As an Olympic sport, competitive Taekwondo (TKD) is one where players accumulate points based on their kicking technique and strategies (Estevan and Falco, 2013). TKD athletes tend to emphasize head-height kicks, which are harder to execute, because they are worth more points. According to current Olympic rules, a valid kick to the head is awarded three points, whereas a kick to the abdomen is awarded only one point (Evelyn Watta, 2020). Thus, head kick strategies are frequently adopted by TKD players, with the round house kick (RHK) being the most popular technique (Ho, 2018; Estevan and Falco, 2013). RHK requires stable support from the pivot leg for the attacking leg to land the kick (Figure 1). A previous study showed that gradually increasing hip displacement of the pivot leg as the target distance increased was associated with the pivot foot rotation (Kim et al., 2010a). This implies that difficult kicks (e.g., kicks aimed at a high target) are especially demanding on the pivot-leg. Because kicks aimed at a high target require mature skill and excellent motor coordination (Kim et al., 2010a), TKD beginners are advised against performing such kicks (Estevan and Falco, 2013).

According to recent sports injury studies, TKD kicks can cause serious injuries to the pivot leg, such as anterior cruciate ligament (ACL) injuries (Kasbparast et al., 2014a; 2014b; Kim et al., 2010b). Some scholars have suggested that this type of injury mechanism may be associated with the requisite planting and spinning motion during the RHK (Kasbparast et al., 2014a; 2014b; Amraee et al., 2013; Park and Song, 2018). Injury-causing motions have been mostly observed in the pivot leg (Kasbparast et al., 2014b). A study from Kobayashi et al. (2010) showed that ACL injuries of the pivot leg commonly occurred in athletes when they were performing knee valgus movements with the toe pointing outward and the knee inward, the toe pointing inward and the knee pointing outward and with knee hyperextension during dynamic alignment. For example, when chasing or performing spinning kicks, players are susceptible to noncontact injuries to the knee joint (Kasbparast et al., 2014a; 2014b). Excessive knee valgus is a major cause of ACL injuries (Cortes et al., 2011; Hewett et al., 2005; Kellis et al., 2004). To avoid such injuries, dynamic stability at the hip and knee are necessary to achieve the desirable joint stability, especially in the performance of difficult kicks (Estevan and Falco, 2013). The co-contraction of muscles must also be increased to provide joint stability and protection against awkward, injury-inducing angles (Kellis et al., 2003; Opar and Serpell, 2014). Muscles that have been confirmed to be effective in protecting the hip and ensuring knee stability are the hamstring muscles and gluteus medias (GM) (Distefano et al., 2009; Petersen et al., 2011). These muscles that exhibit higher activation ratios relative to the quadriceps muscle and adductor muscles also provide high levels of stability protection for the hip and knee of the pivot leg during a kicking motion (Brophy et al., 2010; Monajati et al., 2016; Opar and Serpell, 2014; Ruan et al., 2017).

Despite the risk for noncontact injuries in the knee joint on the pivot leg resulting from TKD kicks, most studies focused on the attacking leg (Estevan and Falco, 2013; Gavagan and Sayers, 2017). Because of the current Olympic rules, head-height kicks have been popular point-scoring strategies (Ho, 2018; Tornello et al., 2014). However, the increase in kick height obviously affects the coordinated movement of the pivot foot in TKD athletes. (Kim et al., 2010a). Few studies have focused on the pivot leg when performing TKD kicks. The relevant mechanism remains unknown. Based on the needs identified, this study analysed the effects of (three instances of) RHK heights on kinetic and neuromuscular functions, at the hip and knee on the pivot leg. It was hypothesised that higher RHKs resulted in increased load on knee dynamic stability in the pivot leg. Such increased load may reflect in the return phase by increased dynamic knee valgus angles as well as the co-activation of the biceps femoris (BF), rectus femoris (RF), GM, and hip adductor muscles.

Twelve junior elite college TKD players (6 males and 6 females) were recruited to participate in this study. Mean subject height, weight and age were 172.08 ± 9.32 cm, 62.17 ± 9.45 kg, and 20.33 ± 1.27 years old. The amount of time dedicated to the sport was 11.25 ± 2.83 years. The participants with severe lower extremity skeletal and nerve disease, a history of ACL injury, a history of surgery, or knee pain or functional limitations for the last two months were excluded. The participant's usual pivot and attacking feet were determined by their self-experience. Participants were asked not to exercise vigorously or perform lower limb training one week before each experimental session to prevent fatigue from affecting the experimental results. This study was reviewed and approved by the Fu Jun Catholic University Human Research Ethics Committee, Taiwan (C106192). This study conforms to the Declaration of Helsinki for studies involving humans.

This study designed three different height dynamic kicking tests (the navel (low), xiphoid (middle) and nose (high) positions of the body, referred to as HT, MT and LT). Each kick was executed randomly 5 times. Before the experiment began, the participant was prepared to execute the kick by assuming a consciously comfortable stride distance. When the preliminary command was issued, the researcher quickly moved the target to the pre-event set height within 1 to 10 seconds. The subject struck the target with the utmost effort. In the experiment participants could perform a kick by pivoting or jumping on their pivot leg. However, for a valid kick, the front foot must not move forward, backward, or to the side more than 5 cm from the start position when the kick is executed. In our study, we use the reflective maker fixed to the second metatarsal head of the pivot leg to detect an accurate distance. The participant could rest for at least 30 seconds after each kick.

The athletes were then required to complete a series of maximum voluntary isometric contractions (MVIC) to quantify the EMG as a normalisation factor for each muscle in accordance with the usual practice. The participant first assumed a sitting position (the hips and knees were flexed 90 degrees). A buckled strap was used to fix the distal tibia to resist knee extension (RF). In the same way, in the subject assumed a lying position (the hips were 0 degree, and the knees were flexed 90 degrees) to resist knee flexion (BF), and hip abduction (GM) and hip adduction (adductor).

The Vicon motion analysis system (VICON Motion Systems Ltd., United States) and six high-speed cameras (Vicon Vero 2.2, United States) were used to record the kinematic data. The sampling frequency was set at 100 Hz. The participants were required to perform a specific anatomical movement by a qualified athletic trainer to adhere to the reflective markers (Cappozzo et al., 1995). The coordinate position during the operation was captured by six high-speed red line cameras. The Delsys SDK 2.5+ TrignoTM system (Delsys, Boston, USA) was used mainly in the capture of human muscle EMG signals, were captured at the sampling rate of 2000 Hz. After the neural signal was converted into an analogue signal via the amplifier, it was transmitted into the Delsys multi-function neural signal capture box. Finally, together with the reflective marker coordinate data, the Neurox 2.5.0 processing software was used to output the ASCII data for analysis and processing.

Before the start of the experiment, the participants performed a complete large muscle stretch procedure, including the lumbar muscles, gluteus maximus, quadriceps, hamstring, gastrocnemius, and tibialis anterior muscles. Each muscle was stretched at least 30 seconds. The researchers then measured the actual leg length data (navel to internal haemorrhoids) with a tape measure as described and used by Um and Bae (2011). This was followed by setting myoelectric receiver devices onto the lower limb muscle belly, including the RF, BF, GM or adductor femoris (AD), was referenced from the book by Criswell (2010). After the device attachment was completed, the maximum isometric contraction value of the electromyogram was collected. Each muscle group was subjected to two 5 second maximum isometric contractions to collect EMG data. Twenty-two reflective markers were then attached to a particular anatomical site on the adhesive sheet body, including the anterior superior iliac spine, the posterior superior iliac spine on both legs, the external and medial femoral condyles, the fibula head, the tibial tuberosity, the fibula styloid process, tibia styloid process, fifth metatarsal base, the calcaneus, and the second metatarsal head on the pivot leg. In addition, a fifth metatarsal styloid was attached to the attacking foot (Cappozzo et al., 1995).

All marker trajectories were filtered using a fourth-order, zero-phase-shift low-pass Butterworth 9 Hz by Matlab 2016a. The 3-dimensional coordinates of the hip joint centre were calculated as described from a previous study (Seidel et al., 1995). The centre of the knee joint was calculated as the midpoint between the medial and lateral femoral condyles, while the centre of the ankle joint was calculated as the midpoint between the medial and lateral malleoli (Cappozzo et al., 1995). The hip joint centre, left, and right anterior superior iliac spine and posterior superior iliac spine were used to define the pelvic reference frame. The hip centre, knee centre, and femoral epicondyle were used to define the femur reference frame. The knee joint centre, the centre of the ankle joint, the fibula head and tibial tuberosity were used to define the tibia reference frame (Cappozzo et al., 1995). The order of rotation through the Euler angle formula between the pelvis-femoral and the femoral-tibia coordinate systems were (1) flexion-extension (x-axis), (2) abduction-adduction (y-axis), (3) internal-external rotation (z-axis). Dynamic anatomical angles were calculated referenced to a static standing neutral position.

This study used the kinematic information to stage each kicking process. The kick attack phase was defined as the duration from the time of initial toe-off to the moment the target was struck. The kick return phase was defined as from the moment the target was struck to the moment the attacking foot touched the ground. In addition, the reflective marker displacement on the attacking leg’s 5th styloid process was used to calculate the velocity and acceleration maximum values. These values were used to determine the athletes kicking quality.

The myoelectric data was bandpass filtered with a 4th-order-zero-butterworth filter of 20 to 450 Hz. The root mean square (RMS) of each period was calculated after deducting the myoelectric activation value at rest. The MVIC took the maximum root mean square activation value obtained from the single-length maximum contraction of each muscle for 5 seconds in a 1-second window range. The muscle myoelectric activation values of each dynamic motion phase were normalised to obtain the myoelectric activation (MVIC). In addition, the co-contraction index (CI) formulas of BF-RF and GM-AD are BF/(RF+BF) and GM/(GM+AD), respectively (Kellis et al., 2003).

This study used the IBM SPSS Statistics 20.0 statistical computing software for statistical testing. The basic participant information, kicking speed, as well as the joint angle, muscle contraction ratio and co-activation values, on pivot foot were taken using repeated measures with the least significant difference (LSD) multiple comparison procedure. The post-hoc test was then used to compare results from the variables at three different kicking heights. The alpha level was set at 0.05.

The maximum kick velocity and acceleration for three different kick heights are presented in Table 1. There was no significant difference between the maximum kick velocity (p = 0.370) and acceleration (p = 0.408) of the players at different heights, indicating that the three RHK heights have similar kick quality.

Table 2 and Table 3 present the dynamic joint angles of the pivot leg, for different RHK heights, during attack and return.

There were significant differences on the pivot leg in the attack phase for angles of average hip flexion (p < 0.05), average hip abduction (p < 0.05), the range of hip frontal motion (p < 0.05). Furthermore, according to a post hoc comparison, HT values were greater than LT values (p < 0.05). The frontal motion hip range at HT significantly exceeded those at MT. On the pivot leg, the angles of average knee valgus (p < 0.05), maximum knee valgus (p < 0.05), the knee range at sagittal (p < 0.05) and frontal motions (p < 0.05) were significantly different. According to a post hoc comparison, the values at HT were all significantly smaller than those at LT during attack (p < 0.05). The other kinetic parameters exhibited no significant difference.

At the return phase, there were significant differences on the pivot leg for angles of average hip flexion (p < 0.05), average hip abduction (p < 0.05), maximum hip flexion (p < 0.05), minimum hip flexion (p < 0.05), maximum hip abduction (p < 0.05), and minimum hip abduction (p < 0.05). Furthermore, according to a post hoc comparison, HT values were greater than LT values (p < 0.05). Maximum hip flexion and minimum hip flexion at HT significantly exceeded those at MT. On the pivot leg, the angles of average knee flexion (p < 0.05), average knee valgus (p < 0.05), average knee external rotation (p < 0.05), minimum knee flexion (p < 0.05), maximum knee valgus (p < 0.05), maximum knee external rotation (p < 0.05), and maximum knee internal rotation (p < 0.05) were significantly different. According to a post hoc comparison, the angles of average knee flexion (p < 0.05), minimum knee flexion (p < 0.05), and maximum knee internal rotation (p < 0.05) at HT were significantly smaller than those at LT during deceleration (p < 0.05). By contrast the angles of the average knee valgus, minimum knee external rotation, maximum knee valgus, and maximum knee external rotation were larger than those at MT. The other kinetic parameters exhibited no significant differences.

Table 4 and Table 5 presented the hip muscle activities on the pivot leg, for different RHK heights, during attack and return.

During the RHK attack at different heights, this study detected no significant differences in all muscle (p = 0.103, p = 0.295), except CI_BF-RF (p < 0.01). A post hoc comparison demonstrated that the CI_BF-RF value at HT was significantly lower than that at LT (p < 0.05). The other EMG parameters exhibited no significant difference.

During RHK return at different heights, this study detected significant differences in RF_EMG (p < 0.05), CI_BF-RF (p < 0.01), and CI_GM-AD (p < 0.05). A post hoc comparison demonstrated that the RF_EMG value at HT was significantly smaller than that at LT (p < 0.05). The CI_BF-RF value at HT was significantly greater than the values at MT (p < 0.05) and LT (p < 0.05). The CI_GM-AD values at MT were significantly greater than values at HT (p < 0.05). The other EMG parameters exhibited no significant difference.

This study investigated the RHK height effect on hip–knee joint kinematics on the pivot leg and activation of the surrounding stabilizer muscles. According to the results, RHK height substantially affected the hip–knee joint angles at the pivot leg as well as co-activation of the surrounding stabilizer muscles. The results more specifically reveal the following: (1) During attack and return, compared with lower kicks, a higher kick resulted in greater maximum and average dynamic hip abduction and knee valgus angles. (2) During attack but not return of a high kick, joints on the pivot leg have a greater range of motion at the hip frontal plane as well as the knee sagittal and frontal planes. At the return phase, a high kick resulted in greater dynamic knee abduction and abduction angles and lesser minimum and average knee flexion angles. (3) During a high kick, the GM activation decreased, but the hamstring muscles co-contraction reaction considerably decreased during attack but substantially increased during return.

During attack, a higher RHK resulted in a larger dynamic hip abduction, knee valgus angles, as well as larger joint ranges of motion at the hip frontal plane, knee sagittal plane, and knee frontal plane. Previous studies observed that during an RHK, the body’s centre of mass tended to move horizontally and vertically toward the target, during which the body moved forward with an axial rotation at the pelvis (Gavagan and Sayers, 2017; Kinoshita and Fujii, 2014). At this moment, the hip joint at the attacking leg bent and extended to achieve knee extension, thus finishing the kick (Gavagan and Sayers, 2017; Kim et al., 2010b; Kinoshita and Fujii, 2014). In the preparation phase, the pivot leg stored potential energy required for the kick, including hip extension and hip adduction (Kinoshita and Fujii, 2014). During attack, the pivot leg produced kinetic energy, which drove pelvis rotation and forward inclination. This process improved kick performance (Gavagan and Sayers, 2017). The present study confirmed that an increase in kick height can greatly enhance the hip abduction and flexion angles on the pivot leg, thus enabling the practitioner to increase the kicking leg angles after pelvis rotation. Under these conditions, the kicking height can be increased without compromising kicking performance (Kim et al., 2010b).

Additionally, when the player attempted to increase their attacking range while maintaining the kicking speed, the pelvis on the pivot leg tended to push even further to increase the hip abduction angles (Cortes et al., 2011; Kim et al., 2010a; 2010b). This study also demonstrated that in the attack phase, the pivot leg was partly affected by increased kicking height, thus increasing the dynamic frontal plane movement (such as hip abduction and knee valgus), but not affecting the range of motion in the horizontal plane (es. Hip and knee external rotation). During this kicking process, the pivot leg was mainly responsible for pivoting and supporting the body. Before the attacking foot touches the target, the pivot hip quickly rotates and drives lower-limb rotation to prevent the pivot leg from being “locked” to the ground (Kasbparast et al., 2014b; Moreira and Paula, 2017). Therefore, before the player kicked the target, the heel of their pivot foot was already pointed forward or to the side (Kim et al., 2010b) to allow the attacking leg to smoothly execute the kick movement. However, the present study found that although a high kick does not impact the pivot leg horizontal angle; it does increase the hip and knee frontal plane angle. This shows that the player's foot was “locked” through the hip rotation and the release axis during the kick. Although the horizontal plane angle can be released, the hip and knee frontal plane angles are still relatively high. Previous scholars verified that the lower limb joints of the axial foot may increase joint mobility in the frontal plane synchronously. The purpose is to compensate for the increase in kicking difficulty (Kasbparast et al., 2014a).

This study observed that during the return, as the target height (and, by implication, kicking height) increased, the dynamic hip abduction, knee valgus and external rotation angles on the pivot leg also increased. Notably, compared with a low kick, a high kick resulted in higher dynamic knee valgus, tibia torsion angles and lower minimum knee flexion and average knee flexion angles. The increase in these processes may be related to the pivot foot return phase, which causes the body to decelerate. This deceleration force from the ground makes the planted pivot foot produce greater impact on the lower extremity kinematics (Cortes et al., 2011; Kasbparast et al., 2014a). Especially for higher kicks, this study found that there are more knee valgus and external rotation angles. This result was consistent with that of a previous study comparing the kinematics involved in three different moving tasks, as performed by Division-I soccer players (Cortes et al., 2011). The study observed that compared with a sidestep or drop-jump task, players had a lower knee flexion angle (pivot task 24.3 ± 5.7°, sidestep 38.8 ± 8.4°, drop-jump 30.3 ± 5.2°) and higher knee valgus angle (pivot task 12.0 ± 7.0°, sidestep 3.8 ± 10.0°, drop-jump 3.9 ± 8.0°) when performing a pivot task involving a 180° directional change at the moment of ground contact (Cortes et al., 2011). Generally, with a decreased angle of knee flexion and increased angles of knee valgus and external rotation, burden on the ACL increases, which increased the risk of ACL (Cortes et al., 2011; Kasbparast et al., 2014a; Norcross et al., 2010; Sam, 2010) and knee sprain in athletes (Ford et al., 2003). According to the angular data of the present study, the maximum knee valgus and average knee flexion of the pivot leg peaked at 12.2° and 16.3°, respectively, when performing a high kick, which were 1.7 degrees higher and 3.7 degrees lower than the valgus (10.5°) and flexion (20°), respectively, when performing a low kick. This result indicated that kicks increased the risk of noncontact injuries at knee joints (Cortes et al., 2011; Sam, 2010). This result also clarified the mechanisms underlying the finding of Kasbparast et al. (2014a), who discovered that common knee injuries among TKD players tended to occur when their pivot leg plants, loads and returns.

This study demonstrated that compared with low kicks, high kicks trigger low levels of hamstring and quadriceps co-activation ratio (HQCR) during the attack phase. However, at the return phase, the HQCR level of high kicks substantially increased. This was possibly because the pelvis rotation and forward inclination generated at the kicking attack phase increasing kicking performance (Gavagan and Sayers, 2017). When the quadriceps exerted force and the hamstring relaxed, lateral pelvis rotation and hip flexion were facilitated. In other words, the co-activation reaction level decreased at the attack phase to create more space for the pivot leg, thereby ensuring that high kicks were as swift as lower kicks. By contrast, at the return phase, the co-activation reaction accelerated with RHK height, indicating that HQCR enhancement was affected by the kicking height. Previous research has argued that the HQCR level increased when the pivot leg sustained a strong impact from the dynamic knee valgus, thus the pivot leg functions as a provider of joint stability and mitigator of the valgus on ACL tension effects (Norcross et al., 2010; Sam, 2010).

This previous finding also implied that players must have a strong HQCR to withstand the impact from foot planting when performing a high kick. When entering the return phase, the ability to quickly induce a strong CI_BF-RF force is crucial for knee joint stabilization (Kellis et al., 2004; Sam, 2010; Thibordee and Prasartwuth, 2014). According to this study’s observation of the activation of each muscle, decreased RF activation on the pivot leg during return was the main cause of the decrease in CI_BF-RF. Studies have also revealed that the force exerted by the hamstring can limit the forward tibia displacement when landing, thereby alleviating the burden sustained by the ACL and absorbing the landing shock (Kellis et al., 2003; Li et al., 1999; Norcross et al., 2010; Opar and Serpell, 2014; Shin et al., 2007). Yu et al. (2006) demonstrated that when the knee joint sustained pressure generated by the dynamic knee valgus, the knee joint was susceptible to injuries if the hamstring muscles did not facilitate the type of desirable contraction that provided dynamic stability. Opar and Serpell (2014) noted that stable hamstring contraction must be ensured to reduce the ACL load at the knee joint, demonstrating more generally that the hamstring muscles are essential stabilizer muscles that resist knee valgus load during high RHK deceleration. Moreover, when the knee joint was at full abduction angle, excessive activation of the quadriceps can intensify the shear force on the ACL (Brophy et al., 2010). Similarly, the present study discovered that when performing a high kick, quadriceps activation at the return phase decreased, possibly due to a protective muscle contraction aimed at lowering the burden on the knee ligaments.

This study observed that the magnitude of CI_GM-AD did not increase with the kicking height at the xiphoid position despite an evident co-activation reaction. Few studies discussed the GM activation effects on different kicking heights. One study compared soccer players of different sexes with respect to their hip motions and hip abductor activation when kicking a ball (Brophy et al., 2010). According to the results from that study, the pivot leg of female players who were susceptible to ACL injuries exhibited less satisfactory lower-extremity alignment and GM–AD co-activation reactions (Brophy et al., 2010) relative to their non-susceptible counterparts. In general, the GM at the pivot leg must provide stability during the soccer kicking process if injury was to be avoided. If such stability was not provided, hip abduction and internal rotation were likely to intensify during support, where such intensification, in turn, increased the knee twisting force (Hanson et al., 2008). Similarly, when the kicking height increased, the body tended to lean backward to increase the pelvic lateral inclination, thereby elevating the pelvis at the attacking leg (Kim et al., 2010b). This posture adjustment allowed the centre of gravity to move backward and help reduce the hip abductor burden at the pivot leg when performing a kick. The present study confirmed that GM activation was the most notable at the return phase of a moderate-height kick, indicating that hip abductor stabilization was required for kicks lower than head height to guarantee hip stability.

The implication of this study was that when players perform high kicks, their pivot leg induced a posture with greater dynamic knee valgus, knee straightening, and a high level of requisite HQCR (for lower-limb joint stabilization). TKD training should focus on hamstring training, where a strong hamstring might form the basis of injury-mitigating high-RHK technique (Opar and Serpell, 2014). Like FIFA 11+, an injury-prevention program designed for soccer players, involved exercises that progressively induced lower-limb neuromuscular functions (Daneshjoo et al., 2012; Monajati et al., 2016); FIFA 11+ has been proven to effectively prevent knee joint injuries (Barengo et al., 2014; Grooms et al., 2013). Accordingly, this study argues that a standard TKD warm-up program should be designed to condition the hamstring muscles, thereby reducing the risks for those injuries to the knee joints caused by pivot leg planting and spinning. Scholars have made similar arguments (Park and Song, 2018) involving a training regimen that facilitates knee injury prevention when kicking in TKD.

This study was limited in that the height of the kick, the displacement distance, and the pivoting foot, were primarily controlled by the athlete. However, a mark on the ground was used to confirm the sliding distance and the height of the kick target was based on the anatomical position of human body. Nonetheless, the situation was still potentially different from the kicking performance of athletes facing real opponents. In addition, the experimental design and laboratory-based nature of the study may have caused the athlete's performance to be over-regulated and distorted, and thereby not representative of the real phenomenon during competition (Gavagan and Sayers, 2017).

Furthermore, another limitation was that only high-level TKD athletes were recruited in the study. Recent studies have found that differences in the level of professional knowledge of TKD athletes were inconsistent with specific neuromuscular responses and technical aspects of high kicks (Moreira et al., 2018). Finally, the the recruitment and analysis of different genders in single group may have influenced the study’s findings. Subjects of different genders may have inconsistent lower-extremity neuromuscular control and knee movement behaviours when performing landing tasks (Brophy et al., 2010; Hewett et al., 2005) and the injury mechanism of female TKD athletes may be different from that of males (Yalfani et al., 2019). Therefore, further research should be done to explore the impact of genders on these currently examined outcome variables.

Elite TKD athletes perform head position kicks with greater hip and knee joint displacement in the pivot leg as compared with other positions, especially during the return period. Maximum dynamic knee valgus angle close to 12 degrees may be considered dangerous in the return phase. The hip hamstring and quadriceps co-contraction neuromuscular reaction decreased in executing head position kicks greater than other lower positions during the attack phase. In the return phase, the co-contraction increased significantly. The results showed that the knee joint load increased when players execute head position kicks in the return phase.