[ research report ]

Philip McClure, PT, PhD1 • Jenna Balaicuis, DPT2 • David Heiland, DPT3
Mary Ellen Broersma, DPT4 • Cheryl K. Thorndike, DPT, ATC5 • April Wood, DPT6

A Randomized Controlled Comparison

of Stretching Procedures for Posterior
Shoulder Tightness
Downloaded from www.jospt.org at on November 19, 2021. For personal use only. No other uses without permission.

P
osterior capsule tightness of the shoulder has been suggested activities. This approximation of the hu-
as a causative or perpetuating factor in shoulder impingement meral head and acromion can lead to
compression of tissues in that region and
syndrome and labral lesions.13,15,23,25,30-32 Harryman et al14 have
may be associated with limited shoulder
shown that selective tightening of the posterior portion of the flexion, internal rotation, and horizontal
Copyright © 2007 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

shoulder capsule causes anterior and superior translation of the humeral adduction.27
head with passive shoulder flexion. The abnormal humeral head motion Warner et al28 found that patients with
can result in a decrease in the subacromial space during overhead shoulder impingement syndrome were
limited in passive internal rotation range
t Study Design: Randomized controlled trial.
of motion (ROM) compared to healthy
stretching exercises on the more limited side only,
subjects and attributed this limitation
t Objectives: To compare changes in shoulder
once daily for 5 repetitions, holding each stretch
for 30 seconds. to posterior capsular tightness. Myers et
internal rotation range of motion (ROM), for 2
t Results: The improvements in internal
al20 demonstrated that throwing athletes
stretching exercises, the “cross-body stretch” and
with symptomatic internal impingement
Journal of Orthopaedic & Sports Physical Therapy®

the “sleeper stretch,” in individuals with posterior rotation ROM for the subjects in the cross-body
shoulder tightness. stretch group (mean 6 SD, 20.0° 6 12.9°) were had reduced glenohumeral internal ro-
tation and also reduced glenohumeral
t Background: Recently, some authors have
significantly greater than for the subjects in the
control group (5.9° 6 9.4°, P = .009). The gains in adduction reflecting posterior shoulder
expressed the belief that the sleeper stretch is
the sleeper stretch group (12.4° 6 10.4°) were not tightness compared to matched asymp-
better than the cross-body stretch to address
significant compared to those of the control group
glenohumeral posterior tightness because the tomatic subjects. Burkhart et al5,7 suggest
(P = .586) and those of the cross-body stretch
scapula is stabilized.
group (P = .148). that contracture of the posterior-inferior
t Methods: Fifty-four asymptomatic subjects t Conclusions: The cross-body stretch in
glenohumeral capsule, evidenced by a
(20 males, 34 females) participated in the study. lack of internal rotation with the arm
individuals with limited shoulder internal rotation
The control group (n = 24) consisted of subjects abducted to 90°, is an essential cause
ROM appears to be more effective than no stretch-
with a between-shoulder difference in internal of superior labral lesions. This assertion
ing in controls without internal rotation asymmetry
rotation ROM of less than 10°, whereas those was based on a large series (n = 124) of
to improve shoulder internal rotation ROM. While
subjects with more than a 10° difference were
the improvement in internal rotation from the throwers with arthroscopically proven
randomly assigned to 1 of 2 intervention groups,
cross-body stretch was greater than for the sleeper superior labral lesion, all of whom dem-
the sleeper stretch group (n = 15) or the cross-
stretch and of a magnitude that could be clinically onstrated at least a 25° lack of shoulder
body stretch group (n = 15). Shoulder internal
significant, the small sample size likely precluded
rotation ROM, with the arm abducted to 90° and internal rotation compared to the non-
statistical significance between groups. J Orthop
scapula motion prevented, was measured before throwing side.5,7 The authors suggest that
Sports Phys Ther 2007;37(3):108-114. doi:10.2519/
and after a 4-week intervention period. Subjects a tightened posterior-inferior capsule
jospt.2007.2337
in the control group were asked not to engage in
t Key Words: internal rotation, shoulder,
pushes the humeral head superiorly and
any new stretching activities, while subjects in
the 2 stretching groups were asked to perform stretching, tightness posteriorly with the arm in the cocked
throwing position, and this concept has

Professor, Arcadia University, Glenside, PA. 2 Physical Therapist, Abington Hospital, Abington, PA. 3 Physical Therapist, Orthopedic and Spine Specialists, York, PA. 4 Physical
1 

Therapist, Howard Head Sports Medicine, Vail, CO. 5 Physical Therapist, Peak Sports and Spine Physical Therapy, Renton, WA. 6 Physical Therapist, PRO Sports Club Physical
Therapy, Bellevue, WA. All subjects read and signed an informed consent document approved by the Arcadia University Institutional Review Board prior to participation in the
study. Address correspondence to Philip McClure, Department of Physical Therapy, Arcadia University, 450 S Easton Rd, Glenside, PA 19038. E-mail: [email protected]

108 | march 2007 | volume 37 | number 3 | journal of orthopaedic & sports physical therapy
 been supported in a biomechanical study presumably due to posterior shoulder between group sizes. Therefore the final
using a cadaver model.13 tightness. sample consisted of 54 subjects (20 males,
Morrison et al17 suggest that adequate 34 females). Exclusion criteria consisted
flexibility of the posterior capsule is im- METHODS of a history of shoulder surgery, shoulder
portant prior to beginning a strengthen- symptoms requiring medical care within

W
ing program. Several different methods e used a randomized design the past year, or shoulder pain greater
of stretching have been described to ad- to compare 2 posterior shoulder- than 5 out of 10 using a numerical pain
dress posterior shoulder tightness. These stretching techniques performed scale. Detailed characteristics of each
include the “towel stretch,” where the gle- for 4 weeks in subjects with unilateral subject group are given in Table 1. All sub-
nohumeral joint is adducted, internally posterior shoulder tightness. We com- jects read and signed an informed con-
rotated, and extended, while the hand pared these groups to a nontreated con- sent document approved by the Arcadia
now located behind the individual’s back trol group without unilateral tightness. University Institutional Review Board
is pulled up by the opposite hand using a prior to participation in the study.
Downloaded from www.jospt.org at on November 19, 2021. For personal use only. No other uses without permission.

towel.3,16 Another popular stretch is the Subjects

“cross-body stretch,” where the shoulder From a convenience sample of college Measurements Procedures
is elevated to approximately 90° of flex- students, 83 individuals were measured All measurements were performed by 1
ion and pulled across the body into hori- to identify 30 with a 10° (right versus of 2 testers who were blind to treatment
zontal adduction with the opposite arm.3 left) asymmetry in shoulder internal ro- group. The same tester performed both
Both of these stretching procedures have tation measured at 90° abduction. These pretest and posttest measurements on
Copyright © 2007 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

been criticized because the scapula is 30 subjects with a 10° or greater differ- a given subject. Both testers established
not stabilized and therefore tissue stress ence were, after stratification, randomly intrarater reliability on a group of 15
is imparted to scapulothoracic tissues assigned to 1 of 2 intervention groups: asymptomatic subjects (30 shoulders)
as well as tissues crossing the glenohu- the sleeper stretch group (n = 15) or the by repeating measurements at least 1
meral joint. More recently, authors2,5,29 cross-body stretch group (n = 15). Sub- day apart. Separate reliability coeffi-
have described a “sleeper stretch” that is jects were first stratified based on gender cients (intraclass correlation coefficients
accomplished by lying on the side to be and involvement in overhead sports be- [ICC3,1] and standard error of measure-
stretched, elevating the humerus to 90° cause these factors are believed to influ- ment [SEM]) were established for each
on the support surface, then passively ence shoulder ROM. The control group rater and each side for each measurement
Journal of Orthopaedic & Sports Physical Therapy®

internally rotating the shoulder with (n = 24) consisted of subjects with a be- (Table 2). Prior to range of motion testing,
the opposite arm. Other authors31 have tween-shoulder difference of less than subjects were asked to warm up by per-
described methods where the scapula is 10° of internal rotation measured at 90° forming 3 active, bilateral shoulder flex-
manually stabilized by the therapist while of abduction. After the initial 24 control ion stretches with hands clasped, holding
the arm is adducted or internally rotated. subjects without a significant asymme- each for 10 seconds.
This manual approach has the obvious try were identified, only subjects with Our primary measure of posterior
disadvantage of requiring a therapist or asymmetry were invited to participate shoulder tightness was passive internal
second person to perform the stretch, in the study to avoid excessive imbalance rotation of the glenohumeral joint with
which limits how often the stretch can be
performed.
TABLE 1 Subject Characteristics
Despite the evidence from biome-
chanical studies suggesting that posterior
Control Sleeper Stretch Cross-Body
shoulder tightness may be a contribut-
(n = 24) (n = 15) Stretch (n = 15)
ing factor to subacromial impingement
Gender (male, female) (n) 10, 14 6, 9 4, 11
and the recommendation of authors for
Engage in overhead sports (n) 5 3 3
prophylactic stretching, we could find
Age (y)* 23.5 (1.8) 23.5 (1.7) 22.9 (1.5)
no studies comparing the effectiveness
Height (cm)* 172.2 (10.1) 168.7 (9.2) 167.5 (8.3)
of these stretching procedures for pos-
Body mass (kg)* 71.1 (15.0) 70.3 (12.2) 68.8 (12.6)
terior shoulder tightness. Therefore, the
Dominant arm (right, left) (n) 24, 0 13, 2 12, 3
purpose of this study was to compare the
Side stretched (right, left) (n) 7, 8 7, 8
sleeper stretch and cross-body stretch
Compliance (%)* 81.0 (16.0) 88.9 (10.0)
techniques to improve passive shoulder
*Values represented as mean (SD). No significant differences were found for age, height, and body mass
internal rotation ROM in subjects with among groups.
limited shoulder internal rotation ROM

journal of orthopaedic & sports physical therapy | volume 37 | number 3 | march 2007 | 109
 [ research report ]
sure was normalized and expressed as a
Intraclass Correlation Coefficients (ICC)
percentage of spine length based on the
TABLE 2 and Standard Error of Measurements (SEM)
following formula: TUB = (distance from
for Shoulder Rotation
T1 to IC – distance from thumb to T1)/dis-
tance from T1 to IC.
Rater 1 Rater 2
Measurement ICC3,1 SEM* ICC3,1 SEM*
Stretching Intervention
IR90 left 0.84 4.8 0.80 5.4
After the initial measurements, the inter-
IR90 right 0.82 4.3 0.75 5.6
vention team assigned subjects to the con-
ER90 left 0.88 4.4 0.89 4.6
trol group if the IR90 difference between
ER90 right 0.78 4.8 0.86 3.7
sides was less than 10°, and randomly
Total Rot left 0.98 3.4 0.96 4.8
assigned subjects to 1 of the 2 stretching
Total Rot right 0.91 6.5 0.89 6.7
groups if the difference between sides was
Downloaded from www.jospt.org at on November 19, 2021. For personal use only. No other uses without permission.

TUB left 0.96 0.03 0.91 0.04

greater than 10°. One group performed
TUB right 0.90 0.04 0.96 0.03
the sleeper stretch by lying on the side
Abbreviations: ER90, external rotation at 90° abduction; IR90, internal rotation at 90° abduction;
Total Rot, total rotation calculated as the sum of IR90 and ER90; TUB, thumb up the back.
to be stretched, elevating the humerus to
* All values represent degrees except TUB measures, which are percent of total spine length. 90° on the support surface, then passively
internally rotating the humerus with the
the arm abducted to 90° in the frontal We also measured the ability to actively opposite arm (Figure 3A). The other group
Copyright © 2007 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

plane (IR90) as shown in Figure 1. The move the thumb up the back (TUB), as performed a cross-body stretch by pas-
inclinometer was placed on the dorsal shown in Figure 2. Rather than using ver- sively pulling the humerus across the body
surface of the forearm with the elbow tebral level as an end point, which may into horizontal adduction with the oppo-
flexed to 90°. We were careful to prevent be unreliable,10 we quantified the mea- site arm (Figure 3B). Subjects in the con-
scapular substitution by watching the an- sure according to spine length. Prior to trol group were asked not to engage in any
terior aspect of the shoulder during the asking the subject to move the thumb new stretching activities while subjects
measurement. Accordingly, the end point up the back, we measured the length of in the 2 stretching groups were asked to
for IR90 measurement was the angle just the spine from the first thoracic spinous perform stretching exercises to a point of
prior to the anterior aspect of the shoul- process (T1) to the level of the iliac crests mild discomfort, on the more limited side
Journal of Orthopaedic & Sports Physical Therapy®

der moving anteriorly, indicating scapular (IC) at midline, which were determined only, once daily for 5 repetitions, holding
motion.1 Shoulder external rotation was by palpation and marked. The TUB mea- each stretch for 30 seconds.
measured in the same fashion, with the Subjects in both stretching groups were
arm abducted to 90° in the frontal plane shown their assigned exercise, which they
(ER90). Total rotation was calculated were asked to demonstrate. They were also
by adding the IR90 and ER90 values. given a sheet with written instructions and
a picture of the stretch to be performed.
All subjects in the stretching groups were
given a daily log to be completed to reflect
exercise compliance and were encouraged
Total to fill them out accurately rather than
spine
overestimating compliance to please the
length
investigators. All subjects were also con-
TUB %
tacted at 2 weeks for encouragement and
to schedule the final test session.
At the final test session all measure-
ments were taken again by the same tes-
ter, who took the original measures and
was blind to treatment group assignment.
FIGURE 1. Measurement of passive internal rotation FIGURE 2. Thumb-up-the-back (TUB) measurement. Compliance logs were collected and per-
with the shoulder abducted 90° in the frontal plane The total length of the spine from the first thoracic spi- centage compliance was computed based
(IR90). The examiner was careful to stop the movement nous process to the level of the iliac crests at midline was on the number of days the subjects com-
at the point where the anterior aspect of the shoulder determined by palpation and marked. The TUB measure pleted the daily stretching program. Sub-
(dotted line) was observed to move anteriorly. is expressed as a percentage of total spine length.
jects in the stretching groups completed a

110 | march 2007 | volume 37 | number 3 | journal of orthopaedic & sports physical therapy
 A RESULTS 35
*†
30

N
25
o significant differences were †

IR (deg)
20
found for age, height, and body 15
mass among groups (Table 1). Only 10
1 out 5 subjects in each group was actively 5
involved in sports requiring overhead use 0
Control Group Sleeper Stretch Cross-Body
of the arm. The 2 stretching groups were Group Stretch Group
not perfectly balanced on gender due to Stretch Side Control Side
B
an error in the stratification process. All FIGURE 4. Changes in IR90 over the 4-week stretch-
dependent variables met the assumptions ing period for all groups. Positive changes represent
for ANOVA testing. The values for all de- increased motion. Control group values represent
pendent variables are shown in Table 3 the average of both right and left sides. Error bars
Downloaded from www.jospt.org at on November 19, 2021. For personal use only. No other uses without permission.

represent standard deviation. *Significant change on

and the change scores for IR90 are shown
stretch side compared to control group (P,.01); †Sig-
graphically in Figure 4. For IR90, the nificant change compared to control side (P,.001).
group-by-time ANOVA revealed a signifi-
cant interaction (P,.001). The cross-body cant increase (P = .028). The gain in TUB
stretch group improved significantly more for the sleeper stretch group (2.7%) was
(mean 6 SD, 20.0° 6 13.0°) than the con- also statistically significant relative to the
Copyright © 2007 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

FIGURE 3. (A) Sleeper stretch, (B) cross-body stretch. trol group (mean 6 SD, 5.8° 6 8.5°, P = control side (P = .048), while the small
The subjects in the 2 stretching groups were asked .009). The gains in sleeper stretch group change in the cross-body stretch group
to perform stretching exercises to a point of mild (mean 6 SD, 12.4° 6 11.9°) were not sig- (1.6%) was not significant compared to
discomfort, on the more limited side only, once daily
nificant compared to controls (P = .586) the control side (Table 3).
for 5 repetitions, holding each stretch for 30 seconds.
and the differences (7.6°) between the
brief form which asked about their subjec- cross-body stretch group and the sleeper DISCUSSION
tive response to the stretching program. stretch group were not statistically signif-

A
icant (P = .148). Both stretching groups lthough both stretch groups
Data Analysis showed a significant increase in IR90 on showed increases in IR90 com-
Journal of Orthopaedic & Sports Physical Therapy®

Descriptive statistics were calculated for the stretched side compared to the non- pared to the nonstretched side, the
all variables and all dependent variables stretched control side (P,.001), but there cross-body stretch appeared to be more
were checked for normal distribution and was no side-by-group interaction, indicat- effective and showed the only significant
homogeneity of variance assumptions. ing that relative to the nonstretched side increase compared to the control group.
Change scores were also calculated for there was no difference between methods This finding is somewhat surprising,
all ROM variables by subtracting pretest of stretching. given that stabilization of the scapula
values from posttest values. To determine There were no significant differences as performed with the sleeper stretch
equivalency among groups, differenc- between the pretest and posttest ER90 would seem to enhance the effectiveness
es for subject characteristics were as- measures in the 3 groups, nor were there of stretching for the posterior shoulder
sessed using a 1-way analysis of variance significant differences between the treat- region. Average self-reported compliance
(ANOVA). To compare the effect of the ed and nontreated arms (Table 3). Total for the cross-body stretch group was 89%
stretching exercises among the 3 groups, rotation showed a significant group-by- compared to 81% for the sleeper stretch
a mixed 2-way analysis of variance (group time interaction (P,.001) and follow-up group, which was not statistically differ-
by time) was performed. If a group-by- testing revealed a significant difference ent. Three out of 15 subjects in the sleeper
time interaction was found, a follow-up between controls and the cross-body stretch group complained that the stretch
between-group 1-way analysis of variance stretch group (P = .026), while other dif- itself was painful, whereas only 1 subject
was performed using change scores as the ferences between groups were not signif- in the cross-body stretch group reported
dependent variable and Tukey post hoc icant. Both stretching groups showed a pain during stretching, which she at-
tests for comparisons among groups. To significant increase in total rotation rela- tributed to a minor injury and not to the
compare between sides (stretched or not tive to the nonstretched side (Table 3). For stretch itself. One subject in each stretch-
stretched) in the 2 stretching groups, a the TUB measure, there was a significant ing group reported new symptoms during
mixed 2-way analysis of variance (group main effect of time (P = .044) and follow- the stretching period, but neither could
by side) was performed on the change up analysis revealed that only the sleeper attribute the symptoms to a particular ac-
scores for each side and group. stretch group made a statistically signifi- tivity. One subject in the sleeper group re-

journal of orthopaedic & sports physical therapy | volume 37 | number 3 | march 2007 | 111
 [ research report ]
Range of Motion Before and After a 4-Week Stretching Program Using
TABLE 3
Either the Sleeper Stretch or the Cross-Body Stretch*

Sleeper (n = 15) Cross-Body (n = 15) Controls (n = 24)†

Measures Preintervention Postintervention Preintervention Postintervention Preintervention Postintervention
IR90, stretch side 48.2 6 8.8 60.6 6 10.4§ 46.6 6 11.5 66.6 6 15.9‡§
52.5 6 9.5 58.3 6 8.8
IR90, control side 61.8 6 9.5 64.8 6 12.8 60.5 6 10.6 68.6 6 13.7
ER90, stretch side 95.1 6 14.5 96.1 6 12.7 99.1 6 9.1 97.5 6 8.0
92.7 6 16.6 93.0 6 15.2
ER90, control side 93.0 6 11.4 95.6 6 12.4 98.2 6 11.0 99.4 6 7.5
Total Rot, stretch side 143.7 6 17.9 156.9 6 19.7§ 145.9 6 17.1 164.3 6 21.2‡§
145.3 6 22.9 151.7 6 19.9
Total Rot, control side 155.0 6 17.0 160.6 6 18.1 158.9 6 17.1 168.3 6 17.2
TUB, stretch side 66.5 6 10.8 69.2 6 9.2§ 68.9 6 13.0 70.5 6 11.3
67.5 6 13.8 68.7 6 13.9
Downloaded from www.jospt.org at on November 19, 2021. For personal use only. No other uses without permission.

TUB, control side 67.2 6 15.5 66.6 6 13.1 71.4 6 11.5 69.8 6 10.8
Abbreviations: ER90, external rotation at 90° abduction; IR90, internal rotation at 90° abduction; Total Rot, total rotation calculated as the sum of IR90 and
ER90; TUB, thumb up the back, values represent percent of total spine length with higher numbers indicating the thumb was higher up the back.
* All values represent mean 6 SD degrees except TUB measures, which are percent of total spine length.
†
Control subjects did not stretch either side, therefore the average of left and right measurements were used.
‡
Change preintervention-postintervention significantly different compared to control group (P,.05).
§
Change preintervention-postintervention significantly different compared to control side (P,.05).
Copyright © 2007 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

ported that the stretch was inconvenient viation found in our sample, the power to glenohumeral joint positions in cadavers
because the position required prevented detect a difference between the stretching and found that the supraspinatus under-
simultaneous reading, while all other groups was only 0.42. To obtain a power went the greatest strain with the humerus
subjects reported that they had adequate of at least 0.8 to detect an 8° difference by the side and maximally extended simi-
time to complete the stretching exercise. between groups would have required lar to the TUB position. However, the in-
Four subjects in the sleeper stretch group a sample size of 36 subjects per group. ferior fibers of the infraspinatus muscle
reported increasing an exercise workout Therefore, the failure to find a statistically were most elongated in 60° of glenohu-
(1 aerobically, 1 more general stretching, significant difference between the experi- meral joint elevation (simulating 90°
2 more intense strengthening) during the mental groups could be attributable to an humerothoracic elevation) and internal
Journal of Orthopaedic & Sports Physical Therapy®

stretching period. Only 1 subject in the inadequate sample size. rotation, similar to the sleeper stretch. In
cross-body stretch group changed his/ Despite clear gains in the IR90 mea- this study, relationship between pretest
her exercise routine by swimming more sure, only minimal changes were ob- IR90 and TUB measures using a Pearson
regularly. Based on this self-report data, served in the TUB measure and only in correlation coefficient was r = 0.52 on the
it is possible that subjects in the sleeper the sleeper stretch group (2.7% spine stretch side and r = 0.36 on the control
stretch group performed the stretch less length change, or 1.1 cm). We believe side, for an average r value of 0.44. This
intensely and for less time because of pain there is an anatomic explanation for this indicates only a weak to moderate rela-
or the inconvenient position required. finding. Both stretching procedures were tionship between the 2 measures and
Small differences in IR90 motion performed with the arm elevated to 90°. suggests that they are capturing differ-
were observed in controls as well as in Cadaver studies have shown this position ent factors related to internal rotation
the nonstretched shoulders of the stretch- stresses the posterior-inferior aspect of ROM. Similar to differential findings for
ing groups. Measurement error or learn- the glenohumeral joint capsule.11,12 The the anterior-superior and anterior-infe-
ing effects are most likely responsible TUB measure assesses internal rotation rior capsule based on arm elevation,24 we
for the differences between pretest and with the arm by the side that stresses the speculate that the posterior-inferior and
posttest scores. Despite these potential posterior-superior capsule, while keep- posterior-superior aspects of the capsule
influences, clear differences in IR90 at- ing the inferior capsule relatively slack. and rotator cuff muscles are stressed
tributable to the stretching were observed Therefore, we believe that the TUB mea- differently based on humeral elevation,
between the cross-body stretch and con- sure is more reflective of the length of the therefore require different measures to
trol groups and between sides for both posterior-superior capsule, which was assess their length and different stretch
stretching groups. The difference in IR90 likely not stretched to a significant degree procedures to induce ROM changes.
gains between the 2 experimental groups with our stretching procedures because of One limitation to our study was the
was about 8°, which could be considered the elevated position of the arm. Muraki use of asymptomatic students rather
a clinically meaningful difference. Using, et al19 studied the strain within the poste- than throwing athletes or a symptom-
this mean difference and the standard de- rior rotator cuff muscles during extreme atic clinical population seeking medical

112 | march 2007 | volume 37 | number 3 | journal of orthopaedic & sports physical therapy
 care. Based on our reliability data, a 90% group for whom these stretching exercises pendent t test comparing IR90 changes
confidence interval for the IR90 SEM could be indicated. Several authors have in males versus females in the stretching
was computed to be 68.1°; therefore, we suggested that glenohumeral joint poste- groups and found no differences based on
considered a 10° asymmetry a meaningful rior capsule tightness, as demonstrated gender (P = .22). Similarly we performed
difference. The average between-side dif- by a lack of shoulder internal rotation, an independent t test comparing IR90
ference in our experimental subjects was may produce superior translation of the changes between those who stretched
almost 14°. We could not say with any cer- humeral head and therefore predispose the dominant side (arm dominance de-
tainty whether the side with lesser motion to subacromial impingement.7,13,14,31 We termined by self-report on an intake
was lacking flexibility or if the side with believe the changes in ROM we found questionnaire) compared with those who
greater motion was showing excessive can be logically attributed to changes in stretched the nondominant side. Again,
flexibility. At pretest, the average IR90 for the posterior glenohumeral joint capsule, no differences were found based on which
control subjects was 52.5° compared to an periarticular tissue, and posterior cuff side was stretched (P = .90).
average of 47.4° and 61.1° in the less mo- muscles. It is difficult to judge whether A limitation of this work is the lack of
Downloaded from www.jospt.org at on November 19, 2021. For personal use only. No other uses without permission.

bile and more mobile sides, respectively, high-level throwing athletes or a symp- any long-term follow-up. It seems unlike-
in the 30 experimental subjects. Selection tomatic patient sample would respond ly that the changes induced would remain
of the less mobile side for stretching was similarly. High-level throwing athletes without some ongoing end range tensile
based on the perceived greater opportu- may have bony changes, such as excessive stress in the form of stretching. It would be
nity for improvement of ROM, which was humeral retroversion, that may limit a re- helpful to know if a “maintenance dose” of
the intent of the 2 stretching procedures. sponse to a stretching intervention.9,21,22 stretching would be required to maintain
Copyright © 2007 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

Tyler et al25 found IR90 differences be- Symptomatic patients may be limited the increased motion achieved in 4 weeks
tween control subjects and subjects with primarily by pain rather than short- and, if so, what dose. Likewise, it would be
dominant-side impingement of about 9° ened periarticular tissue and therefore helpful to know if gains in motion would
on the dominant side and 7° on the non- could respond more dramatically if pain plateau with a standard, minimal stretch-
dominant side.25 These differences were subsided concurrent with a stretching ing program and, if so, when that plateau
less if the impingement symptoms were in program. Alternatively, pain could pre- would occur. Based on the Physical Stress
the nondominant shoulder. They also re- vent adequate end range stretching and Theory proposed by Mueller and Maluf,18
ported an average side-to-side difference therefore limit the effect of the stretch on increasing gains would likely require in-
of about 22° in subjects with dominant periarticular tissue and ROM. creasing levels of end range stress either
Journal of Orthopaedic & Sports Physical Therapy®

side impingement, compared to only a 5° Because we did not allow scapular sub- by increasing intensity, frequency, or du-
difference between sides in subjects with stitution with the IR90 measure, changes ration. These questions related to the time
nondominant side impingement. Myers et in ROM must be secondary to changes in course of gains in motion and the required
al20 found an average between-side differ- tissues crossing the glenohumeral joint. dosage of end range stress are clearly im-
ence in shoulder internal rotation of 11° Both the posterior joint capsule and the portant to clinical practice and worthy of
in asymptomatic throwers, but an average posterior rotator cuff muscles are ori- further study.
20° difference between sides in throwers ented such that they would limit IR90.
with symptomatic internal impingement. In this study, it is not possible to deter- CONCLUSION
Therefore, the differences between sides mine which of these tissues, or if both,

B
in our experimental subjects seemed to be are responsible for the initial difference ased on our findings, the cross-
somewhere between normal asymmetry and subsequent gains in motion. Both body stretch appears to be more ef-
found in throwers and asymmetry found muscles and periarticular connective tis- fective than no stretching in control
in subjects with shoulder injury. sues allow length changes with adequate subjects without internal rotation asym-
A unilateral lack of shoulder internal tensile stress18 and both are stressed metry. While the improvement in inter-
rotation ROM has been found previously with the stretching procedures used in nal rotation from the cross-body stretch
in throwing athletes and patients seek- this study. Increased IR90 motion has was greater than from the sleeper stretch
ing care for shoulder impingement syn- been noted following surgical release of and of a magnitude that could be clini-
drome.1,4,8,16,17,26 In the current study, out the posterior capsule, which implies this cally significant, the small sample size
of 54 subjects, 50 reported being athleti- may be the primary source of limitation, likely precluded statistical significance
cally active or exercising regularly, but at least in patients who require surgical between groups. These findings were in
only 11 were engaged in overhead sports. release.23,27 a group of asymptomatic recreational
Because stretching is a common preven- To determine the effect of gender and athletes and further study is warranted
tative measure, particularly in athletes, hand dominance on response to stretch- in higher-level throwing athletes as well
we believe our sample reflects a relevant ing, we secondarily performed an inde- as in patients with symptoms. t

journal of orthopaedic & sports physical therapy | volume 37 | number 3 | march 2007 | 113
 [ research report ]
rhaphy on the passive range of motion of the and its relationship to glenohumeral rotation in
references glenohumeral joint. J Bone Joint Surg Am. the shoulder of college baseball players. Am J
2003;85-A:48-55. Sports Med. 2002;30:354-360.
1. A wan R, Smith J, Boon AJ. Measuring shoulder 13. Grossman MG, Tibone JE, McGarry MH, Sch- 23. Ticker JB, Beim GM, Warner JJ. Recognition and
internal rotation range of motion: a compari- neider DJ, Veneziani S, Lee TQ. A cadaveric treatment of refractory posterior capsular contrac-
son of 3 techniques. Arch Phys Med Rehabil. model of the throwing shoulder: a possible ture of the shoulder. Arthroscopy. 2000;16:27-34.
2002;83:1229-1234. etiology of superior labrum anterior-to-posterior 24. Turkel SJ, Panio MW, Marshall JL, Girgis FG.
2. Bach HG, Goldberg BA. Posterior capsular lesions. J Bone Joint Surg Am. 2005;87:824-831. Stabilizing mechanisms preventing anterior dis-
contracture of the shoulder. J Am Acad Orthop 14. Harryman DT, 2nd, Sidles JA, Clark JM, Mc- location of the glenohumeral joint. J Bone Joint
Surg. 2006;14:265-277. Quade KJ, Gibb TD, Matsen FA, 3rd. Translation Surg Am. 1981;63:1208-1217.
3. Bandy WD, Sanders B. Therapeutic Exercise: of the humeral head on the glenoid with passive 25. Tyler TF, Nicholas SJ, Roy T, Gleim GW. Quanti-
Techniques for Intervention. Baltimore, MD: Lip- glenohumeral motion. J Bone Joint Surg Am. fication of posterior capsule tightness and mo-
pincott Williams & Wilkins; 2001. 1990;72:1334-1343. tion loss in patients with shoulder impingement.
4. Bigliani LU, Codd TP, Connor PM, Levine WN, 15. Ludewig PM, Cook TM. Translations of the Am J Sports Med. 2000;28:668-673.
Littlefield MA, Hershon SJ. Shoulder motion and humerus in persons with shoulder impinge- 26. Tyler TF, Roy T, Nicholas SJ, Gleim GW. Reliability
laxity in the professional baseball player. Am J ment symptoms. J Orthop Sports Phys Ther. and validity of a new method of measuring pos-
Downloaded from www.jospt.org at on November 19, 2021. For personal use only. No other uses without permission.

Sports Med. 1997;25:609-613. 2002;32:248-259. terior shoulder tightness. J Orthop Sports Phys
5. Burkhart SS, Morgan CD, Kibler WB. The dis- 16. McClure PW, Bialker J, Neff N, Williams G, Kar- Ther. 1999;29:262-269; discussion 270-264.
abled throwing shoulder: spectrum of pathology duna A. Shoulder function and 3-dimensional ki- 27. Warner JJ, Allen AA, Marks PH, Wong P. Ar-
Part I: pathoanatomy and biomechanics. Ar- nematics in people with shoulder impingement throscopic release of postoperative capsular
throscopy. 2003;19:404-420. syndrome before and after a 6-week exercise contracture of the shoulder. J Bone Joint Surg
6. Burkhart SS, Morgan CD, Kibler WB. The dis- program. Phys Ther. 2004;84:832-848. Am. 1997;79:1151-1158.
abled throwing shoulder: spectrum of pathology 17. Morrison DS, Frogameni AD, Woodworth P. 28. Warner JJ, Micheli LJ, Arslanian LE, Kennedy
Non-operative treatment of subacromial im- J, Kennedy R. Patterns of flexibility, laxity, and
Copyright © 2007 Journal of Orthopaedic & Sports Physical Therapy®. All rights reserved.

Part III: The SICK scapula, scapular dyskinesis,

the kinetic chain, and rehabilitation. Arthros- pingement syndrome. J Bone Joint Surg Am. strength in normal shoulders and shoulders
copy. 2003;19:641-661. 1997;79:732-737. with instability and impingement. Am J Sports
7. Burkhart SS, Morgan CD, Kibler WB. The dis- 18. Mueller MJ, Maluf KS. Tissue adaptation Med. 1990;18:366-375.
abled throwing shoulder: spectrum of pathology. to physical stress: a proposed “Physical 29. Weldon EJ, 3rd, Richardson AB. Upper extrem-
Part II: evaluation and treatment of SLAP lesions Stress Theory” to guide physical therapist ity overuse injuries in swimming. A discussion
in throwers. Arthroscopy. 2003;19:531-539. practice, education, and research. Phys Ther. of swimmer’s shoulder. Clin Sports Med.
8. Burkhart SS, Morgan CD, Kibler WB. Shoulder 2002;82:383-403. 2001;20:423-438.
injuries in overhead athletes. The “dead arm” 19. Muraki T, Aoki M, Uchiyama E, Murakami G, Miya- 30. Wilk KE, Arrigo C. Current concepts in the
revisited. Clin Sports Med. 2000;19:125-158. moto S. The effect of arm position on stretching rehabilitation of the athletic shoulder. J Orthop
9. Crockett HC, Gross LB, Wilk KE, et al. Osseous of the supraspinatus, infraspinatus, and posterior Sports Phys Ther. 1993;18:365-378.
adaptation and range of motion at the glenohu- portion of deltoid muscles: a cadaveric study. Clin 31. Wilk KE, Meister K, Andrews JR. Current con-
Journal of Orthopaedic & Sports Physical Therapy®

meral joint in professional baseball pitchers. Am Biomech (Bristol, Avon). 2006;21:474-480. cepts in the rehabilitation of the overhead throw-
J Sports Med. 2002;30:20-26. 20. Myers JB, Laudner KG, Pasquale MR, Bradley ing athlete. Am J Sports Med. 2002;30:136-151.
10. Edwards TB, Bostick RD, Greene CC, Baratta JP, Lephart SM. Glenohumeral range of motion 32. Wilk KE, Reinold MM, Dugas JR, Arrigo CA,
RV, Drez D. Interobserver and intraobserver reli- deficits and posterior shoulder tightness in Moser MW, Andrews JR. Current concepts in
ability of the measurement of shoulder internal throwers with pathologic internal impingement. the recognition and treatment of superior labral
rotation by vertebral level. J Shoulder Elbow Am J Sports Med. 2006;34:385-391. (SLAP) lesions. J Orthop Sports Phys Ther.
Surg. 2002;11:40-42. 21. Osbahr DC, Cannon DL, Speer KP. Retrover- 2005;35:273-291.
11. Gagey OJ, Boisrenoult P. Shoulder capsule sion of the humerus in the throwing shoulder
shrinkage and consequences on shoulder move- of college baseball pitchers. Am J Sports Med.

@
ments. Clin Orthop Relat Res. 2004;218-222. 2002;30:347-353.
12. Gerber C, Werner CM, Macy JC, Jacob HA, 22. Reagan KM, Meister K, Horodyski MB, Werner more information
Nyffeler RW. Effect of selective capsulor- DW, Carruthers C, Wilk K. Humeral retroversion www.jospt.org

114 | march 2007 | volume 37 | number 3 | journal of orthopaedic & sports physical therapy