Structure, Sex, and Strength and Knee and Hip Kinematics During Landing

Participants

We conducted a power analysis using pilot data from previous research in our laboratory, which demonstrated means for knee abduction displacement for male and female participants of 3.04° ± 3.50° and 7.33° ± 6.08°, respectively. Assuming an α level of .05, the inclusion of 45 participants was expected to result in sufficient statistical power (0.80). We enrolled 30 women and 15 men in the study (age = 21 ± 2 years, height = 171.7 ± 9.5 cm, mass = 68.4 ± 9.5 kg). The participant pool consisted of recreationally active people recruited from the community as a sample of convenience. Inclusion criteria for participation were self-reported regular physical activity, no history of lower extremity injury in the 6 months before the study, no history of lower extremity surgery, age between 18 and 25 years, and willingness to participate in the study. We defined regular physical activity as participation in activity for a minimum of 30 minutes 3 times per week.

Shoes (Air Max Challenge; Nike, Inc, Beaverton, OR) were provided for the participants to wear during the study to control for differences in footwear. Participants' height and mass were measured before we instructed them to perform 3 trials of a standing, 2-footed, forward jump landing on 1 lower extremity. The lower extremity on which a participant chose to land in 2 of 3 trials was considered the preferred leg and was the only lower extremity tested during our study. Participants then completed a 5-minute warmup on a stationary exercise bicycle at a self-determined intensity level. All participants provided written informed consent, and the study was approved by the Institutional Review Board of the University of Kentucky.

Evaluation of AUHR

We measured hip range of motion for IR and ER and considered the difference between the measurements to be an estimate of femoral anteversion.15161823–25 Rotation was measured with participants lying prone on the examining table with the hips in extension and knees actively flexed to 90°, with the patellae of both limbs lying even with the end of the table and with the knees spaced approximately 14 in (35.6 cm) apart (Figure 1). We used 14 in of spacing, which was novel to our study, to attempt to standardize limb position and control for variations in hip motion resulting from different amounts of abduction. This distance was chosen based on pilot testing in which we observed that participants often self-selected this position for comfort. A standard goniometer modified with the addition of a bubble level to ensure vertical alignment of the reference arm was used to measure IR and ER in degrees. Participants were instructed to actively flex their knees and to allow their hips to rotate internally and then externally under the force of gravity to their passive limit.25 Both limbs were either internally or externally rotated simultaneously to aid in stabilization of the pelvis.41 Each measurement was recorded 3 times to calculate an average measure of IR and ER from which AUHR was calculated. The same tester (J.S.H.) performed all hip rotation measurements. The intraclass correlation coefficients (2,1) were 0.99 (SEM = 0.5°) and 0.95 (SEM = 1.9°), respectively, for intratester between-days reliability (n = 10 with 1 day between tests) for measures of IR and ER a priori.

View Figure

• Download as Powerpoint
• Download Figure

Evaluation of Longitudinal Arch Mobility

In our study, the foot and medial longitudinal arch were characterized by changes in dorsum height in response to loading.42 Measurements were accomplished using the Arch Height Index Measurement System (JAK Tool and Model, LLC, Matawan, NJ) (Figure 2). Arch height was recorded in 2 stance conditions: 10% (AH10%) of weight bearing and 90% (AH90%) of weight bearing. To calculate RAD during loading, we used the equation proposed by Nigg et al43 and modified by Williams and McClay42:

where body weight (BW) is expressed in newtons. Participants were weighed on a standard scale, and 10% and 90% of each participant's total weight were calculated. Participants stood with their hands resting on an examination table, which they used to assist in controlling their amount of weight bearing. The height of the examination table was raised or lowered as needed for participants to maintain balance. Participants then placed one foot on the scale and the other foot on an even, adjacent surface. The adjacent surface was the same height as the scale and was positioned just posterior and slightly medial to the stance limb being tested. This position was used to prevent leaning by the participants to one side, possibly influencing the foot height during the 2 weight-bearing conditions. The participants were instructed to control the amount of weight bearing by balancing directly over the scale and either touching the foot of the nontest limb to the adjacent surface or supporting their weight on the examination table until the scale showed that 90% of weight bearing had been achieved. The arch height then was recorded. The process was repeated for 10% of weight bearing. Participants were monitored carefully to ensure that their center of mass remained over the stance limb and that the knee was not bent. Intratester reliability of this instrument for calculating an arch height index using AH10% and AH90% has been reported (intraclass correlation coefficient = 0.94).42 The same examiner (J.S.H.) performed all measurements.

View Figure

• Download as Powerpoint
• Download Figure

Kinematic Analysis

Three-dimensional joint kinematics of the hip and knee were collected at 100 Hz using Flock of Birds electromagnetic sensors (Ascension Technology Corporation, Burlington, VT) and The MotionMonitor software (Innovative Sports Training, Inc, Chicago, IL). Electromagnetic sensors were placed on the skin of the preferred landing leg using adhesive pads and tape. One sensor was placed on each participant's sacrum, lateral thigh, and medial tibial plateau. The posterior-superior iliac spine, lateral and medial knee joint lines, and lateral and medial malleoli were digitized per manufacturer recommendations. A standard right-hand coordinate system was used for all joints so that the positive x-axis projected anteriorly and the positive z-axis projected superiorly. The landing task, as described in the jump protocol, was performed by each participant. Time of initial contact was identified using a foot switch placed in the participant's shoe. Foot switches contained toe, midfoot, and heel sensors to accurately document ground contact regardless of landing technique. Foot-switch data were collected at 2000 Hz and synchronized with the kinematic data.

Jump Protocol

Participants performed a double-leg jump with a single-leg landing. Each participant started the task at a distance away from a designated target area that was equivalent to 40% of his or her height; the vertical component of the initial jump was equivalent to 115% of the participant's height.37 To control for the vertical component of the jump, a foam block was hung at the appropriate height, and the participant was instructed to jump so the block just brushed the top of his or her head.37 The participants were instructed to use a single-leg landing strategy and to stabilize quickly. They also were instructed to perform several familiarization repetitions of the jump-landing task with the preferred leg. A minimum of 3 practice trials were performed to ensure comprehension of the task. Additional trials were permitted if a participant did not feel comfortable with the task. Three test trials from each participant were used for analysis. A trial was discarded and an additional one recorded if a participant did not meet the height or distance requirement or if the single-leg landing was not held stable. To be considered stable, participants had to maintain unilateral stance for a minimum of 5 seconds after landing.

Strength Testing

After completing the jump-landing task, all participants underwent isometric strength testing for a motion of combined hip abduction and ER.13 Strength of the preferred leg was assessed with the PrimusRS dynamometer (BTE Technologies, Inc, Hanover, MD). Participants were positioned side lying on a table with the preferred leg on top. The trunk was in neutral alignment, with the hips flexed to 45° and the knees flexed to 90° (Figure 3).13 Participants were allowed to support their heads using the contralateral hand and arm and could stabilize themselves on the table using the ipsilateral hand. Participants remained in this position for the duration of the strength testing.

View Figure

• Download as Powerpoint
• Download Figure

The pad on the resistance arm of the dynamometer was placed over the lateral side of the preferred knee. Participants abducted and externally rotated the preferred leg against the pad. The foot of the preferred leg was not allowed to touch the other foot to prevent pushing off but was required to remain below the level of the knee to prevent internal hip rotation. Three maximal contractions of 5 seconds each with a 30-second rest between trials were averaged to establish the maximal voluntary isometric contraction torque (newton-meters) for each participant, which was then normalized to body mass (kilograms) for further analysis.

Data Analysis

Initial joint angles for 3-dimensional motion at the hip and knee were measured at the time of ground contact as detected by foot switches. Joint excursions for the hip and the knee were calculated as the difference between the joint angle at initial ground contact and the maximal joint angle occurring between contact and maximal knee flexion (maximal angle – initial contact angle = joint excursion).3334 Excursions were considered for each direction of motion within a plane. For example, for frontal-plane hip motion, excursions were calculated in abduction and adduction. All joint angles were calculated using a segmental, local coordinate system such that hip motion was defined as movement of the thigh sensor relative to the sacral sensor and knee motion was defined as movement of the shank sensor relative to the thigh. Euler angle equations were used to estimate joint angles, and the Leardini method was used to estimate the hip joint center.4445 Joint position data were processed using a fourth-order, dual-pass, Butterworth filter with a cutoff frequency of 9 Hz, which was confirmed by residual analysis. Datapac 2K2 (Run Technologies, Mission Viejo, CA) was used for the processing and analyzing all kinematic data.

Statistical Analysis

This was a single-occasion, descriptive laboratory study. Joint angles for 3-dimensional motion at the hip and knee at initial contact and maximal joint excursions were considered the response or dependent variables of interest for this study. The AUHR, RAD, sex, and hip strength were the explanatory or independent variables. Pearson product moment correlations were conducted to determine the relationship between continuous explanatory variables and hip and knee kinematics and to evaluate the relationship between AUHR and RAD. For kinematic variables for which a correlation was present, a multivariable linear regression model was used to determine which combination of factors (AUHR, RAD, sex, or hip abduction– ER isometric peak torque) were predictive of lower extremity kinematics of the hip and knee at ground contact and maximal excursion. All explanatory variables were included initially, then the models were backward reduced. The variables that contributed the least were removed at each step until all variables contributed to a concise model at the α level set a priori (.05). All statistical analyses were conducted with SPSS (version 15.0; SPSS Inc, Chicago, IL).

Sex, Hip Strength, and AUHR

Our results indicated that as AUHR increased (suggesting greater relative femoral anteversion), participants experienced greater knee abduction excursion during landing, with female participants and those with weaker hip abductors and external rotators experiencing the most knee abduction excursion. Fem-oral anteversion often is cited as a possible risk factor for ACL or other knee injuries10–13; however, we were the first to attempt to include an estimate of femoral anteversion as a predictor of knee abduction, which is a kinematic position linked to ACL injury.12546

As hypothesized, hip adduction excursion also was correlated with AUHR (R = 0.36). Although not a primary aim of our study, the high correlation (R = 0.70) observed between knee abduction excursion and hip adduction excursion was not surprising. Clinically, these combined motions are often called knee valgus.47 When they occur in unison, these motions represent a medial collapse of the lower extremity.47 Therefore, it would be functionally challenging for one to occur without the other in such a manner that balance between the hip and knee can be maintained.

To identify the functional influence of femoral anteversion, Nyland et al13 studied the relationship between femoral ante-version and electromyographic (EMG) activation of hip musculature during an isometric exercise of combined hip abduction and ER. They reported that participants with greater IR had lower gluteus medius/hip abductor and vastus medialis/hip abductor EMG ratios than did participants with less rotation. Although we did not collect EMG data, our results demonstrated a potential relationship between AUHR and the frontal-plane movements of hip adduction and knee abduction. Based on the findings of Nyland et al,13 it might be theorized that the positive correlation observed between AUHR and hip adduction is a result of less relative gluteus medius activation. A decrease in relative gluteus medius activation in those with increased femoral anteversion (a deformity in the transverse plane) might result in a loss of frontal-plane hip control, sending the hip into adduction and a corresponding knee abduction position.13 Although this pattern is plausible, further investigation involving structural measures, kinematics, and electromyography is needed to elucidate this potential biomechanical relationship.

The observed negative relationship between hip abduction– ER strength and knee abduction excursion also supports a potential link between AUHR and hip muscular strength and function. These results are supported by the work of Arnold et al,39 in which computer modeling demonstrated that increasing femoral anteversion decreases the abduction moment arm of the gluteus medius. This decreased moment arm could contribute to both variations in activation and a loss of frontal-plane control. This potential biomechanical relationship between AUHR and frontal-plane knee motion is further supported by the observed negative relationship between AUHR and knee adduction excursion (R = −0.33). This result suggests that lower AUHR might have a protective effect against medial knee collapse by increasing knee adduction during landing.

The results presented in Tables 3 and 4 demonstrate that participants with greater AUHR went into more hip adduction and that as AUHR increased and hip abduction–ER strength decreased, participants experienced greater knee abduction excursion. Extreme valgus position (composed, at least in part, of knee abduction) has occurred in landing and cutting actions, resulting in subsequent injury to the ACL.13 Our results showed a relationship among greater AUHR, female sex, lower hip strength, and greater knee abduction, which is an established position of risk.

Our observed trend of women experiencing greater knee abduction excursion during landing is consistent with the results reported by several researchers.30–35 The role of hip strength during landing has not been studied extensively and is less understood. In women, a negative correlation between eccentric hip abductor peak torque and knee abduction (R = −0.61) has been reported during a landing activity similar to the one we used.3637 We observed a similar finding with hip abduction–ER strength and sex as predictors of knee abduction excursion. These same factors also were predictive of knee IR at contact. We found no differences in normalized hip strength between sexes (P = .7) (data not shown), which is a finding that other investigators36 also have observed. However, this finding is in contrast to findings reported by other authors374849 documenting differences in hip abduction–ER strength between sexes. More study is needed to fully understand the role of hip strength and its relationship with sex and the possible implications of these factors for medial knee collapse.

We found a predictive relationship among AUHR, sex, and knee ER excursion (Table 5). Participants with greater AUHR went into more knee ER during landing, with women experiencing the greatest ER excursion. The observed ER excursion might be a function in part of the degree of knee IR at contact. Sex and hip strength were predictive of knee IR at contact, with women who had less hip strength landing in the most internally rotated position (Table 6). Landing in an internally rotated position might have contributed to the greater ER excursion associated with the female sex. It is important to note that the 2 kinematic variables were only moderately correlated (R² = 0.27) (data not shown), suggesting that in addition to contact position and sex, other factors, such as AUHR, might have contributed to knee ER excursion. Overall, this movement from a more internally rotated position at contact to a more externally rotated position during landing is consistent with a medial collapse knee injury mechanism, which might be exaggerated among women, especially those with lower hip ER–abduction strength.

In our study, greater anteversion resulted in greater knee external excursion during landing. In a previous study,50 greater anteversion resulted in less tibial IR excursion (tibial IR excursion = 22.009 + 1.09 [navicular drop] −1.083 [body mass index] −0.0771 [anteversion]). Although the terminology is different, the motions described are similar. The results of our study and the study by Carcia and Houglum50 suggest that fem-oral anteversion might influence transverse-plane knee motion during landing. Despite these similarities, several differences exist between our study and that of Carcia and Houglum.50 For example, the landing task was different, the method for evaluating anteversion was not reported, and although the overall model (including femoral anteversion) was predictive of tibial IR, the contribution of femoral anteversion to the model was not at the α level established a priori (P = .06).50 Each study provided additional information for better elucidating the roles of structure, strength, and function in neuromuscular patterns, such as landing, but additional research clearly is needed to identify consistent relationships between hip asymmetry/structure and knee position during landing tasks.

Despite these findings, the clinical applications of our results are influenced by the limitations of the instrumentation used to measure biomechanical movement. In our experimental design, a segmental local coordinate system was used such that knee IR was identified as rotation of the shank relative to the thigh. With this design, IR of the femur relative to a stable tibia would be reported as external knee rotation. This type of kinematic assessment makes drawing distinct conclusions regarding transverse-plane knee motion difficult because which segment of the kinetic chain is moving is not known. This method might explain our lack of findings in the transverse plane at both the hip and knee despite examining factors thought to primarily influence transverse kinematics.

Arch Mobility

Our hypothesis that RAD would predict medial collapse during landing was not supported by our results. We observed a correlation between RAD and hip abduction at contact. This negative relationship is consistent with the potential for medial collapse of the lower extremity.51 However, the absence of a relationship with hip adduction or knee abduction excursion suggests that RAD does not relate to a medial collapse actually occurring during landing. No factors associated with medial collapse during landing were correlated with RAD. Although RAD might have influenced frontal-plane hip position at contact, this relationship did not continue throughout the landing process.

We observed a negative correlation between RAD and knee flexion excursion, suggesting that those experiencing greater arch deformation moved through a smaller range of knee flexion during landing. This relationship might represent a tradeoff in force absorption between knee flexion and arch mobility. If the impact of landing can be attenuated by arch collapse, knee flexion excursions might be lessened. Gross and Nelson52 suggested that attenuation processes during landing might be learned muscular responses. However, Hargrave et al53 did not report lower ground reaction forces or differences in knee flexion among participants with increased pronation. Further research with kinetic analysis is needed to determine whether this relationship is consistent and how it might influence ground reaction forces and subsequent risk for injury during landing activities.

Limitations

Our results are presented with the assumption that AUHR is an appropriate estimate of femoral anteversion. Although this method has been validated successfully in the literature, it was tested predominantly among 8- and 9-year-old children.15 It is well documented that femoral anteversion decreases with growth from birth to age 16 years.24 The smaller variation in degree of anteversion seen in normal, healthy adults possibly limits the accuracy of a test originally designed to evaluate the larger angles of anteversion seen in youth. In addition, this method was validated using biplane radiography, which is a technique documented as being less valid and more variable than computed tomography imaging.54 We examined screening measures that could be performed in a clinical setting without extensive training or expensive equipment. We believe this is a reasonable method of examining the proposed risk factor for knee injury because greater IR relative to ER has been observed consistently in participants with above-average femoral anteversion.15161823–25

The strength of any conclusions drawn from our study is limited by the inability to fully understand what factors influence static measures of hip rotation. We have not found documentation of the percentage of hip range of motion determined by skeletal structure compared with soft tissue laxity. The ability to differentiate soft tissue laxity from bony structure at the hip or to control for soft tissue laxity at the hip could improve the use of AUHR as an estimate of femoral anteversion. Our use of 14 in (35.6 cm) as a standard spacing between knees during range-of-motion assessment similarly might have influenced hip rotation; in subsequent research, this spacing might need to be normalized to participant height or thigh length.

We defined AUHR as IR – ER. The mean of our sample groups was −6°, suggesting that, on average, they demonstrated more ER. Swanson et al55 suggested that 30° of asymmetry is necessary to produce abnormal alignment. Why our participants demonstrated greater ER than IR is not clear. It might be a function of the age of our participants. Femoral anteversion and AUHR are documented to decrease with age.19 Although we anticipated greater IR than ER, reports of rotational differences have varied in the literature, with AUHR (as defined in our study) reported as low as −21.9°.56 The use of the term AUHR is specific to the equation and does not necessarily imply pathologic asymmetry in our sample.

Finally, an additional limitation of our study was the size and health of our sample. The sample size did not allow a robust comparison between sexes, which might have been helpful to fully eliminate sex as a confounding factor in our results. Because sex-related differences have been reported for hip strength and femoral anteversion, fully controlling for the effect of sex on these variables is difficult. However, we have accounted for sex by including it in the regression model and believe that review of the standardized parameter estimates supports our conclusion that the relationship among AUHR, strength, and landing kinematics is not due to sex alone. One limitation with investigating predictive relationships is that the variability of the data for moderate to small samples is compressed. In our sample, the variables of interest might not have been varied enough to detect some relationships. Although our standard deviations did not suggest this was the case with AUHR or RAD, it might have been the case for hip strength (Table 1). Despite these potential limitations of our sample, we believe our participants represented a physically active population that is clinically relevant.