An irreparable radial head fracture can be treated nonoperatively, with
surgical resection, or with prosthetic replacement of the radial
head1-3.
Several studies have documented short and long-term pitfalls of radial head
resection, including a high prevalence (up to 100%) of radiographic evidence
of arthritic
changes4-10.
While it has been assumed by
some10,11
that such arthritic changes result from the initial trauma, another plausible
explanation is that they are due to altered ulnohumeral kinematics and
increased contact stresses on the cartilage. There has been increasing
interest in the option of implanting a radial head prosthesis, especially in a
medial collateral ligament-deficient elbow. A potential problem with radial
head replacement is the lack of proper instrumentation to make sure that the
correct length and orientation of the radial head are reproduced. To our
knowledge, no published studies have specifically addressed the issue of the
effects of overstuffing (lengthening) or understuffing (shortening) of the
radiocapitellar joint on elbow kinematics and load transfer across the
elbow.
We hypothesized that lengthening and shortening the radial neck would alter
the kinematics and the load transfer through the radiocapitellar joint in the
medial collateral ligament-deficient elbow. Knowledge of these effects is
clinically important as overstuffing increases the risk of pain and early
degenerative disease. In contrast, understuffing leaves the elbow at risk for
valgus instability. The aim of this study was to evaluate the effects of
radial neck lengthening and shortening on the kinematics and load transfer in
the medial collateral ligament-deficient elbow in an in vitro setting.
Specimen Preparation and Positioning
Six cadaveric fresh-frozen upper extremities with no macroscopic evidence
of pathological changes were used for the study. Three of the donors were
female and three were male, and the average age at the time of death was
seventy-eight years (range, sixty-nine to eighty-seven years). The range of
motion and stability of the cadaveric elbows were tested clinically and were
considered to be within normal limits. The specimens were thawed overnight at
room temperature. The humerus was transected in its midportion and the wrist
was disarticulated, preserving the ligaments of the distal radioulnar
joint.
A forearm rotation device was cemented distally into the medullary canals
of both the ulna and the
radius3. The forearm
was considered to be in neutral rotation when the ulnar and radial styloids
were aligned so that they were in the same plane as the humerus with the elbow
flexed to 90°. The humeral shaft was cemented into an acrylic tube and was
clamped into a testing device parallel to the floor, allowing the whole
construct to rotate so that the elbow flexed in either a horizontal or a
vertical
plane12.
Heavy nylon sutures were attached separately to the biceps, brachialis, and
triceps tendons to allow simulated muscle loading along their lines of action.
The triceps tendon was loaded with a static load of 20 N in the vertical plane
and 40 N in the horizontal plane. The biceps and brachialis sutures were tied
together at a distance of approximately 0.75 m from the tendon. A hook was
attached to the loop that was created, and the arm was moved to allow the
force to adjust until equilibrium had been achieved between both tendons. The
other end of the hook was then attached to a motor pulling at a constant
speed, which loaded both flexors simultaneously. The lowest possible speed at
which an arm could move from extension to flexion was used in each case. The
action of the biceps and brachialis was always controlled and resulted in
flexion without a rotatory component on the forearm because forearm rotation
was restrained by a pronation-supination
holder3.
Surgical Approach
A lateral skin incision was used to approach the elbow. The radial head and
neck were exposed through an osteotomy of the lateral humeral epicondyle after
two drill-holes had been made to ensure proper replacement at the end of the
procedure and to maintain the correct anatomical position of the ligament
without disturbing the integrity of the ligament itself. The annular ligament
and the anterior aspect of the capsule were transected to further expose the
radial head and neck. The orientation of the articulating surface of the
radial head was projected and marked as a circumferential line onto the radial
neck. A vertical alignment mark was also made. The radial neck was cut at the
level of the circumferential mark with use of an oscillating saw with
custom-made parallel saw blades to allow a telescoping prosthesis to be
inserted later. Therefore, the orientation of both the proximal and the distal
cut had the same orientation as that of the articulating surface of the radial
head. Next, a custom-made, telescoping prosthesis was implanted. It had a
modular design with telescoping capacities to allow controlled lengthening and
shortening of the radial neck. The markings on the resected radial head and
the telescoping prosthesis were aligned, and the head of the radius was
cemented on the proximal component of the prosthesis (Figs.
1-A and
1-B). A groove in the
prosthesis secured the rotational position of the proximal portion within the
distal stem.
A total of 5 mm of bone was resected at the neck. To reconstruct the
original radial neck length, two 2.5-mm-thick polyethylene spacers were
inserted into the telescoping prosthesis. The whole construct was secured with
a set-screw. In this way, the original length was restored and a
"telescoping prosthesis" was created with the native radial head.
Two cancellous bone screws were used to fix the lateral epicondyle back into
place. The anterior aspect of the capsule, the annular ligament, and the skin
were repaired with sutures. Kinematic measurements (described below) were
performed in this original (restored)-length position.
Next, the medial collateral ligament was transected at the base of the
epicondyle through a medial skin incision and a flexor-pronator
muscle-splitting approach. The muscles and skin were closed with a suture.
Kinematic measurements were again conducted: in the original-length condition,
in the two shortened radial neck conditions (2.5 and 5 mm less than the
original length), and in the two lengthened conditions (2.5 and 5 mm more than
the original length). Shortening and lengthening were performed by removing or
inserting the polyethylene spacers in the telescoping prosthesis to reach the
desired length.
On completion of the testing, the skin was removed from the specimen,
leaving muscles, capsule, and ligaments. A pressure-sensitive sensor (Sensor
6900, I-scan; Tekscan, Boston, Massachusetts), with a sensing area of 14
× 14 mm, was inserted through the anterior side into the radiohumeral
joint space. The sensors were always placed in the same manner and
orientation. Care was taken not to damage the sensor by wrinkling, folding, or
twisting it. Once the sensor was seated in the joint, it was secured by
closing the capsule anteriorly as much as possible.
Data Collection
The six-degrees-of-freedom motion of the ulna and humerus was measured with
an electromagnetic tracking device (3Space Fastrak; Polhemus, Colchester,
Vermont) sampling at a rate of 40 Hz. This system is accurate to within
0.5° for angular rotation. One sensor was attached to the ulna, another
sensor was attached to the humerus, and the source was attached to the table.
Data were recorded with Motion Monitor software (Innovative Sports Training,
Chicago, Illinois). Data obtained from the electromagnetic tracking device
were used to calculate the position of the ulna relative to the humerus with
the aid of an anatomical coordinate system with use of the flexion-extension
axis, the pronation-supination axis, and the varus-valgus axis with the origin
of the coordinate system at the center of the trochlea. The details of this
technique have been published
previously12,13.
Data were collected during simulated active motion of the elbow from
extension to flexion in the vertical plane and then in the horizontal plane
with the elbow subjected to gravity valgus and varus stresses
sequentially12
(Table I). Motions in the
horizontal plane defined the gravity valgus and varus stressed positions and
resulted in the medial epicondyle facing upward and downward, respectively.
All of the measurements were made with the forearm fixed in neutral rotation,
then repeated with the forearm in 80° of pronation, and then repeated with
it in 80° of supination (Table
I).
On completion of the kinematic data collection, the pressure across the
radiocapitellar joint was measured under static loading at 5°, 30°,
60°, and 90° of flexion, in the valgus stressed position, for all
three forearm rotations. Care was taken not to change the sensor orientation
in relation to the capitellum while the forearm was rotated. The triceps
tendon and the combined biceps and brachialis tendons were both loaded with 40
N.
Data Analysis
Total varus-valgus laxity was calculated by subtracting the degree of
displacement of the ulna in the varus stressed position from the degree of
displacement in the valgus stressed position at each corresponding flexion
angle. Similarly, total ulnar axial rotation was calculated by subtracting the
ulnar axial rotation in the varus stressed position from the ulnar axial
rotation in the valgus stressed position. The effect of altered radial neck
lengths on angular displacement was determined by subtracting the measured
value in the altered position from that in the original (restored)-length
position.
A three-factor analysis of variance was used to compare the effects of the
surgical conditions throughout the flexion arc in the three forearm positions
(Table I). The significance
level was set at p < 0.05. A post hoc Tukey test was done to assess
significance.
Radiocapitellar load transfer was measured as total compression across the
sensor in the original condition and after lengthening of the radial neck by
2.5 mm. Data were normalized by dividing the data obtained under the 2.5-mm
lengthened condition by the data obtained under the original-length condition.
Thus, sensor compression was expressed relative to the pressure found in the
original-length condition and shown as averaged values across the specimens. A
three-factor analysis of variance was used to compare the effects of
lengthening of the radial neck throughout the flexion arc in the three forearm
positions. The significance level was set at p < 0.05. A post hoc Tukey
test was done to assess significance. JMP software was used for all
statistical calculations (SAS Institute, Cary, North Carolina).
Kinematics
Although the values in pronation and supination differed significantly from
those in neutral (p < 0.05), similar trends were found in all three forearm
rotations.
In each of the forearm rotations, the valgus-varus position of the ulna
throughout the flexion arc was affected by the length of the radial neck
(Fig. 2). In the gravity valgus
stressed positions, shortening caused the ulna to track in more valgus,
whereas lengthening pushed the ulna into varus (p < 0.05). This effect was
greatest at 30° of flexion. Also, the total varus-valgus laxity increased
consistently with shortening of the radial neck and decreased with lengthening
of the radial neck when compared with the laxity in the original-length
condition (Fig. 2). All of the
shortened and lengthened conditions (p < 0.05) differed significantly from
the original-length condition with respect to total varus-valgus laxity
throughout all flexion angles.
Similarly, in the valgus stressed position, shortening of the radial neck
caused internal rotation of the ulna compared with the ulnar rotation in the
original-length condition (p < 0.05)
(Fig. 3), whereas shortening
resulted in little or no change in axial rotation of the ulna under varus
stress. Lengthening of the radial neck consistently caused external rotation
of the ulna under both valgus and varus stresses (p < 0.05). Also, all of
the altered-length conditions resulted in a significant difference in total
ulnar axial rotation compared with that in the original-length condition (p
< 0.05) (Fig. 3). Total
axial rotation increased with radial shortening and decreased with radial
lengthening. There was no significant difference between the measured total
ulnar axial rotation in the 5-mm lengthened condition and that in the 2.5-mm
lengthened condition (p > 0.05).
Joint Pressure
A pilot study revealed that 5 mm of lengthening caused such severe
overstuffing that the measurements were distorted. As a result of the pressure
exerted on the pressure-sensitive sensor, the sensor itself became
irreversibly deformed and additional pressure measurements were not possible.
Shortening by either 2.5 or 5 mm unloaded the radiocapitellar joint to a level
below the threshold for pressure detection by the film. This latter finding
was established with the first specimen used in the actual experiment.
Measurements performed in the original-length and 2.5-mm lengthened
conditions showed that the axial pressure transmission across the
radiocapitellar joint tended to decrease with the flexion angle; however, this
decrease was not significant (p > 0.05). In all normal specimens (with the
original radial length), at all flexion angles, the radiocapitellar pressure
was highest with the forearm pronated and lowest with it supinated (p <
0.05). With 2.5 mm of lengthening, the axial contact pressure was
significantly higher (p < 0.05) than that in the original-length condition
in all forearm positions (Fig.
4).
If a radial head implant is used without restoration of the proper axial
length of the radius, there is a potential for a number of complications, such
as residual
instability14,
altered load
transfer15, and
induction of cartilage lesions in the
capitellum15-18.
Our study showed that lengthening and shortening of as little as 2.5 mm,
compared with the original radial neck length, affected ulnohumeral kinematics
and radiocapitellar pressures. Shortening of the radial neck by =2.5 mm
caused the ulna to track in a valgus and internally rotated position. Valgus
stress subluxates the proximal part of the ulna. Loading of the ulnohumeral
joint during motion forced the medial portion of the articulating surface of
the proximal part of the ulna onto the medial part of the trochlea. This
effectively resulted in internal rotation of the ulna. The opposite happened
when the radial neck was lengthened: lengthening forced the ulna into a more
varus and externally rotated position even with the application of valgus
stress.
Our in vitro study did not exactly mimic the clinical setting. The arm was
moved from extension to flexion by loading only three muscles, whereas
activities of daily living usually involve complex motion of the arm and the
combined activity of a number of muscles that are constantly regulated.
Proprioception and pain also most likely have an effect on in vivo elbow
motion that cannot be replicated in vitro. Despite these limitations, we
believe that the results of our study convey an important clinical message.
Our goal was to evaluate the hypothesis that restoring the exact length of the
radius is crucial to maintaining normal kinematics and intra-articular
pressure in the elbow joint. In order to do so, we used a native radial head
attached to a telescoping prosthesis and left all other parameters unchanged
compared with those in the intact elbow. Using a metal radial head would have
changed material properties and would have introduced additional variables
with unknown or unpredictable effects. Those include altering the joint
surface contact area and pressure and potentially inaccurate positioning of
the implant. With a native radial head, these potential problems were not a
concern. Therefore, we were able to conclusively attribute the ulnohumeral
maltracking and altered radiocapitellar pressure to changes in the length of
the radial neck.
Maltracking of the ulnohumeral joint could possibly induce degenerative
disease as a result of abnormally high stresses on the cartilage. Experimental
studies have shown that >60% of the axial load transfer across the elbow
goes through the radiocapitellar
joint19-21.
The force normally borne by the radial head is transferred to the ulnohumeral
joint after radial head resection, as predicted by Amis et
al.19,22.
The increased force on the ulnohumeral joint could be further increased by any
malrotation and associated kinematic disturbances
(Fig. 4). This might explain
the observation that 52% to 100% of radial head resections have been followed
by the development of ulnohumeral
arthritis4-10.
However, it has also been hypothesized that such arthritic changes result from
the initial
trauma10,11.
Our experiments were performed after transection of the medial collateral
ligament. The main indication for replacing the radial head with a prosthesis
is an irreparable radial head fracture associated with valgus instability due
to a lesion of the medial collateral
ligament7,23,24.
If the elbow remains unstable after radial head replacement, repair or
reconstruction of the medial collateral ligament can be
indicated23-25.
It is generally assumed that an elbow that has sustained a valgus traumatic
injury causing disruption of the medial collateral ligament and a fracture of
the radial head has an overwhelming propensity for chronic valgus instability
if the radial head is
excised12. This is
thought to be due to inadequate healing of the medial collateral ligament. To
our knowledge, there is no evidence in the literature demonstrating that the
medial collateral ligament can heal adequately without correct restoration of
the radial column. The current study supports the idea that the kinematics of
an elbow with a deficient medial collateral ligament depend on the length of
the radial column. This suggests that the radius has to be restored to within
2.5 mm of its original length in order for the medial collateral ligament to
heal correctly.
Apart from anecdotal discussions of overstuffing of the radiocapitellar
joint, the clinical effects of lengthening and shortening are unknown. Our in
vitro experimental data suggest that reconstruction of the radial head and
neck should restore the axial anatomy as closely as possible, to within a
range of approximately ±2.5 mm. The malarticulation of the ulnohumeral
joint and altered kinematics and pressures across the elbow that can be
expected with overstuffing and understuffing may predispose the elbow to the
development of degenerative arthritis in the long term. Consequently, radial
head replacement must not be thought of as simply the insertion of a spacer.
Rather, it should be performed with the same level of concern for accuracy and
reproducibility of component position and orientation as is needed for any
other prosthesis. ?