- Email: [email protected]
The influence of iliotibial band syndrome history on running biomechanics examined via principal components analysis
Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
Contents lists available at ScienceDirect
Journal of Biomechanics journal homepage: www.elsevier.com/locate/jbiomech www.JBiomech.com
The influence of iliotibial band syndrome history on running biomechanics examined via principal components analysis Eric Foch n, Clare E. Milner Department of Kinesiology, Recreation, and Sports Studies, University of Tennessee, Knoxville, TN, USA
art ic l e i nf o
a b s t r a c t
Article history: Accepted 7 October 2013
Iliotibial band syndrome (ITBS) is a common knee overuse injury among female runners. Atypical discrete trunk and lower extremity biomechanics during running may be associated with the etiology of ITBS. Examining discrete data points limits the interpretation of a waveform to a single value. Characterizing entire kinematic and kinetic waveforms may provide additional insight into biomechanical factors associated with ITBS. Therefore, the purpose of this cross-sectional investigation was to determine whether female runners with previous ITBS exhibited differences in kinematics and kinetics compared to controls using a principal components analysis (PCA) approach. Forty participants comprised two groups: previous ITBS and controls. Principal component scores were retained for the first three principal components and were analyzed using independent t-tests. The retained principal components accounted for 93–99% of the total variance within each waveform. Runners with previous ITBS exhibited low principal component one scores for frontal plane hip angle. Principal component one accounted for the overall magnitude in hip adduction which indicated that runners with previous ITBS assumed less hip adduction throughout stance. No differences in the remaining retained principal component scores for the waveforms were detected among groups. A smaller hip adduction angle throughout the stance phase of running may be a compensatory strategy to limit iliotibial band strain. This running strategy may have persisted after ITBS symptoms subsided. & 2013 Elsevier Ltd. All rights reserved.
Keywords: Eigenvector decomposition Percent variance explained Gait analysis Female
1. Introduction Iliotibial band syndrome (ITBS) is a common knee overuse injury afflicting approximately 8% of runners annually. Furthermore, women are two times as likely to sustain ITBS compared to men (Taunton et al., 2002). It has been postulated that ITBS results from repetitive friction of the iliotibial band sliding over the lateral femoral epicondyle during knee flexion and extension (Noble, 1980; Orchard et al., 1996; Renne, 1975). Based on a previous anatomical investigation, the notion of ITBS being a friction syndrome has been challenged (Fairclough et al., 2006, 2007). Instead of limiting sagittal plane knee motion, the iliotibial band serves to stabilize the lateral hip and knee, as well as resist hip adduction and knee internal rotation (Fredericson et al., 2000). Therefore, secondary plane hip and knee biomechanics must be examined to determine associations between biomechanics during running and ITBS. In addition to lower extremity biomechanics, trunk and pelvis kinematics may be associated with ITBS. It has been postulated n Correspondence to: Department of Physical Therapy & Athletic Training, Boston University, 635 Commonwealth Avenue, Boston, MA 02215, USA. Tel.: þ 1 617 353 7472. E-mail address: [email protected] (E. Foch).
that contralateral pelvic drop and trunk lateral flexion away from the stance limb would increase the internal knee abduction moment. An increase in peak internal knee abduction moment may increase the tensile strain experienced by the soft tissue crossing the lateral knee joint such as the iliotibial band (Powers, 2010). However, both runners with previous ITBS and controls lean their trunk towards the stance limb during the stance phase of running (Foch and Milner, in press). Additionally, peak trunk ipsilateral flexion and contralateral pelvic drop were not different between groups. Therefore, it is not unexpected that runners with previous ITBS also exhibited similar peak knee abduction moment compared to controls. Nevertheless, differences in frontal plane biomechanics during running may be detected by examining the entire time-series waveform rather than peak values. Cross-sectional investigations can provide insight into determining associations in biomechanics during running and their influence on iliotibial band mechanics. It is unknown whether post-injury biomechanics during running reflects runners' biomechanics before their first incidence of ITBS. However, half of runners who have sustained an overuse running injury reported a previous injury to the same anatomical location (Taunton et al., 2003). Although recurrence rates of ITBS are unknown, overuse running injuries commonly recur if underlying causative factors are not addressed. Collectively, the results from cross-sectional
0021-9290/$ - see front matter & 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jbiomech.2013.10.008
Please cite this article as: Foch, E., Milner, C.E., The influence of iliotibial band syndrome history on running biomechanics examined via principal components analysis. Journal of Biomechanics (2013), http://dx.doi.org/10.1016/j.jbiomech.2013.10.008i
E. Foch, C.E. Milner / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
2
studies can better inform researchers designing future prospective studies to determine biomechanical causes of ITBS. Previous cross-sectional investigations have implemented discrete analysis of peak variables to investigate factors associated with ITBS (Ferber et al., 2010; Grau et al., 2011). However, analysis of discrete variables is not sensitive to differences in the underlying movement pattern within a biomechanical waveform. Potentially, a more comprehensive waveform analysis would be able to characterize differences during running among runners with different ITBS injury status. Therefore, a principal component analysis (PCA), which captures the time-varying movement pattern of a waveform, may provide deeper understanding of injury risk factors (Kipp et al., 2011). PCA can be used to identify underlying patterns within angle and moment time-series curves based on the retained principal components. Previous research has shown that analysis of discrete variables was not able to discriminate between workers who developed low back pain and those who did not (Wrigley et al., 2005). However, the principal components derived from a PCA were able to identify differences in the kinematics and kinetics of lifting technique before low back pain developed (Wrigley et al., 2005). Additionally, PCA was able to detect differences in knee biomechanics during a run and cut task between genders that analysis of discrete variables did not (O'Connor and Bottum, 2009). Collectively, the results of these studies suggest that a PCA may enhance our understanding of biomechanical factors that are associated with previous ITBS. Therefore, the purpose of this cross-sectional investigation was to determine whether women with previous ITBS exhibited differences in kinematics and kinetics during running compared to controls using a PCA approach.
2. Methods 2.1. Participant details Approval for all procedures was granted by the Institutional Review Board. Forty female runners between the ages of 18 and 45 provided informed written consent prior to participating. Other data obtained from these runners have been reported previously (Foch and Milner, in press). Participants were excluded if they answered ‘yes’ to any question on a Physical Activity Readiness-Questionnaire (PAR-Q) (Thomas et al., 1992) or previously sustained a major lower-extremity injury. Answering ‘yes’ to a question on the PAR-Q indicated that the participant currently had heart and/or orthopedic problems that limited her ability to exercise. A running history questionnaire was then completed by each participant. All participants were running a minimum of 24 km wk 1 (previous ITBS: median ¼29.7 km wk 1; range¼ 24–85.3 km wk 1; controls: median ¼32.2 km wk 1; range¼24–80.5 km wk 1). Forty female participants were equally divided into two groups: previous ITBS and controls (Table 1). Runners with previous ITBS reported that they had been diagnosed by a health care professional (medical doctor, physical therapist, or athletic trainer) and had been running pain free for at least 1 month prior to data collection.
2.2. Data collection Participants wore running shorts, a tank-top, and neutral laboratory footwear (Bite Footwear, Redmond, WA, USA) for the overground running trials (Barnes et al., 2010). Passive reflective markers were placed on the right lower-extremity for controls. Data
Table 1 Mean (standard deviation) of participant demographics in the previous iliotibial band syndrome (ITBS) and control groups. Each group was comprised of 20 participants. Demographics
Previous ITBS
Controls
P value
Age (years) Height (m) Mass (kg) Weekly distance run (km wk 1)
26.0 1.67 58.8 41.8
23.7 1.68 58.9 38.6
0.187 0.809 0.955 0.645
(5.6) (0.04) (7.4) (25.1)
(5.5) (0.06) (5.7) (18.2)
were collected on the previously injured lower-extremity in the ITBS group. If both sides were injured previously, then data from the right side were collected. Joint coordinate systems were defined by placing passive reflective markers over anatomical landmarks on the lower extremity of interest and trunk (Grood and Suntay, 1983). The anatomical landmarks were: acromion processes, superior iliac crests, greater trochanters, lateral and medial femoral epicondyles, lateral and medial malleoli, and first and fifth metatarsal heads. Molded thermoplastic shells with four non-collinear markers were positioned over the posterior pelvis and posterolaterally on the proximal thigh and distal shank (Cappozzo et al., 1997). The shells on the thigh and shank were secured to the segment via neoprene wraps and hook and loop tape (Manal et al., 2000). Rear-foot motion was indicated by placement of three noncollinear markers directly on the heel. Markers were placed on the manubrium, sternal body, seventh cervical vertebra, and tenth thoracic vertebra to indicate trunk motion. A static calibration trial was recorded with participants standing on a foot placement template. The standardized foot position was 0.17 m between heel centers at an angle of 141 between the anteroposterior axes of the feet (McIlroy and Maki, 1997). After the calibration trial was recorded, all anatomical markers were removed. Overground running trials were collected while participants ran along a 17 m runway at a velocity of 3.5 70.18 m s 1. A nine-camera motion capture system (Vicon, Oxford Metrics, Centennial, CO, USA) sampling at 120 Hz recorded marker trajectories. A force plate located in the middle of the runway was synchronized with the motion capture system (AMTI, Inc., Watertown, MA, USA) and sampled at 1200 Hz. Two photocells linked to a timer were placed three meters apart on either side of the force plate to monitor running velocity. Five acceptable trials were collected, in which participants maintained the specified running velocity and landed on the force plate without altering their stride. 2.3. Data reduction Data were reduced in Visual3D (C-Motion, Rockville, MD, USA). Kinematics and ground reaction forces were filtered with a 4th order Butterworth filter at a cut-off frequency of 8 Hz (Bisseling and Hof, 2006). Joint angles were determined using a Cardan X–Y–Z (mediolateral, anteroposterior, and vertical) rotation sequence (Wu and Cavanagh, 1995). Trunk and pelvis segments were computed with respect to the laboratory coordinate system. The laboratory coordinate system was defined at the posterior left corner of the force plate. Inverse dynamics were computed using a standard Newton–Euler approach. Moments were expressed as internal moments and were normalized to body mass and height (O’Connor and Bottum, 2009). A vertical ground reaction force threshold of 20 N was used to determine the onset and end of stance. The five waveforms of interest were: frontal plane trunk, pelvis, and hip angles and knee moment, as well as transverse plane knee angle. 2.4. Principal components analysis Stance phase of the overground running trials was time normalized to 101 points. The angle and moment data for each participant were an ensemble average of the five trials. For each of the five angle and moment waveforms of interest, a data matrix was created. The 101 data points comprised the columns and 40 participants comprised the rows of each matrix ðX 40101 Þ. The PCA approach used in the current investigation was based on a methodology described previously (Wrigley et al., 2006). The mean was computed for each column of the respective matrix. Then, the mean of each column was subtracted from each row in its respective column. The mean centered matrices were transformed into principal components using an eigenvector decomposition method on the input's covariance matrix ðC 101101 Þ. The PCA produced the eigenvectors (V 101101 Þ and eigenvalues (L1101 Þ. The eigenvector matrix consisted of the coefficients for each of the 101 principal components which defined a new coordinate space for the original waveform data (Wrigley et al., 2006). The eigenvalue matrix indicated the relative contribution each principal component had on the total variance in the data. For each matrix, the first three principal components were analyzed. Principal component score matrices (Z 40101 Þ were then computed by multiplying the mean-centered input matrix by transposing the eigenvector matrix: Z 40101 ¼ ðX 40 101 ð140 1 x1 101 ÞÞ V
0
101 101
ð1Þ
where x1 101 is each time normalized data point. The principal component scores represented how closely a runner's waveform matched the shape of its respective principal component (Wrigley et al., 2006). To determine if the retained principal components represented the original data adequately, a residual analysis was performed using the Q-statistic (Jackson, 1991). The Q-statistic is the sum of the squares of the residuals between participants' original waveform and the reconstructed curve based on the retained principal component (Wrigley et al., 2006). A Q-critical value (Qα) was calculated using an alpha level of 0.05 from a t-distribution (degrees of freedom¼39, Qα ¼2.0211) (Wrigley et al., 2006). Q-critical indicated whether the number of retained components reconstructed the original data adequately. For each participant, a Q-statistic value lower than Qα indicated that the original data were represented adequately by the retained principal components (Jackson, 1991). To aid in the interpretation of the principal components, the influence of a principal component on the combined groups' ensemble mean waveform was
Please cite this article as: Foch, E., Milner, C.E., The influence of iliotibial band syndrome history on running biomechanics examined via principal components analysis. Journal of Biomechanics (2013), http://dx.doi.org/10.1016/j.jbiomech.2013.10.008i
E. Foch, C.E. Milner / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎ computed (Ramsay and Silverman, 1997). A multiple M of the respective principal component's coefficients was added and subtracted to the overall mean curve. The 5th and 95th percentiles of the principal component score distribution were used as M values for the low ( ) and high ( þ) curves (Mantovani et al., 2012). Additionally, to interpret when during stance the retained principal components contributed to the variability in the shape of the waveform, the explained percent variance (r 2ji ) was computed for the ith principal component and the jth time point: pffiffiffiffi V ji Li 100% ð2Þ r 2ji ¼ cj where cj is the standard deviation of C at a given data point of the waveform (Wrigley et al., 2006). Differences in timing and magnitude of the relative contribution of a principal component to the overall waveform can be observed visually in the percent variance explained figures (Wrigley et al., 2006). Waveform matrix construction and all PCA calculations were performed using custom software (MATLAB, MathWorks, Natick, MA, USA).
2.5. Statistical analysis Descriptive statistics (means and standard deviations) were computed for the principal component scores of the waveforms: frontal plane hip, pelvis, and trunk angles, knee moment, as well as transverse plane knee angle. To assess between group differences, principal component scores for each waveform were compared using independent t-tests. Additionally, participant demographics were compared between groups via independent t-tests. Statistical analysis was performed using PASW 20.0 (IBM SPSS Statistics, Chicago, IL, USA). An alpha value of 0.05 was set for all tests. Due to the exploratory nature of the analysis, trends were considered 0.05 o alpha o0.10.
3
the pelvis waveform. The high principal component scores exhibited by runners with previous ITBS indicated that they exhibited less contralateral pelvic drop compared to controls. Principal component one was the primary mode of variation about the mean during 25–70% of the stance phase of running. Similar to principal component one, principal component three corresponded to a variation in the overall magnitude of the frontal plane pelvis angle. The low principal component scores exhibited by runners with previous ITBS indicated that their pelvis was closer to neutral than controls. Whereas, high principal component scores in controls indicated the contralateral pelvis was more elevated. During early ( o20%) and late (4 80%) stances, principal component three captured the variation about the mean of the pelvis in an elevated position. There was a trend towards a difference in principal component one for frontal plane trunk angle between runners with previous ITBS and controls (P ¼0.057). The effect of the first principal component was to add or subtract from the overall magnitude of the trunk waveform. High principal component scores revealed that runners with previous ITBS tended to lean their trunk more towards the stance limb compared to controls throughout the stance phase of running. Principal component one was the primary mode of variation about the mean during 20–70% of the stance phase of running. Lastly, there were no differences in the in principal components scores in transverse plane knee angle and the internal frontal plane knee moment.
3. Results The first three principal components accounted for 93.3–99.4% of the total variance in the five biomechanical waveforms of interest (Table 2). All waveforms were reconstructed using the scores and coefficients of the first three principal components. The Q-statistic indicated that all participants' ensemble mean angle and moment waveforms were sufficiently described by the retained principal components. Group ensemble averages for the original waveforms are depicted (Fig. 1). Principal component one scores in frontal plane hip angle were different between runners with previous ITBS compared to controls during the stance phase of running (Fig. 2; P¼0.004). The effect of the first principal component was to add or subtract from the overall magnitude of the hip angle waveform. The high curve indicated that the hip was more adducted than the low curve. Runners with previous ITBS exhibited negative principal component one scores (Table 3). Thus, runners with previous ITBS exhibited less hip adduction during stance than controls. Additionally, the percent variance explained curve revealed that the principal component one captured the primary variation about the mean during 20–60% of stance. Both groups exhibited similar scores for principal components two (P¼0.413) and three (P¼0.890). There were trends towards group differences for frontal plane pelvis angle in principal component one (P¼ 0.067) and principal component three (P¼ 0.063). The effect of the first principal component was to add or subtract from the overall magnitude of
Table 2 The first three principal components (PC) and Q-critical (Qα) for the angle and moment waveforms during the stance phase of running. Waveforms
Frontal plane trunk angle Frontal plane pelvis angle Frontal plane hip angle Frontal plane knee moment Transverse plane knee angle
o Qα (%)
PC (%) PC1
PC2
PC3
Total
76.3 72.2 65.5 83.7 79.6
18.4 19.0 24.6 6.2 9.4
4.7 5.8 5.5 3.5 7.6
99.4 97.0 95.5 93.3 96.7
100 100 100 100 100
4. Discussion It is unclear from the literature whether discrete peak biomechanical variables during running are associated with previous ITBS. Implementing a PCA may provide insight into potential differences in underlying patterns in kinematic and kinetic waveforms between runners with previous ITBS and controls. Therefore, the purpose of this study was to determine if using a PCA approach could detect differences in trunk and lower extremity time-series waveforms between female runners with previous ITBS and controls. The magnitude of hip adduction was smaller (principal component 1) in runners with previous ITBS compared to controls. Less hip adduction throughout stance may be attributed to either a tight iliotibial band or activation of hip abductor musculature acting to restrict motion into adduction. Additionally, trends towards group differences were observed in pelvic segment orientation which can influence the hip adduction angle. Runners with previous ITBS tended to assume a pelvic position that was closer to neutral compared to controls (principal components 1 and 3). Running with a more neutral pelvis helps to explain the smaller hip adduction angle exhibited by runners with previous ITBS compared to controls. Contralateral pelvic drop moves the thigh and pelvis closer medially. Therefore, increased contralateral pelvic drop exhibited by controls aids in explaining the greater hip adduction they assumed throughout stance. Collectively, there were no temporal differences in the overall hip and pelvis waveform pattern between groups. This suggests that both groups exhibited similar movement patterns but runners with previous ITBS had smaller waveform magnitude throughout the stance phase of running. While PCA data have not been reported in relation to ITBS in runners previously, peak hip angles have. There are contradictory findings in the literature regarding increased peak hip adduction angle as a factor associated with previous ITBS. One previous retrospective study (Ferber et al., 2010) and one previous prospective study (Noehren et al., 2007) both reported greater peak hip adduction angle in female runners with previous ITBS
Please cite this article as: Foch, E., Milner, C.E., The influence of iliotibial band syndrome history on running biomechanics examined via principal components analysis. Journal of Biomechanics (2013), http://dx.doi.org/10.1016/j.jbiomech.2013.10.008i
4
E. Foch, C.E. Milner / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
Fig. 1. Previous iliotibial band syndrome (ITBS; dashed line) and control groups' ensemble average plot of: (A) frontal plane hip angle, (B) frontal plane pelvis angle, (C) frontal plane trunk angle, (D) frontal plane knee moment, and (E) transverse plane knee angle. The error bars indicate one standard deviation of the mean. Moment is expressed as internal moment.
compared to controls. However, runners with previous ITBS have also exhibited similar peak hip adduction angle compared to controls (Foch and Milner, in press; Miller et al., 2007). Participants in previous ITBS studies included only women (Foch and Milner, in press; Ferber et al., 2010; Noehren et al., 2007) and both genders (Miller et al., 2007). However, differences exist in peak hip adduction angle between healthy female and male runners (Ferber et al., 2003). Therefore, only women were included in the current study to avoid gender differences in running mechanics confounding the results. In the current study, it appears group differences in hip pattern are due to less pelvis motion in the ITBS group throughout stance. Smaller peak hip adduction angles have been reported in female and male runners with current ITBS compared to controls (Grau et al., 2011). The authors' suggested less hip adduction may be due to a tight iliotibial band and hip abductors
(Grau et al., 2011). Iliotibial band flexibility was not measured in the present study. Therefore, we can only postulate that decreased hip adduction exhibited by runners with previous ITBS was due to less iliotibial band and hip abductor flexibility. By examining joint biomechanics, interferences can be made as to how the iliotibial band is affected. Excessive hip adduction would cause the iliotibial band to elongate. Since the iliotibial band functions to limit hip adduction, decreasing hip adduction may be a compensatory strategy to limit iliotibial band strain. A decrease in iliotibial strain may reduce compression experienced by innervated adipose tissue between the iliotibial band and lateral femoral epicondyle (Fairclough et al., 2007). Therefore, the pain associated with ITBS during running may be alleviated. A reduction in iliotibial band and hip abductor flexibility could result from trying to limit frontal plane hip motion during
Please cite this article as: Foch, E., Milner, C.E., The influence of iliotibial band syndrome history on running biomechanics examined via principal components analysis. Journal of Biomechanics (2013), http://dx.doi.org/10.1016/j.jbiomech.2013.10.008i
E. Foch, C.E. Milner / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
5
Fig. 2. The ensemble average (combined groups) for frontal plane angle waveforms: hip, pelvis, and trunk angles (solid gray) during the stance phase of overground running are displayed. The principal components (PC) are represented by adding ( þhigh) and subtracting ( low) a multiple M of the respective PC coefficients to the combined groups' ensemble average curve. The mean score contribution for runners with previous ITBS (dash black) and controls (solid black) are depicted for each PC. Additionally, the percent variance explained by PC 1 (dot), PC 2 (dash-dot), PC 3 (dash), and overall percent variance explained by the first 3 PCs (solid) is displayed for each waveform.
running. Potentially, runners in the current study tried to limit hip adduction when injured, and this pattern persisted after injury. No statistically significant differences were observed among frontal plane trunk angle and knee moment, as well as transverse plane knee angle between runners with previous ITBS and controls. The findings of the PCA are in agreement with a discrete analysis of the aforementioned variables (Foch and Milner, in press). However, there was a trend towards increased frontal plane trunk angle towards the stance limb in runners previous ITBS compared to controls which was not detected via discrete analysis (Foch and Milner, in press). Trunk lateral flexion away from the stance limb was postulated to affect the frontal plane knee
moment (Powers, 2010). However, runners in both groups exhibited trunk lateral flexion towards the stance limb. Trunk lateral flexion is likely not associated with previous ITBS (Foch and Milner, in press). Limitations to this study are acknowledged. First, runners in the previous ITBS group reported that they had been running without pain for at least 1 month prior to data collection. However, a specific period of time since injury was not recorded. It has not been investigated whether the duration since alleviation of ITBS symptoms influences biomechanics during running. Since this was a cross-sectional study, whether biomechanics during running postinjury are the same as pre-injury is uncertain. Future prospective
Please cite this article as: Foch, E., Milner, C.E., The influence of iliotibial band syndrome history on running biomechanics examined via principal components analysis. Journal of Biomechanics (2013), http://dx.doi.org/10.1016/j.jbiomech.2013.10.008i
E. Foch, C.E. Milner / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
6
Table 3 The mean (standard deviation) for the principal component (PC) scores of the three retained PCs for each waveform of interest during overground running in women with previous iliotibial band syndrome (ITBS) and controls. The P value indicates the pairwise comparison for each waveform between groups. Controls
Retained PC
Frontal plane trunk angle
PC1 PC2 PC3
4.7 (13.6) 4.7 (16.6) 0.05 (7.2) 0.05 (8.4) 0.3 (3.5) 0.3 (4.3)
0.057 0.965 0.592
Frontal plane pelvis angle
PC1 PC2 PC3
5.9 (18.1) 2.5 (9.9) 1.7 (5.7)
5.9 (21.2) 2.5 (10.6) 1.7 (5.3)
0.067 0.124 0.063
Frontal plane hip angle
PC1 PC2 PC3
11.3 (23.5) 2.1 (18.1) 0.2 (6.6)
11.3 (22.8) 2.1 (12.8) 0.2 (8.2)
0.004a 0.413 0.890
Frontal plane knee moment
PC1 PC2 PC3
0.2 (1.4) 0.4 (0.5) 0.1 (0.4)
0.503 0.596 0.254
Transverse plane knee angle
PC1 PC2 PC3
0.5 (51.9) 2.2 (18.8) 2.9 (16.5)
0.949 0.424 0.231
0.2 (1.8) 0.04 (0.4) 0.1 (0.2) 0.5 (47.7) 2.2 (15.1) 2.9 (13.6)
Previous ITBS
P value
Waveform
a Runners with previous ITBS exhibited significant smaller PC1 scores than controls.
investigations are needed to better understand how running pattern is affected by ITBS injury status. This study assessed differences in kinematic and kinetic timeseries waveforms for variables associated with ITBS in runners via a PCA approach. Female runners with previous ITBS exhibited a frontal plane hip angle that was smaller in magnitude across the stance phase of running compared to controls. No differences were observed between groups in the retained principal components for other trunk and lower extremity waveforms. Reducing hip adduction may have enabled the runners with previous ITBS to reduce compression of a highly innervated fat pad between the femoral epicondyle and iliotibial band. This running pattern may have allowed runners to avoid lower extremity positions that were painful when injured. This pattern may have persisted after ITBS symptoms subsided.
Conflict of interest statement None.
Acknowledgments The authors would like to thank Dr. Jeffrey Reinbolt and Dr. Eugene Fitzhugh for their input on performing a PCA. References Barnes, A., Wheat, J., Milner, C.E., 2010. Use of gait sandals for measuring rearfoot and shank motion during running. Gait Posture 32, 133–135.
Bisseling, R.W., Hof, A.L., 2006. Handling of impact forces in inverse dynamics. J. Biomech. 39, 2438–2444. Cappozzo, A., Cappello, A., Della Croce, U., Pensalfini, F., 1997. Surface-marker cluster design criteria for 3-D bone movement reconstruction. IEEE Trans. Biomed. Eng. 44, 1165–1174. Fairclough, J., Hayashi, K., Toumi, H., Lyons, K., Bydder, G., Phillips, N., Best, T.M., Benjamin, M., 2006. The functional anatomy of the iliotibial band during flexion and extension of the knee: implications for understanding iliotibial band syndrome. J. Anat. 208, 309–316. Fairclough, J., Hayashi, K., Toumi, H., Lyons, K., Bydder, G., Phillips, N., Best, T.M., Benjamin, M., 2007. Is iliotibial band syndrome really a friction syndrome? J. Sci. Med. Sport 10, 74–76. Ferber, R., Noehren, B., Hamill, J., Davis, I.S., 2010. Competitive female runners with a history of iliotibial band syndrome demonstrate atypical hip and knee kinematics. J. Orthop. Sports Phys. Ther. 40, 52–58. Foch, E., Milner, C.E., 2013. Frontal plane running biomechanics in female runners with previous iliotibial band syndrome. J. Appl. Biomech. (in press). Fredericson, M., Cookingham, C.L., Chaudhari, A.M., Dowdell, B.C., Oestreicher, N., Sahrmann, S.A., 2000. Hip abductor weakness in distance runners with iliotibial band syndrome. Clin. J. Sports Med. 10, 169–175. Grau, S., Krauss, I., Maiwald, C., Axmann, D., Horstmann, T., Best, R., 2011. Kinematic classification of iliotibial band syndrome in runners. Scand. J. Med. Sci. Sports 21, 184–189. Grood, E.S., Suntay, W.J., 1983. A joint coordinate system for the clinical description of three-dimensional motions: application to the knee. J. Biomed. Eng. 105, 136–144. Jackson, J.E., 1991. A User's Guide to Principal Components. John Wiley & Sons Inc., New York, pp. 37–38. Kipp, K., McLean, S.G., Palmieri-Smith, R.M., 2011. Patterns of hip flexion motion predict frontal and transverse plane knee torques during a single-leg land-andcut maneuver. Clin. Biomech. 26, 504–508. Manal, K., McClay, I., Stanhope, S., Richards, J., Galinat, B., 2000. Comparison of surface mounted markers and attachment methods in estimating tibial rotations during walking: an in vivo study. Gait Posture 11, 38–45. Mantovani, G., Lamontage, M., Varin, D., Cerulli, G.G., Beaule, P.E., 2012. Comparison of total hip arthroplasty surgical approaches by principal component analysis. J. Biomech. 45, 2109–2115. McIlroy, W.E., Maki, B.E., 1997. Preferred placement of the feet during quiet stance: development of a standardized foot placement for balance testing. Clin. Biomech. 12, 66–70. Miller, R.H., Lowry, J.L., Meardon, S.A., Gillette, J.C., 2007. Lower extremity mechanics of iliotibial band syndrome during an exhaustive run. Gait Posture 26, 407–413. Noble, C.A., 1980. Iliotibial band friction syndrome in runners. Am. J. Sports Med. 8, 232–234. Noehren, B., Davis, I., Hamil, J., 2007. ASB clinical biomechanics award winner 2006 prospective study of the biomechanical factors associated with iliotibial band syndrome. Clin. Biomech 22, 951–956. O’Connor, K.M., Bottum, M.C., 2009. Differences in cutting knee mechanics based on principal components analysis. Med. Sci. Sports Exercise 41, 867–878. Orchard, J.W., Fricker, P.A., Abud, A.T., Mason, B.R., 1996. Biomechanics of iliotibial band friction syndrome in runners. Am. J. Sports Med. 24, 375–379. Powers, C.M., 2010. The influence of abnormal hip mechanics on knee injury: a biomechanical perspective. J. Orthop. Sports Phys. Ther. 40, 42–51. Ramsay, J.O., Silverman, B.W., 1997. Functional Data Analysis. Springer Verlag, New York, pp. 92–94. Renne, J.W., 1975. The iliotibial band friction syndrome. J Bone Jt Surg. 57, 1110–1111. Taunton, J.E., Ryan, M.B., Clement, D.B., McKenzie, D.C., Lloyd-Smith, D.R., Zumbo, B. D., 2002. A retrospective case-control analysis of 2002 running injuries. Br. J. Sports Med. 36, 95–101. Taunton, J.E., Ryan, M.B., Clement, D.B., McKenzie, D.C., Lloyd-Smith, D.R., Zumbo, B. D., 2003. A prospective study of running injuries: the Vancouver Sun Run “In Training” clinics. Br. J. Sports Med. 37, 239–244. Thomas, S., Reading, J., Shepard, R.J., 1992. Revision of the Physical Activity Readiness Questionnaire (PAR-Q). Can. J. Sport Sci. 17, 338–349. Wrigley, A.T., Albert, W.J., Deluzio, K.J., Stevenson, J.M., 2005. Differentiating lifting technique between those who develop low back pain and those who do not. Clin. Biomech. 20, 254–263. Wrigley, A.T., Albert, W.J., Deluzio, K.J., Stevenson, J.M., 2006. Principal component analysis of lifting waveforms. Clin. Biomech. 21, 567–578. Wu, G., Cavanagh, P.R., 1995. ISB recommendations for standardization in the reporting of kinematic data. J. Biomech. 28, 1257–1261.
Please cite this article as: Foch, E., Milner, C.E., The influence of iliotibial band syndrome history on running biomechanics examined via principal components analysis. Journal of Biomechanics (2013), http://dx.doi.org/10.1016/j.jbiomech.2013.10.008i
Contents lists available at ScienceDirect
Journal of Biomechanics journal homepage: www.elsevier.com/locate/jbiomech www.JBiomech.com
The influence of iliotibial band syndrome history on running biomechanics examined via principal components analysis Eric Foch n, Clare E. Milner Department of Kinesiology, Recreation, and Sports Studies, University of Tennessee, Knoxville, TN, USA
art ic l e i nf o
a b s t r a c t
Article history: Accepted 7 October 2013
Iliotibial band syndrome (ITBS) is a common knee overuse injury among female runners. Atypical discrete trunk and lower extremity biomechanics during running may be associated with the etiology of ITBS. Examining discrete data points limits the interpretation of a waveform to a single value. Characterizing entire kinematic and kinetic waveforms may provide additional insight into biomechanical factors associated with ITBS. Therefore, the purpose of this cross-sectional investigation was to determine whether female runners with previous ITBS exhibited differences in kinematics and kinetics compared to controls using a principal components analysis (PCA) approach. Forty participants comprised two groups: previous ITBS and controls. Principal component scores were retained for the first three principal components and were analyzed using independent t-tests. The retained principal components accounted for 93–99% of the total variance within each waveform. Runners with previous ITBS exhibited low principal component one scores for frontal plane hip angle. Principal component one accounted for the overall magnitude in hip adduction which indicated that runners with previous ITBS assumed less hip adduction throughout stance. No differences in the remaining retained principal component scores for the waveforms were detected among groups. A smaller hip adduction angle throughout the stance phase of running may be a compensatory strategy to limit iliotibial band strain. This running strategy may have persisted after ITBS symptoms subsided. & 2013 Elsevier Ltd. All rights reserved.
Keywords: Eigenvector decomposition Percent variance explained Gait analysis Female
1. Introduction Iliotibial band syndrome (ITBS) is a common knee overuse injury afflicting approximately 8% of runners annually. Furthermore, women are two times as likely to sustain ITBS compared to men (Taunton et al., 2002). It has been postulated that ITBS results from repetitive friction of the iliotibial band sliding over the lateral femoral epicondyle during knee flexion and extension (Noble, 1980; Orchard et al., 1996; Renne, 1975). Based on a previous anatomical investigation, the notion of ITBS being a friction syndrome has been challenged (Fairclough et al., 2006, 2007). Instead of limiting sagittal plane knee motion, the iliotibial band serves to stabilize the lateral hip and knee, as well as resist hip adduction and knee internal rotation (Fredericson et al., 2000). Therefore, secondary plane hip and knee biomechanics must be examined to determine associations between biomechanics during running and ITBS. In addition to lower extremity biomechanics, trunk and pelvis kinematics may be associated with ITBS. It has been postulated n Correspondence to: Department of Physical Therapy & Athletic Training, Boston University, 635 Commonwealth Avenue, Boston, MA 02215, USA. Tel.: þ 1 617 353 7472. E-mail address: [email protected] (E. Foch).
that contralateral pelvic drop and trunk lateral flexion away from the stance limb would increase the internal knee abduction moment. An increase in peak internal knee abduction moment may increase the tensile strain experienced by the soft tissue crossing the lateral knee joint such as the iliotibial band (Powers, 2010). However, both runners with previous ITBS and controls lean their trunk towards the stance limb during the stance phase of running (Foch and Milner, in press). Additionally, peak trunk ipsilateral flexion and contralateral pelvic drop were not different between groups. Therefore, it is not unexpected that runners with previous ITBS also exhibited similar peak knee abduction moment compared to controls. Nevertheless, differences in frontal plane biomechanics during running may be detected by examining the entire time-series waveform rather than peak values. Cross-sectional investigations can provide insight into determining associations in biomechanics during running and their influence on iliotibial band mechanics. It is unknown whether post-injury biomechanics during running reflects runners' biomechanics before their first incidence of ITBS. However, half of runners who have sustained an overuse running injury reported a previous injury to the same anatomical location (Taunton et al., 2003). Although recurrence rates of ITBS are unknown, overuse running injuries commonly recur if underlying causative factors are not addressed. Collectively, the results from cross-sectional
0021-9290/$ - see front matter & 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jbiomech.2013.10.008
Please cite this article as: Foch, E., Milner, C.E., The influence of iliotibial band syndrome history on running biomechanics examined via principal components analysis. Journal of Biomechanics (2013), http://dx.doi.org/10.1016/j.jbiomech.2013.10.008i
E. Foch, C.E. Milner / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
2
studies can better inform researchers designing future prospective studies to determine biomechanical causes of ITBS. Previous cross-sectional investigations have implemented discrete analysis of peak variables to investigate factors associated with ITBS (Ferber et al., 2010; Grau et al., 2011). However, analysis of discrete variables is not sensitive to differences in the underlying movement pattern within a biomechanical waveform. Potentially, a more comprehensive waveform analysis would be able to characterize differences during running among runners with different ITBS injury status. Therefore, a principal component analysis (PCA), which captures the time-varying movement pattern of a waveform, may provide deeper understanding of injury risk factors (Kipp et al., 2011). PCA can be used to identify underlying patterns within angle and moment time-series curves based on the retained principal components. Previous research has shown that analysis of discrete variables was not able to discriminate between workers who developed low back pain and those who did not (Wrigley et al., 2005). However, the principal components derived from a PCA were able to identify differences in the kinematics and kinetics of lifting technique before low back pain developed (Wrigley et al., 2005). Additionally, PCA was able to detect differences in knee biomechanics during a run and cut task between genders that analysis of discrete variables did not (O'Connor and Bottum, 2009). Collectively, the results of these studies suggest that a PCA may enhance our understanding of biomechanical factors that are associated with previous ITBS. Therefore, the purpose of this cross-sectional investigation was to determine whether women with previous ITBS exhibited differences in kinematics and kinetics during running compared to controls using a PCA approach.
2. Methods 2.1. Participant details Approval for all procedures was granted by the Institutional Review Board. Forty female runners between the ages of 18 and 45 provided informed written consent prior to participating. Other data obtained from these runners have been reported previously (Foch and Milner, in press). Participants were excluded if they answered ‘yes’ to any question on a Physical Activity Readiness-Questionnaire (PAR-Q) (Thomas et al., 1992) or previously sustained a major lower-extremity injury. Answering ‘yes’ to a question on the PAR-Q indicated that the participant currently had heart and/or orthopedic problems that limited her ability to exercise. A running history questionnaire was then completed by each participant. All participants were running a minimum of 24 km wk 1 (previous ITBS: median ¼29.7 km wk 1; range¼ 24–85.3 km wk 1; controls: median ¼32.2 km wk 1; range¼24–80.5 km wk 1). Forty female participants were equally divided into two groups: previous ITBS and controls (Table 1). Runners with previous ITBS reported that they had been diagnosed by a health care professional (medical doctor, physical therapist, or athletic trainer) and had been running pain free for at least 1 month prior to data collection.
2.2. Data collection Participants wore running shorts, a tank-top, and neutral laboratory footwear (Bite Footwear, Redmond, WA, USA) for the overground running trials (Barnes et al., 2010). Passive reflective markers were placed on the right lower-extremity for controls. Data
Table 1 Mean (standard deviation) of participant demographics in the previous iliotibial band syndrome (ITBS) and control groups. Each group was comprised of 20 participants. Demographics
Previous ITBS
Controls
P value
Age (years) Height (m) Mass (kg) Weekly distance run (km wk 1)
26.0 1.67 58.8 41.8
23.7 1.68 58.9 38.6
0.187 0.809 0.955 0.645
(5.6) (0.04) (7.4) (25.1)
(5.5) (0.06) (5.7) (18.2)
were collected on the previously injured lower-extremity in the ITBS group. If both sides were injured previously, then data from the right side were collected. Joint coordinate systems were defined by placing passive reflective markers over anatomical landmarks on the lower extremity of interest and trunk (Grood and Suntay, 1983). The anatomical landmarks were: acromion processes, superior iliac crests, greater trochanters, lateral and medial femoral epicondyles, lateral and medial malleoli, and first and fifth metatarsal heads. Molded thermoplastic shells with four non-collinear markers were positioned over the posterior pelvis and posterolaterally on the proximal thigh and distal shank (Cappozzo et al., 1997). The shells on the thigh and shank were secured to the segment via neoprene wraps and hook and loop tape (Manal et al., 2000). Rear-foot motion was indicated by placement of three noncollinear markers directly on the heel. Markers were placed on the manubrium, sternal body, seventh cervical vertebra, and tenth thoracic vertebra to indicate trunk motion. A static calibration trial was recorded with participants standing on a foot placement template. The standardized foot position was 0.17 m between heel centers at an angle of 141 between the anteroposterior axes of the feet (McIlroy and Maki, 1997). After the calibration trial was recorded, all anatomical markers were removed. Overground running trials were collected while participants ran along a 17 m runway at a velocity of 3.5 70.18 m s 1. A nine-camera motion capture system (Vicon, Oxford Metrics, Centennial, CO, USA) sampling at 120 Hz recorded marker trajectories. A force plate located in the middle of the runway was synchronized with the motion capture system (AMTI, Inc., Watertown, MA, USA) and sampled at 1200 Hz. Two photocells linked to a timer were placed three meters apart on either side of the force plate to monitor running velocity. Five acceptable trials were collected, in which participants maintained the specified running velocity and landed on the force plate without altering their stride. 2.3. Data reduction Data were reduced in Visual3D (C-Motion, Rockville, MD, USA). Kinematics and ground reaction forces were filtered with a 4th order Butterworth filter at a cut-off frequency of 8 Hz (Bisseling and Hof, 2006). Joint angles were determined using a Cardan X–Y–Z (mediolateral, anteroposterior, and vertical) rotation sequence (Wu and Cavanagh, 1995). Trunk and pelvis segments were computed with respect to the laboratory coordinate system. The laboratory coordinate system was defined at the posterior left corner of the force plate. Inverse dynamics were computed using a standard Newton–Euler approach. Moments were expressed as internal moments and were normalized to body mass and height (O’Connor and Bottum, 2009). A vertical ground reaction force threshold of 20 N was used to determine the onset and end of stance. The five waveforms of interest were: frontal plane trunk, pelvis, and hip angles and knee moment, as well as transverse plane knee angle. 2.4. Principal components analysis Stance phase of the overground running trials was time normalized to 101 points. The angle and moment data for each participant were an ensemble average of the five trials. For each of the five angle and moment waveforms of interest, a data matrix was created. The 101 data points comprised the columns and 40 participants comprised the rows of each matrix ðX 40101 Þ. The PCA approach used in the current investigation was based on a methodology described previously (Wrigley et al., 2006). The mean was computed for each column of the respective matrix. Then, the mean of each column was subtracted from each row in its respective column. The mean centered matrices were transformed into principal components using an eigenvector decomposition method on the input's covariance matrix ðC 101101 Þ. The PCA produced the eigenvectors (V 101101 Þ and eigenvalues (L1101 Þ. The eigenvector matrix consisted of the coefficients for each of the 101 principal components which defined a new coordinate space for the original waveform data (Wrigley et al., 2006). The eigenvalue matrix indicated the relative contribution each principal component had on the total variance in the data. For each matrix, the first three principal components were analyzed. Principal component score matrices (Z 40101 Þ were then computed by multiplying the mean-centered input matrix by transposing the eigenvector matrix: Z 40101 ¼ ðX 40 101 ð140 1 x1 101 ÞÞ V
0
101 101
ð1Þ
where x1 101 is each time normalized data point. The principal component scores represented how closely a runner's waveform matched the shape of its respective principal component (Wrigley et al., 2006). To determine if the retained principal components represented the original data adequately, a residual analysis was performed using the Q-statistic (Jackson, 1991). The Q-statistic is the sum of the squares of the residuals between participants' original waveform and the reconstructed curve based on the retained principal component (Wrigley et al., 2006). A Q-critical value (Qα) was calculated using an alpha level of 0.05 from a t-distribution (degrees of freedom¼39, Qα ¼2.0211) (Wrigley et al., 2006). Q-critical indicated whether the number of retained components reconstructed the original data adequately. For each participant, a Q-statistic value lower than Qα indicated that the original data were represented adequately by the retained principal components (Jackson, 1991). To aid in the interpretation of the principal components, the influence of a principal component on the combined groups' ensemble mean waveform was
Please cite this article as: Foch, E., Milner, C.E., The influence of iliotibial band syndrome history on running biomechanics examined via principal components analysis. Journal of Biomechanics (2013), http://dx.doi.org/10.1016/j.jbiomech.2013.10.008i
E. Foch, C.E. Milner / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎ computed (Ramsay and Silverman, 1997). A multiple M of the respective principal component's coefficients was added and subtracted to the overall mean curve. The 5th and 95th percentiles of the principal component score distribution were used as M values for the low ( ) and high ( þ) curves (Mantovani et al., 2012). Additionally, to interpret when during stance the retained principal components contributed to the variability in the shape of the waveform, the explained percent variance (r 2ji ) was computed for the ith principal component and the jth time point: pffiffiffiffi V ji Li 100% ð2Þ r 2ji ¼ cj where cj is the standard deviation of C at a given data point of the waveform (Wrigley et al., 2006). Differences in timing and magnitude of the relative contribution of a principal component to the overall waveform can be observed visually in the percent variance explained figures (Wrigley et al., 2006). Waveform matrix construction and all PCA calculations were performed using custom software (MATLAB, MathWorks, Natick, MA, USA).
2.5. Statistical analysis Descriptive statistics (means and standard deviations) were computed for the principal component scores of the waveforms: frontal plane hip, pelvis, and trunk angles, knee moment, as well as transverse plane knee angle. To assess between group differences, principal component scores for each waveform were compared using independent t-tests. Additionally, participant demographics were compared between groups via independent t-tests. Statistical analysis was performed using PASW 20.0 (IBM SPSS Statistics, Chicago, IL, USA). An alpha value of 0.05 was set for all tests. Due to the exploratory nature of the analysis, trends were considered 0.05 o alpha o0.10.
3
the pelvis waveform. The high principal component scores exhibited by runners with previous ITBS indicated that they exhibited less contralateral pelvic drop compared to controls. Principal component one was the primary mode of variation about the mean during 25–70% of the stance phase of running. Similar to principal component one, principal component three corresponded to a variation in the overall magnitude of the frontal plane pelvis angle. The low principal component scores exhibited by runners with previous ITBS indicated that their pelvis was closer to neutral than controls. Whereas, high principal component scores in controls indicated the contralateral pelvis was more elevated. During early ( o20%) and late (4 80%) stances, principal component three captured the variation about the mean of the pelvis in an elevated position. There was a trend towards a difference in principal component one for frontal plane trunk angle between runners with previous ITBS and controls (P ¼0.057). The effect of the first principal component was to add or subtract from the overall magnitude of the trunk waveform. High principal component scores revealed that runners with previous ITBS tended to lean their trunk more towards the stance limb compared to controls throughout the stance phase of running. Principal component one was the primary mode of variation about the mean during 20–70% of the stance phase of running. Lastly, there were no differences in the in principal components scores in transverse plane knee angle and the internal frontal plane knee moment.
3. Results The first three principal components accounted for 93.3–99.4% of the total variance in the five biomechanical waveforms of interest (Table 2). All waveforms were reconstructed using the scores and coefficients of the first three principal components. The Q-statistic indicated that all participants' ensemble mean angle and moment waveforms were sufficiently described by the retained principal components. Group ensemble averages for the original waveforms are depicted (Fig. 1). Principal component one scores in frontal plane hip angle were different between runners with previous ITBS compared to controls during the stance phase of running (Fig. 2; P¼0.004). The effect of the first principal component was to add or subtract from the overall magnitude of the hip angle waveform. The high curve indicated that the hip was more adducted than the low curve. Runners with previous ITBS exhibited negative principal component one scores (Table 3). Thus, runners with previous ITBS exhibited less hip adduction during stance than controls. Additionally, the percent variance explained curve revealed that the principal component one captured the primary variation about the mean during 20–60% of stance. Both groups exhibited similar scores for principal components two (P¼0.413) and three (P¼0.890). There were trends towards group differences for frontal plane pelvis angle in principal component one (P¼ 0.067) and principal component three (P¼ 0.063). The effect of the first principal component was to add or subtract from the overall magnitude of
Table 2 The first three principal components (PC) and Q-critical (Qα) for the angle and moment waveforms during the stance phase of running. Waveforms
Frontal plane trunk angle Frontal plane pelvis angle Frontal plane hip angle Frontal plane knee moment Transverse plane knee angle
o Qα (%)
PC (%) PC1
PC2
PC3
Total
76.3 72.2 65.5 83.7 79.6
18.4 19.0 24.6 6.2 9.4
4.7 5.8 5.5 3.5 7.6
99.4 97.0 95.5 93.3 96.7
100 100 100 100 100
4. Discussion It is unclear from the literature whether discrete peak biomechanical variables during running are associated with previous ITBS. Implementing a PCA may provide insight into potential differences in underlying patterns in kinematic and kinetic waveforms between runners with previous ITBS and controls. Therefore, the purpose of this study was to determine if using a PCA approach could detect differences in trunk and lower extremity time-series waveforms between female runners with previous ITBS and controls. The magnitude of hip adduction was smaller (principal component 1) in runners with previous ITBS compared to controls. Less hip adduction throughout stance may be attributed to either a tight iliotibial band or activation of hip abductor musculature acting to restrict motion into adduction. Additionally, trends towards group differences were observed in pelvic segment orientation which can influence the hip adduction angle. Runners with previous ITBS tended to assume a pelvic position that was closer to neutral compared to controls (principal components 1 and 3). Running with a more neutral pelvis helps to explain the smaller hip adduction angle exhibited by runners with previous ITBS compared to controls. Contralateral pelvic drop moves the thigh and pelvis closer medially. Therefore, increased contralateral pelvic drop exhibited by controls aids in explaining the greater hip adduction they assumed throughout stance. Collectively, there were no temporal differences in the overall hip and pelvis waveform pattern between groups. This suggests that both groups exhibited similar movement patterns but runners with previous ITBS had smaller waveform magnitude throughout the stance phase of running. While PCA data have not been reported in relation to ITBS in runners previously, peak hip angles have. There are contradictory findings in the literature regarding increased peak hip adduction angle as a factor associated with previous ITBS. One previous retrospective study (Ferber et al., 2010) and one previous prospective study (Noehren et al., 2007) both reported greater peak hip adduction angle in female runners with previous ITBS
Please cite this article as: Foch, E., Milner, C.E., The influence of iliotibial band syndrome history on running biomechanics examined via principal components analysis. Journal of Biomechanics (2013), http://dx.doi.org/10.1016/j.jbiomech.2013.10.008i
4
E. Foch, C.E. Milner / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
Fig. 1. Previous iliotibial band syndrome (ITBS; dashed line) and control groups' ensemble average plot of: (A) frontal plane hip angle, (B) frontal plane pelvis angle, (C) frontal plane trunk angle, (D) frontal plane knee moment, and (E) transverse plane knee angle. The error bars indicate one standard deviation of the mean. Moment is expressed as internal moment.
compared to controls. However, runners with previous ITBS have also exhibited similar peak hip adduction angle compared to controls (Foch and Milner, in press; Miller et al., 2007). Participants in previous ITBS studies included only women (Foch and Milner, in press; Ferber et al., 2010; Noehren et al., 2007) and both genders (Miller et al., 2007). However, differences exist in peak hip adduction angle between healthy female and male runners (Ferber et al., 2003). Therefore, only women were included in the current study to avoid gender differences in running mechanics confounding the results. In the current study, it appears group differences in hip pattern are due to less pelvis motion in the ITBS group throughout stance. Smaller peak hip adduction angles have been reported in female and male runners with current ITBS compared to controls (Grau et al., 2011). The authors' suggested less hip adduction may be due to a tight iliotibial band and hip abductors
(Grau et al., 2011). Iliotibial band flexibility was not measured in the present study. Therefore, we can only postulate that decreased hip adduction exhibited by runners with previous ITBS was due to less iliotibial band and hip abductor flexibility. By examining joint biomechanics, interferences can be made as to how the iliotibial band is affected. Excessive hip adduction would cause the iliotibial band to elongate. Since the iliotibial band functions to limit hip adduction, decreasing hip adduction may be a compensatory strategy to limit iliotibial band strain. A decrease in iliotibial strain may reduce compression experienced by innervated adipose tissue between the iliotibial band and lateral femoral epicondyle (Fairclough et al., 2007). Therefore, the pain associated with ITBS during running may be alleviated. A reduction in iliotibial band and hip abductor flexibility could result from trying to limit frontal plane hip motion during
Please cite this article as: Foch, E., Milner, C.E., The influence of iliotibial band syndrome history on running biomechanics examined via principal components analysis. Journal of Biomechanics (2013), http://dx.doi.org/10.1016/j.jbiomech.2013.10.008i
E. Foch, C.E. Milner / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
5
Fig. 2. The ensemble average (combined groups) for frontal plane angle waveforms: hip, pelvis, and trunk angles (solid gray) during the stance phase of overground running are displayed. The principal components (PC) are represented by adding ( þhigh) and subtracting ( low) a multiple M of the respective PC coefficients to the combined groups' ensemble average curve. The mean score contribution for runners with previous ITBS (dash black) and controls (solid black) are depicted for each PC. Additionally, the percent variance explained by PC 1 (dot), PC 2 (dash-dot), PC 3 (dash), and overall percent variance explained by the first 3 PCs (solid) is displayed for each waveform.
running. Potentially, runners in the current study tried to limit hip adduction when injured, and this pattern persisted after injury. No statistically significant differences were observed among frontal plane trunk angle and knee moment, as well as transverse plane knee angle between runners with previous ITBS and controls. The findings of the PCA are in agreement with a discrete analysis of the aforementioned variables (Foch and Milner, in press). However, there was a trend towards increased frontal plane trunk angle towards the stance limb in runners previous ITBS compared to controls which was not detected via discrete analysis (Foch and Milner, in press). Trunk lateral flexion away from the stance limb was postulated to affect the frontal plane knee
moment (Powers, 2010). However, runners in both groups exhibited trunk lateral flexion towards the stance limb. Trunk lateral flexion is likely not associated with previous ITBS (Foch and Milner, in press). Limitations to this study are acknowledged. First, runners in the previous ITBS group reported that they had been running without pain for at least 1 month prior to data collection. However, a specific period of time since injury was not recorded. It has not been investigated whether the duration since alleviation of ITBS symptoms influences biomechanics during running. Since this was a cross-sectional study, whether biomechanics during running postinjury are the same as pre-injury is uncertain. Future prospective
Please cite this article as: Foch, E., Milner, C.E., The influence of iliotibial band syndrome history on running biomechanics examined via principal components analysis. Journal of Biomechanics (2013), http://dx.doi.org/10.1016/j.jbiomech.2013.10.008i
E. Foch, C.E. Milner / Journal of Biomechanics ∎ (∎∎∎∎) ∎∎∎–∎∎∎
6
Table 3 The mean (standard deviation) for the principal component (PC) scores of the three retained PCs for each waveform of interest during overground running in women with previous iliotibial band syndrome (ITBS) and controls. The P value indicates the pairwise comparison for each waveform between groups. Controls
Retained PC
Frontal plane trunk angle
PC1 PC2 PC3
4.7 (13.6) 4.7 (16.6) 0.05 (7.2) 0.05 (8.4) 0.3 (3.5) 0.3 (4.3)
0.057 0.965 0.592
Frontal plane pelvis angle
PC1 PC2 PC3
5.9 (18.1) 2.5 (9.9) 1.7 (5.7)
5.9 (21.2) 2.5 (10.6) 1.7 (5.3)
0.067 0.124 0.063
Frontal plane hip angle
PC1 PC2 PC3
11.3 (23.5) 2.1 (18.1) 0.2 (6.6)
11.3 (22.8) 2.1 (12.8) 0.2 (8.2)
0.004a 0.413 0.890
Frontal plane knee moment
PC1 PC2 PC3
0.2 (1.4) 0.4 (0.5) 0.1 (0.4)
0.503 0.596 0.254
Transverse plane knee angle
PC1 PC2 PC3
0.5 (51.9) 2.2 (18.8) 2.9 (16.5)
0.949 0.424 0.231
0.2 (1.8) 0.04 (0.4) 0.1 (0.2) 0.5 (47.7) 2.2 (15.1) 2.9 (13.6)
Previous ITBS
P value
Waveform
a Runners with previous ITBS exhibited significant smaller PC1 scores than controls.
investigations are needed to better understand how running pattern is affected by ITBS injury status. This study assessed differences in kinematic and kinetic timeseries waveforms for variables associated with ITBS in runners via a PCA approach. Female runners with previous ITBS exhibited a frontal plane hip angle that was smaller in magnitude across the stance phase of running compared to controls. No differences were observed between groups in the retained principal components for other trunk and lower extremity waveforms. Reducing hip adduction may have enabled the runners with previous ITBS to reduce compression of a highly innervated fat pad between the femoral epicondyle and iliotibial band. This running pattern may have allowed runners to avoid lower extremity positions that were painful when injured. This pattern may have persisted after ITBS symptoms subsided.
Conflict of interest statement None.
Acknowledgments The authors would like to thank Dr. Jeffrey Reinbolt and Dr. Eugene Fitzhugh for their input on performing a PCA. References Barnes, A., Wheat, J., Milner, C.E., 2010. Use of gait sandals for measuring rearfoot and shank motion during running. Gait Posture 32, 133–135.
Bisseling, R.W., Hof, A.L., 2006. Handling of impact forces in inverse dynamics. J. Biomech. 39, 2438–2444. Cappozzo, A., Cappello, A., Della Croce, U., Pensalfini, F., 1997. Surface-marker cluster design criteria for 3-D bone movement reconstruction. IEEE Trans. Biomed. Eng. 44, 1165–1174. Fairclough, J., Hayashi, K., Toumi, H., Lyons, K., Bydder, G., Phillips, N., Best, T.M., Benjamin, M., 2006. The functional anatomy of the iliotibial band during flexion and extension of the knee: implications for understanding iliotibial band syndrome. J. Anat. 208, 309–316. Fairclough, J., Hayashi, K., Toumi, H., Lyons, K., Bydder, G., Phillips, N., Best, T.M., Benjamin, M., 2007. Is iliotibial band syndrome really a friction syndrome? J. Sci. Med. Sport 10, 74–76. Ferber, R., Noehren, B., Hamill, J., Davis, I.S., 2010. Competitive female runners with a history of iliotibial band syndrome demonstrate atypical hip and knee kinematics. J. Orthop. Sports Phys. Ther. 40, 52–58. Foch, E., Milner, C.E., 2013. Frontal plane running biomechanics in female runners with previous iliotibial band syndrome. J. Appl. Biomech. (in press). Fredericson, M., Cookingham, C.L., Chaudhari, A.M., Dowdell, B.C., Oestreicher, N., Sahrmann, S.A., 2000. Hip abductor weakness in distance runners with iliotibial band syndrome. Clin. J. Sports Med. 10, 169–175. Grau, S., Krauss, I., Maiwald, C., Axmann, D., Horstmann, T., Best, R., 2011. Kinematic classification of iliotibial band syndrome in runners. Scand. J. Med. Sci. Sports 21, 184–189. Grood, E.S., Suntay, W.J., 1983. A joint coordinate system for the clinical description of three-dimensional motions: application to the knee. J. Biomed. Eng. 105, 136–144. Jackson, J.E., 1991. A User's Guide to Principal Components. John Wiley & Sons Inc., New York, pp. 37–38. Kipp, K., McLean, S.G., Palmieri-Smith, R.M., 2011. Patterns of hip flexion motion predict frontal and transverse plane knee torques during a single-leg land-andcut maneuver. Clin. Biomech. 26, 504–508. Manal, K., McClay, I., Stanhope, S., Richards, J., Galinat, B., 2000. Comparison of surface mounted markers and attachment methods in estimating tibial rotations during walking: an in vivo study. Gait Posture 11, 38–45. Mantovani, G., Lamontage, M., Varin, D., Cerulli, G.G., Beaule, P.E., 2012. Comparison of total hip arthroplasty surgical approaches by principal component analysis. J. Biomech. 45, 2109–2115. McIlroy, W.E., Maki, B.E., 1997. Preferred placement of the feet during quiet stance: development of a standardized foot placement for balance testing. Clin. Biomech. 12, 66–70. Miller, R.H., Lowry, J.L., Meardon, S.A., Gillette, J.C., 2007. Lower extremity mechanics of iliotibial band syndrome during an exhaustive run. Gait Posture 26, 407–413. Noble, C.A., 1980. Iliotibial band friction syndrome in runners. Am. J. Sports Med. 8, 232–234. Noehren, B., Davis, I., Hamil, J., 2007. ASB clinical biomechanics award winner 2006 prospective study of the biomechanical factors associated with iliotibial band syndrome. Clin. Biomech 22, 951–956. O’Connor, K.M., Bottum, M.C., 2009. Differences in cutting knee mechanics based on principal components analysis. Med. Sci. Sports Exercise 41, 867–878. Orchard, J.W., Fricker, P.A., Abud, A.T., Mason, B.R., 1996. Biomechanics of iliotibial band friction syndrome in runners. Am. J. Sports Med. 24, 375–379. Powers, C.M., 2010. The influence of abnormal hip mechanics on knee injury: a biomechanical perspective. J. Orthop. Sports Phys. Ther. 40, 42–51. Ramsay, J.O., Silverman, B.W., 1997. Functional Data Analysis. Springer Verlag, New York, pp. 92–94. Renne, J.W., 1975. The iliotibial band friction syndrome. J Bone Jt Surg. 57, 1110–1111. Taunton, J.E., Ryan, M.B., Clement, D.B., McKenzie, D.C., Lloyd-Smith, D.R., Zumbo, B. D., 2002. A retrospective case-control analysis of 2002 running injuries. Br. J. Sports Med. 36, 95–101. Taunton, J.E., Ryan, M.B., Clement, D.B., McKenzie, D.C., Lloyd-Smith, D.R., Zumbo, B. D., 2003. A prospective study of running injuries: the Vancouver Sun Run “In Training” clinics. Br. J. Sports Med. 37, 239–244. Thomas, S., Reading, J., Shepard, R.J., 1992. Revision of the Physical Activity Readiness Questionnaire (PAR-Q). Can. J. Sport Sci. 17, 338–349. Wrigley, A.T., Albert, W.J., Deluzio, K.J., Stevenson, J.M., 2005. Differentiating lifting technique between those who develop low back pain and those who do not. Clin. Biomech. 20, 254–263. Wrigley, A.T., Albert, W.J., Deluzio, K.J., Stevenson, J.M., 2006. Principal component analysis of lifting waveforms. Clin. Biomech. 21, 567–578. Wu, G., Cavanagh, P.R., 1995. ISB recommendations for standardization in the reporting of kinematic data. J. Biomech. 28, 1257–1261.
Please cite this article as: Foch, E., Milner, C.E., The influence of iliotibial band syndrome history on running biomechanics examined via principal components analysis. Journal of Biomechanics (2013), http://dx.doi.org/10.1016/j.jbiomech.2013.10.008i