SIDESTEP AND CROSSOVER LOWER LIMB KINEMATICS DURING A PROLONGED SPORT‐LIKE AGILITY TEST

Abstract

Background:

Non‐contact anterior cruciate ligament (ACL) injuries in athletes occur more often towards the end of athletic competitions. However, the exact mechanisms of how prolonged activity increases the risk for ACL injuries are not clear.

Purpose:

To determine the effect of prolonged activity on the hip and knee kinematics observed during self‐selected cutting maneuvers performed in a timed agility test.

Methods:

Nineteen female Division I collegiate soccer players completed a self‐selected cutting agility test until they were unable to meet a set performance time (one standard deviation of the average baseline trial). Using the 3D dimensional coordinate data the cut type was identified by the principle investigators. The 3D hip and knee angles at 32ms post heel strike were analyzed using a two‐factor, linear mixed model to assess the effect of prolonged activity and cut type on the recorded mean hip and knee angles.

Results:

Athletes performed either sidestep or crossover cuts. An effect of cut type and prolonged activity was seen at the hip and knee. During the prolonged activity trials, the knee was relatively more adducted and both the hip and knee were less flexed than during the baseline trials regardless of cut type. Regardless of activity status, during sidestep cuts, the hip was more internally rotated and abducted, and less flexed than during crossover cuts while the knee was more abducted and less flexed during the sidestep than crossover cuts.

Conclusions:

During a sport‐like agility test, prolonged activity appears to predispose the athlete to position their knee in a more extended and abducted posture and their hip in a more extended posture. This position has been suggested to place stress on the ACL and potentially increase the risk for injury. Clinicians may want to consider the effects of prolonged activity on biomechanical risk factors for sustaining ACL injuries in the design of intervention strategies to prevent ACL injuries.

Level of Evidence:

Level 4

INTRODUCTION

Anterior cruciate ligament (ACL) injuries are one of the most common injuries that occur in competitive sports.^1,2 Greater than seventy percent of all ACL ruptures are caused by non‐contact mechanisms, meaning that there is no external force causing the injury.^2‐9 A high incidence of these injuries are sports related, affecting female athletes two to ten times more frequently than males, incurring significant economic costs.^3,4,8‐13 Both the long‐term physical consequences and high economic costs of ACL injuries point to the importance of achieving a better understanding of the mechanism of injury and development of effective preventive programs.

While there are many factors associated with the increased risk of females sustaining an ACL injury two potentially modifiable risk factors are lower limb biomechanics and fatigue.^{2,6,12,14‐16} Athletes are placed at higher risk for non‐contact ACL injury when performing a sudden change in direction and may be at an even greater risk when muscle fatigue is involved during the maneuver.^17‐19

Non‐contact injuries most commonly occur during landing, pivoting (at or near full knee extension), deceleration, and change of direction maneuvers during play.^3,8,9,12 Through the observation of ACL injury mechanisms in basketball players, Ireland et al have described the position in which the ACL is at the greatest risk of rupture: the hip positioned in adduction, internal rotation, and decreased flexion, and the knee positioned in valgus, external rotation, and decreased flexion.^8,9,13

Different cutting maneuvers have been shown to place the lower limb at varying levels of risk for sustaining an ACL injury.^4,20 In particular, a sidestep cut appears to put the athlete at the greatest risk of injury.^5,21 Cochrane et al²¹ defined sidestep cuts as accelerating toward the direction opposite of the planted leg (Figures 1a, 1b). In contrast, a crossover cut is defined as crossing one leg over the planted leg and accelerating in the same direction of the push off leg (Figures 1c 1d).

Figure 1.

The two types of cuts self‐selected by the athletes during the performance of the agility test. 1a and 1b show the Sidestep cut, and 1c and 1d show the Crossover cut.

Prolonged activity may alter lower limb biomechanics, cutting strategies, and the ability of an individual to maintain trunk control.^{17‐19,22‐26} Furthermore, prolonged activity can affect joint proprioception and neuromuscular responses, thereby decreasing the ability to sense joint position and diminishing the protective roles of muscles.^27,28 Due to these factors the ability to maintain motor control, when planning for precision movements, such as cutting, may be impaired, possibly leading to a decrease in dynamic joint stability and failure to prevent the joint from going into a high‐risk injury posture.²⁷ The inability to maintain muscular control may explain why injury rates tend to increase as athletes fatigue during games.^29‐32 Furthermore, studies that have focused on biomechanics during prolonged activity have used scenarios that do not mimic game‐like environments.^18,26,29,33 Therefore, to fully understand the effect of prolonged activity and its impact on the risk for ACL injury, the use of testing protocols that simulate a game‐like environment should be considered.

Using an agility test, the authors of this study have developed a paradigm that allows athletes to perform self‐selected cut types under different conditions. This paradigm allows for an evaluation of lower limb biomechanics during a repetitive activity that mimics the physical demands of competition. Therefore, the purpose of this study was to determine the effect of prolonged activity on the kinematics of the hip and knee while the subjects performed a self‐selected cutting maneuver during a timed agility test that closely mirrors the demands of soccer. Prolonged activity was hypothesized to have a significant effect on the kinematics of the hip and knee during a timed agility test, and these prolonged activity‐related changes could put athletes in a more at‐risk position for a non‐contact ACL injury.

METHODS

Subjects

Twenty‐one female NCAA, Division I collegiate soccer players were recruited to participate in this study. Of the 21 subjects, 19 completed the study. Two subjects did not complete the study due to sustained injuries unrelated to the study. The subject demographics of the 19 who completed the study are presented in Table 1. To be included in this study, subjects had to be members of the women's Quinnipiac University soccer team, free of injury at the time of enrollment in the study, and for at least 1 year prior to commencement of the study, and not have any musculoskeletal, cardiovascular, or neuromuscular condition that was contraindicated for participation in this study. Subjects were excluded from the study if they did not meet these inclusion criteria. Eight subjects had a prior history of lower extremity injury prior to the study, reported in Table 1. All experimental procedures were approved by Quinnipiac University's Institutional Review Board.

Table 1.

Select demographics of the subjects who completed the study (n =; 19) and mean number of trials required to complete the agility test protocol.

Characteristic	Mean ± SD
Mean Age, years (± 1 S.D.)	19.4 ± 1.3
Mean Height, cm (± 1 S.D.)	164.3 ± 5.6
Mean Mass, kg (± 1 S.D.)	62.2 ± 7.0
Number of subjects with prior history oflower extremity injury*	8
Mean number of agility trials completed by subjects	20.6 ± 9.4

*1 subject‐ bilateral ACL tears and reconstruction, 1subject‐medial collateral ligament tear, 1 subject‐meniscal tear, 3 subjects‐tibial or fibular fracture,2 subjects‐multiple lower extremity injuries..

Procedures

Data collection occurred over two sessions, approximately one week apart for each subject. At the first session, informed consent was obtained after which demographic and anthropometric data consisting of age, height, weight, and previous lower extremity injury history were recorded. At the first session an agility test training session was then provided for each subject. At the second data collection session the agility testing protocol consisted of having the subject complete a modified T‐Test, repetitively. The T‐Test required the subject to run through a T‐shaped obstacle course (Figure 2) as fast as possible utilizing a self‐selected cut type. The modified T‐Test has been shown to be a reliable measure of agility but has not been utilized to simulate sport during biomechanical analysis.³⁴ The training session consisted of an explanation of the agility test protocol, a demonstration of the protocol and an opportunity to run through the obstacle course performing both left as well as right cuts. During this practice session, the subjects were instructed to run through the course, each time through increasing their speed, until they were running at their maximum speed. On average the subjects performed 4 practice runs from each side through the course.

Figure 2.

Overhead view of the agility test with the subject starting on the left. The arrows depict the path the subject took. The dark grey arrow represents the first turn analyzed and the light grey arrow represents the second turn analyzed in this study for each trial.

At the second data collection subjects warmed up for 10 minutes while jogging on a treadmill at a self‐selected pace. After completing the warm‐up, a full body Cleveland Clinic marker set (Figure 3) was applied to the subjects using double‐sided adhesive tape. Once all the markers were in place, the subjects were asked to stand still in the middle of the data collection volume with their feet approximately 10 inches apart. The subjects then completed the agility test protocol running at maximum speed for all trials, while video data was recorded using a 10 camera 3‐dimensional motion analysis system (Motion Analysis Corporation, Santa Barbara, CA) sampling at a frequency of 240 Hz.

Figure 3.

Anterolateral view of a subject with the full body Cleveland Clinic marker set in place. The marker set consists of four rigid marker triads attached to the distal ends of the femurs and lower legs, single markers placed bilaterally over the midpoint of the acromions, lateral humeral epicondyles, midpoint between the radius and styloid processes, midpoint of the heel counter of the shoes, and over head of the second metatarsals. Four additional makers placed bilaterally over the medial and lateral femoral epicondyles and malleoli were placed during the static trials to establish the location of the hip, knee, and ankle joint centers and removed during the dynamic trials.

A timed based decrement in performance model was used to determine the endpoint of the agility test protocol. For the first four trials, the subject had a 60 second rest period between each run. The time for these runs were recorded and averaged and were considered to be the baseline trials. The subject then continued to perform the agility test, continuing to run at maximum speed, while alternating from all left turns and all right turns beginning every 30 seconds (includes rest as well as trial, Figure 2). The subject continued the test until they failed to have two consecutive runs within one standard deviation of the established average baseline time for the trials. The subject was then asked to perform two more runs, one from each side, during which video data was recorded. These last two trials were considered the prolonged activity trials. The average number trials of (± 1 S.D.) to complete the test protocol are reported in Table 1. The subject's time to complete the obstacle course was recorded using a timing gait device (Equine Electronics Timing System, Equine Electronics).

Data Analysis

The X, Y, Z coordinate histories of all of the retro reflective markers were obtained using a commercially available software package (Cortex, Motion Analysis Corporation, Santa Barbara, CA). The coordinate histories were then filtered using a zero lag, 4^th order, low pass Butterworth Filter with a 10 Hz cutoff. The leg making the cut and the type of cut were identified and recorded for each trial using the video data.

The trunk and upper and lower extremities were modeled as an eleven‐segment rigid body system with 10, six degree‐of‐freedom joints interspersed between the segments. A commercially available software package (Orthotrak, Motion Analysis Corporation, Santa Barbara, CA) was employed along with the kinematic model and the coordinate histories of the static and dynamic trials to obtain the three dimensional angular displacement histories for the hip and knee joints bilaterally using an Euler decomposition method (Z,Y,X). The stance phase of the extremity making the cut was extracted and the angular position of the hip and knee joints at 32 ms post foot contact was determined. The time point 32 ms post foot contact was chosen for analysis because previous evidence suggests that ACL injuries occur between 20 and 40 ms post foot contact.³³

Statistical Analysis

For each subject, the first two base line trials and the two prolonged activity trials were chosen for all statistical analyses. The frequency of the different cut types performed by subjects at the center cone (Figure 2) in the completion of the agility test was grouped according to activity status (baseline versus prolonged). A Chi‐Square analysis was conducted to determine if cut type was associated with prolonged activity status. The angular position data was then grouped according to cut type and prolonged activity status and descriptive statistics consisting of means and standard deviations obtained. A two‐factor linear mixed model with fixed effects for cut type and prolonged activity, and a random effect for intercept (unconditioned mean) was employed to determine the effect of cut type and prolonged activity status on the three‐dimensional angular position of the hip and knee joints. The alpha level of significance for this study was set at the 0.05 level. SAS v9.3 (SAS, Cary, NC) was used for all statistical analyses.

RESULTS

During the agility test, subjects performed either a sidestep or crossover cut. In both the baseline and prolonged activity states, the sidestep cut was performed more often than the crossover cut (Figure 4). Within a prolonged activity condition, the mean times to complete the entire agility test were similar for the two cut types (7.1 ± 0.4 and 7.1 ± 0.4 s for the baseline sidestep and crossover cuts, respectively and 7.2 ± 0.3 and 7.5 ± 0.3 s for the prolonged sidestep and crossover cuts, respectively). Subjects completed an average of 20.6 trials (± 9.4) to complete the agility testing protocol (Table 1). Compared to the baseline trials, athletes after prolonged activity performed the sidestep cut at a higher frequency than the crossover cut; however, this difference was not statistically significant (p=;0.074, Figure 4).

Figure 4.

The number of cuts according to prolonged activity status and cut type.

An effect of prolonged activity was present for hip joint flexion and extension (F_1,57 =; 7.11, p =; .010). The hip was more extended after prolonged activity than during the baseline trials (Figure 5). Additionally, an effect of prolonged activity was seen for abduction and adduction (F_1,60 =; 6.07, p =; .017), and flexion and extension (F_1,131 =; 5.90, p =; .016) of the knee. During prolonged activity, the knee was less abducted and more extended than during the baseline trials (Figure 5).

Figure 5.

The three dimensional hip and knee angles of the stance limb lower extremity at 32 ms post foot contact according to prolonged activity status. Positive numbers represent hip and knee joint flexion, adduction and internal rotation.

An effect of cut type was seen for internal and external rotation (F_1,118 =; 90.38, p < .001), abduction and adduction (F_1,135 =; 430.9, p < .001), and flexion and extension (F_1,139 =; 10.89, p =;.001) of the hip. The hip was more internally rotated, abducted, and extended during a sidestep cut than in the crossover cut during which the hip was more externally rotated, adducted, and flexed (Figure 6). At the knee, an effect of cut type was seen for abduction and adduction (F_1,109 =; 30.18, p < .001) and flexion and extension (F_1,148 =; 25.10, p < .001). During the sidestep cut, the knee was abducted and more extended than during the crossover cut (Figure 6). An interaction effect of cut type by prolonged activity status was not present at either joint for any of the joint motions. To determine if past injury affected the above findings all of the models were rerun with injury included as a factor, and there was no significant effect for injury and injury did not interact with either cut type or prolonged activity on any of the outcomes. While not statistically significantly different, subjects performing the sidestep cuts had more hip internal rotation and less hip and knee flexion after prolonged activity. For the crossover cuts, subjects had less hip external rotation, adduction and flexion and less knee flexion after prolonged activity (Table 2).

Figure 6.

The three dimensional hip and knee angles of the lower extremity of the stance limb 32 ms post foot contact according to type of cut. Positive numbers represent hip and knee joint flexion, adduction and internal rotation.

Table 2.

The three dimensional hip and knee joint angles (degrees) of the lower extremity of the stance limb 32 ms post foot contact according to activity state (baseline or prolonged) and cut type. Positive numbers represent hip and knee joint flexion, adduction and internal rotation. The plus or minus one standard deviation are noted in the parentheses (± 1 S.D.)

	Flexion/Extension* (Degrees)		Hip Joint Abduction/Adduction* (Degrees)		Internal/External Rotation* (Degrees)		Flexion/Extension* (Degrees)		Knee Joint Abduction/Adduction* (Degrees)		Internal/ External Rotation* (Degrees)
	SS†	CO	SS†	CO	SS†	CO	SS†	CO	SS†	CO	SS†	CO
Baseline	38.1 (± 10.1)	46.4 (± 11.1)	‐12.7 (± 6.3)	7.2 (± 5.3)	3.7 (± 9.6)	‐11.3 (± 8.8)	36.9 (± 12.1)	51.9 (± 20.9)	‐3.7 (± 5.0)	0.7 (± 4.5)	‐7.9 (± 8.8)	‐5.4 (± 8.6)
Prolonged Activity	34.3 (± 12.3)	40.4 (± 18.0)	‐12.9 (± 5.9)	6.3 (± 7.1)	6.2 (± 11.0)	‐9.1 (± 14.3)	32.3 (± 10.5)	44.9 (± 19.6)	‐1.3 (± 6.5)	2.6 (± 6.5)	‐9.3 (± 13.2)	‐10.2 (± 15.8)

^*Positive numbers represent hip and knee joint flexion, adduction and internal rotation,

^†‐ Sidestep, ‐ Crossover

DISCUSSION

In agreement with the authors' stated hypothesis, prolonged activity placed athletes in a more at risk position. Regardless of which type of cut the subjects performed during the prolonged activity trials, subjects demonstrated less hip and less knee flexion when cutting. This more erect posture during cutting and landing maneuvers has been associated with a greater risk for an ACL injury.^1,3,5,13,17 Various researchers using video analysis and observations from the field have suggested that ACL injuries occur when the knee is in a less flexed state, with knee flexion angles ranging from 0° to 40°.^1,35,36 In this position, the ACL may be subjected to larger anterior shear forces than at a more flexed position.³⁷ In addition to decreased knee flexion angles, decreased hip flexion angles during landing and cutting maneuvers may increase the risk of non‐contact ACL injuries.³

In previous studies, there have been inconsistencies among the effects of prolonged activity on knee flexion angles. Several studies have noted a decrease in knee flexion combined with valgus and tibial external rotation which is similar to the findings of this study, seen during the sidestep cutting prolonged activity trials.^17,19,38 A similar study conducted by Lucci et al²² demonstrated that subjects performing an unanticipated sidestep cut after prolonged activity had lower knee and hip flexion angles compared to their baseline trials. In contrast to these findings, Tsai et al³⁹ found no difference in maximum knee flexion angles between prolonged activity and baseline trials during sidestep cutting. A limitation of these previous studies investigating the effects of prolonged activity and cutting kinematics is that the prolonged activity protocol utilized typically consisted of activities that did not mimic competition such as shuttle running, vertical leap jump, or plyometrics. In the present study, the prolonged activity protocol mimicked cutting maneuvers that subjects perform during competition to attempt to better understand how prolonged sport‐specific activity potentially alters cutting kinematics. To the authors' knowledge, this is the first study to use the same activity for both the prolonged activity protocol and for the evaluation of kinematic changes. By having the subjects use the agility test for the prolonged activity protocol and by evaluating how they cut during the performance of the agility test, it ensures that the task the subjects are performing directly relates to fatigue and mimics a common sport‐specific movement pattern.

A unique feature of the present study was that subjects had the ability to choose which type of cut they performed during the agility test. This allowed for a direct comparison of the kinematic differences between these two cut types and if prolonged activity resulted in any kinematic changes. In the present study, cut type affected lower extremity kinematics regardless of prolonged activity status. At the hip, subjects who performed a sidestep cut demonstrated more hip extension, abduction, and internal rotation, and knee extension and abduction than with the crossover cut. Observations by Besier et al²⁰ support the current findings that sidestep cuts elicited greater knee abduction (valgus) and less knee flexion than crossover cuts. The fact that athletes who performed a sidestep cut versus a crossover cut positioned themselves in a more extended and abducted posture suggests that they may be at a greater risk for sustaining an ACL injury than when not in this lower extremity posture. Several studies have shown that during landing and cutting maneuvers, an increase in the knee abduction moments (valgus torques) and knee abduction angles are potentially the most important mechanism for non‐contact ACL injuries.^17‐19

While sidestep cuts may be considered to be more “risky cuts” they were also the most frequently performed cut used by the subjects, regardless of prolonged activity status. While there were no differences in agility times (time to complete the test) between athletes in the present study performing a sidestep or crossover cut, it is possible, that sidestep cuts were chosen by athletes because they may be the more preferred cut from a performance standpoint.⁴⁰ Therefore, rather than teaching athletes to alter their cutting style future studies aimed at preventing ACL injuries might be best served by allowing athletes to perform sidestep cuts in a more controlled and safe manner that avoids dynamic knee valgus and extension.

In addition, the self‐selection of cut type also allowed for an analysis of the effect of prolonged activity on the decision to perform a crossover or sidestep cut. Although not significant, it is interesting to note that subjects during the prolonged activity trials executed sidestep cuts more frequently than crossover cuts as compared to the cuts chosen by athletes during the baseline trials. Therefore, athletes who changed their cut type from a crossover to a sidestep cut during the prolonged activity may further increase their risk for sustaining a non‐contact ACL injury. Future studies using larger sample sizes should investigate the effects of prolonged activity on the decision to perform a particular cut.

Several limitations of this study should be noted. First, due to the relatively small sample size and the large discrepancy between the number of sidestep and crossover cuts the subjects performed, the statistical power for the study was lower than expected. Given this lower statistical power the potential for making a Type II error and being unable to note a significant difference in some of our outcome variables is a real possibility. Increasing the number of subjects for future studies may allow for a more meaningful statistical comparison between variables. A second limitation is the sample population only included Division I female soccer players, which does not represent other types or skill levels of athletes. This could have altered the outcomes of the kinematic changes because such athletes are trained to perform intense cutting activities. Future studies should look at lower extremity kinematics of athletes performing a cutting maneuver in a variety of athletes who participate in sports such as volleyball, basketball, gymnastics, etc. and at various skill levels. A third limitation of the study was the level and type of fatigue (muscular or central) the subjects achieved during the prolonged activity trials was not determined. In the present study fatigue was defined by a physical decline in performance (slower agility times). The subjects may have achieved central instead of muscular fatigue and may have needed more external motivation through the agility course. However, based on the subject's inability to complete the task and the fact that the subjects were Division I athletes, muscular and not central fatigue was likely achieved. Both types of fatigue may be critical in determining why subjects are more likely to become injured late in competition. Future studies should include a physiological measurement such as oxygen consumption, heart rate, or blood lactate levels to help determine which type of fatigue subjects attained during prolonged activity. Lastly, while the subjects were able to select the type of cut they performed, the cutting maneuvers were still anticipated. This, however, is not the case during competition. In a game scenario, players must react quickly to unanticipated stimuli, such as a defender, and manage other distractions during play. Therefore, future studies should focus on examining the cutting maneuvers chosen by athletes in response to unanticipated stimuli, and if this reaction places them in a position that is more prone to injury.

CONCLUSIONS

The results of the present study demonstrated that prolonged activity appears to predispose an athlete to position the knee in a more extended and adducted posture and the hip in a more extended posture during cutting maneuvers. The extended positions of the hip and knee have been suggested to place greater stress on the ACL and potentially increase the risk for injury. Interestingly, subjects chose to perform sidestep cuts more often than crossover cuts during an agility test regardless of activity status. In agreement with previous studies, the current results suggest that sidestep cuts may place the athlete at greater injury risk due to less hip and knee flexion and greater knee abduction angles compared to crossover cuts. Therefore, future studies need to account for the effects of prolonged activity as well as cut type when evaluating an athletes' risk for injury.