. In Powell et al.’s studies on students participating in high school sports, the incidence of knee injuries was higher than that of other body parts [16]. For instance, a study by Xu et al. (2023) explored biomechanical landing patterns before and after fatigue to accurately predict ACL forces, highlighting critical factors in preventing knee injuries [17]. Similarly, another study presented risk factors and preventive strategies in adolescent knee injuries [18]. Additionally, a wide systematic review underlined the importance of targeted interventions to mitigate injury risks in these specific body parts, particularly during physical activity related to running, team games, tennis, or squash [19].

The risk of injury depends on both the individual characteristics of an athlete (intrinsic) and environmental (extrinsic) risk factors. The first group includes biological factors such as body size (height and weight), leg length, and anthropometric asymmetries [16]. Human beings are built on a plan of bilateral symmetry. However, both morphologically and functionally, the body is lateralized. One side of the body is often preferred, making it more exploited and loaded [20,21]. Moreover, unevenly distributed loads exacerbate asymmetry [22]. Despite bilateral symmetry related to the sagittal plane, morphological differences in the size of the right and left sides and the functional preference for the right or left extremity (upper or lower) establish directional asymmetry [23,24]. Regular sports practice may aggravate asymmetries, particularly in the musculature of the dominant limb [25]. A prospective study demonstrated that skeletal muscle responds even to a short, complex exercise training program, particularly in young adults [26]. Increases in inter-limb asymmetry, referring to both the imbalance and deficit between limbs as a result of training, have been previously reported [27]. Moreover, it was also associated with intralimb asymmetry (e.g., a quadricep–hamstring imbalance) [28,29]. Unilateral and bilateral training interventions have been examined to elucidate the effects of these exercise modalities on physical performance [30]. Some studies have explained the relationship between lean body mass asymmetry in the legs and injuries through differences in the force and power between the lower limbs [31]. However, studies examining the relationship between inter-limb asymmetries in muscles and non-contact injuries have shown equivocal results [32,33,34,35].

Body weight and BMI are well-known intrinsic factors. However, the nature and extent of the relationship between BMI and musculoskeletal (MSK) injuries are still unclear. The significance of the BMI is ambiguous. A previous study conducted on army soldiers who entered basic combat training showed a higher risk of injury among individuals with excessive weight, while the fittest army trainees had a lower risk of injury. Conversely, those with the lowest BMI tended to have a high risk of training-related injuries [36,37]. Some studies showed direct correlations between overweight and obesity and an increased risk of injuries [38,39]. In contrast, other studies have postulated particularly indirect associations, suggesting that physically active people with an increased BMI have a slightly higher chance of sustaining MSK injuries [40]. However, the indirect effects have not yet been verified using adequate statistical methods. There is still a lack of statistical verification of the potential role of the BMI or FMI as a mediating variable in the association between muscle asymmetry and the risk of lower limb injury

The above-mentioned studies explained the determinants of injuries mostly using univariate methods or studied direct effects. Body composition and morphology inter-limb imbalances were used as single predictors. The lack of consistent results for the relationships between BC, morphology asymmetry, and injury risk became the rationale for exploring a possible indirect effect. The procedures of the moderation and mediation analyses allow for insight into the mechanism underlying a known relationship by exploring the role of a third variable (moderator or mediator) by which (potentially) two variables interact in relationship with the third variable. There is still a scarcity of evidence regarding the risk of injuries related to muscle mass imbalance, especially when considering the role of body weight. To date, and to the best knowledge of the author, there are no studies examining the potential moderating or mediating roles of the BMI and FMI in the associations between inter-limb muscle asymmetry and injuries. Therefore, the objective of this study was threefold: (1) to evaluate the importance of asymmetry in the muscle mass of the lower limbs, as well as body composition (BMI and FMI), as predictors of lower limb injuries; (2) to investigate the potential role of the most effective body composition index identified in the associations between muscle inter-limb asymmetry and injury; and (3) to assess the prognostic potential of the combined predictors (asymmetry and body composition index) in detecting the risk of injury.

It was hypothesized that body composition could serve as a third variable in the associations between inter-limb muscle asymmetry and the risk of injury. Additionally, it was hypothesized that the predictive power for lower limb injuries of the combined markers would be greater than that of each individual marker.

2. Materials and Methods

2.1. Ethics

All subjects provided their informed consent for inclusion before participating in the study. The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the ethics committee of Wroclaw University of Health and Sport Sciences (consent number 13/2022, date of approval: 28 March 2022).

All participants were asked to provide informed consent via an online form prior to the study, and the purpose and characteristics of the research were explained to them.

2.2. Study Design

The study employed a cross-sectional design. This approach is the most common in the study of ex-post (retrospective) injury incidents. Examinations were conducted in 2023 at Wroclaw University of Health and Sport Sciences among first-year students. Participants were asked to complete an online survey related to their injury history, followed by measurements of their anthropometric parameters and body composition.

To test the mediation hypotheses, a mediation approach was conducted, which is a variety of hierarchical regressions. This approach was used to verify whether both inter-limb asymmetry and injury (outcome) correlated with the hypothesized mediating variable (body composition index) and whether controlling for the body composition index would explain part of the variance in the inter-limb asymmetry variable [41].

2.3. Sample Size

Before recruitment, a power calculation was conducted to determine the minimum sample size required for the mediation analysis approach [42]. Based on 80% total power, with an a-path and b-path (the two arms of the indirect impact through the mediator) equal to 0.8944, an α level of 0.05, and a minimum effect size of 0.05–0.10, it was calculated that 81–159 participants would be required to detect a between-group difference in the outcome values. The sample size calculation was performed using the G*Power tool v.3.1.9.7 for Windows (Sydenham Institute of Management Studies, Research and Entrepreneurship Education) [43].

2.4. Participants

The participants included 237 healthy individuals, of which 105 (44%) were males. All participants were first-year students at the Faculty of Physical Education and Sport and Physiotherapy in 2023 at Wroclaw University of Health and Sport Sciences. The proportions of males and females represented the true population in these fields of study. The flowchart (Figure 1) presents the full sampling procedure.

The preliminary inclusion criteria required students to attend classroom courses and be less than 22 years of age. The exclusion criteria included students participating in regulated sporting activities through the university and athletes enrolled in sports classes or mastery-level classes. A total of 237 students met the criteria and were accepted to participate in the examinations.

2.5. Anthropometric and Body Composition Measurements and Asymmetry Calculations

The anthropometric and body composition measurements were conducted in the Biokinetics Research Laboratory (part of the Central Research Laboratory) of Wroclaw University of Health and Sport Sciences. This facility held quality management system certificates PN-EN ISO 9001:2015 (certificate reg. no.: PW-15105-22X, validity: 27 May 2025).

Two body height measurements were taken with an accuracy of 0.1 cm using an anthropometer (GPM Anthropological Instruments, DKSH Holding Ltd. Zurich, Switzerland).

The body weight and body fat weight were measured using a body composition analyzer with the InBody230 electronic tool (InBody Co. Ltd., Cerritos, CA, USA). Using the body height, weight, and body fat mass, the body mass index (BMI) and fat mass index (FMI) were calculated using the following formulas:

$$\mathrm{BMI}=\frac{\mathrm{body}\mathrm{mass}\left[\mathrm{kg}\right]}{\mathrm{body}\mathrm{height}{[\mathrm{m}}^2]}$$

$$\mathrm{FMI}=\frac{\mathrm{body}\mathrm{fat}\mathrm{mass}\left[\mathrm{kg}\right]}{\mathrm{body}\mathrm{height}{[\mathrm{m}}^2]}$$

Asymmetry calculation

The standardized absolute asymmetry (AA) was calculated for the inter-limb muscle mass. The AA score is a measure that does not account for the directionality of the asymmetry and is calculated as follows [44]:

AA = (|R − L|)/(1/2(R + L)) × 100%

This formula standardizes the difference between the muscle masses of the right and left limbs, providing an absolute value that represents the degree of asymmetry without indicating the direction. The dominant limb side was not taken into account.

2.6. Recording of the Musculoskeletal Injuries of the Lower Limbs

The injury history questionnaire (IHQ) was used to collect the injury data. The reliability of the IHQ was previously assessed using Cronbach’s alpha analysis, with a calculated coefficient of 0.836 confirming strong reliability [45,46]. The IHQ analyzes the number of injuries in the last 12 months, specifically concerning body parts. For this study, only data related to lower limb injuries were assessed. The injury information was self-reported. Students independently completed surveys using Google Forms. The completeness of the data was verified at the end after compilation with the physical measurements. As previously mentioned (and shown in Figure 1), there were no missing data on the injury questionnaire among the 237 subjects selected for analysis.

The participants were required to complete several questionnaires, including those on physical activity, diet, sleep quality, and injury history. Missing data were noted in some of the questionnaires and were subsequently addressed through the process of missing data imputation. However, this was not the case for the injury questionnaire, as complete information was obtained from all 237 participants.

2.7. Statistics

The Kolmogorov–Smirnov (K–S test) test was used to examine the distributions of the analyzed continuous characteristics. This test is recommended for larger samples. The results showed no reason to reject the null hypothesis for any of the continuous variables: height, weight, BMI, FMI, and asymmetry index (the K–S D values ranged from 0.038 to 0.085 with p-values > 0.1). The continuous variables were presented as means, standard deviations, and 95% confidence intervals for the mean (CI). An unpaired Student’s t-test was used to assess the significance of sex differences in the continuous variables. The categorical variables (the percentages of males and females who had musculoskeletal injuries) were presented as numbers and percentages. The Chi-square test of independence was used to assess whether two categorical variables were likely to be related. Additionally, the contingency coefficient (CC) was calculated to assess the strength of the associations between the categories.

The effect of the independent variables (BMI, FMI, and AA) on the total number of injuries each participant experienced during the year prior to the examinations was tested using zero-inflated Poisson regression (ZIP). ZIP is used to model count data that have an excess of zero counts. According to the theory, excess zeros are generated by a separate process from the count values, and the excess zeros can be modeled independently [47].

Moderated multiple regression was used to assess the potential moderating role of the body mass indices in the relationship between AA and injuries. Multiple zero-inflated Poisson regressions were conducted with two independent variables (one of the body mass/composition indices and AA) and the number of injuries as the dependent variable. An interaction term was also included. A moderator is a variable that affects the direction and/or strength of the relationship between an independent variable and a dependent variable. This modifying effect of the moderator is most often expressed in terms of a significant interaction between the moderator and the independent variable [48].

To study the potential mediation role of the body mass indices, two approaches were used. Injuries, as a continuous dependent variable, were analyzed using a causal mediation analysis for the count and zero-inflated count data [49]. The average causal mediation effects (ACMEs), average direct effects (ADEs), and total effects (TEs), as the sum of ACMEs and ADEs, were calculated. Additionally, the proportion of the effect of the independent variable on the dependent variable that goes through the mediator (prop. mediated) was obtained by dividing the ACME by the total effect. The mediation analysis for injuries, treated as the occurrence of at least one accident (categorical data: yes-no), was assessed using a logistic regression approach. The original [41] proposition was used, with the MacKinnon and Cox (2012) [50] modification for the categorical variable approach. Logistic regression was employed with the procedure described by Newsom [51] web access with process macro [52,53] implemented into the R language.

Powerful predictors are important tools in diagnostic settings. Frequently, diagnostic accuracy can be improved by combining multiple variables. To study the performance of individual predictors as well as their combinations, the combiroc package was used. The area under the curve (AUC), sensitivity, specificity, cut-off points, and Akaike information criterion (AIC) were calculated for each variable and the best combinations. An optimal model approach was conducted, selecting the three or fewer best variants.

The significance level for all statistical tests and procedures was set at an α-value of 0.05. The calculations were conducted using RStudio with the following additional R packages: car [54], pscl and boot [55,56], maczic [57], and combiroc [58] using Statistica 13.5 (StatSoft Poland 2023, Cracow, Poland).

3. Results

Table 1. Overview of participants’ anthropometrical measurements and body mass indices (BMI and FMI). Unpaired Student’s t test t-values and p-values.

			Males				Females
	M	SD	−95%CI	+95%CI	M	SD	−95%CI	+95%CI	t	p
Height [cm]	183.3	7.0	181.9	184.6	168.7	5.2	167.8	169.5	−18.42	<0.001
Weight [kg]	79.0	9.6	77.2	80.9	61.3	8.7	59.8	62.8	−14.84	<0.001
BMI [kg/m2]	23.5	2.4	23.1	24.0	21.6	2.8	21.1	22.0	−5.71	<0.001
FMI [kg/m2]	3.7	1.4	3.5	4.0	5.3	1.8	5.0	5.6	7.52	<0.001
AA [%]	3.5	1.4	3.2	3.7	2.4	1.3	2.1	2.6	−6.29	<0.001

Footnote: BMI: body mass index, FMI: fat mass index, AA: absolute asymmetry.

presents the descriptive statistics of the anthropometric measurements, body mass index (BMI), fat mass index (FMI), and absolute asymmetry index (%) in males and females. The results showed statistically significant differences in all the measurements (p < 0.001). Males were taller, heavier, and had a higher BMI. However, they were leaner, as evidenced by the significantly lower FMI values. Additionally, males had, on average, greater asymmetry of muscle mass in the lower limbs.

Out of 105 males, 61 (58.10%) experienced musculoskeletal injuries of the lower limbs, while out of 132 females, 63 (47.73%) experienced at least one injury during the year before the questionnaire. However, the Chi-squared test results showed no significant differences in the injury incidence between males and females (χ²Yates = 2.12, p = 0.145, CC = 0.10, OR = 1.52 [0.91–2.55, CI95%]).

Among all the recorded lower limb injuries (including multiple and recurrent), the most common injuries in males were muscle strains (58.0%), followed by joint sprains (38.4%) and bone fractures (3.6%). In females, muscle strains were also the most common injuries, though less frequent than in males (45.5%). Joint sprains were similarly frequent in females (40.7%), but females experienced more bone fractures (13.8%). These different frequencies of injury types between the sexes were statistically significant (χ²Yates = 6.02, p = 0.014, CC = 0.15).

The next preliminary step for further analyses was to assess the effect of body mass indices on injuries. Injuries were expressed as the total number of incidents for each participant. Therefore, three simple zero-inflated Poisson regressions were conducted, with BMI, FMI, and AA as the independent variables and the number of injuries as the dependent variable. The results are presented in

Table 2. Estimated parameters (with the standard error, z-value, and probability) for the zero-inflated Poisson model (ZIP), with the BMI, FMI, and AA as predictors of the number of injuries.

		Males				Females
	Estimate	Std Error	z-Value	Pr(>|z|)	Estimate	Std Error	z-Value	Pr(>|z|)
BMI	0.05	0.04	1.31	0.190	0.13	0.03	4.42	<0.001
FMI	0.35	0.05	7.11	<0.001	0.21	0.05	4.33	<0.001
AA	0.28	0.06	4.43	<0.001	0.12	0.08	1.40	0.161

Footnote: BMI: body mass index, FMI: fat mass index, AA: absolute asymmetry.

. Although no differences were found between men and women in injury frequency, significant differences in body mass indices and asymmetry indexes justified the separation of sexes in subsequent analyses.

The models for males contained two significant parameters: FMI (b = 0.35, Pr(>|z|) < 0.001) and AA (b = 0.28, Pr(>|z|) < 0.001), while BMI did not significantly affect the number of injuries (b = 0.05, Pr(>|z|) = 0.190). Similarly, the models for females identified two significant determinants. However, in contrast to males, these were body mass indices (BMI: b = 0.13, Pr(>|z|) < 0.001; FMI: b = 0.21, Pr(>|z|) < 0.001), but not AA (b = 0.12, Pr(>|z|) = 0.161).

From the zero-inflated models, it can be derived that the estimated odds of observing an excess zero with higher FMI is exp(0.35) = 1.42, meaning that males with greater fat mass were 1.42 times more likely to be injured than lean males. Similarly, those with higher asymmetry were 32% more likely to be injured than those with less asymmetry (exp(0.28) = 1.32). For females, those with higher BMI and FMI were 13% and 23% more likely to be injured, respectively.

All three models were compared, assessing the model fitting. Log-likelihood statistics were used. The higher the value of the statistic, the better the fitted model. The first model for males had the lowest log-likelihood value (−181.8), indicating the poorest fit, while the second model was the best fitted (−159.4), with the third model in between (−168.8). For females, the first and second models had similar fits (−163.5 and −163.2, respectively) and offered a better fit to the data than the third model (−175.4).

Given the lack of a significant effect of AA on injury in females and the multicollinearity of BMI and FMI, further analyses for females were not conducted.

Examining the potential moderating role of the FMI in the relationship between AA and injuries began with an analysis of moderation. Both continuous variables (FMI and AA) were included in a multiple ZIP regression, along with an interaction term. The results showed significant effects of the FMI (Pr(>|z|) < 0.001) and AA (Pr(>|z|) = 0.003), but no significant interaction effect (Pr(>|z|) = 0.205). The moderator is a factor that impacts the relationship between the independent variable (in this case, AA) and the independent variable (in this case, injuries). A moderation analysis is usually performed using multiple regression models; in this data, a significant interaction confirmed the moderation effects of the third variable (in this case, FMI). Thus, the FMI did not act as a moderator.

The rationale for introducing a mediating variable is generally to understand the mechanism through which an independent variable affects a dependent variable. Hence, the next analysis evaluated the potential mediating role of the FMI in the associations between AA and injuries. First, a causal mediation analysis for the count and zero-inflated count data was conducted. The total effect, represented by the b-coefficient for the independent variable (AA) without the mediator, was very small (0.03) and not significant (p = 0.140) (

Table 3. Mediation effects: total, direct, and indirect (in the model) with the % of mediation.

Effect	Estimate	95%CI Lower	95%CI Upper	p
Total effect	0.03	−0.19	0.32	0.142
ADE	0.04	−0.15	0.32	0.134
ACME	−0.01	−0.07	0.01	0.303

Footnote: ADE: average direct effect, ACME: average causal mediation effect.

). The direct effect (ADE, 0.04), after accounting for the mediation effect, was also small and not significant (p = 0.130). Finally, the mediation effect (ACME, −0.01) was not significant (p = 0.300). These results confirmed the lack of a mediation effect of the FMI on the number of injuries. The goal of the mediation analysis was not obtained because the indirect effect was not statistically significant.

Secondly, a logistic regression mediation model was conducted to assess the performance of the FMI as a potential mediator in the relationship between AA and the occurrence of at least one injury. In the logistic regression model, both asymmetry (AA, b = 0.62, p = 0.018) and the fat mass index (FMI, b = 0.41, p = 0.037) were significantly but independently associated with injuries. Contrary to the continuous analysis, the bootstrap confidence intervals derived from 5000 samples indicated that the indirect effect coefficient was not significant (b = −0.06, CI = −0.206 to 0.046) (

Table 4. Mediation effects: total, direct, and indirect (in the model) with the % of mediation.

Effect	Estimate	95%CI Lower	95%CI Upper	p
ADE	0.62	0.23	1.01	0.018
ACME	−0.06	−0.21	0.05	>0.05

Footnote: ADE: average direct effect, ACME: average causal mediation effect.

). These results showed a non-significant indirect effect, as the bootstrap confidence interval included zero. Therefore, as before, these results did not support the hypothesis that the relationship between asymmetry and injury is mediated by the fat mass index in males. Both factors independently and significantly affect the occurrence of injury.

The results suggested rather additive, common effects of the FMI and AA on the probability of injury. Thus, the final analysis assessed the performance of each independent variable separately and in combination to detect the risk of injury. The results are presented in

Table 5. Areas under the curve (AUC) and respective Youden indices, sensitivity, specificity, cut-off points, and AIC for each model across the best predictors of injury in an optimal set.

Variable	AUC	Youden	SE	SP	Cut-off	AIC
Combination FMI AA	0.686	0.115	0.787	0.523	0.525	134.9
FMI	0.458	0.082	0.230	0.818	0.536	145.2
AA	0.650	0.083	0492	0.818	0.650	137.2

Footnote: FMI: fat mass index, AA: absolute asymmetry, AUC: area under the curve, SE: sensitivity, SP: specificity, AIC: Akaike information criterion.

and Figure 2. In males, the optimal set of variables predicting injury included the FMI, AA, and their combination, excluding the BMI. The highest AUC was observed for the combination (AUC = 0.686), confirming its superior predictive ability for injuries. The least predictive performance was observed for the FMI alone (AUC = 0.458), while the predictive power of AA alone was greater (AUC = 0.650). An illustration of the ROC curves for all the variables is presented in Figure 2. The model fit was confirmed with the AIC values, with the highest for the combination (134.9), followed by AA (137.2) and FMI (145.2). In females, the optimal set included only the FMI with an AUC of 0.662.

4. Discussion

The aim of the study was threefold: (1) to evaluate the importance of asymmetry in muscle mass of the lower limb, as well as body composition (BMI and FMI), as predictors of lower limb injuries; (2) to investigate the potential mediating role of the most effective body composition index identified in the association between muscle inter-limb asymmetry and injury; and (3) to assess the prognostic potential of combined predictors (asymmetry and body composition index) in detecting the risk of injury. Two working hypotheses were put forward: (A) body composition has the potential to act as a mediating variable in the association between inter-limb muscle asymmetry and the risk of injury, and (B) the predictive power for lower limb injuries of combined markers is greater than that of each individual marker. There was no significant difference in the frequency of at least one injury between men and women. However, a significant difference in injury types was found. Although muscle injuries were most common in both sexes, women experienced bone fractures more frequently than men. Differential determinants of injury were identified with the studied variables. In men, lower limb muscle mass asymmetry and FMI were significantly associated with the number of injuries, whereas in women, both body composition indices but not lower limb asymmetry were significantly associated. The calculated odds ratios showed that individuals with an increased body mass had a higher risk of injury: 42% in men and 13–23% in women (considering the BMI and FMI, respectively), and men with lower limb muscle asymmetry had a 32% higher risk of injury. There was no indirect effect (moderating or mediating) of any body composition indicators in the relationship between lower limb asymmetry and injury. Further studies on the inter-relationship of body composition and asymmetry in relation to injury showed a common additive effect of combined asymmetry and fat mass index in men but an independent effect of FMI in women.

The problem addressed in this paper references previous studies on the predictors of injury in body composition and lower limb morphological asymmetry. The discussion began with a general comparison of our current findings with previous studies to better contextualize our contribution. Chassé et al. (2014) found that higher BMI values increased the injury risk among women but not men [59]. The same results were observed by Hollander et al. (2020) [22]. Our results are partially in agreement with these studies. Our findings showed a stronger prediction of injury performance with the BMI in females while this was true for the FMI in males. However, the BMI may be an insufficient indicator for predicting injuries in women, but it is an accurate predictor for men. Additionally, in line with the results of Haventidis et al. (2017), who identified the FMI as a significant injury predictor in male recruits, our study confirms that the FMI is a strong predictor of injury risk in both men and women [12]. A unique contribution of this study is the analysis of the interaction between the BMI and morphological asymmetry. Our study extends previous findings by demonstrating the power of the combination of asymmetry and the FMI, indicating that this combination of two variables more effectively predicts injury risk, although only in men. By integrating these insights, we emphasize the unique contribution of our study to understanding the complex interplay between body composition, muscle asymmetry, and injury risk.

In detail, the results of this study showed that 58% of males and 48% of females experienced musculoskeletal injuries of the lower limbs. These findings are consistent with general observations regarding the prevalence of injuries in young adult athletes. Previous studies of physically active high school students have similarly demonstrated high annual injury rates, ranging from 41% to 61% [60,61]. Most of the injuries were located in the lower limbs [62]. These frequencies remain consistent today, as confirmed by many current studies [63]. This study noted no significant sex differences in the frequency of injuries. This contrasts with other studies where the prevalence rates of physical activity-related injuries (PARIs) varied between male and female students. The cumulative injury frequency per total number of people surveyed per year was 0.44 for males and 0.18 for females (p < 0.001) [64,65,66]. The explanation is related to the fact that males tend to be more actively involved in physical activity (PA) than females [67]. However, in this study, the participants were students at a university with a focus on physical activity. This could affect the results, as both males and females studying sports and physical education are generally highly active, as demonstrated in previous studies [68]. Moderate and high levels of physical activity (PA) were observed among the students, with no significant differences between the males and females in terms of overall MET minutes per week.

On the other hand, the observed sex differences in the types of injuries align with the findings from other studies. Bone stress injuries are more common in female athletes, while male athletes are more likely to experience soft tissue injuries, such as muscle strains and ACL injuries [69]. The explanation appears to be that female athletes generally have a lower bone mineral density (BMD) compared to male athletes [70,71]. However, some studies reported only partially similar results. For instance, although males were at increased risk for posterior thigh injuries compared to females (relative risk (RR) 5.8, p < 0.05), females were at increased risk for overuse injuries in the ankle compared to males (RR 3.1, p < 0.05) [72]. In this study, the frequencies of joint sprains were similar in males and females, while males more often experienced muscle strains. Hamstring strains are common injuries among physically active male amateurs and athletes. The primary causes are maximal sprinting, kicking, and sudden acceleration [73,74]. Thigh injuries are often explained by muscle fatigue, insufficient warm-up, and hamstring tightness [68]. The greater overall risk for ankle and knee injuries in females compared to males is explained by the higher joint laxity in females [75].

Males had significantly greater asymmetry than females, which is consistent with the findings from other studies [76]. This asymmetry may explain its effect on injuries in males but not in females. Research results in this regard are inconclusive, so these findings partially align with some studies [77,78] while not in others. Some authors have suggested that inter-limb asymmetry is frequently greater in females compared to male athletes concerning strength, coordination, and postural control [79,80]. This observation showed that the FMI, an index closely associated with body fat, is a predictor of injuries in both males and females. Additionally, in females, the BMI also played a role (albeit weaker) in predicting injuries. There is an insufficient number of studies evaluating body composition indices, particularly the FMI, in healthy young adult populations; thus, the comparison of the results related to the predictive power of this index for injuries is very limited. Similar results were observed by Havenetidis et al. (2017); however, their study was conducted on male recruits [12]. More studies have presented the relationship between BMI and injuries, primarily comparing individuals across different ranges of this index. For example, one study’s results showed a 34% increased risk of injury in obese adolescents compared to healthy adolescents (odds ratio (OR) = 1.34, 95% CI: 1.02–1.80) [81]. This study also demonstrated a higher risk of injury in individuals with excess body weight, ranging from 13% to 30%, depending on gender. Other studies have reported similar results [82].

The mechanistic explanations for the observed relationships between muscle mass asymmetry, body composition indices, and injury risk might be as follows. One potential biomechanical mechanism is that muscle mass asymmetry can lead to the uneven distribution of forces across the lower limbs during physical activity, which may increase the strain on specific muscles and joints, thereby elevating the risk of injury. Additionally, individuals with a higher FMI may experience greater mechanical load on their musculoskeletal system, leading to faster fatigue and reduced stability, which could also contribute to a higher injury risk. Physiologically, excess body fat might affect muscle function and coordination, further exacerbating the risk of injuries. These explanations suggest that both structural imbalances and excess body weight can compromise the body’s ability to efficiently manage physical stress, thereby increasing the likelihood of lower limb injuries. Integrating these mechanistic insights into our findings provides a clearer understanding of how these factors interact to influence injury risk. However, there is a need to explore various body composition indices in injury risk prediction.

The same can be postulated regarding the study of the indirect role of the so-called third variable in the relationship between the two others: the dependent and the independent variable. Indeed, no work to date has analyzed the potential moderating or mediating role of body composition in the relationship between lower limb morphological asymmetry and injury. This paper addresses this issue for the first time. Therefore, it is not possible to directly compare our results with those of other authors. Results regarding the relationship between lower limb muscle mass asymmetry or body mass composition indices and injury are inconclusive [35,37,83,84]. Some studies have indicated that, for both sexes, the association of BMI with injury risk is bimodal, with the lowest risk in the group with an average BMI (middle quintile). Others suggest that the BMI acts unidirectionally, with risk increasing as the BMI or other indicators increase. However, there are emerging results indicating no association between body composition indices and injury. The study group composition may play a significant role in these findings, as participants included military recruits, athletes, physically active individuals, and even clinical patients. Perhaps this is a consequence of the choice of static methods that are inadequate or whose effectiveness is limited due to the issue of multiple zeros in the injury survey results. Classical methods may not be sufficiently sensitive to capture correlations. Hence, there are emerging calls for more sophisticated methods, such as multivariate modeling like Cox proportional hazards models, nested frailty models, and generalized estimating equations (Poisson and logistic) to analyze asymmetry injury data [85]. In this paper, another approach was applied: a variant of Poisson regression analysis designed for data with a large number of zeros, known as zero-inflated Poisson (ZIP) regression. This method was used in both simple correlation analyses and moderation and mediation analyses. Using this advanced approach, the role of the body composition indices as a third variable in the relationship between asymmetry and injury was excluded. This approach is in line with Helme et al. (2021), who derived such a postulate after analyzing a large number of works with very different methodologies in a systematic review [32]. Mediation analyses help to identify the mechanisms that determine which additional variable (mediator) does or does not affect the independent dependent variable. Similarly, the moderator is a variable that determines the context (under which circumstances or for which types of people) in which an effect exists or does not and in what magnitude [52]. It is only through this complex approach that we could gain an in-depth insight into the problem of the relationship between the many inter-related determinants of injury. Such an approach informs the understanding of whether the associations between asymmetry and injury observed were the result of the asymmetry itself or the result of other risk factors, thereby avoiding spurious findings [32].

Rejecting the role of the third variable, the possible cumulative effect of both independent variables on injuries was investigated. In women, this approach did not yield significant results, while in men, it was shown that the additive joint effect of asymmetry and fatigue better predicted injuries than both factors separately. Consequently, the results cannot be directly compared with other studies. All the analyses of the performance of different markers in detecting injuries known to the author were analyzed in relation to a single variable. The performance of the BMI or general skeletal muscle mass (SMI) in predicting injury was mostly at an acceptable and statistically significant level and demonstrated itself as a reliable tool in predicting injury prevalence for young, physically active men [12]. However, there were differences between men and women in the performance of individual indicators. Chassé, Fergusson, and Chen (2014) indicated an increased risk of injury with higher BMI values among women but not among men [59]. The same was observed previously by Hollander et al. (2020) [22]. This study showed that the BMI may be inadequate for detecting injury in males, and a more appropriate measure could be the FMI, but not in females. A reasonable explanation may be related to the ratio of muscle to body fat. On the contrary, Richmond, Kang, and Emery (2013) confirmed that maintaining an adequate weight-to-height ratio is associated with a lower risk of injury [85]. Other studies have identified the performance power of free fat mass rather than body fat indices. However, FFMI, in the relevant research, effectively distinguished injured and non-injured males from females [12]. In this case, men had higher visibility. The possible mechanisms related to musculoskeletal injuries could be low muscular strength, low muscular endurance, or low neuromuscular coordination [86,87]. New light on the possibility of injury prediction based on morphological or functional variables is provided by the marker combining method and multiple ROC analysis. This study confirmed, although only for men, the higher injury detection performance of the combined variable that includes both asymmetry and the body fat index.

Despite the usefulness of these findings, the present study has some limitations which must be acknowledged. Firstly, the effect of muscle mass asymmetries and body composition indices on physical activity-related injuries should be studied according to the different forms of activity. In this study, due to the relatively low frequencies of respondents in individual disciplines, the analyses were carried out without such a breakdown. Secondly, while the selected group was age-coherent, which is a strength, further research should also be conducted with a focus on greater age diversity, considering age as a factor in intrinsic injury risk. Thirdly, the muscle mass to inter-limb asymmetries analysis was conducted using the impedance method, which is not the gold standard. Further studies should also include data retrieved from DEXA. In addition, future research could study asymmetries and body composition indices from a longitudinal perspective. Both morphological asymmetry and body composition are not stable but change over time. Relationships between changes and injury risk would be of interest. Finally, this study did not take into account variables such as the use of warm-ups or the circumstances of accidents. Additionally, other body parts apart from the lower limbs were not analyzed. Future research should consider including different age groups and types of physical activity to broaden the applicability of the study. Moreover, researchers should consider including variables such as the use of warm-ups, training intensity, and the circumstances of injuries. These factors could provide a more comprehensive understanding of injury risks. Additionally, including data on other body parts apart from the lower limbs could offer a broader perspective on musculoskeletal injuries.

5. Conclusions

The results of the present study provide evidence that morphological (muscular) asymmetry of the lower limbs is a risk factor for injuries in men but not in women. Additionally, in both sexes, the fat component of body mass, expressed as a proportion of body height (FMI), is also a significant risk factor. The hypothesis regarding the mediating role of body composition indicators was rejected, as no moderation or mediation by the FMI was detected in the relationship between absolute asymmetry (AA) and injuries. However, the hypothesis that the predictive power for lower limb injuries of the combined markers is greater than that of each individual marker was confirmed, but only in men. The effect of both predictors was cumulative and additive, confirming the higher performance of the combined markers in the detection of lower limb injuries. Individuals leading an active lifestyle, those initiating physical activity, and practitioners involved in the promotion of physical activity should consider the increased risk of injuries associated with muscle mass asymmetry and excessive fat. Further research should focus on identifying the safest forms of activity for individuals who, paradoxically, need to be physically active. Considering the necessity for individuals with an increased body mass to lead an active lifestyle, it is recommended to introduce additional forms of strength exercises aimed at compensating and equalizing the differences between both lower limbs.