**Inter-Segmental Coordination during a Unilateral 180**◦ **Jump in Elite Rugby Players: Implications for Prospective Identification of Injuries**


Received: 6 December 2019; Accepted: 27 December 2019; Published: 7 January 2020

**Abstract:** Musculoskeletal injuries often occur during the execution of dynamic sporting tasks that involve rotation. The prescription of appropriate prevention strategies of musculoskeletal injury relies on assessments to identify risk, but current assessment tools focus on uniplanar movements. The purpose of this paper is to demonstrate the utility of the unilateral 180◦ jump as a potential assessment tool for injury risk in the lower body by (1) providing descriptive kinematics of the knee, thigh, and pelvis (2) conducting inter-segmental coordination analysis, and (3) comparing the knee kinematics between the dominant and non-dominant limb (NDL) during the loading (LOP) and landing phase (LAP). Elite rugby players completed one session, performing five 180◦ unilateral jumps on each limb while collecting kinematic data. Independent *t*-tests were used to compare peak angles of DL and NDL. Continuous Relative Phase (CRP) plots were constructed for thorax and pelvis in the transverse plane. At the loading phase, the non-dominant limb had greater peak knee abduction (ABD) (*p* = 0.01). At the landing phase, the dominant limb had greater peak knee adduction (ADD) (*p* = 0.05). At the landing phase, the non-dominant limb had greater peak knee ABD (*p* = 0.01). CRP plots indicate participants can utilize a thorax-led, pelvis-led, or synchronized rotational method. Bilateral asymmetries were observed, indicated by significant differences in the bilateral landing phase peak ADD/ABD, which is of particular interest considering all participants were healthy. Therefore, additional research is needed to determine thresholds for injury risk during rotational tasks.

**Keywords:** functional; movement; evaluation; assessment; screen

#### **1. Introduction**

Athletes in sports that involve jumping, pivoting, forceful change of direction, and contact are at a higher risk of lower-body injury [1–5]. Non-contact anterior cruciate ligament (ACL) tears are one of the most common lower-body injuries and have been largely attributed to the torsional forces associated with cutting and rotational tasks [4–6]. Rugby incorporates all of the aforementioned characteristics and it has been reported that lower limb joint/ligament injuries are the most common location and type of injury [7] occurring in both contact and non-contact situations [8]. Knee injuries result in players missing the most days from training, with ACL injuries accounting for the greatest proportion [9]. Clinicians and practitioners attempt to pre-emptively identify modifiable risk factors and address through strength, flexibility, and neuromuscular training. The identification of risk factors requires the utilization of a physical assessment tool with valid and reliable prognostic value for sporting tasks.

Pre-participation screenings are designed to determine potential intrinsic injury risk factors by identifying characteristics of the musculoskeletal system that may predispose an athlete to injury or identify incomplete recovery from a previous injury [10]. Traditionally, screening methods require a battery of tests, including static and dynamic protocols [10–16]. Static tests focus on measurements of joint range of motion, muscle strength, and muscle flexibility. The limitations of static tests are their applicability to dynamic situations, yielding limited meaningful information in the context of a dynamic sporting environment [17]. Previous research has demonstrated that scores obtained during static balance tests do not reflect scores obtained during dynamic balance tests, further indicating their limitations as injury predictors in a sporting context [16,18]. The recognition of these limitations led to the emergence of more functionally relevant or task-specific screening tests; where the functional aspect refers to the adoption and inclusion of assessments that more closely replicate activities of daily living and dynamic sports movement [19].

The purpose of functional and dynamic testing is to evaluate movements similar to sports actions that require the muscles to co-activate in integrated patterns to control a multi-joint movement [20]. A failure of the muscles to activate in a coordinated manner in the control of a movement often results in the development of a compensatory but predictively variable strategy to complete the task, which, after a period of time, can lead to injury and pain around the affected joint [11,17,21]. Functional testing is intended to enable the tester to identify 'weak links', suggesting an uncontrolled movement system (joint), within the series of linked joints [17]. Identification of these movement dysfunctions may serve as a predictor of injury and can be addressed by the appropriate professionals prior to the emergence of symptoms within a sporting context. Specifically, Mottram and Comerford [17] stated that the ability to identify uncontrolled movement strategies is imperative to enable risk management strategies to be developed, which can mitigate the propensity for (re)injury through increased localized strength and/or enhanced motor control.

Several mechanical dysfunctions have been associated with functional/dynamic assessments as injury predictors. Multiple studies have suggested that uncontrolled knee motion in the frontal plane when landing is a good predictor of ACL injury [22–24]. Additionally, compensatory strategies have been demonstrated in jump landings following injury [25,26]. The neuromuscular control associated with this form of dysfunctional, maladaptive movement pattern has been suggested as a potentially modifiable risk factor [27]. The utilization of unilateral tasks further enhances the assessment through the identification of limb asymmetry when compared to bilateral assessments [15,16,18]. However, research regarding cutting tasks (change of direction) suggests that large frontal plane knee excursions are necessary in order to complete the task and may not be indicative of uncontrolled motion [28–31]. Therefore, drop jumping landing tasks may not be appropriate for assessing injury risk during rotational sporting tasks. Designing an appropriate task to identify the predictors of rotary injury risk during dynamic sports-specific activity must include several components: high-velocity loading force, a certain degree of motor control complexity to complete the task, bilateral assessment, and multi-planar movement. A novel unilateral 180◦ jump task could potentially fulfill all the aforementioned requirements and better mimics the dynamic multi-planar nature of sports that involve jumping, pivoting, and forceful changes in direction (i.e., rugby, soccer, basketball, American football). Additionally, as athletes performing rotational movements are the target population for this assessment, using elite-level athletes, such as rugby players, will provide greater external validity of findings.

Considering the unilateral 180◦ jump as a potential physical assessment tool for rotational injury prediction in the lower body first requires a demonstration of its utility. Therefore, the purpose of this investigation is to analyze the unilateral 180◦ jump in elite-level rugby players by (1) providing descriptive kinematic of the knee, thigh, and pelvis, (2) conducting inter-segmental coordination analysis between the thorax and pelvis, and (3) comparing the knee kinematics between the dominant and

non-dominant limb during the loading and landing phase. It is hypothesized that the greater rotational component will demonstrate greater frontal plane kinematics (greater peak abduction/adduction angles) of the knee during landing to compensate for the high torsional energy transfer through the kinetic chain. Similarly, greater peak frontal plane knee kinematics will be associated with a shoulder led transverse thorax–pelvis relative phase angle. More simply, we believe that greater knee adduction angles will be observed in individuals who utilize a shoulder led rotation pattern.

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

#### *2.1. Participants*

Fourteen male, elite rugby players from a National Union Academy (age: 20.0 ± 1.9 years; height: 184.5 ± 7.2 cm; mass 94.7 ± 12.0 kg) volunteered their participation. The inclusion criteria required that all participants were free of any current injury and had no history of serious lower limb injuries (injuries requiring surgery). To ascertain technique limb dominance, participants were asked which leg they preferred to kick a ball with. All participants indicated that their preferred kicking leg was their right leg. As a result, the right leg will be referred to as the dominant leg and the left as the non-dominant leg. All participants were 18 years of age or older. The study adhered to the guidelines set by the Declaration of Helsinki, ethical approval was obtained from the local University Ethics Committee (Edinburgh Napier University) prior to the investigation, and all participants provided written informed consent.

#### *2.2. Procedures*

Participants were requested to refrain from strenuous activity at least 24 h before attending the testing session. All participants were given verbal instructions indicating how to complete the jumping task. These included to initiate a unilateral stance, perform a countermovement action followed by a 180◦ jump as high as possible (maximal effort) landing on the same leg. Participants were asked to jump in a clockwise direction from their right leg and counterclockwise from their left leg. Participants were instructed not to use any arm movements during the jumps by placing "their hands on their hips". Following a demonstration of the jump, all participants practiced during an individualized warm-up to familiarize themselves with the protocol. Subsequent to the warm-up, each participant rested for 3 min before performing five single-leg 180◦ jumps either with their dominant or non-dominant leg (with a balanced pause between each jump), followed by five additional jumps with the other leg. The order of the starting jump leg was randomized to avoid learning effects.

#### *2.3. Motion Analysis*

A three-dimensional analysis was carried out using a 12-camera high speed (240 Hz) motion capture system (ProReflex, Qualisys AB., Gothenburg, Sweden). All participants wore fitted clothing and were barefoot to permit the accurate attachment of 25 retro-reflective markers (19 mm diameter) on the following anatomical landmarks (on both left and right sides): head of the first and fifth metatarsal bones, lateral and medial malleolus, posterior calcaneus, lateral and medial femoral epicondyles, anterior superior iliac spine (ASIS), posterior superior iliac spines (PSIS), acromion process, 4th lumbar vertebra, 10th thoracic vertebra, 7th cervical vertebra, sternum jugular notch, and xiphoid process. To reduce skin movement artifact error, clusters of 4 retro-reflective markers fixed to lightweight rigid plastic plates (Qualisys AB, Gothenburg, Sweden) were attached to both thighs and shanks to track the leg movement during the unilateral 180◦ jump, and a cluster of 4 markers placed on the skin around T7 was used to track thorax motion. Following the marker placement procedure, a static calibration file in the anatomical position was collected for each participant for the purpose of 3D model building and data generation.

#### *2.4. Data Analysis*

Marker trajectories were smoothed with a 6 Hz fourth-order low-pass Butterworth filter and kinematic data processed using Visual 3DTM software (C-Motion Inc., Rockville, MD, USA). A power spectrum density calculation was conducted and 99% of the power was contained in the first 6 Hz, and, therefore, a 6 Hz cutoff frequency was chosen. All kinematic variables were quantified within the loading and landing phases of the unilateral 180◦ jump. The loading phase was defined as the instant the knee commenced flexion (following a stationary single-leg stance) until the instant of maximum knee flexion of the same leg. The landing phase was defined as the instant the landing foot made contact with the ground (determined from kinematic data when the metatarsal markers stopped their downward trajectory) until maximum knee flexion was attained.

Individual segment pose and segment coordinate systems (SCS) for each participant was generated with Visual 3DTM from the static calibration trial using a multi-anatomical landmark optimized method [32] and utilizing the markers as follows: The pelvis was defined using the X-Y plane passing through the ASIS and PSIS, the origin of the SCS at the midpoint between the 2 ASIS markers, which allowed the segment X axis to be determined by the vector between the origin and the right ASIS marker. The SCS Z axis was determined by the axis perpendicular to the X-Y plane in the vertical direction, and, lastly, the SCS Y axis calculated as the cross product of the X and Z axes. All remaining segments were defined using the principle of creating a frontal plane using medial and lateral proximal and distal markers and then determining the SCS Z (vertical) axis as the unit vector directed from the distal segment end to the proximal segment end. The SCS Y axis was then perpendicular to the frontal plane and Z axis, and, finally, the X axis determined by the application of the right-hand rule. The origin of each SCS was located at the proximal end for each segment.

Knee joint angles were calculated according to the International Society of Biomechanics recommendations as the shank segment relative to the thigh segment resolved into the proximal segment SCS, using the Cardan sequence XYZ, where movement in the X plane denotes flexion (−)/extension (+), Y plane abduction (−)/adduction (+), and Z plane axial internal (+)/external (−) rotation [33]. Knee alignment in the Y plane was defined as zero when the long axes of the thigh and shank were aligned. Axial rotation of the thigh segment (Z plane) was calculated relative to the pelvis segment [33].

For both adduction/abduction and axial rotation, the non-dominant leg was multiplied by negative one to allow comparison between limbs. Pelvis motion was assessed in terms of flexion/extension, adduction/abduction, and axial rotation, which was computed as the angle of the pelvic segment relative to the fixed laboratory/global coordinate system (GCS) using Cardan sequence XYZ and the right-hand thumb rule [33]. Given the dynamic nature of the activity and magnitude of axial rotation involved in this activity, the sequence XYZ was selected instead of ZYX as described by Baker [34]. All variables were averaged across the five trials per leg for each participant; these means were then used to calculate group means.

The continuous relative phase (CRP) was calculated as a representative measure of inter-segmental coordination between the thorax and pelvis segmental rotations about the vertical (z) axis. To calculate the CRP between two segments, the phase-angle (Φ) from the phase-plane portraits of each segment were calculated. Phase-portraits were constructed with angular displacement (θ) on the horizontal (*x*) axis and the first derivative, angular velocity (ω) on the vertical (*y*) axis. Prior to calculating Φ, the phase-portrait values were normalized to the minimum and maximum values found in each axis using the protocol outlined by Li et al., [35], and to 101 data points. The normalization procedures minimize the influence of different segmental movement amplitudes [35] and allow comparisons of jumps with different temporal structures. The Φ was defined as the angle between the right horizontal and a line drawn to a specific data point (θ*i*,ω*i*) from the origin (0,0). The CRP was calculated as the difference between the thorax and pelvis segment angles at each of the 101 data points. Ensemble curves were produced for the individual CRP profiles by determining the mean CRP value at each of the 101 data points from the 5 jump trials. The variability in the CRP was displayed as the standard

deviation of the 5 trials at each data point. The CRP provides a means to interpret both the coordination between the relative segments and its variability. This information can give insight into the relative stability (change in variability) of a pattern of movement over time, helping to identify which, if any, coordination patterns are important and common across multiple individuals. Measuring the relative phase between limb/segment movements (oscillations) has been regularly employed to examine the organization of a system at a synergistic level, as phase differences reflect the fundamental cooperation and competition evident within a movement system. The tendency of the synchronization of the reversal points of frequency-locked coupled oscillators is to adopt either an in-phase (0◦) or anti-phase (180◦) relationship with the movements initiating and/or terminating simultaneously [36]. According to Swinnen et al., [37] synchronization of the reversal points can be interpreted as intermittent loci of control, where reversal points act as anchors for the organization of the system. In contrast, asynchronous phase differences (i.e., 90◦, 270◦, etc.) are more difficult to maintain, requiring effort and considerable practice [38].

#### *2.5. Statistical Analysis*

All statistical analyses were performed using the Statistical Package for Social Sciences 14.0 (SPSS Inc., Chicago, IL, USA, 2004). The Kolmogorov–Smirnov test was used to determine the normality of the data distribution for each variable. Measures of central tendency and distribution of the data were reported as means and sample standard deviations. Paired sample *t*-tests were used to determine if any statistically significant differences existed between the mean values of the dominant and non-dominant legs for peak knee angle (flexion/extension and adduction/abduction), time to reach peak knee angle, axial rotation of the femur with respect to the pelvis and axial rotation of the pelvis with respect to the GCS in both the loading and landing phases following a 180◦ unilateral jump. Significance was accepted at *p* ≤ 0.05 for all statistical tests. To measure the magnitude of the difference between the dominant and non-dominant legs relative to the variability, effect size (d) calculations were performed across all variables. Interpretation of the data was based on Cohen's (1992) guidelines, whereby effect sizes greater than 0.2 and less than 0.5 are considered small, greater than 0.5 and less than 0.8 are moderate, and greater than 0.8 are large. The intra-class correlation coefficient (ICC) was calculated to determine the reliability of each variable across the repeated trials. The standard error of measurement (SEM) was also calculated to assess the test's reliability, for example, a larger SEM indicates a lower test reliability and was calculated as follows:

$$SEm = SD \circ \left(\sqrt{1} - ICC\right)$$

#### **3. Results**

The results are presented with respect to the knee joint motion, thigh motion, pelvis motion, and inter-segmental coordination employed during the unilateral 180◦ jump. The ICC values (see Table 1) show a large range (ICC's Dominant Limb: Pelvis 0.03–0.83; Knee 0.34–0.92; Non-Dominant Limb: Pelvis 0.13–0.74; Knee 0.32–0.89) of test–retest reliability scores across the variables of interest, with a tendency for the dominant limb to demonstrate greater reliability across a greater number of variables.

#### *3.1. Knee Joint Motion*

During both the loading and landing phases, peak knee abduction was found to differ significantly between the legs, with the knee of the non-dominant leg abducting more than the dominant leg (see Table 2). Peak knee adduction between the legs differed significantly in the landing phase, with the knee of the dominant leg adducting more than the knee of the non-dominant leg. Peak knee adduction and peak knee abduction occurred within the early and late phases, respectively, during both the loading and landing phases. All other knee variables did not differ between the dominant and non-dominant leg within the loading and landing phases.


**Table 1.** Reliability measures.

ICC 0.5–0.75 = moderate reliability, ICC 0.75–0.90 = good reliability, & ICC > 0.90 = excellent reliability. Larger SEM indicates less reliability.

**Table 2.** Knee motion for both loading and landing phases. Abd—Abduction; Add—Adduction.


Mean ± Standard Deviation. \* Significant at *p* ≤ 0.05. Cohen's d calculated for effect size (0.2–0.5 are considered small, 0.5–0.8 considered moderate, and >0.8 considered a large effect).

#### *3.2. Thigh Motion*

There was no difference in axial rotation of the thigh between the dominant and non-dominant leg during both the loading (*p* = 0.89, d = 0.04) and landing phases (*p* = 0.77, d = 0.14). During the loading phase thigh rotation for the dominant leg was 2.2 ± 5.3 degrees and the non-dominant leg was 2.4 ± 5.0 degrees. During the landing phase, thigh rotation for the dominant leg was 6.1 ± 4.5 degrees and the non-dominant leg was 5.6 ± 4.1 degrees. In preparation for executing and following the single-leg 180◦ jump, the thigh segments of the dominant and non-dominant legs internally rotated.

#### *3.3. Pelvis Motion*

During both the loading and landing phases, pelvis motion was not significantly different between dominant and non-dominant legs (Table 3). In preparation for taking-off, the pelvis was found to extend (sagittal plane), adduct (frontal plane), and externally rotate (transverse plane). On landing, the pelvis was found to extend, abduct, and internally rotate.


**Table 3.** Pelvis alignment for both loading and landing phases. Flex—Flexion; Ext—Extension; Add—Adduction; Abd—Abduction; Rot—Rotation.

Mean ± Standard Deviation. Cohen's d calculated for effect size (0.2–0.5 are considered small, 0.5–0.8 considered moderate and >0.8 considered a large effect).

#### *3.4. Inter-Segmental Coordination*

The three exemplar CRP plots presented (Figure 1) denote differences in pelvis/thorax segmental coordination for both the dominant and non-dominant take-off legs within the 180◦ jump. As shown in Figure 1, all participants tended to remain in-phase with minimal variability during the loading/take-off phase of the jump irrespective of the take-off leg. All three participants remain in-phase with low variability throughout the unilateral 180◦ jump when taking off from the non-dominant leg. Participant C employs a similar pattern of inter-segmental coordination (and variability) irrespective of the take-off leg. Participants A and B show a tendency to move out-of-phase and demonstrate high levels of CRP variability throughout the landing phase of the unilateral 180◦ jump.

**Figure 1.** Exemplar continuous relative phase plots.

The solid red line represents the mean relative phase angle of all five jumps with a phase angle of 0 representing completely in phase or both the thorax and pelvis rotating together at the same rate. A phase angle > 0 (positive angle) indicates that the thorax is rotating ahead or leading the pelvis. A phase angle < 0 (negative angle) indicates the pelvis is rotating ahead or leading the thorax. The loading phase is represented by the first ~20%. The landing phase is represented by the last ~50%. Vertical error bars indicate the standard deviation at points of the CRP plot. Therefore, larger vertical bars indicate greater movement variability jump to jump, while smaller bars represent less variability.

In phase is represented by a relative phase angle of 0. Out of phase (thorax-led) is represented by a positive relative phase angle.

#### **4. Discussion**

The principal aim of this study was to demonstrate the utility of the unilateral 180◦ jump as a potential physical assessment tool by providing descriptive kinematic data. All participants demonstrated similar knee and thigh motion and pelvis motion. During the loading phase there were no significant differences between limbs, with all participants executing internal rotation of the thigh while extending, adducting, and externally rotating the pelvis. These movements portray a "counter movement" pattern opposite to the desired action of rotation, by loading the system up similar to the downward movement (knee flexion) preceding forceful knee extension one would observe in a vertical jump. This rotational "counter movement" pattern may be performed in an attempt to load the associated musculature and utilize strain energy to facilitate the stretch-shortening cycle. Previous research has indicated hip flexion with hip internal rotation has a greater correlation with large knee valgus moments [29], therefore the observed external rotation of the hips in healthy athletes may be a subconscious attempt to reduce potentially deleterious forces at the knee.

During the landing phase there were no significant differences between limbs for thigh rotation (*p* = 0.77, d = 0.14) or pelvis movement (Flex/Ext [*p* = 0.95, d = 0.03], Add/Abd [*p* = 0.6, d = 0.22], Axial Rotation [*p* = 0.82, d = 0.08]). For all participants, the thigh internally rotated and the pelvis extended, abducted and internally rotated. For cutting tasks with up to a 110◦ change in direction, hip internal rotation in conjunction with hip abduction was associated with larger knee valgus moments [30,31]. The larger the degree of rotation the greater the rotational forces [30], indicating a necessity to execute the task and not necessarily a greater risk of injury. Since the participants of this study were healthy elite-level athletes, the pelvis moving in line with the rotation of the jump may have been a strategy to reduce torsional forces experienced at the knee and ankle by dissipating energy at a link in the kinetic chain with a greater range/number of degrees of freedom (DOF).

There was a significant difference in the frontal plane knee motion during the landing phase between limbs (Add [*p* = 0.05, d = 0.79], Abd [*p* = 0.01, d = 1.11]). Greater knee adduction was observed in the dominant limb (~4.2◦ greater), whereas greater knee abduction was observed in the non-dominant limb (~6.5◦ greater). Moreover, it should be noted that these specific findings had some of the strongest ICC values and lowest SEM values (see Table 1). This finding indicates that a bilateral asymmetry is observable even in healthy athletic populations. Due to the cross-sectional nature of this study, it is impossible to indicate which strategy is more beneficial. Uncontrolled knee movement during landing has been associated with greater ACL injury risk in uniplanar drop jump tasks [23] but rotational tasks require greater knee frontal plane excursions to complete [29–31]. Therefore, prospective evidence is needed to indicate which motion contributes to a greater risk of injury. Furthermore, one motion is most likely not good or bad, but stresses different soft tissue structures more greatly (for example, abduction: ACL, adduction: MCL). Likewise, how the system (body) moves as a whole is more important for understanding the potential injury risk.

Visual qualitative assessment of CRP plots indicated that there was little variation between participants during the loading/take-off phase (refer to the first 20% in Figure 1). Moreover, the relative phase angle between the shoulders and the pelvis for the majority of the participants was very close to 0◦. A phase angle at or close to 0◦ indicates that the shoulders and pelvis were in-phase or moving together. Coordinated coupling of the upper and lower body segments during the loading phase may be a result of the population tested. Healthy athletes presumably have better motor control and coordinate their torsos in a synergistic fashion with their lower extremities to improve mechanical efficiency by reducing counter-productive movement. Furthermore, this may demonstrate torso control by the participants to reduce torsional forces experienced at the knee that has been observed in cutting tasks when torso lean/excursion is excessive [28]. Likewise, the amelioration of the torsional force and modulation of body segments may help maintain the center of gravity within the base of support so that the individual successfully executes the task (does not fall over). Prospective research should be conducted on unhealthy or non-athletic populations to determine if uncontrolled torso movement can be observed during the loading phase of a unilateral 180◦ jump task. Investigating 'less athletic' or impaired populations will further elucidate motor control capabilities, differentiating key differences in task execution compared to elite-level athletes when performing this specific task.

Whereas the loading/take-off phase exhibited relative uniformity between participants, the greatest amount of variation was observed during the landing phase of the 180◦ jump (refer to last 50% of Figure 1). CRP plots displayed three common landing strategies amongst participants: thorax (shoulder)-led movement, pelvis-led, or simultaneous in-phase landing. Although these three strategies were performed, it should be noted that each individual only utilized one of them and with varying degrees of phase coupling. In some cases, individuals were close to a relative phase angle of 180◦, which represents an anti-phase movement (similar to arms swinging in separate directions at the same rate during normal gait). Furthermore, the phase coupling between the thorax and pelvis was not constant, meaning the rates at which the two segments rotated in conjunction with one another were at different rates throughout the second half of the unilateral 180◦ jump task, suggesting a dissociation between the segments during the landing portion of the task.

CRP best describes the passage of energy through a system [39,40], in this case, the human body. Each participant utilized a slightly different strategy to dissipate the high torsional energy in order to execute the task and land successfully. The fact that many strategies with different degrees of variation were observed demonstrates the principle of equifinality of movement solutions due to the vast number of DOF [41,42]. Although there are multitudes of strategies to complete the task, it stands to reason a shoulder-led strategy may be the most deleterious for structures around the knee. If the shoulders lead with the pelvis following, upon landing when a high load rate force is transmitted up the kinetic chain the shoulders will have little range left for further excursion to enable energy dissipation. Instead, it could likely result in a resultant force transmitted back down to attempt to be dissipated by the pelvis movement. This could result in greater hip flexion and hip internal rotation, which has been linked to greater knee valgus moments [29–31]. A pelvis-led strategy may enable better dissipation of excessive energy because it can pass up through the entire kinetic chain. However, because this study is cross-sectional this theory is conjecture on the part of the authors. Similar to research involving drop jumps [23,43,44], future research could yield information regarding which strategy portends greater injury risk and what relative phase angles indicate problematic upper torso and lower body coupling during rotational sporting tasks.

Further evidence to support this approach can be observed in the test–retest reliability (ICC) scores (see Table 1). Large variability can be observed across the ICC scores, and can be interpreted as being indicative of the presence of variability in specific movements, thus, enabling the production of more constrained, and reliable movements in key structures (depending on the movement strategy employed). There are some limitations to this study, mainly, the small sample size (n = 14). However, the use of elite-level athletes (rugby players) not only strengthens the external validity of this investigation (the application for athletes), the smaller sample size is representative of the smaller percentage of the population that achieved elite athlete level status (professional). Furthermore, for significant findings the effect sizes were moderate to large (>0.79) (see Table 2) and the reliability measures (ICC [0.79–0.92] and SEM [0.44–1.26]) for those that were significant were moderate to strong (see Table 1). Considering the knee is a major focal point of injury risk assessment, these aforementioned values support the unilateral 180◦ jump as a potential rotary injury risk assessment tool.

#### **5. Practical Application**

The unilateral 180◦ jump provides a plethora of biomechanical data. Firstly, the unilateral 180◦ jump task can be performed in a confined space making it a more feasible alternative to performing cutting tasks when assessing lower body rotary risk. Requiring the participant to jump as high as possible could provide kinetic data and a simultaneous metric on vertical jump performance. Although jump heights will not be as large in magnitude as a traditional countermovement jump, frequently, jumps are made off a single leg and unbalanced (not having time to properly align and set the body) during play, thus, the 180◦ jump task replicates scenarios experienced in sport. Furthermore, the requirement of maximal jump height forces the participant to utilize muscle contraction velocities that are similar to game conditions, unlike other previously mentioned assessments [2,11,12,18,45]. Likewise, landing from a maximal vertical height while rotating will exert a multi-planar loading force (vertical and torsional) typical to athletes in sports such as soccer, rugby, football, or basketball. Moreover, landing unilaterally while rotating more closely mimics one of the most common mechanisms of ACL injury as compared to landing from a uniplanar jump task [4].

Future studies should investigate the unilateral 180◦ jump with pathological and non-pathological populations prospectively (over the course of an athletic season) to associate movement patterns with injury predictors. Due to the multiple variables that can be collected at once with the unilateral 180◦ jump and the ability to visually assess the movement, the task has the potential to be a valid and efficient objective prognostic physical assessment tool that may provide greater sensitivity for identifying rotational injury risk.

**Author Contributions:** C.C. conceived the concept of this investigation and supervised the project which was carried out by C.M. Data was reduced and analyzed by S.B. K.T.K. wrote the manuscript and created the tables/figures with the support A.M.S., G.L.M. and H.S.L. All authors have read and agreed to the published version of the manuscript.

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

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

#### **References**


© 2020 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 (http://creativecommons.org/licenses/by/4.0/).

*Article*

### **Do Grade II Ankle Sprains Have Chronic E**ff**ects on the Functional Ability of Ballet Dancers Performing Single-Leg Flat-Foot Stance? An Observational Cross-Sectional Study**

**Bruno Dino Bodini 1, Giacomo Lucenteforte 2, Pietro Serafin 1, Lorenzo Barone 2, Jacopo A. Vitale 1,\*, Antonio Serafin 1, Valerio Sansone 1,3 and Francesco Negrini <sup>1</sup>**


Received: 19 November 2019; Accepted: 18 December 2019; Published: 24 December 2019 -

**Abstract:** Ballet dancers have a higher risk than the general population of ankle sprains. Ankle proprioception is of the utmost importance for executing static and dynamic positions typical of ballet dancing. Ankle sprains can create changes in functional ability that may affect ballet performance. The aim of this cross-sectional observational study is to evaluate if non-professional ballet dancers that were previously injured with a grade II ankle sprain carry a long-term stability deficit in ballet specific positions (passé, arabesque) and in single-leg flat-foot stance, thereby affecting ballet performance. We enrolled 22 amateur female ballet dancers, 11 who previously had a grade II ankle injury and 11 who had no history of ankle injury. Stabilometric data (Center of Pressure Speed and Elipse Area) were assessed with the postural electronic multisensory baropodometer in normal, arabesque, and passè positions with both open and closed eyes. Using an unpaired *t*-test, we compared healthy and pathological feet of the ankle injury group for a standard monopodalic position and two ballet-specific positions. No difference between pathological and healthy feet of non-professional ballet dancers who suffered grade II ankle injury was detected. According to the parameters considered in this study, grade II ankle sprains seem to have a favorable prognosis in the sample that we evaluated.

**Keywords:** stabilometry; ballet dancers; ankle injury; chronic ankle instability; balance

#### **1. Introduction**

During performance, ballet dancers have a great need for stability and balance control [1,2], along with a high level of general "flexibility", which are peculiar characteristics that improve with ballet training [1,3–5]. Ankle control is of the utmost importance for executing static and dynamic positions typical of ballet dancing [6,7]. In 2017, Vassallo et al. reported the epidemiology of dance-related injuries between 2000 and 2013, defining ankle sprains as the most common injury [8]. Later, Smith et al. and Smith et al. collected data about their incidence and prevalence [9,10]. Ankle sprains could cause chronic functional ability changes, as is common in groups of dancers and non-dancers alike [11,12], also increasing the risk of re-injury [13]. In addition, it was seen that people with functional ankle instability had a less refined kinesthesia, in particular for inversion movements [14]. However, a prospective study conducted in 1996 by Leanderson et al., based on stabilometry, demonstrated only short-term consequences after ankle sprain in ballet dancers [15]. Data about chronic ankle instability (CAI) in the general sporting population are, to date, not satisfactory to determine the presence of common aspects of CAI within individual sports [16]. However, an exploratory study conducted by Simon et al. revealed a condition of CAI in 75.9% of university dance majors [17]. Despite new wearable technologies, such as Inertial Measurament Unit IMU [18,19] and wireless electromyography (EMG) [20], having been proposed to investigate balance and postural control, stabilometry remains the most commonly used tool [21]. To date, there is a lack of studies about the chronic effects of an ankle injury in ballet dancers using stabilometry evaluation, but there is an increasing body of literature concerning this method [22,23], including not previously injured ballet dancers [24]. A study performed by Lin et al., using stabilometry to assess dancers with ankle injuries, found that non-acute ankle injuries could worsen postural stability of professional ballet-dancers during ballet-specific postures. However, they did not focus on a specific injury nor on a precise time-frame for the post-injury evaluation [25]. Another relevant aspect in the evaluation of balance control is represented by the visual system's role in proprioception, which has the potential to mask proprioceptive issues in athletes [26]. For this reason, it can be argued that in order to obtain reliable proprioceptive measures, visual occlusion is mandatory [27,28]. As ankle control is crucial for successfully executing ballet specific positions and it could be impaired in ankle sprains, a position specific testing with closed eyes is needed. Moreover, some studies showed that athletic and sport training is task specific and skills acquired in a specific task did not completely translate to the same skills in different tasks [29–31]. Recently, Thalassinos et al. discovered that there were many sport skill-specific bias about sensory inputs for spatial orientation and postural control between experienced soccer athletes, ballet dancers, and nonathletes [32]. Considering the high incidence and prevalence of ankle injuries in professional and non-professional ballet dancers (from 3% to 25.6% of all musculoskeletal ballet injuries [4,8,10,33]), and the primary role of ankle control to perform ballet-specific tasks, the aim of our study was to understand if non-professional ballet dancers, previously injured with a grade II ankle sprain, compared to dancers who did not suffer any ankle injury, carry a long-term stability deficit in ballet specific positions (passé, arabesque), thereby affecting ballet performance.

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

This was a cross-sectional observational multi-center study. We enrolled 22 subjects (N = 11 in the pathological group and N = 11 in the control group) from six different schools of classical ballet dance in Italy. The inclusion criteria were: age between 15 and 25 years old; classical ballet dancer; at least 10 years of training (average of 2–3 sessions of training a week); right footed in ballet practice; unilateral ankle injuries; clinical diagnosis of grade II inversion right ankle sprain that occurred at least 6 months before our evaluations made by an orthopedic surgeon [34]; and informed consent signed (parents' consent was collected if subjects were under 18 years old). The exclusion criteria were: professional dancer; another lower extremity injury; history of neurological or motor deficits; and any visual impairment. The control group had the same inclusion and exclusion criteria, except for the history of ankle injury. For grading the severity of the ankle injury, we used the classification proposed by Malliaropoulos et al. [34], and classified grade II ankle injury patients positive to anterior drawer test but negative to talar tilt test. We defined grade II ankle injury as a partial tear of the lateral ligament complex of the ankle without decreased motion and loss of function. All pathological subjects were treated conservatively with soft bandages applied for 2 weeks. No patient underwent a specific rehabilitation protocol. At the moment of inclusion in the study, all patients were symptom free and fully participating in dance at their pre-injury level. Before entering the study, the participants were fully informed about the study's aims and procedures, and written informed consent was obtained before testing. The study protcol was approved by the Ethics Committee of the University of Milan

(approved on 12/10/15, Prot. N. 54/15) in accordance with current national and international laws and regulations governing the use of human subjects (Declaration of Helsinki II).

#### **3. Sample Size Calculation**

In order to determine the sample size of the study, we carried out an a-priori power analysis on the basis of the scientific literature. Since our main outcome was to determine if there was a difference in monopodalic stance ellipse area between healthy and previously pathological ballet dancers, we used data from the 2017 paper by de Mello et al. [24]. With a reported monopodalic ellipse area in ballet dancers of <sup>155</sup> <sup>±</sup> 64 mm2 and an expected difference of 70 mm2 (effect size <sup>=</sup> 1.1), we calculated that we would need a sample of 22 subjects (11 for each group) to detect a significant change with a power of 0.80 and an alpha of 0.05. To calculate the sample size, we used the GPower software (Universitat Dusseldorf-Germany).

#### **4. Tools and Procedures**

The internationally approved stabilometric parameters [35] we evaluated in this paper are: (1) Ellipse area (mm2), contains 90% of the positions sampled of the center of pressure (CoP) and represents the dispersion of the oscillations and the precision of the system; and (2) CoP speed (mm/s), the average speed related to the CoP oscillations. The acquisition of stabilometric data was performed with the postural electronic multisensory baropodometer Diasu®, equipped with Milletrix® interface, which has previously been shown to be a valid and reliable tool [36] This tool features a scanning frequency of 100 frames per second in real time, acquisition surface of 40 cm2 per module, accuracy of +/−5%, and maximum point pressure of 150 N/cm2. The evaluations were carried out in six different gyms with a dedicated medical room, characterized by the absence of external perturbations, especially acoustic. Prior to the test, each dancer stretched and warmed up the muscles of the lower limbs via routine self-selected exercises for 5 min. Each dancer wore comfortable and adherent clothes, and a pair of cotton socks.

During the test, an expert podiatrist and a nurse were present in the room. The baropodometer was placed at a distance of two meters from a homogeneous color wall that the subjects faced during the whole time of examination. Each evaluation session was held in the early afternoon and each subject had to be at rest from training in the previous day. The subjects were evaluated with both eyes open and closed. To ensure a reference point to be observed during acquisitions, a mark was positioned on the wall at the same height and equidistant to the patient's eyes, and at a distance of 2 m. We acquired each measurement three times with eyes open and three times with eyes closed. The first try was considered as a demonstration and we did not record it. For each foot we tested, the patients in the following conditions both with eyes opened and closed.

Monopodalic stance with the contralateral limb slightly detached from the ground (5 s length) (Figure 1).

**Figure 1.** Monopodalic position.

Monopodalic with the limb in contact with the ground in "en dehors" position, and the contralateral in "arabesque" position, defining a 45◦ angle between them (5 s length) (Figure 2).

**Figure 2.** Arabesque.

Monopodalic with the limb in contact with the ground in "en dehors" position and the contralateral in "passé" position with a back "retiré" (5 s length) (Figure 3).

**Figure 3.** Passè.

#### **5. Statistical Analysis**

For each subject, parameters were determined by calculating the mean of the trials; subsequently, the grand average and the standard deviation over the subjects in each position were computed. Preliminary tests for normality (Kolmogorov-Smirnov test) and for equality of sample variances (Levene's test) provided the basis for using parametric statistics. All parameters were normally distributed. First, we checked if there were any significant statistical differences between the two groups as far as age, body mass index (BMI), years of dance practice, or age when they started practicing using an unpaired Student's *t*-test. Second, we compared stabilometric parameters, of each position, between the left and right foot in healthy subjects, and between injured and non-injured feet in the post-injury patients by the unpaired Student's *t*-test. Third, the difference between open and closed eyes in each condition, for each foot, was calculated; then, we checked if the differences in stabilometric parameters between closed and open eyes were larger in pathological vs. healthy feet

using the unpaired Student's *t*-test. The level of significance was set to *p* < 0.05. Statistical analysis was performed using IBM SPSS 20.

#### **6. Results**

Demographical data of the study population are shown in Table 1. We could not find any statistical difference between the two groups for the parameters of age, BMI, years of dance practice, or age when they started practicing.


**Table 1.** Demographical data of participants.

Data are reported as Mean ± SD. Abbreviation: BMI, body mass index.

First, we analyzed the healthy subjects. We compared stabilometric parameters of the right and left feet for each parameter and no statistical differences for all parameters were detected (Table 2).

**Table 2.** Comparison between right and left foot in healthy subjects (N = 11) for stabilometric parameters acquired with open eyes.


The pathological group was then analyzed. We compared stabilometric parameters between the injured feet and the healthy feet and no statistically significant differences were observed (Table 3).

**Table 3.** Comparisons between injured and healthy feet in pathological subjects (N = 11) for stabilometric parameters acquired with open eyes.


Data are reported as Mean ± SD. Abbreviation: CoP, Centre of Pressure.

We calculated the differences between open and closed eyes for all the parameters, and with reference to the pathological group, data collected with the feet that suffered a grade II ankle injury were compared with healthy contralateral limbs. As reported in Table 4, no significant differences between the sides were observed.

Lastly, we also compared right foot parameters of healthy and pathological subjects, both with open eyes (Table 5) and the difference between open and closed eyes (Table 6), and no significant differences were detected.


**Table 4.** Comparison between injured and healthy feet in pathological subjects (N = 11). Mean delta values (difference between open and closed eyes) of stabilometric parameters.

Data are reported as Mean ± SD. Abbreviation: CoP, Centre of Pressure.

**Table 5.** Comparison between right feet of pathological and healthy subjects for stabilometric parameters acquired with open eyes.


Data are reported as Mean ± SD. Abbreviation: CoP, Centre of Pressure.

**Table 6.** Comparison between right foot of pathological and healthy subjects. Mean delta values (difference between open and closed eyes) of stabilometric parameters.


Data are reported as Mean ± SD. Abbreviation: CoP, Centre of Pressure.

#### **7. Discussion**

Our study showed that there is no ballet specific long-term balance deficit for grade II ankle injuries in ballet dancers. In order to have this specific conclusion, we checked if our sample of non-professional, but experienced, ballet dancers had any difference in balance while performing ballet specific positions on the dominant or non-dominant foot. We did not find any difference between the two sides. This finding is in contrast with a classic study performed by Leanderson et al. in 1996, where they tested professional athletes and non-athletes in the monopodalic position [15]. However, the study cited is 22 years older, so training techniques and materials such as shoes have changed, and are performed in a different, monocentric, and professional setting. This could very well explain the differences found. Moreover, the verified parameters of validity and reliability of the tools we used [36] and the power set at 80% for the sample size calculation, could be some of the reasons of the differences found between the outcomes of our study and the study published by Leanderson et al. Since the right and left foot were found to have no difference in the healthy group, we compared injured and non-injured side of our pathological group. We found out that, after six months, there was no significant difference in balance between the injured and non-injured sides. The internal (same subject) comparison is of the utmost importance, as any kind of difference in training could be an

important bias [37]. In fact, using the contralateral healthy feet as a control, we were able to eliminate biases caused by potential differences in skills, body types, and other physical characteristics between athletes. Recently, a few studies pointed out that a crossover effect of injuries can exist, especially tendinitis, from one side to the other [38,39]. While it is not clear the mechanism of action of this crossover effect (e.g., central nervous system or anticipatory postural adjustments [40]), it still could be a possible confounder of our results as the pathological and healthy side could be similar because of it. Our study arrived to a different conclusion than a similar 2011 study [25]; however, we focused on a more selective group of subjects who had only grade II ankle injury. We supposed that patients with more severe ankle sprains could have long-term impairment in postural control.

One strength of our study was the use of stabilometry in a new way, comparing ballet-specific static positions for determining ankle injury effects. We found just two studies that examined ballet-specific positions; one examined the single-leg retiré position [37] and the other the passé en demi-pointe position [24]. However, neither of the studies focused on post-injury athletes. Another strength of the study was the comparison between healthy and injured feet in the same athletes that eliminate the role of possible differences in skills between different subjects, allowing us to better study the role of the previous ankle injury. Furthermore, we recruited the subjects in six different ballet schools. Different centers have different training strategies, so the results obtained are probably not linked to a peculiar type of training but to the natural history of grade II ankle injuries.

A limitation of our study is the small number of subjects involved; however, we enrolled 22 patients (11 for each group) as resulted with the sample size calculation. Our inclusion criteria were very strict to guarantee the internal validity of our study, and despite screening six different centers, we were able to find 11 patients meeting the required inclusion criteria for the pathological subjects group. Furthermore, because of our study design, we cannot exclude an important role of specific rehabilitation protocol in speeding up the recovery process, as we tested our patients long after their injury.

In the future, it could be important to study in a similar fashion different grades of ankle injuries, classified using imaging support, as well as different types of ankle injuries, such as inversion and eversion. Moreover, it could be interesting to evaluate the parameters used in our study in relevé, another ballet-specific position. One more possible future study could investigate the role of treatment (e.g., immobilization (functional or rigid) vs. no immobilization) and rehabilitation in the time to return to optimal performance as well as potential long-term effects in ankle injuries of higher clinical grading.

Our finding could encourage many dancers all over the world that suffered and will suffer ankle injuries [8]. Dancers with a grade II ankle injury should be reassured by clinicians that their functional ability as ballet dancers performing single-leg flat-foot stance will likely not be hindered in the long term.

#### **8. Conclusions**

Our findings show that grade II ankle injury does not compromise the ability of a ballet dancer, even if treated with just soft bandage applied for two weeks, even without a specific rehabilitation protocol.

**Funding:** This work was supported by the Italian Ministry of Health (Ricerca Corrente).

**Conflicts of Interest:** The authors have 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.

**Author Contributions:** B.D.B., G.L., P.S., and F.N. designed the present study, checked the literature, and wrote the manuscript; J.A.V. performed the statistical analysis; L.B., J.A.V., A.S., and V.S. critically revised the manuscript; all authors gave the approval for the final version of the review submitted for publication. All authors have read and agreed to the published version of the manuscript.

#### **References**


© 2019 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 (http://creativecommons.org/licenses/by/4.0/).
