Eight unpreserved cadaveric upper extremities with normal elbow joints from
individuals who had been a mean of seventy-four years old (range, sixty-five
to seventy-eight years old) at the time of death were amputated through the
middle part of the humerus and were stored at -20°C prior to use. The
specimens were thawed for eighteen hours at room temperature (22°C
± 2°C) and were prepared for mounting on a custom elbow-testing
apparatus11,12.
Stainless-steel cables with a diameter of 0.8 mm were attached to the distal
intact tendons of three elbow flexors (the biceps, brachialis, and
brachioradialis), the primary extensor (the triceps), and a forearm pronator
(the pronator teres). The muscle bellies of the biceps, brachialis, and
triceps were transected 10 cm proximal to the elbow, and the remaining portion
of the humeral shaft was cleared of soft tissues. Skin incisions were closed,
and the specimen was kept moist with use of 0.9% normal saline solution
throughout testing.
The humerus was mounted in the testing apparatus in a neutral position with
use of a clamp that rigidly held the arm while allowing unconstrained elbow
motion13,14
(Fig. 1). The tendon cables
were attached to computer-controlled pneumatic actuators, which allowed for
simulated active joint motion when load was applied. The lines of action for
the biceps, brachialis, and triceps cables were controlled by running the
cables through a unit that aligned them and controlled their direction of
pull. The cables for the pronator teres and brachioradialis were routed
through the humeral canal by means of Delrin sleeves that were inserted into
the medial and lateral supracondylar ridges, respectively. A hinge on the
base-plate of the testing device allowed placement of the arm in the dependent
position for active elbow simulation and in varus and valgus gravity-loaded
positions (Fig. 1).
The tester (D.M.B.) carried out passive motion in the dependent position by
grasping the wrist and first fully pronating or supinating the forearm until
resistance was felt. The arm was then brought from full extension to full
flexion. Active elbow motion in the dependent position was initiated from the
fully extended position. The pneumatic actuators that were attached to the
cables applied prescribed forces to the pronator teres or biceps to fully
pronate or supinate the forearm, respectively. Prescribed forces were then
applied to the flexor tendons (i.e., the biceps, brachialis, and
brachioradialis) to initiate the flexion sequence, thus producing motion
through an arc from full extension to full flexion. The muscle-loading
protocol was based on electromyographic data and a previously validated in
vitro testing system that was developed in our
laboratory11,12.
We tested varus-valgus laxity by adjusting the base-plate on the testing
system to place the elbow in the varus or valgus gravity-loaded position. For
each motion, the forearm was maintained in pronation or supination with the
tester grasping the wrist without supporting the weight of the arm. The elbow
was then brought through an arc of motion from full extension to full
flexion.
Medial and lateral epicondylar osteotomies were performed to simulate
medial and lateral collateral ligament deficiency, respectively
(Fig. 2). Care was taken to
leave the ligaments and common flexor-extensor origins attached to the
osteotomy fragments. Some dissection of the anterior and posterior elbow
capsule was necessary to improve exposure. The osteotomy sites were securely
repaired with use of 3.5-mm cortical screws (Synthes Canada, Mississauga,
Ontario, Canada) to reattach the collateral ligaments and muscle origins. The
screw-holes were reinforced with bone cement to ensure that rigid repair was
obtained repeatedly throughout testing. The testing protocol was performed
with the elbow intact and for each of four ligamentous states: (1) the stable
elbow (with both osteotomy sites repaired), (2) the medial collateral
ligament-deficient elbow, (3) the lateral collateral ligament-deficient elbow,
and (4) the medial and lateral collateral ligament-deficient elbow.
The radial head was approached through the lateral osteotomy site, and no
additional soft-tissue dissection was needed for exposure. For each of the
aforementioned ligamentous states, testing was carried out first with the
native radial head intact, then with the radial head resected, and finally
after the performance of a radial head arthroplasty with use of a metallic
implant. The radial head was excised at the junction of the head and neck,
with the radial neck left intact. The excised radial head was used as a
template to determine the appropriate size of implant. The implant system that
was used (EVOLVE modular radial head implant system; Wright Medical
Technology, Arlington, Tennessee) allowed us to directly template the diameter
and thickness of the radial head implant from the excised radial head. Stem
diameter was determined by hand reaming the radial neck with the supplied
rasps.
One receiver of an electromagnetic tracking device (Flock of Birds;
Ascension Technology, Burlington, Vermont) was attached to the ulna, and a
transmitter was securely mounted on the base-plate of the testing apparatus
(Fig. 1). A second receiver,
attached to a stylus, was used to digitize landmarks on the humerus, radius,
and ulna in order to construct independent bone-coordinate systems to describe
the motion
pathways8. This
testing system allowed for the measurement of motion of the ulna relative to
the humerus in six degrees of
freedom13. Elbow
kinematics, expressed in degrees of valgus angulation and external rotation of
the ulna relative to the humerus, were determined during passive and active
motion with the specimen in the dependent position with the forearm maintained
in both pronation and supination. Data were analyzed at 30°, 60°,
90°, and 120° of elbow flexion. Varus-valgus laxity was determined by
calculating the difference in angulation of the ulna relative to the humerus
when the elbow was in the varus gravity-loaded position as compared with the
valgus gravity-loaded position. Maximum varus-valgus laxity was defined as the
greatest difference in varus-valgus laxity measured throughout the arc of
flexion.
Statistical Methods
Data were analyzed with use of one-way and two-way repeated-measures
analyses of variance and Student-Newman-Keuls multiple-comparison procedures,
with the level of significance set at p < 0.05.
Model Reliability
Fully intact elbows were compared with elbows in which the native radial
head was intact and both osteotomy sites had been repaired in order to verify
the reliability of our model in restoring the ligamentous integrity of the
elbow. There was no significant difference in dependent-position kinematics
between the intact specimens and the specimens in which the osteotomy sites
had been repaired (p > 0.3). There was a significant difference in maximum
varus-valgus laxity between the intact and repaired specimens (p = 0.009);
however, these differences were small with the forearm in both pronation
(1.87°) and supination (1.28°). Therefore, all statistical comparisons
were made with use of the repaired specimens.
Elbow Kinematics-Valgus Angulation
Two-way repeated-measures analyses of variance indicated that radial head
excision and arthroplasty had an overall effect on valgus angulation of the
elbow during simulated active motion (p < 0.001), but this effect was
dependent on the angle of flexion; that is, there was a significant
interaction (p < 0.001) (Fig.
3). In general, the surgical interventions had less of an effect
on kinematics at higher angles of flexion. More specifically, in pronation
(Fig. 3, A), the
magnitude of valgus angulation after radial head excision was significantly
larger than that with the elbow intact or after radial head arthroplasty at
30°, 60°, and 90° of flexion (p < 0.004) but not at 120° of
flexion (p = 0.6). In supination (Fig. 3,
B), there also was no difference between the three radial
head conditions (i.e., with the radial head intact, after radial head
excision, and after radial head arthroplasty) at 120° of flexion (p =
0.8). For illustration purposes, the behavior at 60° is considered
throughout the remainder of this section.
Passive Motion
No significant change in valgus angulation after radial head excision or
radial head arthroplasty was noted in association with any of the ligamentous
conditions with the forearm maintained in pronation or supination (p >
0.05) (Fig. 4).
Active Motion, Ligaments Intact
Valgus angulation increased after radial head excision when the forearm was
maintained in pronation (p = 0.001) or supination (p = 0.006)
(Fig. 5). Valgus angulation
after radial head arthroplasty was similar to that seen with the native radial
head intact when the forearm was maintained in pronation (p = 0.9) or
supination (p = 0.4) (Fig.
5).
Active Motion, Lateral Collateral Ligament Insufficiency
Valgus angulation increased after radial head excision with the forearm
maintained in pronation (p = 0.004) and was similar to that seen with the
native radial head after radial head arthroplasty (p = 0.1). A similar trend
was seen after radial head excision with the forearm maintained in supination,
but this increase was not significant (p = 0.2)
(Fig. 5).
Active Motion, Medial Collateral Ligament Insufficiency
Valgus angulation increased after radial head excision when the forearm was
maintained in pronation (p < 0.001) or supination (p = 0.001). Valgus
angulation after radial head arthroplasty was similar to that seen with the
native radial head when the forearm was maintained in pronation (p = 0.9) or
supination (p = 0.1) (Fig.
5).
Active Motion, Lateral and Medial Collateral Ligament
Insufficiency
Valgus angulation increased after radial head excision when the forearm was
maintained in pronation (p < 0.001) or supination (p = 0.005). Valgus
angulation after radial head arthroplasty was similar to that seen with the
native radial head intact when the forearm was maintained in pronation (p =
0.2) or supination (p = 0.1) (Fig.
5).
Elbow Kinematics-External Rotation
Overall, elbow kinematics were less affected by radial head excision at
higher flexion angles. Again, for illustration purposes, the behavior at
60° is considered throughout this section.
Passive Motion, Ligaments Intact
There was no significant change in rotation after radial head excision when
the forearm was maintained in pronation (p > 0.05). External rotation
increased after radial head excision when the forearm was maintained in
supination (p = 0.004). External rotation after radial head arthroplasty was
similar to that seen with the native radial head (p = 0.4)
(Fig. 6).
Active Motion, Ligaments Intact
There was a small increase in relative internal rotation (i.e., a decrease
in external rotation) after radial head excision with the forearm maintained
in pronation (p = 0.002) that was improved but was not completely corrected
after radial head arthroplasty (p = 0.03). There was no significant change in
rotation after radial head excision with the forearm maintained in supination
(p = 0.1) (Fig. 7).
Passive Motion, Lateral Collateral Ligament Insufficiency
There was no significant change in rotation after radial head excision with
the forearm maintained in pronation (p = 0.5). External rotation increased
after radial head excision with the forearm maintained in supination (p =
0.002), and it was not restored to that seen with the native radial head after
radial head arthroplasty (p = 0.01) (Fig.
6).
Active Motion, Lateral Collateral Ligament Insufficiency
There was a small increase in relative internal rotation after excision of
the radial head with the forearm maintained in pronation (p = 0.009) that was
corrected after radial head arthroplasty (p = 0.5). There was no significant
change in rotation after excision of the radial head with the forearm
maintained in supination (p = 0.2) (Fig.
7).
Passive Motion, Medial Collateral Ligament Insufficiency
There was no significant change in rotation after radial head excision with
the forearm maintained in pronation (p = 0.4) or supination (p = 0.4). When
pronation was compared with supination, there was significant internal
rotation of the ulna when the forearm was maintained in pronation (p = 0.008)
(Fig. 6).
Active Motion, Medial Collateral Ligament Insufficiency
There was a small increase in relative internal rotation after radial head
excision with the forearm maintained in pronation (p = 0.001) or supination (p
= 0.03) that was corrected after radial head arthroplasty with the forearm in
pronation (p = 0.3) and supination (p = 0.6)
(Fig. 7).
Passive Motion, Lateral and Medial Collateral Ligament
Insufficiency
There was no significant change in rotation after radial head excision with
the forearm maintained in pronation at any of the flexion angles (p = 0.6).
There was a small increase in relative internal rotation after radial head
excision with the forearm maintained in supination (p = 0.008) that was not
fully corrected after radial head arthroplasty (p < 0.001)
(Fig. 6).
Active Motion, Lateral and Medial Collateral Ligament
Insufficiency
There was a small increase in relative internal rotation after radial head
excision with the forearm maintained in pronation (p < 0.001) that was
corrected after radial head arthroplasty (p = 0.7). There was no significant
change in rotation with the forearm maintained in supination (p = 0.2)
(Fig. 7).
Varus-Valgus Laxity
During testing, the elbows were observed to be unstable in the valgus
position when the medial collateral ligament was disrupted, unstable in the
varus position when the lateral collateral ligament was disrupted, and
unstable in both the valgus and varus positions when both ligaments were
disrupted. This instability was quantified as an increase in varus-valgus
laxity when one or both ligaments were disrupted as compared with that seen
when both ligaments were intact for all radial head conditions (p < 0.001)
(Fig. 8).
Forearm Pronation
There was an increase in maximum varus-valgus laxity after radial head
excision with both the lateral and medial collateral ligaments intact (p <
0.001) and with the lateral collateral ligament disrupted (p = 0.01)
(Fig. 8). Stability was
improved but was not restored after radial head arthroplasty with both the
lateral and medial collateral ligaments intact (p = 0.02) and with the lateral
collateral ligament disrupted (p = 0.03). There was an increase in
varus-valgus laxity after radial head excision with the medial collateral
ligament disrupted, but this increase was not significant (p = 0.3). Stability
was similar to that seen with the native radial head after radial head
arthroplasty with the medial collateral ligament disrupted (p = 0.3). There
was a decrease in varus-valgus laxity after radial head excision with both
ligaments disrupted (p = 0.01), and stability was similar to that seen with
the native radial head after radial head arthroplasty (p = 0.1).
Forearm Supination
There was an increase in varus-valgus laxity after radial head excision for
all ligament conditions (p < 0.001)
(Fig. 8). Stability was similar
to that seen with the native radial head after radial head arthroplasty for
all ligament conditions (p = 0.07).
The present study demonstrated that radial head excision alters the
kinematics and varus-valgus laxity of the elbow with intact ligaments and that
stability is improved after radial head arthroplasty. An increase in external
rotation of the ulna with respect to the humerus during passive motion with
the forearm in supination was measured after radial head excision when both
ligaments were intact. Dunning et
al.15 demonstrated
a similar increase in external rotation in elbows with an intact radial head
but a disrupted lateral collateral ligament complex. Clinically, O'Driscoll et
al.16 described
posterolateral rotatory instability of the elbow following radial head
excision. We therefore postulate that the lateral ligament complex is
effectively under reduced tensile stress after radial head excision. Similar
to the elbows in the report by Dunning et
al.15, the elbows
in the present study tended to be more stable at higher flexion angles.
Morrey et al.6
demonstrated only a small change in valgus elbow stability in association with
radial head excision alone. We also noted a small but significant alteration
in stability after radial head excision in the presence of intact ligaments.
Jensen et al.17
demonstrated an alteration of elbow kinematics after radial head excision in
the presence of intact ligaments, but they did not account for muscular
stabilizers. Although the elbows in this study generally were more stable
during active as compared with passive motion, there was still increased
valgus angulation following radial head excision in the presence of intact
ligaments. These findings were most likely due to the valgus moment applied by
the simulated elbow flexors.
In the present study, medial and lateral epicondylar osteotomies were used
to model ligamentous insufficiency. Although dependent-position kinematics
were the same in the intact and osteotomized specimens, there was a small
(<2°) increase in varus-valgus laxity in the osteotomized specimens.
This finding was not surprising as some soft-tissue dissection was required to
carry out the osteotomies. We believed that we could not repeatedly carry out
accurate soft-tissue repair in addition to fixation of the osteotomy sites.
Therefore, we accepted this small amount of instability and used the specimens
with repaired osteotomy sites as our "intact" controls. In our
model, the flexor pronator mass was detached with the medial collateral
ligament and the common extensor muscles were detached with the lateral
collateral ligament. Therefore, the instability that was induced in the
present study is more severe than the instability that is typically seen in
elbows with more minor injuries. This model, therefore, can be considered to
represent a worst-case scenario, as is often seen in elbows with higher-energy
injuries, such as fracture-dislocations, in which disruption of muscle origins
is commonly noted in association with ligamentous
injuries18.
We are not aware of any previous studies on the effect of radial head
excision and arthroplasty in elbows with disruption of the lateral collateral
ligament. In the current study, a significant decrease in elbow stability was
noted in association with radial head excision in elbows with disruption of
the lateral collateral ligament. Elbow laxity was improved following
arthroplasty; however, these elbows were still unstable relative to those with
intact ligaments. These findings suggest that repair of the disrupted lateral
collateral ligament complex is essential in order to restore elbow stability
following radial head arthroplasty. Similar to the elbows in the study by
Dunning et al.15,
the elbows in the present study were more stable with the forearm in
pronation. Therefore, the findings of the present study suggest that patients
with these injuries should be rehabilitated with the forearm in the pronated
position.
Pomianowski et
al.9 reported that
laxity was increased after radial head excision in elbows with disruption of
the medial collateral ligament. Other authors have shown that radial head
arthroplasty restores valgus stability in elbows with disruption of the medial
collateral ligament to a state similar to that seen in elbows with a native
radial
head8,9.
In our model of medial collateral ligament insufficiency, the kinematics and
stability of the elbow improved following radial head arthroplasty as compared
with those observed following radial head excision, but they did not return to
normal. Armstrong et
al.19 observed
improved stability during active as compared with passive motion following
disruption of the medial collateral ligament in elbows with a native radial
head. They postulated that the improvement in stability that was afforded by
muscle activation would be reduced following excision of the radial head in
the presence of medial collateral ligament insufficiency. We observed
decreased elbow stability during simulated active motion following radial head
excision and medial collateral ligament disruption. However, the amount of
instability was very small, possibly because of the stabilizing effect of the
biceps and brachialis.
The current study also modeled the most severe injury scenario of radial
head fracture combined with both medial and lateral collateral ligament
disruption. To our knowledge, this combination of injuries has not been
previously reported in a biomechanical study. Ligament disruption had a
greater effect on elbow instability than did radial head excision itself. The
elbows were still unstable following radial head arthroplasty, emphasizing the
importance of ligament repair in elbows with combined medial and lateral
soft-tissue disruption.
Radial head fractures often involve the entire radial head but rarely may
extend distally to involve a substantial portion of the radial neck. With use
of the modular implant that was employed in this study, it was possible to
make up for a large radial neck deficiency. Although gross valgus instability
was evident in association with combined radial head excision and medial
collateral ligament deficiency with the arm in the valgus gravity-loaded
position, this instability might have been even more pronounced with a more
distal excision of the radial neck as contact of the radial neck with the
capitellum would have been reduced. Radial head arthroplasty in the setting of
a substantial concomitant radial neck fracture is challenging and may require
fixation of the neck before radial head implantation. Our conclusions
regarding radial head arthroplasty only apply if the length of the proximal
part of the radius is accurately restored with an appropriately sized
implant.
In conclusion, this cadaveric study demonstrated that elbow kinematics and
stability are altered following radial head excision and improved following
radial head arthroplasty in elbows with intact ligaments. Radial head excision
decreases the stability of elbows with disruption of the medial collateral
ligament, those with disruption of the lateral collateral ligament, and those
with disruption of both the medial and the lateral collateral ligament. Radial
head arthroplasty in elbows with these ligamentous injuries improves stability
(i.e., to a level similar to that seen in elbows with a native radial head)
but does not return it to normal. Radial head arthroplasty alone may be
insufficient for the treatment of complex elbow instability, and concomitant
repair of ligaments and muscular origins should be considered when possible.