Trunk and Lower Extremity Kinematics During Stair Descent in Women With or Without Patellofemoral Pain

Participants

Because of the high incidence and prevalence of PFP in this population, our participant recruitment was limited to young, physically active women.^1,26,27 We used a cross-sectional research design for this study. Pilot data were collected on a convenience sample of healthy participants to perform an a priori power analysis and determine an appropriate sample size. Effect sizes were estimated based on these trunk kinematic data (trunk flexion = 5.90° ± 1.52°, trunk lateral flexion = 3.35° ± 0.60°, trunk rotation = 7.73° ± 2.78°) and the expectation that PFP participants would demonstrate 34% to 66% difference in these variables. These percentage differences have been previously observed for hip kinematics during a similar stair-descent task by McKenzie et al²⁰ in their comparison of PFP and control groups. Based on our anticipated effect sizes for trunk kinematics, we determined that 20 participants per group would provide a priori power of 0.80 to detect a difference in trunk kinematics between groups at an α level of .05.

Using the criteria in Table 1, we recruited 40 female participants from the local university population, of whom 20 constituted the PFP group and 20 served as the control group. We matched the control group to the PFP group based on age, height, mass, and leg dominance.^28–31 The principal investigator (B.G.S.), a certified athletic trainer, evaluated each participant before testing to determine compliance with the inclusion and exclusion criteria provided in Table 1.

Table 1.  Inclusion and Exclusion Criteria for the Patellofemoral Pain and Control Groups

View Fullscreen

Each participant read and signed an informed consent approved by the university institutional review board before data collection. We sampled data from the affected leg for the PFP group. If PFP participants experienced pain bilaterally, we tested the most affected leg, determined as the most painful limb subjectively reported by the participant.^28,29 For the control group, the test leg was the same as for the matched participant with PFP.

Instrumentation

Trunk and lower extremity kinematic data were collected with a 7-camera optical motion-capture system (model MX; Vicon Motion Systems, Los Angeles, CA) during a stair-descent task at a sampling frequency of 120 Hz. The stair-descent task consisted of 4 steps. We constructed the steps based on standard building code specifications for step height and tread, allowing for a step height of 20 cm between steps and a tread depth of 30.5 cm. We used 2 force plates (model FP4060-10; Bertec Corp, Columbus, OH) to collect ground reaction force data and identify the stance phase of stair descent for analysis. We collected force plate data synchronously with the kinematic data at a sampling frequency of 1200 Hz. The force plates were located under the second and third steps (Figure 1).^29,32–34

View Figure

• Download as Powerpoint
• Download Figure

Procedures

Each participant attended 2 sessions: an initial screening session to ensure she met inclusion and exclusion criteria for her group and a second session during which all testing was conducted. During the testing session, we recorded demographic information that included age, height, mass, test leg, and leg dominance. We then asked participants to perform a stair-descent task.

Stair-Descent Task

We instructed each participant to descend 4 steps in a step-over-step fashion,³⁴ leading with the unaffected leg (Figure 2), which allowed data to be collected as she lowered herself using the affected leg from the second to the fourth step. Each participant had to complete a minimum of 2 strides immediately after descent to maintain a continuous movement pattern.²⁹ Stepping was performed in time to a metronome set at 96 beats per minute to control differences in gait velocity.^29,36,37 Each participant was allowed to practice 5 trials. Then participants performed 5 acceptable trials with 30 seconds of rest between trials. We collected 5 trials per participant to guard against marker occlusion during testing and ensure 3 adequate trials for data analysis.

View Figure

• Download as Powerpoint
• Download Figure

Acceptable trials included those during which the participant (1) walked with the specified cadence, (2) took a minimum of 2 strides after the stair-descent task, (3) made contact with the second step with the appropriate foot, and (4) completed the task in a step-over-step fashion.

Data Processing and Reduction

All data were imported into The MotionMonitor software (version 8.0; Innovative Sports Training Inc, Chicago, IL) for processing. A local axis system was embedded into each segment of interest. Each was defined based on a right-hand convention with the positive direction of the x-axis corresponding with the participant's anterior direction, the positive z-axis in the superior direction, and the positive y-axis as a vector oriented at positive 90° rotation from the x-axis about the z-axis. We defined motion of the trunk relative to the global axis system using a Cardan-Euler sequence of X, Y, Z. We defined motion about the hip as the thigh relative to the pelvis and about the knee as the shank relative to the thigh. We calculated hip- and knee-joint angles using a Cardan-Euler sequence of Y, X, Z. The Cardan-Euler sequences of the trunk, hip, and knee all corresponded with a first rotation to define sagittal-plane motion, a second rotation to define frontal-plane motion, and a third rotation to define transverse-plane motion. The difference in Cardan-Euler sequences between trunk and lower extremity kinematics was a result of trunk motion being referenced to the global axis system. During stair descent, each participant was facing and moved in the direction of the positive y-axis of the global axis system. Therefore, sagittal-plane motion of the trunk occurred about the x-axis of the global axis system, frontal-plane motion of the trunk occurred about the y-axis of the global axis system, and transverse-plane motion of the trunk occurred about the z-axis of the global axis system. Motion about the local and global x-axes corresponded to knee valgus (−)/varus (+), hip abduction (−)/adduction (+), and trunk flexion (+)/extension (−), respectively. Motion about the local and global y-axes corresponded to knee flexion (+)/extension (−), hip flexion (−)/extension (+), and trunk lateral flexion toward the stance leg (+)/trunk lateral flexion away from the stance leg (−), respectively. Motion about the local and global z-axes corresponded to knee internal rotation (+)/external rotation (−), hip internal rotation (+)/external rotation (−), and trunk rotation toward the stance leg (−)/trunk rotation away from the stance leg (+), respectively.

Kinematic data were filtered using a fourth-order Butterworth low-pass filter with a cutoff frequency of 12 Hz. Variables of interest included 3-dimensional joint displacements of the trunk, hip, and knee during the stance phase of stair descent, defined as the point of initial contact (first time point at which vertical ground reaction force exceeded 10 N) to toe off (first time point at which vertical ground reaction force dropped below 10 N) for the involved limb. Joint displacements were defined as the differences between the joint angles at initial contact and the peak joint angle in each direction of motion for each plane of motion, except for the sagittal plane, where only flexion displacement was calculated. This allowed us to quantify the amount of displacement that each joint displayed and the direction in which the displacement occurred. This was necessary to avoid producing an inaccurate representation of joint displacement as a result of sign conventions (ie, +, −) used as directional, rather than quantitative, measures when calculating trial and group means. Average joint displacement for each dependent variable was calculated across 3 trials. Although we collected 5 trials of data, we selected the 3 middle trials of the 5-trial sequence for each participant and used the first and last trials only if 1 of the 3 middle trials was not acceptable. All data reduction was performed using a custom MATLAB program (The MathWorks, Inc, Natick, MA).

Statistical Analysis

We used independent-samples t tests (15 total) to compare mean trunk, hip-joint, and knee-joint displacement between the PFP and control groups. Differences in VAS scores between groups before and after stair descent were determined using a 2 × 2 (group = PFP, control; time = pretest, posttest) mixed-model analysis of variance. Post hoc analysis consisted of independent-samples t tests (group) and paired t tests (time) with a Bonferroni correction (0.0125) for any significance interaction or main effect. All statistical analyses were performed using SPSS statistical software (version 18.0; SPSS Inc, Chicago, IL). Statistical significance was established a priori as α ≤ .05.

Kinematics

No differences were seen for displacement of the trunk and hip between groups. However, we did observe a difference in knee internal-rotation displacement, with the PFP group having approximately 4° greater displacement than the control group (Figure 3). This represented a 30% difference between groups. No other differences were noted for displacement of the knee between groups. Statistical information supporting these results can be found in Table 3.

View Figure

• Download as Powerpoint
• Download Figure

Visual Analog Scale Scores

Descriptive statistics for VAS pain scores are presented in Table 4. There was a significant group × time interaction. Post hoc analyses revealed that VAS pain scores were greater in the PFP group at pretest and posttest as compared with the control group. The VAS pain scores increased from pretest to posttest for the PFP group but not for the control group. Statistical information supporting these findings can be found in Table 4.

Table 4.  Pretest and Posttest Visual Analog Scale Scores

View Fullscreen

Trunk Kinematics

We observed no difference in trunk kinematics—specifically trunk flexion and ipsilateral trunk lean—as we initially hypothesized. These are movement patterns we expected to observe in our sample of patients with PFP, as they may increase contact pressure at the patellofemoral joint by increasing sagittal- and frontal-plane moments of the knee.¹⁹ In addition, in the limited research that has examined 3-dimensional trunk kinematics for those with PFP, increased trunk flexion and lateral flexion toward the stance leg compared with healthy controls have been seen.^17,18 Nakagawa et al^17,18 noted increased ipsilateral trunk lean in PFP participants compared with healthy control participants during a single-leg–squat task and a stepping maneuver. A possible explanation for this digression may be that it is the result of a mixed-sex cohort, as evidence has indicated differences in trunk kinematics between sexes.⁴⁰ The differences may also be attributed to the task, as the stepping maneuver and single-leg squat may have required greater demand to be placed on the gluteus medius and a resulting movement compensation using the trunk. The stair descent used in our study may not have been challenging enough to elicit differences between groups, but we chose it because individuals with PFP often complain of pain with stair descent.

Hip Kinematics

Although the results did not support our hypotheses, our finding of no difference in hip adduction and internal rotation between groups agrees with previous research.^21,24,28,29 Similar to our findings, Bolgla et al²⁹ did not observe differences in hip adduction or internal rotation between females with and without PFP during stair descent. Grenholm et al²⁸ also observed no difference in hip adduction between females with and without PFP during a stair-descent task.

We believe the reason for the lack of difference in hip adduction and internal rotation in our study, as in the others, may be that stair descent is not challenging enough to elicit differences between groups. The total number of steps used for these analyses may not be adequate to elicit pain or kinematic alterations such as would be elicited by descending several flights of steps. Souza and Powers²¹ demonstrated this point, as they observed greater peak hip internal rotation for females with PFP when averaged across the 3 tasks of a drop jump, running, and step down but no significant interaction for the tasks. In addition, although the authors did not demonstrate a significant difference for hip adduction during these tasks, those with PFP displayed lateral trunk lean toward the stance leg and attributed this to a compensatory strategy to reduce hip adduction in the presence of hip-abductor weakness.²¹ As discussed earlier, contralateral pelvic elevation with ipsilateral trunk lean is a common compensation for hip-abductor weakness in those with PFP.¹⁹

Knee Kinematics

Our current observation of greater knee internal-rotation displacement during stair descent in those with PFP adds to the current literature, as few authors have examined knee internal rotation in this population.^41,42 To better explain our observations, we performed additional post hoc analyses and noted no difference between groups for knee-rotation angle at initial contact (PFP = −13.72° ± 6.08°, control = −10.54° ± 5.62°; t₃₈ = −1.72; P = .09). We also found no difference in peak knee internal-rotation angle between groups (PFP = −0.92° ± 8.49°, control = −1.68° ± 5.41°; t₃₈ = 0.34; P = .74) but did identify a difference between groups for peak knee external rotation (PFP = −15.21° ± 6.05°, control = −11.48° ± 5.26°; t₃₈ = −2.08; P = .04). Our observed difference in knee internal-rotation displacement may have been attributed in large part to our PFP patients making initial contact in slightly more knee external rotation and achieving a slightly greater peak knee internal-rotation angle. It is important to note that although we did see greater knee internal-rotation displacement, our PFP patients maintained an externally rotated position of their knees during stair descent. Barton et al⁴¹ examined knee internal rotation in a mixed-sex cohort during a walking task and found no difference in knee internal rotation between those with PFP and healthy controls. This suggested that assessing movement displacement may provide information that is not gained by assessing peak kinematic values alone, as we would have made similar observations had we limited our analysis to peak kinematic values.

Knee internal rotation is not typically associated with PFP. Research conducted on cadaver specimens has shown that knee external rotation increases lateral patellar contact pressure, whereas knee internal rotation has little to no effect on medial or lateral patellar contact pressure.^43–45 Increased knee internal-rotation displacement may be a compensatory mechanism to move the knee out of an externally rotated position. After reviewing the data collected for the PFP participants during the screening session, we noted that they had experienced knee pain on average for 4 to 5 years. It is possible that over time these participants began compensating to decrease their knee pain.

Although we found that participants with PFP had a significant increase in VAS scores from before to after stair descent, the change in VAS scores was only 5.4 mm. Research studies assessing pain,^46–51 patient satisfaction,⁵² and sleep quality⁵³ have found that a minimal clinically significant difference in VAS scores is between 9 and 13 mm, with the lowest reported minimal clinically significant difference being 7 mm⁵² and the highest being 30 mm.⁵⁴ Therefore, although the change in VAS scores for the PFP group was statistically significant, it was likely not clinically meaningful. The relatively small change in VAS scores supports knee internal rotation as a compensatory mechanism.

Limitations

Our study is not without limitations. We included only female participants in the study because this population has a higher incidence and prevalence of PFP compared with males.^1,26,27 We did not include women over the age of 35 to reduce the likelihood of osteoarthritic changes within the patellofemoral joint. Furthermore, we did not include adolescents in this study because the causes of PFP within this population may differ from the causes of PFP within an adult population.

Our study was restricted to a single task of stair descent because participants with PFP most often complain of pain with stair descent. The task may not have been challenging enough to reveal altered trunk or lower extremity kinematics, as we only asked participants to descend 4 steps. During initial screening, many participants reported that their knee pain would be greater if they had to descend several flights of stairs as opposed to 3 to 4 steps. We could not add additional steps to the task because of limited laboratory space and safety concerns, but future researchers could use a fatigue protocol to elicit knee pain before a stair-descent task.

Another limitation of our study was the 10-cm VAS pain scale criterion. We found it difficult to locate PFP participants who rated their average pain as at least 3 cm within the past week. We modified the criteria and included individuals who were able to rate their pain as at least 3 cm with at least 2 of the following activities: ascending/descending stairs, hopping/jogging, prolonged sitting with flexed knees, kneeling, or squatting. Because many of the PFP participants had experienced knee pain for several years, it is possible that they had become accustomed to their pain, as they often expressed during the screening, and rated their pain at a low level. The relatively low amount of pain experienced by the PFP participants could also be a result of mild lower extremity dysfunction. It is plausible that the PFP participants did not exhibit a severe enough alteration in lower extremity kinematics to elicit pain, which could explain why we did not see differences in kinematic variables other than knee internal-rotation displacement.

Clinical Significance

We observed no difference in trunk displacement for women with and without PFP during stair descent. Therefore, neuromuscular control of the trunk may not play a role in a low-demand task such as stair descent for the assessment of patients with PFP. Evaluating joint displacement during movement tasks may provide better information about those with PFP, as indicated by our observation of differences in knee internal-rotation displacement.