J Shoulder Elbow Surg (2013) 22, 102-107

www.elsevier.com/locate/ymse

Electromyographic analysis of the rotator cuff

in postoperative shoulder patients during passive
rehabilitation exercises
Cynthia A. Murphy, MPTa,*, William J. McDermott, PhDb, Roger K. Petersen, PTa,
Scott E. Johnson, DOc, Stephanie A. Baxter, BSb
a

Department of Physical Therapy, The Orthopedic Specialty Hospital, Murray, UT, USA
Department of Sports Science, The Orthopedic Specialty Hospital, Murray, UT, USA
c
Department of Orthopedic Specialty Group, The Orthopedic Specialty Hospital, Murray, UT, USA
b

Background: Numerous rehabilitation protocols exist for postoperative rotator cuff repairs. Because the
goal of early rehabilitation is to prevent postoperative adhesions while protecting the repaired tendons,
it would be advantageous to know which range-of-motion exercises allow the rotator cuff to remain the
most passive in a painful, guarded, postsurgical shoulder.
Methods: Twenty-six subjects who had undergone subacromial decompression, distal clavicle resection,
or a combination of both procedures volunteered to participate within the first 4 days after surgery.
Fine-wire electrodes were inserted into the subjects supraspinatus (SS) and infraspinatus (IS). Muscle
activity was recorded at resting baseline (BL) and during 14 exercises that have been found in the passive
phase of rotator cuff protocols and tested in healthy subjects. Each exercise was compared with BL activity
as well as with other exercises in the same movement group.
Results: The SS remained as passive as BL during therapist- and self-assisted external rotation, therapistassisted elevation, pendulums, and isometric internal rotation and adduction. The IS was activated greater
than BL for all 14 exercises studied.
Conclusion: Of the 14 exercises studied, 6 allowed the SS and 0 allowed the IS to remain as passive as
quiet-stance BL in postsurgical subacromial decompression/distal clavicle resection patients.
Level of evidence: Basic science study, In vivo electromyography.
2013 Journal of Shoulder and Elbow Surgery Board of Trustees.
Keywords: Rotator cuff; passive; postoperative; shoulder; rehabilitation

Whereas rehabilitation protocols for rotator cuff repair

(RCR) are inconsistent regarding specific exercises, it is an
accepted goal that the rotator cuff (RTC) should remain
The Intermountain Health Care Institutional Review Board approved this
study (No. 1001046).
*Reprint requests: Cynthia A. Murphy, MPT, The Orthopedic Specialty
Hospital, 5848 S 300 E, Murray, UT 84107, USA.
E-mail address: cindy.murphy@imail.org (C.A. Murphy).

passive during the early postoperative phase. A rational

approach to postoperative therapy involves early, safe motion
to allow for joint mobility and optimal tendon healing, with
minimal stress placed on the repaired tendons.2,7,23 Restoring
range of motion (ROM), strength, and pain relief after RCR
are best achieved with a customized rehabilitation protocol to
optimize postoperative ROM while protecting and maintaining the integrity of the RTC.2,12 Retearing rates of 25% to

1058-2746/$ - see front matter 2013 Journal of Shoulder and Elbow Surgery Board of Trustees.
doi:10.1016/j.jse.2012.01.021

EMG during passive rotator cuff exercises

94% have been reported in the literature for rotator cuff
repairs6,21; therefore, protecting the repair needs to be the
first priority. On the one hand, it has been suggested that
premature or excessive activation of the RTC after surgery
may contribute to retearing of the repair,7 and on the other
hand, persistent stiffness has been found to be the most
common postoperative complication for rotator cuff repairs,
with rates ranging from 2.7% to 15%.1,9,12,23 Taking these
two factors into account, the primary goal of the physical
therapist in the first few weeks after surgery is to protect the
repair and promote healing while gradually restoring ROM to
prevent persistent postoperative stiffness.3-5,7,11-16,19,22
Previous research quantifying the amount of muscle
activity associated with postoperative exercises using
healthy subjects has reported that supine self-assisted
elevation, internal and external rotation (IR) (ER), therapist-/helper-assisted IR and ER, wand ER, pendulums,
continuous passive motion in 30
to 60
of elevation,
isometric IR, and active scapular depression and protraction
allowed the supraspinatus (SS) and infraspinatus (IS) to
remain the most passive and that pulleys, the remaining
isometric exercises, and the remaining active scapular
motions activate the SS and IS to a higher level than that
which would be considered passive.3,4,14,15,17,19,22 It is not
known how valid it is to take data collected on healthy
subjects and extrapolate to postoperative patients,3 because
the literature suggests that patients with painful, symptomatic shoulders activate muscles differently and are unable to
remain as passive as healthy control subjects.3,4,10,11
Therefore, the purpose of this study was to determine the
relative amounts of electromyographic (EMG) activity of
the most commonly torn RTC muscles, the SS and IS,18
during typical early postoperative exercises in postsurgical patients.

Materials and methods

Subjects were recruited from 3 participating surgeons offices
during a preoperative visit. In total, 26 participants (8 women and
18 men) who were scheduled for subacromial decompression
(SAD) surgery, distal clavicle resection (DCR) surgery, or both
procedures volunteered to participate (mean age [standard
deviation], 53  10 years). The only exclusion criterion was if any
other procedures were performed during surgery. All participants
visited the laboratory on one occasion between 1 and 4 days after
surgery and signed a written informed consent form before testing.
While patients were standing at rest, they were asked to rate their
current pain level on a scale from 0 to 10 (with 10 being the worst
pain they have ever felt in their lives).
A Noraxon TeleMyo 2400 transmitter and receiver system
(Noraxon USA, Scottsdale, AZ, USA) set at a sampling rate of 3
kHz, with a gain of 1000, a low-pass filter of 1 kHz, and fine-wire
adaptor leads, was used to collect EMG data. Fine-wire electrodes
(Motion Lab Systems, Baton Rouge, LA, USA) consisted of a pair
of 0.051-mm wires with 6-mm bare-wire terminations and were
inserted with a 50-mm 25-gauge needle. A single physician
(S.E.J.) skilled in the insertion of fine-wire electrodes inserted the

103
electrodes into the subjects IS and SS muscles using sterile
techniques. Proper placement and gain settings were checked by
viewing the EMG signals on the computer during submaximal
isometric contractions in the plane of the scapula for the SS and
ER for the IS. The EMG signals were transmitted to the receiver
unit of the Noraxon system, which was interfaced with the analogto-digital conversion board of the Motion Analysis system
(Motion Analysis, Santa Rosa, CA, USA), allowing synchronization of the motion and EMG signals. The subject was then fitted
with reflective markers to define the 3-dimensional positions and
orientations of the trunk, surgical-side upper arm, and forearm
during each exercise. Marker positions were recorded with an 8camera real-time Motion Analysis system. Marker positions were
sampled at 300 Hz synchronized with the EMG data, which were
resampled at 24 kHz.
A 2-second, quiet-stance calibration trial was recorded for
baseline (BL) measures of muscle activity and reference orientations of the segments. After the calibration trial, 3 repetitions of
each exercise were performed in random order under the direction
of 1 of 2 experienced physical therapists. The specific exercises are
presented in Table I and are classified into movement groups
according to common goals (ie, ER, elevation, scapular retraction,
and isometrics). Each exercise was performed for a count of 5
seconds. For all nonisometric exercises, subjects were instructed to
keep their surgical arm as relaxed as possible while moving through
a tolerable ROM that did not significantly increase their pain. For
the isometric exercises, they were instructed to push into the barrier
with a level of force that did not significantly increase pain.

Data analysis
Raw marker position data were low pass filtered (5 Hz, fourth
order, Butterworth). The 3-dimensional kinematics of the surgical
shoulder was calculated as the orientation of the upper arm relative to the trunk and expressed relative to the calibration trial. Raw
EMG signals were band pass filtered (35-500 Hz, fourth order,
Butterworth), rectified, and then low pass filtered to create a linear
envelope. Activity for each muscle was determined by the
maximal activity observed throughout the entire movement. It is
a common procedure to normalize EMG values to a standardized
contraction to account for variations in activity due to factors such
as electrode placement. Previous research in this area, using
healthy subjects, has reported EMG levels as a percentage of
maximal voluntary isometric or dynamic contractions. Because
this was not possible in the current postsurgical patient group and
given that our primary purpose was to identify exercises that elicited greater activity than found at rest, activity levels for each
repetition were reported as a percentage of the activity observed
during quiet stance at BL. Because the normalized activity levels
were relatively low and naturally bounded in one direction (below
1  BL at 0), a natural log transformation was applied to the data
before all statistical procedures. Two sets of comparisons were
made. The first was between activity levels for each exercise
compared with that at BL with 2-tailed, independent-samples t
tests, which resulted in 14 comparisons per muscle. The second
was a set of comparisons between exercises within each movement group (Table I). This was performed for each muscle by
means of analyses of variance with repeated measures (1 withinsubject factor [ie, exercise]). Significance was set at P < .05.
No adjustments were made to the criterion P value for multiple
comparisons because these would only adjust for the likelihood of

104
Table I

C.A. Murphy et al.

Exercises tested and displayed in movement groups based on the goal of each specific exercise

External rotation (supine)

Elevation

Scapular retraction

Isometric

Other

Therapist-assisted external
rotation
Self-assisted external rotation

Therapist-assisted elevation
(supine)
Self-assisted elevation (supine)
Table flexion
Pulleys

Scapular retraction with tubing

IR

BL (quiet stance)

Scapular retraction with sling

Adduction
Extension
ER
Abduction

Pendulums

making a type I error (finding a difference between an exercise

and BL when one does not exist) at the expense of an increased
chance of making a type II error (finding no difference between an
exercise and BL when one, in fact, exists). For this research
question, there are greater consequences to the patient if a type II
error is made than if a type I error is made.

Results
Figure 1 depicts the results for muscle activity of the SS
and IS during each exercise compared with BL. Of
importance, the SS had no more activity than BL during
therapist- and self-assisted ER, therapist-assisted elevation,
pendulums, and isometric IR and adduction (confidence
interval [CI] included 0, P < .05). All other exercises
studied activated the SS above that of BL (lower CI > 0,
P > .05). The IS did not remain as passive as BL during any
of the 14 exercises studied (lower CI > 0, P > .05).
Table II depicts each exercise in relation to the other
exercises in the same movement group for the SS and IS. In
the ER group, therapist assistance produced significantly
less activity than self-assistance for both muscles. In the
elevation group, pulleys produced significantly more
activity in each muscle than the other 3 exercises in that
movement group, with no difference between each of the
other 3 exercises. There was no difference between the
exercises in the scapular retraction movement group for
either muscle. In the isometric movement group, the SS
showed no significant difference between adduction and IR
or between abduction and ER. The IS showed no significant
difference between adduction and extension or between
adduction and IR. ER and abduction were significantly
greater than IR, adduction, and extension for both muscles.
The mean pain score at the time of testing was 3.1  2.3
on a scale from 0 to 10.

Discussion

Figure 1 Mean muscle activity and 95% CIs (across subjects)

during each exercise for SS (top) and IS (bottom) expressed
relative to BL activity (ie, 1 1  BL activity). Red dots indicate
exercises that were significantly greater than BL, and green dots
indicate exercises that were significantly less than BL. TAER,
therapist-assisted external rotation; SAER, self-assisted
external rotation; TAE, therapist-assisted elevation (supine); SAE,
self-assisted elevation (supine); TF, table flexion; Pul, pulleys;
Pend, pendulums; SRT, scapular retraction with tubing; SRS,
scapular retraction with sling; IR, internal rotation; ADD,
adduction; EXT, extension; ER, external rotation; Abd, abduction.

The purpose of this study was to determine the relative

amounts of EMG activity of the SS and IS during typical
early postoperative exercises in postsurgical patients. The
majority of scientific information that has been generated to
evaluate whether shoulder rehabilitation exercises are safe
to perform in the early postoperative phases of RTC

rehabilitation has been yielded from the evaluation of

young healthy subjects, the normalization of EMG levels to
varying contractions, and the labeling of activity levels as
passive based on criteria that are arguably arbitrary.20
In this study, we evaluated levels of RTC activity in early
postoperative SAD/DCR patients who were aged 53  10

EMG during passive rotator cuff exercises

Table II
group

P values resulting from analyses of variance and post hoc tests (if applicable) comparing exercises within each movement
External
rotation: SAER

SS
TAER
TAE
SAE
TF
SRT
IR
ADD
EXT
ER

.0171)

TAER
TAE
SAE
TF
SRT
IR
ADD
EXT
ER

.0318)

IS

105

Elevation
SAE

TF

Pull

.1302

>.9999
.1184

<.0001)
.0070)
<.0001)

Scapular
retraction: SRS

Isometric
ADD

.8069

.2621

.9671
.1070

EXT

ER

ABD

>.9999

.0311)
.0245)

<.0001)
<.0001)
.0006)

<.0001)
<.0001)
.0001)
.9864

.2230

<.0001)
.0660

<.0001)
<.0001)
<.0001)

<.0001)
<.0001)
.0237)
.0023)

.0001)
.0401)
<.0001)
.0486

TAER, therapist assisted external rotation; TAE, therapist assisted elevation; SAE, self-assisted elevation; TF, table flexion; SRT, scapular retraction with
tubing; IR, internal rotation; ADD, adduction; EXT, extension; ER, external rotation.
) Statistically significant.

years and were currently having postoperative shoulder pain.

These patients were chosen because there is evidence
pointing to the potential difficulty that pathologic patients
may have when attempting to remain relaxed during passive
exercises, despite evidence indicating that the exercises are
reported to minimally affect shoulder muscle activity in
healthy volunteers.4 Therefore, these results are the first, to
our knowledge, to evaluate postsurgical patients and thus
may provide a more useful assessment of passive exercises
in postsurgical patients compared with data collected from
young, healthy control subjects.
There are 2 factors that make comparing the studies in
the literature with each other, as well as with this project,
difficult: the variety of means of normalizing EMG data by
use of different standard contractions, and the resulting
thresholds chosen to indicate passive exercise. The various
standard contractions that have been used include maximal
voluntary contraction (MVC), submaximal voluntary
dynamic contraction (VDC), and MVC of the contralateral
shoulder for pathologic subjects.3,4,15,19,22 The thresholds
that have been used to define passive exercise in the
literature include less than 20% MVC/VDC,4,15,19 less than
5% MVC,3 and less than 10% MVC.22 Because these types
of maximal contractions are not safe to assess in a postoperative patient population, they do not provide functional
thresholds for clinical use. As a general rule, the standardized contraction should be reliable and consistent with

the level of muscular activity of the exercises being studied

in order to be sensitive to small differences. Without
knowing exactly how much force will damage a RCR and
with our knowledge that the critical force level will vary
among patients, we not only normalized EMG data in our
study to the activity observed during quiet stance at BL but
we also used BL activity as our criterion threshold. To put
this in perspective, quiet-stance activity has been found to
be less than 1.5% MVC in previous studies.20 This BL
comparison not only provides a valid normalization
procedure to account for variations in the EMG recording
process but also provides a valid reference activity for
passive exercises.
Pulleys, scapular retraction, isometric ER, and isometric
Abd produced more activity than BL for both the SS and IS
in our study and have likewise been considered active
exercises in the literature.3,15,19
This study has found that therapist-assisted elevation,
self-assisted ER, and isometric IR elicit no greater activity
than BL for the SS but elicit greater activity than BL for
the IS. The literature includes the described exercises in the
passive phase of RCR protocols based on differing
thresholds of muscle activity for the SS and IS.3,15
Our findings indicate that both muscles are activated to
a greater extent than BL during table flexion, yet the
literature studying healthy subjects has reported table
flexion to be passive for both muscles.22

106
For therapist-assisted ER, there is conflicting evidence in
the literature regarding the SS but there is agreement that it
is passive for the IS.3,15 The variations among studies could
be the result of some differences in methodology previously
mentioned. Our study found therapist-assisted ER to activate the IS more than BL whereas the SS was activated no
more than BL. In fact, it was the only exercise in this study
that elicited less activity than BL for either muscle,
possibly indicating that an experienced physical therapist
may be able to manually cue a patient to relax his or her SS
even more during this exercise than when the patient is
standing at rest.
Whereas some studies concur that self-assisted elevation
and pendulums are passive exercises for one or both
muscles,3,15,22 others deem them active exercises.3,4,15 Our
study found the SS to activate no more than BL during
pendulums and to activate greater than BL during selfassisted elevation. The IS was activated more than BL
during both of these exercises.
In our study, both the SS and IS showed greater values
than BL during isometric extension. McCann et al15 found
the SS to be less than 20% VDC and the IS to be greater
than 20% VDC during isometric extension in a healthy
population.
In designing an individual rehabilitation protocol, one
considers the size of the tear; tissue quality; location, type,
and chronicity of the tear; concomitant surgical procedures;
preoperative limitations; age; lifestyle; and general
health.3,7,8,12,13,16,23 In addition to taking these patientspecific factors into account, the therapist and surgeon can
use the results of this project to individually select passive
interventions based on research studying postoperative
patients rather than healthy control subjects. Further
research on postsurgical patients would improve our
knowledge on how to safely rehabilitate them.
There are several limitations to this study worth noting.
First, intra-muscular fine-wire electrodes, while optimizing
specificity, reducing muscular crosstalk, and reducing the
influence of other confounders, are limited to a small
recording area and can be influenced by motor unit recruitment ordering. In addition, exercises that incur a large ROM
can also influence the motor unitelectrode orientations. We
believe that these limitations should be minimized through
the within-subject statistical design and the use of fine-wire
electrodes instead of needle electrodes. The second limitation is that electromyography is not a direct measurement of
the potential damaging force incurred at the site of a RCR.
As a result, EMG studies cannot provide definitive guidelines on safe versus not safe exercises applicable to all
patients without assumptions regarding the force-EMG
relationships, as well as the force levels that will cause
damage to a repair. To minimize these assumptions, we have
provided very conservative estimates on the relative safety
of each of the exercises studied by comparing them with
quiet-stance BL activity. Finally, although this is the first
study to examine postoperative patients, RCR patients may

C.A. Murphy et al.

respond differently to these exercises than the SAD and DCR
patients studied.

Conclusion
Our results showed that in postsurgical SAD/DCR
patients, the SS remains as passive as quiet-stance BL in
the following 6 exercises: therapist- and self-assisted
ER, therapist-assisted elevation, pendulums, and
isometric IR and adduction. The IS does not remain as
passive as quiet-stance BL in any of the 14 exercises
studied. We examined postsurgical patients rather than
the younger, healthy control subjects examined in
previous studies, and thus, these results may provide
a more realistic pictures of postoperative patients ability
to remain as passive during specific rehabilitation exercises as they are at quiet stance BL.

Acknowledgments
We acknowledge Stephen C. Swanson for his contribution to the writing of the grant and Gregory Snow for his
feedback on the statistics.

Disclaimer
Support for this project came from a Deseret Foundation
grant.
The authors, their immediate families, and any
research foundations with which they are affiliated have
not received any financial payments or other benefits
from any commercial entity related to the subject of this
article.

References
1. Brislin KJ, Field LD, Savoie FH. Complications after arthroscopic
rotator cuff repair. Arthroscopy 2007;23:124-8. doi:10.1016/j.arthro.
2006.09.001
2. Cofield RH. Current concepts review. Rotator cuff disease of the
shoulder. J Bone Joint Surg Am 1985;67:974-9.
3. Dockery ML, Wright TW, LaStayo PC. Electromyography of the
shoulder: an analysis of passive modes of exercise. Orthopedics 1998;
21:1181-4.
4. Ellsworth AA, Mullaney M, Tyler TF, McHugh M, Nicholas SJ.
Electromyography of selected shoulder musculature during unweighted and weighted pendulum exercises. N Am J Sports Phys
Ther 2006;1:73-9.
5. Escamilla RF, Yamashiro K, Paulos L, Andrews JR. Shoulder muscle
activity and function in common shoulder rehabilitation exercises.
Sports Med 2009;39:663-85.

EMG during passive rotator cuff exercises

6. 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:
219-24.
7. Ghodadra NS, Provencher MT, Verma NN, Wilk KE, Romeo AA.
Open, mini-open, and all-arthroscopic rotator cuff repair surgery:
indications and implications for rehabilitation. J Orthop Sports Phys
Ther 2009;39:81-9. doi:10.2519/jospt.2009.2918
8. Harryman DT, Hettrich CM, Smith KL, Campbell B, Sidles JA,
Matsen FA. A prospective multipractice investigation of patients with
full thickness rotator cuff tears. J Bone Joint Surg Am 2003;85:690-6.
doi:10.1067/mse.2000.108962
9. Huberty DP, Schoolfield JD, Brady PC, Vadala AP, Arrigoni P,
Burkhart SS. Incidence and treatment of postoperative stiffness
following arthroscopic rotator cuff repair. Arthroscopy 2009;25:88090. doi:10.1016/j.arthro.2009.01.018
10. Kelly BT, Williams RJ, Cordasco FA, Backus SI, Otis JC,
Weiland DE, et al. Differential patterns of muscle activation in patients
with symptomatic and asymptomatic rotator cuff tears. J Shoulder
Elbow Surg 2005;14:165-71. doi:10.1016/j.jse.2004.06.010
11. Klintberg IH, Gunnarsson AC, Svantesson U, Styf J, Karlsson J. Early
loading in physiotherapy treatment after full-thickness rotator cuff
repair: a prospective randomized pilot-study with a two-year followup. Clin Rehabil 2009;23:622-38. doi:10.1177/0269215509102952
12. Koo SS, Burkhart SS. Rehabilitation following arthroscopic rotator
cuff repair. Clin Sports Med 2010;29:203-11. doi:10.1016/j.csm.2009.
12.001
13. Lastayo PC, Wright T, Jaffe R, Hartzel J. Continuous passive motion
after repair of the rotator cuff. J Bone Joint Surg Am 1998;80:1002-11.
14. Long JL, Thiele RA, Skendzel JG, Jeon J, Hughes RE, Miller BS,
et al. Activation of the shoulder musculature during pendulum

107

15.

16.

17.

18.

19.

20.

21.

22.

23.

exercises and light activities. J Orthop Sports Phys Ther 2010;40:2307. doi:10.2519/jospt.2010.3095
McCann PD, Wootten ME, Kadaba MP, Bigliani LU. A kinematic and
electromyographic study of shoulder rehabilitation exercises. Clin
Orthop Relat Res 1993;288:179-88.
Millett PJ, Wilcox RB, OHolleran JD, Warner JJ. Rehabilitation of
the rotator cuff: an evaluation-based approach. J Am Acad Orthop
Surg 2006;14:599-609.
Misamore GW, Ziegler D, Higginbotham G. Passive range of motion
exercises of the shoulder: an EMG analysis. J Shoulder Elbow Surg
1993:2(1)Part 2:S27.
Mochizuki T, Sugaya H, Uomizu M, Maeda K, Matsuki K, Sekiya I,
et al. Humeral insertion of the supraspinatus and infraspinatus. New
anatomical findings regarding the footprint of the rotator cuff. J Bone
Joint Surg Am 2009;91(Suppl 2, Pt 1):1-7. doi:10.2106/JBJS.H.01426
Smith J, Dahm DL, Kaufman KR, Boon AJ, Laskowski ER,
Kotajarvi BR, et al. Electromyographic activity in the immobilized
shoulder girdle musculature during scapulothoracic exercises. Arch
Phys Med Rehabil 2006;87:923-7. doi:10.1016/j.apmr.2006.03.013
Smith J, Padgett DJ, Dahm DL, Kaufman KR, Harrington SP,
Morrow DA, et al. Electromyographic activity in the immobilized
shoulder girdle musculature during contralateral upper limb movements.
J Shoulder Elbow Surg 2004;13:583-8. doi:10.1016/j.jse.2004.03.010
Thomazeau H, Boukobza E, Morcet N, Chaperon J, Langlais F.
Prediction of rotator cuff repair results by magnetic resonance
imaging. Clin Orthop Relat Res 1997;344:275-83.
Uhl TL, Muir TA, Lawson L. Electromyographical assessment of
passive, active assistive, and active shoulder rehabilitation exercises.
PM R 2010;2:132-41. doi:10.1016/j.pmrj.2010.01.002
Warner JJ, Greis PE. The treatment of stiffness of the shoulder after
repair of the rotator cuff. Instr Course Lect 1998;47:67-75.