Are patellofemoral pain and quadriceps femoris muscle torque associated with locomotor function?

Key Words: Gait, Patellofemoral pain, Quadriceps femoris muscle torque.

During the stance phase of gait, the knee is believed to be the principal determinant of limb stability.[1] The quadriceps femoris muscles act as the primary stabilizers of the knee, especially during loading response, when the knee flexion moment is the greatest.[2] Activity of these muscles is necessary to support the flexed knee posture.[3]

Reduced knee flexion during loading response is generally thought to be an action aimed at limiting joint forces and may be indicative of knee pathology.[1] For example, pain and weakness are commonly associated with patellofemoral joint pathology,[4] and the avoidance of knee flexion during stance has been found in persons with patellofemoral joint pathology.[5] Berchuck et al[6] used the term "quadriceps avoidance pattern" for persons with anterior cruciate ligament (ACL) deficiency to describe a gait pattern that minimizes the knee flexion moment during the loading response and therefore the demand of the knee extensors. Persons with PFP also may adopt a similar strategy to reduce the patellofemoral joint reaction forces associated with increased knee flexion and quadriceps femoris muscle activity. A quadriceps femoris muscle avoidance pattern could be deleterious to the patient with PFP, however, if further quadriceps femoris muscle atrophy results from disuse. This avoidance pattern may contribute to patellar instability, which is commonly believed to be at least partly the result of weakened dynamic stabilizers.[4,7,8]

Although gait patterns have been described for various knee pathologies such as degenerative joint disease,[9,10] rheumatoid arthritis,[9-11] and ACL insufficiency,[6] little is known about subjects with PFP and the relationship between pain and weakness. The relationship between knee pain and quadriceps femoris muscle inhibition, however, has been discussed previously in the literature. Reflex inhibition has been demonstrated in subjects with knee pathology[12] and is reported to occur when afferent stimuli from receptors in or around the knee joint result in inhibition of alpha motoneurons in the anterior horn of the spinal cord.[13] A1though in clinical practice pain and inhibition have been associated,[14] decreased motor unit recruitment of the quadriceps femoris muscle appears to be linked to knee joint effusion[12,13,15,16] and has been reported to be independent of pain.[12,13] Young et al[17] reported that afferent block by local anesthesia was not effective in reducing quadriceps femoris muscle inhibition, despite a complete reduction in pain. Additionally, Stratford[12] did not observe a relationship between pain and inhibition that would explain reduced electromyographic activity of the quadriceps femoris muscle during a maximal isometric contraction in persons with acutely effused knees. In contrast, deAndrade and colleagues[15] presented evidence that pain reduction through lidocaine injection reduced quadriceps femoris muscle inhibition in knees that were artificially distended. These observations, however, were made on only four subjects.

Despite the current state of knowledge regarding the cause of quadriceps femoris muscle inhibition, many functionally related questions remain. For example, are compensatory gait patterns a result of pain, weakness, or both. Do gait adaptations associated with patellofemoral joint pathology differ among persons with varying degrees of pain? What is the relationship between PFP and quadriceps femoris muscle weakness? The purpose of our investigation was to determine the influence of PFP and quadriceps femoris muscle weakness on stride characteristics and the amount of knee flexion during the loading response in different gait conditions (level walking, ascending and descending stairs, ascending and descending ramps). Functional assessment scores were also correlated with actual gait characteristics. We hypothesized that there would be a correlation between either pain or quadriceps femoris muscle weakness and the limitations in gait function associated with PFP. This information could assist in identifying variables associated with gait limitations in this population and could aid in guiding treatment programs aimed at improving function.

Method

Subjects

Nineteen female subjects between the ages of 14 and 46 years with a diagnosis of PFP participated in this study (Tab. 1). Subjects were recruited from the Southern California Orthopaedic Institute (Van Nuys, Calif) and were screened to rule out ligamentous instability, internal derangement, and patellar tendinitis. In addition, subjects did not have any other orthopedic or neurologic impairments, as determined by physical examination and questionnaire, that would adversely affect gait. Each subject's pain originated from the patellofemoral joint (as determined through their complaints and a physical examination), and only patients with histories relating to overuse (ie, symptoms related to repetitive activity) or insidious onset were accepted. The physical examination consisted of passive range of motion, active range of motion, palpation of the patella and related structures, and a patellar grind test. In addition, each subject s pain was readily reproducible with at least two of the following activities: stair ascent or descent, squatting, kneeling, prolonged sitting, or isometric quadriceps femoris muscle contraction. The subjects with PFP were varied with respect to the severity and duration of symptoms. Subjects were excluded from the study if they reported having either knee surgery or acute traumatic patellar dislocation.

Table 1. Subject Characteristics

                PFP(a) Group       Comparison Group
                   (n=19)            (n=19)               P
Age (y)
  [bar]X           25.4                 27.5
  SD                8.2                  4.7              3.5
  Range            14-46                23-38
Height (cm)
  [bar]X]         165.1                165.3
  SD                7.6                  7.7               .94
  Range           151.1-177.2          149.9-183.5
Weight (kg)
  bar[X]           62.4                 59.2
  SD                9.3                  7.5               .25
  Range            42.0-82.7            46.8-74.1

(a) PFP=patellofemoral pain.

Nineteen female subjects between the ages of 23 and 38 sears served as a comparison group (Tab. 1). These subjects had no history or diagnosis of knee pathology or trauma, and they were free of any current knee pain. In addition, these subjects did not report discomfort with any of the activities described as criteria for the subjects with PFP, and they had no other limitations that would alter their gait.

Instrumentation

Isometric knee extensor torque was recorded using a Lido dynamometer.(*) Prior to testing, compensation for limb weight and the effects of gravity was made automatically by the dynamometer's computer software program (Version 3.8, 1989). Reliability of the data used for correction was not assessed. Torque data were recorded by a DEC 11/23 computer([dagger]) at a rate of 2,500 Hz. The DEC computer was interfaced with the dynamometer.

Knee pain was recorded using a visual analog pain scale (VAS). The VAS consisted of a 10-cm horizontal line, the ends of which defined the minimum ("no pain") and maximum ("extreme pain") of perceived pain. Each subject placed a mark on the line to indicate the intensity of pain. The amount of pain indicated on the line was convened to a numerical value based on the distance (in centimeters) from the minimal possible pain to the mark on the line. The VAS has been shown by Chesworth et al[18] to be a valid indicator of pain changes in patients with PFP.

To evaluate symptoms and functional limitations in the subjects with PFP, a functional assessment questionnaire (FAQ) developed by Kujala et al[19] was used. The validity and reliability of measurements obtained with the FAQ have not been reported. This questionnaire contained 13 multiple-choice questions relating to patellofemoral joint symptoms. Scoring was based on a numerical scale depending on question response, with some items being weighted more than others. The maximum possible score was 100, which represented no pain and no functional deficits. This scoring system has been demonstrated to differentiate between different classifications of patellofemoral disorders.[19]

Stride characteristics were recorded with a microprocessor-based Foots which Stride Analyzer System.([double dagger]) This system consisted of compression-closing footswitches taped to the soles of the subjects bare feet. The footswitches contained sensors at the heel, the first and fifth metatarsal heads, and the great toe that responded to compressive loads equal to or greater than 3 psi. Stride characteristics calculated from this system included: speed, stride length, cadence, single- and double-limb support times, and stance and swing durations.

Sagittal-plane motion of the ankle, knee, and hip joints was measured with a Vicon motion analysis system.([sections]) Six infrared cameras operating at a 50-Hz sampling rate were used.

A 10-m walkway was used for free- and fast-walking trials. with data being collected over the middle 6 m. Analysis of stair use was done with a four-step staircase with a slope of 33.7 degrees, a step height of 20.3 cm, and a tread depth of 30.0 cm. Ramp walking was assessed with a 12-degree incline that was 6.1 m in length.

Procedure

All data collection was performed at the Pathokinesiology Laboratory, Rancho Los Amigos Medical Center. Downey, Calif. Before testing, all procedures were explained to each subject and informed consent was obtained. Subjects were then asked to complete the FAQ based on their current symptoms and limitations.

Prior to gait analysis, maximal isometric knee extensor torque and knee pain were measured. Subjects were seated on the Lido dynamometer chair with the hips flexed to 90 degrees and the knee flexed to 60 degrees. The axis of rotation of the dynamometer was then positioned in line with the axis of rotation of the knee, with the resistance arm cuff placed just proximal to the malleoli. A Velcro[R] strap([parallel]) was placed across the pelvis to ensure proper stabilization. Sixty degrees of knee flexion was used because this position has been found to result in the greatest torque output in female subjects without impairments.[20]

Isometric torque during a 5-second maximal contraction was then recorded. Verbal encouragement was given to all subjects during the trial. After torque was measured, the subjects with PFP were asked to rate, using the VAS, their knee pain during the maximal contraction. Our rationale for assessing pain during contraction rather than during the locomotor tasks was that we expected that pain scores obtained during ambulation would not reflect true symptoms. We believed that the subjects would most likely adopt gait strategies to reduce or eliminate pain.

Following the torque and pain assessment, subjects were prepared for gait analysis. Footswitches were taped to both of the subjects' bare feet, and the reflective markers that were used to determine sagittal-plane motion were placed at the designated landmarks (posterior heel, fifth metatarsal head, dorsum of the foot, medial and lateral malleoli, anterior tibia, medial and lateral femoral epicondyles, anterior thigh, greater trochanter, bilateral anterior superior iliac spines, and sacrum). One practice trial of both free and fast walking allowed the subjects to become familiar with the instrumentation. For free walking, subjects were instructed to walk at their normal speed. For fast walking, subjects were instructed to walk at a speed as if they were in a hurry. Joint motion and stride characteristics were then assessed simultaneously during free and fast level walking, ascending and descending stairs, and ascending and descending ramps.

Data Analysis

Sagittal-joint motion of the ankle, knee, and hip was calculated for all conditions tested. Raw motion data were filtered at 6 Hz using a fourth-order. Butterworth recursive filter.[21] The data were then digitized and linearly interpolated to 0.01-second internals. The stance phase of each stride of motion collected was normalized to 62% of the gait cycle in order to average data from multiple strides and different subjects. We believe this value to be representative of normal walking,[1] and it was consistent with the average stance phase demonstrated by our subjects for all conditions. Maximum and minimum motion for each joint were analyzed for each phase of the gait cycle Analog signals obtained from the individual footswitch sensors were synchronized with the motion data and were used as event markers to determine the different phases of the gait cycle.

Torque data were integrated at 0.1-second intervals. The torque produced by the limb weight (as determined by the gravity compensation test, was added to the raw torque to account for the effects of gravity. The greatest value over the 5-second trial was recorded for each subject. To control for the effects of subject size, all torque data were normalized by body weight and expressed in newton-meters per kilogram.

The BMDP statistical software(#) was used for all data analyses The data were tested for normality of distribution using the Wilks-Shapiro W statistic. All significance levels were set at P [is lesser than] 0.5.

Subject characteristics (age, height, and weight) were compared between groups using two-sample t tests. Comparison of isometric torque values between groups also was made using a two-sample t test.

To determine whether stride characteristics differed between groups and conditions, a 2 X 6 (group X condition) two-way analysis of variance (ANOVA) for repeated measures on one variable (condition) was performed. This analysis was repeated for each stride characteristic. Data for stride length and cadence during stair ambulation were omitted from the analysis due to the limitation imposed on these variables as a result of the fixed stair height and depth. Peak motion at each joint also was compared between groups and conditions using a 2 X 6 (group X condition) two-way ANOVA for repeated measures. This analysis was repeated for each phase of the gait cycle.

To assess the association among PFP, quadriceps femoris muscle torque, and FAQ score, we used the Pearson product-moment correlation coefficient. We used separate analyses to assess the linear relationship between PFP and torque, PFP and FAQ score, and torque and FAQ score.

Stepwise regression analyses were performed to determine whether any of the independent variables (pain, quadriceps femoris muscle torque, or FAQ score) were predictive of any of the stride characteristics or the amount of knee flexion during loading response (dependent variables). This analysis was performed for the subjects with PFP only and was repeated for all six walking conditions.

Results

Relationship Among Knee Extensor Torque, Pain, and Functional Assessment Score

After normalizing by bodyweight, the maximum knee extensor torque of the PFP group was less than that of the comparison group (2.30 N.m/kg versus 3.04 N.m/kg, P [is less than] .05) (Tab. 2). During the maximal isometric test, the PFP group reported an average pain level of 4.4 out of 10 on the VAS (Tab. 3). The mean score on the FAQ for the PFP group was 67.0 out of a possible 100 (Tab. 3).

Table 2. Maximum Knee Extension Torque (Normalized by Body Weight)

                   PFP(a) Group     Comparison Group     P
Torque (N.m/kg)
  bar[X]            2.35             3.04
  SD                0.78             0.69                .03
  Range             1.28-3.92        1.96-4.02

(a) PFP=patellofemoral pain.

Table 3. Individual Values for Knee Extension Torque, Visual Analog Pain Scale (VAS), and Functional Assessment Questionnaire(19) (FAQ) for Subjects With Patellofemoral Pain

Knee Extension VAS (10= FAQ (100= Subject Torque Maximum Maximum No. (N.m/kg) Pain) Function)

1 2.54 8.6 53 2 2.09 7.6 35 3 1.45 9.6 37 4 3.90 6.5 70 5 1.43 3.4 85 6 2.14 0 73 7 2.74 4.1 84 8 1.78 4.8 73 9 2.21 0.2 73 10 1.23 6.8 38 11 2.73 0.8 73 12 1.79 1.1 75 13 1.23 3.2 62 14 3.16 5.1 68 15 2.42 0 100 16 2.21 8.6 45 17 2.92 6.4 83 18 3.64 3.8 74 19 3.48 3.0 82 X 2.35 4.4 67.5 SD 0.78 3.1 18.1

The VAS pain score was not correlated with knee extensor torque in the PFP group (r=.03). In addition, knee extensor torque was not correlated with the FAQ score (r=.20). The VAS pain score, however, demonstrated a correlation with the FAQ score in the PFP group (r=.72, P [is less than] .001) (Fig. 1).

[Figure 1 ILLUSTRATION OMITTED]

Stride Characteristics

There was a difference between the PFP and comparison groups for walking speed when the data were averaged across all conditions (significant group effect, no interaction) (Fig. 2). Similarly, there was a difference in cadence and in stride length when the data were averaged across all conditions (except data for ascending and descending stairs, which were omitted from the analysis) (Figs. 3, 4). In general, the PFP group demonstrated decreased values for these stride characteristics compared with the other group.

[Figures 2 to 4 ILLUSTRATION OMITTED]

The average walking speed of the PFP group (for all conditions) was 81% of the average walking speed of the comparison group (56.5 m/min versus 69.7 m/min, P [is less than] .001) (Fig. 2). The average stride length of the PFP group across all conditions was 88% of the average stride length of the comparison group (1.22 m versus 1.38 m. P [is less than] .001) (Fig. 3). Cadence of the PFP group was 91% of that of the comparison group when averaged across all conditions (114.1 steps/min versus 125.2 steps/min, P [is less than] .001) (Fig. 4). There were no differences between groups for time spent in single-limb support, double-limb support, swing, and stance.

Of the three variables measured (pain, torque, and FAQ score), knee extensor torque was the only predictor of speed, with higher torque values being associated with higher walking speeds. This association was evident for five of the six conditions (free walking: r=.59. P [is less than] .05; fast walking: r=.59, P [is less than] .05; ascending stairs: r=.50, P [is less than] .05; ascending ramps: r=.62, P [is less than] .05: descending ramps: r=.67, P [is less than] .05) (Tab. 4). Knee extensor torque also was the only predictor of stride length for four of the six conditions (free walking: r=.73. P [is less than] .05; fast walking: r=.61, P [is less than] .05; ascending ramps: r=.62. P [is less than] .05; descending ramps: r=.76, P [is less than] .05) (Tab. 4). No other associations were found between any of the three variables and the remaining stride characteristics.

Table 4.
Stepwise Regression Results for Predicting Walking Speed,
Stride Length, and Cadence

               Stride           Independent Variable
Condition(a)   Characteristic   (Predictor)             r(b)

FR             Walking speed    Knee extension torque   .59
FT             Walking speed    Knee extension torque   .59
AS             Walking speed    Knee extension torque   .50
DS             Walking speed    None                    ...
AR             Walking speed    Knee extension torque   .62
DR             Walking speed    Knee extension torque   .67
FR             Stride length    Knee extension torque   .73
FT             Stride length    Knee extension torque   .61
AR             Stride length    Knee extension torque   .62
AR             Stride length    Knee extension torque   .76
FR             Cadence          None                    ...
FT             Cadence          None                    ...
AR             Cadence          None                    ...
DR             Cadence          None                    ...

(a) FR = free walking, FT = fast walking, AS = ascending stairs, DS = descending ending stairs, AR = ascending ramps, DR = descending ramps Stride length and cadence for ascending and descending stairs omitted due to the limitations imposed by the fixed stair height and depth.

(b) All r values significant at the P<. 05 level. Ellipsis indicates not applicable.

Joint Motion

There was a significant group effect and a significant interaction for ankle joint motion during the terminal stance phase of gait. When conditions were analyzed separately between the two groups, the PEP group demonstrated greater ankle dorsiflexion compared with the other group for fast walking (9.9 [degrees] versus 7.0 [degrees], P[is less than to] .05), descending stairs (27.6 [degrees] versus 18.9 [degrees], P[is less than to] .001), and descending ramps (15.8 [degrees] versus 11.9 [degrees], P[is less than to] .01), (Fig. 5). No other differences for ankle motion were found. There were no differences in knee motion between groups for any phase of the gait cycle, regardless of the condition (Fig. 6). Similarly, hip joint motion was not different between groups, regardless of the condition, for any phase of the gait cycle (Fig. 7).

Pain, quadriceps femoris muscle torque, and FAQ score were not predictors of the amount of knee flexion during loading response. This finding was consistent for all conditions tested.

Discussion

We found a decrease in knee extensor torque in the PFP group (77% of the knee extension torque of the comparison comparison group), as well as an average pain score of 4.4 out of a possible 10 during testing. These associated findings suggest that pain may have played a role in reducing quadriceps femoris muscle torque. When pain was correlated with knee extensor torque, however, this inference did not hold true. These two variables appeared to be completely independent of one another (r = .03). This finding would imply that knee extensor torque was not affected by pain, which is consistent with the observations of Stratford[12] and Young et al.[17]

The lack of an association between knee extensor torque and pain may have been related to numerous factors. For example, patients dealing with persistent pain might tend to protect themselves during an activity in which they would expect to experience pain. Possibly, in order to avoid pain, patients would not produce a maximum torque value that truly reflects their strength. This concept is supported by the fact that 5 of the 19 subjects with PFP reported little or no pain during the maximal isometric quadriceps femoris muscle test. Furthermore, inhibition of quadriceps femoris muscle activity as a result of effusion also could have contributed to the reduction in quadriceps femoris muscle torque. Because we did not assess swelling, we cannot determine whether this really occurred. None of the subjects with PFP, however. demonstrated gross joint effusion.

An alternative explanation for the lack of a correlation between pain and quadriceps femoris muscle torque could be related to the testing position used to elicit knee pain. Our procedure assessed the maximum isometric knee extension torque at 60 degrees of flexion, which placed the quadriceps femoris muscle at its greatest length-tension advantage[20] but ma! have been inadequate in reproducing the amount of patellar pain that would inhibit normal function. Although the high quadriceps femoris muscle forces produced at this l;nee flexion angle also would have resulted in substantial patellofemoral joint reaction forces,[22] the modest pain scores reported by our subjects suggest that this compression was reasonably tolerated. These relatively low pain scores may have been the result of the increase in contact surface area between the patella and femur, which has been reported by Mathews and colleagues[23] to be approximately 40% more at 60 degrees of l;nee flexion as compared with 15 degrees of flexion. Increased contact area would have reduced the joint contact pressure, as the joint forces would have been distributed over a greater area. In addition, because PFP has been linked to patellar subluxation[24] and because patellar subluxation has been shown radiographically to occur at angles of less than 30 degrees of knee flexion,[25] it is possible that testing the subjects with the knee less flexed (ie, 0 [degrees]-30 [degrees]) would have yielded greater pain scores. This position, however, would have placed the quadriceps femoris muscle at a mechanical disadvantage[20] and therefore would have resulted in lower torque values. Given this paradox between testing position and the pain-torque relationship, as well as the need to assess both variables simultaneously for correlation purposes, we believe that it is not surprising that no relationship was found.

An inverse linear association was found between pain and the FAQ score, indicating that the FAQ may be sensitive to individual pain levels. The fact that many of the FAQ questions related to the reproduction of symptoms may explain the correlation between these two scores. Care must be taken in interpreting these results, however, as we did not assess pain during functional activities. Whether the pain associated with the maximal isometric quadriceps femoris muscle contraction has a relationship to the pain that ma! be present during gait is not known at this time and is a limitation that should be addressed in future studies. The lack of an association between quadriceps femoris muscle torque and the FAQ score was not surprising because most of the possible responses to items in the questionnaire pertained primarily to pain during functional activities.

We did not find a reduction in knee flexion during the loading response in the PFP group, indicating that these subjects did not alter the normal knee joint kinematics during early stance. This finding is contrary to the conclusions of Dillon and colleagues, [5]who reported that subjects with PFP reduce knee flexion during the stance phase to minimize the patellofemoral joint reaction force. Our kinematic data indicate that this gait adaptation cannot be generalized to persons with PFP. Our findings also suggest that quadriceps femoris muscle torque in the PFP group, although reduced. was capable of providing stability during this phase of the gait cycle.

The primary gait adaptation in the PFP group was a reduction in walking speed, which was consistent across all conditions. The greatest differences between groups occurred during the more vigorous tasks of fast walking and ascending ramps, which suggests that the higher-demand activities required greater speed attenuation. Winter[26] has demonstrated that a slower gait speed reduces the demand of the quadriceps femoris muscle during initial stance by decreasing the flexion moment. The reduction of the knee flexion moment during slower walking is most likely the result of the reduced vertical component of the ground reaction force, which is the predominant external force contributing to the knee flexion moment. The influence of walking speed on the magnitude of the vertical ground reaction force has been demonstrated by Powers et al,[27] who found a linear relationship between these two variables. Therefore, a decrease in walking speed could allow for a reduction of muscular demand, without a compromise in l;nee kinematics, and is concordant with previous findings of decreased electromyographic activity of the vastus muscles of subjects with PFP.[28]

Although it would appear that persons with PFP may adopt a slower gait speed as a possible way of reducing the patellofemoral joint reaction force, there was no relationship between the amount of knee pain and walking speed for an; of the conditions. There was a relationship, however, between quadriceps femoris muscle torque and walking speed for five of the six conditions (descending stairs excepted), with increased quadriceps femoris muscle torques resulting in faster walking speeds. This association suggests that persons with greater levels of quadriceps femoris muscle torque tend to demonstrate greater ambulation speeds, which may be related to the higher quadriceps femoris muscle demand associated with accelerated speed.

The reduction in gait speed in the PFP group was a function of reduced stride length and cadence, both of which were less in the PFP group than in the comparison group in all conditions. The tendency toward decreased terminal swing hip flexion in the PFP group contributed to this decreased stride length by limiting the forward position of the limb at initial contact. As with walking speed, quadriceps femoris muscle torque was the only predictor of stride length in four of the six conditions tested, further supporting the relationship between quadriceps femoris muscle torque and stride variables.

Conclusion

The results of our stud!- have potential clinical implications. Conservative care for individuals with PFP typically involves both pain management and strengthening of the extensor mechanism.[4,7,14] The fact that greater isometric quadriceps femoris muscle torque was associated with increased walking speed and stride length suggests that strength of this muscle group may be an important factor in determining the gait characteristics of persons with PFP. Quadriceps femoris muscle strengthening, therefore, ma! be useful for improving functional ability, a clinical practice already used for persons with PFP. Quadriceps femoris muscle strengthening may be particularly important for individuals who want to return to higher-demand activities such as running or other athletic activities.

[Figures 1 to 7 ILLUSTRATION OMITTED]

(*) Loredan Biomedical Inc, 1632 Da Vinci Ct, PO Box 1154, Davis, CA 95617.

([dagger]) Digital Equipment Corp, 146 Main St, Maynard, MA 01754.

([double dagger]) B&L Engineering, 8807 Pioneer Blvd, Suite C, Santa Fe Springs, CA 90670.

([sections]) Oxford Metrics Ltd. Unit 14, 7 West Way, Botley, Oxford, England OX2 0JB.

([parallel]) Velcro USA Inc, PO Box 5218, 406 Brown Ave, Manchester, NH 03108.

(#) BMDP Statistical Software Inc, 1440 Sepulveda Blvd, Suite 316, Los Angeles, CA 90025.

References

[10] Stauffer RN, Chao EYS, Gyory AN. Biomechanical gait analysis of the diseased knee joint. Clin Orthop. 1977;126:246-253.

[11] Kettlekamp DB, Leaverton PE, Misol S. Gait characteristics of the rheumatoid knee. Arch Surg. 1972;104:30-34.

[12] Stratford PW. Electromyography  of the quadriceps femoris muscles in subjects with normal knees and acutely effused knees Phys Ther. 1981;62:279-283.

[13] Stokes M. Young A. Investigations of quadriceps inhibition: implications for clinical practice. Physiotherapy. 1984;425-428.

[14] McConnell J. The management of chondromalacia patellae: a long-term solution. A Australian Journal of Physiotherapy. 1986;32:215-223.

[15] deAndrade JR, Grant C, Dixon A. Joint distension and reflex muscle inhibition in the knee. J Bone Joint Surg Am. 1965;47:313-322.

[16] Spencer JD, Haves KC, Alexander IJ. Knee joint effusion and quadriceps reflex inhibition in man Arch Phys Med Rehabil. 1984;65:171-177.

[17] Young A, Stokes M, Shakespeare DT, Sherman KP. The effect of intra-articular bupivicaine on quadriceps inhibition after meniscectomy. Med Sci Sports Exerc. 1983:15:154. Abstract.

[18] Chesworth BM, Culham EG, Tata GE, Peat M. Validation of outcome measures in patients with patellofemoral syndrome. J Orthop Sports Phys Ther. 1989;10:302-308.

[19] Kujala UM, Jaakkola LH, Koskinen SK, et al. Scoring of patellofemoral disorders. Journal of Arthroscopy and Related Surgery 1993:9:159-163.

[20] Lieb IJ. Perry J. Quadriceps function: an electromyographic study under isometric conditions. J Bone Joint Surg Am. 1971;53:749-758.

[21] Winter DA. Biomechanics and Motor Control of Human Movement. New York, NY John Wiley & Sons Inc; 1990.

[22] Maquet PG. Biomechanics of the Knee. 2nd ed. New York. NY Springer-Verlag New York Inc: 1984.

[23] Mathews LS, Sonstegard DA, Henke JA. Load-bearing characteristics of the patellofemoral joint. Acta Orthop Scand. 1977;48:511-516.

[24] Heywood WB. Recurrent dislocation of the patella. J Bone Joint Surg Br. 1961;43:508-517.

[25] Brossmann J, Muhle C, Schroder C, et al. Patellar tracking patterns during active and passive knee extension: evaluation with motion-triggered cine MR imaging. Radiology 1993;187:205-212.

[26] Winter DA. Kinematic and kinetic patterns in human gait: variability and compensating effects. Human Movement Science. 1984;3:51-76.

[27] Powers CM, Rao S, Perry J. Loading characteristics in subjects with patellofemoral pain Gait and Posture. 1995;3:84. Abstract.

[28] Powers CM, Landel R, Perry J. Timing and intensity of vastus muscle activity during functional activities in subjects with and without patellofemoral pain. Phys Ther. 1996;76:946-955.

CM Powers, PhD, PT, is Assistant Professor, Department of Biokinesiology and Physical Therapy, University of Southern California, 1540 E Alcazar St, CHP 155, Los Angeles, CA 90033 (USA) ([e-mail]). Address all correspondence to Dr Powers.

J Perry, MD, is Chief, Pathokinesiology Service, Racho Los Amigos Medical Center, Downey, Calif, and Professor, Department of Biokinesiology and Physical Therapy, University of Southern California, Los Angeles.

A Hsu, PhD, PT, is Assistant Professor, Department of Biokinesiology and Physical Therapy, University of Southern California, Los Angeles.

HJ Hislop, PhD, PT, FAPTA, is Professor and Chair, Department of Biokinesiology and Physical Therapy, University of Southern California, Los Angeles.

This study was approved for human subjects by the Los Amigos Research and Education Institute Inc of Rancho Los Amigos Medical Center.

This study was supported in part by a grant from the Foundation for Physical Therapy Inc.

This article was submitted August 8, 1996, and was accepted April 25, 1997.