[ research report ]

Joy L. Long, MD1 • Ramon A. Ruberte Thiele, MS2 • Jack G. Skendzel, MD3 • Jongeun Jeon4
Richard E. Hughes, PhD5 • Bruce S. Miller, MD6 • James E. Carpenter, MD5

Activation of the Shoulder

Musculature During Pendulum
Exercises and Light Activities
Downloaded from www.jospt.org at on June 4, 2019. For personal use only. No other uses without permission.

T
here are a variety of rehabilitation protocols following rotator waist and allows the arm to swing like
cuff repair. Many surgeons allow patients to begin gentle passive a pendulum, moving in circles gener-
ated by trunk motion. Burkhart et al5 ob-
motion within a day of surgery, but active motion is generally
Copyright © 2010 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

served, however, that patients sometimes

restricted to prevent retear of the rotator cuff.30 Nevertheless, mistakenly make this an active exercise,
Galatz et al15 reported retearing of the rotator cuff after surgical performing it incorrectly by using their
repair in over 75% of patients with large tears. While not all patients shoulder muscles to swing the arm rather
with failure of their repair experience One of the more commonly used ex- than simply allowing it to hang in a re-
functional deficits equivalent to their ercises after rotator cuff repair is the laxed state. This type of muscle activity
preoperative status, it is imperative to pendulum, which is also referred to as would not be recommended immediately
optimize postoperative care, including Codman’s exercise.6 To perform this ex- after rotator cuff repair.
rehabilitation, to limit recurrence. ercise, a patient leans forward at the Patients are usually interested in
Journal of Orthopaedic & Sports Physical Therapy®

returning to activities of daily living

t STUDY DESIGN: Prospective, single-group,
repeated-measures design.
t OBJECTIVES: To evaluate electromyographic
(EMG) signal amplitude in the supraspinatus, in-
fraspinatus, and deltoid muscles during pendulum
exercises and light activities in a group of healthy
subjects.
t BACKGROUND: There are numerous rehabilita-
tion protocols used after rotator cuff repair. One
of the most commonly used exercises in these
protocols is the pendulum. Patients can easily
perform these exercises incorrectly, and may
also perform light activities of daily living without
knowing that they may be putting excessive stress
on the repair. The effect of improperly performed
pendulum exercises and light activities after rota-
tor cuff repair is unknown.
t METHODS: Muscle activity was recorded
in 13 subjects performing pendulum exercises
incorrectly and correctly in both large (51-cm) and
small (20-cm) diameters, and while typing, drink-
ing, and brushing their teeth.
t RESULTS: Incorrect and correct large
pendulums and drinking elicited more than 15%
maximum voluntary isometric contraction in the
supraspinatus and infraspinatus. The supraspi-
natus EMG signal amplitude was greater during
large, incorrectly performed pendulums than
during those performed correctly. Both correct and
incorrect large pendulums resulted in statistically
higher muscle activity in the supraspinatus than
the small pendulums.
t CONCLUSION: Larger pendulums may require
more force than is desirable early in rehabilitation
after rotator cuff repair. J Orthop Sports Phys Ther
2010;40(4):230-237. doi:10.2519/jospt.2010.3095
t KEY WORDS: EMG, infraspinatus, rotator cuff,
supraspinatus
(ADLs) as soon as possible after surgery.
Although patients are asked to limit ac-
tive use of the operative extremity im-
mediately after surgery, they may still
perform light activities, believing that
their shoulder is protected in the sling.
Light ADLs that are thought to be harm-
less by the patient and that may not be
addressed in the postoperative instruc-
tions by the surgeon may activate the
shoulder muscles and increase the ten-
sion on the repair. Little is known about
muscle activation in the rotator cuff
muscles when simple ADLs, such as typ-
ing, brushing teeth, and drinking water
from a bottle, are performed.
The purpose of this study was to
measure the EMG signal amplitude of

Sports Medicine Fellow, Department of Orthopaedic Surgery, University of Michigan, Ann Arbor, MI. 2 Research Specialist, Department of Orthopaedic Surgery, University of
1 

Michigan, Ann Arbor, MI. 3 House Officer, Department of Orthopaedic Surgery, University of Michigan, Ann Arbor, MI. 4 Undergraduate Student, Department of Orthopaedic
Surgery, University of Michigan, Ann Arbor, MI. 5 Associate Professor, Department of Orthopaedic Surgery, University of Michigan, Ann Arbor, MI. 6 Assistant Professor, Department
of Orthopaedic Surgery, University of Michigan, Ann Arbor, MI. This study was approved by The University of Michigan Institutional Review Board. Address correspondence to Dr
James E. Carpenter, Department of Orthopaedic Surgery, The University of Michigan, 1500 E Medical Center Drive, Ann Arbor, MI 48109-0328. E-mail: jcarp@med.umich.edu

230 | april 2010 | volume 40 | number 4 | journal of orthopaedic & sports physical therapy

06 Long.indd 230 3/23/10 5:29:15 PM

 the supraspinatus, infraspinatus, and
deltoid muscles in a group of healthy
subjects who performed pendulums cor-
rectly (using trunk motion to cause the
arm to move) versus incorrectly (using
shoulder musculature to cause the arm
to move), and 3 ADLs that included typ-
ing, brushing teeth, and drinking from a
water bottle. We hypothesized that per-
forming pendulum exercises incorrectly
would elicit more muscle activity than
performing pendulum exercises correctly,
and that pendulum exercises and select
light activities would not elicit more than
15% of the maximum voluntary isomet-
Downloaded from www.jospt.org at on June 4, 2019. For personal use only. No other uses without permission.
ric contraction (MVIC) in the shoulder
muscles.
METHODS
Copyright © 2010 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.
A
pproval was obtained from the
University of Michigan Institu-
tional Review Board prior to be-
ginning the study, and all subjects signed
an informed consent document and were
compensated. The subjects’ rights were
protected during the course of the study.
We recruited healthy, right-hand–domi-
Journal of Orthopaedic & Sports Physical Therapy®
nant individuals aged 18 years or older.
Exclusion criteria included any history
of shoulder pain, injury, or surgery, use
of an assistive device for ambulation,
and inability to tolerate the study proto-
col. The right shoulder was tested in all
trials. Seventeen subjects were recruited
to participate in this study. Two subjects
were unable to complete the study due
to vasovagal episodes, and 2 others were
excluded due to technical problems ob-
taining EMG data. Of the 13 subjects in
our final analysis, 7 were males and 6
females, with an average age of 29 years
(range, 20-57 years).
A physician sterilized and inserted
1 set of 50-mm, disposable, paired-
wire electrodes (VIASYS Healthcare
Inc, Madison, WI) through a 25-gauge
needle, for EMG monitoring of the su-
praspinatus and infraspinatus muscles.
Each paired-wire electrode consisted of
2 strands of insulated nickel alloy wire.
The wires were arranged so that 2 mm
journal of orthopaedic & sports physical therapy | volume 40 | number 4 | april 2010 |
06 Long.indd 231
FIGURE 1. Location of wire electrodes for
electromyographic data collection of the
supraspinatus and the infraspinatus muscles.
and 5 mm of the first and second wire
exited the needle, respectively. The last 2
mm of insulation was stripped from each
wire. The wires for the supraspinatus
were placed approximately 2 cm superi-
or to the midpoint of the scapular spine,
and the wires for the infraspinatus were
placed approximately 2 cm inferior to the
midpoint of the scapular spine. Bipolar,
pediatric-sized Ambu Blue Sensor Ag-
AgCl electrodes (Ambu Inc, Glen Burnie,
MD) were used for the surface recordings
of the middle deltoid muscle. These oval-
shaped electrodes were placed approxi-
mately 3 cm distal to the lateral aspect of
the acromion, with an interelectrode dis-
tance of 2.2 cm. The technique for place-
ment of the wire and surface electrodes
was consistent with that described by
Basmajian and De Luca.2 The locations of
the electrodes are shown in FIGURES 1 and
2. Proper placement of the electrodes was
verified by noting an appropriate electri-
cal response with resisted action of the
muscle. One of the surface electrodes
was also placed on the right olecranon as
a ground.
Data were recorded at 1000 Hz on a
Noraxon Myosystem 2000 EMG system
(Noraxon, Inc, Scottsdale, AZ), collected
on a computer using MotionMonitor
software (Innovative Sports Training,
Inc, Chicago, IL), and stored for offline
data analysis. The system uses differen-
tial amplifiers with a gain of 1000. There
was an input impedance of 10 MΩ, a
signal-to-noise ratio of less than 1 mV
root-mean-square (RMS), and a com-
mon-mode rejection ratio of 135 dB. The
FIGURE 2. Location of surface electrodes for
electromyographic data collection of the deltoid
muscle.
data-processing approach was designed
to represent the contraction dynamics of
muscle, based on published EMG-driven
muscle force prediction models.27,28 Signal
processing began with full-wave rectifica-
tion, followed by a 2-pass, fourth-order,
low-pass Butterworth filter with a cutoff
frequency of 3 Hz to obtain a signal for
the ith muscle, ei (t). Then a recursive fil-
ter was applied: uj (t) = αej (t – d) – β1uj
(t – 1) – β2uj (t – 2), where uj(t) was the
processed signal at time t, α was a gain
for the muscle (α = .02), d was electrome-
chanical delay (d = 40 ms), and β1 and β2
were coefficients for the recursive filter (β1
= –1.46 and β2 = 0.48). Muscle activation
was then estimated from ui(t), assuming
a linear relationship. The end product of
this processing was an estimate of muscle
activation.
One dependent measure was com-
puted for each muscle for each trial:
maximum normalized muscle activation.
Normalization was done by dividing the
activation of a muscle by its maximum
muscle activation, computed from EMG
data recorded during MVIC trials. The
subject was asked to perform maximal
isometric external rotation and abduc-
tion movement to elicit MVIC for each
231
3/23/10 5:29:16 PM
 [ research report
muscle. The isometric actions were per-
formed by having the patient pull on a
secured rope while standing. For abduc-
tion, the elbow was in extension, for ex-
ternal rotation, the elbow was in 90° of
flexion and held against the torso (by the
patient, not by a restrictive device). No
sling was used. A rest EMG measurement
was also obtained, with the patient sit-
ting quietly at rest for at least 4 seconds.
The average baseline muscle activation
level at rest was determined and sub-
tracted from all other values. The peak
muscle activation was determined for
each trial so that the peak percent MVIC
Downloaded from www.jospt.org at on June 4, 2019. For personal use only. No other uses without permission.
could be calculated for each activity. The
peak percent MVIC over the entire trial
for each muscle during each activity was
recorded for the individual subjects. The
Copyright © 2010 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.
mean for all of the subjects was then ob-
tained from these peak values, and the
standard deviations were calculated.
These methods are consistent with prior
studies.20,22
The subject was then placed in a stan-
dard sling and asked to complete the 3
ADLs in a previously determined ran-
domized order. Typing was performed
Journal of Orthopaedic & Sports Physical Therapy®
using a standard keyboard, and the
height of the chair was adjusted for in-
dividual comfort. Subjects were asked to
type continuously, while three 4-second
data sets were recorded. Similarly, each
subject was given a toothbrush and asked
to brush their teeth continuously while
three 4-second data sets were recorded.
For the drinking task, each subject was
given an identical 0.5-L bottle of water.
The subject was asked to take 2 sips of
water, placing the bottle back on the table
between sips. EMG activity was recorded
for 3 trials of 2 sips each. No time limit
was used. The subjects were not timed for
this activity, as the intention was to allow
the subjects to perform the task as they
would normally.
Subjects were then instructed, using
a combination of verbal explanation and
demonstration by a physician, in how
to perform pendulum exercises. In this
study, correct pendulum exercises were
the swinging or stooping exercises that
232 | april 2010 | volume 40 | number 4 | journal of orthopaedic & sports physical therapy
06 Long.indd 232
FIGURE 3. Subject performing a pendulum exercise.
The laser pointer falls within the outer dark
concentric circle (arrow), indicating a large pendulum
exercise.
Codman6 originally described for the
early postoperative period, which were
meant to be a passive exercise for the
shoulder. Alternatively, incorrect pendu-
lum exercises were similar swinging ex-
ercises performed using active motion of
the shoulder muscles. The subjects were
given as much time as they felt necessary
to practice the pendulum exercise before
the actual trials commenced. The subject
performed the 4 types of pendulum exer-
cises (large incorrect, large correct, small
incorrect, and small correct) in a previ-
ously determined, randomized order. The
pendulum exercises were standardized by
using a diagram of 10-cm-wide concen-
tric circles, the largest of which was 51
cm and the smallest 20 cm in diameter,
placed on the floor (FIGURE 3). The sub-
ject performed the pendulum exercise
in either the incorrect or correct method
for each size. Three revolutions of each
pendulum were completed without any
time limit. The diameter of the circle of
the pendulum was assured by taping a
laser pointer to the dorsal surface of the
subject’s wrist and asking the subject to
keep the laser essentially within a 10-cm-
wide concentric circle. It was emphasized
to the subject during the correct exer-
cises that it was more important for the
exercise to be passive than for the laser
pointer to remain perfectly within the
concentric circle diagram. The subject
was also instructed to stop any correct
pendulum trial if the subject perceived
that the motion was active rather than
passive.
]
Statistical Analysis
A power analysis was performed based on
data collected in a pilot study of 3 sub-
jects. It was based on detecting a differ-
ence of 10% MVIC in the supraspinatus
muscle, as the pilot data showed a stan-
dard deviation of the difference between
large active and passive pendulum exer-
cises of 10.9% MVIC. It was determined
that 12 subjects would be sufficient to
provide 80% power, when controlling for
type I error at 0.05 using a 2-sided test.
Statistical analyses were performed
for both pendulum and ADL data. Nor-
mal probability plots were prepared for
each dependent variable, and it was de-
termined that they were not sufficiently
normal to justify using parametric statis-
tical methods. Therefore, a Friedman test
was used to implement a nonparametric
repeated-measurements analysis of the
pendulum data. This was a 2-by-2 model,
with 1 factor being the experimental fac-
tor of interest and the other factor being
a subject. Significance was set at 0.05 for
assessing main effects. When a main ef-
fect was found, post hoc analyses were
performed to assess pairwise differences
using the method described by Pett.32 The
method compares differences in average
ranks, and it limits the overall type I er-
ror to 0.05.
RESULTS
T
hree activities (large cor-
rect pendulums, large incorrect
pendulums, and drinking from a
water bottle) showed a mean peak per-
cent MVIC greater than 15% in the su-
praspinatus and infraspinatus muscles.
Mean peak percent MVIC and standard
deviations for all activities are reported
in the TABLE. A graphic representation is
included in FIGURES 4 and 5.
Supraspinatus percent MVIC was af-
fected by pendulum exercise (P.001).
Post hoc analyses showed that the differ-
ence in percent MVIC in the supraspi-
natus between incorrect and correct
pendulum exercises reached statistical
significance for large-diameter pendu-
3/23/10 5:29:17 PM
 lums, but not for small-diameter pen- analysis was based on large pendulum about the lack of differences found be-
dulums. Because the statistical power data, no strong conclusions can be made tween correct and incorrect performance
of small pendulum exercises. The differ-
Mean (SD) EMG Signal Amplitude ence in percent MVIC in the supraspina-
TABLE tus for large versus small pendulums was
as Percent of MVIC for Activities
significant for those performed both cor-
Supraspinatus Infraspinatus Deltoid rectly and incorrectly.
Small correct pendulum 8.0 (6.3)*,† 14.4 (37.4)*,† 2.8 (2.0)*,† The statistical analysis also showed
Small incorrect pendulum 8.7 (5.1)
§
12.9 (25.8)
§,‡
4.0 (2.4)§ that drinking from a bottle of water had
Large correct pendulum 13.7 (12.9)* ,||
22.4 (39.0)* ,‡
4.4 (2.7)* significantly higher muscle activation
Large incorrect pendulum 18.8 (15.6)†,§,|| 24.5 (41.2)§,† 6.0 (3.1)†,§ than typing for both the supraspinatus
Typing 7.5 (5.1)¶ 12.0 (13.2)¶,** 2.1 (2.3)¶,** and the infraspinatus muscles.
Drinking 21.0 (20.4) ¶,#
18.3 (14.8)
¶
4.4 (2.6)¶
Brushing teeth 12.1 (10.1) #
20.2 (21.3)** 4.1 (3.8)** DISCUSSION
Downloaded from www.jospt.org at on June 4, 2019. For personal use only. No other uses without permission.

Abbreviations: EMG, electromyographic; MVIC, maximum voluntary isometric contraction.

C
* Significant difference between small correct pendulum and large correct pendulum, P.05. odman6 was one of the first to
†
Significant difference between small correct pendulum and large incorrect pendulum, P.05.
‡
Significant difference between small incorrect pendulum and large correct pendulum, P.05. emphasize the complex nature of
§
Significant difference between small incorrect pendulum and large incorrect pendulum, P.05. the shoulder joint, which allows for
||
Significant difference between large correct pendulum and large incorrect pendulum, P.05. equally complex motion. In 1944, Inman
Copyright © 2010 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

¶
Significant difference between typing and drinking, P.05.
#
Significant difference between drinking and brushing teeth, P.05. et al19 performed a study of shoulder mor-
** Significant difference between typing and brushing teeth, P.05. phology and function, as well as EMG
analysis of shoulder musculature. They
emphasized that shoulder muscle groups
act synchronously during certain move-
ments. This has important implications
in rehabilitation after rotator cuff repair,
Small correct
pendulum because the desired effect is to maintain
passive range of motion without stress-
Journal of Orthopaedic & Sports Physical Therapy®

ing the repair. The effect of improperly

performed postoperative pendulum ex-
ercises and light ADLs after rotator cuff
Small incorrect
pendulum
repair is largely unknown. Measuring
the amount of muscle activation during
tasks may provide information about
the effect of activity on the rotator cuff.
Our results indicate that the EMG activ-
Large correct ity of the supraspinatus is higher when
pendulum
healthy subjects perform pendulum ex-
ercises incorrectly in larger (51-cm diam-
eter) circles, and a significant difference
in activation of all 3 muscles was found
Large incorrect between the large and small diameters
pendulum for both the correctly and incorrectly
performed pendulums.
Many previous researchers have exam-
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 ined activation of the shoulder muscula-
ture during active and isometric exercises
MVIC (%)
for rehabilitation.1,8-10,14,17,20,23,33,41,44,47 Few,
Deltoid Infraspinatus Supraspinatus however, have looked at EMG data for ac-
tivities that are meant to be passive.13,31,37,38
Dockery et al13 found approximately 5%
FIGURE 4. Percent maximum voluntary isometric contraction (MVIC) for the pendulum exercises, with standard error.
MVIC in the rotator cuff by examining

journal of orthopaedic & sports physical therapy | volume 40 | number 4 | april 2010 | 233

06 Long.indd 233 3/23/10 5:29:18 PM


Brushing teeth
Typing
Downloaded from www.jospt.org at on June 4, 2019. For personal use only. No other uses without permission.
Drinking
Copyright © 2010 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.
0.0 5.0 10.0
Deltoid
Journal of Orthopaedic & Sports Physical Therapy®
FIGURE 5. Percent maximum voluntary isometric contraction (MVIC) for tested activities of daily living, with
standard error.
surface EMG data for pendulum exercis-
es, but he did not include fine-wire EMG
nor did he study any ADLs. McCann et
al29 studied multiple passive exercises,
including the pendulum, using wire elec-
trode EMG techniques. They found mini-
mal activity in the supraspinatus, deltoid,
and infraspinatus with pendulum exer-
cises presumably performed correctly. No
mention is made of the specific instruc-
tions to the patient on how the exercises
were to be performed.29 Smith et al36-38
have published EMG data for the immo-
bilized shoulder to determine the safety
of various exercises and activities after
shoulder surgery. They recommended
that certain motions, such as a backwards
pulling motion, even if performed with
the contralateral arm, should be avoided
after rotator cuff repair, because the mus-
cles in the operative extremity are not
234 | april 2010 | volume 40 | number 4 | journal of orthopaedic & sports physical therapy
06 Long.indd 234
[ research report
15.0 20.0 25.0
MVIC (%)
Infraspinatus Supraspinatus
quiescent.38 In several published papers,
this group has studied various types of
exercises, including scapulothoracic and
contralateral limb movements; however,
pendulum exercises and specific ADLs
were not included, although some of the
resisted motions were meant to simulate
ADLs, such as opening a door or lifting
a bag.36-38
Although standardizing ADLs is more
difficult than standardizing specific re-
habilitation exercises, we felt that it was
important to have subjects perform ac-
tual tasks rather than less-complicated
motions associated with different activi-
ties. While there have been some ergo-
nomic studies involving keyboards that
measure the EMG signal amplitude in
some of the shoulder musculature dur-
ing typing and similar activities, none of
these studies look at the EMG data for
]
the supraspinatus, which is the muscle
most commonly involved in rotator cuff
tears.7,11,12,40,46 Kelly et al23 investigated
EMG data for more intense ADLs, such
as lifting 1- to 10-pound (0.5- to 4.5-kg)
objects and washing the back, in subjects
with symptomatic versus asymptomatic
rotator cuff tears compared to healthy
controls. The results of that study showed
that the symptomatic patients had more
EMG signal amplitude for the muscles
of the torn rotator cuff tendon than the
asymptomatic patients, which could ex-
plain the disability of the symptomatic
patients who were not compensating by
using the surrounding shoulder muscula-
ture.23 Interestingly, Kelly et al23 showed
that the percent MVIC of the supraspi-
natus was higher for all studied activities
in patients with rotator cuff tears than in
normal controls. For example, the activity
of lifting a 1-pound (0.5-kg) weight to the
level of the shoulder required less than
30.0
20% MVIC for controls but over 50%
MVIC for patients with symptomatic ro-
tator cuff tears.23 Although our tasks were
not identical, the findings of Kelly et al23
do bring into the forefront the differences
in EMG findings between healthy sub-
jects and patients with rotator cuff tears.
It is unclear whether patients who have
had a rotator cuff repair would be more
similar to a control subject or a patient
with a rotator cuff tear.
While there is no definitive consensus
opinion, previous reports of EMG data
for the shoulder have considered activity
levels of less than 20% to 25% MVIC to
be low or minimal.20,29,36-38 Based on these
reports, our study shows that there is at
least moderate activity produced in the
supraspinatus muscle by large incorrect
pendulums and drinking in healthy sub-
jects. However, the published EMG activ-
ity designations are not standardized for
percentage levels or for the definition of
MVIC. For example, McCann et al29 used
a different method to obtain MVIC than
we used. Rather than a maximum iso-
metric contraction, subjects in their study
lifted a 2.25-kg weight, and the EMG out-
put for this action was used as MVIC.29
3/23/10 5:29:19 PM
 Thus, their results may represent a lower
overall muscle activity than studies that
use a maximal isometric contraction, as
we did. In an effort to be conservative,
we designated muscle activation greater
than 15% MVIC to be potentially indica-
tive of higher loads than desirable in a
newly repaired rotator cuff. Although our
study involves healthy subjects, the goal
is to help identify activities that may po-
tentially be safely performed by a patient
who has recently undergone rotator cuff
repair. Fifteen percent MVIC correlates
to 30 N of force, based on our estimates.
This is lower than the lowest force (50 N)
Downloaded from www.jospt.org at on June 4, 2019. For personal use only. No other uses without permission.
that we found in the literature for failure
after cyclic loading of the rotator cuff.3-
5,16,24,25,34,35,39,42,45,48
We chose a threshold of 15% of maxi-
Copyright © 2010 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.
mum activation for our data interpreta-
tion, based on biomechanical calculations.
One study using in vitro biomechanical
testing of sutured supraspinatus defects
indicated a 50% loss of repair at (mean
 SD) 206  88 cycles of loading at 44
 15 N.3 A conservative value of 30 N was
chosen as an upper limit of safe loading
for the limited-repetition activities used
Journal of Orthopaedic & Sports Physical Therapy®
in this study. Thirty-one N was estimated
to be a percent of maximal contraction
by the following procedure. Assuming
supraspinatus muscle force can be mod-
eled as the product of MVIC fraction (be-
tween 0 and 1), muscle cross-sectional
area (3.36 cm2),26 specific tension (62 N/
cm2),18 and cosine of the fiber pennation
angle (7.0°),26 15% MVIC corresponds to
31 N of force. While many assumptions
were made in the determination of 15%
MVIC and 31 N, we believe these num-
bers to be conservative, yet useful for this
and, potentially, future studies.
In analyzing our results, there are
several reasons that large pendulum
exercises could produce more muscle
activity than small pendulum exercises.
Performing the larger-diameter exercises
may be more difficult than the smaller ex-
ercises, making it harder for the subject
to perform the pendulum passively. Al-
ternatively, pendulum exercises may not
be truly passive at all. If so, then a larger
journal of orthopaedic & sports physical therapy | volume 40 | number 4 | april 2010 |
06 Long.indd 235
exercise would require more muscle ac-
tivity. Furthermore, although we focused
on the supraspinatus, as this tendon is
most frequently involved in rotator cuff
tears, we did note that the data for the
infraspinatus was less consistent overall.
It is unclear whether even passive pen-
dulum exercises may elicit relatively high
muscle activation in the infraspinatus, at
least in some individuals.
This study has several limitations.
First, our instrumentation was limited
to a 1000-Hz bandwidth. It has been
reported that 5% of the power of the
supraspinatus signal is greater than
510 Hz, when measured using fine-wire
electrodes43; therefore, a small amount
of signal may be aliased, given the sam-
pling rate we used. Second, EMG is not
a direct measurement of muscle force.
While our EMG processing method was
designed to include important elements
of models to estimate muscle force from
EMG, the method lacks a nonlinear
transformation from muscle activation to
muscle force. However, our nonparamet-
ric statistical analysis is based only on the
weak assumption of an ordinal scale of
measurement.32 Because EMG-to-force
mappings are monotonically increasing,
our statistical results are independent of
the exact functional relationship between
computed muscle activation and muscle
force.27 Third, the activity of the shoulder
muscles in healthy subjects may not be
representative of the activity of individu-
als with a repaired rotator cuff muscle.
Fourth, the average age of the volunteers
is younger than the average age of pa-
tients undergoing rotator cuff surgery.
If we anticipate that the findings in
healthy individuals may approximate the
findings in patients who have undergone
rotator cuff repair, then some concern is
raised about the rehabilitation of these
patients postoperatively. If patients are
going to perform pendulums without su-
pervision, then it is important that they
understand and remember to do the
exercises passively in small circles (ap-
proximately 20 cm in diameter). Those
involved in the postoperative care of pa-
tients who have undergone rotator cuff
repair may wish to consider these rec-
ommendations when prescribing home
exercise programs for individual patients.
Currently, some clinicians and therapists
instruct patients to perform pendulum
exercises with the opposite arm sup-
ported by a table. It is possible that sup-
port of the contralateral limb may affect
muscle activity in the affected shoulder
during pendulum exercises. In our study,
subjects were permitted to use a table for
support of the contralateral limb, if pre-
ferred, though most did not. Codman’s
original description of the exercise did
not mention contralateral arm support.
There are some patients who will lack
the ability to perform pendulum exer-
cises as prescribed, so supervised or as-
sisted range-of-motion exercises may be
safer for these individuals. Just as indi-
vidual patients may perform pendulums
incorrectly or differently, our subjects,
although they underwent specific instruc-
tion and were observed during the exer-
cises, had some level of variability in the
amount of muscle activity recorded with
the different exercises. With regard to
ADLs, those patients who wish to return
to typing in the early phase after rotator
cuff repair may be able to do so without
placing excess strain on the repair. How-
ever, even though the patient is able to
perform a task, such as drinking from a
bottle of water or brushing teeth, with the
operative arm in a sling, the activity may
not be entirely safe. Thus, patients may
need specific instructions (restrictions)
postoperatively for ADLs.
CONCLUSION
L
arge pendulums and pendulums
performed incorrectly generate more
supraspinatus muscle activity in the
shoulder than smaller, correctly per-
formed pendulum exercises. Based on
this finding, we believe that it is possible
that larger pendulum exercises may not
be desirable early in rehabilitation after
rotator cuff repair. In healthy individuals,
drinking water from a bottle produces
235
3/23/10 5:29:20 PM
 [ research report
more supraspinatus muscle activity than About the Subacromial Bursa. Malabar, FL: G.
Miller and Co Medical Publishers, Inc; 1984.
typing. Drinking water from a bottle may
7. Cook C, Burgess-Limerick R, Papalia S. The
also not be desirable immediately after effect of upper extremity support on up-
rotator cuff repair. t per extremity posture and muscle activity
during keyboard use. Applied Ergonomics.
2004;35:285-292.
KEY POINTS 8. Cordasco FA, Wolfe IN, Wootten ME, Bigliani LU.
FINDINGS: Large pendulums and pendu- An electromyographic analysis of the shoulder
lums performed incorrectly generate during a medicine ball rehabilitation program.
more supraspinatus muscle activity than Am J Sports Med. 1996;24:386-392.
9. Dark A, Ginn KA, Halaki M. Shoulder muscle
smaller, correctly performed pendulum recruitment patterns during commonly used ro-
exercises. tator cuff exercises: an electromyographic study.
IMPLICATIONS: Smaller pendulums (20 cm Phys Ther. 2007;87:1039-1046. http://dx.doi.
in diameter) may be safer for patients org/10.2522/ptj.20060068
10. David G, Magarey ME, Jones MA, Dvir Z, Turker
who have had rotator cuff repair. KS, Sharpe M. EMG and strength correlates of
CAUTION: EMG is not a direct measure-
Downloaded from www.jospt.org at on June 4, 2019. For personal use only. No other uses without permission.
]
1996;24:477-485.
21. J obe FW, Moynes DR, Tibone JE, Perry J. An
EMG analysis of the shoulder in pitching. A sec-
ond report. Am J Sports Med. 1984;12:218-220.
22. Kelly BT, Kadrmas WR, Speer KP. The manual
muscle examination for rotator cuff strength. An
electromyographic investigation. Am J Sports
Med. 1996;24:581-588.
23. Kelly BT, Williams RJ, Cordasco FA, et al. Differ-
ential patterns of muscle activation in patients
with symptomatic and asymptomatic rotator
cuff tears. J Shoulder Elbow Surg. 2005;14:165-
171. http://dx.doi.org/10.1016/j.jse.2004.06.010
24. Kim DH, Elattrache NS, Tibone JE, et al. Bio-
mechanical comparison of a single-row versus
double-row suture anchor technique for rotator
cuff repair. Am J Sports Med. 2006;34:407-414.
http://dx.doi.org/10.1177/0363546505281238

selected shoulder muscles during rotations of
ment of muscle force, so translation the glenohumeral joint. Clin Biomech (Bristol,
of EMG results to imply muscle force Avon). 2000;15:95-102.
11. Delisle A, Lariviere C, Plamondon A, Imbeau
should be done with caution. Further, D. Comparison of three computer office work-
findings in healthy subjects may not stations offering forearm support: impact on
Copyright © 2010 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.
25. Klinger HM, Steckel H, G. S, Buchhorn GH,
Baums MH. Biomechanical comparison of
double-loaded suture anchors using arthroscop-
ic Mason-Allen stitches versus traditional
transosseous suture technique and modified
Mason-Allen stitches for rotator cuff repair. Clin

replicate the findings in patients with
rotator cuff repair.
ACKNOWLEDGEMENTS: The authors thank
Chris Gatti for his assistance in MATLAB
programming and data analysis, and Kristi
Overgaard for her assistance with manuscript
preparation and revision.
Journal of Orthopaedic & Sports Physical Therapy®
upper limb posture and muscle activation.
Ergonomics. 2006;49:139-160. http://dx.doi.
org/10.1080/10610270500450739
12. Dennerlein JT, Johnson PW. Different com-
puter tasks affect the exposure of the upper
extremity to biomechanical risk factors.
Ergonomics. 2006;49:45-61. http://dx.doi.
org/10.1080/00140130500321845
13. Dockery ML, Wright TW, LaStayo PC. Elec-
tromyography of the shoulder: an analysis
of passive modes of exercise. Orthopedics.
Biomech (Bristol, Avon). 2007;22:106-111.
26. Langenderfer J, Jerabek SA, Thangamani
VB, Kuhn JE, Hughes RE. Musculoskeletal
parameters of muscles crossing the shoul-
der and elbow and the effect of sarcomere
length sample size on estimation of optimal
muscle length. Clin Biomech (Bristol, Avon).
2004;19:664-670. http://dx.doi.org/10.1016/j.
clinbiomech.2004.04.009
27. Lloyd DG, Besier TF. An EMG-driven musculosk-
eletal model to estimate muscle forces and knee

1998;21:1181-1184. joint moments in vivo. J Biomech. 2003;36:765-

references
1. A lpert SW, Pink MM, Jobe FW, McMahon PJ,
Mathiyakom W. Electromyographic analysis of
deltoid and rotator cuff function under vary-
ing loads and speeds. J Shoulder Elbow Surg.
2000;9:47-58.
2. Basmajian JV, De Luca CJ. Muscles Alive: Their
Functions Revealed by Electromyography. 5th
ed. Baltimore, MD: Williams and Wilkins; 1985.
3. Bicknell RT, Harwood C, Ferreira L, et al. Cy-
clic loading of rotator cuff repairs: an in vitro
biomechanical comparison of bioabsorbable
tacks with transosseous sutures. Arthroscopy.
2005;21:875-880. http://dx.doi.org/10.1016/j.
arthro.2005.03.009
4. Burkhart SS, Diaz Pagan JL, Wirth MA, Atha-
nasiou KA. Cyclic loading of anchor-based
rotator cuff repairs: confirmation of the tension
overload phenomenon and comparison of su-
ture anchor fixation with transosseous fixation.
Arthroscopy. 1997;13:720-724.
5. Burkhart SS, Johnson TC, Wirth MA, Athanasiou
KA. Cyclic loading of transosseous rotator cuff
repairs: tension overload as a possible cause of
failure. Arthroscopy. 1997;13:172-176.
6. Codman EA. The Shoulder: Rupture of the
Supraspinatus Tendon and Other Lesions in or
14. Fujisawa H, Suenaga N, Minami A. Electromyo-
graphic study during isometric exercise of the
shoulder in head-out water immersion. J Shoul-
der Elbow Surg. 1998;7:491-494.
15. Galatz LM, Ball CM, Teefey SA, Middleton WD,
Yamaguchi K. The outcome and repair integrity
of completely arthroscopically repaired large
and massive rotator cuff tears. J Bone Joint
Surg Am. 2004;86-A:219-224.
16. Goradia VK, Mullen DJ, Boucher HR, Parks BG,
O’Donnell JB. Cyclic loading of rotator cuff re-
pairs: A comparison of bioabsorbable tacks with
metal suture anchors and transosseous sutures.
Arthroscopy. 2001;17:360-364.
17. Hintermeister RA, Lange GW, Schultheis JM, Bey
MJ, Hawkins RJ. Electromyographic activity and
applied load during shoulder rehabilitation exer-
cises using elastic resistance. Am J Sports Med.
1998;26:210-220.
18. Ikai M, Fukunaga T. Calculation of muscle
strength per unit cross-sectional area of human
muscle by means of ultrasonic measurement.
Int Z Angew Physiol. 1968;26:26-32.
19. Inman VT, Saunders M, Abbott LC. Observations
on the function of the shoulder joint. J Bone
Joint Surg Am. 1944;26:1-30.
20. Jenp YN, Malanga GA, Growney ES, An KN. Acti-
vation of the rotator cuff in generating isometric
shoulder rotation torque. Am J Sports Med.
776.
28. Manal K, Gonzalez RV, Lloyd DG, Buchanan TS.
A real-time EMG-driven virtual arm. Comput
Biol Med. 2002;32:25-36.
29. McCann PD, Wootten ME, Kadaba MP, Bigliani
LU. A kinematic and electromyographic study
of shoulder rehabilitation exercises. Clin Orthop
Relat Res. 1993;179-188.
30. Millett PJ, Wilcox RB, 3rd, O’Holleran JD, Warner
JJ. Rehabilitation of the rotator cuff: an evalua-
tion-based approach. J Am Acad Orthop Surg.
2006;14:599-609.
31. Misamore GW, Ziegler D, Higgin botham G.
Passive range of motion exercises of the shoul-
der: an EMG analysis. J Shoulder Elbow Surg.
1993;2:S27.
32. Pett MA. Nonparametric Statistics for Health
Care Research: Statistics for Small Samples
and Unusual Distributions. Thousand Oaks, CA:
SAGE Publications, Inc; 1997.
33. Reinold MM, Wilk KE, Fleisig GS, et al. Elec-
tromyographic analysis of the rotator cuff and
deltoid musculature during common shoulder
external rotation exercises. J Orthop Sports
Phys Ther. 2004;34:385-394. http://dx.doi.
org/10.2519/jospt.2004.0665
34. Rossouw DJ, McElroy BJ, Amis AA, Emery RJ. A
biomechanical evaluation of suture anchors in
repair of the rotator cuff. J Bone Joint Surg Br.

236 | april 2010 | volume 40 | number 4 | journal of orthopaedic & sports physical therapy

06 Long.indd 236 3/23/10 5:29:21 PM

 1997;79:458-461.
35. Smith CD, Alexander S, Hill AM, et al. A biome-
chanical comparison of single and double-row
fixation in arthroscopic rotator cuff repair.
J Bone Joint Surg Am. 2006;88:2425-2431.
http://dx.doi.org/10.2106/JBJS.E.00697
36. Smith J, Dahm DL, Kaufman KR, et al. Electro-
myographic activity in the immobilized shoulder
girdle musculature during scapulothoracic exer-
cises. Arch Phys Med Rehabil. 2006;87:923-927.
http://dx.doi.org/10.1016/j.apmr.2006.03.013
37. Smith J, Dahm DL, Kotajarvi BR, et al. Elec-
tromyographic activity in the immobilized
shoulder girdle musculature during ipsilateral
kinetic chain exercises. Arch Phys Med Rehabil.
2007;88:1377-1383. http://dx.doi.org/10.1016/j.
apmr.2007.07.028
38. Smith J, Padgett DJ, Dahm DL, et al. Electro-
Downloaded from www.jospt.org at on June 4, 2019. For personal use only. No other uses without permission.
Surg Br. 1992;74:585-588.
40. S zeto GP, Straker LM, O’Sullivan PB. A compari-
son of symptomatic and asymptomatic office
workers performing monotonous keyboard
work--1: neck and shoulder muscle recruitment
patterns. Man Ther. 2005;10:270-280. http://
dx.doi.org/10.1016/j.math.2005.01.004
41. Takeda Y, Kashiwaguchi S, Endo K, Matsuura
T, Sasa T. The most effective exercises for
strengthening the supraspinatus muscle: evalu-
ation by magnetic resonance imaging. Am J
Sports Med. 2002;30:374-381.
42. Tingart MJ, Apreleva M, Lehtinen J, Zurakowski
D, Warner JJ. Anchor design and bone mineral
density affect the pull-out strength of suture
anchors in rotator cuff repair: which anchors are
best to use in patients with low bone quality?
Am J Sports Med. 2004;32:1466-1473.
Sports Phys Ther. 2003;33:109-117.
45. W altrip RL, Zheng N, Dugas JR, Andrews JR.
Rotator cuff repair. A biomechanical com-
parison of three techniques. Am J Sports Med.
2003;31:493-497.
46. Westad C, Mork PJ, Westgaard RH. Firing pat-
terns of low-threshold trapezius motor units in
feedback-controlled contractions and vocational
motor activities. Exp Brain Res. 2004;158:465-
473. http://dx.doi.org/10.1007/s00221-004-
1925-y
47. Worrell TW, Corey BJ, York SL, Santiestaban J.
An analysis of supraspinatus EMG activity and
shoulder isometric force development. Med Sci
Sports Exerc. 1992;24:744-748.
48. Zheng N, Harris HW, Andrews JR. Failure analy-
sis of rotator cuff repair: a comparison of three

myographic activity in the immobilized shoulder
girdle musculature during contralateral upper
limb movements. J Shoulder Elbow Surg.
2004;13:583-588. http://dx.doi.org/10.1016/
S1058274604001259
43. Tomlinson DP, Vanadurongwan B, Henn F, et al. A
spectral analysis of healthy rotator cuff muscula-
ture electromyographic activity: surface and fine
wire. Gait and Clinical Movement Analysis Soci-
ety 13th Annual Meeting. Richmond, VA: 2008.
double-row techniques. J Bone Joint Surg Am.
2008;90:1034-1042. http://dx.doi.org/10.2106/
JBJS.G.00049

@ more information
39. Sward L, Hughes JS, Amis A, Wallace WA. The 44. Uhl TL, Carver TJ, Mattacola CG, Mair SD, Nitz
Copyright © 2010 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

strength of surgical repairs of the rotator cuff. A AJ. Shoulder musculature activation during up-
biomechanical study on cadavers. J Bone Joint per extremity weight-bearing exercise. J Orthop www.jospt.org
Journal of Orthopaedic & Sports Physical Therapy®

GO GREEN By Opting Out of the Print Journal

JOSPT subscribers and APTA members of the Orthopaedic and Sports
Physical Therapy Sections can help the environment by “opting out” of
receiving the Journal in print each month as follows. If you are:

· A JOSPT subscriber: Email your request to jospt@jospt.org or call the

Journal office toll-free at 1-877-766-3450 and provide your name and
subscriber number.
· An APTA Orthopaedic or Sports Section member: Go to www.apta.org
and update your preferences in the My Profile area of myAPTA.
Select “myAPTA” from the horizontal navigation menu (you’ll be asked
to login, if you haven’t already done so), then proceed to “My Profile.”
Click on the “Email & Publications” tab, choose your “opt out”
preferences and save.

Subscribers and members alike will continue to have access to JOSPT

online and can retrieve current and archived issues anytime and anywhere
you have Internet access.

journal of orthopaedic & sports physical therapy | volume 40 | number 4 | april 2010 | 237

06 Long.indd 237 3/23/10 5:29:21 PM