PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
J Orthop Res. Author manuscript; available in PMC 2010 April 26.
Published in final edited form as:
PMCID: PMC2859715
NIHMSID: NIHMS188744

KNEE INSTABILITY AFTER ACUTE ACL RUPTURE AFFECTS MOVEMENT PATTERNS DURING THE MID-STANCE PHASE OF GAIT

SUMMARY

Purpose

The purpose of this study was to identify gait asymmetries during the mid-stance phase of gait among subjects with knee instability (non-copers) after acute anterior cruciate ligament (ACL) rupture.

Methods

Twenty-one non-copers with acute, isolated, unilateral ACL injury ambulated at their intentional walking speed as kinetic, kinematic, and EMG data were collected bilaterally. Lower extremity movement patterns and muscle activity were analyzed during the mid-stance and weight acceptance phases of stance.

Results

During mid-stance subjects exhibited lower sagittal plane knee excursions and peak knee extension angles, and higher muscle co-contraction on the injured limb compared to the uninjured limb. There was also a lower knee flexion moment at peak knee extension, a trend for the knee contribution to the total support moment to be lower, and a higher ankle contribution to the total support moment on the injured limb in comparison to the uninjured limb. Differences in the magnitude of muscle activity included higher hamstring activity and lower soleus activity on the injured limb compared to the uninjured limb. Changes in quadriceps, soleus, and hamstring muscle activity on the injured limb were identified during weight acceptance that had not previously been reported.

Conclusions

Subjects with knee stability after ACL rupture consistently stabilize their knee with a stiffening strategy involving less knee motion and higher muscle contraction. The variable combination of muscle adaptations that produce joint stiffness, and the ability of both the ankle and hip to compensation for lower knee control, indicate the non-coper neuromuscular system may be more malleable than previously believed.

Keywords: Gait, Knee, ACL, Neuromuscular

INTRODUCTION

Alterations in injured limb kinematics, kinetics, and muscle activity are an almost universal finding early after anterior cruciate ligament (ACL) injury, yet reports in the literature are often inconsistent. Differences in methodology and data analysis can contribute to differences in findings15, as may a non-homogenous study sample may. Individuals who are able to dynamically stabilize their knees after ACL injury (copers) demonstrate distinctly different movement patterns from individuals who experience knee instability (non-copers).1, 68 Inclusion of both copers and non-copers in the same study sample may obscure genuine differences in movement patterns.1

At the University of Delaware we have conducted a series of investigations evaluating the movement patterns of individuals who are ACL deficient. A central concept in the design of each of these studies is establishing subject groups based on functional abilities.1, 610 Copers are defined as individuals who have asymptomatically resumed all pre-injury activities, including high level sports, for at least one year after the index injury.11 Copers are rare, comprising less than 14% of the entire population of patients who are ACL deficient,11 and implement strategies involving more coordinated muscle activation that stabilize the knee without compromising knee motion.1, 8 In contrast, non-copers are individuals who experience knee instability after ACL rupture.11, 12 Comprising the majority of the population sustaining an ACL injury, non-copers necessitate surgical stabilization of the knee in order to make a full return to pre-injury activities. Distinct gait adaptations are evident in non-copers early after ACL injury, including lower sagittal plane knee motion and knee moments, a lower knee and higher hip contribution to the total support moment, and higher muscle co-contraction in comparison to the contralateral limb.1, 6 This pattern is robust as non-copers exhibit similar movement and muscle activity during jogging, hopping, and stepping tasks.1, 68

Previous investigations of individuals with ACL deficiency have focused on movement patterns during the weight acceptance phase of gait (heel strike to peak knee flexion). Muscle activity and knee motions during weight acceptance create an unstable weight bearing posture.13 Thus, evaluation of movement patterns during this period of stance has provided tremendous insight to the neuromuscular strategies of the ACL deficient knee. Comparatively, few investigations have evaluated movement patterns of the ACL deficient knee during mid-stance. The period of stance occurring from peak knee flexion to peak knee extension, mid-stance is the first half of single limb support. Limb stability is therefore a major objective during this phase of gait. Mechanisms in addition to concentric quadriceps muscle action are employed to attain optimum extension in single limb stance. Tibial stabilization by the soleus is a major determinant of knee extension. Soleus muscle contraction restrains the tibia, allowing the femur to progress forward more quickly than the tibia and induce passive knee extension.13 This role as a tibia stabilizer when the foot is in contact with the ground also makes the soleus an ACL agonist, and therefore may play a significant role in the maintenance of knee stability after ACL rupture. Given the unique biomechanical demands placed on the lower extremity during mid-stance, evaluating movement patterns during this phase of gait may provide further insight to the neuromuscular strategies of patients with knee instability after ACL rupture.

The purpose of this study was to investigate the movement patterns of patients with knee instability early after ACL rupture (non-copers) during the mid-stance phase of gait. We hypothesized non-copers would exhibit gait asymmetries during mid-stance including 1) significantly lower sagittal plane knee excursions and angles, and higher muscle co-contractions on the injured versus uninjured limb, 2) significantly lower knee moments, with lower knee and higher ankle contributions to the total support moment on the injured limb compared to the uninjured limb, and 3) higher soleus and hamstring, and lower quadriceps muscle activity on the injured limb compared to the uninjured limb. Data from the weight acceptance interval were also evaluated to assess whether our cohort of non-copers was comparable to those previously described in the literature. We expected the non-copers to exhibit movement patterns during this portion of the gait cycle that included lower sagittal plane knee excursions, angles, and moments, a lower knee and higher hip contribution to the total support moment, and higher muscle co-contraction on the injured limb compared to the uninjured limb.

METHODS

Subjects

A total of 21 subjects (Male N=15; Female N=6; X age=31.38) with acute ACL injury who had been classified as non-copers with the University of Delaware screening examination14 were recruited for the study (X Time from injury=11.4 weeks). Inclusion criteria included regular pre-injury participation in level I or II activities12, 15 (e.g., jumping, cutting, pivoting), and ACL insufficiency based on magnetic resonance imaging results and a minimum 3 mm side-to-side difference during KT-1000 testing(MEDmetric Corp., San Diego, CA).12 Exclusion criteria included bilateral knee involvement and any lower extremity or low back injuries that prevented the completion of the screening examination. Patients were also ineligible if they had concomitant or symptomatic grade III injury to other knee ligaments, a repairable meniscus tear, or full thickness articular cartilage defect. All patients provided informed consent approved by the University of Delaware Institutional Review Board prior to testing.

Motion Analysis

Kinematic data were collected with a passive, six camera three-dimensional motion analysis system (VICON, Oxford Metrics Ltd., London, England) at 120 Hz and low-pass filtered with a bidirectional 2nd order Butterworth filter with a cutoff frequency of 6 Hz. Retroreflective markers were used to identify joint centers, measure limb length, and track lower extremity limb motion. Kinetic data were collected with a six-component force plate (Bertec Corporation, Worthington, OH) embedded within the walkway. Force data were collected at 1080 Hz and filtered with a 2nd order bidirectional, phase corrected Butterworth filter using a low-pass frequency of 50 Hz. Data were collected as subjects walked along a 13-meter walkway at their intentional walking speed. Only trials in which the speed did not vary by ±5% were accepted. Data were collected unilaterally per trial, with a total of 5 usable trials collected for each limb.

Electromyography

Electromyography (EMG) was collected at 1080 Hz with dual, active surface electrodes (Motion Lab Systems, BatonRouge, LA) and synchronized with motion analysis data. The electrodes were placed bilaterally over the vastus lateralis, vastus medialis, tibialis anterior, medial gastrocnemius, lateral gastrocnemius, soleus, medial hamstring, and lateral hamstring mid-muscle bellies and secured to the skin with adhesive tape (Leukotape®, Beiersdorf-Jobst Inc., Rutherford College, NC) and an elastic overwrap (Superwrap, Fabrifoam, Inc., Exton, PA). Four seconds of resting data and 2 seconds of maximum voluntary isometric contraction (MVIC) data for each muscle group were collected prior to motion analysis. The maximum EMG signal was defined as the highest level of EMG data found during any of the dynamic or MVIC trials.

EMG data were hardware filtered (20–1000 Hz) during motion testing and subsequently post-processed using custom written software (Labview, National Instruments, Austin, TX). All EMG data were band pass filtered in the software between 20–350 Hz. A linear envelope was then created from the EMG signals by full-wave rectification and low-pass filtering with a phase-corrected 8th order Butterworth filter using a cut-off frequency of 10 Hz. Linear envelope data were normalized to the maximum EMG signal.

Data Management and Analysis

Data from the 5 trials were averaged and the average was used for analysis. The intervals of interest for all kinematic, kinetic, and EMG data included the weight acceptance and mid-stance portions of the stance phase of gait. Stance was normalized to 100 data points to allow limb to limb comparisons. Weight acceptance was defined as the period of stance from heelstrike to peak knee flexion, and mid-stance from peak knee flexion to peak knee extension.

Lower extremity motions and moments were calculated using rigid body analysis with Euler angles, and joint moments were calculated using inverse dynamics (C-Motion, Inc. Rockville, MD). Joint moments were normalized to body mass (Nm/kg) x height (m) to reduce inter-subject variability and represent net internal moments. Hip, knee, and ankle extensor moments were summed to calculate the total support moment. The relative contribution of each joint to the support moment (percentage support moment) was calculated by dividing the individual joint extensor moment by the total support moment.

EMG variables of interest included magnitude of muscle activity and co-contraction during weight acceptance and mid-stance intervals. Magnitude of individual muscle activity is expressed as the average muscle activity across each interval (Average Rectified Value (ARV)). Co-contraction was defined as the simultaneous activation of antagonistic muscles and calculated using a technique previously described by Rudolph.1

Paired t-tests were performed to identify differences between the injured and uninjured limbs. Significance was established at p<0.05 for kinematic and kinetic variables and p<0.1 for EMG variables. A higher level of significance was established for EMG variables in an effort to avoid a type I error given the highly variable nature of EMG data.16 All statistical analyses were performed with commercially available software (SPSS 14.0, Chicago, IL).

RESULTS

Mid-Stance

Kinematics and Muscle Co-contraction

Sagittal plane knee excursion during mid-stance was lower on the injured limb (p<0.001) resulting in a more flexed knee at peak knee extension (p=0.028) than the uninjured knee (Fig. 1). Muscle co-contraction was greater among quadriceps-hamstring muscle pairs (VLLH: p=0.04, VMMH: p=0.009) (Fig. 2A) on the injured limb than the uninjured limb. There were no differences in quadriceps-gastrocnemius co-contraction indexes between limbs.

Figure 1
Sagittal Plane Knee Kinematics
Figure 2
Muscle Co-Contraction Indexes During Mid-Stance (A) and Weight Acceptance (B)

Moments

The knee moment (p=0.038) at peak knee extension was lower on the injured limb compared to the uninjured limb (Fig. 3). There was not a side to side difference in the hip contribution to the total support moment. There was a trend for the knee (p=0.064) contribution to be lower, and there was a significantly greater ankle (p=0.008) contribution to the total support moment on the injured limb compared to the uninjured limb (Fig. 4A).

Figure 3
Sagittal Plane Knee Moments
Figure 4
Percent Support Moment Contribution For The Hip, Knee, And Ankle

Muscle Activity

During mid-stance the magnitude of muscle activity differed between limbs for the hamstring and soleus muscles. Compared to the uninjured limb, hamstring activity was higher (lateral hamstring: I X=5.32 (4.38), U X=2.88 (2.37), p=0.027); medial hamstring: I X=3.47 (2.97), U X=2.19 (2.56), p=0.056) on the injured limb while soleus activity was lower (I X=26.87 (9.77), U X=31.97 (7.58), p=0.003). There were no side to side differences in quadriceps, gastrocnemius, or tibialis anterior muscle activity.

Weight Acceptance

Kinematics and Muscle Co-contraction

Knee flexion angles for both the injured and uninjured limbs were similar at initial contact. After initial contact the injured knee underwent significantly less sagittal plane excursion (p=0.006) resulting in a lower peak knee flexion angle compared to the uninjured knee (p=0.41) (Fig. 1). Muscle co-contraction on the injured limb was greater in quadriceps-hamstring muscle groups (VLLH: p=0.078, VMMH: p=0.095) but not quadriceps-gastrocnemius muscle pairs in comparison to the uninjured limb (Fig. 2B).

Moments

The knee moment at peak knee flexion was lower on the injured limb (p=0.003) in comparison to the uninjured limb (Fig. 3). The percent contribution of the hip and knee extensors to the total support moment differed between limbs (Fig. 4B). There was a significantly lower contribution from the knee (p=0.003) in conjunction with a greater contribution from the hip (p=0.009) compared to the uninjured limb.

Muscle Activity

The magnitude of muscle activity differed between limbs during weight acceptance. Quadriceps muscle activity was lower (vastus lateralis: I X=22.53 (11.36), U X=29.24 (14.14), p=0.023; vastus medialis: I X=20.32 (11.06), U X=27.88 (15.07), p=0.026) as was soleus activity (I X=13.22 (4.64), U X=15.07 (6.67), p=0.091) on the injured limb compared to the uninjured limb. In contrast the hamstring muscles were more active on the injured limb (lateral hamstring: I X=10.15 (6.04), U X=7.4 (3.45), p=0.047; medial hamstring: I X=10.9 (8.96), U X=6.69 (5.84), p=0.077) in comparison to the uninjured limb. There were no differences between limbs in gastrocnemius or tibialis anterior muscle activity.

DISCUSSION

There are many consistencies in the movement patterns of non-copers across different tasks and when comparing mid-stance to weight acceptance. As we hypothesized, there was a characteristic stiffening strategy observed during mid-stance on the injured limb that included lower sagittal plane knee motion, and higher quadriceps-hamstrings muscle co-contraction compared to the uninjured limb. Our hypothesis regarding a shift in the percent contribution to the total support moment was partially confirmed. The ankle contribution to the total support moment was significantly higher on the inured limb compared to the uninjured limb. There was a trend for the knee contribution to the total support moment to be lower on the injured limb, but insufficient power (β=0.49) prevented side to side differences from reaching statistical significance. Our hypotheses regarding the magnitude of muscle activity were also partially supported by the results. As we predicted, hamstring muscle activity was higher and soleus activity lower on the injured limb compared to the uninjured limb. There were no differences between limbs in quadriceps muscle activity. Kinematic, kinetic, and muscle co-contraction results in weight acceptance confirmed our population was similar to those previously described in the literature.

The stiffening strategy identified during mid-stance is similar to the one implemented by non-copers during the weight acceptance phases of walking, hopping, jogging, and stepping tasks.1, 68 Observation of this movement pattern during mid-stance further confirms stiffening as the predominant stabilization strategy adopted by non-copers. Whether a stiffening strategy may be resolved with rehabilitation interventions is a topic for future investigations.

Muscle co-contraction has been viewed by some as a positive adaptation after ACL injury. While co-contraction may help stabilize the knee in the absence of ligamentous support, high muscle co-contraction may be detrimental. After ACL rupture tibio-femoral articular cartilage contact points are altered.17 Wu 18 suggested osteoarthritis may be initiated after ACL injury when these areas of the knee joint that normally have little or no contact become overloaded due to instability and altered mechanics. Muscle co-contraction may further contribute to joint loading and hasten the degenerative process that is triggered with ACL rupture. Though co-contraction indexes observed during mid-stance were substantially lower than those in weight acceptance, it is unlikely co-contraction of any magnitude would be conducive to the maintenance of joint integrity after ACL rupture. Future investigations are necessary to quantify the forces imparted to the knee joint as a result of muscle co-contraction to clarify the role of muscle activity in the degenerative process after ACL injury.

The total support moment during mid-stance provides further insight to the movement strategies of non-copers. There was a trend for control to be transferred away from the unstable knee to the ankle, indicating mid-stance, though not as demanding as weight acceptance, still challenges knee stability. In an uninjured subject the ankle is typically the primary contributor to the total support moment, as the plantarflexors are contracting eccentrically to control dorsiflexion and muscles crossing the hip and knee are comparatively less active. In contrast, the hip provides the largest contribution to the total support moment during weight acceptance. The non-coper strategy of compensating for a loss of knee control is one of greater reliance on the joint that is already the primary contributor to limb stability. The invariant presence of a hip strategy during variable activities1, 8, 19 prompted Houck19 to propose non-copers rely on a single control strategy. In each of these conditions only the weight acceptance interval was evaluated, when muscle activity and joint demands are distinct from those in mid-stance. The current study indicates the non-coper control strategy does vary, but follows a pattern of joint contribution to limb stability that reflects the different demands of the mid-stance phase of gait.

A combination of higher hamstring and lower soleus muscle activity was observed during mid-stance. The hamstring complex was the only muscle group found to have higher muscle activity compared to the contralateral limb. In an interval when hamstring muscles typically are relatively inactive, persistence of higher muscle activity from weight acceptance into mid-stance suggests the hamstrings are a key component of the non-coper stabilization strategy. Because the hamstring muscles attach on the proximal tibia and exert a posterior force, they are considered the primary ACL agonists.2022 Higher hamstring muscle activity after ACL injury has therefore been viewed as a positive adaptation. High magnitude hamstring activity in mid-stance may represent a negative adaptation. Quadriceps-hamstring co-contraction values and less knee extension may be attributed to the hamstring muscle action, as quadriceps activity was not different between limbs during mid-stance. Lower soleus activity may have further contributed to greater knee flexion during mid-stance by failing to control tibia advancement. An alternative means of obtaining knee stability without stiffening the knee may have included lower hamstring and higher soleus muscle activity, as the soleus can function as an ACL agonist by holding the tibia posteriorly when the foot is on the ground 23 and allowing the knee to extend.

Interestingly, the magnitude of quadriceps muscle activity was not different between limbs during mid-stance. Lower knee (internal) flexion moments may therefore not be attributed to the quadriceps muscles during this interval. In the absence of EMG data, lower knee extension moments have been attributed to low quadriceps muscle activity (“quadriceps avoidance”) as part of a strategy to avoid knee instability.2, 4 The theoretical basis for a quadriceps avoidance strategy is that contraction of the quadriceps muscles will draw the tibia anteriorly when the knee is near full extension20, 22, making the knee extensors ACL antagonists. In studies that have included EMG data, however, there is evidence that quadriceps activity is not ubiquitously mitigated after ACL injury to avoid knee instability.1, 5, 2426 Although we observed lower quadriceps muscle activity during weight acceptance in this study, in mid-stance quadriceps activity was not lower. The absence of a quadriceps avoidance strategy during mid-stance in this study suggests tempering quadriceps activity is not a strategy for non-copers to control tibial translation during this phase of gait as the knee extends.

The movement patterns exhibited during weight acceptance suggests our sample of non-copers is similar to those who have previously been described in the literature. Our subjects’ alterations in knee kinematics, kinetics, co-contraction indexes, and shift from the knee to the hip in percent contribution to the total support moment resembled Rudolph’s previous description of non-coper movement patterns during weight acceptance.1, 68 Lower quadriceps and soleus, and higher hamstring muscle activity during the weight acceptance phase of gait have not previously been described among non-copers, yet the outcome remains a pattern of higher co-contraction, less knee excursion and less knee flexion. Muscle adaptations after ACL injury that effect knee stiffening appear to be variable, and not a fixed response.

Evaluation of mid-stance is useful for identifying gait abnormalities in patients with knee instability after ACL injury. Reflecting the different demands on and actions of the knee in mid-stance, we identified a control strategy novel to this phase of the gait cycle that includes a shift in the support moment from the knee to the ankle. Joint stiffening as a strategy for maintaining knee stability consistently includes less knee motion and higher muscle contraction during weight acceptance and persists into mid-stance. The variable combination of muscle adaptations that produce joint stiffness, and the ability of both the ankle and hip to compensation for lower knee control, indicate the non-coper neuromuscular system may be more malleable than previously believed. Future investigations will be necessary to determine if non-coper movement and muscle patterns during mid-stance can be altered by rehabilitation.

Acknowledgments

Funding for this project was provided by the National Institutes of Health (5R01AR048212-02), and the Foundation for Physical Therapy. The authors would like to thank Kyle Fennemore, Ben Keeton, and Jennifer Lipschultz, for their assistance during data collections; and Sarah Trager for the development of processing software.

References

1. Rudolph KS, Axe MJ, Buchanan TS, et al. Dynamic stability in the anterior cruciate ligament deficient knee. Knee Surg Sports Traumatol Arthrosc. 2001;9:62–71. [PubMed]
2. Wexler G, Hurwitz DE, Bush-Joseph CA, et al. Functional gait adaptations in patients with anterior cruciate ligament deficiency over time. Clin Orthop Relat Res. 1998:166–75. [PubMed]
3. Teixeira da Fonseca S, Silva PL, Ocarino JM, et al. Analyses of dynamic co-contraction level in individuals with anterior cruciate ligament injury. J Electromyogr Kinesiol. 2004;14:239–47. [PubMed]
4. Berchuck M, Andriacchi TP, Bach BR, Reider B. Gait adaptations by patients who have a deficient anterior cruciate ligament. J Bone Joint Surg Am. 1990;72:871–7. [PubMed]
5. Knoll Z, Kiss RM, Kocsis L. Gait adaptation in ACL deficient patients before and after anterior cruciate ligament reconstruction surgery. J Electromyogr Kinesiol. 2004;14:287–94. [PubMed]
6. Rudolph KS, Eastlack ME, Axe MJ, Snyder-Mackler L. 1998 Basmajian Student Award Paper: Movement patterns after anterior cruciate ligament injury: a comparison of patients who compensate well for the injury and those who require operative stabilization. J Electromyogr Kinesiol. 1998;8:349–62. [PubMed]
7. Rudolph KS, Axe MJ, Snyder-Mackler L. Dynamic stability after ACL injury: who can hop? Knee Surg Sports Traumatol Arthrosc. 2000;8:262–9. [PubMed]
8. Rudolph KS, Snyder-Mackler L. Effect of dynamic stability on a step task in ACL deficient individuals. J Electromyogr Kinesiol. 2004;14:565–75. [PubMed]
9. Chmielewski TL, Hurd WJ, Snyder-Mackler L. Elucidation of a potentially destabilizing control strategy in ACL deficient non-copers. J Electromyogr Kinesiol. 2005;15:83–92. [PubMed]
10. Chmielewski TL, Rudolph KS, Fitzgerald GK, et al. Biomechanical evidence supporting a differential response to acute ACL injury. Clin Biomech (Bristol, Avon) 2001;16:586–91. [PubMed]
11. Eastlack ME, Axe MJ, Snyder-Mackler L. Laxity, instability, and functional outcome after ACL injury: copers versus noncopers. Med Sci Sports Exerc. 1999;31:210–5. [PubMed]
12. Daniel DM, Stone ML, Dobson BE, et al. Fate of the ACL-injured patient. A prospective outcome study. Am J Sports Med. 1994;22:632–44. [PubMed]
13. Perry J. Gait analysis. Normal and pathological function. Thorofare, NJ: SLACK, Inc; 1992.
14. Fitzgerald GK, Axe MJ, Snyder-Mackler L. A decision-making scheme for returning patients to high-level activity with nonoperative treatment after anterior cruciate ligament rupture. Knee Surg Sports Traumatol Arthrosc. 2000;8:76–82. [PubMed]
15. Hefti F, Muller W, Jakob RP, Staubli HU. Evaluation of knee ligament injuries with the IKDC form. Knee Surg Sports Traumatol Arthrosc. 1993;1:226–34. [PubMed]
16. Winter DA, Eng P. Kinetics: our window into the goals and strategies of the central nervous system. Behav Brain Res. 1995;67:111–20. [PubMed]
17. Li G, Moses JM, Papannagari R, et al. Anterior cruciate ligament deficiency alters the in vivo motion of the tibiofemoral cartilage contact points in both the anteroposterior and mediolateral directions. J Bone Joint Surg Am. 2006;88:1826–34. [PubMed]
18. Wu JZ, Herzog W, Epstein M. Joint contact mechanics in the early stages of osteoarthritis. Med Eng Phys. 2000;22:1–12. [PubMed]
19. Houck JR, Duncan A, De Haven KE. Knee and hip angle and moment adaptations during cutting tasks in subjects with anterior cruciate ligament deficiency classified as noncopers. J Orthop Sports Phys Ther. 2005;35:531–40. [PubMed]
20. Renstrom P, Arms SW, Stanwyck TS, et al. Strain within the anterior cruciate ligament during hamstring and quadriceps activity. Am J Sports Med. 1986;14:83–7. [PubMed]
21. MacWilliams BA, DesJardins JD, Wilson DR, et al. A repeatable alignment method and local coordinate description for knee joint testing and kinematic measurement. J Biomech. 1998;31:947–50. [PubMed]
22. Li G, Rudy TW, Sakane M, et al. The importance of quadriceps and hamstring muscle loading on knee kinematics and in-situ forces in the ACL. J Biomech. 1999;32:395–400. [PubMed]
23. Elias JJ, Faust AF, Chu YH, et al. The soleus muscle acts as an agonist for the anterior cruciate ligament. An in vitro experimental study. Am J Sports Med. 2003;31:241–6. [PubMed]
24. Ciccotti MG, Kerlan RK, Perry J, Pink M. An electromyographic analysis of the knee during functional activities. II. The anterior cruciate ligament-deficient and -reconstructed profiles. Am J Sports Med. 1994;22:651–8. [PubMed]
25. Boerboom AL, Hof AL, Halbertsma JP, et al. Atypical hamstrings electromyographic activity as a compensatory mechanism in anterior cruciate ligament deficiency. Knee Surg Sports Traumatol Arthrosc. 2001;9:211–6. [PubMed]
26. Roberts C, Rash G, Honaker J, et al. A deficient anterior cruciate ligament does not lead to quadriceps avoidance gait. Gait Posture. 1999;10:189–99. [PubMed]