Linked semiconstrained total elbow prostheses have been used successfully
to treat a wide variety of conditions but may be at higher risk for implant
loosening than unlinked implants
are1-3.
Over the past two decades, reports on total elbow arthroplasty with use of the
Coonrad-Morrey linked prosthesis have demonstrated relatively low rates of
aseptic loosening, particularly in patients with rheumatoid
arthritis4-7.
Gill and Morrey4, in
a study of forty-six patients who were managed with the Coonrad-Morrey
prosthesis for the treatment of rheumatoid arthritis, reported a ten-year
survival rate, free of revision, of 92%. Clinical experience has suggested
that aseptic loosening still does occur in association with this device but is
more likely to involve the ulnar component than the humeral
component4,5.
Understanding the potential causes of mechanical failure is important to
prevent loosening. We recently have become aware of a mechanism of failure
that is probably preventable and therefore important to recognize.
The present study includes ten patients in whom the mechanism responsible
for ulnar component failure was pistoning caused by a distraction force at the
limit of elbow flexion. The purpose of the present report is to describe this
mechanism of failure as well as its clinical diagnosis, treatment, and
potential for prevention.
Between January 1998 and December 2004, ten patients underwent revision of
the Coonrad-Morrey total elbow prosthesis at the Mayo Clinic because of
painful mechanical loosening of the ulnar component caused by pistoning within
the ulna. All revisions were performed by the senior author (S.W.O'D.). The
patients were assessed both clinically and radiographically.
The study group included six women and four men with an average age of
fifty-seven years (range, thirty-two to seventy-six years). Eight of the ten
elbows were in the non-dominant extremity. Seven of the ten patients had
preexisting rheumatoid arthritis, and the other three had posttraumatic
arthritis. Five of the seven patients with rheumatoid arthritis had been
taking disease-modifying antirheumatic medication or prednisone prior to the
revision arthroplasty. The mean interval from the primary arthroplasty to the
revision arthroplasty was fifty months (range, seventeen to ninety-six
months). One patient had had two previous revisions at the site of the total
elbow arthroplasty.
All patients had previously undergone a primary Coonrad-Morrey linked total
elbow arthroplasty through a posterior incision and the Mayo
triceps-reflecting
approach8. In all
cases, the proximal aspect of the ulnar component had a polymethylmethacrylate
precoat. The polymethylmethacrylate precoat became a part of the standard
design of the ulnar component in 1994, was discontinued in 2001 because of
reports of debonding of the polymethylmethacrylate precoat from the stem, and
was changed to a titanium plasma-sprayed
surface9.
History
At the time of presentation, five of the ten patients complained of audible
squeaking and a sensation of the ulnar component moving within the ulna, three
had the sensation of the ulnar component moving within the ulna but without
audible squeaking, and one complained of squeaking without a sensation of
ulnar component movement. All nine of these patients had pain; the pain was
subjectively rated as mild in one patient, moderate in three, and severe in
five. The remaining patient had severe pain, without audible squeaking or the
sensation of ulnar component movement.
Physical Examination
On physical examination, the pistoning could be recreated by axially
loading the ulna with manual compression and distraction forces in three
patients and could be recreated by manual flexion and extension in an
additional three patients. In the latter three patients, clunking at the
extremes of flexion and extension could be recreated, with the distinct
sensation on the part of the examiner of a shift of the component within the
bone. Audible squeaking could be reproduced during the examination of two of
the six patients who had reported a history of squeaking. An effusion was
evident in two patients. Tattooing of the skin from metal debris produced from
the implant was present in one patient. Tenderness to palpation of the ulna
was present in four patients. Triceps deficiency was present in one patient.
The average preoperative range of motion was from 30° to 135° (flexion
contractures ranged from 0° to 70°, and flexion ranged from 100°
to 150°). The average pronation was 75° (range, 60° to 90°),
and the average supination was 68° (range, 15° to 90°).
Radiographs
A complete radiolucent line between the ulnar prosthesis and the cement
mantle was noted in three of the ten patients. In one additional patient,
there was a complete radiolucent line between the cement mantle and bone
(Fig. 1-A). In the remaining
six patients, the radiolucent line between the implant and the cement mantle
was incomplete. A radiolucent gap between the cement and the tip of the ulnar
prosthesis was noted in six patients. Two patients had evidence of proximal
migration of the ulnar component within the cement mantle when the position of
the component on the lateral flexion radiograph was compared with that on the
lateral extension radiograph. Fracture of the cement mantle was seen in two of
the ten patients. Three of the ten patients had evidence of anterior
impingement, defined as abutment of the coronoid process or cement mantle
against the anterior flange of the humeral implant, on the lateral flexion
radiograph.
Operative Findings
All ten patients underwent a revision arthroplasty, during which
intraoperative culture specimens were obtained. One patient had positive
growth of Corynebacterium species on culture at five days. This was thought to
be a contaminant, and treatment showed no subsequent evidence of
infection.
The ulnar component was found to be loose in all ten patients. In nine
patients, the prosthesis was loose in the cement. In the remaining patient,
the cement-bone interface was grossly loose and the ulnar component with its
surrounding cement mantle was easily extracted from the ulnar canal
(Fig. 1-B). One finding that we
observed repeatedly was that the ulnar component, while grossly loose, could
not be extracted from the ulna without being tapped out. In other words, it
could piston in and out over a distance of 1 to 2 cm, but resistance would be
encountered during extraction when we attempted to pass the curved stem
through the cement mantle. Burnishing of the stem of the ulnar component as a
result of the pistoning effect between the implant and the cement was seen in
the cases of four patients. Macroscopic evidence of worn bushings was seen in
the cases of three patients. Metal synovitis with subcutaneous fluid
collection was present in one patient
(Figs. 2-A and 2-B).
Intraoperatively, after the ulnar component had been removed, the distal
cement mantle was removed with use of an ultrasonic cement removal instrument
or by drilling through it with a Steinmann pin and then the canal was reamed
with a series of flexible reamers in order to create a space for a new ulnar
prosthesis. Range of motion was checked intraoperatively with the trial
component in place to ensure that adequate motion was possible without
impingement. A new prosthesis was then cemented in place. Impaction grafting
of the ulna was performed if the cortex of the ulna was thinned but intact.
Strut grafting of the ulna with use of cerclage wires was performed if there
was cortical perforation or a periprosthetic fracture.
At the time of most recent follow-up, nine of the ten ulnar components were
intact and stable. Three of the ten patients subsequently required another
surgical procedure on the elbow during the follow-up period; only one of these
three patients required re-revision of the ulnar component. The first patient
was a thirty-two-year-old man with posttraumatic arthritis who required a
revision total elbow arthroplasty with use of a custom implant at five months
postoperatively. The second patient was a sixty-four-year-old woman with
rheumatoid arthritis who required revision of the humeral component because of
loosening at twenty-eight months postoperatively. The third patient was a
forty-four-year-old woman with juvenile rheumatoid arthritis who required
repair of an avulsed triceps mechanism at fourteen months postoperatively.
This report describes an important and potentially preventable cause of
loosening of the ulnar component of a total elbow prosthesis. Distraction
forces on the ulnar component occur when the elbow is flexed past a limit set
by any impinging structure anteriorly. This impinging structure can be the
coronoid hitting the anterior humeral flange, protruding cement, heterotopic
bone, a malaligned component, or just bulky soft tissue and scar. In our
surgical experience, bulky soft tissue is more likely in patients with a flail
elbow due to loss of the distal part of the humerus or in patients with a
severe deformity that is corrected by shortening of the limb. Essentially, the
risk of creating an anterior fulcrum and a distraction force on the prosthesis
is present whenever there is an object (e.g., a flange, bone, or cement)
anterior to the hinge or when the humeral or ulnar components are inserted
further into the bone than would be required for anatomic recreation of the
axis of rotation.
This problem can be prevented at the time of implantation by trimming away
excess anterior osteophytes or a prominent coronoid or cement after checking
for anterior impingement during a trial of range of motion. The mere presence
of the anterior flange in fact makes coronoid impingement a likelihood with
the Coonrad-Morrey design of elbow prosthesis. Therefore, more recently, it
has been our routine practice to trim the tip of the coronoid down to the
level of the brachialis insertion as a prophylactic measure to prevent such
anterior impingement. Doing so also helps to achieve greater degrees of
flexion.
Postoperatively, all patients should be counseled concerning the avoidance
of forced passive hyperflexion of the elbow past 135°. This can happen as
the patient leans forward over a table onto the operatively treated extremity
while reaching for something with the contralateral extremity
(Fig. 3). The weight of the
upper body forces the elbow into hyperflexion, creating a distraction force on
the prosthetic components.
A history of squeaking, clunking, or grinding, along with a clinical
examination suggesting pistoning of the ulnar component, should raise
suspicion of this diagnosis. Lateral radiographs made with the elbow in full
flexion and extension may confirm pistoning. Documented pistoning of the ulnar
component should be treated with revision arthroplasty and elimination of
anterior impingement. We believe that failure to do so can lead to particulate
osteolysis caused by abrasion during pistoning.
The ulnar components in the present series were all precoated with
polymethylmethacrylate and were without a porous surface. We believe they had
a much lower threshold for pullout from the cement mantle than previous
components, which were porous-coated. More specifically, they are likely to
become loose as a result of breakage of the bond between the cement and the
prosthesis. Micromotion can then cause wear debris from the textured titanium
surface and, eventually, the characteristic patterns of osteolysis seen in
some of our patients. We have not recognized this mode of failure in
association with porous-coated Coonrad-Morrey components, and, to our
knowledge, this mechanism of failure (that is, component pullout) has not been
described previously in the literature on the Coonrad-Morrey total elbow
prosthesis.
Unfortunately, we are not able to determine the prevalence of this
complication on the basis of these data. However, during the same time-period,
the senior author had performed forty-five revisions of Coonrad-Morrey total
elbow prostheses. Therefore, 22% of the revisions in that period were
performed because of ulnar loosening due to this mechanism. We believe that
the key design feature that would prevent this problem is having a surface
texture offering a high resistance to pullout forces.
In summary, the increasing success of total elbow arthroplasty over the
past two decades has been due to understanding the reasons behind
complications leading to alterations of technique and implant design.
Distraction forces on the ulnar component can be experienced as a result of
anterior impingement due to the creation of a lever. This mechanism of
failure, which causes pistoning of the ulnar component, should be minimized at
the time of implantation of the prosthesis and must be considered when a
patient presents with pain at the site of a total elbow arthroplasty.
?