2D AND 3D KINEMATICS DURING LATERAL STEP-DOWN TESTING IN INDIVIDUALS WITH ANTERIOR CRUCIATE LIGAMENT RECONSTRUCTION

Abstract

Background

The lateral step-down test is an established clinical evaluation tool to assess quality of movement in patients with knee disorders. However, this test has not been investigated in individuals after anterior cruciate ligament reconstruction (ACLR) in association with quantitative 3D motion analysis.

Purposes

The purpose of this study was to determine the strength of association between visually-assessed quality of movement during the lateral step-down test and 3D lower limb kinematics in patients with history of ACLR. A second purpose was to compare kinematics between subgroups based on the presence or absence of faulty alignments during the task. The final purpose was to compare visually-assessed quality of movement scores between box heights during lateral step-down testing.

Methods

Twenty subjects at least one year status post-ACLR (18 females, age of 24.5 ± 4.6 years and body mass index of 23.4 ± 2.3 kg/m²) performed the lateral step-down test unilaterally on the surgical limb atop four and six inch boxes. A board-certified orthopedic physical therapist scored overall quality of movement during the lateral step-down test using established criteria during 2D video playback. Lower limb kinematics were simultaneously collected using 3D motion capture. An alpha level of 0.05 was used for all statistical treatments.

Results

Overall 2D quality of movement score significantly correlated (r =0.47-0.57) with 3D hip adduction and hip internal rotation across box heights. Across box heights, the presence of faulty pelvic alignment differentiated a subgroup exhibiting less peak knee flexion, and the presence of faulty knee alignment differentiated a subgroup exhibiting greater peak hip adduction. The six inch box elicited worse quality of movement compared to the four inch box.

Conclusions

These results suggest that visually-assessed quality of movement is associated with several kinematic variables after ACLR. 2D movement deviations at the pelvis appear to consistently relate to less knee flexion, and 2D deviations at the knee appear to suggest greater hip adduction. Generally, poorer quality of movement was observed for the six inch box height. Clinically, these data suggest that interventions targeting hip abductor and knee extensor strength and neuromuscular control may be useful in the presence of poor quality of movement during lateral step-down testing.

Level of Evidence

2b

INTRODUCTION

Anterior cruciate ligament reconstruction (ACLR) is the predominant standard of care for a ruptured ACL in the United States. The incidence of ACLR increased from 86,687 patients in 1994 to 129,836 patients by 2006.¹ The incidence rate for secondary ACL injury has been shown to be six times greater than for primary injury.² Movement patterns in these individuals remain abnormal years after ACLR.³ Perhaps of greatest concern, a three-fold increased risk of developing knee osteoarthritis has been shown in the ACLR knee compared to the contralateral knee 14-years post-surgery.⁴ Based on these data, contemporary rehabilitation practices likely remain suboptimal. As a component of the rehabilitative process, clinical evaluation tools capable of assessing quality of movement after ACLR merit further exploration.

The lateral step-down (LSD) test is a well-established clinical assessment of lower extremity quality of movement.^5-16 This test aims to identify faulty movements at the trunk, pelvis and knee during a step-down maneuver off a box. Several rating systems have been adapted for visually scoring the LSD test.^5,6,8 Rabin et al. used a modified version¹³ of previously established criteria⁵ and yielded excellent inter-rater reliability for visually assessing quality of movement in patients with patellofemoral pain syndrome. It has been reported that quality of movement scores may be affected by decreased ankle dorsiflexion range of motion.¹⁴

Associations between visually-assessed quality of movement using the LSD test and findings from an objective measure are relevant in establishing clinical usability of this test. Jones and colleagues¹¹ revealed observational ratings of frontal plane knee position during the LSD test using 2D video were related to the frontal plane projection angle. Rabin et al.¹⁶ showed that faulty pelvis and knee alignments during the LSD test were associated with greater peak knee external rotation, contralateral pelvic drop, and hip adduction during 3D motion capture. However, these studies were conducted on healthy individuals, and it is unclear how visually-assessed quality of movement relates to 3D movement in the ACLR population.

Therefore, the purpose of this study was to determine the strength of association between visually-assessed quality of movement during the lateral step-down test and 3D lower limb kinematics in patients with history of ACLR. A second purpose was to compare kinematics between subgroups based on the presence or absence of faulty alignments during the task. The final purpose was to compare visually-assessed quality of movement scores between box heights during lateral step-down testing.

METHODS

Subjects

Twenty individuals between the ages 18-40 with a history of unilateral ACL reconstruction at least one year prior were recruited from a university setting via electronic and paper advertisements. All subjects completed formal rehabilitation and had been cleared by a physician for return to sport if an athlete. Subjects with any other recent (six months) or current spinal or lower extremity pathology were excluded. This study was approved by the university Institutional Review Board and written informed consent was obtained for all subjects. Height was then measured using a stadiometer upon arrival to the laboratory environment.

Ankle Dorsiflexion Range of Motion

Weight-bearing ankle dorsiflexion range of motion was measured using a lunge test as previously described and found to be highly reliable.¹² In brief, a 1 m piece of tape was fixed on the floor perpendicular to a wall. Subjects stood in bare feet while the researcher placed a black sticker 15 cm distal to the tibial tuberosity. Next, subjects faced the wall and positioned the test foot such that the second toe and heel lay directly over the tape. Subjects then placed the palms of their hands against the wall in front of them and were instructed to lunge as far forward as possible without lifting the heel of the test foot off the ground. The researcher then positioned a smartphone running a digital inclinometer application (iHandy Level, iHandySoft Inc, NY, USA) over the tibial sticker and dorsiflexion range of motion was recorded. Use of this particular inclinometer application has been found to have good intra-rater and inter-rater reliability for spinal measures (ICC > 0.80).­­­¹⁷ Three measures were recorded and averaged for the test limb.

Motion Capture Preparation

An eight-camera motion capture system (VICON, Centennial, CO, USA) was used to collect kinematic data (150 Hz). In preparation for motion capture, 47 reflective markers were placed on the subject's bilateral lower extremities and pelvis: L5-S1 interspinous space, iliac crests, anterior superior iliac spines, greater trochanters, medial and lateral femoral condyles, medial and lateral tibial plateaus, medial and lateral malleoli, first and fifth metatarsal heads, the distal feet, and the proximal, distal, and lateral heels. Four rigid clusters of four tracking markers were placed bilaterally on the distal posterolateral thighs and shanks. Static standing and functional hip motion trials were captured to later build and define segment coordinate systems for the pelvis, thighs, shanks and feet.

Lateral Step-down Test

The LSD test was conducted according to the procedure used by Piva and colleagues.⁵ Subjects performed the test unilaterally on both a four and six inch box. The two conditions were tested in random order. Prior to the first condition, the investigator demonstrated the LSD maneuver. Subjects were instructed to place hands on their hips and stand with the test limb on the edge of the box. The contralateral limb hung off the side of the box with the knee in full extension and foot fully dorsiflexed. To initiate the movement, subjects were instructed to bend the test limb and lower themselves until the contralateral heel tapped a floor-level force plate (BERTEC Corp., Worthington, OH, USA) and return to the starting position. The force data were captured (1500 Hz) to confirm heel touch on each repetition. Subjects were allowed to practice the LSD for familiarization prior to each condition as needed. Six continuous repetitions were performed for each condition at a self-selected pace, with the middle four repetitions extracted for analysis. Two minute rest periods between conditions were used.

Data Processing

Marker trajectories were labelled and gap-filled in Vicon Nexus. Trials were then exported in c3d file format and post-processed using Visual 3D (C-motion, Bethesda, MD, USA) software. Angle data were derived using an X (flexion/extension) -Y (adduction/abduction) -Z (internal/external rotation) Cardan rotation sequence. Kinematic variables of interest included frontal, sagittal, and transverse plane peak hip and knee joint angles, as well as peak ankle dorsiflexion, and peak contralateral pelvic drop. Custom software was used to extract the discrete variables from the data (Labview 2010, National Instruments, Austin, TX, USA).

2D Digital Video Assessment

The LSD test was video recorded at 30 Hz using the camera from an iPhone 5c (Apple Inc., Cupertino, CA, USA). The camera was mounted on a tripod positioned 3 m in front of the subject at a height of 31 inches and in portrait orientation. Quality of movement was scored using a modified version of previously established criteria⁵ assessing five aspects of the movement: arm strategy, trunk alignment, pelvis plane, knee position, and steady stance. Each aspect of the movement was individually scored. A score of 0 was given if the aspect of movement was deemed “not faulty”. A score of 1 was given if movement deviations occurred, deeming the aspect of movement as “faulty”. Movement deviations included: hand removed from waist, trunk leaning in any direction, contralateral pelvic drop or rotation, or wavering on tested limb and/or stepping down on non-tested limb. For knee position, subjects received a score of 1 if the tibial tuberosity was medial to the second toe and 2 if the tibial tuberosity was medial to the entire medial border of the foot (Figure 1). The possible range in aggregate score was 0 to 6. A physical therapist with board certification as an orthopedic clinical specialist rated both LSD conditions.

Figure 1.

“Not faulty” knee alignment scored as a 0.

Figure 2.

“Faulty” knee alignment scored as a 1 as the tibial tuberosity was medial to the 2nd toe but not medial to the medial border of the foot.

Self-report questionnaires

Each subject completed the International Knee Documentation Committee (IKDC) Subjective Knee Evaluation Form, Tegner Activity Scale (TAS), and an intake form. In patients with various knee disorders, the IKDC has been found to be a valid and reliable questionnaire to assess knee symptoms, function, and sports activity.¹⁸ The TAS is an activity rating system with established reliability, validity, and responsiveness for individuals with ACL injuries.¹⁹ The intake form gathered further descriptive information including limb dominance, graft type, surgical and rehabilitation history, age, and sex.

Statistical Analysis

Associations between the aggregate score for quality of movement and the target kinematic variables were assessed using Spearman's rank correlation coefficients. Separate Mann-Whitney U tests were used to compare each kinematic variable by groups determined by the observational ratings for the pelvis and knee.¹⁶ Paired t-tests were used to determine differences in the kinematic variables between the two box heights. A Wilcoxon Signed Ranks test was used to compare the number of movement deviations exhibited between box heights. An alpha level of 0.05 was used for all tests. Data were analyzed using SPSS 23.0 (IBM Corp, Armonk, NY, USA).

RESULTS

Subjects

Descriptive data are presented in Table 1. Weight bearing ankle dorsiflexion range of motion was consistent with previously reported means for healthy individuals with at least a score of 2 on the Piva criteria.^12,13 The average time since surgery was nearly five years. Average current activity level represented participation in recreational or competitive sports at least five times per week.

Table 1.

Subject demographics (mean ± standard deviation)

Sex Age (years)	18 females; 2 males 24.55 ± 4.61
Height (m)	1.71 ± 0.08
Weight (kg)	68.35 ± 9.73
Body mass index (kg/m2)	23.36 ± 2.32
Ankle dorsiflexion range of motion (degrees)	47.48 ± 6.84
Time since surgery (months)	56.80 ± 38.20
International Knee Documentation Committee score	83.16 ± 7.22
Tegner Activity Scale score	6.90 ± 1.94

Spearman's Correlations

Spearman's correlations between overall quality of movement score and kinematic variables are displayed in Table 2. Positive correlations between visually-assessed quality of movement and peak hip adduction and internal rotation were seen during performances on both the four and six inch box. Additionally, correlations were seen between overall quality of movement and contralateral pelvic drop, peak knee flexion, abduction, and internal rotation only on the four inch box.

Table 2.

Spearman's correlations between overall quality of movement score and kinematic variables for the pelvis, hip, and knee at different box heights

	4-inch box		6-inch box
Kinematic variables	Correlation Coefficient	p-value	Correlation Coefficient	p-value
Contralateral pelvic drop	-0.51	0.02*	-0.15	0.54
Hip flexion	0.04	0.86	0.09	0.72
Hip adduction	0.57	< 0.01*	0.51	0.02*
Hip internal rotation	0.55	0.01*	0.41	0.07
Knee flexion	0.75	< 0.01*	0.26	0.26
Knee adduction	0.15	0.52	-0.05	0.82
Knee abduction	0.56	0.01*	0.35	0.14
Knee internal rotation	0.55	0.01*	0.12	0.61

^*indicates statistically significant difference

Kinematics by Knee Position Score(s)

Kinematic variables among individuals with “not faulty” and “faulty” knee positions are displayed in Table 3. A knee position score of 1 did not differentiate any kinematic variables for either box height. A knee position score of 2 differentiated greater peak pelvic drop, hip adduction, and hip and knee internal rotation on the four inch box. A knee position score of 2 differentiated greater hip adduction and less knee abduction on the six inch box.

Table 3.

Kinematic variables presented as mean (SD) among individuals with “not faulty” and “faulty” knee positions, by box height

	Score of 1						Score of 2
Kinematic Variables	4-inch box			6-inch box			4-inch box			6-inch box
	Not Faulty (n = 8)	Faulty (n = 12)	p	Not Faulty (n = 5)	Faulty (n = 15)	p	Not Faulty (n = 17)	Faulty (n = 3)	p	Not Faulty (n = 18)	Faulty (n = 2)	p
Contralateral pelvic drop	−2.0 (2.3	−3.2 (3.2)	0.32	−6.4 (3.6)	−6.1 (3.5)	0.90	−2.1 (2.7)	−6.3 (0.3)	0.02*	−5.7 (3.3)	−9.9 (2.7)	0.10
Hip flexion	29.1 (10.7)	28.5 (7.0)	0.82	39.4 (10.5)	38.6 (7.0)	0.90	28.2 (9.0)	31.3 (4.2)	0.37	38.3 (8.0)	43.3 (2.9)	0.31
Hip adduction	14.9 (3.9)	18.7 (3.7)	0.06	20.8 (3.3)	24.1 (5.7)	0.32	16.3 (3.9)	21.9 (1.2)	0.02*	22.4 (4.9)	30.6 (3.3)	0.04*
Hip internal rotation	5.2 (5.2)	8.3 (5.9)	0.22	6.4 (5.2)	10.0 (5.1)	0.18	5.8 (5.2)	14.3 (0.9)	0.02*	8.3 (4.9)	16.2 (1.4)	0.06
Knee flexion	−47.9 (6.1)	−45.2 (4.1)	0.25	−54.3 (5.2)	−57.9 (6.3)	0.28	−46.7 (5.3)	−43.7 (2.3)	0.32	−56.8 (6.3)	−58.7 (6.1)	0.71
Knee adduction	1.2 (2.4)	−0.2 (2.7)	0.28	1.2 (3.4)	0.6 (2.6)	0.76	0.2 (2.8)	0.8 (1.7)	0.79	0.7 (2.8)	1.6 (2.0)	0.61
Knee abduction	−4.6 (3.5)	−4.8 (2.8)	0.82	−3.8 (4.0)	−4.6 (2.5)	0.63	−5.0 (3.1)	−2.8 (1.6)	0.19	−4.9 (2.5)	0.0 (0.3)	0.03*
Knee internal rotation	4.7 (5.1)	8.3 (5.9)	0.17	4.3 (5.6)	4.1 (4.8)	0.76	5.6 (5.2)	14.3 (0.9)	0.02*	4.7 (4.7)	−0.4 (4.9)	0.13

^*Indicates statistically significant difference

Kinematics by Pelvis Position Score

Kinematic variables among individuals with “not faulty” and “faulty” pelvis positions are displayed in Table 4. Faulty pelvis alignment differentiated less knee flexion for both the four and six inch box conditions, as well as greater hip internal rotation and less knee abduction for the six inch box.

Table 4.

Kinematic variables presented as mean (SD) among individuals with “not faulty” and “faulty” pelvis positions, by box height

	4-inch box			6-inch box
Kinematic variables	Not Faulty (n = 7)	Faulty (n = 13)	p	Not Faulty (n = 6)	Faulty (n = 14)	p
Contralateral pelvic drop	−1.5 (2.8)	−3.4 (2.8)	0.10	−5.3 (3.2)	−6.5 (3.6)	0.56
Hip flexion	27.4 (7.9)	29.4 (8.9)	0.84	42.0 (5.3)	37.4 (8.4)	0.22
Hip adduction	15.3 (4.1)	18.1 (4.0)	0.10	20.1 (2.5)	24.6 (5.7)	0.08
Hip internal rotation	5.2 (6.1)	8.0 (5.5)	0.29	4.3 (2.1)	11.1 (4.8)	<0.01*
Knee flexion	−51.3 (3.8)	−43.6 (3.3)	< 0.01*	−61.2 (4.4)	−55.2 (6.0)	0.03*
Knee adduction	0.1 1.9)	0.4 (3.0)	0.61	−0.2 (2.1)	1.2 (2.9)	0.14
Knee abduction	−6.3 (3.8)	−3.8 (2.1)	0.10	−6.7 (1.7)	−3.4 (2.6)	<0.01*
Knee internal rotation	5.2 (6.1)	7.7 (5.6)	0.36	4.7 (2.1)	3.9 (5.7)	0.87

^*indicates statistically significant difference

Comparisons by Box Height

Kinematic differences between the four and six inch box heights are displayed in Table 5. The six inch box condition displayed significantly greater pelvic drop, hip and knee flexion, hip and knee adduction, and hip internal rotation compared to the four inch box condition. The median scores on the Piva et al.⁵ criteria for the four and six inch box height conditions were 2 and 3, respectively.

Table 5.

Differences in kinematic variables between the 4- and 6-inch lateral step-downs

	4-inch box		6-inch box
Kinematic Variables	Mean	Standard Deviation	Mean	Standard Deviation	p-value
Contralateral pelvic drop	−2.7	2.9	−6.1	3.5	<0.01*
Hip flexion	28.7	8.4	38.8	7.7	< 0.01*
Hip internal rotation	7.1	5.7	9.1	5.2	< 0.01*
Hip adduction	17.1	4.2	23.3	5.3	< 0.01*
Knee flexion	−46.3	5.1	−57.0	6.1	< 0.01*
Knee internal rotation	6.9	5.8	4.2	4.8	0.21
Knee adduction	0.3	2.7	0.8	2.7	0.02*
Knee abduction	−4.7	3.0	−4.4	2.8	0.42

^*indicates statistically significant difference

DISCUSSION

The purposes of the present study were to determine 1) the strength of association between 2D visual observation for quality of movement during the LSD test and lower limb kinematic variables using 3D motion capture in the ACLR population; 2) to compare kinematics between subgroups based on the presence or absence of faulty pelvis and knee alignments; and 3) to compare the number of visually-assessed movement deviations during LSD testing between box heights. Consistent with the first hypothesis, scores for overall quality of movement were shown to be correlated with several kinematic variables associated with ACL injury at the pelvis, hip, and knee. For the subgrouping hypothesis, across both box heights, greater 3D hip adduction and internal rotation was observed in those with faultier 2D knee alignments, and less 3D knee flexion was seen in those with faulty 2D pelvis alignments. Finally, more 2D movement deviations were identified for the six inch test condition, suggesting box height is influential on total movement, reflected in deviation scoring.

As expected, overall 2D quality of movement was associated with 3D joint kinematics. On the four inch box, overall 2D quality of movement score moderately correlated with most of the 3D variables related to ACL injury risk. Interestingly, on the six inch box, overall 2D quality of movement related to only two of the 3D parameters. Generally, across all variables, stronger relationships between overall quality of movement and all kinematic variables were seen for the four compared to six inch box height. This weaker association may be partly explained by the limitations of 2D assessment. As most 3D joint rotations were greater in magnitude (Table 5) on the six inch box, particularly for the proximal assessments at the hip and knee, the potential to amplify 2D perspective errors may have also increased. Of interest, Jones et al.¹¹ reported that the knee abduction angle did not differ across 2D observational rating groups for quality of movement during LSD testing in healthy individuals using box heights of 6-10 inches. This lack of relationship was also evident in the current study. In another study evaluating single-leg squatting, no relationship was found between the 2D knee frontal plane projection angle and 3D knee abduction angle, but rather an association with 3D hip adduction.²⁰ These findings are again supported by the current data on the six inch step. No comparable data using four inch step downs were found in the literature. The only 3D parameters that related to 2D overall quality of movement across box heights were peak hip adduction and peak hip internal rotation, both of which are important in lower extremity injury risk and function. The consistency of these relationships across box heights suggests that 2D assessment of overall quality of movement is sensitive to 3D peak hip adduction and internal rotation movement patterns. Including the current findings, most evidence suggests that 2D perspective errors associated with visual assessment of these tasks are of concern when compared to 3D analysis, but that more robust relationships may exist at the hip when compared to the knee.

During the 2D assessment, the observation that the tibial tuberosity moved past the medial border of the foot yielded a knee position score of 2. This event differentiated individuals after ACLR with increased hip adduction, hip internal rotation, knee internal rotation, and contralateral pelvic drop for the four inch box height and increased hip adduction and decreased knee abduction on the six inch box height. Visualization of a medialized knee was consistently associated with greater 3D hip adduction across box heights. A similar analysis¹⁶ reported greater amounts of knee external rotation with “faulty” knee alignment using a 15 cm box, which is similar to the six inch box height used in the current study. However, Rabin and colleagues did not evaluate the knee position relationship to 3D hip mechanics, did not observe any knee medialization past the medial border of the foot, and their study was conducted on a healthy sample.¹⁶ For the four inch box, individuals after ACLR with knee medialization past the medial border of the foot were differentiated by greater amounts of knee internal rotation. However, no comparable analyses on four inch LSD testing are available.

The presence of “faulty” pelvic alignments consistently differentiated individuals with decreased knee flexion for both the four and six inch box height and increased hip internal rotation and decreased knee abduction for the six inch box height. These results suggest that increased contralateral pelvic drop may be used as a compensatory movement during the LSD test in order to avoid positions of deeper knee flexion. Diminished quadriceps strength has been shown to be associated with decreased knee flexion angles during single-legged hop and landing tasks in individuals after ACLR at the time of return to activity.^21,22 Thus, decreased knee flexion angles seen in this study may be indicative of decreased quadriceps strength of the involved limb after ACLR.

A greater number of overall movement deviations were observed on the six compared to four inch box height, as hypothesized. Median frequencies of scores for overall quality of movement were 2 and 3 for the four and six inch box heights, respectively. For the four inch box height, 25% of individuals received scores of 0 and 1. Only 5% of individuals received scores of 0 or 1 for the six inch box height. These findings suggest that taller box height elicited a greater number of movement deviations and may be useful in eliciting subtle deviation patterns. Other studies that have applied this rating system to assess quality of movement in healthy individuals during the LSD test reported that 60% and 31% of participants received a score of 0-1.^12,16 The only other study to test a pathologic population reported 38% of individuals with patellofemoral pain syndrome received a score of 0-1.¹³ These comparative data suggest that individuals with a history of ACLR often exhibit worse quality of movement than healthy individuals, and perhaps even worse than those with patellofemoral pain syndrome.

This study has several limitations. First, it is unknown whether kinematic variables or visually-assessed movement deviations during the LSD test are associated with risk for ACL re-injury. Thus, future work should aim to link movement patterns during the LSD test with kinematics relevant to ACL re-injury during other dynamic tasks such as the drop vertical jump. Second, data from this study was limited to the involved limb. Incorporation of a healthy control group may be useful in identifying movement deviations unique to the ACLR population during the LSD test. Finally, this study did not explore the frontal plane projection angle. As this is another readily available 2D assessment approach that may be used to evaluate movement, it is possible that a combined assessment approach using the LSD scoring and frontal plane projection angle would have greater clinical utility than either approach alone. Finally, the study utilized a smaller sample, which may have inhibited the subgroup comparisons.

CONCLUSION

In conclusion, these results suggest that poorer visually-assessed quality of movement during the LSD test is associated with increased hip adduction and hip internal rotation after ACLR. 2D movement deviations at the pelvis appear to consistently relate to less knee flexion, and 2D deviations at the knee appear to suggest greater hip adduction. Generally, poorer quality of movement was observed during LSD performance using the six inch box height. Clinically, these data suggest that interventions targeting hip abductor, hip lateral rotator, and knee extensor strength and neuromuscular control may be useful in the presence of poor quality of movement during lateral step-down testing.