Explosive Quadriceps Strength Symmetry and Landing Mechanics Limb Symmetry After Anterior Cruciate Ligament Reconstruction in Females

This investigation was part of a larger study of the relationships between (1) quadriceps RTD and landing mechanics during functional landing tasks and (2) quadriceps RTD and arthrogenic muscle inhibition in females with or without ACLR.

A total of 38 females (19 with ACLR and 19 serving as the control group) between the ages of 16 and 30 years were recruited. Participants completed a general screening questionnaire, the 2000 International Knee Documentation Committee Subjective Knee Evaluation Form (IKDC 2000), the Knee Outcome Survey–Activities of Daily Living Scale (KOS-ADLS), and the Tegner Activity Scale to verify that they met the eligibility criteria.

Inclusion criteria were no history of low back, hip, knee, or ankle surgery except for ACLR and ≤1 subsequent minor procedure for hardware removal, debridement, or arthroscopic assessment after the primary ACLR. All participants with ACLR had undergone unilateral primary ACLR surgery and been cleared for unrestricted activities within the past 2 years by their surgeons. If an additional operation after the primary ACLR was performed, the participant must have been cleared for unrestricted activity by the operating surgeon.

All recruits were recreationally active, which we defined as self-reporting participation in moderate to vigorous physical activity (64%–95% of maximum heart rate) for at least 150 min/wk. Exclusion criteria were an injury to the back or lower extremity within 6 months of the study that limited her physical activity, a neurologic or cardiopulmonary disorder, or a history of multiple ACLRs or graft failure. Any individual who scored <2 on question 7 of the IKDC 2000, <3 on any item in question 9 of the IKDC 2000, or <4 on any question on the KOS-ADLS was also excluded. Females who met these criteria but did not have a history of ACLR were placed in the control group. All participants and their legal guardians (if the participants were <18 years of age) provided written informed assent or consent as appropriate, and the study was approved by the Institutional Review Board of Oregon State University.

After verifying eligibility, we measured participants' height and mass and determined limb dominance by identifying the limb that was used to complete at least 2 of the following 3 tasks: kicking a ball for distance, stepping up onto a small step, and recovering from a small perturbation from behind.¹⁸  Participants were then instructed to perform a warm-up on a stationary bicycle or treadmill for 5 minutes at submaximal intensity.

Participants were positioned on the Biodex System 3 dynamometer (Biodex Medical Systems Inc) in a sitting position with the trunk reclined to 70° from the horizontal plane and the knee of the test limb flexed to 70°. The lateral femoral condyle was aligned with the axis of rotation of the dynamometer. Shoulder, waist, and thigh straps were used to secure each person and minimize any excessive movement. The distal shank of the test limb was attached to the dynamometer arm with a hook-and-loop strap placed over the musculotendinous junction of the gastrocnemius.

For familiarization, participants were instructed to perform 1 quadriceps voluntary isometric muscle contraction each at 25%, 50%, and 75% of their self-perceived maximal effort. We then asked them to place their arms across their chests and isometrically contract the quadriceps muscles 2 to 3 seconds “as hard and fast as possible,” with a 60-second rest between trials (Figure 1). No oral encouragement was provided during trials. A successful trial occurred when no excessive movement of the hip flexors or trunk extensors or initial countermovement of the torque-time curve was observed.

After collecting the quadriceps strength measurements, we attached a standard retroreflective cluster set to each participant, who wore spandex shorts, a sleeveless top, and her own athletic shoes (Figure 2). Cluster position data were recorded during 3 landing tasks using an 8-camera, motion-capture system (model Optitrack Prime 13; NaturalPoint, Inc) and streamed using a real-time plug-in to The MotionMonitor software (Innovative Sports Training, Inc). Streamed kinematic data were also time synchronized with data from 2 force plates (Bertec Corp) sampled directly using The MotionMonitor software. After all cluster sets were attached, we digitized anatomic landmarks to generate a biomechanical model using The MotionMonitor software. To generate the biomechanical model, segment position data obtained using the clusters were combined with virtual joint center markers created using a standardized digitization protocol. The 3-dimensional coordinates of the hip-joint center were estimated based on digitization of the anterior-superior iliac spine using the method described by Bell et al.¹⁹  The ankle-joint and knee-joint centers were identified as the middle points between the digitized medial and lateral malleoli and between the digitized medial and lateral epicondyles, respectively. Local coordinate systems for the shank, thigh, and pelvis were defined with the positive x-axis directed anteriorly, positive y-axis directed to the left, and positive z-axis directed superiorly. Cluster positions and force-plate data were recorded as participants completed 3 landing tasks after performing at least 3 practice trials of each task. Three successful trials of each task were collected, with at least 60 seconds of rest allowed between trials.

Participants stood atop a 30-cm box placed at a distance equal to 50% of their height from the force plate. They were instructed to jump forward using both feet, perform a double-legged (DL) landing with each foot completely positioned on a separate force plate, and then jump vertically for maximal height.²⁰  Trials were considered successful if these conditions were met and the movement was completed smoothly (Figure 3).

Using both feet, participants jumped from the same box positioned at 25% of their height from the force plate, performed a single-legged (SL) landing with the foot of the test limb placed completely on a force plate, and jumped vertically for maximal height. Trials were considered successful if these conditions were met and the movement was completed smoothly (Figure 4).

During the side-cutting (SC) task, participants were instructed to run 7 m at a self-selected speed and then perform a side-step cutting maneuver “as quickly as possible” at a 45° angle.²¹  Trials were considered successful if they ran with a natural movement pattern and cut immediately after the foot of the test limb contacted the force plate (Figure 5).

For the quadriceps strength measurements and SL jump-landing and SC tasks, the order of the involved or uninvolved limb in the ACLR group and nondominant or dominant limb in the control group was counterbalanced. Participants completed the 3 tasks in counterbalanced order.

The raw voltage signal from the Biodex System 3 dynamometer was sampled at 2000 Hz and stored on a personal computer equipped with a data-collection system (model MP100; BIOPAC Systems, Inc). LabVIEW (National Instruments) custom computer software was used to analyze the data. The torque signals were filtered using a fourth-order, low-pass Butterworth filter with a cutoff frequency of 10 Hz. We calculated the RTD by inserting a line of best fit to the torque-time curve between torque onset, which was defined as the point when torque exceeded 2.5% of the peak torque of that trial,¹³  and 100 milliseconds after onset. The RTD and maximal voluntary isometric contraction values were normalized by body mass (× kg⁻¹) and averaged across the 3 trials before statistical analysis.

Kinematic and kinetic data were collected at 150 Hz and 1500 Hz, respectively, and filtered using a fourth-order, low-pass Butterworth filter with a cutoff frequency of 12 Hz.²²  Kinematic data were resampled at 1500 Hz and time synchronized to the kinetic data. The knee-joint angles were determined as Euler angles based on the shank reference frame relative to the thigh reference frame rotated in an order of flexion-extension (y-axis), valgus-varus (x-axis), and internal-external rotation (z-axis). Net internal knee-joint moments were calculated in The MotionMonitor software by using filtered kinematic, kinetic, and anthropometric data via the inverse-dynamics approach described by Gagnon and Gagnon.²³ 

We applied a cubic spline interpolation for missing data. Custom computer software (LabVIEW) was used to identify the biomechanical variables of interest during the initial 100 milliseconds after IC, which was defined as the time when the vGRF exceeded 10 N. Loading rate was calculated as the peak vGRF divided by the time from IC to peak vGRF. Loading rate, vGRF, and net sagittal-plane internal knee-joint moments were normalized by body mass (× kg⁻¹).⁹ Cutting velocity along the line of progression was identified during each SC trial and operationally defined as the velocity of the center of mass of the sacrum at IC.

The average of the 3 trials was used for statistical analysis. Limb symmetry for each variable during each functional task and SC velocity was calculated as the ACLR limb minus the uninvolved limb for the ACLR group and the nondominant limb minus the dominant limb for the control group.

Participant characteristics were compared between the ACLR and control groups using independent-samples t tests. We performed separate paired-samples t tests to evaluate whether any differences in quadriceps RTD magnitude or any of the landing mechanics variables of interest existed between the nondominant and dominant limbs of the control group.

Given that cutting velocity has been reported to possibly affect lower extremity movement patterns and biomechanics,²⁴  we controlled for this during the SC task by adding the difference in SC velocity between limbs as a covariate to the models. All statistical analyses were performed using SPSS (version 25.0; IBM Corp), and the α level was set a priori at ≤.05.

Participant characteristics, descriptive statistics, and comparisons between groups are presented in Table 1. Height, mass, and body mass index were not different between groups. The control group was on average 2 years older (P = .04) and had better self-reported knee function than the ACLR group (P < .001), as indicated by their IKDC 2000 and KOS-ADLS scores. No differences in quadriceps RTD or any landing mechanics variables of interest were identified between the nondominant and dominant limbs of the control group (P values > .05). The descriptive statistics for quadriceps maximal voluntary isometric contractions and RTD magnitude and limb symmetry are shown in Table 2. The descriptive statistics for sagittal-plane knee moment at IC, peak vGRF, and loading rate during 3 different landing tasks are provided in Table 3.

The stepwise linear regressions indicated that limb symmetry in the sagittal-plane knee moment at IC, peak vGRF, and loading rate during the 3 landing tasks could not be predicted by quadriceps RTD magnitude, group, or their interaction, nor could landing symmetry during the SL task. In contrast, we found an interaction between group and limb symmetry in quadriceps RTD, which predicted 21% of the variance in limb symmetry in the sagittal-plane knee moment at IC during the DL jump landing (model: R² = 0.21, P = .004; sagittal-plane knee moment at IC = 0.005–0.016 [group × limb symmetry in quadriceps RTD]; Figure 6). We estimated that in the ACLR group, a 1-Nm × s⁻¹ × kg⁻¹ increase in limb symmetry in the quadriceps RTD (ie, reducing the difference in the quadriceps RTD between the ACLR and uninvolved limbs) was associated with a 0.016-Nm × kg⁻¹ reduction in the magnitude of the difference in the sagittal-plane knee moment at IC between limbs.

We investigated the influences of quadriceps RTD magnitude and limb symmetry on limb symmetry in landing mechanics, which have been linked to a greater risk for a second ACL injury,^7–9  in females with or without ACLR during 3 landing tasks. In the ACLR group, greater limb symmetry in the quadriceps RTD was associated with a more symmetric sagittal-plane knee moment at IC during the DL but not the SL task. Contrary to our hypothesis, the quadriceps RTD magnitude, previous injury status, and an interaction between quadriceps RTD magnitude and previous ACLR did not predict higher-risk limb asymmetries in landing mechanics.

Limb asymmetries in peak quadriceps muscle strength²⁵  and explosive quadriceps strength¹⁶  have been reported after ACLR. The influence of limb symmetry in peak quadriceps muscle strength on limb symmetry in landing mechanics has been studied previously.²⁵  Schmitt et al²⁵  found that restoring peak quadriceps strength symmetry (limb symmetry index ≥90%) in individuals with ACLR facilitated more symmetric landing strategies that closely mirrored those of individuals without ACLR. In contrast, individuals with ACLR who had less peak quadriceps strength symmetry (limb symmetry index <85%) demonstrated a compensatory landing strategy wherein the peak knee-extension moment and peak vGRF were reduced in the ACLR limb but peak vGRF and loading rate were increased in the uninvolved limb.²⁵  Considering the importance of generating a knee-extension moment quickly through quadriceps muscle contraction, developing sufficient and symmetric explosive quadriceps muscle strength may facilitate safer landing mechanics. To our knowledge, the influence of explosive quadriceps strength magnitude or limb symmetry on limb symmetry in landing mechanics has not been explored before. Based on our analysis, for the ACLR group, greater limb symmetry in quadriceps RTD was associated with a more symmetric sagittal-plane knee moment at IC during the DL but not the SL jump landing. We estimated that in the ACLR group, a 1-Nm × s⁻¹ × kg⁻¹ increase in limb symmetry in quadriceps RTD was associated with a 0.016-Nm × kg⁻¹ reduction in the magnitude of the differences in the sagittal-plane knee moment at IC between limbs.

A quadriceps RTD capacity that was closer to that of the uninvolved limb in the ACLR group was associated with greater symmetry in sagittal-plane knee-moment magnitude at IC during the DL landings. This result was particularly clinically relevant because Paterno et al⁹  reported that greater limb asymmetry in the sagittal-plane knee moment at IC during DL landings predicted a second ACL injury following the return to play after a primary ACLR. They observed that individuals with ACLR who went on to sustain a second ACL injury (second-injury group) exhibited an internal knee-extension moment at IC on the uninvolved limb. The ACLR limb in these participants and in both limbs of the females who did not sustain a second ACL injury (first-injury group) demonstrated an internal knee-flexion moment at IC.⁹  Thus, the second-injury group had 4.1-fold greater asymmetry in the magnitude of the internal sagittal-plane knee moment at IC (0.123 Nm × kg⁻¹) than in the first-injury group (0.03 Nm × kg⁻¹).⁹ 

Given that our result supports the concept that having a more symmetric quadriceps RTD is associated with greater symmetry in the sagittal-plane knee moment at IC, interventions aimed at increasing quadriceps RTD symmetry—generally by improving quadriceps RTD deficits in the ACLR limb compared with the uninvolved limb—might contribute to a reduction in the second ACL injury risk. Moreover, the magnitude of the RTD symmetry improvement required to potentially reduce the second ACL injury risk is likely clinically attainable. Using the results of our regression equation, an increase in quadriceps RTD symmetry of 5.5 Nm × s⁻¹ × kg⁻¹ would reduce the magnitude of limb asymmetry in the sagittal-plane knee moment at IC by 0.093 Nm × kg⁻¹, which is the magnitude of the limb asymmetry difference between participants who did and those who did not sustain a second ACL injury in the study of Paterno et al.⁹  Although an improvement of this magnitude of quadriceps symmetry would be substantial, it does appear possible given the range of quadriceps RTD symmetry noted in our participants.

Among our participants with ACLR, 53% exhibited quadriceps RTD deficits in their ACLR limb that ranged from 0.11 to 13.86 Nm × s⁻¹ × kg⁻¹ compared with their uninvolved limb. Quadriceps RTD deficits greater than 5 Nm × s⁻¹ × kg⁻¹ were evident in the ACLR limbs of 3 participants compared with their uninvolved limb; their quadriceps RTD asymmetry ranged from 5.12 to 13.86 Nm × s⁻¹ × kg⁻¹. As discussed earlier, an increase in quadriceps RTD symmetry of 5.5 Nm × s⁻¹ × kg⁻¹ would reduce the magnitude of limb asymmetry in the sagittal-plane knee moment at IC by 0.093 Nm × kg⁻¹, which is the magnitude of the limb asymmetry difference between participants who did and those who did not sustain a second ACL injury in the research by Paterno et al.⁹  Given that more than half of our participants with ACLR exhibited quadriceps RTD symmetry and 3 of them demonstrated quadriceps RTD deficits in their ACLR limb compared with their uninvolved limb, the examination of quadriceps RTD in this population is indicated. Our findings support the need to evaluate quadriceps RTD in order to achieve limb symmetry during rehabilitation and before return to play after ACLR. After ACLR, we recommend that clinicians include interventions such as whole-body vibration training²⁶  and plyometric training²⁷  aimed at increasing quadriceps RTD to improve and achieve limb symmetry with the goal of preventing a second ACL injury.

Although greater limb symmetry in quadriceps RTD in the ACLR group was associated with a more symmetric sagittal-plane knee moment at IC during the DL jump landings, we did not identify quadriceps RTD asymmetry as a predictor of asymmetry in DL landing mechanics in control participants. Furthermore, quadriceps RTD asymmetry was not a predictor of any of the 3 asymmetry variables of interest during the SL tasks in either the ACLR or control group. During the DL landings, none of the asymmetry in landing mechanics in our control participants was predicted by quadriceps RTD asymmetry, which likely resulted from the lack of quadriceps RTD asymmetry: 95% of the control group exhibited limb differences in quadriceps RTD within 5 Nm × s⁻¹ × kg⁻¹. For the unexpected findings during the SL tasks, it is possible that the lack of a relationship between quadriceps RTD symmetry and symmetry in landing mechanics was driven by the nature of the SL task. During the SL landings, females with ACLR and limb symmetry in quadriceps RTD may not have been able to use a compensatory landing mechanism by shifting the load to the uninvolved limb, as is possible during DL landings.

Unexpectedly, we did not find that quadriceps RTD magnitude predicted higher-risk limb asymmetries in landing mechanics during the DL and SL tasks. Thus, measuring the magnitude of quadriceps RTD is not useful for detecting limb asymmetry in landing mechanics. We observed a wide range of quadriceps RTD magnitudes in the ACLR limbs and the nondominant limbs of the females without ACLR. Moreover, participants with quadriceps RTD symmetry could be on either end of the spectrum, with low or high quadriceps RTD magnitude in the ACLR limbs or the nondominant limbs of the females without ACLR (Figure 7). In contrast, other participants exhibited either high or low quadriceps RTD magnitude on 1 side but quadriceps RTD asymmetry (Figure 7). During the DL landings, for instance, females with low quadriceps RTD magnitude in the ACLR limb could have either similar quadriceps RTD magnitude on the uninvolved side (symmetric quadriceps RTD) or higher quadriceps RTD magnitude on the uninvolved side (asymmetric quadriceps RTD) that was associated with an asymmetric sagittal-plane knee moment at IC. Therefore, knowing only the quadriceps RTD magnitude on a limb is insufficient for predicting limb asymmetry in landing mechanics.

Generating a sufficient internal knee-extension moment is essential for deceleration during landing.¹⁰  As discussed earlier, during SL landings, females with ACLR and limb symmetry in quadriceps RTD may not have been able to use a compensatory landing mechanism by shifting the load to the uninvolved limb, as was possible during the DL landings. During the SL landing tasks, the capacity to generate a sufficient internal knee-extension moment quickly enough (ie, quadriceps RTD magnitude) to meet the demand of the task may have had a greater influence on landing mechanics than RTD symmetry. For example, individuals with limb asymmetry in quadriceps RTD may have had sufficient quadriceps RTD magnitude in both limbs to meet the given landing demand such that they may not have exhibited asymmetry in landing mechanics during SL landings. On the other end of the spectrum, individuals with symmetric quadriceps RTD of insufficient magnitude in both limbs might have completed the SL landings using symmetric, but less safe, landing mechanics that reduced the magnitude of the internal knee-extension moment that must have been generated quickly. In this case, measuring limb symmetry in quadriceps RTD may not have been appropriate for detecting individuals who used symmetric landing mechanics but whose mechanics during the SL landings might have been associated with a greater ACL injury risk. Instead, greater quadriceps RTD magnitude is likely more critical for landing safely on the limb during the SL task, and future research is needed to investigate the relationship between quadriceps RTD magnitude and the mechanics of the landing limb during SL tasks.

One possible limitation of this study is that we chose to match the nondominant limbs of control participants with the operated limbs of participants with ACLR. Although matching to the nondominant limb instead of the dominant limb of control participants might have confounded our results, we do not believe this to be the case. Our control participants exhibited no differences in quadriceps RTD or any of the landing mechanic variables of interest, which was consistent with the results of a systematic review²⁸  in which limb differences in landing mechanics among uninjured individuals were examined. Given these findings, our choice to not control for the limb dominance of the participants with ACLR likely did not introduce a confounding factor into our analysis. In addition, whereas our focus was exclusively on knee sagittal-plane landing mechanics, ACL injuries are probably the result of a multijoint or multiplanar mechanism of injury.²⁹  Carcia et al²⁹  reported that females who had a longer time to peak force as measured during an isometric closed kinetic chain unilateral squat test were more likely to display valgus kinematic patterns during landing. Also, Cronin et al³⁰  showed that females who had greater hip-extensor RTD demonstrated lower hip-adduction and knee-valgus displacements during an SL jump cut. These outcomes suggested that being able to generate force rapidly through triple joint extension (ie, hip, knee, and ankle) can provide dynamic knee stability in the frontal plane. Future studies should evaluate the influence of quadriceps RTD as well as hip-extensor and ankle plantar-flexor RTD on frontal- and transverse-plane biomechanics to determine if RTD improvements in these joints can facilitate the use of safer landing mechanics. Similarly, we did not evaluate landing mechanics in other joints (eg, ankle, hip, and trunk), so further work is needed to assess whether quadriceps RTD or limb symmetry in quadriceps RTD is associated with compensatory movement patterns that manifest in potentially deleterious movements and loadings on other parts of the kinetic chain.

Quadriceps RTD magnitude in the ACLR limb and the nondominant limb of control participants did not predict asymmetries in the DL and SL jump-landing mechanics associated with higher ACL risk. However, greater limb symmetry in quadriceps RTD in the ACLR group was associated with landing with a more symmetric net sagittal-plane knee moment at IC during the DL jump landings, a finding that was consistent with a lower risk of a second ACL injury. Therefore, clinicians should consider including explosive quadriceps muscle-strengthening exercises after ACLR to promote limb symmetry in quadriceps RTD with the aim of preventing a second ACL injury.

This research was funded by grant No. 1718DGP08 from the National Athletic Trainers' Association Research and Education Foundation (Dr Huang). We thank Emma Ruth Gibbs, Daniel V. Radu, and Noah J. Levine for their help with participant recruitment and data collection.