[ research report ]

KEISUKE MATSUKI, MD, PhD1 KEI O. MATSUKI, MD, PhD2 SATOSHI YAMAGUCHI, MD, PhD2 NOBUYASU OCHIAI, MD, PhD3
TAKAHISA SASHO, MD, PhD3 HIROYUKI SUGAYA, MD, PhD4 TOMOAKI TOYONE, MD, PhD5 YUICHI WADA, MD, PhD5
KAZUHISA TAKAHASHI, MD, PhD6 SCOTT A. BANKS, PhD7

Dynamic In Vivo Glenohumeral

Kinematics During Scapular Plane
Abduction in Healthy Shoulders

S
houlder subacromial impingement syndrome is one of the most symptomatic patients and to result in
common causes of shoulder pain and is thought to be due to additional narrowing of the subacro-
Downloaded from www.jospt.org at on July 22, 2017. For personal use only. No other uses without permission.

mial space, which could aggravate symp-

inadequate space for clearance of the rotator cuff tendons during
toms.9,15 These changes could include
arm elevation.9 Various theories have been proposed regarding altered glenohumeral joint kinematics,
the primary mechanism of shoulder impingement, including kinematic such as excessive superior translation of
changes, anatomic abnormality, and degeneration of the rotator cuff the humeral head and inadequate exter-
Copyright 2012 Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

tendons.9,29,39 Kinematic changes are thought to occur primarily in
TTSTUDY DESIGN: Controlled laboratory study.
TTOBJECTIVES: To measure superior/inferior
translation and external rotation of the humerus
relative to the scapula during scapular plane
abduction using 3-D/2-D model image registration
techniques.
TTBACKGROUND: Kinematic changes in the
Journal of Orthopaedic & Sports Physical Therapy
TTRESULTS: On average, there was 2.1 mm of
initial humeral translation in the superior direction
from the starting position to 105 of humeral eleva-
tion. Subsequently, an average of 0.9 mm of trans-
lation in the inferior direction occurred between
105 and maximum arm elevation. The average
amount of external rotation of the humerus was
14 from the starting position to 60 of humeral
nal rotation of the humerus.31,34
Many studies have quantified shoul-
der and glenohumeral joint kinematics.
Most earlier studies used conventional
single-plane radiographs,5,7,8,31 which
could not describe 3-D motion of the
shoulder and did not provide dynamic
kinematics. Three-dimensional dynamic
assessments have been employed in more

glenohumeral joint, including excessive superior
translation of the humeral head and inadequate ex-
ternal rotation of the humerus, are believed to be a
possible cause of shoulder impingement. Although
many researchers have analyzed glenohumeral
kinematics with various methods, few articles have
assessed dynamic in vivo glenohumeral motion.
TTMETHODS: Twelve healthy males with a mean
age of 32 years (range, 27-36 years) were enrolled
in this study. Fluoroscopic images of the dominant
shoulder during scapular plane elevation were
taken, and computed tomography-derived 3-D
bone models were matched with the silhouette
of the bones in the fluoroscopic images using
3-D/2-D model image registration techniques. The
kinematics of the humerus relative to the scapula
were determined using Euler angles.
elevation. The humerus then rotated internally
an average 9 by the time the shoulder reached
maximum elevation. These changes in superior/
inferior translation and external/internal rotation
were statistically significant (P<.001 and P = .001,
respectively), based on 1-way repeated-measures
analysis of variance.
TTCONCLUSION: The observed glenohumeral
translations and rotations characterize healthy
shoulder function and serve as a preliminary
foundation for quantifying pathomechanics in the
presence of glenohumeral joint disorders. J Orthop
Sports Phys Ther 2012;42(2):96-104, Epub 25
October 2011. doi:10.2519/jospt.2012.3584
TTKEY WORDS: 3-D/2-D registration, arthro-
kinematics, computed tomography, imaging,
impingement
recent works to describe shoulder kine-
matics via radiostereometric analysis,12
magnetic resonance imaging,10,11,32 elec-
tromagnetic tracking devices,6,19-21,25,26
or computerized 3-D motion analysis
systems.37 These 3-D assessments have
provided significant new information on
shoulder motions but still have techni-
cal shortcomings. Magnetic resonance
imaging studies do not provide dynamic
kinematics, motion capture with skin-
affixed markers can suffer from soft tis-
sue artifacts (skin movement relative to
the bones),6,19,20,25 and studies using bone
pins have limited application outside the
research laboratory setting.4,21,26

1
Lecturer, Teikyo University Chiba Medical Center, Department of Orthopaedic Surgery, Ichihara, Chiba, Japan. 2Staff Surgeon, Chiba University Graduate School of Medicine,
Department of Orthopaedic Surgery, Chiba, Chiba, Japan. 3Assistant Professor, Chiba University Graduate School of Medicine, Department of Orthopaedic Surgery, Chiba, Chiba,
Japan. 4Director of Shoulder and Elbow Surgery, Funabashi Orthopaedic Sports Medicine Center, Funabashi, Chiba, Japan. 5Professor, Teikyo University Chiba Medical Center,
Department of Orthopaedic Surgery, Ichihara, Chiba, Japan. 6Professor, Chiba University Graduate School of Medicine, Department of Orthopaedic Surgery, Chiba, Chiba, Japan.
7
Associate Professor, University of Florida, Department of Mechanical and Aerospace Engineering, Gainesville, FL. Approval of the study was obtained from the Institutional
Review Board of Chiba University Graduate School of Medicine. The authors and their families did not receive and will not receive any benefits or funding from any commercial
party related directly or indirectly to the subject of this article. Address correspondence to Dr Keisuke Matsuki, Department of Orthopaedic Surgery, Teikyo University Chiba
Medical Center, 3426-3 Anesaki, Ichihara, Chiba 2990111 Japan. E-mail: kmatsuki@med.teikyo-u.ac.jp

96 | february 2012 | volume 42 | number 2 | journal of orthopaedic & sports physical therapy

42-02 Matsuki.indd 96 1/18/2012 3:32:47 PM

 Previous studies have resulted in con-
flicting observations of superior/inferior
translation of the humeral head during
arm elevation. One study demonstrated
monotonic superior translation of the
humeral head,33 2 other studies observed
the humerus moving first superiorly and
then inferiorly during arm elevation,10,20
and 3 others observed the humerus be-
ing essentially fixed with respect to the
glenoid during arm elevation.5,7,12 The
absence of glenohumeral translation is
surprising and counterintuitive, as one
would expect the deltoid to produce su-
Downloaded from www.jospt.org at on July 22, 2017. For personal use only. No other uses without permission.
perior humeral translation in the early
abduction arcthe so-called setting
phasebefore the rotator cuff muscles
stabilize the joint.7,14
There have been only a few articles
Copyright 2012 Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.
reporting on the movement of external
rotation of the humerus relative to the
scapula during arm elevation.12,19,21 They
indicate that the humerus externally ro-
tates an average of 30.
Three-D/two-D model image regis-
tration is a 3-D assessment technique
utilizing fluoroscopic images and was
originally developed to measure in vivo
Journal of Orthopaedic & Sports Physical Therapy
total knee arthroplasty kinematics.1,24
This technique has been applied to natu-
ral knee joints,17,27 anterior cruciate liga-
ment-deficient knees,36 and osteoarthritic
knees.13 Three-D/two-D model image
registration techniques can provide suf-
ficient accuracy for noninvasive, dynamic
kinematic analysis but they use ionizing
radiation.1,24,27 Several studies of shoulder
kinematics using 3-D/2-D model image
registration techniques recently have
been published.2,18,23,30 Nishinaka et al30
analyzed healthy shoulders and reported
that the humeral head monotonically
translated 1.7 mm superiorly during arm
elevation. However, Bey et al2 used bi-
plane fluoroscopy and reported that the
humeral head first moved superiorly and
then inferiorly in shoulders with repaired
rotator cuffs and in the contralateral
shoulders of the same individuals. Data
on external rotation of the humerus were
not reported.
The purpose of this study was to mea-
journal of orthopaedic & sports physical therapy | volume 42 | number 2 | february 2012 | 97
42-02 Matsuki.indd 97
FIGURE 1. Anatomic coordinate systems of the
humerus and the scapula.18,30 Elevation of the
humerus was defined as rotation about the scapular
z-axis and internal/external rotation was defined as
rotation about the humeral y-axis. Superior/inferior
translation was defined as the motion of the humeral
origin relative to the scapular origin along the y-axis
of the scapula.
sure superior/inferior translation and
external rotation of the humerus relative
to the scapula during dynamic scapular
plane elevation. We hypothesized that the
humeral head would translate superiorly,
at least in the early phase of arm elevation,
and that humeral external rotation would
be 30 during arm elevation, similar to
what has been reported in previous studies.
METHODS
Participants
T
welve healthy males with a
mean age of 32 years (range, 27-36
years) were enrolled in this study.
They had no history of shoulder pain,
and a radiologist confirmed that plain
radiographs of their shoulders did not
show any radiographic abnormalities. All
participants provided informed consent,
and the protocol for this study was ap-
proved by the Institutional Review Board
of Chiba University. All participants were
right-handed.
Image Acquisition and 3-D Modeling
Fluoroscopic images of scapular plane
abduction were recorded at 30 Hz (di-
ameter field of view, 406 mm; pixel
size, 0.48 0.48 mm) for the dominant
shoulder of each participant (Infinix Ac-
tiv; Toshiba, Tochigi, Japan). The par-
ticipants sat on a chair with the torso at
an angle approximately 30 to the X-ray
beam, so that the plane of the scapula
was parallel to the image intensifier. El-
evation in the scapular plane was per-
formed from the arm at the side, along
the trunk, to maximum elevation, while
keeping the elbow fully extended and
the arm externally rotated (thumbs-up
position), at a rate of approximately 2
seconds per cycle. One cycle was defined
as full arm elevation followed by return
to the initial position. To allow natural
arm motion, the participants were not
constrained and the speed of motion was
not strictly controlled. The participants
practiced the activity until they felt com-
fortable, then 3 cycles of shoulder eleva-
tion were recorded for each shoulder. A
short pause was taken between cycles to
allow complete relaxation of the shoulder
muscles. Fluoroscopic images of a cali-
bration object were acquired to permit
computation of the radiographic projec-
tion parameters and to correct geometric
image distortion.
Computed tomography scans (Infinix
Activ) of the shoulder were acquired with
a 0.5-mm slice pitch (image matrix, 512
512; pixel size, 0.62 0.62 mm). The
computed tomography images were seg-
mented, and 3-D surface models of the
proximal humerus and scapula were cre-
ated (ITK-SNAP; Penn Image Comput-
ing and Science Laboratory, Philadelphia,
PA).38 Anatomic coordinate systems were
embedded in each model, according to
reported conventions (Geomagic Studio;
Geomagic USA, Morrisville, NC).18,30 In
brief, the humeral origin was placed at
the centroid of the humeral head. The y-
axis was parallel to the humeral shaft and
the z-axis was defined as a line through
the intertubercular groove from the ori-
gin. The scapular origin was defined as
the midpoint of the line connecting the
most superior and inferior bony edges of
the glenoid, and the y- and z-axes were
pointed superiorly and anteriorly, respec-
tively (FIGURE 1).
1/18/2012 3:32:49 PM
 [ research report ]
A
0

Humeral Position, mm
2

3
R2 = 0.86

FIGURE 2. Three-D/two-D model image registration

Downloaded from www.jospt.org at on July 22, 2017. For personal use only. No other uses without permission.

techniques were used to determine the 3-D position

and orientation of the shoulder bones using the 5
custom, open-source software JointTrack (http:// 20 0 20 40 60 80 100 120 140 160 180
www.sourceforge.net/projects/jointtrack).
Humeral Elevation Angle, deg
Copyright 2012 Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

Model Image Registration and Data

B
Processing 120

The 3-D position and orientation of the

humerus and scapula were determined
via model image registration techniques,
Humeral External Rotation Angle, deg

110
using the open-source software (Joint-
Track; www.sourceforge.net/projects/
jointtrack) developed in our labora-
tory (FIGURE 2).1,24 The bone model was 100
Journal of Orthopaedic & Sports Physical Therapy

projected onto the distortion-corrected

fluoroscopic image, and its 3-D pose was
iteratively adjusted to match its silhou-
90
ette with the silhouette of the fluoro-
scopic image. The best-case accuracy of
this matching method was 0.53 mm for
in-plane translation (parallel to image 80
plane), 1.6 mm for out-of-plane transla- 0 20 40 60 80 100 120 140 160

tion (perpendicular to image plane), and

0.54 for rotations reported in a previous
study on natural knee joints.27 Model im-
age registration measurements were per-
formed by a single operator (K.M.) using
a series of fluoroscopic images from the
second cycle of the activity.
The kinematics of the humerus rela-
tive to the radiographic coordinate sys-
tem and to the scapula were determined
using Cardan angles (z-x-y order).4,19 El-
evation of the humerus was defined as
rotation about the z-axis, and internal/
external rotation of the humerus was de-
fined as rotation about its y-axis. Supe-
rior/inferior translation was defined as
Humeral Elevation Angle, deg
FIGURE 3. Polynomial regression line fits for a representative participant. (A) Humeral position relative to the
center of the glenoid. Superior translation is indicated by values moving toward zero. (B) External/internal rotation
of the humerus.
the motion of the humeral origin rela- to 0.96 for translational data (FIGURE 3A
tive to the scapular origin along the y- and 3B).
axis of the scapula. Kinematic data were
individually plotted as a function of the Statistical Analysis
humeral elevation angle, and polynomial Statistical analysis was performed using
regression lines were used to calculate PASW Version 17.0 (SPSS Inc, Chicago,
interpolating values at each 15 incre- IL), and the level of significance was set
ment of humeral elevation. R2 values of at P<.05. One-way repeated-measures
the polynomial regression line fits were analyses of variance and post hoc pair-
0.90 to 0.99 for rotational data and 0.47 wise comparisons (Tukey) were used for

98 | february 2012 | volume 42 | number 2 | journal of orthopaedic & sports physical therapy

42-02 Matsuki.indd 98 1/18/2012 3:32:50 PM

 A
35

30

25

Translation, mm
20

15
FIGURE 4. Fluoroscopic image of the cadaveric
Downloaded from www.jospt.org at on July 22, 2017. For personal use only. No other uses without permission.

shoulder. Five and 6 metallic beads (2 mm, lead)

were placed in the scapula and proximal humerus,
10
respectively.

the analysis of rotational and translation- 5

Copyright 2012 Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

al changes within subjects.

Validation Study With Cadaver 0

To study the accuracy of the 3-D/2-D 1 2 3 4 5 6 7 8 9 10

model image registration technique for B

70
the shoulder, we conducted a validation
study with a cadaveric specimen. One
fresh-frozen shoulder specimen was 60
used for this study. No skin or soft tissue
Journal of Orthopaedic & Sports Physical Therapy

was removed from the specimen, and 5

and 6 metallic beads (2 mm, lead) were 50

placed in the scapula and proximal hu-

merus, respectively. Computed tomogra-
40
phy scans with 0.75-mm slice thickness
Rotation, deg

(image matrix, 512 512; pixel size, 0.74

0.74 mm) were acquired for the speci-
30
men (SOMATOM Sensation 16; Siemens
Medical Solutions, Malvern, PA), and
3-D surface models of the bones and im- 20
planted beads were reconstructed. The
scapula was mounted to a radiolucent
fixture, and the humerus was manually 10
abducted at approximately 2 seconds per
cycle. The range of glenohumeral abduc-
tion was from 6 to 37. Images of pas- 0

sive abduction were recorded at 15 frames 1 2 3 4 5 6 7 8 9 10

Image Number
per second (field of view, 380 380 mm;
x value, bead model y value, bead model z value, bead model
pixel size, 0.37 0.37 mm) with a biplane x value, bone model y value, bone model z value, bone model
flat-panel imaging system (Allura Xper
FD20/20; Philips Healthcare, Andover, FIGURE 5. Comparison of biplane bead-based and single-plane bone model-based kinematic measurements.
MA) (FIGURE 4). Ten images from 1 abduc- Translations (A) and rotations (B) relative to the image coordinate system are shown.
tion sequence were selected; model image
registration was performed with single-

journal of orthopaedic & sports physical therapy | volume 42 | number 2 | february 2012 | 99

42-02 Matsuki.indd 99 1/18/2012 3:32:51 PM

 [ research report ]
plane images and bone-surface models A
1
and with biplane images and bead mod-
els, separately, by a single operator (K.M.).
Root-mean-square errors between the
0
single-plane bone model kinematics and
the biplane bead model kinematics were

Humeral Position, mm
calculated, taking the biplane kinematics 1
as the reference measure.

RESULTS 2

Accuracy of 3-D/2-D Model Image

Registration 3

T
he kinematic differences be-
Downloaded from www.jospt.org at on July 22, 2017. For personal use only. No other uses without permission.

tween single-plane bone matching

and biplane bead matching for a 4
single fresh-frozen cadaver specimen are Start 30 45 60 75 90 105 120 135 Max
shown in FIGURE 5. Root-mean-square er-
B
rors in glenohumeral translation were 2
Copyright 2012 Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.

0.47 mm and 1.53 mm for in-plane and

out-of-plane translations, respectively.
Root-mean-square errors in glenohumeral
0
rotation were 0.76 and 3.72 for in-plane
and out-of-plane rotations, respectively.
Humeral Position, mm

Superior/Inferior Translation of the

2
Humerus
On average, a humeral translation of 2.1
Journal of Orthopaedic & Sports Physical Therapy

mm (range, 0.4 to 3.2 mm) in the su-

perior direction occurred from the start-
ing position to 105 of humeral elevation.
This was followed by, on average, a 0.9-
mm (range, 0.9 to 2.9 mm) translation
in the inferior direction throughout the
rest of the motion toward maximum arm
elevation (FIGURE 6A). Changes in humer-
al position throughout the movement of
humeral elevation were statistically sig-
nificant (P<.001). Post hoc pairwise tests
revealed significant differences between
the starting position and all other report-
ed angles, except at 30 of humeral eleva-
tion (P<.001 at each point). The humeral
superior/inferior position at maximum
elevation was significantly different from
the positions at 90 and 105 of humeral
elevation (P = .03 and P = .02, respective-
ly). Individual patterns of humeral trans-
lation are shown in FIGURE 6B. There was
a large amount of variation in the initial
humeral position; however, the humerus
translated superiorly in the early phase of
4
6
Start 30 45 60 75 90 105 120 135 Max
Humeral Elevation Angle, deg
FIGURE 6. Humeral position relative to the center of the glenoid during shoulder scapular abduction. Superior
translation is indicated by values moving toward 0. (A) Means and standard deviations for all 12 shoulders. (B) The
individual patterns.
elevation and inferiorly in the late phase rotated 9 (range, 1 to 20) from 75 to
in 10 of the 12 shoulders (83%). Two maximum elevation (FIGURE 7A). Changes
shoulders showed continuous superior in humeral external/internal rotation
translation with humeral elevation. throughout the movement of humeral el-
evation were statistically significant (P =
External/Internal Rotation of the .001). Post hoc tests revealed significant
Humerus pairwise differences in humeral rotation
On average, 14 (range, 2 to 26) of hu- between the starting position and every
meral external rotation occurred from point between 30 and 120 of elevation (P
the starting position to 60 of humeral = .02 at 120 and P<.001 at other points).
elevation. The humerus then internally Humeral rotation at every point between

100 | february 2012 | volume 42 | number 2 | journal of orthopaedic & sports physical therapy

42-02 Matsuki.indd 100 1/18/2012 3:32:52 PM

 A plane fluoroscopy have recently appeared
120
110
Humeral External Rotation Angle, deg
100
90
80
70
Downloaded from www.jospt.org at on July 22, 2017. For personal use only. No other uses without permission.
60
Start
B head translates superiorly in the early
120
Copyright 2012 Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.
Humeral External Rotation Angle, deg
100
80
Journal of Orthopaedic & Sports Physical Therapy
60
40
Start
FIGURE 7. External/internal rotation of the humerus during shoulder scapular abduction. (A) Means and standard
deviations for all 12 shoulders. (B) The individual patterns.
45 and 90 of elevation was significantly
different from rotation at 135 (P = .005, P
= .001, P = .002, and P = .04, respectively)
and at maximum elevation (P = .005, P =
.001, P = .002, and P = .04, respectively).
The individual patterns of external rota-
tion are shown in FIGURE 7B. There was
great variation in the amount of humeral
external rotation at the initial position;
however, the humerus rotated externally
in the early phase of elevation and inter-
nally in the late phase in all shoulders, and
42-02 Matsuki.indd 101
in several published articles18,23,30; but
the accuracy of these techniques for the
shoulder has not yet been reported. Our
results comparing kinematics from bi-
plane images with metal bead markers
to single-plane images with bone surface
models were comparable with the pre-
viously reported accuracy for imaging
of the knee joint.27 These techniques al-
low kinematic analysis of dynamic joint
motion via conventional devices found
in most hospitals and provide accurate
kinematic data without physically inva-
sive procedures, such as insertion of bone
30 45 60 75 90 105 120 135 Max
pins.
Our results show that the humeral
phase of arm elevation, then inferiorly
in the late phase, consistent with our
hypothesis. Although there was great
variation in the absolute amount of trans-
lation, most shoulders showed similar
motion patterns. The large variation in
translations may be due to differences in
shoulder size and variability in definitions
of bone coordinate systems. The initial
amount of humeral rotation affects ten-
sion in the rotator cuff muscles, capsule,
and glenohumeral ligaments, which also
could affect glenohumeral translations.
In the early phase of elevation (the setting
phase), if rotator cuff muscles do not fully
30 45 60 75 90 105 120 135 Max
offset deltoid force, then dynamic stabil-
ity of the glenohumeral joint may not be
Humeral Elevation Angle, deg fully achieved14 and the humeral head
may translate superiorly.7,20 The result of
the present study seems to be consistent
with this concept. The humeral head may
position inferiorly at the starting position
11 of the 12 shoulders (92%) reached their due to gravity and translate superiorly in
maximum external rotation between 40 the early phase of elevation. Then, the
and 75 of humeral elevation. humeral head may be centralized and
FIGURE 8 shows 3-D renderings of a stabilized within the glenoid fossa as the
representative shoulder from the lateral rotator cuff muscles are activated. The
perspective. small amount of inferior translation in
the late phase of shoulder elevation may
DISCUSSION be an artifact related to the rotation of
the humerus. If the humeral coordinate
K
inematic analyses of the shoul- origin is not the center of rotation, there
der using 3-D/2-D model image will be apparent humeral translation with
registration techniques with single- humeral rotation. However, extreme care
journal of orthopaedic & sports physical therapy | volume 42 | number 2 | february 2012 | 101
1/18/2012 3:32:53 PM
 [ research report
Downloaded from www.jospt.org at on July 22, 2017. For personal use only. No other uses without permission.
Copyright 2012 Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.
Journal of Orthopaedic & Sports Physical Therapy
FIGURE 8. Three-dimensional kinematics of the glenohumeral joint viewed from a lateral direction at (A) 30, (B)
60, (C) 90, and (D) 120 of humeral elevation.
was taken to numerically find the loca-
tion of the best-fit sphere for the articular
surface of the humeral head, to place the
humeral origin at the center of rotation.
There have been only a few studies ad-
dressing external rotation of the humerus
relative to the scapula during shoulder
elevation. Hallstrm and Karrholm12
reported that 35 of external rotation of
the humerus occur during active abduc-
tion in healthy shoulders, as measured
by radiostereometric analysis. Ludewig
et al21 measured healthy shoulders using
electromagnetic tracking devices fixed to
bones with pins and documented 30 of
external rotation of the humerus during
scapular plane abduction. Our results
show that the humerus rotated 15 ex-
ternally from the starting position to 60
of humeral elevation, followed by a small
102 | february 2012 | volume 42 | number 2 | journal of orthopaedic & sports physical therapy
42-02 Matsuki.indd 102
reversal toward internal rotation. This is
not consistent with the results of previ-
ous studies nor with our hypothesis about
humeral rotation.
This discrepancy may be due to dif-
ferences in measurement methods,
participants (eg, age, gender, or ethnic
background), or the activity performed
(eg, speed of motion, external rotation
position of the humerus at the initial
position, or the plane of elevation). How-
ever, differences in anatomic coordinate
systems may be the biggest reason for
the discrepancy. Our anterior axis of the
humerus was directed anterolaterally
relative to the axis recommended by the
International Society of Biomechanics, 35
which might have had some effect on
the degree of rotation. We suspect that
details of the joint-angle calculation are
]
largely responsible for the difference,
because rotational movement is usually
decomposed into rotations about 3 mu-
tually orthogonal axes in a specific order
to describe 3-D motion. It is well docu-
mented that the choice of rotation se-
quence affects the computed rotations.16
Several studies have reported glenohu-
meral kinematics using the same rota-
tion sequence as that used in the present
study.4,19 Braman et al4 analyzed uncon-
strained elevation in healthy shoulders
using electromagnetic tracking devices
fixed to bones with pins and reported
that the humerus rotated 45 externally,
followed by a slight amount of internal
rotation. Ludewig and Cook19 studied
shoulder kinematics during scapular
plane abduction in subjects with and
without impingement syndrome, using
electromagnetic tracking devices at-
tached to the skin, and reported that the
humerus rotated 6 externally from 60
to 90 of humeral elevation, followed by
a small reversal in the direction of rota-
tion. Although the amount of humeral
external rotation differed among studies,
the studies we did, which employed the
same rotation sequence, showed a similar
pattern of humeral external/internal ro-
tation. In particular, the results reported
by Ludewig and Cook19 were quite simi-
lar to ours. These observations support
our speculation that choice of rotation
sequence is largely responsible for the
differences in humeral rotation among
studies. Recently, Bonnefoy-Mazure et
al3 studied 3 different rotation sequences
for shoulder motion analysis and found
that the sequence of abduction, flexion,
and external rotation was the best from
the perspective of both gimbal lock and
amplitude coherence. On this basis, we
chose the abduction-flexion-external ro-
tation sequence for this study.
We calculated external rotation of
the humerus using several common ro-
tation sequences, and the results varied
dramatically, with the amount of exter-
nal rotation ranging from 30 to 130 for
the same shoulder motion (FIGURE 9). The
rotation-sequence dependence of joint
1/18/2012 3:32:54 PM
 angles is an unavoidable consequence of
the method and demands that close at-
tention be paid to this choice when com-
paring results among studies.
Three study limitations should be con-
sidered. First, the participants were all
men within a small age range. Although
gender and age may affect glenohumeral
kinematics, we had to exclude younger
healthy individuals and women because
of exposure to ionizing radiation. Sec-
ond, 3-D/2-D model image registration
techniques using single-plane fluorosco-
py provide poor measurement accuracy
Downloaded from www.jospt.org at on July 22, 2017. For personal use only. No other uses without permission.
for out-of-plane translations.27 Because
all reported kinematic parameters are
insensitive to out-of-plane translation
errors, there should be no effect on the
kinematics we described. Biplane imag-
Copyright 2012 Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.
ing techniques provide smaller and more
uniform measurement errors,2 but single-
plane methods provide a less restricted
field of view and permit more natural
dynamic activities and less radiation ex-
posure. Third, variation in glenohumeral
translation was relatively large compared
to the magnitude of translation, and
this makes it more difficult to interpret
Journal of Orthopaedic & Sports Physical Therapy
the results. The variation could be due
largely to the initial humeral position.
Participants were instructed to move
their arms in the thumbs-up position,
but humeral external rotation can vary
with forearm pronation/supination. To
allow the participants to move as natu-
rally as possible, however, we did not
strictly control the position. Despite the
variation, we observed a biphasic pattern
of humeral superior/inferior translation
in most shoulders and believe that it is
representative of healthy shoulder kine-
matics during arm elevation.
In this study, we analyzed only
healthy shoulders. Our next step will
be to evaluate shoulders with disorders
such as shoulder impingement and ro-
tator cuff tears. It has been suggested
that superior translation and inadequate
external rotation of the humerus are as-
sociated with shoulder impingement syn-
drome.9,12,15,19,20,22 Contrasting pathologic
kinematics to those in healthy shoulders
journal of orthopaedic & sports physical therapy | volume 42 | number 2 | february 2012 | 103
42-02 Matsuki.indd 103
250
200
Humeral External Rotation Angle, deg
150
100
50
Start
FIGURE 9. The differences in external/internal rotation values calculated with 3 different rotation sequences.
This study used the z-x-y order.3,4,19 The y-z-y rotation order is recommended by the International Society of
Biomechanics.35
will provide an objective basis on which
to plan therapeutic interventions and as-
sess their outcomes. Ultimately, quanti-
tative assessment of dynamic shoulder
kinematics will lead to better treatment
of shoulder disorders.
CONCLUSION
W
e analyzed dynamic glenohu-
meral kinematics in healthy male
subjects using 3-D/2-D model
image registration techniques. The hu-
meral head translated superiorly in the
early phase of abduction and then inferi-
orly in the late phase. The humerus rotat-
ed externally from the starting position
to 60 of arm elevation and then rotated
slightly internally. These observations
will contribute to the understanding of
normal and pathological glenohumeral
kinematics. t
KEY POINTS
FINDINGS: The humeral head translated
superiorly in the early phase of scapular
plane abduction and then translated
30 45 60 75 90 105 120 135 Max
Humeral Elevation Angle, deg
z-x-y y-z-y y-x-z
inferiorly. The humerus rotated exter-
nally from the starting position to 60 of
arm elevation and then rotated slightly
internally.
IMPLICATION: Data on normal gleno-
humeral joint kinematics provide a
comparative basis to understand the
pathology of shoulder disorders related
to kinematic changes, such as shoulder
impingement syndrome.
CAUTION: This study was conducted
with 3-D/2-D model image registra-
tion techniques that have not yet been
fully validated with regard to accuracy
of measuring shoulder motions. In 3-D
kinematic studies, anatomic rotation
values are strongly affected by the math-
ematical rotational sequence used to
describe 3-D shoulder motions.
ACKNOWLEDGEMENT: We thank Hideyuki Kato,
RT for data collection. We also thank Shang
Mu, PhD for developing computer programs
to analyze the data.
1/18/2012 3:32:55 PM
 [ research report
REFERENCES
1. B  anks SA, Hodge WA. Accurate measurement of
three-dimensional knee replacement kinemat-
ics using single-plane fluoroscopy. IEEE Trans
Biomed Eng. 1996;43:638-649. http://dx.doi.
org/10.1109/10.495283
2. Bey MJ, Kline SK, Zauel R, Lock TR, Kolowich PA.
Measuring dynamic in-vivo glenohumeral joint ki-
nematics: technique and preliminary results. J Bio-
mech. 2008;41:711-714. http://dx.doi.org/10.1016/j.
jbiomech.2007.09.029
3. Bonnefoy-Mazure A, Slawinski J, Riquet A, Leveque
JM, Miller C, Cheze L. Rotation sequence is an
important factor in shoulder kinematics. Applica-
tion to the elite players flat serves. J Biomech.
2010;43:2022-2025. http://dx.doi.org/10.1016/j.
Downloaded from www.jospt.org at on July 22, 2017. For personal use only. No other uses without permission.
jbiomech.2010.03.028
4. Braman JP, Engel SC, Laprade RF, Ludewig PM. In
vivo assessment of scapulohumeral rhythm during
unconstrained overhead reaching in asymptomatic
subjects. J Shoulder Elbow Surg. 2009;18:960-967.
Copyright 2012 Journal of Orthopaedic & Sports Physical Therapy. All rights reserved.
http://dx.doi.org/10.1016/j.jse.2009.02.001
5. Chen SK, Simonian PT, Wickiewicz TL, Otis JC,
Warren RF. Radiographic evaluation of glenohumer-
al kinematics: a muscle fatigue model. J Shoulder
Elbow Surg. 1999;8:49-52.
6. Dayanidhi S, Orlin M, Kozin S, Duff S, Karduna A.
Scapular kinematics during humeral elevation in
adults and children. Clin Biomech (Bristol, Avon).
2005;20:600-606. http://dx.doi.org/10.1016/j.
clinbiomech.2005.03.002
7. Deutsch A, Altchek DW, Schwartz E, Otis JC, War-
ren RF. Radiologic measurement of superior dis-
Journal of Orthopaedic & Sports Physical Therapy
placement of the humeral head in the impingement
syndrome. J Shoulder Elbow Surg. 1996;5:186-193.
8. Freedman L, Munro RR. Abduction of the arm in
the scapular plane: scapular and glenohumeral
movements. A roentgenographic study. J Bone
Joint Surg Am. 1966;48:1503-1510.
9. Fu FH, Harner CD, Klein AH. Shoulder impingement
syndrome. A critical review. Clin Orthop Relat Res.
1991;162-173.
10. Graichen H, Hinterwimmer S, von Eisenhart-
Rothe R, Vogl T, Englmeier KH, Eckstein F. Effect
of abducting and adducting muscle activity on
glenohumeral translation, scapular kinematics
and subacromial space width in vivo. J Biomech.
2005;38:755-760. http://dx.doi.org/10.1016/j.
jbiomech.2004.05.020
11. Graichen H, Stammberger T, Bonel H, Karl-Hans E,
Reiser M, Eckstein F. Glenohumeral translation dur-
ing active and passive elevation of the shoulder - a
3D open-MRI study. J Biomech. 2000;33:609-613.
12. Hallstrm E, Karrholm J. Shoulder kinematics in
25 patients with impingement and 12 controls. Clin
Orthop Relat Res. 2006;448:22-27. http://dx.doi.
org/10.1097/01.blo.0000224019.65540.d5
13. Hamai S, Moro-oka TA, Miura H, et al. Knee
kinematics in medial osteoarthritis during in
vivo weight-bearing activities. J Orthop Res.
2009;27:1555-1561. http://dx.doi.org/10.1002/
jor.20928
104 | february 2012 | volume 42 | number 2 | journal of orthopaedic & sports physical therapy
42-02 Matsuki.indd 104
]
14. Inman VT, Saunders JBM, Abbott LC. Observations
on the function of the shoulder joint. J Bone Joint
Surg Am. 1944;26:1-30.
15. Kamkar A, Irrgang JJ, Whitney SL. Nonoperative
management of secondary shoulder impinge-
ment syndrome. J Orthop Sports Phys Ther.
1993;17:212-224.
16. Karduna AR, McClure PW, Michener LA.
Scapular kinematics: effects of altering the
Euler angle sequence of rotations. J Biomech.
2000;33:1063-1068.
17. Komistek RD, Dennis DA, Mahfouz M. In vivo
fluoroscopic analysis of the normal human knee.
Clin Orthop Relat Res. 2003;69-81. http://dx.doi.
org/10.1097/01.blo.0000062384.79828.3b
18. Kon Y, Nishinaka N, Gamada K, Tsutsui H, Banks
SA. The influence of handheld weight on the
scapulohumeral rhythm. J Shoulder Elbow Surg.
2008;17:943-946. http://dx.doi.org/10.1016/j.
jse.2008.05.047
19. Ludewig PM, Cook TM. Alterations in shoulder ki-
nematics and associated muscle activity in people
with symptoms of shoulder impingement. Phys
Ther. 2000;80:276-291.
20. Ludewig PM, Cook TM. Translations of the humerus
in persons with shoulder impingement symptoms.
J Orthop Sports Phys Ther. 2002;32:248-259.
21. Ludewig PM, Phadke V, Braman JP, Hassett DR,
Cieminski CJ, LaPrade RF. Motion of the shoulder
complex during multiplanar humeral elevation.
J Bone Joint Surg Am. 2009;91:378-389. http://
dx.doi.org/10.2106/JBJS.G.01483
22. Ludewig PM, Reynolds JF. The association of scap-
ular kinematics and glenohumeral joint patholo-
gies. J Orthop Sports Phys Ther. 2009;39:90-104.
http://dx.doi.org/10.2519/jospt.2009.2808
23. Mahfouz M, Nicholson G, Komistek R, Hovis D,
Kubo M. In vivo determination of the dynam-
ics of normal, rotator cuff-deficient, total, and
reverse replacement shoulders. J Bone Joint
Surg Am. 2005;87 Suppl 2:107-113. http://dx.doi.
org/10.2106/JBJS.E.00483
24. Mahfouz MR, Hoff WA, Komistek RD, Dennis DA.
A robust method for registration of three-dimen-
sional knee implant models to two-dimensional
fluoroscopy images. IEEE Trans Med Imaging.
2003;22:1561-1574. http://dx.doi.org/10.1109/
TMI.2003.820027
25. McClure PW, Michener LA, Karduna AR. Shoulder
function and 3-dimensional scapular kinematics
in people with and without shoulder impingement
syndrome. Phys Ther. 2006;86:1075-1090.
26. McClure PW, Michener LA, Sennett BJ, Karduna
AR. Direct 3-dimensional measurement of scapu-
lar kinematics during dynamic movements in vivo.
J Shoulder Elbow Surg. 2001;10:269-277. http://
dx.doi.org/10.1067/mse.2001.112954
27. Moro-oka TA, Hamai S, Miura H, et al. Can mag-
netic resonance imaging-derived bone models
be used for accurate motion measurement with
single-plane three-dimensional shape registration?
J Orthop Res. 2007;25:867-872. http://dx.doi.
@ MORE INFORMATION
org/10.1002/jor.20355
28. Moro-oka TA, Hamai S, Miura H, et al. Dynamic ac-
tivity dependence of in vivo normal knee kinemat-
ics. J Orthop Res. 2008;26:428-434. http://dx.doi.
org/10.1002/jor.20488
29. Nirschl RP. Rotator cuff tendinitis: basic con-
cepts of pathoetiology. Instr Course Lect.
1989;38:439-445.
30. Nishinaka N, Tsutsui H, Mihara K, et al. Determi-
nation of in vivo glenohumeral translation using
fluoroscopy and shape-matching techniques. J
Shoulder Elbow Surg. 2008;17:319-322. http://
dx.doi.org/10.1016/j.jse.2007.05.018
31. Poppen NK, Walker PS. Normal and abnormal
motion of the shoulder. J Bone Joint Surg Am.
1976;58:195-201.
32. Sahara W, Sugamoto K, Murai M, Tanaka H, Yoshi-
kawa H. The three-dimensional motions of gleno-
humeral joint under semi-loaded condition during
arm abduction using vertically open MRI. Clin
Biomech (Bristol, Avon). 2007;22:304-312. http://
dx.doi.org/10.1016/j.clinbiomech.2006.04.012
33. Teyhen DS, Miller JM, Middag TR, Kane EJ. Rota-
tor cuff fatigue and glenohumeral kinematics
in participants without shoulder dysfunction.
J Athl Train. 2008;43:352-358. http://dx.doi.
org/10.4085/1062-6050-43.4.352
34. Warner JJ, Micheli LJ, Arslanian LE, Kennedy J,
Kennedy R. Scapulothoracic motion in normal
shoulders and shoulders with glenohumeral insta-
bility and impingement syndrome. A study using
Moire topographic analysis. Clin Orthop Relat Res.
1992;191-199.
35. Wu G, van der Helm FC, Veeger HE, et al. ISB
recommendation on definitions of joint coordinate
systems of various joints for the reporting of hu-
man joint motion--Part II: shoulder, elbow, wrist
and hand. J Biomech. 2005;38:981-992.
36. Yamaguchi S, Gamada K, Sasho T, Kato H, Sonoda
M, Banks SA. In vivo kinematics of anterior cruci-
ate ligament deficient knees during pivot and
squat activities. Clin Biomech (Bristol, Avon).
2009;24:71-76. http://dx.doi.org/10.1016/j.
clinbiomech.2008.08.007
37. Yoshizaki K, Hamada J, Tamai K, Sahara R, Fuji-
wara T, Fujimoto T. Analysis of the scapulohumeral
rhythm and electromyography of the shoulder
muscles during elevation and lowering: compari-
son of dominant and nondominant shoulders. J
Shoulder Elbow Surg. 2009;18:756-763. http://
dx.doi.org/10.1016/j.jse.2009.02.021
38. Yushkevich PA, Piven J, Hazlett HC, et al.
User-guided 3D active contour segmenta-
tion of anatomical structures: significantly
improved efficiency and reliability. Neuroimage.
2006;31:1116-1128. http://dx.doi.org/10.1016/j.
neuroimage.2006.01.015
39. Zuckerman JD, Kummer FJ, Cuomo F, Simon J,
Rosenblum S, Katz N. The influence of coracoacro-
mial arch anatomy on rotator cuff tears. J Shoulder
Elbow Surg. 1992;1:4-14. http://dx.doi.org/10.1016/
S1058-2746(09)80010-4
WWW.JOSPT.ORG
1/18/2012 3:32:55 PM