Patellofemoral pain (PFP) syndrome is the most common overuse injury that runners sustain (32). Runners that experience PFP typically describe the pain as having a subtle onset (i.e., little-to-no pain at the beginning of a running session) and originating in the anterior aspect of the knee. The pain tends to progressively worsen throughout the running session with the repetitive action of single-leg weight bearing over a flexed knee and may result in the inability to continue the run. If individuals continue to run without addressing the syndrome, the severity may intensify over time and could lead to a more severe condition such as osteoarthritis. Despite the high incidence rate and severity of PFP, the mechanisms that contribute to the syndrome remain relatively unclear. Although some authors have speculated that abnormal kinematics may play a role in the cause of PFP, few studies have documented these relationships.
Excessive eversion has often been proposed as a mechanism that contributes to PFP (8,13,17,18,26,27,33). Owing to the tight articulation in the ankle joint mortise (16,22-24,30,31), excessive eversion during weight bearing is likely to influence motion of the tibia, which can then disrupt patellofemoral mechanics. For example, with excessive eversion, there is a greater medial displacement of the foot, which may lead to tibial abduction and a valgus position at the knee (13,27,35). Knee valgus is typically associated with an increase in the quadriceps angle, creating a larger lateral force acting on the patella and predisposing the patella to maltracking (27). Patellofemoral maltracking can result in increased contact pressure between the lateral facet and the lateral femoral condyle, leading to articular cartilage degradation and pain over time.
The patellofemoral joint may also be influenced proximally by abnormal motions of the femur (14,27,28). Powers et al. (28) demonstrated that during a single-leg squat, femoral internal rotation occurred independently from patellar motion in subjects with patellar instability. It was concluded that the observed patellofemoral maltracking was the result of the femur rotating against the patella, rather than the patella against the femur. Increased femoral internal rotation would likely result in a larger quadriceps angle and lead to increased contact pressure between the lateral facets and the lateral femoral condyles. In addition to abnormal transverse plane motion, femoral adduction in the frontal plane may also contribute to PFP. With an increase in femoral adduction, there is likely an increase in knee valgus. This can be the result of hip adduction or some combination of hip adduction and tibial abduction (27). Thus, knee valgus may increase the quadriceps angle and lead to increased joint contact pressure.
Only a handful of studies have examined lower extremity kinematics in runners with PFP, leading to relatively inconclusive findings. Messier et al. (18) found no difference between runners with PFP and controls for peak eversion, eversion excursion, and peak eversion velocity. However, this may have been a result of the two-dimensional method used to assess rear foot motion, as the authors stated that rear foot motion should continue to be evaluated as a discriminatory measure for runners with knee pain. In a preliminary prospective study of only nine subjects, McClay Davis et al. (17) reported no difference in peak eversion, peak knee adduction, and peak knee internal rotation between runners who later developed PFP and uninjured controls. However, amid the lower subject size, the authors did find an increase in both peak hip adduction and hip internal rotation for the PFP group. During fast walking, those with PFP have been reported to display decreased peak knee flexion and femoral rotation when compared with controls (7,25,26). It was suggested that decreased knee flexion was a compensatory mechanism aimed at decreasing knee joint loading, thus reducing the pain.
A unique aspect of PFP is that runners with the syndrome generally do not experience pain at the beginning of a running session. Yet no studies have examined the mechanics of runners with PFP at the end of a prolonged run where pain levels are greatest, and only a handful of studies have examined uninjured runners in an exerted state. Of these studies, increases have generally been found at the end for kinematic variables of rear foot eversion and velocity (3,5) and knee flexion and velocity (5,21). In addition, increases have been reported at the end of a run for peak impact acceleration of the tibia, which represents the impulsive loading along the shank that is transmitted to the knee (5,19-21,34). However, these increases have occurred with increased knee flexion (5,21). Derrick et al. (5) suggested that decreasing knee flexion, which may be a compensatory mechanisms used by runners with PFP, would reduce the effective mass that is accelerated, thus reducing peak impact acceleration. Yet it remains unclear how foot, knee, and hip kinematics in the same population, be it runners with PFP or uninjured runners, are influenced when running in an exerted state.
The purpose of this study was to investigate the influence of exerted state running on lower extremity kinematics during the stance phase in runners with PFP while compared with controls. It was hypothesized that runners with PFP would generally display larger peaks, excursions, and peak velocities for eversion, tibial internal rotation, hip adduction, and hip internal rotation at the end of the run and when compared with controls. It was also expected that the PFP group would demonstrate smaller values for knee flexion, knee adduction (varus), knee internal rotation, and for peak tibial impact acceleration. Lastly, it was hypothesized that for the PFP group, exerted state running, where pain is the greatest, would have a more profound influence on the kinematic variables such that larger changes in these variables would be observed compared with controls.
A total of 20 runners with PFP and a control group of 20 healthy, uninjured runners were recruited for the study (Table 1). Sample size was based on an a priori sample size calculation (α = 0.05, β = 0.20, and desired effect size = 0.66) using variability from kinematic pilot data, which resulted in an estimation of 16 to 20 runners for adequate statistical power. All runners were between the ages of 18 and 45 yr and were free of any cardiovascular pathology, as determined by a physical activity readiness questionnaire. All were recreational runners (≥10 miles·wk−1), ran with an observed heel-strike running pattern, and were nonorthotic users. The PFP group consisted of runners who met all PFP inclusion criteria on completion of a thorough evaluation by a licensed physical therapist. This included a complaint of anterior knee pain when running that had persisted for a minimum of 2 months. The intensity of the knee pain during running was at least a 3 (moderate pain) on a 0 (no pain) to 10 (maximum pain) point visual analog scale for pain (2). The knee pain was retropatellar, was believed to originate from the patellofemoral joint, and was confirmed on palpation that the pain was peripatellar. Subjects were excluded if it was determined that ligament instability, meniscal pathology, or patellar tendinitis was present, if the knee pain resulted from acute trauma, or if they were receiving physical therapy. The control group consisted of runners who had no known lower extremity pathologies and were free of any lower extremity injury at the time of data collection. The two groups were matched according to sex. All subjects signed an informed consent form approved by the University of Delaware's Human Subjects Review Board.
For the prolonged run, each subject was fitted with a neutral, cushioning running shoe (Nike Air Pegasus, Beaverton, OR). The tested limb during the run was the side with the most painful knee for the PFP group, whereas the control group was chosen at random. Retroreflective markers were positioned on the tested limb at the sites of: the sides of the shoe at the first and fifth metatarsal heads, the tip of the shoe, the medial and lateral malleoli, the medial and lateral femoral condyles, both greater trochanters, and on the pelvis at the right anterior superior iliac spine, right iliac crest, and the L5-S1 interspace (Fig. 1A). These markers were used to develop the anatomical coordinate systems. Additional tracking markers were positioned in clusters of four on the femur and tibia. These markers were affixed to thermoplastic shells and attached to straps with Velcro and secured around the thigh and shank. Lastly, three rear foot tracking markers were attached directly to the calcaneus, as it has been previously reported that markers placed directly on the heel counter of the shoe overestimate rear foot motion (30). These markers protruded through windows cut in the heel counter of the shoes (Fig. 1B). The anatomical markers of the pelvis also served as tracking markers. With the markers affixed to the lower extremity, a standing calibration trial was collected. The anatomical markers were then removed leaving the tracking markers for the prolonged run.
Before performing the prolonged run, subjects underwent a 5-min warm-up session on the treadmill (model 18-60; Quinton, Bothell, WA). This consisted of low-intensity jogging (HR approximately 120 beats·min−1) and allowed the subjects to accommodate to the treadmill environment (1). A Polar A3 HR monitor (Polar Electro Inc, Woodbury, NY) was used to monitor HR throughout the data collection. After the warm-up, subjects were given the opportunity to stretch before performing the prolonged run. The prolonged run was performed at a self-selected pace that best represented the subject's typical training pace. To determine the self-selected pace, each subject was allowed to increase or decrease the treadmill speed until a comfortable running pace was found. The prolonged run ended when one of three events was reached: 1) 85% of the subject's HR maximum (1), 2) a score of 17 (very hard) on the rating of perceived exertion scale (2), and 3) for the PFP group only, a score of 7 (very strong) out of 10 on the visual analog scale for pain (2). Three-dimensional kinematic data were collected during the prolonged run using a VICON passive, six-camera, three-dimensional motion analysis system (model MCAM 1; VICON, Oxford Metrics, UK) at a sampling frequency of 120 Hz. The axial acceleration of the tibia was measured using a piezoelectric accelerometer (PCB model 353B17) that was mounted to the anteromedial aspect of the distal tibia and sampled at a frequency of 1080 Hz. When one of the three stopping criteria was met, the run ended, and subjects performed a cool-down session consisting of gradual diminishing walking intensities until the HR was below 120 beats·min−1 (1).
Data processing and reduction.
Kinematic data were collected at the beginning of the run and at the end when one of the three stopping criteria was met, with each collection consisting of 20 consecutive footfalls. The three-dimensional marker coordinates were filtered using an 8-Hz low-pass, fourth-order, 0 lag Butterworth filter. A 75-Hz low-pass filter was used for the accelerometry data. Data reduction was performed using Visual3D software (C-Motion Inc, Rockville, MD). Joint angles were decomposed about joint coordinate systems and referenced to the proximal segment (10). To determine the stance phase for each footfall, an algorithm was used to predict heel-strike and toe-off (11). Heel-strike was identified as the time of the minimum acceleration of the foot, whereas toe-off occurred at the minimum acceleration of the shank. With the stance phase identified, the angular data during stance were interpolated to 100 points to normalize stance to 100%. Customized computer programs (LabView 6i; National Instruments, Austin, TX) were used to evaluate the kinematics.
Kinematic joint motions of interest in the frontal plane included 1) rear foot eversion (EV); 2) knee adduction (KADD), referenced as tibia relative to femur; and 3) hip adduction (HADD), referenced as femur relative to pelvis. The transverse plane included 4) tibial internal rotation (TIR), inferred as tibial motion because the foot is fixed to the ground during stance; 5) knee internal rotation (KIR), referenced as tibia relative to femur; and 6) hip internal rotation (HIR), referenced as femur relative to pelvis. Lastly, the sagittal plane included 7) knee flexion (KF). For all joint motions, peak angles (°) and peak velocities (°·s−1) were determined as the maximum value that occurred during stance. Excursions were calculated as the peak angle during the first half of stance minus the minimum angle preceding the peak. Peak impact acceleration, obtained from the tibial mounted accelerometer, was determined as the peak positive acceleration value that occurred during the stance phase.
On visual inspection of the data, it was noted that several of the runners with PFP displayed unexpected frontal plane patterns at the knee and the hip. It was expected that all runners would display movements of adduction at both the knee and the hip during the first half of stance and then reverse this movement into abduction during the second half of stance (9). However, several of the runners in the PFP group demonstrated the opposite patterns (Fig. 2). Therefore, as these subjects seemed to exhibit distinctly different movement patterns, the PFP group as a whole was subdivided into three groups: 1) those who exhibited a knee abduction (valgus) type of movement, 2) those who exhibited a hip abduction type of movement, and 3) those who generally displayed typical knee and hip adduction patterns. As a preliminary analysis, the peak angles of these three subgroups were descriptively compared with the uninjured control group.
For KADD, a typical pattern would be to begin stance with the knee in a slightly abducted position, then decrease abduction such that a peak on the KADD graph was reached at midstance, and then return to a position of slight abduction at the end of stance (Fig. 2A) (9). However, 5 (4 women and 1 man) of the 20 runners with PFP did not display this KADD pattern, whereas only 1 of the 20 uninjured runners did not display the pattern. Rather, these five subjects with PFP began stance in a position of slight knee abduction, then increased knee abduction (i.e., moved toward greater knee valgus) such that a minimum on the KADD graph was reached at midstance, and then returned to a position of slight abduction at toe-off. As such, for data analysis, the dynamic knee frontal plane angle that occurred during the first half of stance for these runners was represented by the peak knee abduction angle, which is represented as the minimum on the KADD graph. This angle, along with the peak KADD angle for the remaining subjects, typically occurred between 40% and 60% of stance.
For HADD, a typical pattern would be to begin stance with the hip in a slightly adducted position, then increase adduction to a peak angle at midstance, and then return to a position of slight adduction at the end of stance (Fig. 2B) (9). However, 8 (4 women and 4 men) of the 20 runners with PFP did not display this HADD pattern during stance, whereas only 1 of the 20 uninjured runners did not (different subject than that of the knee abduction). Instead, these subjects began stance in a position of slight hip adduction, then decreased hip adduction (i.e., moved toward hip abduction) such that a minimum was reached at midstance, and then returned to a position of slight adduction at toe-off. As such, for data analysis, the dynamic hip frontal plane angle that occurred during the first half of stance for these runners was represented by the minimum HADD angle. This angle, along with the peak HADD angle for the remaining subjects, typically occurred between 40% and 60% of stance.
For each subject, all kinematic variables were extracted from individual trials, and then averaged across the 20 trials, creating an individual-subject mean. Group means were calculated using the individual subject means. Between-subject variability was calculated as the SD of the group means. Kinematic variables were compared at the beginning and end of the run using a two-way (group × time) mixed-model ANOVA with time as the repeated measure. For all analyses, the α level was set at 0.05. For the PFP subgroups, we operationally identified differences as 15% or more between a PFP subgroup and the control group and considered this to be clinically meaningful for the purposes of this article.
Subject characteristics and anthropometric measures.
The group results for the average treadmill running velocity, the average run time, and the stopping criteria along with anthropometric measures appear in Table 1. All variables were similar between groups with the exception of run time, where the total time of the run was 10 min shorter in the PFP group compared with the control group. For the runners in the PFP group, the knee pain score at the beginning of the run ranged between 0 (no pain) and 1 (very weak) on the 10-point (max pain) visual analog pain scale (2).
Although no statistically significant group by time interaction effects were detected for any of the peak angles or peak impact acceleration, significant main effects for both group and time were observed (Table 2). Group main effects were observed for both peak KF (P = 0.034) and HADD (P = 0.044) to where significantly smaller peaks were displayed in the PFP group when compared with controls. For the main effect of time, the peak angles of EV (P < 0.001), TIR (P < 0.001), and KIR (P < 0.001) all significantly increased at the end of the run when compared with the beginning. Conversely, peak KADD (P = 0.020) significantly decreased (moved toward knee valgus) at the end of the run.
No statistically significant group by time interaction effects were detected for any of the excursion variables (Table 2). However, significant main effects for time were observed for several of the excursions. Significant increases were observed at the end of the run when compared with the beginning for EV (P = 0.002), TIR (P = 0.026), KIR (P < 0.001), KADD (P = 0.006), and HIR (P = 0.007). No statistically significant main effects for group were detected for any of the excursions.
No statistically significant group by time interaction effects were detected for any of the peak velocity variables (Table 2). However, significant main effects for both group and time were detected for several of the peak velocity variables. Group main effects were observed for HADD (P = 0.008) and HIR (P = 0.028) to where the PFP group displayed smaller peaks compared with controls. For the main effect of time, significant increases were observed at the end of the run for the peak velocities of EV (P = 0.011), KIR (P < 0.001), and KADD (P = 0.016). Conversely, a significant decrease in peak velocity at the end of the run was observed for KF (P = 0.003).
For the PFP subgroups, all clinically meaningful differences between a subgroup and the uninjured group were generally a magnitude of two degrees or more. The runners in the knee abduction (valgus) PFP subgroup (n = 5) displayed smaller peak angles for EV, TIR, KIR, and HIR, in addition to KADD, when compared with controls (Table 3). The runners in the hip abduction subgroup (n = 8) demonstrated smaller peaks for EV, KF, and HIR, in addition to HADD, when compared with controls. This subgroup also displayed a larger peak KADD compared with controls. Lastly, the runners in the PFP subgroup that displayed typical frontal plane motions at the knee and the hip (n = 7) demonstrated a smaller peak KIR and larger peaks for KADD and HIR compared with controls.
This study investigated the lower extremity kinematics in runners with PFP compared with uninjured runners at the beginning and at the end of a prolonged run. No significant interactions were observed between the PFP group as a whole and the control group. This suggests that both groups responded similarly to the prolonged run, which is contrary to our hypothesis where we expected greater differences in the PFP group at the end because of the combination of running with pain and exertion. However, a significant main effect for group indicated that the PFP group displayed less peak knee flexion compared with controls, which was expected (Table 2). Previous studies have suggested that decreasing knee flexion may be a protective mechanism against knee pain in individuals with PFP (7,25,26). By reducing knee flexion, patellofemoral compressive forces are reduced, which should result in a reduction in pain. On the basis of the significant main effect for time, this mechanism was used from the very beginning of the run because no change in knee flexion was observed. However, knee pain levels were greatest at the end of the run. Thus, although adopting this mechanism at the start of the run did not reduce pain, it may have delayed the onset and slowed the progression throughout the run.
Aside from the decreased knee flexion in the PFP group, the kinematic differences found between groups were opposite to what was expected. It is commonly theorized that PFP is related in some capacity to increases in eversion, knee valgus (from hip adduction, tibial abduction, or a combination of both), and excessive internal rotations of the tibia, knee, and hip (8,13,17,18,26,27,33). However, the PFP group generally displayed less motion and lower peak velocities when compared with the uninjured group (Table 2). One possible explanation may be that this reduced motion was a strategy used in an attempt to minimize dynamic malalignment and reduce pain. Although this strategy might have been effective initially, the ability to compensate through this mechanism likely lessened as exertion levels increased. The significant main effects for time indicated increases in most of the joint motions and peak velocities at the end of the run, which were associated with the greatest levels of exertion and pain in the PFP group. Thus, while a strategy aimed at reducing joint motion might reduce knee pain, in the presence of exertion it becomes difficult to maintain this strategy, but it may still slow the progression of pain.
Although a compensatory mechanism is one explanation for the unexpected results, further inspection of the data revealed some interesting findings. Based on a visual inspection of the data, it appeared that there were three distinct subgroups within the PFP group. Analyses of these subgroups provided an alternate explanation for the unexpected findings in the PFP group as a whole. The PFP subgroup that demonstrated knee valgus during stance also displayed less joint motion in the peaks of EV, TIR, KIR, and HIR compared with controls (Table 3). This is counterintuitive given that the knee valgus is often associated with increased lower extremity joint motions in the frontal and transverse planes, resulting in a tendency for the leg to collapse medially (13,35). Thus, the runners in this PFP subgroup seem to have been compensating by limiting motion in their joints in an attempt to stabilize the leg to prevent further dynamic malalignment and reduce pain. However, the peaks of these motions, including knee valgus, increased at the end of the run compared with the beginning, which corresponded with increased pain at the end of the run.
The knee valgus that was exhibited by the runners in the valgus PFP subgroup may be related to hip strength. Recreational runners and other athletes with PFP have been found to demonstrate weakness in the hip abductor and external rotator musculature relative to uninjured controls (4,6,12,15,29). This weakness has been suggested as a mechanism that predisposes the femur to increased adduction and internal rotation during dynamic activity, contributing to a more valgus position of the knee (4,12,15,27,29). Exerted state running has been shown to intensify this mechanism for runners with PFP. Dierks et al. (6) reported a relationship to where the weaker the hip abductor muscles, the greater the HADD peak angle at the start of a run. But this relationship became even stronger at the end when runners were exerted and experiencing the most pain (6). Thus, in response to weaker hip muscles, a compensatory strategy of limiting joint motion may have the effect of preventing dynamic malalignment and further increases in knee pain when running.
The most surprising PFP subgroup was the hip abduction subgroup. This subgroup seems to have accounted for the decreased HADD seen in the PFP group as a whole, as the remaining two subgroups did not differ from controls (Table 3). Hip abduction typically occurs from the leg abducting relative to the pelvis. However, on closer inspection of the data, the pelvis was noted to elevate on the contralateral side, whereas the femur remained relatively stable in the frontal plane. This may also be a mechanism related to hip strength weakness in that elevation of the pelvis on the opposite side may have been secondary to leaning the trunk to the ipsilateral side in compensation for weaker hip abductor muscles. This may have been a compensatory mechanism used to reduce the demands on the weaker hip abductors to stabilize the pelvis in the frontal plane.
Another interesting finding for the hip abduction PFP subgroup was that this was the only subgroup that seemed to adopt the decreased KF mechanism. Although the PFP group as a whole displayed less peak KF compared with controls, subdividing the PFP group revealed that this difference resulted from the hip abduction subgroup (Table 3). The decreased KF may have been a compensatory mechanism aimed at decreasing knee joint loading, thereby reducing the pain (7,25,26). In addition, increased KADD (knee varus) and decreased EV were both observed in this subgroup. These findings may also be secondary to a lateral trunk lean in that as the center of mass moves laterally over the support leg, inversion would increase along with knee varus in an effort to maintain balance. In terms of slowing the progression of knee pain, these potential compensatory mechanisms exhibited by this PFP subgroup seem to have been successful as five of the eight runners in this PFP subgroup did not exceed a visual analog pain score of 5 at the end of the run. In comparison, of the remaining 12 runners with PFP, 3 reported a pain score of 6 and the remaining 9 reported a 7 at the end of the run. It was also interesting to note that 80% of the males (4/5 runners) and only 27% of the females (4/15 runners) with PFP displayed movement toward hip abduction, which may be an indication that running mechanics related to PFP are sex specific.
The runners with PFP in the subgroup that displayed typical frontal plane patterns at the knee and hip generally demonstrated similar kinematics compared with controls (Table 3). However, they did display increased HIR, which is thought to be related to weakness of the hip external rotators. This weakness has been shown to be related to recreational athletes with PFP (4,6,12,15,29). Also observed in this PFP subgroup was a decreased peak KIR, which is not surprising in light of the increased HIR. With increased HIR, there is most likely an increase in internal rotation of the femur. Because peak TIR remained unchanged in this subgroup, the increased femoral internal rotation on the tibia would have the effect of decreasing peak KIR.
Owing to the cross-sectional design of the current study, conclusions regarding cause and effect cannot be drawn. As the runners in the PFP group were symptomatic at the time of data collection, it is unknown if the kinematic differences are associated with the development of PFP or are the result of PFP. The current study has provided preliminary evidence of three distinct groups of runners with PFP that exhibit unique kinematic patterns compared with uninjured runners. However, caution is urged when interpreting these results as the differences were based on relatively low PFP subgroup sample sizes (<10), whereas comparisons were made using a minimum 15% difference. Thus, we cannot truly state that any of the observed group differences were statistically significant, although they may have clinical relevance. It is also possible that because of the symptomatic nature of the runners with PFP, these subgroups demonstrate various kinematic strategies used to compensate for the pain when running. Further investigation is needed to determine whether there are several distinct mechanisms of PFP and whether these are sex related. This information would provide a better guide to the treatment and prevention processes for this multifactorial syndrome. Future research is also needed to gain insight into where the changes in kinematics occur throughout the run and how they are related to increasing pain levels. However, the kinematics alone may not be enough to clearly delineate the differences between runners with and without PFP, thus the inclusion of inverse dynamics is warranted. Future studies should then consider that examining kinematics alone may lead to an inability to find definitive answers regarding questions related to runners with PFP. Prospective and longitudinal studies are needed to more completely delineate the role of kinematic mechanisms as they relate to the development and progression of knee pain during running.
The PFP group as a whole displayed less overall motion compared with uninjured controls. This may be indicative of a strategy aimed at limiting motion of the lower extremity in response to knee pain. However, at the end of the run, increases in joint motion occurred in both groups. Three distinct subgroups of runners with PFP emerged. When compared with controls during the prolonged run, each subgroup demonstrated unique kinematic mechanisms that may be associated with PFP. In the PFP subgroup that demonstrated knee valgus, increased knee valgus and decreased peak motions were noted in other joints. In the PFP subgroup that demonstrated hip abduction, less knee flexion and overall motion were noted. In the PFP subgroup that demonstrated the expected frontal plane motions at the knee and hip, increased hip internal rotation and decreased knee internal rotation were observed. These results suggest that several different kinematic mechanisms associated with PFP may exist.
This study was funded by the Department of Defense grant DAMD17-00-1-0515.
The authors thank the dedicated and professional efforts of Dr. John Scholz for his critical reading of the article, John Willson for his assistance with subject diagnoses and with data collections, and Kevin McCoy for his assistance with data collections.
The results of the present study do not constitute endorsement by the American College of Sports Medicine.
1. American College of Sports Medicine. Guidelines for Exercise Testing and Prescription
. 7th ed. Philadelphia (PA): Williams & Wilkins; 2000. p. 74, 102, 137-138.
2. Borg G. Borg's Perceived Exertion and Pain Scales
. Champaign (IL): Human Kinetics; 1998. p. 104.
3. Christina KA, White SC, Gilchrist LA. Effect of localized muscle fatigue on vertical ground reaction force and ankle joint motion during running. Hum Mov Sci
4. Cichanowski HR, Schmitt JS, Johnson RJ, Niemuth PE. Hip strength in collegiate female athletes with patellofemoral pain. Med Sci Sports Exerc
5. Derrick TR, Dereu D, Mclean SP. Impacts and kinematic adjustments during an exhaustive run. Med Sci Sports Exerc
6. Dierks TA, Manal KT, Hamill J, Davis IS. Proximal and distal influences on hip and knee kinematics in runners with patellofemoral pain during a prolonged run. J Orthop Sports Phys Ther
7. Dillon PZ, Updyke WF, Allen WC. Gait analysis with reference to chondromalacia patellae. J Orthop Sports Phys Ther
8. Eng JJ, Pierrynowski MR. The effect of soft foot orthotics on three-dimensional lower-limb kinematics during walking and running. Phys Ther
9. Ferber R, McClay Davis I, Williams DS. Gender differences in lower extremity mechanics during running. Clin Biomech
10. Grood ES, Suntay WJ. A joint coordinate system for the clinical description of three-dimensional motions: application to the knee. J Biomech Eng
11. Hreljac A, Stergiou N. Phase determination during normal running using kinematic data. Med Biol Eng Comput
12. Ireland ML, Willson JD, Ballantyne BT, McClay Davis I. Hip strength in females with and without patellofemoral pain. J Orthop Sports Phys Ther
13. James SL, Bates BT, Osternig LR. Injuries to runners. Am J Sports Med
14. Lee TQ, Morris G, Csintalan RP. The influence of tibial and femoral rotation on patellofemoral contact area and pressure. J Orthop Sports Phys Ther
15. Leetun DT, Ireland ML, Willson JD, Ballantyne BT, McClay Davis I. Core stability measures as risk factors for lower extremity injury in athletes. Med Sci Sports Exerc
16. McClay I, Manal K. Coupling parameters in runners with normal and excessive pronation. J Appl Biomech
17. McClay Davis I, Dierks TA, Ferber R, Hamill J. Lower extremity mechanics in patients with patellofemoral joint pain: a prospective study. Proceedings of the 27th Annual Meeting of the American Society of Biomechanics
. Toledo (OH): 2003. p. 23.
18. Messier SP, Davis SE, Curl WW, Lowery RB, Pack RJ. Etiologic factors associated with patellofemoral pain in runners. Med Sci Sports Exerc
19. Mizrahi J, Verbitsky O, Isakov E. Fatigue-related loading imbalance on the shank in running: a possible factor in stress fractures. Ann Biomed Eng
20. Mizrahi J, Verbitsky O, Isakov E. Shock accelerations and attenuation in downhill and level running. Clin Biomech
21. Mizrahi J, Verbitsky O, Isakov E, Daily D. Effect of fatigue on leg kinematics and impact acceleration in long distance running. Hum Mov Sci
22. Nawoczenski DA, Saltzman CL, Cook TM. The effect of foot structure on the three-dimensional kinematic coupling behavior of the leg and rear foot. Phys Ther
23. Nigg BM, Cole GK, Nachbauer W. Effects of arch height of the foot on angular motion of the lower extremities in running. J Biomech
24. Norkin CC, Levangie PK. Joint Structure and Function
. 4th ed. Philadelphia (PA): FA Davis Company; 1992. p. 447-9.
25. Powers CM, Heino JG, Rao S, Perry J. The influence of patellofemoral pain on lower limb loading during gait. Clin Biomech
26. Powers CM, Chen PY, Reischi SF, Perry J. Comparison of foot pronation and lower extremity rotation in persons with and without patellofemoral pain. Foot Ankle Int
27. Powers CM. The influence of altered lower-extremity kinematics on patellofemoral joint dysfunction: a theoretical perspective. J Orthop Sports Phys Ther
28. Powers CM, Ward SR, Fredericson M, Guillet M, Shellock FG. Patellar kinematics during weight-bearing and non-weight bearing movements in persons with patellar subluxation. J Orthop Sports Phys Ther
29. Robinson RL, Nee RJ. Analysis of hip strength in females seeking physical therapy treatment for unilateral patellofemoral pain syndrome. J Orthop Sports Phys Ther
30. Stacoff A, Nigg BM, Reinschmidt C, van den Bogert AJ, Lundberg A. Tibiocalcaneal kinematics of barefoot versus shod running. J Biomech
31. Stacoff A, Nigg BM, Reinschmidt C, et al. Movement coupling at the ankle during the stance phase of running. Foot Ankle Int
32. Taunton JE, Ryan MB, Clement DB, McKenzie DC, Lloyd-Smith DR, Zumbo BD. A retrospective case-control analysis of 2002 running injuries. Br J Sports Med
33. Tiberio D. The effect of excessive subtalar joint pronation on patellofemoral mechanics: a theoretical model. J Orthop Sports Phys Ther
34. Verbitsky O, Mizrahi J, Voloshin A, Treiger J, Isakov E. Shock transmission and fatigue in human running. J Appl Biomech
35. Williams DS, McClay IS, Hamill J. Arch structure and injury patterns in runners. Clin Biomech