The deleterious impact of noncontact anterior cruciate ligament (ACL) injuries on both short- and long-term athlete health and performance is well documented (2). Although precise injury mechanisms remain unclear, altered lower-limb neuromuscular control during sporting maneuvers incorporating landing or cutting is increasingly implicated as a primary risk factor (21). These movement abnormalities are believed to culminate in specific dynamic joint movement and loading states, such as extreme knee abduction and/or anterior tibial shear loads (21), which compromise ligament integrity. Neuromuscular training programs designed to modify high-risk lower-limb control patterns continue to evolve (33). Despite the ever-increasing complexity and number of these programs, however, neither ACL injury rates nor the associated gender disparity have diminished (1). Thus, if neuromuscular control is indeed central to noncontact ACL injury risk, then current training programs evidently continue to omit key factors associated with the injury mechanism.
One factor that likely influences lower-limb-joint control and, hence, injury risk, is neuromuscular fatigue. Fatigue stems from combined physiological mechanisms occurring at the central and peripheral levels (6,39). It can affect afferent neuromuscular pathways, commonly observed as proprioceptive deficiency (19,25,27,28,39), as well as efferent neuromuscular pathways, as evidenced via delayed muscle response (42,43,50). Studies also have shown that fatigue-induced changes in these pathways, particularly proprioception, can be influenced both by age (44) and gender (45).
In sports such as football, soccer, and basketball, where high-intensity efforts are often required for extended time periods, athlete fatigue is inevitable. It seems feasible, therefore, that fatigue effects incurred cumulatively through the course of a game may compromise neuromuscular control responses to the point that abnormal and potentially hazardous movement strategies are inevitable. Recent epidemiological data tend to support this contention, where in sports such as American football and rugby league, for example, game-related injuries tend to occur more frequently at the beginning or end of the season, where the cumulative effects of an intense preseason or regular season may be most evident (5,14). Similarly, practice-related injuries often occur at times corresponding to extremely demanding training intensities (5,15). Regarding the temporal aspects of the game itself, Hawkins et al. (17,18) have demonstrated that a large percentage of noncontact knee injuries occur in the last 15 min of the first half and in the last 30 min of the second half of soccer matches. These temporal events most likely correspond to times during which athletes succumb to the mental and physiological demands of the game. Similar observations have been made for other sports such as rugby league (14).
Despite the apparent epidemiological link between fatigue and noncontact ACL injury, few studies have evaluated the precise role of neuromuscular fatigue within the noncontact ACL injury mechanism. Wojtys and associates (50) have, however, found that isolated fatigue of both the quadriceps and hamstring musculature in young healthy subjects precipitated potentially hazardous anterior tibial translation increases in the presence of an externally applied anterior loads. Furthermore, although the order of muscle activation did not change under fatigue, muscle premotor or reaction phases were noticeably greater, suggesting a possible compromise in their protective role. Localized quadriceps and hamstring fatigue also have been found to induce significant changes in female lower-limb control during crossover cutting tasks (42). An explicit link to noncontact ACL injury, however, was not examined here. More recently, generalized neuromuscular fatigue has been suggested to increase ACL injury risk during stop jump tasks, primarily via promotion of potentially hazardous anterior tibial shear loading, particularly in females (8). This study focused solely on the knee joint, however, which may limit elucidation of the precise role of fatigue within the ACL injury mechanism. This point seems particularly pertinent because hazardous knee loads are known to stem from altered control elsewhere in the lower limb, particularly in women (37). Examining neuromuscular fatigue effects across the entire lower limb during high-risk sports movements thus seems crucial to determining the role of fatigue within ACL injury and to enhancing the efficacy of current injury-prevention strategies.
Considering the current state of research in this area, this study aimed to evaluate the effects of fatigue on lower-limb (hip, knee, and ankle) kinematics and kinetics during the landing phase of a drop jump task. A second purpose was to examine the potential for gender specificity in these effects, with the hope of providing further insight into why noncontact ACL injuries are more common in women.
Our recent research comparing male and female knee-joint (rotational) kinematic data for side jumping tasks found that the ratio of within-group to between-group differences was 0.6 (38). Using this ratio, a power analysis revealed that to achieve 80% statistical power with an exploratory alpha level of 0.05, a minimum of seven subjects per group (male and female) was required. Ten female and ten male NCAA Division I athletes (basketball, soccer, and volleyball) were thus recruited as subjects for the purpose of the current study. Before experimentation, approval for the research through the institutional review board of the Cleveland Clinic Foundation, and written informed consent for all subjects was obtained. Subject inclusion was based on no history of cardiovascular or respiratory disease, no previous ACL injury, and no operable lower-limb-joint injury that would prevent them from taking part in either the jumping or fatigue protocols. Subjects had limb dominance ascertained before data collection, with the dominant leg being the one that could kick a ball the farthest. These data were considered within the statistical treatments presented later. A series of standard anthropometric measures were also recorded before testing. A summary of these and generic subject data are presented in Table 1.
For each subject, three-dimensional (3D) lower-limb-joint kinematic and 3D ground-reaction force data were recorded bilaterally during the stance (ground contact) phase of 10 jump landing tasks, both before and after exposure to a fatigue protocol. A successful jump trial necessarily required subjects to initially step off a 50-cm-high platform and land with each foot making complete contact with a separate AMTI forceplate (OR6-5 #4046 and #4048, Advanced Mechanical Technology Inc, Watertown, MA) within the field of view of a six-camera high-speed (240 frames per second) motion-analysis system (Motion Analysis Corp, Santa Rosa, CA). After landing, subjects were required to immediately jump vertically, with the intent of achieving maximum height as quickly as possible. Only data from the ground contact phase (heel-strike to toe-off) of the movement were used in the ensuing analyses (see dependent measures below).
After completion of the first set of 10 drop jump trials, subjects took part in a generalized lower-limb fatigue protocol. The protocol lasted exactly 4 min, consisted of a series of continuous drills that loosely reflected tasks synonymous with actual game play, and was of a similar intensity and magnitude to general fatigue models adopted previously (8,39) (Fig. 1). Specifically, subjects were first required to perform a series (N = 20) of step-up (and down) movements as quickly as possible, onto a step of height 20 cm. Then, they immediately performed a series of continuous plyometric bounding movements for a distance of 6 m, initiated a direction change (180°), and completed a second series of bounds back to the starting point. For each bounding movement, subjects were instructed to land and move into a deep knee-flexion position as the body was rapidly decelerated. After reaching their maximum flexion position, they immediately and rapidly rebounded into the next jump phase, similar to a plyometric bounding task. After the bounding drills, subjects again performed the step-up task, with this entire sequence being repeated as many times as possible within the 4-min period.
Throughout the fatigue protocol, each subject's heart rate was continuously monitored and recorded via a Polar heart rate monitor secured around the chest (S520; Polar Electro Inc., Lake Success, NY) to provide a general subject-specific measure of fatigue. Heart rate data were downloaded to a laptop computer on completion of all testing, from which the maximum heart rate attained throughout the protocol was calculated. The ratio between this value and the subjects' resting heart rate, recorded while the subjects sat comfortably for 10 min after arriving at the laboratory, was subsequently calculated.
Kinematic data were generated for each landing trial via the 3D coordinates of 29 precisely attached external reflective markers (19.5 mm in diameter) secured with strapping tape to predefined anatomic locations (Fig. 2). Attachment sites were first shaved and attachment over areas of large muscle mass was avoided to reduce the potential for excessive marker movement. A static (stationary) trial was initially recorded via the high-speed video system with the subject standing in the neutral position (38). To maximize the alignment of this position with that of anatomic (or lab) neutral, subjects' foot placements were standardized by having them stand precisely on a template covering each force plate. Sagittal and frontal alignments were also visually monitored and adjusted (if necessary) before capturing this stationary shot. The left and right anterior superior iliac spine (ASIS) and bilateral medial femoral condyle and medial and lateral malleoli markers were subsequently removed before the recording of the jump landing trials (Fig. 2a).
From the standing trial, a kinematic model comprising nine skeletal segments (one pelvis, and bilateral thigh, shank, talus, and foot segments) and 24 df was defined using Mocap Solver 6.17 (Motion Analysis Corp., Santa Rosa, CA). This same methodology has been presented in detail previously (35) and has been used to successfully quantify 3D lower-limb-joint motions associated with the stance phase of sidestepping (38), jump landing (41), and shuttle run tasks (38).
For the current study, the pelvis was assigned 6 df relative to the global (laboratory) coordinate system, with the hip, knee, and ankle joints of each limb defined locally (35) and assigned three rotational degrees of freedom, respectively (Fig. 2b). Hip, knee, and ankle-joint centers were also defined according to our previous work (35). The 3D marker trajectories recorded during each sidestepping trial were then processed by the Mocap Solver software to solve the generalized coordinates for each frame. Joint rotations at the hip, knee, and ankle were expressed relative to each subject's standing (neutral) position (38). These and the 3D ground-reaction force data were then low-pass filtered with a cubic smoothing spline at a 12-Hz cutoff frequency (52).
Three-dimensional intersegmental forces and moments were obtained by submitting the filtered kinematic and GRF data to a conventional inverse dynamics analysis, similar to our recent work (37). Segment inertial characteristics were defined in accordance with the work of de Leva (11). The 3D intersegmental forces at the hip, knee, and ankle were transformed to the femoral, tibial, and talar reference frames, respectively, to obtain anterior-posterior, medial-lateral, and compression-distraction forces. Intersegmental moments at the hip and knee were expressed as flexion-extension, adduction-abduction, and internal-external rotation moments with respect to the cardanic axes of their respective joint-coordinate systems (37). Similar to kinematic data, intersegmental ankle moments were expressed as plantar-dorsiflexion, internal-external rotation, and supination-pronation. Joint moments represent the external loads applied at each joint. For instance, the term "knee abduction moment" is used for an external load that acts to move the knee into a abduction posture (37). Kinematic and kinetic data were time normalized to 100% of stance, with initial contact and toe-off being defined as the instants when the vertical GRF first exceeded and went below 10 N, respectively (37). Considering the high loading rates associated with these movements, this enabled reliable detection of contact points while avoiding the potential for erroneous detection attributable to force-plate signal noise, which may otherwise occur at a lower threshold.
Maximum jump heights were also recorded for each trial, being defined as the maximum height of the sacral marker in the postlanding jump. As noted above, subjects were instructed to jump vertically as high as possible after landing, so we used the vertical (z) marker coordinate only. On completion of all jump trials, subject-based mean jump heights were calculated for the prefatigue data. The maximum jump heights recorded for each postfatigue jump trial were subsequently calculated, with each represented as a percentage of the subject's mean prefatigue value. Individual subject data were then combined, enabling gender-based mean prefatigue and trial-by-trial postfatigue jump heights to be compared (Fig. 3). From these data, a viable cutoff for the number of postfatigue trials to be used in ensuing statistical treatments was determined for both males and females. This cutoff was adopted because of concerns relating to a deterioration in fatigue effects and was denoted by the first postfatigue trial for which maximum jump height fell within 1 SD of the mean prefatigue values.
For each trial, bilateral hip, knee, and ankle angles exhibited at initial contact and the subsequent peak stance-phase values of each of these angles were determined. The peak degree-of-freedom value chosen for statistical evaluation (e.g., flexion or extension) was that which deviated the most from the stationary (standing) value (38). The associated normalized (mass × height) peak bilateral hip, knee, and ankle-joint peak moments and the peak anterior-posterior tibial shear force were similarly calculated for each trial. The time to each peak stance-phase knee moment and the peak stance-phase anterior-posterior tibial force was also quantified (ms). These data were not incorporated within the statistical model, however, because of the already large number of dependent factors. Instead, they were used to assist in interpretation of peak moment and force data with regard to injury potential. Contact times (ms) were also calculated from the initial ground contact and toe-off times defined above. On the basis of the analysis of maximum jump-height data described above, mean postfatigue data were calculated from the first five trials (Fig. 3). The maximum jump height obtained for each subject in the fifth trial was also recorded and was represented as a percentage of their mean prefatigue jump height. Mean subject-based measures of the nine initial-contact rotations, nine peak rotations, nine peak moments, peak force, contact time, and subject heart rate ratio and jump-height percentage data were submitted to a series (N = 31) of three-way mixed-design ANCOVA to determine the main effects of, and possible interactions between, leg (dominant and nondominant), gender (male and female), and fatigue (pre and post) conditions. Both the heart rate ratio and maximum jump-height percentage measures were initially considered as covariates within the analyses, specifically examining whether potential variations in fatigue levels influenced relationships between main effects and the remaining dependent measures. Fatigue condition was treated as a repeated measure within each subject, with subject nested within the fatigue condition and considered as a random factor. In instances where significant differences or interactions between conditions were observed, Tukey post hoc analyses were used to determine precisely where they occurred. Considering the extensive array of dependent measures examined within the study, we adopted a Bonferroni corrected alpha level of 0.002 for all statistical evaluations.
Heart rate ratio measures were similar between male (268.1 ± 34.8%; CI = ± 20.0%) and female (264.7 ± 18.1%; CI = ± 17.9%) subjects (P = 0.792, observed power = 0.057) and did not influence the remaining statistical outcomes. Peak jump-height percentages for the fifth jump trial were also found to be consistent between genders, with males (87.4 ± 11.2%; CI = ± 6.9%) and females (86.9 ± 9.4%; CI = ± 6.9%) exhibiting similar maximum heights with respect to mean prefatigue values. Furthermore, leg (P = 0.923, observed power = 0.051), gender (P = 0.388, observed power = 0.137), and fatigue (P = 0.890, observed power = 0.052) conditions did not influence mean ground-contact time data (Table 2).
A number of initial-contact and peak stance-phase lower-limb-joint rotations were influenced by the main effect of gender during execution of jump landing tasks (Tables 3 and4). Specifically, females demonstrated statistically significant increases in initial-contact ankle plantar flexion (P < 0.001) and in peak knee abduction (P = 0.001), peak knee internal rotation (P = 0.001), and peak ankle supination (P < 0.001) angles compared with males (Fig. 4). The main effect of fatigue also produced significant increases in peak knee abduction (P < 0.001) and peak knee internal rotation (P < 0.001) measures across all (gender × leg) conditions (Fig. 4). The main effect of leg did not have a statistically significant impact on any kinematic measures. Statistically significant interactions between main effects of gender, fatigue, and/or leg were also not observed for any kinematic parameters (Tables 3 and 4).
The main effect of gender was found to have a statistically significant impact on several kinetic parameters (Table 5). Specifically, mean normalized peak knee abduction (P < 0.001) and knee internal rotation (P = 0.002) moments were found to be statistically greater in females compared with males across all (fatigue × leg) conditions (Fig. 5). Females also demonstrated significantly smaller peak ankle dorsiflexion moments compared with males (P = 0.001). The main effect of fatigue produced significant increases in normalized peak knee abduction (P < 0.001) and peak knee internal rotation (P < 0.001) moments in both men and women (Fig. 5). Additionally, a statistically significant interaction between the main effects of fatigue and gender was observed for peak knee abduction moments, with fatigue-induced increases in this parameter being more pronounced in females compared with males (P = 0.002) (Fig. 6). A posteriorly directed proximal tibial shear force was evident throughout the entire landing phase, regardless of test condition (Fig. 6). Peak posterior shear force magnitudes were also observed to be similar between male (7.98 ± 1.00 N·kg−1) and female (7.73 ± 1.27 N·kg−1) (P = 0.378, observed power = 0.141), pre (7.79 ± 1.07 N·kg−1) and post (7.90 ± 1.25 N·kg−1) fatigue (P = 0.578, observed power = 0.085), and dominant (7.92 ± 1.18 N·kg−1) and nondominant (7.77 ± 1.15 N·kg−1) leg (P = 0.617, observed power = 0.078) conditions. Similarly, neither gender (P = 0.342, observed power = 0.156), fatigue (P = 339, observed power = 0.158) nor leg (P = 0.324, observed power = 0.165) had a significant main effect on normalized peak stance-phase knee-flexion moments.
Specific trends were observed on the basis of gender and/or fatigue when the times to peak stance-phase knee moments and forces were considered (Table 6). Specifically, the peak external knee-flexion moment and peak posterior tibial shear force occurred noticeably earlier in females than males across all leg and fatigue conditions. Both of these measures did occur earlier in males after fatigue, but they were still noticeably later than corresponding female peaks. Peak stance-phase external knee-abduction moments were also distinctively earlier in stance for females after fatigue, occurring very close to initial contact under this condition. For males, however, peak abduction moments occurred toward the end of stance, regardless of fatigue condition.
We set out to examine the combined effects of fatigue and gender on lower-limb biomechanics during drop jump tasks. Previous studies have suggested that fatigue-induced changes in knee mechanics may precipitate increased ACL injury risk, particularly in women (8,42,50). Although ACL loading is ultimately governed by loads experienced at the knee, failure to include the combined dynamic lower-limb system within analyses may compromise understanding of how fatigue truly manifests within the complete injury mechanism, thus limiting successful application to screening and prevention strategies.
Fatigue models implemented to investigate the impact of fatigue on both sensory and neuromuscular function have varied considerably within the literature (6,39). Research pertaining to the knee joint has examined fatigue effects via either local, as in the case of targeted isokinetic exercise (42,50), or general load models (8,32). Considering evolving links between noncontact ACL injury risk and neuromuscular control of the entire lower limb, however, initial research seems best served by the use of a general fatigue protocol (37). Further, such models employ drills that may more effectively replicate actual game play (6) and, unlike localized models, directly affect limb proprioception (39). In saying this, we acknowledge that the most effective means of studying the precise manifestations of fatigue within the noncontact ACL injury mechanism remains unclear. Similar to Chappell et al. (8), we have implemented a protocol that impacts both the cardiovascular and motor systems. Their definition of fatigue based on volitional exhaustion, however, may potentiate between-subject fatigue variations that are large enough to erroneously impact ensuing biomechanical data comparisons. We chose subjects of similar (NCAA Division I) fitness levels and submitted them to a temporally standardized fatigue drill to minimize such a likelihood. From a cardiovascular standpoint, we managed to fatigue gender groups equally, as evidenced by the consistent between-gender heart rate ratio data. One may intuitively expect the heavier and taller male subjects to perform more physiologic and mechanical work and, hence, elicit higher heart rates than their female counterparts. This anthropometric disparity, however, in conjunction with the nature of fatigue tasks themselves (e.g., rapid repetitive step-ups), most likely enabled the smaller, more agile females to complete more repetitions than (and hence perform equal work to) males during the 4-min period. Although we did not count bout repetitions explicitly during the fatigue protocol, this phenomenon was indeed observed for most cases.
The fact that postfatigue relative jump-height measures were similar between genders also suggests that consistency was achieved in peripheral neuromuscular fatigue, at least as it relates to performance. Of course, unlike isolated fatigue models (39,42,50), general fatigue models such as that currently adopted do not afford explicit measurement of a decrement in muscle strength and/or power; hence, our assumption of comparable fatigue effects at a muscular level remains largely speculative. Future work should address this limitation by developing standardized tasks that potentially enable a temporal stratification of precise local fatigue effects within a general fatigue paradigm. It is also possible that subjects experienced varying levels of central and/or sensory fatigue, which may have equally confounded results. Further assessment of the potential contributions of central fatigue to the ACL injury mechanism thus seems warranted. Finally, although we are confident that we fatigued subjects to a reasonable level, the rapid deterioration observed in fatigue effects suggests that a pretest-posttest model may not be ideal. Others have examined changes in landing biomechanics cumulatively by quantifying these parameters in parallel with the progression of fatigue (32). Such an approach largely negates the above concern and seems to provide a viable basis for our future work.
Patterns and magnitudes of initial-contact and peak kinematic and kinetic data were in general agreement with those reported previously for landing tasks from similar jump heights (12,26). Gender differences observed in our data, particularly in the frontal plane, where females consistently demonstrated greater abduction rotations and external abduction moments, were also consistent with previous investigations of these (26) and other sports postures (37,38). It has been shown in cadaveric (34) and computer (3) models that increases in abduction and/or adduction loading contribute to concomitant increases in ACL loading. Thus, it is frequently suggested that this typically female landing characteristic contributes directly to their increased risk of ACL injury (21). As is true with many of these studies, however, we have only characterized lower-limb-joint biomechanics during normal, safe movements in healthy subjects. It would be somewhat naive and inappropriate to immediately infer causality from these data. Hewett et al. (20), however, have shown prospectively that abduction motions and loads elicited during landing tasks may indeed predict ACL injury risk in young female athletes. Further examination of the interplay between frontal-plane knee and ACL loading during a wider variety of dynamic sports postures will add further insight here.
This is the first study to report a gender difference in knee axial rotation biomechanics during jump landing tasks. Internal tibial rotation motions and loads are known to contribute directly to ACL loading (34). Current results thus suggest that female tibial rotation patterns when landing from a jump may indeed contribute to their increased risk of ACL injury. However, the fact that for other movements, females elicit increased external, rather than internal, tibial rotation (8,38), suggests that if it indeed contributes to ACL injury risk, it may do so only for landings similar to the one currently examined. Again, it is impossible to successfully test this tenet on the basis of normative data only obtained during safe movements. Regardless, further research into the relationship between ACL loading and the combined 3D joint-loading scenarios associated with high-risk dynamic sports postures seems warranted.
Similar to previous studies (12,26), we have found that females exhibited greater plantar-dorsiflexion motion ranges compared with males. A landing strategy incorporating a large range of plantar-dorsiflexion affords greater shock attenuation at the ankle joint (48) and, hence, minimizes energy transfer to the knee joint and the ACL (12). Our observation of synchronous decreases in female external dorsiflexion moments suggests an altered activation strategy in the ankle musculature. Whether this strategy presents a direct means for reducing ACL injury risk is difficult to elucidate from the current data. In any case, the consistency with which this phenomenon has been observed warrants further investigation into possible links with ACL injury risk.
We are unsure why females exhibited greater peak ankle supination angles compared with males. Kernozek et al. (26) propose that increased frontal-plane ankle rotations represent a landing strategy that further minimizes energy propagation to the knee joint. In that study, however, females exhibited increased pronation rather than supination. Between-study differences in subject skill levels have been suggested previously to explain concomitant differences in joint biomechanical profiles for the same movement task (8,38). This may be the underlying cause of the current disparity, because Kernozek et al. (26) tested recreational athletes only. It is equally plausible, however, that it stems from erroneous talocrural-axis definitions based on mean population data, especially considering the extreme variations evident in human foot morphology (24).
We did not find any evidence of gender dimorphic sagittal-plane motion and/or load patterns elsewhere in the lower limb. There currently is a great deal of conjecture within the literature regarding the potential contributions of sagittal-plane biomechanics to female ACL injury risk. It is suggested by some, for example, that females consistently land in a more extended posture in stop jumping and cutting movements, which, in conjunction with increased quadriceps loading, precipitates anterior tibial shear loads that are large enough in isolation to cause injury (2,8). Recent modeling studies, however, suggest that injurious sagittal-plane loads are not possible during such movements in either men or women (36,46). Currently, there are at least as many studies finding that females land from a box in a more extended position (12) as there are that do not (26). If the gender disparity in ACL injury rates stems from an isolated sagittal-plane loading mechanism, then it may, at the very least, be task, if not population specific.
The links between neuromuscular fatigue and resultant athletic performance are well documented. The potential contributions of fatigue to ACL injury risk, however, remain largely unclear. Increases in knee internal rotation (42) and abduction loading (8) have been observed previously in the presence of fatigue during crossover cutting and stop jump tasks, respectively. The fact that we observed pronounced increases in both of these loading parameters may again be attributable to variations in the movement tasks and/or test populations. The fatigue model adopted in each case may also be an important contributing factor here, because, as noted above, local (42) and general (8) fatigue models have varying effects on proprioceptive control (28,39).
Neuromuscular fatigue did not influence peak sagittal-plane knee loads during the stance phase of the drop jump tasks. We noted earlier that ACL injury via an isolated sagittal-plane loading mechanism was unlikely in either men or women, and it seems that this tenet holds true in the presence of fatigue. This is not to say, however, that the interaction between fatigue and sagittal-plane loading is not still an important contributing factor to injury risk, especially if one considers injury to present through an extreme combined 3D knee-loading state (36). For example, Kernozek and associates (26) highlight that for tasks similar to those tested here, the temporal characteristics of the combined 3D female knee-joint load state may be an important contributing factor to their increased risk of ACL injury compared with that of males. Specifically, it has been speculated that the synchrony of specific joint motion and load parameters may precipitate cumulative joint-load states large enough to compromise ligament integrity (26).
Similar to the work of Kernozek et al. (26), we did observe gender differences in the timing of specific stance-phase knee-moment and force peaks. A noticeable temporal shift in peak knee abduction moments was evident in females after fatigue, for example, being much closer to initial contact when in this state. For males, however, peak abduction moments consistently occurred at the end of stance, both before and after fatigue. Recent work has suggested that the noncontact ACL injury occurs relatively early in stance, perhaps within the first 50 ms (23,46). If we consider this to be true, then it seems that the temporal shift in the peak abduction moment in the presence of fatigue may be an important contributing factor to female ACL injury risk. Of course, we are only considering the timing of single maximum stance-phase moment magnitudes here, which may limit interpretation in terms of injury. More insight may be gained into potential injury mechanisms if the timings of other secondary load peaks are considered, especially those evident early in stance. The fatigue-induced coupling of the peak knee abduction moment and the relatively large internal rotation moment peak evident within the first 20% of stance, for example (Fig. 5), which are both known in isolation to induce reasonable ACL loading (34), may present as a combined manifestation within the female injury mechanism.
As noted earlier, ACL loading dramatically increases in the presence of an anterior tibial shear load (34); hence, addition of an anterior tibial load to the combined out-of-plane load state discussed above would indeed create a high-risk scenario. For the movements tested in the current study, however, both before and after fatigue, the stance phase is dominated by a large, posteriorly directed tibial load (Fig. 5). Furthermore, females elicited peak posterior tibial shear forces much earlier in stance than males, coinciding more closely with these coupled out-of-plane moment peaks. In fact, males demonstrated a similar tendency when fatigued. We have suggested recently that female lower-limb landing mechanics elicited during normal pivoting and landing movements may represent an adaptive rather than predisposing strategy, accommodating their underlying anatomic and neuromuscular differences (37). Thus, rather than contributing to ACL injury risk, the magnitude and timing of the female posterior tibial shear force may exist to counter hazardous load states. It seems that, at least for normal jump landing movements, the temporal variations observed between genders in 3D knee-load peaks may exist to protect rather than compromise the female ACL. Further, this protective temporal load balance does not seem to be compromised in the presence of fatigue for men or women, at least for the levels of fatigue that we have induced. This is not to say, however, that detrimental changes in the timing and magnitudes of knee-joint loads are not possible in a more fatigued state. It may be that athletes experiencing extreme levels of fatigue elicit sagittal-plane loading patterns that can no longer protect the ACL from likely even greater cumulative out-of-plane knee loading. Further investigation into the impact of neuromuscular fatigue on the temporal aspects of knee-joint loading and the resultant risk of injury thus seems well warranted. In particular, a detailed focus on knee loading characteristics occurring in early stance may present with substantial benefit.
We did not quantify muscle activations, so we cannot comment on the precise fatigue-induced changes in these parameters. Considering the intensity and nature of our fatigue protocol, however, we are confident that such changes did occur and were an important contributor to our observed kinematic and kinetic adaptations. Selective activation of medial lower-limb muscle groups, for example, have been shown previously to counter external abduction loads simultaneously applied to the knee joint during the landing phase of cutting movements (4,42). Our fatigue model, involving repetitive plyometric bounds in the sagittal plane, likely impacted directly on the contractility of both the medial quadriceps and hamstrings, reducing their capacity to oppose such loads. Similarly, the large decelerations required about the ankle joint during landing likely effected soleus, gastrocnemius, and deep posterior-compartment calf-muscle contractility, which assist in limiting internal tibial rotation during dynamic landing tasks (42). Future work would benefit from the exploration of fatigue effects on explicit muscle responses during such movements, to gain greater insight into these interactions and, hence, injury risk. Considering this fact in conjunction with our earlier statements regarding the apparent benefits of a general fatigue model, it may well be that the precise role of fatigue within the sports-related ACL injury mechanism may be best evaluated through models integrating both local and general concepts.
Another potential reason for the fatigue-induced increases observed in out-of-plane knee motions and loads may stem from concomitant increases in knee-joint laxity. Exercise at or near exhaustion has been shown previously to significantly increase knee laxity (50), which, in turn, may compromise ligament mechanoreceptor feedback and, hence, muscle contributions to joint stability (28). Fatigue also has been shown to induce proprioceptive deficits in the muscle spindles (16) and golgi tendon organs (28), again promoting joint instability in the presence of reasonable loading. The above factors may also explain why we found that fatigue-induced increases in external knee-abduction loading were more pronounced in females than in males. Females generally possess increased knee laxity, for example (51), and it seems intuitive that this gender-dimorphic laxity behavior would remain evident with at least the same magnitude in the presence of fatigue. Females also have different muscle-activation strategies than males during the landing phase of jumping and cutting movements-in particular, one of quadriceps dominance (21) and a lack of effective muscle synchrony to offset potentially hazardous anterior knee loads (10). They also have been shown to more effectively recruit the quadriceps when at or near exhaustion (9). Thus, females may still be able to successfully perform sagittal-plane landing movements, which are controlled and driven primarily by the knee extensors (8), despite other major muscle groups such as the hamstrings and external hip rotators being fatigued. Under such a scenario, the potential for increased knee abduction seems reasonable because the external knee-abduction moment is known to be sensitive to both hamstring and hip rotator muscle control (4,22,30,37). Again, incorporating an assessment of muscle activation strategies during fatigued motions would add significant insight to this theory and should be considered in future work.
We have shown that in the presence of fatigue, knee motions and loads are possible during jump landing tasks in a pattern suggestive of increased noncontact ACL injury risk, particularly in women. Of course, we are no closer to elucidating the precise means through which fatigue manifests within this mechanism. It remains unclear, for example, whether our fatigue-induced changes represent degradation in central processing, the afferent sensory pathways, the efferent motor pathways, or a combination of each. Differentiation of fatigue effects within each of these facets of control, although definitely challenging, is not only imperative in terms of understanding its implications within the injury mechanism, but also with regard to injury prevention. This last point is particularly pertinent because it remains uncertain whether learned neuromuscular adaptations, stemming from current intervention programs (33), are in fact maintained in the presence of fatigue. Regardless of one's level of fitness, for example, peripheral fatigue is inevitable and, hence, cannot truly be combated. Through the inclusion of more complex, challenging decision-making tasks within neuromuscular training regimens, it may be possible to counter the deleterious impact of fatigue on central processing, decision making, and, thus, the resultant movement response. Of course, one must consider the potential tradeoff between increased complexity and an actual injury-causing event. Exposing individuals to movement tasks incorporating combinations of decision making and fatigue, for example, will almost certainly enhance their potential to elicit effective and safe movement responses when exposed to such tasks within the true game environment. It is possible, however, that integrating such tasks within what are already considered extremely challenging training regimens (40) will, ironically, precipitate the injuries that these programs are designed to prevent. Regardless, we feel that further exploration of this topic remains well warranted.
The underlying theory behind our kinematic and kinetic methods is that markers placed on the skin enable the underlying rigid body motions to be successfully quantified. Reasonable potential exists, however, for erroneous data arising from excessive skin-movement artifact (7,13,29). Skin marker methods have been used extensively in examinations of in vivo joint motions during gait and, more recently, in high-impact sports movements (8,26,37,38). Other methods are available that provide more accurate descriptions of in vivo joint motion (47,49). However, they are either extremely invasive or carry significant radiation effects, which do not lend themselves to the mass-testing requirements of studies such as this one. The impact of skin marker movement error is largely unknown for dynamic sports postures, but we have taken several steps to minimize this problem. First, a single investigator (S.G.M.) identified and placed markers on anatomic landmarks for all subjects, negating the potential for significant intertester errors that have been acknowledged previously (13). Second, our Mocap Solver software uses a model-based least squares global optimization technique (31), which makes use of the fact that, on the scale of gross movement, joints have fewer than 6 df. This assumption makes results less sensitive to errors in marker trajectories (31) because redundancy in the marker set is better exploited. The least squares approach also enables markers that may be more prone to impact error, such as the anterior thigh marker, to be removed from the analyses, with little change in the ensuing rotational output (38). Hence, although we acknowledge that errors arising from excessive skin movement are largely unavoidable, particularly when assessing dynamic motions with a pronounced impact phase, we are confident that these errors have not confounded the outcomes of the current study.
A point of contention in studies examining potential links between lower-limb biomechanics and ACL injury risk is whether the recording and analyses of a vast array of dependent factors is warranted. Others have previously compared a similar number of kinematic and kinetic measures for jump movements, justifying their approach by the need to evaluate the combined lower-limb system in totality (26). Such an approach seems feasible, considering recent links between knee-joint motions and loads associated with ACL injury and biomechanics elicited elsewhere in the lower limb (22,37,53). The quest for adequate statistical power, however, may often result in a number of potentially important variables being excluded from the statistical treatment, affording incomplete understanding of the combined lower-limb biomechanical contributions to injury. Considering that the true mechanism(s) of noncontact ACL injury remains largely unknown, we feel that our holistic approach presents an important initial step in the ultimate elucidation of more explicit factors contributing to ACL injury. In saying this, however, we are not ignorant of the need for statistical integrity, and thus we encourage further exploration of novel statistical methods in this area.
On the basis of the research outcomes obtained for the population tested, the following conclusions can be drawn:
* Women execute jump landing movements with more initial-contact ankle plantar flexion, peak stance-phase ankle supination, peak knee abduction, and peak knee internal rotation compared with men.
* Women also execute jump landing movements with larger peak stance-phase external knee-adduction, knee abduction, and knee internal rotation moments and smaller peak external ankle-dorsiflexion moments compared with men.
* Fatigue causes large increases in initial-contact and peak stance-phase knee abduction and knee internal rotation motions and in peak external knee-adduction, abduction, and internal rotation moments.
* Fatigue-induced increases in external knee-abduction moments occur noticeably earlier and are more pronounced in females than in males, suggesting a potential link with the increased risk of noncontact ACL injuries observed in women.
This research was funded by the National Institutes of Health (1R01-AR47039) and NFL charities.
1. Agel, J., E. A. Arendt, and B. Bershadsky. Anterior cruciate ligament injury in national collegiate athletic association basketball and soccer: a 13-year review. Am. J. Sports Med.
2. Bahr, R., and T. Krosshaug. Understanding injury mechanisms: a key component of preventing injuries in sport. Br. J. Sports Med.
3. Bendjaballah, M. Z., A. Shirazi-Adl, and D. J. Zukor. Finite element anlysis of human knee joint in varus-valgus. Clin. Biomech.
4. Besier, T. F., D. G. Lloyd, and T. R. Ackland. Muscle activation strategies at the knee during running and cutting maneuvers. Med. Sci. Sports Exerc.
5. Bradley, J. P., J. J. Klimkiewicz, M. J. Rytel, and J. W. Powell. Anterior cruciate ligament injuries in the National Football League: epidemiology and current treatment trends among team physicians. Arthroscopy
6. Cairns, S. P., A. J. Knicker, M. W. Thompson, and G. Sjogaard. Evaluation of models used to study neuromuscular fatigue. Exerc. Sport Sci. Rev.
7. Cappozzo, A., U. Della Croce, A. Leardini, and L. Chiari. Human movement analysis using stereophotogrammetry. Part 1: theoretical background. Gait Posture
8. Chappell, J. D., D. C. Herman, B. S. Knight, D. T. Kirkendall, W. E. Garrett, and B. Yu. Effect of fatigue on knee kinetics and kinematics in stop-jump tasks. Am. J. Sports Med.
9. Clark, B. C., S. R. Collier, T. M. Manini, and L. L. Ploutz-Snyder. Sex differences in muscle fatigability and activation patterns of the human quadriceps femoris. Eur. J. Appl. Physiol.
10. Cowling, E. J., and J. R. Steele. Is lower limb muscle synchrony during landing affected by gender? Implications for variations in ACL injury rates. J. Electromyogr. Kinesiol.
11. de Leva, P. Adjustments to Zatsiorsky-Seluyanov's segment inertia parameters. J. Biomech.
12. Decker, M. J., M. R. Torry, D. J. Wyland, W. I. Sterett, and J. Richard Steadman. Gender differences in lower extremity kinematics, kinetics and energy absorption during landing. Clin. Biomech. (Bristol, Avon)
13. Della Croce, U., A. Leardini, L. Chiari, and A. Cappozzo. Human movement analysis using stereophotogrammetry. Part 4: assessment of anatomical landmark misplacement and its effects on joint kinematics. Gait Posture
14. Gabbett, T. J. Influence of training and match intensity on injuries in rugby league. J. Sports Sci.
15. Gabbett, T. J. Reductions in pre-season training loads reduce training injury rates in rugby league players. Br. J. Sports Med.
16. Gandevia, S. C. Spinal and supraspinal factors in human muscle fatigue. Physiol. Rev.
17. Hawkins, R. D., and C. W. Fuller. A prospective epidemiological study of injuries in four English professional football clubs. Br. J. Sports Med.
18. Hawkins, R. D., M. A. Hulse, C. Wilkinson, A. Hodson, and M. Gibson. The association football medical research programme: an audit of injuries in professional football. Br. J. Sports Med.
19. Hiemstra, L. A., I. K. Lo, and P. J. Fowler. Effect of fatigue on knee proprioception: implications for dynamic stabilization. J. Orthop. Sports Phys. Ther.
20. Hewett, T. E., G. D. Myer, K. R. Ford, et al. Biomechanical measures of neuromuscular control and valgus loading of the knee predict anterior cruciate ligament injury risk in female athletes: a prospective study. Am. J. Sports Med.
21. Hewett, T. E., G. D. Myer, and K. R. Ford. Anterior cruciate ligament injuries in female athletes: Part 1, mechanisms and risk factors. Am. J. Sports Med.
22. Hewett, T. E., K. R. Ford, G. D. Myer, K. Wanstrath, and M. Scheper. Gender differences in hip adduction motion and torque during a single-leg agility maneuver. J. Orthop. Res.
23. Ireland, M. L. The female ACL: why is it more prone to injury? Orthop. Clin. North Am.
24. Isman, R. E., and V. T. Inman. Anthropometric studies of the human foot and ankle. Bull. Prosthet. Res.
25. Johnston, R. B., III, M. E. Howard, P. W. Cawley, and G. M. Losse. Effect of lower extremity muscular fatigue on motor control performance. Med. Sci. Sports Exerc.
26. Kernozek, T. W., M. R. Torry, H. van Hoof, H. Cowley, and S. Tanner. Gender differences in frontal and sagittal plane biomechanics during drop landings. Med. Sci. Sports Exerc.
27. Lattanzio, P. J., R. J. Petrella, J. R. Sproule, and P. J. Fowler. Effects of fatigue on knee proprioception. Clin. J. Sport Med.
28. Lattanzio, P. J., and R. J. Petrella. Knee proprioception: a review of mechanisms, measurements, and implications of muscular fatigue. Orthopedics
29. Leardini, A., L. Chiari, U. Della Croce, and A. Cappozzo. Human movement analysis using stereophotogrammetry. Part 3. Soft tissue artifact assessment and compensation. Gait Posture
30. Leetun, D. T., M. L. Ireland, J. D. Willson, B. T. Ballantyne, and I. M. Davis. Core stability measures as risk factors for lower extremity injury in athletes. Med. Sci. Sports Exerc.
31. Lu, T. W., and J. J. O'Connor. Bone position estimation from skin marker co-ordinates using global optimisation with joint constraints. J. Biomech.
32. Madigan, M. L., and P. E. Pidcoe. Changes in landing biomechanics during a fatiguing landing activity. J. Electromyogr. Kinesiol.
33. Mandelbaum, B. R., H. J. Silvers, D. S. Watanabe, et al. Effectiveness of a neuromuscular and proprioceptive training program in preventing anterior cruciate ligament injuries in female athletes: 2-year follow-up. Am. J. Sports Med.
34. Markolf, K. L., D. M. Burchfield, M. M. Shapiro, M. F. Shepard, G. A. Finerman, and J. L. Slauterbeck. Combined knee loading states that generate high anterior cruciate ligament forces. J. Orthop. Res.
35. McLean, S. G., A. Su, and A. J. van den Bogert. Development and validation of a 3-D model to predict knee joint loading during dynamic movement. J. Biomech. Eng.
36. McLean, S. G., X. Huang, A. Su, and A. J. van Den Bogert. Sagittal plane biomechanics cannot injure the ACL during sidestep cutting. Clin. Biomech.
37. McLean, S. G., X. Huang, and A. J. van den Bogert. Association between lower extremity posture at contact and peak knee valgus moment during sidestepping: implications for ACL injury. Clin. Biomech.
38. McLean, S. G., K. Walker, and A. J. van den Bogert. Effect of gender on lower limb kinematics during rapid deceleration changes: an integrated analysis of three sports movements. J. Sci. Med. Sport
39. Miura, K., Y. Ishibashi, E. Tsuda, Y. Okamura, H. Otsuka, and S. Toh. The effect of local and general fatigue on knee proprioception. Arthroscopy
40. Myer, G. D., K. R. Ford, and T. E. Hewett. Methodological approaches and rationale for training to prevent anterior cruciate ligament injuries in female athletes. Scand. J. Med. Sci. Sports
41. Myer, G. D., K. R. Ford, S. G. McLean, and T. E. Hewett. The effects of plyometric versus dynamic stabilization and balance training on lower extremity biomechanics. Am. J. Sports Med.
42. Nyland, J. A., R. Shapiro, D. N. Caborn, A. J. Nitz, and T. R. Malone. The effect of quadriceps femoris, hamstring, and placebo eccentric fatigue on knee and ankle dynamics during crossover cutting. J. Orthop. Sports Phys. Ther.
43. Nyland, J. A., D. N. Caborn, R. Shapiro, and D. L. Johnson. Fatigue after eccentric quadriceps femoris work produces earlier gastrocnemius and delayed quadriceps femoris activation during crossover cutting among normal athletic women. Knee Surg. Sports Traumatol. Arthrosc.
44. Petrella, J. K., J. S. Kim, S. C. Tuggle, S. R. Hall, and M. M. Bamman. Age differences in knee extension power, contractile velocity, and fatigability. J. Appl. Physiol.
45. Pincivero, D. M., C. M. Gandaio, and Y. Ito. Gender-specific knee extensor torque, flexor torque, and muscle fatigue responses during maximal effort contractions. Eur. J. Appl. Physiol.
46. Pflum, M. A., K. B. Shelburne, M. R. Torry, M. J. Decker, and M. G. Pandy. Model prediction of anterior cruciate ligament force during drop-landings. Med. Sci. Sports Exerc.
47. Reinschmidt, C., A. J. van den Bogert, B. M. Nigg, A. Lundberg, and N. Murphy. Effect of skin movement on the analysis of skeletal knee joint motion during running. J. Biomech.
48. Self, B. P., and D. Paine. Ankle biomechanics during four landing techniques. Med. Sci. Sports Exerc.
49. Tashman, S., D. Collon, K. Anderson, P. Kolowich, and W. Anderst. Abnormal rotational knee motion during running after anterior cruciate ligament reconstruction. Am. J. Sports Med.
50. Wojtys, E. M., B. B. Wylie, and L. J. Huston. The effects of muscle fatigue on neuromuscular function and anterior tibial translation in healthy knees. Am. J. Sports Med.
51. Wojtys, E. M., L. J. Huston, H. J. Schock, J. P. Boylan, and J. A. Ashton-Miller. Gender differences in muscular protection of the knee in torsion in size-matched athletes. J. Bone Joint Surg. Am.
52. Woltring, H. J., R. Huiskes, A. deLange, and F. E. Veldpaus. Finite centroid and helical axis estimation from noisy landmark measurements in the study of human joint kinematics. J. Biomech.
53. Zeller, B. L., J. L. McCrory, W. B. Kibler, and T. L. Uhl. Differences in kinematics and electromyographic activity between men and women during the single-legged squat. Am. J. Sports Med.