Shoulder Kinematics in Subjects With Frozen Shoulder


13 Νοε 2013 (πριν από 3 χρόνια και 4 μήνες)

85 εμφανίσεις

Shoulder Kinematics in Subjects With Frozen Shoulder
Peter J.Rundquist,PT,PhD,Donald D.Anderson,PhD,Carlos A.Guanche,MD,Paula M.Ludewig,PhD,PT
ABSTRACT.Rundquist PJ,Anderson DD,Guanche CA,
Ludewig PM.Shoulder kinematics in subjects with frozen
shoulder.Arch Phys Med Rehabil 2003;84:1473-9.
Objectives:To describe 3-dimensional humeral motion in
subjects with frozen shoulder and to determine whether a
consistent capsular pattern of restriction was present.
Design:Descriptive study including repeated measurements
of shoulder kinematics.
Setting:Motion-analysis laboratory.
Participants:Ten (9 women,1 man) volunteers with a diag-
nosis of idiopathic adhesive capsulitis and 10 (9 women,1 man)
subjects with asymptomatic shoulders as comparison subjects.
Interventions:Not applicable.
Main Outcome Measures:Electromagnetic tracking sen-
sors monitored the 3-dimensional position of the trunk,scap-
ula,and humerus throughout active shoulder motions.Peak
humeral positions relative to the trunk and scapula were deter-
mined for shoulder ßexion,abduction,scapular plane abduc-
tion,external rotation (ER),and internal rotation (IR).Descrip-
tive statistics (means,standard deviations,percentage of
normal) were calculated and capsular patterns described.
Results:For humeral position relative to the trunk,subjectsÕ
mean peak motion was as follows:abduction,98.4¡;ER at the
side,4.5¡;ER with the armabducted,33.5¡;ßexion,116.9¡;IR
at the side,54.3¡;IR with the arm abducted,17.8¡;and scap-
ular plane abduction,113.4¡.For humeral position relative to
the scapula,subjectsÕ mean peak motion was as follows:ab-
duction,46.4¡;ER at the side,34.7¡;ER with the arm ab-
ducted,45.3¡;ßexion,70.5¡;IR at the side,10.3¡;IR with the
arm abducted,￿6.4¡;and scapular plane abduction,61.7¡.
Conclusions:Symptomatic subjects demonstrated substan-
tial kinematic deÞcits during humeral range of motion.No
single capsular pattern emerged.
Key Words:Adhesive capsulitis;Biomechanics;Kinemat-
ics;Range of motion,articular;Rehabilitation;Shoulder joint.
© 2003 by the American Congress of Rehabilitation Medi-
cine and the American Academy of Physical Medicine and
HE SIGNS AND SYMPTOMS of frozen shoulder have
been recognized since 1872.
However,to this day,more
questions than answers exist regarding this condition.The
difÞculty in studying frozen shoulder was initially acknowl-
edged by Codman in 1934:ÒThis is a class of cases which I Þnd
it difÞcult to deÞne,difÞcult to treat and difÞcult to
This observation is still true today.
A deÞnition of frozen shoulder was presented by Reeves
in 1975,who has called it Òan idiopathic condition of the
shoulder characterized by the spontaneous onset of pain in
the shoulder with restriction of movement in every
proposed dividing frozen shoul-
der into primary and secondary types.He stated that primary
frozen shoulder presents as an idiopathic decreased range of
motion (ROM) in which no systemic diagnosis,precipitating
shoulder condition,or radiographic explanation can be
found.Secondary frozen shoulders are believed to result
from a predisposing condition.
Primary frozen shoulder
was the focus of the present investigation.
was the Þrst author to address the progression of
frozen shoulder.He observed 49 cases for up to 10 years.He
documented 3 phases:pain for 2
to 9 months,stiffness for 4
to 12 months,and recovery for 12 to 42 months.Phase I was
identiÞed by pain and a decrease in capsular volume,phase II
by stiffness and discomfort,and phase III by gradual recovery.
Several others have proposed theories on frozen shoulder pro-
gression.The primary difference presented was in the time
frame for each phase.
The clinical diagnosis of frozen shoulder is based on pa-
tientsÕ subjective history and cliniciansÕ objective Þndings.
Most authors
cite a decrease in motion of at least 3 months
in duration as 1 diagnostic criterion.Quantifying the ROM
diagnostic of frozen shoulder has not been consistent.Maximal
abduction and maximal external rotation (ER) reported vary
Rotation values have also been
described as percentages of normal,with maximal ER ranging
from 50% to 60%.
Reported maximal internal rota-
tion (IR) percentages range from 45% to 50%.
The pathogenesis of primary frozen shoulder is unknown.In
was the Þrst to attempt to implicate shoulder
capsule adhesions as the etiology of frozen shoulder.Since
then,several others
have agreed with his proposal.
More recently,arthrography and arthroscopy have been used to
investigate the involved tissue(s).Adhesive capsulitis,
loss of
dependent fold,
decreased capsular volume,
capsular contractions
have been demonstrated.Addition-
ally,contracture of the coracohumeral ligament,
of the subacromial bursa,rotator interval thickening and Þbro-
sis,and capsular and intraarticular subscapularis tendon thick-
ening have all been reported.
initially proposed that tightness in a joint capsule
would result in a pattern of proportional motion restriction.He
used the concept of a capsular pattern to differentiate in diag-
nosis between loss of motion secondary to bony and/or muscle
or joint changes and that caused by the capsule.He believed
that an irritated capsule would restrict motion in a predictable
pattern.For the shoulder,he proposed that ER would be more
limited than abduction,which would be more limited than IR.
Alternatively,different areas of capsular adhesions (superior,
may result in no consistent capsular
pattern across subjects.
The techniques used to determine the ROMs cited in subjects
with frozen shoulder have not been consistent.The majority of
From the Program in Physical Therapy,Department of Physical Medicine and
Rehabilitation,University of Minnesota,Minneapolis,MN (Rundquist,Ludewig);
Department of Orthopaedic Surgery,University of Iowa,Iowa City,IA (Anderson);
and University of Minnesota,Minneapolis Sports Medicine Center,Minneapolis,MN
Preliminary data were presented to the American Physical Therapy AssociationÕs
Combined Sections Meeting,February 16,2001,San Antonio,TX.
Supported in part by the Minnesota Medical Foundation (equipment grant).
No commercial party having a direct Þnancial interest in the results of the research
supporting this article has or will confer a beneÞt upon the author(s) or upon any
organization with which the author(s) is/are associated.
Reprint requests to Paula M.Ludewig,PhD,PT,Program in Physical Therapy,
MMC 388,University of Minnesota,420 Delaware St,Minneapolis,MN 55455,
Arch Phys Med Rehabil Vol 84,October 2003
did not identify the measurement tech-
nique they used.One documented by using visual estimation.
Two studies
measured IR by the level of the lumbar spine
reached.Three other investigators measured goniometri-
Only Shaffer
documented the reliability of the
technique he used.
There have been 2 previous studies
with a primary
objective to deÞne the ROMof frozen-shoulder subjects.Both
used 2-dimensional analyses.Projection errors fromevaluating
a 3-dimensional movement with a 2-dimensional technique are
possible.Further,neither study discussed the validity of the
measurements.Recent technical advances permit accurate and
reliable 3-dimensional analysis for the evaluation of shoulder
Although 3-dimensional techniques can better cap-
ture the true movement of the shoulder and separate glenohu-
meral and scapulothoracic motion,no quantitative 3-dimen-
sional ROM data for frozen-shoulder subjects are currently
available in the literature.
The 2 purposes of the present investigation were to provide
a 3-dimensional kinematic description of the motion of the
humerus relative to the trunk and scapula for subjects with
frozen shoulder and to determine whether a capsular pattern of
restriction was consistently present.
The population of interest was people with a diagnosis of
idiopathic frozen shoulder.Symptomatic volunteers were
recruited through a local orthopedic surgeon and physical
therapy clinics.The asymptomatic shoulders from volun-
teers with full motion and no shoulder symptoms were used
for comparison purposes.All subjects were at least 18 years
old.Symptomatic shoulders demonstrated active (AROM)
and passive range of motion (PROM) losses of 25% or
greater compared with the noninvolved shoulder in at least
2 of the following shoulder motions relative to the trunk:
abduction,ER,or IR.
Current symptoms were used to identify symptomatic sub-
jects.Subjects were included if they were in the Reeves
II (stiffness) or phase III (recovery) of frozen shoulder.Sub-
jects were excluded if their pain and/or stiffness had increased
in the past month.This approach was taken to avoid subjects
who were in Reeves
phase I (pain) of frozen shoulder.Asymp-
tomatic control subjects were similar in age range and gender
distribution to the symptomatic subjects.Additional inclusion
criteria for asymptomatic subjects were full pain-free shoulder
motion and no history or current symptoms of shoulder pathol-
Additional speciÞc exclusion criteria for both groups in-
cluded a history of (1) stroke with residual upper-extremity
involvement,(2) rheumatoid arthritis,(3) documented rotator
cuff tear,(4) surgical stabilization of the shoulder,(5) non-
healed fracture of the shoulder complex,(6) osteoporosis,or
(7) severe skin allergies,sensitivities,or other dermatologic
problems in the examination area.Also,subjects whose symp-
toms were exacerbated during a cervical screening examination
were excluded.All subjects reviewed and signed institutional
review boardÐapproved consent forms for human subjects be-
fore participating.
The 3-dimensional position and orientation of each subject Õs
humerus,scapula,and thorax were tracked (40-Hz sampling
rate) by the Polhemus FASTRAK
electromagnetic motion-
capture system.An additional sensor attached to a stylus man-
ually digitized anatomic coordinates.Within a source-to-sensor
separation of 76cm,a root mean square (RMS) accuracy of.15 ¡
for orientation and 0.3 to 0.8mmfor position has been reported
by the manufacturer.
Experimental Procedure
Three FASTRAKsensors were used.With adhesive tape,we
attached 1 sensor to the sternum and another to the skin
overlying the ßat superior bony surface of the scapular acro-
mion process.The third was attached to a thermoplastic cuff
secured to the distal humerus with Velcro straps.The scapular
sensor was placed to minimize the movement error caused by
deltoid contraction.This method closely tracks underlying
scapular movement.
For humeral motion,previous validity
data compared with an external humeral Þxator found less than
a 4¡ RMS error for all motions except for long-axis rotation.IR
and ER resulted in an RMS error of 7.5¡,with the greatest error
at the end ROM.Range of worst case errors were from less
than 1¡ to 15.6¡ (for long-axis rotation),with smaller errors in
the middle of the range and larger errors at the end of the
Subjects were in a standing position throughout testing.
Digitization of bony landmarks on the humerus,scapula,and
thorax enabled transformation of sensor data to local anatom-
ically based coordinate systems (Þg 1).
Kinematic data were
then collected for each subjectsÕfull AROM into ßexion,
abduction,scapular plane abduction,ER,and IR.The scapular
plane was deÞned as 40¡ anterior to the coronal plane.
ER and
IR were collected both with the arm at the side (ER1,IR1) and
with the arm at as close as possible to 90¡ of coronal plane
abduction (ER2,IR2).The starting position for abduction,
ßexion,and scapular plane abduction was with the arm at the
side.The starting position for IR1 and ER1 was with the arm
adducted to the side and the elbow ßexed to 90¡.The starting
position for ER2 and IR2 was with the armabducted to as close
to 90¡ as possible with the elbow ßexed to 90¡ and the forearm
parallel to the ßoor.
Subjects were instructed to move their armas far as possible
for each motion at a self-selected slow,steady speed.Five
repetitions of each motion were collected.Subjects were al-
lowed to rest,if necessary,between each set of motion repe-
Data Reduction and Analysis
Digitized anatomic points were used to deÞne clinically
relevant local anatomic coordinate systems for each segment
based on previously described methods for the shoulder ( Þg
Matrix transformations were performed to describe the
position and orientation of the humerus in relation to the thorax
and scapula.
Humeral orientation relative to the thorax for ßexion,
abduction,and scapular plane abduction was described as
rotation about z
(orients the humerus in a plane of eleva
tion),rotation about y￿
(elevation angle),and rotation about
(axial rotation) (z,y￿,z￿ Euler angles;see Þg 1).Humeral
orientation relative to the thorax for IR and ER was de-
scribed as rotation about y
about x￿
(ßexion,extension),and rotation about z￿
ER) (y,x￿,z￿ Cardan angles;see Þg 1).The alternate
sequence for ER and IR was used to avoid singular or
Ògimbal lockÓ positions.With a Euler rotation sequence
where the second rotation equals or approaches 0 ¡,as would
occur with the arm adducted at the side,the other 2 rotations
are undeÞned,and such a position is referred to as a singular
Also,when interested in clinically interpretable
Arch Phys Med Rehabil Vol 84,October 2003
values of long-axis rotation,the use of a y,x￿,z￿ sequence
is not complicated by the 2 long-axis rotations inherent to a
z,y￿,z￿ description.
A y,x￿,and z￿ Cardan angle sequence was used to describe
all humeral orientations relative to the scapula (see Þg 1).
These rotation sequences (z,y￿,z￿ equivalent
and y,x￿,
) permit clinically relevant descriptions of humeral motion
and are consistent with those previously published.For all
motions using these axis descriptions,the neutral or 0¡,0¡,0¡
3-dimensional positions are deÞned when the humeral axes are
parallel with the axes of the proximal reference frame (either
the cardinal axes of the trunk or the scapular axes).For exam-
ple,for ER and IR relative to the trunk,the humeral axis
parallel with a line through the medial and lateral epicondyles
would be aligned with the frontal plane of the trunk in a neutral
From the 3 trials of each motion,an individual subject Õs
overall peak position value (single trial) was identi Þed.Posi-
tion values were described relative to the 0¡,0¡,0¡ position
(90¡,0¡,0¡ for IR2 and ER2),rather than an ROM from an
individualÕs self-selected position with the arm at the side.
Descriptive statistics (mean,standard deviation [SD],percent-
age of maximum) were calculated across subjects for each
motion.Dependent variables included peak elevation of the
humerus in relation to the trunk and scapula in ßexion,scapular
plane abduction,and abduction and peak long-axis humeral IR
and ER.Symptomatic shoulder rotation values that did not
reach neutral were recorded as negative values.
The normative values used for the calculations of the per-
centage of maximum were the mean peak values from the
asymptomatic shoulder group data.Each symptomatic sub-
jectÕs peak values were expressed as a percentage of the
asymptomatic group data.Average percentages were then cal-
culated across subjects.Negative ROMvalues resulted in neg-
ative percentage values.Patterns of motion loss were deter-
mined from the percentage reduction numbers.
Subject demographics are in table 1.Nine of the 10 subjects
in both groups were women.The right shoulder was the in-
volved shoulder for 4 subjects.Means,SDs,and ranges across
all subjects with frozen shoulder for the 6 motions and total
long-axis rotation relative to the trunk and scapula are in table
2.In general,for elevation of the arm,the maximum ROM
progressively decreased for subjects from ßexion to scapular
plane abduction to coronal plane abduction.For ER,greater
range was generally available with the arm abducted than with
the arm at the side.For IR,however,the available ROM
decreased when the armwas abducted.The variability between
subjects was substantial,as noted by the range and SD values.
Additionally,IR1 was limited by contact with the trunk for all
asymptomatic and most symptomatic subjects.
Mean peak asymptomatic values and average percentages of
motion relative to asymptomatic controls are in table 3 for the
humerus relative to the trunk and scapula.The pattern of
humerus relative to trunk ROM restriction was dependent on
the test position for IR and ER (Þg 2).With the humerus
adducted to the subjectsÕsides,7 of 10 subjects demonstrated
the Cyriax
pattern of proposed capsular restriction ( Þg 2).
One subject showed IR most limited,followed by ER,followed
by abduction.One showed ER most limited,followed by IR,
followed by abduction.One showed abduction most limited,
followed by ER,followed by IR.With the humerus abducted as
close to 90¡ as possible,the predicted Cyriax
capsular pattern
Fig 1.Local coordinate systems for the thorax,scapula,and hu-
merus.NOTE.Trunk coordinate systems are aligned with the coro-
nal planes.Abbreviations:AC,acromioclavicular joint;AD/AB,ad-
duction/abduction;C7,spinous process of the C7 vertebrae;DR/UR,
downward/upward rotation;Flex/Ext,flexion/extension;IA,infe-
rior angle of the scapula;LE,lateral epicondyle;ME,medial epicon-
dyle;PT/AT,posterior/anterior tipping;RS,root of the spine of the
scapula;SN,suprasternal notch;T8,spinous process of the T8
vertebrae;XP,xiphoid process.Reprinted with adaptations from
Ludewig and Cook
with permission of the American Physical
Therapy Association.
Table 1:Subject Demographics
Group Weight (kg) Height (m) Age (y)
Duration of
Symptoms (mo)
Symptomatic 75.95￿18.28 1.70￿.11 52.9￿10.49 7.9￿4.5
Asymptomatic 75.38￿11.26 1.67￿.08 51.0￿10.55 NA
NOTE.Values are mean ￿ SD.
Abbreviation:NA,not applicable.
Arch Phys Med Rehabil Vol 84,October 2003
was present in 4 subjects.Four showed IR most limited,fol-
lowed by ER,followed by abduction.Two showed IR most
limited,followed by abduction,followed by ER (see Þg 2).
The mean actual coronal plane abduction for the symptom-
atic shoulders was 86.2¡ for peak ER2 and 66.7¡ for peak IR2.
For the asymptomatic shoulders,subjects showed the same
tendency to reduce the abduction position for IR2.The average
abduction position was 97.4¡ for peak ER2 and 70¡ for peak
The results of this investigation showed substantial ROM
deÞcits in subjects with frozen shoulder compared with the
comparison group data.Frozen-shoulder subjects showed de-
creased humeral motion as a percentage of normal relative to
the trunk in all planes of motion investigated,except for 3
subjects for IR1,1 for IR2,and 1 for ßexion.Excluding these
values,restricted motions in relation to the trunk ranged be-
tween ￿76%and 99%of the normative values.Comparison of
the percentage-of-normal value relative to the scapula was
more variable.
The descriptive humerus abduction data relative to the scap-
ula and the trunk closely followed the 2-dimensional ßuoro-
scopic results previously published by Eto,
who studied 17
symptomatic subjects with a diagnosis of Òperiarthritis scapulo-
humeralis.ÓInclusion criteria for that study required subjects to
have a maximum elevation angle relative to the trunk of less
than 120¡.In relation to the trunk,Eto
documented maximum
elevation ranging from 65¡ to 117¡ across subjects,with an
average of 92.8¡.In the present study,the peak abduction
values ranged from 57¡ to 134¡,with an average of 98.4¡.In
relation to the scapula,Eto
documented maximum elevation
ranging from 21¡ to 67¡,with an average of 38.3¡.
In the
present study,the peak abduction values ranged from 27 ¡ to
89¡,with an average of 46.4¡.
Comparison with the hydrogoniometer data from Clarke et
for 30 symptomatic frozen-shoulder subjects was less
similar.A hydrogoniometer is a ßuid-Þlled goniometer that
uses gravity as a reference point.Our ER values were less than
those reported by Clarke
(men averaged 23¡;women,28¡),
and our abduction values were generally greater (men averaged
42¡;women,51¡).EtoÕs methods
(standing AROM) were
more similar to those of the present study.Clarke
PROM and did not provide subject position information.
A difÞculty in comparing the results of the present study
with those from previous investigations of frozen-shoulder
ROM is the lack of humeral data described relative to the
scapula.Only Eto
provides glenohumeral motion data for
subjects with ROM deÞcits.The technology used for this
investigation does allow for differentiation of glenohumeral
motion.In general for our results during arm elevation,the
average humerus to scapula percentages of normal are some-
Table 3:Motion Relative to the Trunk and Scapula as a
Percentage of Normal
Mean Peak
Asymptomatic Values
Mean % of
Symptomatic to
Trunk Scapula Trunk Scapula
Flexion 147.9 97.2 79 73
Abduction 151.1 99.8 65 46
Scapular plane
abduction 150.7 98.8 75 62
ER1* 24.9 50.8 18 49

62.1 65.4 54 68
IR1* 59.6 10.4 91 99

41.5 15.2 43 ￿42
*ER1 and IR1 performed with the humerus adducted at the side.

ER2 and IR2 performed with the humerus abducted to as close to
90° as possible.Negative values indicate that a neutral position was
not attained.
Table 2:Humeral ROM (deg)
Relative to Trunk Relative to Scapula
Mean ￿ SD Range Mean ￿ SD Range
Flexion 116.9￿22.1 80–165 70.5￿16.4 41–102
Abduction 98.4￿25.0 57–134 46.4￿18.9 27–89
Scapular plane abduction 113.4￿18.7 80–145 61.7￿17.0 38–97
ER1* 4.5￿12.3 ￿19 to 20 34.7￿15.8 14–64

33.5￿15.5 8–54 45.3￿18 15–70
IR1* 54.3￿13.6 39–73 10.3￿16.2 ￿18 to 29

17.8￿17.9 0–50 ￿6.4￿16.6 ￿29 to 20
Total rotation 1 65.4￿16.2 42–92 51.4￿10.5 32–69
Total rotation 2 51.3￿16.9 28–78 59.5￿23.2 27–84
NOTE.Values are mean ￿ SD.
*ER1 and IR1 performed with the humerus adducted at the side.Negative values indicate that a neutral position was not attained.

ER2 and IR2 performed with the humerus abducted to as close to 90° as possible.Negative values indicate that a neutral position was not
Fig 2.Distribution of motion patterns across subjects and arm
Arch Phys Med Rehabil Vol 84,October 2003
what less than the average humerus to trunk percentages,
suggesting scapular substitution of increased motion to obtain
maximum elevation ROM (table 3).These results are sup-
ported by Vermeulen et al,
who demonstrated an increased
scapular contribution to the scapulohumeral rhythm in subjects
with frozen shoulder.
However,an approximation of the contribution of glenohu-
meral and scapulothoracic components of elevation motion can
be made by review of the average data in table 2.During
abduction,the asymptomatic groupÕs humerus to thorax motion
was 151¡,and the humerus to scapular motion was 100¡.For
the symptomatic group,humerus to thorax abduction was 98 ¡,
and humerus to scapular motion was 46¡.Therefore,on aver-
age,53¡ of abduction was lost overall,whereas 54¡ was lost at
the glenohumeral joint,indicating only minimal,if any,scap-
ular substitution.By using this same approximation during
ßexion,on average,31¡ of overall motion was lost,and 26¡ of
this loss occurred at the glenohumeral joint,suggesting no
positive scapular substitution.During scapular plane abduction,
on average,38¡ of motion was lost,and the glenohumeral
motion loss was nearly equivalent,at 37¡.Therefore,this
approximation method does not support the theory of any
consistent substantive scapular substitution to maximize end-
range elevation motion.It should be noted,however,that
because this method is essentially a 2-dimensional approxima-
tion of a 3-dimensional motion,it must have some inherent
error.This method also is based on average rather than on
individual results and considers only end ROM.Scapular sub-
stitution earlier in the ROM would not be detected with this
method and may occur in some subjects.Further investigation
of this issue is warranted in future studies.
It is interesting to note the changes in ER and IR ROMwith
the arm abducted as compared with the arm adducted (table 2).
Greater ROM for ER was obtained when the arm was ab-
ducted.The IR ROM,however,decreased when the arm was
abducted.This same pattern was apparent for healthy shoulders
(table 3) and is consistent with an abducted position resulting
in differential capsular tightness as compared with a humeral
adducted position.
However,the percentages of normal
motion are greater for ER2 than for ER1 and for IR1 than for
IR2.The pattern of loss for ER supports a premise of coraco-
humeral ligament restrictions.
The coracohumeral ligament
is believed to limit ER ROM in an adducted arm position to a
greater degree than when the arm is in an abducted posi-
Other potential factors limiting ER are the rotator
interval and the superior glenohumeral ligament,which have
greater resistance to ER as the arm is abducted.The rotator
interval resists ER at 60¡,
and the superior glenohumeral
ligament resists ER at 90¡.
The pattern of loss for IR is
consistent with capsular tightness in the posterior band of the
inferior glenohumeral ligament complex,which restricts IR in
abducted but not adducted humeral positions.
Tightness in
any or all of these structures may lead to the development of
frozen shoulder and has been implicated by various au-
These positional inßuences on the available rotation ROM
contributed to a lack of consistent support for a single capsular
pattern.If the shoulder capsular pattern proposed by Cyriax
exists,ER should be most limited,followed by abduction,
followed by IR.This pattern was not consistently supported by
these results.With the humerus adducted to the side,the
capsular pattern proposed by Cyriax
was present for 7 of the
10 subjects.However,this Þnding decreased to 4 of 10 subjects
when the arm was abducted.In the abducted position,IR was
limited more than ER,and both were limited more than abduc-
tion for 4 subjects.The IR was limited more than abduction,
and both were limited more than ER for 2 subjects.Cyriax
did not specify the position of the arm when he described the
capsular pattern.
Considering these inconsistent capsular pattern results,we
have 2 possibilities to consider.First,different areas of the
capsule may tighten in different subgroups of frozen-shoulder
patients.If so,the ROMchanges would be different across the
different groups.This would support different treatment inter-
vention emphasis in terms of joint mobilization and stretching
for each distinct group.For example,interventions with the
arm adducted to the side might focus on regaining abduction
and ER.With the arm abducted,the focus might be on regain-
ing ER and IR.For subjects with the greatest restriction for IR
with the arm abducted,greater emphasis on mobilizing the
posterior capsule might be used,whereas for subjects with the
greatest restriction for ER with the arm at the side,emphasis
might be placed on mobilizing the anterior capsule.Currently,
very limited research data on the effectiveness of stretching or
mobilization for frozen shoulder are available in the peer-
reviewed literature.
Alternatively,consistent capsular pat-
terns may not exist at all,necessitating individualized interven-
tion strategies for each subject.
The measurement method used in the present study has both
advantages and limitations as compared with standard clinical
goniometric methods.Besides providing the advantage of 3-di-
mensional measurement,a trunk sensor also enables better
separation of humeral motion from trunk movement.Further,
the moving axis is embedded in the humerus,providing a more
direct measure of humeral long-axis rotation,rather than being
aligned with the forearm,as in clinical goniometry.However,
both IR and ER in all subjects may be underrepresented by
using the surface thermoplastic cuff around the humerus.The
cuff may not fully track humeral motion at the end ROM.
Validation of this surface cuff technique demonstrated a 7.5 ¡
average error in long-axis rotation measurements,with the
largest discrepancies occurring at end range,whereas a 3 ¡
average error was described for all other planes of humeral
Additionally,IR1 was limited by contact with the
trunk for all asymptomatic and most symptomatic subjects,
thus possibly underrepresenting full motion.In future studies,
collection of IR beginning with the forearm in the small of the
back and moving away from the body might clarify the end
range available.All of these factors,plus the average age of the
subjects,contribute to lower IR and ER values than often
reported in the literature for healthy subjects.
Two different rotation sequences were used in the present
study to describe humeral motions.Optimally,to avoid any
possible confounding of mathematical sequence effects in ßu-
encing the interpretation of humerus to scapula versus humerus
to thorax motion,a single sequence would be desired.How-
ever,as identiÞed in the Methods section,the alternate se-
quence for ER and IR was used to avoid singular or gimbal
lock positions for IR1 and ER1.
In addition,when interested
in clinically interpretable values of long-axis rotation,the use
of a y,x￿,z￿ sequence is not complicated by the 2 long-axis
rotations inherent in a z,y￿,z￿ description.Alternatively,using
only a y,x￿,z￿ sequence is not possible because of singular
positions with humeral ßexion relative to the trunk.Despite the
use of these 2 sequences across motions,all comparisons
across subjects and groups are made without confounding.The
same rotation sequences are used across subjects.
A further limitation of the present study is the sample size
and proportion of female subjects.The study sample may not
fully represent the spectrum of patients with a frozen-shoulder
diagnosis.This factor may be particularly true if subgroups of
patients exist.Also,9 of 10 symptomatic subjects were women.
Arch Phys Med Rehabil Vol 84,October 2003
The incidence of frozen shoulder is higher in women than
Thus,our data may not be fully generalizable to men
with frozen shoulders.Despite these study limitations,this
study provides the Þrst 3-dimensional description of shoulder
ROM deÞcits for this subject population.
There are several avenues to pursue in future investigations.
Further studies focusing on humeral motion relative to the
scapula could more directly measure the inßuence of a tight
capsule on glenohumeral motion.Correlation of ROM de Þcits
to surgically identiÞed capsular adhesions would be optimal;
however,surgical treatment approaches are not common with
these patients.Additional avenues of investigation might in-
clude treatment efÞcacy studies,patient classiÞcation studies,
and longitudinal investigations of motion loss and recovery.
Longitudinal studies would provide greater insight into how
frozen shoulder develops and resolves clinically.
Humeral ROMdeÞcits relative to the trunk and scapula were
present in subjects with a diagnosis of idiopathic frozen shoul-
der.The pattern of motion loss for IR and ER was dependent
on arm position and was not consistent across all subjects.The
results raise questions about the validity of a theorized single
capsular pattern of motion in these subjects.
1.Duplay S.Dela periarthrite scapulohumerale.Med Week 1896;4:
2.Codman EA.The shoulder.Rupture of the supraspinatus tendon
and other lesions in or about the subacromial bursa.Boston:
Thomas Todd;1934.
3.Reeves B.The natural history of the frozen shoulder syndrome.
Scand J Rheumatol 1975;4:193-6.
4.Lundberg BJ.The frozen shoulder:clinical and radiographical
observations.The effect of manipulation under general anesthesia.
Structure and glycosaminoglycan content of the joint capsule.
Local bone metabolism.Acta Orthop Scand Suppl 1969;119:1-59.
5.Anton HA.Frozen shoulder.Can Fam Physician 1993;39:1773-8.
6.Neviaser RJ,Neviaser TJ.The frozen shoulder:diagnosis and
management.Clin Orthop 1987;Oct(223):59-64.
7.Neviaser TJ.Adhesive capsulitis.Orthop Clin North Am 1987;
8.Bridgeman JF.Periarthritis of the shoulder and diabetes mellitus.
Ann Rheum Dis 1972;31:69-71.
9.Bulgen DY,Binder AL,Hazleman BL.Frozen shoulder:a pro-
spective clinical study with an evaluation of three treatment re-
gimes.Ann Rheum Dis 1984;43:353-60.
10.Dickson JA,Crosby EH.Periarthritis of the shoulder:analysis of
200 cases.JAMA 1932;99:2252-7.
11.Hill JJ,Bogumill H.Manipulation in the treatment of frozen
shoulder.Orthopedics 1988;11:1255-60.
12.Hsu SY,Chan KM.Arthroscopic distension in the management of
frozen shoulder.Int Orthop 1991;15:79-83.
13.Loyd JA,Loyd HM.Adhesion capsulitis of the shoulder:arthro-
graphic diagnosis and treatment.South Med J 1983;76:879-83.
14.Mulcahy KA,Baxter AD,Oni OO,Finlay D.The value of
shoulder distension arthrography with intraarticular injection of
steroid and local anesthetic:a follow-up study.Br J Radiol 1994;
15.Neviaser RJ.Painful conditions affecting the shoulder.Clin Or-
thop 1983;Mar(173):66-9.
16.Nobuhara K,Sugiyama D,Ikeda H,Makiura M.Contracture of
the shoulder.Clin Orthop 1990;May(254):105-10.
17.Ellenbecker TS,Roetert EP,Piorkowski PA,Schulz DA.Gleno-
humeral joint internal and external range of motion in elite junior
tennis players.J Orthop Sports Phys Ther 1996;24:336-41.
18.Rizk TE,Christopher RP,Pinais RS.Adhesive capsulitis (frozen
shoulder):a new approach to its management.Arch Phys Med
Rehabil 1983;64:29-33.
19.Rizk TE,Gavant ML,Pinals RS.Treatment of adhesive capsulitis
(frozen shoulder) with arthrographic capsular distension and rup-
ture.Arch Phys Med Rehabil 1994;75:803-7.
20.Shaffer B.Frozen shoulder:a long-term follow-up.J Bone Joint
Surg Am 1992;74:738-46.
21.Sharma RK,Bejekal RA,Bhan S.Frozen shoulder syndrome.Int
Orthop 1993;17:275-8.
22.Soren A,Fetto JF.Contracture of the shoulder joint.Arch Orthop
Trauma Surg 1996;115:270-2.
23.Mao C,Jaw W,Cheng H.Frozen shoulder:correlation between
the response to physical therapy and follow-up shoulder arthrog-
raphy.Arch Phys Med Rehabil 1997;78:857-9.
24.Neviaser JS.Adhesive capsulitis of the shoulder:a study of the
pathological Þndings in periarthritis of the shoulder.J Bone Joint
Surg 1945;27:211-22.
25.Warner JJP.Frozen shoulder:diagnosis and management.J Am
Acad Orthop Surg 1997;5:130-40.
26.Parker RD,Froimson AI,Winsberg DD,Arsham NZ.Frozen
shoulder.Part I:chronology,pathogenesis,clinical picture,and
treatment.Orthopedics 1989;12:869-73.
27.Bunker TD.Frozen shoulder:unravelling the enigma.Ann R Coll
Surg Engl 1997;79:210-3.
28.Pearsall AW,Osbahr DC,Speer KP.An arthroscopic technique
for treating patients with frozen shoulder.Arthroscopy 1999;15:
29.Cyriax J.Textbook of orthopedic medicine.Vol 1:Diagnosis of
soft tissue lesions.7th ed.New York:Macmillan Publishing;
30.Grey RG.The natural history of ÒidiopathicÓfrozen shoulder.
J Bone Joint Surg Am 1978;60:564.
31.Miller MD,Wirth MA,Rockwood CA.Thawing the frozen shoul-
der:the ÒpatientÓpatient.Orthopedics 1996;19:849-53.
32.Clarke GR,Willis LA,Fish WW,Nichols PJ.Preliminary studies
in measuring range of motion in normal and painful stiff shoul-
ders.Rheumatol Rehabil 1975;14:39-46.
33.Eto M.Analysis of the scapulo-humeral rhythm for periarthritis
scapulohumeralis.J Jpn Orthop Assoc 1991;65:693-707.
34.Karduna AR,McClure PW,Michener LA,Sennett B.Dynamic
measurements of three dimensional scapular kinematics:a valida-
tion study.J Biomech Eng 2001;123:184-90.
35.Ludewig PM,Cook TM.Alterations in shoulder kinematics and
associated muscle activity in people with symptoms of shoulder
impingement.Phys Ther 2000;80:276-91.
36.McQuade KJ,Smidt GL.Dynamic scapulohumeral rhythm:the
effects of external resistance during elevation of the arm in the
scapular plane.J Orthop Sports Phys Ther 1998;27:125-31.
37.Polhemus Inc.3Space Fastrak user Õs manual.Rev F.Colchester
38.Ludewig PM,Cook TM,Shields RK.Technical Note:comparison
of surface sensor and bone-Þxed measurement of humeral motion.
J Appl Biomech 2002;18:163-70.
39.Craig JJ.Introduction to robotics:mechanics and control.2nd ed.
New York:Addison-Wesley;1989.
40.Meskers CG,Vermeulen HM,de Groot JH,van der Helm FC,
Rozing PM.3D shoulder position measurements using a six-
degree-of-freedomelectromagnetic tracking device.Clin Biomech
(Bristol,Avon) 1998;13:280-92.
41.An KN,Browne AO,Korinek S,Tanaka S,Morrey BF.Three-
dimensional kinematics of glenohumeral elevation.J Orthop Res
42.Vermeulen HM,Stokdijk M,Eilers PH,Meskers CG,Rozing PM,
Vliet Vlieland TP.Measurement of three dimensional shoulder
movement patterns with an electromagnetic tracking device in
patients with a frozen shoulder.Ann Rheum Dis 2002;61:115-20.
43.Terry GC,Hammon D,France P,Norwood LA.The stabilizing
function of passive shoulder restraints.Am J Sports Med 1991;
Arch Phys Med Rehabil Vol 84,October 2003
44.Harryman DT II,Sidles JA,Harris SL,Matsen FA III.The role of
the rotator interval capsule in passive motion and stability of the
shoulder.J Bone Joint Surg Am 1992;74:53-66.
45.Pearsall AW,Osbahr DC,Speer KP.An arthroscopic technique
for treating patients with frozen shoulder.Arthroscopy 1999;15:2-11.
46.Warner JJ,Answorth A,Marks PH,Wong P.Arthroscopic release
for chronic,refractory adhesive capsulitis of the shoulder.J Bone
Joint Surg Am 1996;78:1808-6.
47.Vermeulen HM,Obermann WR,Burger BJ,Kok GJ,Rozing PM,
van Den Ende CH.End-range mobilization techniques in adhesive
capsulitis of the shoulder joint:a multiple-subject case report.
Phys Ther 2000;80:1204-13.
a.Polhemus Inc,40 Hercules Dr,PO Box 560,Colchester,VT 05446.
Arch Phys Med Rehabil Vol 84,October 2003