Introduction

Generalized anterior knee pain (AKP) or patellofemoral

joint pain is one of the most common afflictions of an ac-
tive population, especially among women [1, 9]. Causes
of AKP can be insidious or have a traumatic onset. The
goal of most therapeutic interventions is to restore normal
lower extremity function without exacerbating knee pain.
Treatment options for insidious onset of generalized AKP
may include selective strengthening of the vastus medi-
alis, patellar taping, illiotibial band stretching, use of foot
orthotics, patellar mobilizations, bracing, or a combina-
tion of any of the above [7, 12, 16, 26, 27]. If the goal is
to strengthen the lower extremity without exacerbating
knee pain, functional rehabilitation using closed kinetic
chain activity may be the intervention of choice [4, 8, 12].
Fluid, pain-free movement is the result of synchronous
interconnected body segments. If one segment of the ki-
netic chain is not functioning properly, referred pain or
Abstract Ascending and descending
stairs is a provocative activity for an-
terior knee pain (AKP) patients. The
gluteus medius (GM) acts on the
lower extremity in the frontal plane
and can affect forces at the knee. De-
termining activation patterns of the
GM in patients with AKP can help
identify efficacy of training the GM
in this population. This study exam-
ined electromyographic (EMG) firing
patterns in lower extremity muscles
in subjects with AKP while ascending
and descending stairs. Subjects in the
AKP group (n=16) demonstrated gen-
eral AKP for at least 2 months com-
pared to the control group (n=12);
neither group had any history of knee
trauma. Subjects were instrumented
with EMG electrodes on the vastus
medialis oblique (VMO), vastus later-
alis (VL), and GM. Retroreflective
markers were placed on lower ex-
tremities to determine knee flexion
angle, and frontal plane pelvis orien-
tation at toe contact. Subjects then
performed a series of five stair
(height=18 cm) ascent and descent tri-
als. Repeated measures analyses of
variance were performed on EMG
and kinematic variables, between the
two groups and between the symp-
tomatic and asymptomatic sides. In
the AKP group the GM demonstrated
delayed onset and shorter durations
for stair ascent and shorter duration
during descent. There were no signifi-
cant differences between sides in the
AKP group. Consistent with previous
studies, subjects in the AKP group
demonstrated no difference in the
VMO onsets relative to VL onsets
compared to the control group.
Changes in neuromuscular activity
patterns may be a result of a compen-
sations strategy due to AKP. Training
of GM and other hip muscles is war-
ranted during rehabilitation of AKP
patients.
Keywords Kinetic chain
Electromyography Temporal
patterns Patellofemoral pain
Kinematics
KNEE
Knee Surg Sports Traumatol Arthrosc
(2003) 11 : 244251
DOI 10.1007/s00167-003-0353-z
Timothy J. Brindle
Carl Mattacola
Jean McCrory
Electromyographic changes
in the gluteus medius
during stair ascent and descent
in subjects with anterior knee pain
Received: 1 July 2002
Accepted: 7 January 2003
Published online: 15 April 2003
Springer-Verlag 2003
T. J. Brindle ()
Physical Disabilities Branch,
Rehabilitation Medicine Department,
Warren Grant Magnuson Clinical Center,
National Institute of Child Health
and Human Development,
National Institutes of Health,
Bethesda, MD 20892-1604, USA
Tel.: +1-301-4514927,
Fax: +1-301-4809896,
e-mail: Tbrindle@cc.nih.gov
C. Mattacola
Division of Athletic Training,
University of Kentucky,
Lexington, KY 40506, USA
J. McCrory
Neuromuscular Research Laboratory,
University of Pittsburgh,
Pittsburgh, PA 15203, USA
dysfunction may result at a location remote from the pain
site. Grelsamer and McConnell [7] report that dysfunction
of the gluteus medius (GM) can result in contralateral
dropping of the pelvis or increased internal rotation at the
hip. Increased internal rotation of the hip can contribute to
a greater valgus force vector at the knee, thereby con-
tributing to patellofemoral joint pain [7]. In the lower ex-
tremity kinetic chain hip muscle imbalances have been
demonstrated to play a role in low back pain, but there has
not been any reported evidence of these hip muscle im-
balances contributing to knee pain [5].
Anecdotal clinical reports stress the importance of
strengthening the hip musculature in order to stabilize the
hip and avoid increased valgus load at the knee. Hip
strengthening is generally performed non-weight bearing;
however, the automatic carryover of improved hip muscle
strength to functional activities has not been proven defin-
itively. Moreover, Grelsamer and McConnell [7] stress the
importance of training the GM while the patient is weight
bearing in order to control the frontal orientation of the
pelvis and concomitant internal hip rotation. Diminished
capacity of the GM, which stabilizes the pelvis in the
frontal plane, may cause dropping of the contralateral side
of the pelvis causing a Trendelenberg sign [10]. To our
knowledge, there is no clinical literature which character-
izes muscle activity at the hip during lower leg functional
activities.
Exacerbation of pain while ascending and descending
stairs is one common clinical symptom of AKP; descend-
ing stairs tends to be more painful than ascending stairs
[11, 16]. Grelsamer and McConnell suggest [7] utilizing a
mirror for visual feedback of the pelvis with shorter step
heights, so that patients can maintain a level pelvis while
ascending and descending these low height stairs. As sub-
jects demonstrate adequate pelvic control, step height is
increased and visual feedback is gradually removed. Clin-
ically this treatment method appears to have positive re-
sults. However, there are no studies to evaluate hip mus-
cle activity during the execution of this functional task in
patients with AKP.
Grelsamer and McConnell [7] also utilize patellar tap-
ing to reduce stress on the lateral patellar fascet and/or to
influence VMO muscle activity. They postulate that aber-
rant VMO activity could be responsible for AKP. How-
ever, Sheehy et al. [24] demonstrate no significant differ-
ences between VMO and vastus lateralis (VL) onset times
in subjects with AKP compared to control subjects while
ascending stairs. Given the lack of difference between
VMO and VL electromyographic (EMG) activation pat-
terns and the influence GM muscle activity has on lower
extremity function, we chose to study muscle activity pat-
terns of subjects with AKP during stair ascent and de-
scent.
The purpose of this study was to identify differences in
muscle (EMG) activity patterns between subjects with
AKP and age-matched control subjects. Of particular in-
terest are the firing patterns of the GM relative to the
VMO and the VL during stair ascent and descent. Identi-
fication of these temporal differences may help clinicians
isolate abnormalities that may be unique to individuals
with AKP. Such abnormalities may give rise to rehabilita-
tion options that can be used in the treatment of AKP pa-
tients. In lieu of using terminology such as patellofemoral
pain we use a broad definition of generalized AKP be-
cause we cannot substantiate true patellofemoral pathol-
ogy. In doing so we acknowledge that some of our sub-
jects could suffer from AKP that does not originate from
the patellofemoral joint.
Materials and methods
Subjects
All subjects were between the ages of 18 and 35 years and re-
cruited from the University of Kentucky Sports Medicine program,
the general student population at the University of Kentucky, and
the local community. Exclusion criterion for all subjects included
history or trauma to the knee, with any indication of ligamentous
or cartilage damage by either report or physical examination. The
AKP group consisted of 16 subjects (12 women, 4 men) who had
had generalized anterior knee pain for at least 2 months which was
exacerbated by ascending or descending stairs, running, or pro-
longed sitting. Six of the 16 subjects in the AKP group presented
with bilateral anterior knee pain. Twelve age-matched subjects
(7 women, 5 men) served as controls. Each subject provided a brief
history to rule out possible ligament, cartilage, and muscle pathol-
ogy. Subjects provided written informed consent. Subjects in the
AKP group also filled out visual analog to rate their pain from 0 to
10 prior to the task of stair ascent and descent.
Protocol
This study was performed in two phases. The initial phase was per-
formed on the control group to determine whether there was a dif-
ference in the temporal activation patterns of the GM, VMO, and
VL in the lower extremity. Each control subjects right lower ex-
tremity was instrumented to measure sagittal knee plane motion,
and EMG onset and durations for the VMO, VL, and GM. Once
this phase was completed and demonstrable differences were found
between GM vs. VMO and GM vs. VL activity, we expanded our
protocol to include subjects with AKP. For the AKP subjects we
instrumented (EMG electrodes and retroreflective markers) both
lower extremities in the AKP group and compared differences be-
tween symptomatic and asymptomatic extremities. We also placed
markers on the pelvis (described below) to measure frontal plane
pelvis orientation and added a measure of EMG magnitude to com-
pare muscle activity between symptomatic and asymptomatic
sides.
Instrumentation
Prior to electrode placement the skin was cleaned with isopropyl
alcohol. Subjects in both groups were instrumented with silver-sil-
ver chloride EMG surface electrodes (Therapeutics Unlimited,
Iowa City, Iowa, USA) over the following muscles: VMO, VL,
and the GM (Fig. 1). Proper electrode placement was confirmed by
observing the electrical signal on an oscilloscope during common
manual muscle testing procedures [10]. Retro-reflective markers
245
were placed bilaterally on the AKP subjects and the right lower ex-
tremity of the control subjects on the toe, heel, lateral mallelous of
the ankle, lateral condyle of the knee, and greater trochanter. These
retroreflective markers determined toe contact with the stair and
knee flexion angle at toe contact. For subjects in the AKP group,
retroreflective markers were also placed on bilateral anterior supe-
rior iliac spines and on the sacrum [22]. This enabled the calcula-
tion of frontal plane pelvis orientation. All retroreflective markers
were placed over prominent bony landmarks of the lower extrem-
ity that are easily identifiable and reproducible, where subcuta-
neous tissue is thin, to minimize movement artifact. Once fully in-
strumented, subjects sat quietly in order to record a quiet file of
EMG that was used as a reference to determine muscle onsets and
muscle duration. The AKP subjects also performed 1-s periods of
maximum voluntary isometric contractions of the VMO, VL, and
GM for a reference value of peak muscle activation magnitude.
Experimental procedure
Subjects were instructed to ascend stairs at a normal pace after the
command, ready, set, go. A manual switch initiated the collec-
tion of kinematic and EMG data just prior to the beginning of
movement. The stairs were custom made with a height of 18 cm
and depth of 22 cm and did not include handrails. Each AKP sub-
ject was instructed to start with his/her symptomatic leg; control
subjects started with the right leg (Fig. 1a). Subjects were told to
stop at the top of the stairs and wait for instructions to descend.
Subjects were instructed to descend the stairs following the same
verbal command. Subjects with AKP stepped down first with their
asymptomatic leg so that the symptomatic extremity would contact
the final step prior to the floor (Fig. 1b). Subjects in the control
group stepped down first with their left leg so the right foot would
contact the final step prior to the floor. Only data for the first step
up and the last step down were analyzed (Fig. 1). Subjects per-
formed five trials in each direction. Following the collection of
five trials in each direction (ascent and descent), subjects in the
AKP group were asked to rate their knee pain with a visual analog
scale during both ascent and descent of the stairs.
Data collection and analysis
Kinematic data were recorded by six high-speed Falcon Cameras
(Motion Analysis, Santa Rosa, Calif., USA), with a sampling rate
of 60 Hz, and processed with a 4th order, zero-lag, low-pass But-
terworth filter with a cutoff frequency of 6 Hz. Spatial coordinates
from the retroreflective markers were combined to represent the
knee angle in both groups and frontal plane pelvis orientation in
the AKP group. Calculation of pelvis orientation in the frontal
plane has been proven reliable at self-selected speeds [25]. Three-
dimensional coordinate data were computed using direct linear
transformation as modified by Motion Analysis Corporation Soft-
ware (Motion Analysis). Velocity was calculated from the first de-
rivative of the toe marker [13]. Kinematic data were synchronized
with the EMG data (960 Hz) and saved in 3-s epochs to ensure
capture of the complete task of stair ascent or descent. Kinematic
data were used to determine the timing of toe contact with the stair,
knee flexion angle, and frontal plane pelvis orientation at toe con-
tact.
Toe contact with the first stair was defined to be the moment
when the velocity of the toe was zero (Fig. 2). A two-dimensional
sagittal model of the shank (lateral mallelous to lateral condyle of
knee) and thigh (lateral condyle of knee to greater trochanter) was
constructed. The knee angle was computed as the difference be-
tween the thigh and shank at when the toe contacts the step. The
frontal plane pelvis orientation angle was computed from the angle
between the thigh segment (lateral condyle of knee to greater
trochanter) and pelvis segment (ipsilateral to contralateral anterior
superior iliac spines) when the toe contacts the step. These angles
are created from the retroreflective markers and are approxima-
tions of joint angles.
The EMG data were amplified (100) and were band-pass fil-
tered (101000 Hz) during collection. The EMG data then under-
went an analog to digital conversion (16-bit) and were stored on a
246
Fig. 1 a Toe contact with first step while ascending stairs. b Toe
contact with last step while descending stairs. Arrows Toe contact
with stair as a reference for kinematic and EMG analysis. EMG
Electrodes not seen, but position of electrode is indicated (X) for
VMO in a, for VL in b, and for GM in b
PC. Analysis to determine onset of muscle activity was performed
by a custom program in Matlab (Mathworks, Natick, Mass., USA).
All EMG files were rectified and were filtered using a 4th order,
low-pass Butterworth filter with a cutoff frequency of 15 Hz to cre-
ate a linear envelope. Onset of muscle activity was determined
from the quiet file as follows: The threshold voltage (Vo) re-
quired for muscle onset was calculated from five standard devia-
tions above the resting mean of the quiet file. The onset of the
muscle activity was determined by comparing discrete data point
values (Vi) in a point-by-point fashion to the threshold voltage.
When the mean voltage of a 25-ms window of data immediately
adjacent to the point (50 ms total) exceeded Vo, the initial data
point value was considered to represent the onset of muscle activ-
ity. Termination of muscle activity was defined when the second
discrete data point value (Vi
2
) of a 25-ms window of data immedi-
ately adjacent to the point (50 ms total) dropped below Vo. The du-
ration of the muscle activity was then determined to be the time
from Vi to Vi
2
. The magnitude of muscle activity was computed
from the area under the linear envelope, or integrated EMG, for the
duration of muscle activity. Linear envelopes are both reliable and
commonly used to determine magnitude of muscle activity during
functional tasks [15, 28]. The area under the linear envelope was
calculated using a trapezoidal estimation technique. This magni-
tude of EMG activity during stair ascent and descent was calcu-
lated as a percentage of the maximum voluntary isometric contrac-
tion.
Statistical analysis
In order to determine temporal differences in onset and duration
means during ascent and descent a one-way analysis of variance
with repeated measures was performed on data for the control
group for mean onsets and durations of the EMG activity from the
GM, VMO, and VL, and mean knee flexion angle at the point of
toe contact. Next a 223 analysis of variance with repeated mea-
sures was performed to determine whether there were differences
in mean EMG firing patterns between the control and AKP groups.
Data from subjects symptomatic lower extremity, for unilateral
symptoms were combined with data from the right lower extrem-
ity (side-matched to the control group) from subjects who demon-
strated bilateral symptoms. The between-subject factor was group
(control, AKP) and the within-subject factors were condition (stair
ascent, stair descent) and muscle (VMO, VL, GM). The dependent
variables for this analysis were mean EMG onset and duration for
the VMO, VL, and GM, and knee flexion angle. The right lower
extremity was used or side-matched for the six subjects who pre-
sented with bilateral knee pain in the AKP group.
Last, a 223 analysis of variance with repeated measures was
used to compare differences between the symptomatic and asymp-
tomatic extremity on the ten subjects with unilateral symptoms
within the AKP group. The factors for this analysis include side
(symptomatic, asymptomatic), condition (stair ascent, stair de-
scent), and muscle (VMO, VL, GM). The dependent variables for
this analysis were mean EMG onset and duration for the VMO,
VL, and GM, and mean EMG magnitude for all three muscles,
mean knee flexion, and frontal plane pelvis orientation. Post-hoc
pairwise comparisons were performed to determine differences
among the three muscles during stair ascent or descent. Signifi-
cance was set at a P<0.05 probability level for all the above analy-
ses, with appropriate adjustments to during post-hoc pairwise
comparisons.
Results
Difference between the GM and VMO/VL activity
in control subjects
When the control subjects ascended the stairs, the GM was
activated later, or closer to the point at which the toe con-
tacted the stair, than the VMO (F
1,11
=11.79, P=0.004) and
VL (F
1,11
=19.50, P=0.002). The GM duration was signifi-
cantly shorter than the durations of the VMO (F
1,11
=6.59,
P=0.02) and VL (F
1,11
=11.80, P=0.004) when subjects as-
cended the stairs (Table 1). When the control subjects de-
scended the stairs the onset of the GM was similarly acti-
vated later than the VMO (F
1,11
=9.31, P=0.013) and the
VL (F
1,11
=9.48, P=0.015). The duration of GM EMG was
significantly shorter than the durations of the VMO (F
1,11
=
9.65, P=0.011) and the VL (F
1,11
=11.9, P=0.005) when
subjects descended the stairs (Table 2). In general, for
control subjects the GM was activated after the VMO and
VL and for a shorter duration than the VMO and VL for
both ascent and descent of stairs. Establishing these tem-
poral patterns for the asymptomatic control group would
permit comparisons between the control subjects and sub-
jects with AKP. In addition the control subjects also demon-
strated knee flexion angles of 69.2(8.8) and 27.6(9.1)
at toe contact during stair ascent and descent, respectively
(Table 3).
247
Fig. 2 Example of temporal relationship between toe kinematic
and EMG data. Zero velocity of toe marker following swing phase
indicates toe contact with step and serves as reference point for
EMG analysis. Onset is prior to toe contact and is indicated by
negative values. EMG duration is between EMG onset and offset.
Calculation of linear envelope is area under the curve during the
duration of the EMG activity
Control group vs. anterior knee pain group
On a scale from 0 (no pain) to 10, subjects in the AKP group
reported, with visual analog scales, minimal amount of
discomfort at rest (111.3) prior to performing the stair
ascent and descent task. The AKP group reported moder-
ate knee pain for both stair ascent (4.4 3.0) and descent
(4.25 1.8). The control group did not report any discom-
fort with either stair ascent or descent. These AKP subjects
demonstrated later onset of GM EMG during stair ascent
(F
1,26
=4.96, P=0.035) and a shorter duration of activity
(F
1,26
=5.15, P=0.032) than control subjects (Table 1). When
subjects in the AKP group descended the stairs, the
GM EMG was significantly shorter (F
1,26
=4.24, P=0.049)
than the control group. Compared to the control group dur-
ing stair descent the AKP group also demonstrated shorter
EMG duration of VMO (F
1,26
=5.88, P=0.023) and VL
(F
1,26
=4.42, P=0.045). There was no significant difference
between the two groups with regard to knee flexion angles
at toe contact (Table 3).
Symptomatic vs. asymptomatic comparison
within the AKP group
Within the AKP group there were six subjects with bilat-
eral AKP. There were no significant differences (P>0.05)
between muscle activation patterns of the asymptomatic
and symptomatic lower extremities of the AKP group.
There were also no differences in knee flexion angle and
frontal plane pelvis orientation at toe contact. This could
be due to a small number of subjects in this subset.
Discussion
Previous studies of subjects with patellofemoral pain and
AKP have demonstrated no differences compared to nor-
mal subjects between the VMO and the VL onsets times
in stretch reflexes, during isometric exercise, or during
stair ascent or descent [14, 17, 24]. Boucher et al. [2] also
demonstrated no significant difference in magnitude of
EMG activity between the VMO and VL during isometric
knee extension activity. The VMO and VL activity can af-
fect patellar dynamics predominantly from 0 to 40 of
248
Table 1 EMG activation pat-
terns while ascending stairs
(AKP anterior knee pain, GM
gluteus medius muscle, VMO
vastus medialis oblique mus-
cle, VL vastus lateralis muscle,
MVIC maximum voluntary iso-
metric contractions)
*P<0.05 GM vs. VMO and
GM vs. VL within the control
group
**P<0.05 AKP group vs. con-
trols
VMO VL GM
Onset (ms)
Controls 204.8193.0* 191.152.0* 182.1110.5*
,
**
AKP 167.9136.8 150.4116.9 88.1110.3
Duration (ms)
Controls 724.2131.7* 1461.5213.0* 758.8115.7*
,
**
AKP 757.7139.2 762.1110.1 608.1206.4*
Linear envelope (%MVIC), AKP
Involved 1.361.36 0.6340.67 0.870.84
Uninvolved 1.161.51 1.732.66 0.930.61
Table 2 EMG activation pat-
terns while descending stairs
(AKP anterior knee pain, GM
gluteus medius muscle, VMO
vastus medialis oblique mus-
cle, VL vastus lateralis muscle,
MVIC maximum voluntary iso-
metric contractions)
*P<0.05 GM vs. VMO and
GM vs. VL within the control
group
**P<0.05 AKP group vs. con-
trols
VMO VL GM
Onset (ms)
Controls 366.969.2* 394.881.8* 333.9144.9*
AKP 289.5177.7 349.7234.1 238.3202.4
Duration (ms)
Controls 913.6154.2*
,
** 913.2121.0*
,
** 712.5314.4*
,
**
AKP 777.1143.5** 792.7168.5** 494.2246.5**
Linear envelope (%MVIC), AKP
Involved 1.160.998 0.4950.56 0.5130.4
Uninvolved 3.51.1 1.211.9 0.5190.46
Table 3 Knee flexion angle and frontal pelvis orientation at toe
contact (AKP anterior knee pain); no statistical significance be-
tween control and AKP groups or between symptomatic and
asymptomatic extremities in the AKP group
Upstairs Downstairs
Knee flexion angle (
o
)
Controls 69.28.8 27.69.1
AKP
Symptomatic 70.04.9 26.35.9
Asymptomatic 71.64.9 29.44.0
Pelvis orientation (
o
)
AKP
Symptomatic 95.26.2 84.93.0
Asymptomatic 97.02.8 85.64.7
flexion; beyond 40 of knee flexion the shape of the inter-
condylar groove dictates patellar position [4, 20, 21].
Our findings are consistent with those of previous stud-
ies, which found no significant difference between VMO
and VL activity in healthy subjects during stair ascent and
descent [23, 24]. None of the prior studies evaluated the
function of lower extremity muscles other than the VMO
and VL during activities. Our study is the first to evaluate
hip muscle EMG activity in conjunction with quadriceps
activity during lower extremity tasks, making it hard to
compare GM EMG activity with previous investigations,
involving subjects with AKP. Only one study to date has
demonstrated statistically significant differences between
onset times of the VMO and VL while subjects ascended
and descended stairs [3]. In that study the VL preceded
the VMO onset by 15.8 and 19.4 ms while subjects as-
cended and descended the stairs, respectively. Statistical
significance was achieved with both 33 subjects in the AKP
pain and control group. With such small differences be-
tween VMO and VL onset might be significant statisti-
cally but not clinically significant.
Nadeau et al. [19] suggested a hip strategy for patients
with AKP while walking in which an increased hip exten-
sor moment would compensate for a decreased knee ex-
tensor moment. They demonstrated a decreased knee flex-
ion angle within the first 20% of the gait cycle, but knee
moments remained unchanged. They believed that a more
challenging task would be a better way to examine this
lower extremity function when subjects have AKP. Sal-
sich et al. [22] report a decreased knee extensor moment
for both stair ascent and descent in subjects with AKP.
However, no concomitant increase in hip extensor mo-
ment was observed. This would support the hip-strategy
proposed by Nadeau et al. [19] for patients with AKP.
Salisch et al. [22] also reported no major change in me-
dial-lateral (frontal plane) moments during stair ascent or
descent. Nadeau et al. [19] demonstrated diminished me-
dial-lateral hip moments during level walking in subjects
with AKP pain, compared to control subjects but statisti-
cal analysis was not reported. Because the GM is able to
control frontal plane moments at the hip, it is likely the
GM is part of a compensation strategy in patients with
AKP. In our study the GM demonstrated altered activity
patterns during both stair ascent and descent. Altered
EMG activity of the GM can affect the dynamics of the
rest of the lower extremity [7]. This suggests the need for
further study of the role of the GM in compensatory strate-
gies during functional tasks.
Duration of EMG activity was shorter in the VMO, VL,
and GM for AKP subjects as they descended the stairs.
Subjects with AKP have typically demonstrated slower
cadences during ambulation and stair ascent and descent
[19, 22]. While we did not control or measure movement
speed, slower velocities tend to generate longer periods of
muscle activation, not the shorter EMG durations we ob-
served during stair descent. Moreover, altered EMG dura-
tions were not observed while subjects ascended the
stairs. In particular the GM is the only muscle that demon-
strated a consistent difference between the AKP group and
the control group. There are certainly other muscles in the
region of the hip that can influence lower extremity kine-
matics; regardless, GM muscle activity appears to be al-
tered in the presence of AKP. We agree with others who
believe that the GM does play a role in AKP because its
action can produce forces indirectly across the knee [7].
McConnell [18] and Gilleard et al. [6] have proposed
both taping the patella and training hip muscles to allevi-
ate AKP pain during activities. Taping or pulling the patella
medially is thought in the short term to off-load the lateral
patella fascet and provide long-term changes in muscle
activation patterns that could decrease pain during routine
activities [6, 18]. In a recent study subjects with AKP
demonstrated no significant changes in magnitude of VL
activity when the patella was taped during either ascent or
descent of stairs [23]. Salsich et al. [23] demonstrated in-
creased knee extensor moment but a decrease in knee pain
in subjects with AKP while the patella is taped. Increased
knee extensor moments could increase patellofemoral com-
pression exacerbating pain, but pain was diminished in
their study. Hip moments and EMG activity, other than the
VL, were not evaluated in the study by Salsich et al. This
is the first report of hip muscle activity in subjects with
AKP while doing functional activity.
This study required subjects to ascend and descend
stairs, which is both common and provocative in terms of
exacerbating AKP. Subjects in the AKP group demon-
strated symptoms similar to patellofemoral pain for at least
2 months, differentiating them from the control group.
A brief history was used to exclude other joint patholo-
gies, without diagnostic imaging studies such as radiogra-
phy and magnetic resonance imaging, we cannot assert that
there were no subjects with subtle anatomical disruption in
the knee joint of the AKP group. More important is the
nebulous nature of general AKP lack of homogeneity within
the AKP group that could confound study of this patient
population.
The kinematic measures (knee flexion angle and frontal
pelvis orientation) were used to validate our data collec-
tion methods and provide an opportunity to determine
whether patients with AKP demonstrated kinematic dif-
ferences at toe contact. In addition to the inherent vari-
ability in retroreflective marker placement on anatomical
landmarks, the kinematic data used in this study was not
comprehensive enough to identify compensation strate-
gies. In our study we identified only sagittal knee joint an-
gle and pelvis position in the frontal plane at toe strike.
We did not measure kinematic changes in position as the
subjects ascended and descended the stairs, where muscle
activity can further influence kinematic characteristics.
Another limitation of this study includes only evaluating
three muscles. Two-joint muscles and other muscles about
the hip may provide greater information about compensa-
249
250
tion strategies. The rectus femoris and hamstring muscles
in particular act at both the hip and knee joints, while the
gluteus maximus is a powerful hip extensor muscle. In
fact any muscle in the lower extremity or hip could be af-
fected by the development of a compensation strategy.
Further study is needed to identify possible compensa-
tion strategies in patients with AKP. Future studies should
include evaluation of more lower extremity muscles and
their relation to lower extremity kinematics and kinetics.
Once these compensation strategies are identified treat-
ments can be developed to restore normal muscle activity
and joint kinetics. Treatment interventions for AKP and
patellofemoral pain could then be judged on their efficacy
of affecting the compensation strategy.
Conclusion
This study demonstrated significant temporal differences in
lower extremity EMG firing patterns of the GM between
AKP subjects and a control group without knee pain dur-
ing ascent and descent of the stairs. In particular the GM
demonstrated delayed onset and diminished duration for
both stair ascent and descent. There were no significant
differences between VMO and VL during either stair as-
cent or descent between subjects with AKP and the con-
trol group during either stair ascent or descent, which is
consistent with previous investigations of VMO and VL
activity. We believe that this is a necessary initial step in
identifying possible compensations in the lower extremity
due to AKP, and that a more thorough biomechanical
analysis of this task is needed. Because the GM activity
does appear to be altered due to AKP, rehabilitation strate-
gies to train the GM are warranted.
Acknowledgements We thank Southeastern Athletic Trainers
Association (SEATA) for partial funding for this project, and
J. Kruger, J. Fowler, L. Canty, A. Soules, R. Creighton, and R. Os-
borne for assistance in data collection and analysis. This project
complies with all regulations regarding testing on human subjects.
We also thank Dr. Barbara Myklebust for critical review of this
manuscript. The opinions presented in this report reflect the views
of the authors and not those of the National Institutes of Health or
the United States Public Health Service.
1. Almeida SA, Trone DW, Leone DM,
Shaffer RA, Patheal SL, Long K (1999)
Gender differences in musculoskeletal
injury rates: a function of symptom re-
porting? Med Sci Sports Exerc 31:
18071812
2. Boucher JP, King MA, Lefebvre R,
Pepin A (1992) Quadriceps femoris
muscle activity in patellofemoral pain
syndrome. Am J Sports Med 20:527
532
3. Cowan SM, Bennell KL, Hodges PW,
Crossley KM, McConnell J (2001) De-
layed onset of electromyographic ac-
tivity of vastus medialis obliquus rela-
tive to vastus lateralis in subjects with
patellofemoral pain syndrome. Arch
Phys Med Rehabil 82:183189
4. Doucette S, Child D (1996) The effect
of open and closed chain exercise and
knee joint position on patellar tracking
in lateral patellar compression syn-
drome. J Orthop Sports Phys Ther 23:
104110
5. Ellison JB, Rose SJ, Sahrmann SA
(1990) Patterns of hip rotation range of
motion: a comparison between healthy
subjects and patients with low back
pain. Phys Ther 70:537541
6. Gilleard W, McConnell J, Parsons D
(1998) The effect of patellar taping on
the onset of vastus medialis obliquus
and vastus lateralis muscle activity in
persons with patellofemoral pain. Phys
Ther 78:2532
7. Grelsamer RP, McConnell J (1998)
The patella: a team approach. Aspen,
Gaithersburg
8. Hung YJ, Gross MT (1999) Effect of
foot position on electromyographic ac-
tivity of the vastus medialis oblique
and vastus lateralis during lower-ex-
tremity weight-bearing activities.
J Orthop Sports Phys Ther 29:93102
9. Hutchinson MR, Ireland ML (1995)
Knee injuries in female athletes. Sports
Med 19:288302
10. Kendall FP, McCreary EK (1983)
Muscles: testing and function.
Williams and Wilkins, Baltimore
11. Kernohan WG, Dodd FJ, Dowey KE,
McConnell LA (1993) Patient assess-
ment of treatment following knee in-
jury. Br J Sports Med 27:131134
12. Irrgang J (1994) Rehabilitation for
nonoperative and operative manage-
ment of knee injuries. In: Fu F, Harner
C, Vince K (eds) Williams and Wil-
kins, Baltimore, pp 485501
13. Karara HM, Abdel-Aziz YI (1974) Ac-
curacy aspects of non-metric imageries.
Photogrammetric Eng 40:11071117
14. Karst GM, Willett GM (1995) Onset
timing of electromyographic activity in
the vastus medialis oblique and vastus
lateralis muscles in subjects with and
without patellofemoral pain syndrome.
Phys Ther 75:813823
15. Kleissen RF (1990) Effects of elec-
tromyographic processing methods on
computer-averaged surface elec-
tromyographic profiles for the gluteus
medius muscle. Phys Ther 70:716722
16. LaBrier K, ONeill DB (1993) Patello-
femoral stress syndrome. Current con-
cepts. Sports Med 16:449459
17. Laprade J, Culham E, Brouwer B
(1998) Comparison of five isometric
exercises in the recruitment of the vas-
tus medialis oblique in persons with
and without patellofemoral pain syn-
drome. J Orthop Sports Phys Ther 27:
197204
18. McConnell BE (1976) A dynamic
transpatellar approach to the knee.
South Med J 69:557560
19. Nadeau S, Gravel D, Hebert LJ, Arse-
nault AB, Lepage Y (1997) Gait study
of patients with patellofemoral pain
syndrome. Gait Posture 5:2127
20. Powers CM (2000) Patellar kinematics.
I. The influence of vastus muscle activ-
ity in subjects with and without patello-
femoral pain. Phys Ther 80:956964
21. Powers CM (2000) Patellar kinematics.
II. The influence of the depth of the
trochlear groove in subjects with and
without patellofemoral pain. Phys Ther
80:965978
22. Salsich GB, Brechter JH, Powers CM
(2001) Lower extremity kinetics during
stair ambulation in patients with and
without patellofemoral pain. Clin Bio-
mech 16:906912
References
251
23. Salsich GB, Brechter JH, Farwell D,
Powers CM (2002) The effects of
patellar taping on knee kinetics, kine-
matics, and vastus lateralis muscle ac-
tivity during stair ambulation in indi-
viduals with patellofemoral pain. J Or-
thop Sports Phys Ther 32:310
24. Sheehy P, Burdett RG, Irrgang JJ,
VanSwearingen J (1998) An electro-
myographic study of vastus medialis
oblique and vastus lateralis activity
while ascending and descending steps.
J Orthop Sports Phys Ther 27:423429
25. Taylor N, Evans O, Goldie P (2001)
Reliability of measurement of angular
movements of the pelvis and lumbar
spine during treadmill walking. Physio-
ther Res Int 6:205223
26. Werner S (1995) An evaluation of
knee extensor and knee flexor torques
and EMGs in patients with patello-
femoral pain syndrome in comparison
with matched controls. Knee Surg
Sports Traumatol Arthrosc 3:8994
27. Werner S, Knutsson E, Eriksson E
(1993) Effect of taping the patella on
concentric and eccentric torque and
EMG of knee extensor and flexor mus-
cles in patients with patellofemoral
pain syndrome. Knee Surg Sports
Traumatol Arthrosc 1:169177
28. Yang JF, Winter DA (1984) Electro-
myographic amplitude normalization
methods: improving their sensitivity as
diagnostic tools in gait analysis. Arch
Phys Med Rehabil 65:517521