Single-Legged Triple-Hop Propulsion Strategies in Females With and Those Without a History of Anterior Cruciate Ligament Reconstruction

Participants

A total of 38 female participants (ACLR group = 19; control group = 19) between the ages of 16 and 30 volunteered for this study. Participants with ACLR had undergone unilateral ACLR (n = 16 with ipsilateral hamstring autograft, 1 with ipsilateral bone-patellar tendon-bone autograft, and 2 with allograft); had received full clearance for activities within the past 2 years (20.05 ± 9.50 months post–ACLR; they were 10.68 ± 6.47 months after clearance) from their operating surgeon; and had passed a comprehensive screening battery.¹⁷ All individuals were recreationally active (ie, 150 min/wk of moderate to vigorous physical activity) and free of any injury or condition that limited their physical activity.¹⁷ All participants and their legal guardians provided written informed assent and consent, as appropriate, and the study was approved by the Institutional Review Board of Oregon State University.

Procedures

After the screening process, the height and mass of each participant were measured and recorded before limb dominance was identified. Limb dominance was determined as the limb that participants selected to complete 2 of the following 3 tasks: (1) box step-up, (2) kicking a ball, and (3) stepping strategy after an unexpected perturbation.¹⁷ After limb dominance was identified, participants completed a 5-minute warmup on a stationary bicycle at a self-selected speed.¹⁷

Participant Preparation and Digitization

After the warmup period, we affixed 8 standard retroreflective marker clusters over the following anatomic segments: thoracic spine; sacrum; and bilaterally over the anterolateral thigh, shank, and dorsum of the foot.¹⁷ Lower extremity kinematics and kinetics were assessed using an 8-camera, optical motion-capture system (model Optitrack Prime 13; Natural-Point, Inc) interfaced with a single force plate (model 3210012; Bertec Corp).¹⁷ Kinematic data were then time synchronized with the kinetic data collected using 2 force plates (Bertec Corp). We created a segment-linkage model of the lower extremities and pelvis using The MotionMonitor software (version 9.32; Innovative Sports Training Inc) by digitizing the medial and lateral malleoli, tip of the second phalanxes and medial and lateral femoral epicondyles bilaterally and the anterior-superior iliac spine. The ankle- and knee-joint centers were identified using the midpoints of the digitized medial and lateral malleoli and femoral epicondyles, respectively.¹⁷ The hip-joint center was identified using the method of Bell et al.¹⁸ The global and local coordinate systems for the shank, thigh, and pelvis were aligned with an anteriorly and forward-directed positive x-axis, a left-directed positive y-axis, and a superiorly and upward-directed positive z-axis.¹⁷

Single-Legged Triple-Hop Assessment

Participants stood with a standardized foot position alongside the start of a 6-m line and were instructed to hop 3 consecutive times without pause in a straight line for maximal distance (Figure 1). Trials were considered successful if they completed the 3 hops without the other limb contacting the ground and by “sticking” and holding the final landing for at least 2 seconds.¹⁹ Participants completed the single-legged triple hop under 2 conditions: (1) starting on the force plate to assess propulsion of the first hop and (2) landing on the force plate from the first hop to assess the subsequent propulsion of the second hop. For trials beginning on the force plate, a researcher (C.M.S.M.) marked the location of the heel after the landings from the initial hop and final hop and then measured the initial hop distance and total hop distance in centimeters. For trials landing on the force plate, the average of the first hop distance was used to reposition the start line so that the participant landed on the force plate from the first hop. All were given the same instructions as for the first condition. If they did not successfully land on the force plate, the trial was discarded, and if necessary, the start line was repositioned. Hop distance demonstrated excellent reliability²⁰ between conditions (first hop distance: intraclass correlation coefficient [2,k] = 0.983, SEM = 0.013 body heights; total distance: intraclass correlation coefficient [2,k] = 0.987, SEM = 0.046 body heights). Participants were given at least 1 practice trial bilaterally before completing 3 successful trials under each condition with each limb. Limb sequence was counterbalanced, and participants rested for ≥60 seconds between trials.

View Figure

• Download as Powerpoint
• Download Figure

Data Sampling, Processing, and Reduction

Kinematic and kinetic data were collected at 150 and 1500 Hz, respectively, and filtered using a fourth-order, low-pass Butterworth filter with a cutoff frequency of 12 Hz.¹⁷ Joint angles were calculated as Euler angles (YX′Z″ rotation sequence) based on the distal reference frame relative to the proximal reference frame rotated in an order of flexion-extension (y-axis), valgus-varus (x-axis), and internal-external rotation (z-axis). Net internal hip, knee, and ankle moments were calculated via an inverse-dynamics approach described by Gagnon and Gagnon²¹ in The MotionMonitor software using filtered kinematic, kinetic, and anthropometric data.

A custom computer software program (LabVIEW, National Instruments Corp) was used to calculate mechanical joint work and the contributions from the hip, knee, and ankle joints during the first and second hops of the single-legged triple hop. The propulsive phase was defined as the time from peak knee flexion to toe-off, which was operationally defined as the instant that the vertical ground reaction force was <10 N. Joint power (P) was calculated by multiplying the sagittal-plane net joint moment (M) and angular velocity (ω) for each trial (P = M × ω). The positive portions of the joint power curves (ie, when the angular velocity and net joint moment were in the same direction, indicating concentric muscle action) were integrated to quantify positive mechanical joint work during each propulsive phase.²² Total mechanical joint work was calculated by summing the work across joints with individual joint contributions to total work expressed as a percentage. Individual-joint moment impulses were calculated by integrating the area under each joint’s moment-time curve during which the joint exhibited a net extensor–plantar-flexor moment. Total support moment impulse was calculated by summing the calculated extensor–plantar-flexor–moment impulses across joints. Joint contributions to total support moment impulse were expressed as a percentage. All mechanical joint work and moment-impulse values were normalized to the product of body weight and height (× [N·m]⁻¹), while initial and total hop-distance values were normalized by body height (× cm⁻¹). Limb symmetry index values were quantified by dividing the hop distance of the reconstructed and nondominant limbs by the hop distance of the uninvolved and dominant limbs for participants with and those without ACLR, respectively. These values were multiplied by 100 and expressed as a percentage. Outcome variables were averaged across the 3 hop trials before statistical analysis.

Statistical Analysis

We used dependent-samples t tests to compare initial hop distance, total hop distance, total mechanical joint work, and total support moment impulse between the reconstructed and uninvolved limbs of participants with ACLR. Independent-samples t tests were calculated to assess group differences in participant characteristics, hop distance LSI, single-legged triple-hop performance, total mechanical joint work, and total support moment impulse between the reconstructed limb in participants with ACLR and the nondominant limb in control participants. Separate 2 (limb: ACLR-involved [reconstructed] and ACLR-uninvolved [uninvolved])-by-3 (joint: hip, knee, and ankle) repeated-measures analysis-of-variance (ANOVA) models were conducted to determine the influence of previous ACLR, joint, and their interaction on the magnitude of mechanical joint work, moment impulse, and joint contributions to total work and moment impulse. Additionally, separate 2 (group: ACLR and control)-by-3 (joint: hip, ankle, and knee) mixed-model ANOVAs were used to evaluate the influence of previous ACLR, joint, and their interaction on these same outcome variables between the reconstructed limb in participants with ACLR and the nondominant limb of control participants. We selected the nondominant limb based on evidence that no biomechanical or performance differences were present in similar functional tasks between limbs in healthy individuals.²³ Interaction or joint main effects were identified via planned pairwise comparisons with a Tukey honestly significant difference test. The α was set a priori at ≤.05. All statistical analyses were performed using SPSS (version 26; IBM Corp).

Between-Limbs Comparisons: ACLR Group

After ACLR, participants jumped approximately 4% farther on the first hop (t₁₈ = −2.62, P = .03) and in total (t₁₈ = −3.14, P = .04) with their uninvolved limb than their reconstructed limb (Figure 2). Furthermore, the reconstructed limb generated less total mechanical joint work than the uninvolved limb during the propulsive phase of the first (t₁₈ = −3.73, P = .002) and second (t₁₈= −2.55, P = .02) hops (Figure 2). Total support moment impulse did not differ between limbs during the first (t₁₈ = −0.44, P = .67) or second (t₁₈ = −0.32, P = .76) hops (Figure 2). Results from the repeated-measures ANOVA for mechanical joint work and moment-impulse magnitudes and the relative joint contributions are presented in Tables 1 and 2 and Figures 3 and 4, respectively. The reconstructed limb generated 11.3% less work at the ankle than the uninvolved limb during the first hop (t₁₈ = −4.31, P = .002; Figure 3A). Additionally, the knee of the reconstructed limb produced 19.3% (t₁₈ = −2.33, P = .03) and 27.3% (t₁₈ = −4.47, P < .001) less work during the first and second hops, respectively, than the uninvolved knee (Figure 3B). In addition, the knee joint of the reconstructed limb contributed less to the total mechanical work than the knee of the uninvolved limb during the second hop (t₁₈ = −4.02, P = .01; Table 3, Figure 4B). In both limbs, the ankle plantar flexors produced more mechanical joint work and moment impulse and contributed more to the mechanical work and total support moment impulse than the knee and hip extensors (P < .001; Table 3, Figures 3C and D and 4).

View Figure

• Download as Powerpoint
• Download Figure

Table 1 Mechanical Joint Work and Total Support Moment Impulse Magnitudes and Joint Contributions During Single-Legged Triple-Hop Propulsion in Females With and Those Without Anterior Cruciate Ligament Reconstruction

View Fullscreen

Table 2 Post Hoc Analysis for Significant Limb × Joint Interactions Identified During Single-Legged Triple-Hop Propulsion in the Anterior Cruciate Ligament Reconstruction Group

View Fullscreen

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Table 3 Relative Joint Contributions to Mechanical Joint Work and Moment Impulse During Single-Legged Triple-Hop Propulsion, Mean ± SD

View Fullscreen

Between-Groups Comparisons: ACLR Versus Control Group

The LSI of total hop distance did not differ between groups (ACLR: 96.2% ± 6.3%; control: 98.8% ± 5.3%; t₃₆ = −1.35, P = .19). Single-legged triple-hop performance did not differ between the reconstructed limb in the ACLR group and the nondominant limb in the control group in initial (t₃₆ = 0.13, P = .72) or total (t₃₆ = −0.20, P = .77) hop distance. No differences in total work were identified between groups in either first-hop (t₃₆ = −0.74, P = .47) or second-hop (t₃₆ = −0.98, P = .33) propulsion. Similarly, no differences were seen in total support moment impulse between groups during the first (t₃₆ = −0.54, P = .70) or second (t₃₆ = −0.39, P = .51) hop (Figure 2). No group-by-joint interaction or group main effects for the mechanical joint work and moment impulse magnitudes and joint contribution were identified in the first or second hop of a single-legged triple hop (Table 1). However, in the first and second hops, the ankle generated more work and moment impulse and contributed more to total work and moment impulse than the knee and hip in both the ACLR and control groups (P < .001; Table 3, Figures 3 and 4).

Between-Limbs Comparisons: ACLR Group

We found a slight difference between hop distances in the ACLR group, with participants hopping approximately 4% farther on the uninvolved limb (0.69 ± 0.09 and 2.27 ± 0.36 body heights for first and second hops, respectively) than the reconstructed limb (0.67 ± 0.11 and 2.19 ± 0.34 body heights for first and second hops, respectively). Yet these differences were not clinically meaningful, as mean LSI values exceeded 95% in the ACLR group, surpassing clinical recommendations of 90% for return to play.^10,24 The reconstructed limb exhibited differences in the magnitude of total work during both propulsive phases of the single-legged triple hop, generating approximately 11% less total work during the initial propulsive phase and approximately 12% less total work during the second propulsive phase (Figure 2). These differences were driven by the reconstructed knee generating 19.3% and approximately 27.3% less work than the uninvolved knee during the propulsive phase of the first and second hop, respectively (Figure 3B). In addition to generating less mechanical joint work, the reconstructed knee contributed less to the total work than the uninvolved knee during the propulsive phase of the second hop of a single-legged triple hop (Table 3, Figure 4). With the mean LSI exceeding 95% and differences in mechanical joint work identified between limbs, our results provide further evidence that the LSI does not detect the presence of mechanical differences in the reconstructed limb and is an insufficient basis for clinical decisions regarding restored lower extremity function in females post–ACLR.

Given that single-legged hop performance is commonly used to assess lower extremity function and has been moderately associated with quadriceps strength,^7,25 it was an unexpected finding that the ankle plantar flexors were such disproportionate contributors to the total work and support moment during single-legged triple-hop propulsion. Interestingly, this ankle-dominant strategy remained consistent in both limbs and across the first and second hops of the task. Because distance traveled is largely a function of one’s ability to generate work to propel oneself forward, the ankle-dominant strategy may explain why clinically meaningful deficits in mechanical joint work in the reconstructed knee during the second hop did not result in clinically detectable deficits in hop performance.

No differences were identified between limbs in moment impulse or relative joint contributions during the propulsive phase of the first or second hop of a single-legged triple hop (Figures 2 and 3). During the first hop, the reconstructed knee generated only approximately 81% of the work while producing approximately 95% of the moment impulse compared with the uninvolved knee (Figure 3A and C). Similarly, the reconstructed knee produced only approximately 73% of the work despite generating approximately 89% of the moment impulse during the second propulsive phase (Figure 3B and D). Although our study could have been underpowered to detect a difference in moment impulse during the second hop, the 11% difference in moment impulse identified bilaterally accounts for just more than one-third of the approximately 27% decrease in work at the knee. Together, our results suggest not that decreases in mechanical joint work were driven by less net joint-moment production but that less joint angular velocity in the reconstructed limb during propulsion was the greatest contributor to less mechanical joint work. Whereas peak quadriceps strength (which has been reported previously¹⁷) and knee-extensor moment impulse did not differ between limbs in females with ACLR, it is possible they did not harness this capacity and extend their knee as quickly during the task, resulting in decreased mechanical joint work. Therefore, it appears that deficits in mechanical joint work at the knee reflect not simply a peak strength concern but an underlying functional deficit in which strength is not effectively used during single-legged triple-hop propulsion. Based on our findings, clinicians should focus on addressing not only peak strength but also explosive and plyometric strength to help better restore lower extremity function post–ACLR.²⁶

Our research substantiates the conclusions of Barfod et al⁸ that hop-test LSI was not a valid indicator of restored quadriceps function post–ACLR. Furthermore, given the widespread use of the LSI to interpret hop-test performance, our results may provide context to previous work from Wellsandt et al,¹¹ who found that 8 of 11 athletes who sustained a second ACL injury achieved ≥90% LSIs in 4 functional hop tasks. These athletes may have had clinically meaningful mechanical deficits that were not detected via the LSI during a functional hop test to determine lower extremity function. Also, our results are largely consistent with those of Kotsifaki et al,¹⁶ who investigated mechanical joint work during a single-legged hop for distance in male athletes with a history of ACLR. Even with a different functional hop task in male athletes, a consistent finding was that variations in mechanical joint work were identified bilaterally that were not detected using the LSI. This reaffirms that hop-test distance may not be an appropriate indicator of restored lower extremity function. Interestingly, our results differ from those of Kotsifaki et al¹⁶ in the joint-to-joint contributions to propulsion. Our participants demonstrated a consistent ankle-dominant strategy throughout, while participants in the study of Kotsifaki et al¹⁶ generated work predominantly through the hip and ankle joints. Although variations in the demands of the task may contribute to these differences, it is possible that females use their hip extensors to a lesser extent than their male counterparts. Researchers²⁷ have suggested that a lack of neuromuscular control in the hip musculature may be a contributing factor to the increased ACL injury risk in females. However, we did not evaluate hip-extension strength or activation patterns, and this notion should be examined in future work.

Between-Groups Comparisons: ACLR Versus Control Group

Contrary to our hypothesis, no differences in work, moment impulse, or relative joint contribution were observed between the ACLR and control groups. Both groups displayed an ankle-dominant propulsion strategy during both propulsive phases, but no differences were present in work or moment impulse at the knee. One possible explanation is that the ACLR group was exposed to a rehabilitation protocol in which both the reconstructed and uninvolved limbs received training in functional tasks such as the single-legged triple hop. Whereas all participants in this study were recreationally active, whether the control group had been exposed to this type of training in the past is unknown. Nonetheless, the ankle-dominant propulsion strategy during both hops in the control group indicates that the ankle-dominant propulsion strategy used by the ACLR group was not attributed to injury status.

Clinical Implications and Future Research

Our results lend support to previous work^8,11,13,16 and further highlight the need for clinicians to exercise caution when interpreting results from a functional hop task post–ACLR. Functional hop tests are easy to administer, yet simply evaluating single-legged triple-hop distance and the LSI alone does not provide the appropriate insight into lower extremity function on which to base clinical return-to-play decisions. In addition to focusing on explosive and plyometric strength throughout rehabilitation,²⁶ clinicians should consider assessing single-legged vertical jump performance to better identify functional deficits post–ACLR.²⁸ Furthermore, a quality-of-movement assessment during various sport-specific plyometric and agility tasks should also be considered to better make informed decisions regarding return to play post–ACLR. Given the deceleration demands during the landing phases of a single-legged triple hop, this task could be useful in a return-to-play test battery if the quadriceps play a larger role during landing. Future researchers should determine whether certain modifiable movement characteristics (eg, knee position at initial contact, knee-flexion displacement) during the landing phase of a single-legged triple hop are associated with improved lower extremity function, as shown in other tasks.^29,30 If such a relationship exists, clinicians would be advised to evaluate these modifiable movement characteristics using a quality-of-movement assessment instead of quantifying the LSI from the distance traveled during a single-legged triple hop.

Limitations

The primary limitation of this study was that the propulsive phase during the first and second hops was measured during separate trials. However, because of the excellent reliability demonstrated between these conditions, we do not believe that capturing the different propulsive phases in separate trials affected our results. It is possible that the quadriceps are used to a greater extent during the landing phase, and negative mechanical joint work should be examined in a future study. In addition, we did not evaluate the third hop of the single-legged triple hop. Propulsion strategies during the third hop may differ from the strategies observed in the first 2 hops and should also be evaluated in a future study. Moreover, only females were included in this investigation. Provided that recurrent ACL injury affects both sexes, future investigators should see if these results can be replicated in male athletes. However, our findings and those of Kotsifaki et al¹⁶ suggest these results may not be sex specific. Finally, we did not match our control participants on a one-to-one basis or control for the types of activity in which participants engaged. Still, no differences in activity level or demographics other than age (on average, 1.9-year difference) were identified. Therefore, we are confident that the main difference between our groups was a previous history of ACLR.