e s
co
Ma
E.K.J. Chadwick a, F.C.T. van der Helm a
a
Interpretation. The ability to perform, or not perform a task appeared to be related to a compensatory movement implementation
ohumeral elevation angle of 100 (Gill et al., 1999;
Levy and Copeland, 2001; Torchia et al., 1997), where
have investigated (2D) motion patterns after shoulder
arthroplasty (Barrett et al., 1987; Boileau et al., 1992;
Friedman, 1995) during arm elevation. These stud-
ies found an increase in scapular motion contribut-
ing to arm elevation, or a 2:3 relationship for the
.
* Corresponding author.
E-mail address: h.e.j.veeger@wbmt.tudelft.nl (H.E.J. Veeger).
Clinical Biomechanics 21 (20268-0033/$ - see front matter
2005 Elsevier Ltd. All rights reservedby means of clavicular retraction. It is concluded that the functional outcome after arthroplasty is limited due to a lack of gleno-
humeral Range of Motion but that it is possible to compensate for this restriction.

2005 Elsevier Ltd. All rights reserved.
Keywords: Shoulder; ADL; Arthroplasty; Kinematics; Range of motion; Rheumatoid arthritis
1. Introduction
Shoulder arthroplasty is generally recognized as suc-
cessful in terms of pain relief, with an improved, but still
limited Range of Motion (RoM). Patients with rheuma-
toid arthritis achieve an average post-operative thorac-
160 might be expected for healthy subjects. Function
is also still limited: after shoulder arthroplasty 75% is
able to perform perineal care, but only 55% are able
to comb their hair and to reach shoulder level. How
these poor results relate to limitations in glenohumeral
motion is not clear. To our knowledge, only few studiesDelft University of Technology, Design, Construction and Production, Mechanical Engineering, Man Machine Systems,
Mekelweg 2, 2628 CD Delft, The Netherlands
b Institute of Fundamental and Clinical Movement Sciences, Department of Human Movement Sciences,
Vrije Universiteit Amsterdam, The Netherlands
c Leiden University Medical Center, The Netherlands
Abstract
Background. After shoulder arthroplasty, post-operative Range of Motion is usually compromised. It is, however, unclear to
what extent limitations in Range of Motion are related to functional outcome in terms of Activities of Daily Living.
Methods. The upper extremity motions of 13 patients (16 shoulders) and a control group (N = 24) during four Range of Motion
tasks and Activities of Daily Living were measured using a six degree-of-freedom electromagnetic tracking device. Based on the
results for the Activities of Daily Living task
hair combing, the patient groups was divided into a group that could perform this
task (
Able, N = 8, 10 shoulders) and a group that could not perform the task (
Unable, N = 6, six shoulders).
Results. Both patient groups showed considerable limitation in glenohumeral Range of Motion, when compared to controls, but
between patient groups only axial rotation Range of Motion was different: the
Able group has a larger external rotational Range of
Motion, but less internal rotation. During
combing hair the Able group appeared to successfully perform the task through a larger
clavicular retraction.A kinematical analysis of th
during a hair
H.E.J. Veeger a,b,*, D.J.doi:10.1016/j.clinbiomech.2005.09.012houlder after arthroplasty
mbing task
germans a, J. Nagels c,
www.elsevier.com/locate/clinbiomech
006) S39–S44

healthy control group (N = 24) participated in this
study. The patient groups mean age was 66.6 ± 16.2
Biomyears, versus 36.8 ± 11.8 years for the controls, which
was significantly older (t = 7.7, p < 0.00). Seven shoul-
ders had undergone a Hemi Shoulder Arthroplasty
(HSA), nine shoulders a Total Shoulder Arthroplasty
(TSA). All subjects gave written informed consent prior
to the experiment.
2.2. Measurement device
A six degree-of-freedom electromagnetic tracking
device, the Flock of Birds (Ascension Technology Inc.,
Burlington, Vermont, USA) was used for the recording
of kinematic data. We used a set-up with five sensors,
which comprised a 60-mm long pointer, and sensors fix-
ated to the sternum, arm, forearm and a scapula locator
(Johnson et al., 1993). The sensor on the sternum was
attached to the skin with double-sided adhesive tape.
Arm and forearm sensors were fitted on a cuff. All three
sensors were additionally covered with Fixomull stretch
self-adhesive bandage (Beiersdorf AG, Hamburg, Ger-
many). The arm sensor was attached on the dorso-lat-
eral side of the distal humerus. The forearm sensor
was attached dorsally on the distal forearm, approxi-scapula–humeral rhythm, apparently compensating for
a loss in glenohumeral elevation, instead of 1:3 relation-
ship as found in healthy subjects (Groot et al., 1998). In
a previous study on the movement requirements during
selected Activities of Daily Living (Magermans et al.,
2005) we concluded that the reasons for the poor results
for hair combing were, however, most likely not found
in a limitation in elevation, but might be due to the large
external glenohumeral rotation observed during this
task in healthy controls.
The aim of this study is to determine the relationship
between RoM and the ability to perform combing hair
after shoulder arthroplasty and to compare these with
results from a study on healthy subjects (Magermans
et al., 2005). It is hypothesized that despite the reduction
in pain, glenohumeral elevation is still limited after
implantation of a prosthesis, when compared to a
healthy norm population. In addition, we expect that
differences between patients that are and are not able
to perform a key task such as combing hair will become
visible in differences in glenohumeral external rota-
tion as quantified during active Range of Motion
measurements.
2. Materials and methods
2.1. Subjects
Thirteen patients (N = 13, 16 shoulders) and a
S40 H.E.J. Veeger et al. / Clinicalmately 5 cm proximal to the ulnar and radial styloid.To link the sensors to local anatomical coordinate
systems, 16 bony landmarks were digitized relative to
their sensors using the pointer. These measurements
were performed with the subjects in a standing position
with the arms hanging aside the body. From the combi-
nation of local coordinate systems constructed from
these anatomical landmarks and the sensor motions,
both segment and subsequently also joint rotations
could be calculated with an inter-subject variability of
no more than seven degrees for all three scapula rota-
tions (Meskers et al., 1998). Local coordinate systems,
segment and joint rotations were defined following the
ISB standardization proposal for the upper extremity
(Wu et al., 2005). To correct for underestimation of
the axial rotation of the humerus due to sensor-to-skin
displacement the axial rotation of the humerus was
determined by fitting the orientation of the forearm
sensor relative to the arm sensor to a two degree-of-
freedom arm model.
2.3. Procedure
Four Range of Motion (RoM) tasks: forward flexion,
abduction, internal and external rotation with humeral
elevation, and six Activities of Daily Living (ADL) were
measured. From these ADL, the activity
combing hair
was selected for further analysis. Since dynamic tracking
of the scapula is very difficult, measurements were per-
formed in a quasi-static mode, implying that after a
few practice session, the motion was performed in up
to ten steps, during each of which the sensor orienta-
tions, including the orientation of the scapula, were
recorded.
2.4. RoM
For most RoM tasks the subjects were instructed to
reach a maximal joint angle. This means that e.g. for
the forward flexion and abduction tasks, the subject
was instructed to elevate the arm as high as possible.
Internal rotation is defined as positive axial rotation of
the humerus and external rotation as negative axial rota-
tion. The axial rotation task and the pronation task
started differently from the other tasks. The internal
rotation with scapular abduction started with 90 of
humeral elevation, the humerus making an angle of
30 with the frontal plane (scapular plane) and in max-
imum external rotation. The pronation task started with
90 of elbow flexion and the forearm maximally
supinated.
2.5. ADL
Subjects were instructed to start in a neutral position
with the arms hanging beside the body, but were free to
echanics 21 (2006) S39–S44choose their way of performance. Within the control

3. Results
3.1. RoM
In general, glenohumeral RoM was reduced for the
patient groups in comparison to the healthy population
(Fig. 1). Glenohumeral elevation (i.e. both forward flex-
ion and abduction) was approximately 85 for the con-
trol group versus 35–55 for the patient groups.
Maximal elevation was not different between the Able
and Unable group.
There were no significant differences between the
healthy controls and Able patient group for internal
and external rotation, but external rotation was signifi-
cantly reduced for the Unable group in comparison to
both the controls and the Able group. The Able group
showed a peak external rotation of
67 ± 18 versus

38 ± 46 for the Unable group. Peak internal rotation
was significantly smaller for both patient groups com-
Biomechanics 21 (2006) S39–S44 S41population, the right shoulder was measured. For
patients, the operated shoulder was measured. In case
a patient had two implants, both shoulders were mea-
sured. For calculations all left shoulders were mirrored
to the right. All subjects performed the ADL without
objects.
2.6. Angle definitions
In this study the motions of clavicle, scapula and of
the glenohumeral joint were taken into account. Joint
angles were defined based on the International Society
of Biomechanics standardization proposal of the Inter-
national Shoulder Group (Wu et al., 2005). Glenohu-
meral and thoracohumeral motions are then described
according to the globe system (Doorenbosch et al.,
2003): ‘‘plane of elevation’’ – ‘‘elevation’’ – ‘‘internal/
external rotation’’. For the scapula this implied a rota-
tion order ‘‘pro/retraction’’ – ‘‘latero/mediorotation’’ –
‘‘tipping backward/forward’’. For the clavicle the order
was ‘‘pro/retraction’’ – ‘‘elevation/depression’’ – ‘‘inter-
nal/external rotation’’.
The forearm angles are rather straightforward: elbow
flexion and pronation, where 0 coincides with the
anatomical position.
2.7. Data analysis
On the basis of the ADL results, the patient popula-
tion was divided into two groups: a patient group
(N = 7) which was able to perform the hair combing
task and a patient group (N = 6) which was not
able to reach the target position. Mean RoM values
for the glenohumeral joint were compared between the
control group and the two patient groups on the basis
of a simple ANOVA. Using a LSD post-hoc test,
differences among the three groups (controls, able
patients and unable patients) could subsequently be
identified.
For the ‘‘hair combing’’ task mean peak values for
clavicular, scapular and humeral motions were calcu-
lated, as well as the 5–95% percentiles, assuming a nor-
mal distribution. Since, however, peak values cannot
uniquely describe differences between the Able and
Unable group, because the Unable group did, by defi-
nition, not finish the task, additional information on
the mean motion patterns between groups was
obtained using a B-spline fitting method with penalties
(Eilers and Marx, 1996). This fitting method can also
be used as an exploratory method to find statistical dif-
ferences in the motion patterns of different groups. The
95% confidence intervals of the mean motion patterns
were calculated and when these intervals did not inter-
sect each other, it may be assumed that motion
patterns differ significantly during that part of the
H.E.J. Veeger et al. / Clinicaltrajectory.pared to the healthy controls (
8 ± 17 and 4 ± 33
for the patient groups versus 36 ± 26 for the controls,
F = 12.0, p < 0.05). The negative internal rotation for
the Able group implied that this group did on average
reach slightly externally rotated values.
3.2. ADL
Combing hair required a thoracohumeral elevation
angle of about 100–120 (Fig. 2, x-axes) and was per-
formed with about a peak glenohumeral angle between
73 and 102 (95% confidence interval, Fig. 3) for the
healthy group versus 46–67 for the Able group
(Fig. 3). The Unable patient group managed to reach
37–62 glenohumeral elevation, but did of course not
finish the task. As expected, these peak elevation values
were significantly different between groups (p < 0.000),
-100
-80
-60
-40
-20
0
20
40
60
80
100
Healthy
Able
Unable
Forward flexion Internal rotationExternal rotationAbduction
An
gl
e
(de
gre
e
s)
Glenohumeral Range of Motion angles
*
*
*
*
*
*
*
*
Fig. 1. Glenohumeral RoM for healthy subjects and for both patient
groups (Able: able to comb their hair, Unable: not able to comb hair).
Asterisks indicate significant differences between groups.

S42 H.E.J. Veeger et al. / Clinical Biomechanics 21 (2006) S39–S44but trajectories did, however, not differ significantly
between groups (Fig. 2C). Peak glenohumeral external
Fig. 2. Mean motion patterns with 95% confidence intervals for clavicular p
and glenohumeral axial rotation (D), for healthy subjects and subjects who w
were not (Unable patients).
-100
-50
0
50
100
150
Able
Laterorotation Elevation External rotationProtraction
ScapulaClavicle Humerus Humerus
An
gl
e
(˚)
95% percentiles for peak angles during
the "Combing Hair" task
*
*
*
*
*
*
*
*
Healthy
Unable
Fig. 3. Fifth to 95th intervals for the peak angles reached during the
task ‘‘combing hair’’. The healthy control group (N = 24) and Able
patient group (N = 10) were able to perform the task, the Unable
group could not.rotation during hair combing differed significantly
between the three groups (Fig. 3). This was also visible
in the glenohumeral axial rotation trajectories
(Fig. 2D). The Unable group showed on average less
rotraction (A), scapular laterorotation (B), glenohumeral elevation (C)
ere able to perform the comb hair task (Able patients) and those who
Sternoclavicular joint
*
*
*
*
-80
-60
-40
-20
0
20
40
60
80
Healthy
Able
Unable
Retraction Elevation
An
gl
e
(de
gre
e
s)
Peak angles in combing hair
Axial rotation
Fig. 4. Results for peak angles in the sternoclavicular joint. Asterisks
indicate significant differences between groups.

plane of elevation did not differ between both patient
co-contraction of alternative muscles, which would sub-
sequently limit glenohumeral motion.
Biomechanics 21 (2006) S39–S44 S43groups (47 ± 14 for the Able group, versus 51 ± 18
for the Unable group). Also, trajectories did not differ.
It is therefore not likely that this alternative explanation
is valid.
Difference between both patient groups became
particularly visible in the sternoclavicular joint. The suc-
cessful patient group showed significantly more clavicu-
lar retraction than the patient group that was unable to
comb their hair. This compensatory motion appears to
be necessary to allow for additional thoracohumeral
external rotation and therefore compensate for the
lower than normal external rotation RoM and glenohu-external rotation (
29 ± 43) than the Able group
(
59 ± 17), or the healthy controls (
70 ± 19).
Scapular laterorotation (Fig. 2B) was not different
between groups. A large difference was, however, found
in protraction/retraction of the clavicle (Figs. 2A and 3).
Able patients appeared to retract the clavicle about 15
further than the Unable group. This can also be seen in
the peak retraction values reached during the motion:
Clavicular retraction was
55 ± 6 for the Able group
versus
43 ± 5 for the group that was unable to per-
form the task (Fig. 4).
4. Discussion
When compared with results for our control group,
glenohumeral RoM was lower for both patient groups.
In contrast to our expectation, there was no significant
difference in glenohumeral elevation between both
patient groups. Peak glenohumeral external differed,
however, significantly between both patient groups.
Patients that were able to perform the combing hair task
appeared to have more external rotation than the unable
group (Fig. 1). This might be one of the explanations for
the difference in task performance for the ‘‘combing
hair’’ task. It is likely that these patients could keep their
arm sufficiently externally rotated to perform the comb-
ing task, especially since their peak external rotation
angle was significantly smaller that for the healthy
group. This difference suggests that either external rota-
tion reached by the able patient group is the minimally
necessary rotation, or that the
Able’’ subjects show
other compensatory motions.
A possible alternative explanation for the observed
difference in external rotation during hair combing
might have been the effect of the interdependency of
Euler decompositions. In this case, a more sagittal (clo-
ser to 90) plane of elevation for the humerus would
require less external rotation of the humerus. This effect
is found for the total population, as indicated by a cor-
relation of
0.42 (p < 0.05) between peak plane of eleva-
tion and peak external rotation. However, the peak
H.E.J. Veeger et al. / Clinicalmeral elevation RoM.Another possible explanation for why glenohumeral
RoM might be limited is that due to the implantation
of the prosthesis the glenohumeral rotation centre with
respect to the humeral shaft was changed. According
to de Leest et al. (1996) the retroversion angle is an
important aspect in the positioning of the rotation cen-
tre. A change in orientation of the humeral head pros-
thesis would cause a change in the moment arms of
muscles that cross the glenohumeral joint.
Glenohumeral external rotation is important for
activities of daily living that require high glenohumeral
elevation angles. After shoulder arthroplasty the
amount of glenohumeral motion was found to be
restricted, but it appeared to be possible to compensate
for this limitation, in this case by means of clavicular
retraction. It must be taken into account that compen-
sating strategies could of course cause secondary prob-
lems in other joints, which will ultimately affect total
motion. Therefore, improving glenohumeral RoM
should of course remain the most important aspect in
shoulder arthroplasty. To assist the clinician in diagno-
sis and evaluation, it is advisable to measure glenohu-
meral RoM instead of thoracohumeral RoM, since
this will give more insight into joint status and possible
functional outcome. Future research should indicate
why glenohumeral RoM is limited and how it can be
improved. Important aspects that need attention will
be the effect of muscle status and the version angle of
the prosthesis implant on functional outcome.
References
Barrett, W.P., Franklin, J.L., Jackins, S.E., Wyss, C.R., Matsen III,
F.A., 1987. Total shoulder arthroplasty. J. Bone Joint Surg. [Am]
69, 865–872.
Boileau, P., Walch, G., Liotard, J.P., 1992. [Radio-cinematographic
study of active elevation of the prosthetic shoulder]. Rev. Chir.
Orthop. Reparatrice Appar. Mot. 78, 355–364.
de Leest, O., Rozing, P.M., Rozendaal, L.A., Van der Helm, F.C.T.,
1996. Influence of glenohumeral prosthesis geometry and place-
ment on shoulder muscle forces. Clin. Orthop., 222–233.
Doorenbosch, C.A.M., Harlaar, J., Veeger, H.E.J., 2003. The globe
system: an unambiguous description of shoulder positions in dailyA possible explanation for the observed limitation in
glenohumeral RoM might be that rotator cuff muscles
are atrophied to such an extent that glenohumeral
motion is not possible. Passive RoM measurements
would have been useful to obtain insight into the maxi-
mal achievable glenohumeral motion. If passive RoM
had been higher than active RoM this would mean that
more glenohumeral motion is in principle possible. Also,
lack of rotator cuff force might have resulted in reduced
stability of the joint. Stabilizing the glenohumeral joint
with insufficient rotator cuff force will probably causelife movements. J. Rehabil. Res. Dev. 40 (2), 147–156.

Eilers, P.H., Marx, B.D., 1996. Flexible smoothing with B-splines and
penalties. Stat. Sci. 11 (2), 89–121.
Friedman, R.J., 1995. Biomechanics of total shoulder arthroplasty: a
preoperative and postoperative analysis. Semin. Arthroplasty 6,
222–232.
Gill, D.R., Cofield, R.H., Morrey, B.F., 1999. Ipsilateral total shoulder
and elbow arthroplasties in patients who have rheumatoid arthritis.
J. Bone Joint Surg. [Am] 81, 1128–1137.
Groot, J.H., Valstar, E.R., Arwert, H.J., 1998. Velocity effects on the
scapula–humeral rhythm. Clin. Biomech. 13, 593–602.
Johnson, G.R., Stuart, P.R., Mtchell, S., 1993. A method for the
measurement of three dimensional scapular movement. Clin.
Biomech. 8, 269–273.
Levy, O., Copeland, S.A., 2001. Cementless surface replacement
arthroplasty of the shoulder. 5- to 10-year results with the
Copeland mark-2 prosthesis. J. Bone Joint Surg. [Br] 83, 213–221.
Magermans, D.J., Nagels, J., Chadwick, E.K.J., Veeger, H.E.J., Van
der Helm, F.C.T., 2005. Requirements for upper extremity motions
during activities of daily living. Clin. Biomech. 20, 591–599.
Meskers, C.G., Vermeulen, H.M., Van der Helm, F.C.T., De Groot,
J.H., Rozing, P.M., 1998. 3D shoulder position measurements
using a six-degree-of-freedom electromagnetic tracking device.
Clin. Biomech. 13, 280–292.
Torchia, M.E., Cofield, R.H., Settergren, C.R., 1997. Total shoulder
arthroplasty with the Neer prosthesis: long-term results. J. Shoul-
der Elbow Surg. 6, 495–505.
Wu, G., Van der Helm, F.C.T., Veeger, H.E.J., Makhsous, M., Van
Roy, P., Anglin, C., Nagels, J., Karduna, A.R., McQuade, K.,
Wang, X., et al., 2005. ISB recommendation on definitions of joint
coordinate systems of various joints for the reporting of human
joint motion—Part II: Shoulder, elbow, wrist and hand. J.
Biomech. 38, 981–992.
S44 H.E.J. Veeger et al. / Clinical Biomechanics 21 (2006) S39–S44