|Home | About | Journals | Submit | Contact Us | Français|
Rose Smith, PT; Kevin R. Ford, MS; Gregory D. Myer, MS; Adam Holleran, PT; Erin Treadway, PT; and Timothy E. Hewett, PhD, contributed to conception and design; acquisition and analysis and interpretation of the data; and drafting, critical revision, and final approval of the article.
Context: The recent increase in women's varsity soccer participation has been accompanied by a lower extremity injury rate that is 2 to 6 times that of their male counterparts.
Objective: To define the differences between lower extremity biomechanics (knee abduction and knee flexion measures) and performance (maximal vertical jump height) between National Collegiate Athletic Association Division I and III female soccer athletes during a drop vertical jump.
Design: Mixed 2 × 2 design.
Setting: Research laboratory.
Patients or Other Participants: Thirty-four female collegiate soccer players (Division I: n = 19; Division III: n = 15) participated in the study. The groups were similar in height and mass.
Intervention(s): Each subject performed a maximal vertical jump, followed by 3 drop vertical jumps.
Main Outcome Measure(s): Kinematics (knee abduction and flexion angles) and kinetics (knee abduction and flexion moments) were measured with a motion analysis system and 2 force platforms during the drop vertical jumps.
Results: Knee abduction angular range of motion and knee abduction external moments were not different between groups (P > .05). However, Division I athletes demonstrated decreased knee flexion range of motion (P = .038) and greater peak external knee flexion moment (P = .009) compared with Division III athletes. Division I athletes demonstrated increased vertical jump height compared with Division III (P = .008).
Conclusions: Division I athletes demonstrated different sagittal-plane mechanics than Division III athletes, which may facilitate improved performance. The similarities in anterior cruciate ligament injury risk factors (knee abduction torques and angles) may correlate with the consistent incidence of anterior cruciate ligament injury across divisions.
Colleges and universities have expanded female athletic programs since the inception of Title IX of the Educational Assistance Act (1972). Women's varsity soccer programs have grown 48% over the 5-year period from 1992– 1993 to 1998–1999.1 This increase in participation has been accompanied by lower extremity injury rates that are twice as high in female soccer players as in their male counterparts.2,3
The knee is one of the most frequently injured body parts for female soccer players, with major knee injuries accounting for 42% of injuries resulting in more than 30 days of time loss.4 The anterior cruciate ligament (ACL) accounted for more than half of those injuries during one soccer season.4 The collegiate athlete falls into the 15-year through 25-year age range for female athletes, which has the highest incidence of ACL injuries in landing and pivoting sports, with as many as 70% of these injuries being noncontact.5,6
Female college athletes are at a 2-fold to 6-fold greater risk of noncontact ACL injuries compared with male athletes playing the same sport (eg, soccer, basketball).7–9 Gray et al8 reported a 10-fold greater incidence of ACL injuries in professional female basketball players, whereas Malone et al10 showed a 6-fold to 8-fold higher rate in female collegiate basketball players. Authors7–9 have consistently found higher ACL injury rates among women. An early review of data from the National Collegiate Athletic Association (NCAA) Injury Surveillance System7 demonstrated that female collegiate soccer players had a 3-fold greater risk of a noncontact ACL injury than male soccer players. Agel and Arendt9 extended the review of collegiate soccer and basketball injuries from the Injury Surveillance System to include the 13 years prior to 2003.9 Female soccer players consistently sustained noncontact ACL injuries at a rate 3 times that of male soccer players.9 Female collegiate soccer athletes had a 59% greater frequency of noncontact ACL injuries compared with contact ACL injuries.9 The higher rate of injury for the female soccer athlete is consistent across collegiate divisions of play. According to the NCAA Web site,1 ACL injury rates for Division I female soccer players in 2002–2003 were 1.01 injuries per game versus a slightly lower rate of 0.66 injuries per game for Division III.
Although most activities during a soccer game involve walking, jogging, and running, Withers et al11 documented that jumping and pivoting are also frequently performed in soccer. For every 2 turns a player performs during a game, a landing from a jump or header occurs.11 The execution of 100 to 200 landings by a team during 1 game may help account for the 25% to 35% incidence of noncontact ACL injuries in women's soccer that occur during landing.6,11,12
Noncontact ACL injuries typically occur with a combined mechanism of knee valgus near full extension with rotation of the tibia and the foot firmly planted on the ground.13 Tasks associated with ACL injury are rapid decelerations that occur with landing from a jump or a change in direction during a pivot or cutting motion. Hewett et al14 reported a high sensitivity and specificity for knee abduction moments measured during a drop vertical jump in the prediction of ACL injury risk. Several groups15–18 have reported a greater risk of noncontact ACL injury in postpubertal females because they land jumps with greater knee valgus than prepubertal females or males.
Noncontact ACL injury may be related to sagittal-plane mechanics in female athletes. Females who land with decreased knee flexion may increase the forces on the ACL.5,19 In addition to valgus collapse, Boden et al6 reported that most ACL injuries occur with the knee near full extension during a sharp deceleration or while landing a jump. This extended position, along with eccentric contraction of the quadriceps muscle, increases the strain on the ACL. Conflicting reports19–22 describe greater, equal, and less initial-contact and maximal knee flexion in female athletes than in males. Landing with a smaller knee flexion angle has been suggested to increase the potential for noncontact ACL injuries, as the extended knee may have less dynamic control and ability to absorb forces.23 Chappell et al19 suggested that a decreased knee flexion angle in postpubertal females may increase anterior knee shear at landing due to increased quadriceps force and decreased hamstrings firing, which are associated with an increased risk of injury. In addition, Hewett et al14 reported a maximal knee flexion that was 10.5° less in athletes who went on to injure their ACLs as compared with uninjured female adolescent athletes during landing from a drop vertical jump (DVJ).
Comparing males and females, as well as females prepuberty and postpuberty, may lead one to conclude that knee flexion angle may be a measure for performance and less of a predictor of ACL risk. However, these reported findings are not unequivocal. Hass et al16 related an increase in ACL injury risk postpuberty with a decrease in knee flexion angle. However, Ford et al24 reported that male soccer athletes demonstrated a decreased knee flexion angle and increased performance (vertical jump height) compared with female collegiate players while performing the DVJ with an overhead target.
Limited comparisons of performance levels across NCAA divisions are reported in the literature. However, Arnason et al25 demonstrated a significant relationship between team average jump height and team success (defined as final league standing). Fry and Kraemer26 ran a battery of tests (vertical jump, bench press, power clean, and a 36.6-m sprint) that effectively differentiated among Division I, II, and III football players' abilities. Garstecki et al27 also reported that vertical jump was a predictor of different performance between Divisions I and II football players. Wisloff et al28 correlated strength and short-burst skills with vertical jump height in male soccer players.
Interestingly, although performance has been documented to be greater in the higher divisions, injury rates in females were not reported to be different among divisions. Harmon and Dick29 used data from the NCAA Injury Surveillance System (1989 through 1997) to compare ACL injury rates among divisions. The ACL injury rate was not different among divisions and was consistently greater for female than male athletes. This finding may help to explain the similarities in ACL injury incidence rates across female collegiate divisions reported by the NCAA.1 After a review of the literature, Wong and Hong30 concluded that low skill level and high competitive levels increased injury rate. Risk increased with match exposures and not the level of skill.30,31 Jacobson and Tegner32 reported similar findings among elite Swedish female soccer players at various skill levels.
Neuromuscular risk factors, specifically increased knee abduction and decreased knee flexion, may be related to ACL injury and the sex differences in ACL injury incidence. However, how these biomechanics are related to skill or participation level in female soccer athletes is unknown. Reportedly, the injury risk for collegiate athletes is similar for all divisions, including Division I, where the highest levels of athletic performance are expected.25–27,29
Our purpose was to evaluate kinematic and kinetic factors linked to increased ACL injury risk and to compare performance in female athletes from different collegiate divisions. Specifically, we were interested in whether sagittal-plane and coronal-plane movement biomechanics and vertical jump height differed between Division I and Division III female soccer athletes. The first hypothesis was that Division I and Division III female soccer players would not demonstrate differences in knee abduction and knee flexion measures during a DVJ. The second hypothesis was that Division I female soccer athletes would demonstrate higher performance measures than Division III female soccer athletes.
Thirty-four female collegiate soccer players (19 from Division I, 15 from Division III) volunteered to participate in the study. They were of similar height (Division I = 166.7 ± 5.4 cm, Division III = 162.6 ± 5.9 cm) and mass (Division I = 63.3 ± 7.4 kg, Division III = 62.9 ± 10.5 kg). All participants read and signed the informed consent form approved by the institutional review board, which also approved the study. After subjects signed the written consent form, we measured their height and mass. Leg dominance was determined by the leg the athlete stated that she would use to kick the ball the farthest.
Maximal vertical height was determined using an MX-1 vertical jump trainer (MXP Sports Inc, Reading, PA) with a basketball as a target to encourage maximal jump height.33 The athlete was instructed to grab the ball with both hands as she jumped as high as possible. The height of the ball was adjusted until she failed to grab the ball in 3 consecutive trials. The greatest vertical jump height was recorded.33 This target height was then set for the DVJ trials. A total of 37 retroreflective markers (Figure 1) were secured to the sacrum, left posterior superior iliac spine, and sternum and bilaterally to the shoulders, elbows, wrists, anterior superior iliac spines, greater trochanters, mid thighs, medial and lateral knees, tibial tubercles, mid shanks, distal shanks, medial and lateral ankles, and heels and dorsal surfaces of the midfoot, lateral foot, and toes (between metatarsals 2 and 3) with double-sided tape. A static trial was conducted with the subjects standing as positioned in Figure 1. Kinematic measures were determined in relation to the static position. Each athlete was shown the DVJ maneuver and given adequate practice to successfully complete the task (Figure 2). The DVJ consisted of dropping off the box and jumping immediately to grab the target.15 Three DVJs were performed from a box height of 31 cm; 8 digital cameras (Eagle cameras; Motion Analysis Corp, Santa Rosa, CA), and 2 force platforms (model OR6-5; AMTI, Watertown, MA) were used to record the DVJ maneuver. The data were time synchronized (video = 240 Hz, force = 1200 Hz) and collected with EVaRT (version 4, Motion Analysis Corp). The motion analysis system was calibrated to recommended specifications before each data collection session.
Visual3D (version 3.65; C-Motion, Inc, Rockville, MD) was used for data reduction and analysis. The 3-dimensional marker trajectories from each trial were filtered at a cutoff frequency of 12 Hz (low-pass, fourth-order Butterworth filter). The 3-dimensional knee joint angles were calculated according to the Cardan/Euler rotation sequence.34 Kinematic data were combined with force data to calculate knee joint moments using inverse dynamics.35,36 The ground reaction force was filtered through a low-pass, fourth-order Butterworth filter at a cutoff frequency of 12 Hz in order to minimize possible impact peak errors.37,38 Net external moments are described in this paper and represent the external load on the joint. Center of mass was estimated for the entire body from each segment within Visual3D. Vertical jump height was calculated from the body center of mass as the difference between the maximal vertical height during the jump and the standing height (Figure 3).
We calculated means and SDs for knee flexion and abduction angular range of motion and maximal joint moment. A 2 × 2 mixed-design multiple analysis of variance was performed with side (dominant or nondominant) and collegiate division (Division I or III) as independent variables. A post hoc univariate analysis was performed to identify the kinematic and kinetic dependent variables that were significantly different between divisions. An independent t test was used to identify differences in vertical jump height between divisions. An exploratory α level of .05 was determined a priori to indicate statistical significance. Statistical analyses were conducted in SPSS (version 12.0; SPSS Inc, Chicago, IL).
A power analysis was performed a priori with measures of vertical jump height and maximal knee abduction moments. It was based on the differences Fry et al39 identified among collegiate divisions (I through III) in vertical jump performance, and these data were adjusted based on female soccer players.33 In order to achieve 80% power (α level = .05), a minimum of 14 subjects was required in each group based on vertical jump performance. A minimum of 12 subjects in each group was necessary to achieve 80% power (α level = .05) based on knee abduction moments during DVJ landings.40
Multivariate differences between collegiate divisions were found in the knee kinematic and kinetic variables (collegiate division main effect: F4,29 = 3.2, P = .027). No multivariate interaction was noted between collegiate division and side (dominant versus nondominant, F4,29 = 0.2, P = .92), and no main effect of side was demonstrated (F4,29 = 2.0, P = .13) between the dependent variables. The Table shows the univariate analysis of knee kinematic and kinetic variables in the coronal and sagittal planes. Knee abduction angles for Divisions I and III female soccer players were 8.8 ± 3.4° and 10.0 ± 3.9°, respectively (P = .292; Figure 4). Correspondingly, maximal knee abduction moment (torque) was not significantly different between groups (P = .220; Table, Figure 5). Hence, in the coronal plane, no differences were identified in measured knee kinematics or kinetics between Division I and Division III female soccer players.
Kinematic and kinetic variables at the knee in the sagittal plane, in contrast to coronal-plane measures, were different between the collegiate divisions of female soccer athletes (Table). Division III female athletes demonstrated a significantly greater knee flexion range of motion than Division I athletes (P = .038; Figure 4). In addition, Figure 5 illustrates a 29% greater peak external knee flexion moment in Division I athletes than Division III athletes (P = .009). Hence, significant differences in knee motion and torque, which were related to performance, were observed in the sagittal plane. Maximal vertical jump was measured as an additional performance measure of whole body power. As shown in the Table, a significant difference in vertical jump height was found between divisions (P = .008). Division I players jumped a mean of 4.5 cm higher than Division III players. This difference represents a relative 12% higher level of performance of whole body power in Division I female soccer players.
Our purpose was to identify whether the biomechanics of high-risk movement and performance measures differed between Division I and Division III female athletes. The first hypothesis was that Division I and Division III female soccer players would not demonstrate differences in knee abduction and knee flexion measures during a DVJ. This hypothesis was partially supported. Knee abduction angular range of motion and knee abduction moments in this study were similar between Division I and Division III female soccer players. This finding may help to explain the similarities in ACL injury incidence rates across female collegiate divisions as reported by the NCAA.1
Wong and Hong30 concluded, after a review of the literature, that low skill level and high competitive levels increased injury rate. Risk increased with match exposures and not the level of the skill.30,31 Jacobson and Tegner32 reported similar findings among elite Swedish female soccer players of various skill levels. Harmon and Dick29 compared the rate of noncontact ACL injuries among divisions in female basketball and soccer players. No significant difference was observed among divisions (ie, skill levels). Chomiak et al41 reported similar findings among soccer players between 14 and 42 years of age. Greater velocities and forces may be expected among Division I athletes, which may result in higher joint forces and torques. Although this may be the case, these factors do not correlate with the epidemiologic findings.
The predictive value of knee abduction angle and moment was reported by Hewett et al.14 A total of 205 female high school basketball, volleyball, and soccer players were screened before participation. The 9 who subsequently injured an ACL had a significant 6.4 times greater knee abduction moment, 20% greater ground reaction force, and 8° greater knee valgus angle than the uninjured female athletes. The knee abduction moment had a high sensitivity and specificity for the prediction of noncontact ACL injuries.14 Cowley et al40 compared high school basketball and soccer players and reported similar initial-contact and maximum-during-stance knee abduction (valgus) angles during a DVJ. In addition, McLean et al42 demonstrated that a change in valgus motion of as little as 2° can increase the abduction moment by 40 Nm. Computer simulations have documented lower extremity moments high enough to rupture the ACL.42,43 High valgus knee torques have been correlated with increases in ground reaction forces,14,21,44 further linking knee valgus to the risk of ACL injury.14,15,42,45
To reduce this risk, Beynnon and Fleming46 emphasized the need for active muscular restraints to decrease joint loads. Markolf et al47 reported a 3-fold decrease in valgus laxity with muscular contraction, and Fagenbaum and Darling22 showed that healthy females adapt compensatory hamstrings activity to decrease joint laxity. Lloyd et al23 stated that landing with the knee in valgus can potentially injure the knee and that people use different activation patterns of the quadriceps and hamstrings muscles to control the valgus forces at the knee.
Ekstrand and Gillquist48 listed conditioning as a risk factor that could be altered in order to decrease the risk of ACL injuries. Neuromuscular training has been integrated in training programs to aid in dynamic knee joint control. A 6-week neuromuscular training program to reduce the valgus moments and angles has also demonstrated a reduction in the ground reaction force during landing in high school athletes.21 Hewett et al44 reduced the rate of ACL injury with similar training as compared with untrained individuals by reducing the ground reaction forces. The similarity in knee valgus angles and moments demonstrated between Division I and Division III female soccer athletes in this study may help to explain the similarity in risk for ACL injury observed among NCAA divisions, concurring with the study of Harmon and Dick.29 Contrary to the similarity in knee abduction angles, knee flexion angles were significantly different.
Sagittal-plane knee flexion may not be as reliable as knee valgus for predicting ACL injury risk, at least when comparing Division I and Division III female athletes. Male collegiate soccer players, who demonstrate decreased noncontact ACL injury rates, had a smaller knee flexion angle than female athletes during the landing phase of a DJV.24 Kernozek et al18 reported similar findings with a young adult recreational athletic population performing a 60-cm drop landing. Chappell et al19 reported decreased knee flexion angles and increased extension moments while landing from stop-jump tasks. Huston et al20 studied drop jumps from 20-cm, 40-cm, and 60-cm heights and observed greater knee flexion angles in men on landing at the greater heights, but at the 20-cm height, no significant difference was evident. Hewett et al21 reported no difference in knee flexion angles between female volleyball players and untrained males, whereas Fagenbaum and Darling22 demonstrated that women landed with increased knee flexion, whether from a DVJ or maximal vertical jump task. Ford et al24 tested a cutting task and also found no difference in knee flexion angle between male and female basketball players. Neuromuscular training has also been shown to reduce coronal-plane moments in basketball and soccer players.14,49 Neuromuscular training utilizes exercises that target the eccentric contraction that prestretches the muscle before a concentric contraction. According to Vanezis and Lees,50 superior jump performance is due to greater muscle capability in terms of strength and rate of strength development, rather than technique, as different strategies are used to emphasize either the knee or the hip. This variability may account for the significant difference demonstrated in knee flexion angles, which warrants further study of the role of knee flexion as a risk factor.
The second hypothesis was that Division I and Division III would be different in performance measures between the 2 levels of collegiate participation. Our results demonstrated that Division I female soccer athletes jumped nearly 5 cm higher than Division III athletes during a vertical jump, which supported the second study hypothesis. This difference was expected, because players are recruited into higher divisions based on power and skill. The correlation of NCAA division with player skill level in the study of Harmon and Dick29 concurs with this principle. Higher skill levels do correlate with a higher level of performance. Fry and Kraemer26 and Garstecki et al27 used the vertical jump as one test that effectively differentiated athletes according to divisional play. Silvestre et al51 demonstrated a significant correlation between total body power during the vertical jump in college soccer players to differentiate starters and nonstarters.39 Markovic et al52 confirmed that the countermovement jump was a reliable (estimated between .90 and .99) and valid (validity coefficient = .78) field test for the estimation of explosive power.53
The plyometric exercise component of neuromuscular training designed to prevent ACL injury utilizes the enhanced power produced by the rapid muscle stretch that occurs immediately before an explosive concentric contraction. Stretch of the series elastic component of the muscle that uses stored energy, stretch of the muscle spindle that provides neural feedback regarding length and rate of stretch, and the optimal length of each muscle fiber for enhanced actomyosin interaction may all be contributing factors to the plyometric effect of the vertical jump.53 Vanezis and Lees50 related enhanced jumping performance to the strength development capabilities of the muscle and rate of strength development, rather than technique. Bosco and Komi54 showed that a concentric contraction can be magnified when performed quickly after a short-range eccentric contraction. A 6-week neuromuscular training program for ACL prevention increased jump height 10%.21 Authors23,55,56 studying other plyometric training programs also reported improvements in vertical jumping performance. Therefore, the observed performance differences in Division I relative to Division III athletes may be the result of enhanced training, innate ability, or a combination of the two.
This study has potential limitations. The biomechanical patterns demonstrated by collegiate female soccer athletes may not be generalized to athletes in other sports. In addition, use of the DVJ to assess soccer players is a potential limitation in soccer athletes, because 65% to 75% of noncontact injuries result from maneuvers other than landing (such as running and cutting), although a soccer player is predicted to land up to 20 times per game.11 Therefore, the measured task used for this study may not be as relevant to soccer players because human performance is task specific. The measured tasks we used may be more appropriately applied to athletes in sports that primarily involve jumping (eg, basketball and volleyball).9 In addition, although our findings support the position of Harmon and Dick29 that collegiate divisions are a marker that can be used to measure performance, level of division play may not be determined solely by skill. Other social, cultural, financial, and environmental factors may account for or influence an athlete's decision to play at the school of his or her choice. These additional influences may need to be studied before the relationship between the skill level of the athlete and the level of divisional play can be understood.
The experimental findings indicate that Division I female soccer athletes demonstrated different sagittal-plane mechanics than Division III players. Division I athletes landed from a vertical drop with a decreased knee flexion angle and increased vertical jump height relative to Division III athletes. The observed similarities in coronal-plane knee abduction range of motion and moments between Division I and III female soccer athletes may be related to the consistent incidence of ACL injury across divisions as documented by the NCAA. Ekstrand et al57 listed conditioning as a factor that can be manipulated to decrease the risk of ACL injuries. Prevention programs designed to decrease the abduction angle at the knee during landing may be recommended for female soccer players at all performance levels to reduce their risk of injury.
We acknowledge funding support from the National Institutes of Health/NIAMS (grant R01-AR049735). We also acknowledge the women's soccer teams from the University of Cincinnati and the College of St. Thomas More for their participation in the study. We thank team physicians Angelo Colosimo, MD; Jon Divine, MD; and Michael Miller, MD, for their contributions. We also thank Robert Mangine, EdD, PT, ATC, for his critical reading of the manuscript.