Biomechanics of lower limb in badminton lunge: a systematic scoping review

Search strategy

Searches were designed and conducted by DWCW. Electronic literature searches of electronic databases included ISI Web of Science (excluding patents, from 1970), Scopus (from 1960), SPORTDiscus (from 1830), and PubMed (from 1975) was performed on 3 September 2020. The searches were made using a combination of the following keywords linked by the AND function: “badminton”, “lunge”, and “biomechan* OR kinematics OR kinetics” in the topic field of the databases. The title, abstract, and full-text of the included articles were screened based on the following inclusion criteria: (1) published in English; (2) in a peer-reviewed journal; (3) conducted experiments to investigate the badminton lunge manoeuvre of the lower limb; (4) biomechanical investigation in nature.

Screening and study selection

The authors (DWCW and WKL) conducted screening on the abstract and full-text and subsequent data extraction. Any disagreements were resolved by discussion with the third author (WCCL), and all authors conducted a final check of the review.

An initial search identified 86 articles, and 54 duplicates were removed. Four articles were identified from grey literature. Fifteen articles were excluded because of non-original research articles, including review and survey only (n = 6); non-English article (n = 2); not related to lower limb biomechanics of lunge manoeuvre (n = 5); conducted on participants with pain or pathological conditions (n = 3). Finally, there were 20 articles eligible for inclusion after screening.

The search and study selection process is summarized in Fig. 1. Studies were excluded if they did not constitute any biomechanical outcome measures. There was no disagreement between authors in the selection of studies eligible for the review. The following data were extracted and grouped into the research context: bibliographic details, sample size, characteristics of participants, inclusion and exclusion criteria of studies, experimental settings, and outcome measures.

Data extraction

Information of the 20 included studies qualified for further review is summarized based on the review context in Table 1 (population and selection criteria), Table 2 (experimental protocols), and Table 3 (outcome measures and key findings).

Methodological quality assessment

The methodological quality of each study was assessed using the modified Downs and Black Quality Index Tool (Downs & Black, 1998). The scale consists of 27 questions evaluating the quality of reporting, external and internal validity, and statistical power. It was proven to demonstrate high internal consistency (Downs & Black, 1998). In this study, seven questions (items 8, 9, 13, 17, 19, 25, 26) were not applicable due to the study design of this review (Richmond et al., 2013). Additionally, the power subscale (item 27) was discarded due to ambiguity, as recommended in the literature (Deeks et al., 2003). Therefore, a 19-item scale was used to evaluate the study quality.

The quality test was conducted by the authors (DWCW & WKL) independently. Any disagreements found were resolved by discussion with the third author (WCCL). The authors agreed to refine and remark some items, as stated in Table 4, to avoid ambiguity. The findings of the methodological quality assessment were addressed in the discussion section.

Participant characteristics

There were a total of 314 participants (264 males, 50 females) in the 20 included articles (Table 1). Sixteen of them included male participants only. One article investigated only female participants. Only one article had participants with male-to-female at a one-to-one ratio (Lam et al., 2018). The sample size of each subject group ranged from 4 to 17. All participants were young healthy adults with an average age ranged from 19.9 to 27.1 years old. Participants were generally excluded from musculoskeletal problems, previous or current injuries, or surgeries. Moreover, some included articles required the participants to have an uncorrected vision and free of medications and alcohol. The skill levels of participants were classified based on the years of experience and levels of badminton competitions.

Experimental protocol

Lunge manoeuvres can be performed in five directions, including three forward directions (forehand, backhand, and in-front) and two backward directions (forehand and backhand). Almost all included articles examined lunge biomechanics in forward direction, while one did not specify the direction of lunge (Nagano et al., 2020). Three included articles considered backward lunges, and six compared the movement mechanics among lunge directions. Hong et al. (2014) examined the differences in the impact of vertical and horizontal ground reaction forces (GRF) among lunges in four diagonal directions, while some included articles compared forehand and backhand performance in forward direction (Chen et al., 2020; Nielsen et al., 2020; Valldecabres et al., 2018). Moreover, Hu et al. (2015) compared the plantar loading pattern amongst three forward lunges (forehand, backhand, and in-front).

Variation in lunge manoeuvre

The kick lunge, consisting of three steps, was commonly investigated in badminton studies on lower limb biomechanics (Table 2). The first step was to start from the initial position in standing or before two running steps. The next step was to launch the lunging step striking on the force plate and hit the target (shuttlecock). Two included articles required the participants to explicitly extend the knee (i.e., lunging leg) in the second step to achieve the maximum lunge distance (Lam, Ding & Qu, 2017; Lam et al., 2017), while one included article evaluated the sensitivity of ankle abduction angle (Choi & Lee, 2017). The third step (recovery phase) was to return to the initial position as fast as they could in gameplay. Before the experiment, the coaches (or elite athletes) who were able to consistently reproduce the manoeuvre demonstrated the footwork and placement to less skilled participants. In addition, the badminton coach and the experimenters confirmed whether a trial was successive and valid (Hong et al., 2014; Hu et al., 2015).

Furthermore, there were some other testing requirements when performing the lunge manoeuvre. Lam et al. (2018) controlled the total completion time of the move (forward lunge + recovery) within 3 s. The lunge distance between the starting position and force plate was determined as 2.5 times the foot length during a forehand forward lunge and three times of the foot length during a forehand backward lunge, respectively (Hu et al., 2015).

Three included articles investigated different variations of lunges. Nadzalan et al. (2017) asked the participants to perform both step-lunge and jump-lunge, which were achieved by bending the trunk anteriorly at 45° with thigh maintained horizontal, and by jumping explosively during the lunge and return, respectively. In addition, Kuntze, Mansfield & Sellers (2010) compared the step-in and hop lunges to the kick lunge. Step-in lunge was characterized by pulling the non-dominant limb towards the dominant limb at recovery to raise the body, while the hop lunge incorporated a hop after the strike and before the recovery. Lam, Ding & Qu (2017) highlighted that athletes were always moving back and forth during a game, and considering a single isolated lunge may not be realistic, despite that there were challenges to define and segment phases and events of movement accurately. Nagano et al. (2020) considered the more realistic aspect of the game. They extracted and classified the frequency of different movements in the game that included landing after an overhand stroke, lunging during an underhand stroke, cutting from a split step, take-off before the overhand stroke, pre-split steeping, split step, stopping, backstepping, and cutting. One of the drawbacks was that the movement could not be controlled, and the directions of the lunge were not specified to allow comparison across studies.

Shuttlecock position/task

The assignments of the hitting target (shuttlecock) were also different among studies, which might alter both upper and lower limb biomechanics of a lunge. During the experiments, a shuttlecock was often suspended at 0.6 m height as the hitting target (Lam, Ding & Qu, 2017; Lam et al., 2018; Lam et al., 2017; Nielsen et al., 2020; Valldecabres et al., 2018). On the other hand, Kuntze, Mansfield & Sellers (2010) and Mei et al. (2017) manually threw the shuttlecock over the net to a dedicated court area. Five included studies did not mention the use of shuttlecock nor address how it was positioned. Athletes were asked to perform an underhand stroke to resemble a frontcourt shuttle-drop or drop-shot serve conditions. More precisely, Lam, Ding & Qu (2017) instructed the participants to grip the racket backhand or thumb-up in front of the body and executed a drop-shot by lifting motion that only allowed a small amount of shoulder follow-through (Grice, 2008).

Footwear construction

The influence of footwear construction is undeniable since it is associated with the plantar loading pattern and landing kinematics during lunges. Some included articles did not report if footwear was standardized. Valldecabres et al. (2018) reported that their participants wore their own badminton shoes, whereas Fu, Ren & Baker (2017), Mei et al. (2017), and Huang et al. (2019) provided the same brand and series of badminton shoes used in the experiments. Several included articles utilized the badminton shoes from the Li Ning (Hu et al., 2015; Lam, Ding & Qu, 2017; Lund et al., 2017; Park et al., 2017), Mizuno, and Yonex company (Hu et al., 2015; Wei et al., 2009). Only one included article attempted to compare between barefoot and shod conditions that introduced a prototype designed for the foot shape of Asian players (Wei et al., 2009).

Isolated footwear design/constructions were also investigated in heel curvature design (Lam et al., 2017), midsole thickness (Lin et al., 2016), sole hardness (Mei et al., 2014), insole hardness (Lund et al., 2017), and forefoot bending stiffness (Park et al., 2017). Lam et al. (2017) examined the effect of the geometry of shoe heel curvatures (rounded, standard, and flattened heel) during lunge in elite and intermediate athletes. Lin et al. (2016) hypothesized that thicker midsoles might decrease joint instability and performance and thus tested this contention with shoes with three different midsole thickness. Moreover, Lund et al. (2017) divided the insoles into three plantar regions (rearfoot, midfoot, and forefoot) to further investigate the effect of the forefoot and rearfoot hardness on joint mechanics during lunges. While keeping the midfoot region as medium hardness, Lund et al. (2017) evaluated the insoles with five combinations of regional hardness with three hardness levels (hard, medium, and soft). Regarding midsole hardness, Mei et al. (2014) compared two hardness of soles (58 vs 68) using a Durometer (Type C) on lower limb biomechanics. On the other hand, Park et al. (2017) reduced the forefoot bending stiffness of a shoe by reducing the midsole thickness by 2 mm, while increased the stiffness by filling the outsole flexing grooves. In summary, some reviewed articles did not provide details on the selection of footwear in the experiment, while some attempted to use commercially available badminton shoes. For articles comparing footwear construction factors, midsole thickness, midsole hardness, and heel curvature were of concern in badminton lunge.

Outcome measures

Differences in playing levels

Five included articles investigated biomechanical differences between playing levels. While both amateur and professional badminton athletes showed similar maximum lunge distances, professional athletes demonstrated larger footstrike angles and faster approaching speeds that could characterize as better performance (Lam et al., 2017). Better heel impact attenuation during the landing of a lunge step was another attribute of professional athletes, which is demonstrated by smaller peak horizontal GRF and loading rates, although comparable vertical GRF was noted compared to the amateurs (Lam et al., 2017). Another included article from the same research group indicated that professional athletes produced significantly smaller peak vertical and horizontal GRF, whilst loading rates were mainly governed by the gender factor (Lam et al., 2018). Professional athletes landed at the lateral heel of the lunging foot and gradually shifted plantar load to the medial forefoot and hallux regions from landing to the drive-off phase (Kuntze, Mansfield & Sellers, 2010). They showed significantly higher peak pressures in the medial forefoot and hallux regions, and higher force-time integrals in the medial forefoot. This was compared to amateur athletes who presented significantly higher peak pressures and force-time integrals in the lateral forefoot and lesser toe regions (Mei et al., 2017). Abnormal high loading at the lateral side of the foot might indicate an inverted ankle landing posture, which could be a potential contributor for ankle inversion sprain and associated ligament injuries (Shariff, George & Ramlan, 2009).

Professional athletes generally elicited greater mechanical outputs at the knee and ankle that aided to transfer body mass effectively (Fu, Ren & Baker, 2017). The professional players demonstrated greater knee joint moments in both sagittal and frontal planes (Fu, Ren & Baker, 2017; Huang et al., 2019; Lam et al., 2017), greater peak knee adduction, and internal rotation angles (Mei et al., 2017). To allow quick recovery from lunges, athletes exhibited smaller knee and hip flexion angles, which were associated with shorter stance times (Brahms, 2014; Mei et al., 2017). On the other hand, interestingly, opposite findings indicated that professional athletes demonstrated both larger (Lam et al., 2017) and smaller (Lam et al., 2018) peak knee flexion moment and impact loading rate. The former was explained by greater power production while the latter was explained by better efficiency in elite athletes. Similarly, there are conflicting results related to ankle joint variables across included studies. Fu, Ren & Baker (2017) found that amateurs produced greater ankle inversion and internal rotation angles, whereas Mei et al. (2017) found amateurs had smaller peak ankle inversion and ankle internal rotation, but larger peak plantarflexion angle and RoM. Huang et al. (2019) found that professional athletes produced greater ankle eversion moment during weight acceptance phase. Poorer muscle control yielding ankle instability may lead to a larger range of motion (Fu, Ren & Baker, 2017). In contrast, Mei et al. (2017) and Huang et al. (2019) attributed the findings due to poor landing techniques and fatigue of their participants. Valldecabres et al. (2018) showed that fatigue caused a significant reduction in knee angular velocity during heelstrike, range of knee angular velocity in the coronal plane, and knee abduction moment. Discrepancies could be due to the differences in specific tasks, difficulty, or instructions. For example, the direction of lunges, repeated or isolated lunges (Lam et al., 2018; Lam et al., 2017), instructions on whether the lunge would perform maximally or replicate according to a model participant (Fu, Ren & Baker, 2017; Mei et al., 2017). Further investigation is warranted, and the interaction with footwear could be a plausible explanation (Lam et al., 2017).

Influence of footwear

Our included articles did not demonstrate any promising evidence of the influence of midsole hardness or forefoot bending stiffness on lunge biomechanics. A softer midsole is suggested to provide better cushioning, lower peak vertical GRF, and longer landing time, but some included articles did not show significant differences among various midsole hardness conditions (Mei et al., 2014). Similar findings were obtained from a computational model investigating different insole hardness (Lund et al., 2017). Moreover, different shoe bending stiffness did not impose observable alteration on the ankle and foot kinematics, even though perceived cushioning can be altered (Park et al., 2017). It could be explained by “comfort filter” and “preferred movement path” paradigms, which suggested that the human body could unconsciously select comfortable footwear based on their own perception to allow for individual preferred movement path (Nigg et al., 2015). Athletes could optimize muscle forces to complete the lunge movement successfully and to reduce excessive impact forces (Lees & Hurley, 1994).

Compared to the barefoot condition, wearing sports shoes reduced metatarsophalangeal joint flexion and elongated length of plantar muscles that may enhance push-off efficiency (Wei et al., 2009). However, different badminton shoe models used did not significantly change the footstrike angles, GRF, and plantar load distribution (Hong et al., 2014; Hu et al., 2015). On the other hand, Lin et al. (2016) found the shoes with thicker midsoles were unable to influence shock attenuation, but in another study, thicker midsoles were found to be associated with impaired proprioception (Robbins et al., 1994), leading to higher peak midfoot pronation at the early stance that may be related to overuse injuries. In a similar vein, heel geometry design was identified to be one plausible construction to improve impact attenuation capacity during lunge (Lam et al., 2017). Rounded heel shoes could reduce maximum vertical loading rates compared to flat or standard heels shoes. However, such a situation was only applicable to elite athletes but not to amateurs, which may imply that professional athletes were more sensitive and adapted to very small changes in footwear. In summary, softer or thicker footwear may not be effective in shock absorption despite providing better comfort perception. The impact of forefoot bending stiffness and heel curvature on lower limb biomechanics may be influenced by some confounding factors, such as the level and accommodation capability of the athletes. Future studies should investigate different material and structural designs that could compromise impact attenuation, proprioception, and performance.

Lunge directions and variations

The left (backhand/non-racket side) forward lunge is suggested to be the most critical as this lunge direction had the highest plantar loading than other lunge directions (Hong et al., 2014). Left forward lunge produced significantly higher first and second peaks of vertical impacts compared to backward lunges, and higher maximum anterior-posterior shear forces compared to left backward lunges (Hong et al., 2014). Backhand forward lunges were also related to larger trunk rotation, greater demand in core control and dynamic postural stability compared to forehand forward lunges (Lin et al., 2015). However, compared to backward lunge, Chen et al. (2020) reported that forward lunge produced higher compressional ankle contact, faster touchdown hip abduction, and larger horizontal deceleration of the mass center and torso. Among forward lunges, participants performing in-front forward lunge experienced significantly smaller load on hallux but a larger load on the lateral midfoot (Hu et al., 2015). However, it should be noted that performing lunge in a single movement would be biomechanically different from that of the repeated movements. Participants performing repeated lunge demonstrated shorter foot contact time and smaller impact, peak knee anteroposterior force, peak knee sagittal movement, but larger peak horizontal GRF (Lam, Ding & Qu, 2017). The differences in movement strategies may allow for more consistent execution of consecutive ballistic movements (Cormack et al., 2008).

Among the three lunge variations (kick, step-in, and hop), step-in lunge was considered as a more energy-saving/efficient move, since involvement of the non-lunging (non-racket side) leg aided in support and reduced GRF of the leading foot (Kuntze, Mansfield & Sellers, 2010). It also generated less hip joint power and, thus, a smaller contribution to the recovery phase (Kuntze, Mansfield & Sellers, 2010). In contrast, hop lunge produced significantly higher joint power output for quicker movement, as indicated by larger GRF during loading response and drive-off phases, larger peak ankle moment, and ankle and knee joint powers (Kuntze, Mansfield & Sellers, 2010). The greater plantarflexor power was associated with improved mechanical efficiency by enhancing the muscular stretch-shortening cycle and subsequently dissipating energy to recover back to the initial position quickly (Kuntze, Mansfield & Sellers, 2010).