Asprain or separation of the acromioclavicular joint is a
common injury in young, athletic individuals and typically occurs as
a consequence of direct trauma to the superior aspect of the shoulder1-4. The mechanism of injury usually
includes inferior and anterior translation of the acromion with
respect to the distal aspect of the clavicle. The resultant partial
or complete rupture of the acromioclavicular joint capsule is classified
as either type I or type II4,
depending on the degree of capsular disruption. Almost 45% of
all shoulder injuries sustained by athletes are one of these types5. With such injuries, the coracoclavicular
ligaments remain intact.
More than thirty-five different types of treatment for these injuries,
including compressive bandages, slings, tape, braces, harnesses,
and other traction techniques, have been reported. Many patients
have persistent pain after the injury because of posttraumatic osteolysis
of the clavicle or torn capsular ligaments trapped in the joint
space. Clinical studies6,7 have
suggested that untreated type-II injuries may lead to more chronic
disability than was previously recognized. Those investigations
showed that 23% to 35% of patients with a type-II
injury had nuisance symptoms and that 13% to 42% of
the patients had important symptoms from six months to five years
following the injury. Some patients have also been found to have
residual pain and stiffness, with a 24% decrease in horizontal
abduction strength at high rates of joint motion8.
Improvement in the treatment and rehabilitation protocols for these
injuries requires a comprehensive understanding of the contribution
of each structure to overall joint function in the intact and injured
states.
Previous studies4,9-11 have
demonstrated that the acromioclavicular ligaments control anterior-posterior
stability, while the coracoclavicular ligaments control superior-inferior
stability. More recently, other investigations12,13 have
shown that all of the soft tissues at the acromioclavicular joint
function synergistically, in a complex manner, to provide joint
stability. Thus, traumatic disruption of the acromioclavicular joint
capsule is thought to result in abnormal joint kinematics and load
transmission, factors that increase the possibility of postinjury
pain, instability, and degenerative joint disease14.
However, the effect of capsular disruption on the translations at
the articular surface and the degree to which the intact coracoclavicular
ligaments compensate for loss of the acromioclavicular ligaments
have not been thoroughly investigated on a quantitative basis.
Therefore, the objective of the present study was to quantify the
effect of transection of the acromioclavicular joint capsule on
the kinematics and the in situ forces in the coracoclavicular ligaments
under externally applied anterior, posterior, and superior loading
conditions. We hypothesized that the coracoclavicular ligaments
cannot adequately compensate for the absence of the acromioclavicular
joint capsule in restraining anterior-posterior translation and
that such an alteration in joint motion leads to an increase in
the in situ forces of the coracoclavicular ligaments.
Eleven fresh cadaveric shoulders from donors who were fifty-seven
to eighty-one years old at the time of death were wrapped in gauze
that had been soaked in saline solution and stored in plastic bags
at —20°C. Prior to the day of the test, each specimen was
thawed overnight at room temperature. Each shoulder was disarticulated
at the glenohumeral joint and dissected free of all soft tissues
except for the acromioclavicular joint capsule and the coracoacromial
and coracoclavicular ligaments. Specimens that showed evidence of degenerative
joint disease or previous injury on gross and radiographic examination
were removed from the study. The clavicle was then potted in epoxy
putty, secured within a thick-walled aluminum cylinder, and fixed
in a custom clamp mounted to the base of the robotic manipulator
(Fig. 1).
The scapula was fixed in a block of epoxy putty and rigidly attached
to the end-effector of a robotic manipulator through a specially
designed clamp. The clavicle and scapula were each potted in as
much epoxy putty as possible to prevent deformation of these bodies
during joint motion.
A robotic/universal force-moment sensor testing system
was utilized to determine the motion of the joint in response to externally
applied loads15-17. The robot
is a six-axis serial-articulated manipulator (Puma 762; Unimate,
Duncan, South Carolina) with a repeatability of 0.2 mm and 0.2°
for position and orientation, respectively. Force-moment data collection
was achieved with use of a universal force-moment sensor (model
4015; JR3, Woodland, California), which has a repeatability of better
than 0.2 N and 0.1 Nm for forces and moments, respectively.
This testing system was able to operate in two modes to determine
the joint kinematics and the in situ forces in the ligaments. In
the force-control mode, the testing system determined the joint
kinematics in response to applied loads, while individual degrees
of freedom were restricted to allow joint motion only in specified
planes. With use of the same loading conditions, the kinematic changes
resulting from transection of the acromioclavicular joint capsule
were studied. The paths of motion determined for both the intact
and the transected-capsule conditions were saved and were reproduced
with use of the testing system in position-control mode. This mode
allowed application of the principle of superposition to determine
the magnitude and direction of the in situ force for portions of
the capsule or the coracoclavicular ligaments. With use of this
principle, the difference in force measured by the universal force-moment
sensor before and after cutting of a ligament represented the force
in that ligament. This methodology also required the clavicle and
scapula to be effectively rigid compared with the soft tissue around
the acromioclavicular joint. This assumption was found to be reasonable
for load levels of up to 70 N during preliminary experiments with
use of specimens that possessed good cortical bone, on the basis
of radiographic evaluation before testing, and were free from any
observable disease.
A coordinate system associated with the scapula was used to describe
motion of the clavicle with respect to the scapula18,19. The x axis was perpendicular
to the scapular plane and directed anteriorly. The y axis was parallel
to the scapular plane and directed superiorly. The z axis was directed
medially and obtained from the cross product of the x and y axes. The
origin of the coordinate system was located at the center of the
articular surface on the medial aspect of the acromion. A Euler
angle system was used to describe the motion of the clavicle with
respect to the scapula. The first rotation was about the long axis
of the clavicle and corresponded to axial rotation. The second rotation,
about the y axis, corresponded to protraction-retraction, and the
final rotation, about the x axis, corresponded to elevation.
An experimental protocol was developed to obtain an axial rotation
position of the clavicle that served as a standard reference position
for all loading tests. The joint was initially positioned in the
testing system at 0° of elevation and 0° of protraction-retraction.
Elevation and protraction-retraction were held constant throughout
the entire experimental protocol while the scapula was free to translate
along all three axes in order to maintain contact between the distal
end of the clavicle and the acromion. Force control was then used
to apply a 10-N compressive load (medially directed) to the clavicle while
the forces in the two orthogonal directions were minimized. While
the previously described force conditions were maintained, the clavicle
was axially rotated in the positive and negative directions (in
1 increments) until the angle of rotation that minimized the moment
about the longitudinal axis of the clavicle was achieved. The testing
system learned and then recorded the position of the joint at this
axial rotation angle.
Each loading test then applied a maximum of 70 N to the scapula
in the anterior, posterior, or superior direction with use of the
previously obtained axial rotation position as the starting position
for all tests. During the loading protocol, the testing system attempted
to satisfy two force targets: 10 N of joint compression, to maintain
contact between the distal end of the clavicle and the acromion,
and 10% increments, to the maximum load of 70 N. The scapula
was allowed to translate along each of the three axes to meet the
required force targets while the rotational degrees of freedom were
held constant. These constraints were placed on the joint motion
to simulate injury mechanisms and to obtain repeatable results (the
acromioclavicular joint has a high laxity in two rotational degrees of
freedom). The testing system recorded the anterior-posterior, superior-inferior,
and medial-lateral translations of the intact joint resulting from
the application of each load, while the resultant forces and moments
at each loading position were recorded by the universal force-moment
sensor.
To assess possible interaction between the superior and inferior
acromioclavicular ligaments, we initially separated them horizontally
along the anterior and posterior aspects of the joint. The previously
determined paths of motion of the intact acromioclavicular joint
were repeated by the testing system while operating in the position-control
mode. A new set of forces and moments was measured by the universal
force-moment sensor for each increment of loading in each of the three
directions. The difference in forces before and after separation
of the acromioclavicular joint capsule indicated the amount of interaction
between these structures during application of each loading condition.
The superior and inferior acromioclavicular ligaments were then
sequentially transected by a scalpel in random order. The previously
determined paths of motion for the intact acromioclavicular joint
were repeated by the testing system, as already described, and a
new set of forces and moments was measured by the universal force-moment
sensor for each loading condition. The difference in force between
these two tests represented the in situ forces in the acromioclavicular
capsule and its ligamentous thickenings.
Externally applied loads of 70 N in the anterior, posterior,
and superior directions were then applied to the transected-capsule
specimen, and the resulting kinematics were recorded for each loading
condition. The conoid and trapezoid ligaments were then sectioned
in random order. After each structure was cut, the previously determined
kinematics of both the intact and the transected-capsule condition
were repeated by the robot for each loading condition. For each
increment of joint motion, the universal force-moment sensor measured
a new set of force and moment data. Once again, the decrease in force
observed by the universal force-moment sensor between these two
tests with identical shoulder positions represented the in situ
force in each of the coracoclavicular ligaments.
The data obtained from this experimental protocol included the
acromioclavicular joint kinematics and the in situ forces in the
conoid and trapezoid, in both the intact and transected-capsule
shoulders, for each loading condition. The in situ forces in the
superior and inferior capsular ligaments were also determined during
the three loading conditions for the intact shoulder. Statistical
analysis was performed with use of a two-factor repeated-measures
analysis of variance to assess the effects of loading and the joint
condition (intact or after transection of the capsule) on the amount
of translation in the direction of loading. A two-factor repeated-measures
analysis of variance was utilized to assess the effect of joint
condition and ligament on the magnitude of the in situ force in
the coracoclavicular ligaments. Both of these analyses were followed by
multiple contrasts, and the significance was set at p < 0.05.
Transection of the acromioclavicular joint capsule resulted in a
nearly 100% increase (p < 0.05) in anterior translation
(6.4 mm) and posterior translation (3.6 mm), while capsule transection
had no significant effect on superior translation, which increased
by only 1.6 mm (p > 0.05, power = 85%)
(Fig. 2).
After transection of all soft-tissue structures, the forces attributable
to osseous contact were found to be much smaller than the applied
loads. This finding indicated that the clavicle moved smoothly over
the articular facet of the acromion. Coupled translations were also
found to occur in response to these externally applied loads.
The magnitude of the difference between forces measured before
and after separation of the capsule into its superior and inferior
components was found to be <10 N in response to an anterior
or posterior load. However, in response to a superior load of 70
N, the magnitude of the difference between forces reached almost
20 N because of the separation of the capsule. This finding suggests
that there is some interaction between the superior and inferior
acromioclavicular ligaments with the application of a superior load.
The effect of capsule transection on the in situ forces in the coracoclavicular
ligaments was significant (p < 0.05) when each specimen
was subjected to an anterior or posterior load but not when it was
subjected to a superior load. With the capsule intact, the mean
in situ force (and standard deviation) in the superior acromioclavicular
ligament (35 ± 18 N) was larger than that in either
the conoid ligament (15 ± 14 N) or the trapezoid
ligament (14 ± 14 N) (p < 0.05) in response
to an anterior load of 70 N (Fig. 3, A). Transection of the capsule
resulted in a significant increase (p < 0.05) in the mean
in situ force in the conoid ligament (to 49 ± 23
N) of >200%. The mean force in the conoid ligament
was also significantly greater (p < 0.05) than that in
the trapezoid (25 ± 19 N) after transection of the
capsule.
In contrast, no significant difference in the in situ forces between
the acromioclavicular and coracoclavicular ligaments could be demonstrated
in response to a posterior load of 70 N with the capsule intact
(Fig. 3, B).
However, after transection of the capsule, the mean in situ force
in the trapezoid significantly increased (p < 0.05) from
23 ± 15 N to 38 ± 23 N, or 66%,
in response to the posterior load. The resultant force in the trapezoid
was also found to be significantly (approximately 50%) greater
than that in the conoid (24 ± 22 N) (p < 0.05), which
had increased only 9%.
Under loads applied in the superior direction, the magnitude of
the in situ force was greatest in the conoid ligament for both the
intact and the transected-capsule conditions. The force in the conoid
was found to be approximately 50% greater (p < 0.05)
than the forces in the superior and inferior acromioclavicular ligaments
and that in the trapezoid. In contrast to the other loading conditions,
the in situ force did not increase significantly in either of the
coracoclavicular ligaments in response to a 70-N load applied in
the superior direction after transection of the capsule (Fig. 3, C).
The degree to which a ligament served as a restraint against
an applied load in a given direction was assessed by determining the
percentage of the in situ force contributed by the vector component
in the direction of loading. In response to an anterior load of
70 N, the superior acromioclavicular ligament was found to contribute
the largest amount (p < 0.05) of ligamentous restraint
(mean and standard deviation, 46% ± 26%)
to the intact shoulder (Table I). After transection of the capsule,
however, the conoid contributed the largest amount (p < 0.05)
of ligamentous restraint (58% ± 33%)
against an anterior load. In contrast, no significant differences
in ligamentous restraint against a posterior load of 70 N could
be found for either of the acromioclavicular or coracoclavicular
ligaments in the intact shoulder. After transection of the capsule,
the trapezoid provided a significantly greater (p < 0.05)
amount of ligamentous restraint than did the conoid. The percent
contribution of the trapezoid to posterior restraint also increased
(p < 0.05) from the intact to the transected-capsule condition.
In response to a superior load of 70 N, the conoid demonstrated
the greatest contribution of any ligament in both the intact and
the transected-capsule condition, contributing nearly 70% in
the transected-capsule shoulders.
The robotic/universal force-moment sensor testing system allowed
simultaneous quantification of the in situ forces in the acromioclavicular
and coracoclavicular ligaments and of joint motion during three-degrees-of-freedom
joint motion in response to anterior, posterior, and superior loading
conditions. The results of this study confirmed our hypothesis that disruption
of the acromioclavicular capsule results in a significant increase
in translation in the anterior-posterior plane but not in the superior
plane. This finding agrees with the generally accepted principle
that horizontal stability is mediated by the acromioclavicular ligaments
while vertical stability is mediated by the coracoclavicular ligaments.
The increase in anterior and posterior translation after transection
of the capsule was accompanied by a significant increase in the
in situ forces in the coracoclavicular ligaments in response to
anteriorly and posteriorly directed loads. This increase in force
suggests that the coracoclavicular ligaments partially compensate for
the injured capsule in resisting these loading conditions. Nevertheless,
the relative vertical orientation of the coracoclavicular ligaments
prevents their effective restraint against anterior-posterior instability
in light of the significant increase in translation.
The magnitude of the in situ force in and percent contribution by
each ligament revealed that the conoid and trapezoid function differently
in resisting applied loads, depending on the direction of the applied
load. These distinctions became exaggerated after transection of
the capsule, when the in situ forces in the coracoclavicular ligaments
increased. With transection of the capsule, the conoid served as
the primary restraint against anterior and superior loading, while
the trapezoid functioned as the primary restraint against posterior
loading. The relative orientations of these two ligaments has been thought
to account for their different functions12.
These functional differences in providing stability to the acromioclavicular
joint have implications concerning the surgical reconstruction of
displaced acromioclavicular joint separations. Procedures that reconstruct
only the coracoclavicular ligaments with use of either suture or
synthetic graft material typically treat the coracoclavicular ligaments
as a single structure when reducing the superiorly displaced distal
part of the clavicle. The results of our study suggest that such
methods may not be sufficient to prevent anterior-posterior translation at
the acromioclavicular joint despite preventing superior translation.
Residual anterior-posterior instability after such operative procedures
may contribute to persistent postoperative pain and to inferior
outcomes, especially in patients who engage in overhead throwing
activities20. Furthermore, these
suture or graft materials as well as the intact coracoclavicular
ligaments may be subjected to higher forces and thus may be at risk
for early failure in the absence of supplemental fixation across
the acromioclavicular joint.
Previous investigators have suggested that the acromioclavicular
joint capsule is responsible for anterior-posterior stability or
that the trapezoid maintains posterior stability when the joint
is intact9,13. The differences
between the results in their studies and those in ours can be partly
attributed to the number of constraints (or degrees of freedom)
placed on the resulting joint motion and the magnitude of load applied
to the joint. These comparisons suggest that kinematic constraints
placed on the acromioclavicular joint during loading are important
and that the in situ force in each ligament is affected by the coupled
motions that occur during loading. Therefore, the force in the soft-tissue
structures is redistributed during application of an external load
when a greater number of degrees of freedom of motion are allowed,
as has been shown in studies of other joints21-23.
Other mechanisms, including individual muscles (the deltoid and
the trapezoid) and osseous contact, have been shown to contribute
to joint stability. The large variance found in the magnitude of
the force vector representing each ligament could also be caused
by the three distinct types of joint geometry described previously24. In addition, the resultant force
at the joint due to separation of the superior and inferior sections
of the capsule (the interaction between these portions of the capsule)
may contribute to joint stability12.