Abstract
Background: Fractures of total elbow arthroplasty components are
uncommon, and the literature provides little guidance regarding the management
and outcomes of treatment of these complications. The goal of this report was
to investigate the prevalence and management of fractures of ulnar and humeral
components following total elbow arthroplasty and to review our experience
with cement-within-cement reconstruction for revision following such
fractures.
Methods: Between 1979 and 2003, twenty-four patients with a total of
twenty-seven fractured total elbow arthroplasty components (seventeen ulnar
and ten humeral) of different designs presented to our institution. Twenty-six
implants underwent subsequent revision elbow arthroplasty at our institution.
Fourteen of those revisions were done with a cement-within-cement technique,
and twelve, with traditional methods. Twenty-one patients (twenty-three
implants) were available for final follow-up, and data that had been acquired
prospectively and entered into the institutional arthroplasty database were
reviewed retrospectively. At the time of final follow-up, the Mayo Elbow
Performance Score (MEPS) was calculated and preoperative, postoperative, and
most recent radiographs were examined for bone loss, bushing wear, and
integrity of the bone-cement interface.
Results: The prevalences of humeral and ulnar component fracture
following primary total elbow arthroplasties performed at our institution were
0.65% and 1.2%, respectively. At a mean of 5.1 years following revisions for
those fractures, the MEPS was excellent for eight patients, good for five,
fair for six, and poor for two. The average MEPS was 82 points following the
revision total elbow arthroplasties done with the cement-within-cement
technique and 78 points following the revisions done with the traditional
method of cement removal and insertion of a revision component. Complications
included seven intraoperative cortical perforations; five nerve injuries, two
of which were permanent; three triceps avulsions; and one deep infection.
Conclusions: Implant fractures following total elbow arthroplasty
are uncommon. They occur for several reasons, such as notch sensitivity,
component design, and high stresses due to bone deficiency. Revision
techniques, such as cement-within-cement reimplantation, are reliable for
relieving pain and restoring function; however, the rate and spectrum of
complications are a cause for concern.
Level of Evidence: Therapeutic Level IV. See Instructions
to Authors for a complete description of levels of evidence.
The most common modes of failure necessitating revision total elbow
arthroplasty are loosening, infection, periprosthetic fracture, and
instability. Component fractures are uncommon following total elbow
arthroplasty, and the literature provides little information regarding the
prevalence, revision techniques, and outcomes of treatment. We are aware of
only a single case report of an ulnar component
fracture1. The
purpose of this study was to investigate the prevalence, etiology, and
management of ulnar and humeral component fractures after total elbow
arthroplasty. We also compare two revision techniques and describe our
preferred surgical approach. The results and complications of these techniques
are reported in a consecutive series of patients treated at our
institution.
Patients
Between December 1979 and December 2003, twenty-four patients with a total
of twenty-seven fractured total elbow arthroplasty components (seventeen ulnar
and ten humeral) presented at our institution. The fractured humeral implants
included one London prosthesis (Dow Corning Wright, Arlington, Tennessee) and
nine Coonrad-Morrey components (Zimmer, Warsaw, Indiana) (two 4-in [10.2-cm]
small, three 6-in [15.2-cm] small, two 6-in regular, and two 8-in [20.3-cm]
small). The fractured ulnar implants included one Ewald prosthesis (Johnson
and Johnson, Raynham, Massachusetts) and sixteen Coonrad-Morrey implants
(twelve small, two extra-small/extra-long, and two regular). Twenty-two
patients had had the index prosthesis implanted at our institution, and two
patients were referred to our institution for treatment of the implant
fracture. Twenty-six of the twenty-seven components were revised at our
institution by four surgeons with expertise in elbow arthroplasty. One patient
elected to undergo revision surgery at another institution closer to home. A
Coonrad-Morrey semiconstrained total elbow arthroplasty was used for all
revisions but one, in which a Coonrad implant (Zimmer) was inserted.
The study cohort comprised sixteen women and eight men, with an average age
of sixty-four years (range, thirty-two to eighty-six years) at the time of the
revision surgery. Nineteen fractured implants were in right extremities, and
fifteen were in dominant extremities. The underlying diagnosis leading to the
index elbow arthroplasty was rheumatoid arthritis in six cases, pathologic
fracture in one, and sequelae of trauma in twenty; the sequelae of trauma
consisted of posttraumatic arthritis in nine cases, distal humeral fracture in
six, and distal humeral nonunion in five. The mean age of the patients at the
time of the primary elbow arthroplasty was fifty-five years (range,
twenty-three to seventy-eight years). An average of 4.0 prior procedures
(range, one to twelve prior procedures) and 1.6 prior arthroplasties (range,
one to six prior arthroplasties) had been performed on the affected elbows
prior to the revision because of the component fracture.
The predominant presenting symptom of all twenty-seven component fractures
was pain. Eleven fractures were in patients who described the spontaneous
onset of pain with no history of trauma and who stated that they had complied
with the restrictions recommended following the elbow
arthroplasty2. Six
fractures were in patients who also described the spontaneous onset of pain
without trauma but stated that they had been regularly noncompliant with the
elbow arthroplasty restrictions and had routinely lifted >50 to 100 lb
(>23 to 45 kg). Five patients described a fall associated with the onset of
elbow pain, and five described hearing a "snap," followed by elbow
pain, while they were engaged in heavy lifting.
The patients presented with considerable prerevision pathological
conditions. Four patients had preexisting ulnar nerve symptoms, one of which
was related to a traumatic ulnar nerve laceration at the time of the index
open distal humeral fracture. One patient had preexisting triceps weakness due
to a high radial nerve injury sustained in childhood. Four patients had
insufficiency of the extensor mechanism due to multiple prior operations on
the elbow (mean, eight prior elbow operations; range, five to twelve). One
patient had a history of deep infection after revision total elbow
arthroplasty, which was successfully treated with a two-stage revision. One
patient had a fibrous union at the site of an olecranon osteotomy that had
been used to expose a distal humeral fracture.
The average time between the insertion of the implant and the revision
surgery due to the implant fracture was 8.2 years (range, 2.4 to 13.8 years)
for the humeral components and 4.6 years (range, 1.1 to eleven years) for the
ulnar components. Three patients sustained fractures of both the ulnar and the
humeral component. All three patients initially presented with a fracture of
the ulnar component, which was revised and followed by a fracture of the
humeral component at a mean of 4.7 years (3.4, five, and 5.8 years) later. All
three patients were in high-demand occupations (waitress, truck driver, and
farmer) at the time of the index arthroplasty, and they had continued in these
jobs against medical advice.
Surgical Technique
Patients were positioned supine with a tourniquet applied to the upper arm.
The previous skin incision was used in most patients; however, if there was
more than one previous incision, the one closest to the posterior midline was
utilized. The ulnar nerve was identified in all patients. The ulnar nerve had
been transposed anteriorly in twenty-three cases. In two cases the nerve was
identified posterior to the medial epicondyle and was transposed anteriorly at
the time of the revision surgery, and in one case the location of the nerve
could not be discerned from the operative report. The joint was exposed
through a Bryan-Morrey approach in sixteen
cases3, and in six
cases the triceps was left intact on the olecranon and the joint was exposed
through medial and lateral arthrotomies. In two cases a triceps-splitting
approach4 was
utilized, and in two cases an extended Kocher
approach5 was used.
Pathologic specimens were always sent for histologic examination and culture.
No signs of acute or chronic infection were found at the time of revision
surgery.
All component fractures occurred in a portion of the stem, adjacent to the
linkage or articular apparatus, where the implant was unprotected by host
bone. All humeral component fractures occurred at the junction between the
well-fixed proximal part of the stem and the less-well-supported distal part
of the stem. The London humeral component fractured at the interface between
the intramedullary pegs and the articular segment. Similarly, the ulnar
component fractures occurred at the junction of the well-fixed distal part of
the stem and the less-well-fixed proximal portion of the stem. In the
Coonrad-Morrey ulnar component, this high-stress area corresponded to the
porous-coated portion of the implant. Preoperative periarticular bone loss was
also evident: at the time of the primary total elbow arthroplasty, nine of the
ten humeri exhibited distal bone loss due to preexisting fracture or nonunion.
Proximal ulnar bone deficiency was noted at the time of three primary total
elbow arthroplasties.
Once the joint was exposed, the loose, fractured component that remained
linked to the intact component was disarticulated. The remaining, intact
component was examined and was tested for stability. In twenty-five of the
twenty-six revisions, the remaining implant was found to be stable. In
twenty-four of these revisions, only the fractured implant was revised. In one
patient with a fractured Ewald prosthesis, the fractured ulnar component and
the stable humeral component were both replaced with a Coonrad prosthesis.
A well-fixed fractured stem within the medullary cavity was revised in one
of two ways. In fourteen cases, the fractured component was removed and a new
component was cemented into an intact, expanded cement mantle, a technique
referred to as "cement-within-cement" or
"tap-out-tap-in" (Figs. 1-A and
1-B, 1-C,
1-E,
1-E). This technique was used
when preoperative radiographs had demonstrated an intact cement mantle with
preservation of the cement-bone interface without osteolysis. These findings
were confirmed intraoperatively by testing the cement mantle to ensure that it
was intact and stable. The well-fixed stem was removed by carefully excavating
a minimal amount of cement from around the fractured portion of the stem in
order to allow application of a needle-nosed vise-grip. The vise-grip was
linked to a slap-hammer, which allowed extraction of the fractured stem with
preservation of the surrounding cement mantle. Once the fractured stem was
removed, the cement mantle was evaluated; if it was intact and stable, the
mantle was enlarged with flexible reamers and drills. Attempts were made to
use the longest and largest-diameter revision component possible. Long trial
components were inserted; if the trial component was too long to fit into the
cement mantle, the component to be implanted was shortened intraoperatively
with a burr. Two of seven revision humeral components were modified
intraoperatively, to remove 15 and 20 mm of stem length. Two of seven revision
ulnar components were shortened intraoperatively, by 20 and 25 mm.
Reaming and drilling of the cement mantle expanded the diameter of the
mantle to allow insertion of a larger component, and it also created surface
imperfections that would enhance bonding between the new and old
cement6. Once the
canal was prepared, it was thoroughly irrigated with pulsatile lavage to
cleanse it of blood, bone, and cement debris, which have been shown to weaken
a new-old cement
interface7. The
cement mantle was then dried with suction and gauze packing. Antibiotic-laden
cement (1 g of gentamicin per 40 g of cement) was introduced, in the liquid
phase to prevent lamination, into the cleaned and dried mantle
cavity8. A venting
tube was not used. The new revision component was then inserted. The average
operative time for the "cement-within-cement" technique was 167
minutes (range, 129 to 243 minutes).
A more traditional approach was undertaken in the remaining twelve
revisions (Figs. 2-A and 2-B,
2-C, 2-D and 2-E). The entire
cement mantle was removed, either with the assistance of a cortical window
(three cases) or without it (nine cases). The revision was then done with
impaction bone allografting (one case) or with cementation into a reamed,
cement-free, prepared medullary cavity (eleven cases) with or without strut
allograft augmentation (three and eight cases, respectively). The average
operative time for this method was 213 minutes (range, 140 to 325 minutes).
Three humeral components were inserted in this fashion, and none required
intraoperative modification; however, two of nine ulnar components required
shortening, of 5 and 15 mm.
The polyethylene bushings were also examined at the time of revision and
were revised in twenty-one of the twenty-six cases. Five bushings were revised
because of wear and the remaining sixteen were revised prophylactically.
Evaluation
Patients were identified through our institution's computerized total joint
replacement database. The department protocol calls for all patients treated
with a total elbow arthroplasty to return for an interview, clinical
examination, and radiographic assessment. These evaluations are scheduled at
two to three months after the arthroplasty; at one year, two years, and five
years; and then at each subsequent five-year interval until revision or death.
Patients who are unable to return for evaluation are sent standardized
questionnaires or are interviewed by telephone. In addition, patients are
asked to forward all records and radiographs from subsequent orthopaedic or
medical consultations to the institutional database office. These records are
then reviewed and documented by the consulting orthopaedist. The accuracy and
completeness of the database have been shown previously to be approximately
95%9,10.
The total joint replacement database allowed retrieval of patient
demographics, the date of the revision surgery, and the implant type. A
retrospective chart review was then conducted to determine the presenting
findings and symptoms, type of revision, prior and subsequent procedures,
complications, and functional scores. Five of the patients in the original
cohort died from unrelated causes. One, who died at thirteen months and had
been last evaluated less than a year after the revision, was excluded from the
analysis, and four, who died at twenty-five months, six years, eight years,
and eleven years, were included in the study. Another patient refused to
participate, and still another underwent revision surgery at an institution
closer to home. This left twenty-one patients (twenty-three implants) followed
for a mean of 5.1 years (range, twelve months to twenty-one years). Ten
patients underwent clinical evaluation, nine patients were interviewed by
telephone and responded to a mail-in questionnaire, and two patients were
assessed by their local physician.
The Mayo Elbow Performance Score (MEPS), a performance index based on
subjective, objective, and functional
characteristics11,
was calculated at the final assessment. The MEPS assigns a maximum score of 45
points for pain, 25 points for daily functional activities, 20 points for
motion, and 10 points for stability. An outcome was considered to be excellent
if the score was =90 points, good if it was between 75 and 89 points, fair
if it was between 60 and 74 points, and poor if it was <60 points.
Radiographs made after the primary total elbow arthroplasty, after the
component fracture, immediately after the revision, and at the time of final
follow-up were assessed. The radiographs of the component fracture were
reviewed to determine the integrity of the distal humeral and proximal ulnar
bone and the location of the fracture. Distal humeral bone loss was classified
as grade I when the subchondral architecture was intact, grade II when the
medial and lateral supracondylar columns were preserved, grade III when either
the medial or the lateral supracondylar column was absent, and grade IV when
the entire distal part of the humerus was
absent12,13.
Ulnar bone loss was described with respect to the radiographic appearance of
the olecranon and the coronoid process. The olecranon was recorded as present
or absent, and the coronoid process was recorded as completely present, as
present but with only a thin cortex of bone remaining, or as absent.
The most recent follow-up radiographs were available for twenty patients
and were reviewed to assess bone loss, polyethylene bushing wear, and the
bone-cement interface. The bone-cement interface was classified as one of five
types on the basis of its appearance on the lateral radiograph, as described
by one of us (B.F.M.) and
Adams11. Type 0
indicated no radiolucency at the bone-cement interface; type 1, a
nonprogressive radiolucent line involving <50% of the interface; type 2, a
nonprogressive radiolucent line involving >50% of the interface; type 3, a
progressive radiolucent line involving <50% of the interface; type 4, a
progressive radiolucent line involving >50% of the interface; and type 5,
gross loosening of the implant. Polyethylene bushing wear was assessed on the
anteroposterior radiograph as described by Gill and one of us
(B.F.M.)2.
Statistical evaluation was performed with use of an unpaired t test, and a
p value of <0.05 was considered significant. This study was carried out
under a protocol approved by our institutional review board, and informed
patient consent was obtained.
Prevalence of Component Fracture
Between December 1979 and December 2003, 927 primary total elbow
arthroplasties were performed at our institution; 392 of them were done for
the sequelae of trauma (acute fracture, nonunion, or posttraumatic arthritis)
and 448, for rheumatoid arthritis. Eleven primary ulnar components that had
been inserted at our institution fractured, resulting in a 1.2% prevalence of
ulnar component fracture following primary total elbow arthroplasty. Six
primary humeral components that had been inserted at our institution
fractured, resulting in a 0.65% rate of humeral component fracture following
primary total elbow arthroplasty. The rates of ulnar and humeral component
fracture were 1.8% and 1%, respectively, following primary total elbow
arthroplasties performed for the sequelae of trauma and 0.9% and 0% following
those performed for rheumatoid arthritis. During the same study period, 346
revision total elbow arthroplasties were done at our institution, and the
component fracture rate following those revisions was 2.6%.
Clinical Assessment
Preoperatively, all patients had moderate-to-severe pain and had difficulty
using the affected upper extremity for daily activities.
At the time of final follow-up, at an average of 5.1 years (range, twelve
months to twenty-one years), the MEPS averaged 79 points (range, 50 to 100
points) and was excellent for eight patients, good for five, fair for six, and
poor for two (see Appendix). At that time, eight patients had no pain in the
elbow, seven had mild pain, and six had moderate pain. Of the patients with
moderate pain, one was being treated with suppressive antibiotics because of
deep infection, one had had recurrence of metastatic disease in the olecranon
with an associated triceps tendon avulsion, one had a triceps avulsion with a
failure of a repair that had been done elsewhere, and one had severe
rheumatoid arthritis with global involvement of the upper extremities and
cervical spine. Two patients rated their pain as moderate without explanation.
The average score for the pain component of the MEPS was 32 points (range, 15
to 45 points). The average final arc of motion was 108° (range, 70° to
145°), with flexion averaging 131° (range, 105° to 150°) and
extension averaging 23° (range, —15° to 45°). The mean final
score for the functional component of the MEPS was 21 points (range, 0 to 25
points) out of the total of 25 points.
Two subgroup analyses were done. In the first, the final outcomes of the
revisions due to the humeral component fractures were compared with those of
the revisions due to the ulnar component fractures. At a mean of 4.0 years
(range, twelve months to 6.2 years) after the revisions due to the humeral
component fractures, the MEPS score averaged 82 points (range, 55 to 95
points) and was excellent for three patients, good for four, and poor for one.
At a mean of 5.6 years (range, twelve months to twenty-one years) after the
revisions due to the ulnar component fractures, the MEPS averaged 79 points
(range, 50 to 100 points) and was excellent for six patients, good for two,
fair for six, and poor for one.
The second subgroup analysis compared the outcomes of the two different
revision techniques. The average MEPS score following the revision total elbow
arthroplasties done with the cement-within-cement technique was 82 points
(range, 55 to 100 points), whereas the average score following the revisions
done with the more traditional method of complete cement removal followed by
insertion of the revision component was 78 points (range, 50 to 100 points).
Ten of thirteen patients treated with the cement-within-cement procedure and
five of the ten treated with the traditional revision technique had a
satisfactory MEPS score. The mean operative time for the cement-withincement
procedures was significantly less (p = 0.009) than that needed for the more
traditional method.
Radiographic Assessment
All ten humeral component fractures were associated with preoperative bone
loss, which was grade IV in nine cases and grade II in one. Preoperatively,
three ulnar component fractures were associated with complete absence of the
olecranon; two, with complete absence of the coronoid; twelve, with partial
absence of the coronoid; and two, with minimal to no bone loss.
As graded on the radiographs made immediately following the revisions due
to the humeral component fractures, the humeral bone-cement interface was type
0 in five cases, type 1 in two, and type 2 in one. At the time of the final
radiographic evaluation, five of the humeral bone-cement interfaces had not
changed and three had increased in grade (from type 0 to 1, from type 2 to 3,
and from type 0 to 4). The patient with a type-4 interface underwent repeat
revision of the humeral component because of aseptic loosening at fifty-five
months. The index revision had been performed with the cement-within-cement
technique in that patient. Two of the six cement-within-cement humeral
revisions and one of the two traditional humeral revisions were associated
with an increase in the grade of the interfaces between the immediate
postoperative and final follow-up evaluations.
The radiographs made immediately following the ulnar component revisions
revealed a type-0 bone-cement interface in eight cases and a type-1 interface
in six. At the time of the final radiographic follow-up, ten of these
interfaces were unchanged, one had increased from type 0 to 1, two had
increased from type 1 to 2, and one had increased from type 1 to 3 (in a
patient who was taking oral suppressive antibiotics for the treatment of deep
infection). Recent radiographs were unavailable following one ulnar component
revision. Five ulnar bone-cement interfaces were unchanged and two had
increased in grade in both the group treated with cement-within-cement ulnar
revision and in the group treated with the traditional ulnar revision.
At the final radiographic examination, bushing wear was rated, as described
by Gill and Morrey2,
as normal/no wear in seventeen patients, partial wear in two patients, and
complete wear in one patient. The complete wear was seen at the time of a
105-month follow-up in a patient who subsequently died at eleven years
postoperatively.
Complications
Nineteen complications occurred in fourteen patients. There were seven
intraoperative complications, all cortical perforations measuring less than 1
× 1 cm, affecting six patients. Three perforations involved the humerus
(one was distal posterior and two were distal anterior), and four involved the
ulna (two were proximal and two were diaphyseal). Five of the seven
perforations were treated with bone-grafting (four with allograft struts and
one with cancellous bone chips).
Nerve complications developed in five patients. Three of them had transient
ulnar nerve injuries that resolved completely at six, eight, and twelve months
postoperatively. One patient had ulnar nerve paresthesias, with intact motor
function, that persisted at the time of final follow-up, at thirty-six months.
In one patient, who was lost to follow-up, a radial nerve palsy developed
after humeral revision surgery even though the radial nerve had been
identified and protected intraoperatively. Three patients sustained a triceps
tendon avulsion; one underwent successful reconstructive surgery at our
institution, another underwent unsuccessful repair elsewhere, and one refused
surgery.
One patient sustained an olecranon fracture during a manipulation,
performed with anesthesia, for the treatment of early postoperative stiffness
of the elbow. The fracture required open reduction and internal fixation.
Another patient sustained a stable periprosthetic humeral shaft fracture as a
result of a fall forty-three months after humeral revision. The fracture
healed uneventfully with functional bracing.
Three patients underwent additional revision surgery after the revision for
the component fracture. One patient who had isolated polyethylene bushing wear
with stable components had a bushing exchange at forty-one months. The second
patient had mechanical pain and was diagnosed with aseptic loosening of the
revised humeral component. This patient underwent humeral revision at
fifty-five months and was doing well sixteen months later. In the third
patient, who had had the index revision because of an ulnar component
fracture, a deep infection developed, which was treated with surgical
débridement and six weeks of intravenous antibiotics followed by
chronic oral antibiotic suppression. This patient eventually underwent
revision of the humeral component because of gross loosening and mechanical
pain at nineteen months. At fifty-six months after the index revision, the
patient was still taking suppressive oral antibiotics and had loosening of
both the humeral and the ulnar component.
Fatigue fractures of total joint arthroplasty components are rare, and most
of those reported in the English-language literature have involved femoral
components8,14-16.
However, there have been reports of fractures of shoulder hemiarthroplasty
components17 and
ulnar
components1,18.
Crowninshield et
al.16 found that
loss of proximal bone support for the implant was the initiating event in
fatigue fractures of femoral stems, and the same mechanism of failure probably
applies to fractures of elbow arthroplasty components. All of the humeral and
fifteen of the seventeen ulnar component fractures in our series were in
patients with deficient bone at the time of the index arthroplasty. Cantilever
loading of the well-fixed intramedullary stem occurs from the unsupported
periarticular portion of the stem, leading to excessive tensile stress at the
junction between the two.
The process of fatigue fracture consists of three stages: (1) fracture
initiation, (2) fracture propagation across the part, and (3) sudden final
fracture of the remaining cross section, which occurs when the applied load
surpasses the remaining strength of the
component19. In
general, the preliminary microfractures initiate and propagate under tensile
stresses that are concentrated at the junction between the well-supported and
unsupported portions of the stem. This mode of failure is unique in that
unexpected catastrophic failure occurs as a result of repetitive loading below
the static yield strength of the implant. There are several effective methods
for enhancing fatigue performance, such as insertion of the largest-possible
diameter of implant to distribute tensile stress and increase performance.
Preservation of bone or augmentation of deficient bone provides global implant
support and decreases the concentration of stresses. Stresses are also reduced
or eliminated by "streamlining" the construct or avoiding sharp
interfaces. For example, anterior and posterior allograft struts placed on the
distal part of a humerus should be staggered, as opposed to all struts ending
at the same level, to decrease stress concentration.
In our series, twenty-five of the twenty-seven fractured components were
constructed of titanium alloy. Titanium has several advantages, such as a
lower modulus of elasticity, superior biocompatibility, and enhanced corrosion
resistance, over more conventional stainless steels and cobalt-based
alloys20. Despite
these advantages, the fatigue life of titanium is substantially decreased by
scratches or notches on its surface, a trait termed "notch
sensitivity." The original design of the Coonrad-Morrey implant employed
titanium beads in order to enhance cement fixation or to provide the option
for cementless application. Unfortunately, the sintering process weakened the
substrate of the titanium, which, when coupled with the notch-sensitivity
effect, predisposed the implant to fatigue failure. This deficiency was
recognized, and in 1993 a new ulnar component was introduced. This markedly
increased the strength of the device. However, the new design employed a
precoat surface of polymethylmethacrylate, which introduced an unintended and
unforeseen consequence, as so often happens. It was observed over time that,
while the surface resulted in a stronger implant, it was vulnerable to
micromotion and associated with an increased rate of osteolysis. Currently,
the Coonrad-Morrey ulnar component, which has been employed since 2000,
incorporates a plasma-spray surface. This device can be manufactured without
weakening the substrate of titanium. To our knowledge, there have been no
reported fractures of the ulnar implant since 2000 and the problem of
osteolysis associated with the precoat has not been noted since the
introduction of the new component.
The mean age of our patients at the time of the primary elbow arthroplasty
was fifty-five years (range, twenty-three to seventy-eight years), which is
younger than the mean ages of patients undergoing primary total elbow
arthroplasty for posttraumatic arthritis (fifty-seven
years18),
rheumatoid arthritis (sixty-six
years2), and distal
humeral fracture (sixty-seven
years21), as
reported in the literature from our institution. This younger cohort may have
placed higher demands on the prostheses, increasing their susceptibility to
fatigue failure. The rate of component fracture was highest following
revisions (2.6%), perhaps as a result of the greater bone deficit, fewer
soft-tissue stabilizers, and weakened remaining bone support. Implant
fractures were also more prevalent following arthroplasties done for
trauma-related causes than they were following arthroplasties in patients with
rheumatoid arthritis. Patients who have an arthroplasty for trauma-related
reasons tend to place higher demands on the limbs and have a tendency for
increased and excessive use of a previously functionless joint after it has
been rendered pain-free and stable by the total elbow
arthroplasty18.
Two methods of revision—the more traditional approach and the
cement-within-cement technique—were employed in this study cohort.
Cement-within-cement revision was used when the cement mantle was intact and
stable. To our knowledge, this technique has not been previously reported for
revision elbow arthroplasty, although it has been used for revision total hip
arthroplasty with good
outcomes6,8,22.
When intraoperative conditions were conducive to use of the
cement-within-cement technique, the revision elbow arthroplasties were faster
than, and the outcomes were similar to, the more traditional revision
techniques. The complications were also similar in type and distribution
between the two techniques.
In this series, nineteen complications occurred in fourteen patients.
Revision total elbow arthroplasty is technically demanding and is associated
with a high complication
rate12,23-26.
Implant fractures are infrequent after total elbow arthroplasty, and their
likely etiology is fatigue failure from loading with inadequate osseous
support. Effective methods for enhancing fatigue performance include insertion
of the largest-diameter implant possible, avoidance of notching of components,
reduction of stress risers by streamlining the reconstruction, and
preservation of bone stock and/or augmentation of bone stock with allograft as
necessary. Patients who are noncompliant with restrictions necessitated by
total elbow arthroplasty must be advised that they are probably at higher risk
for catastrophic component failure.
Several different reconstructive techniques can be used to revise fractured
components, with successful outcomes. We found the cement-within-cement
technique to be effective and efficient, and we believe that it can be added
to the armamentarium of possible reconstructive techniques for revision total
elbow arthroplasty.
A table showing the demographic, clinical, and follow-up data on all
patients is available with the electronic versions of this article, on our web
site at
(go to the article citation and click on "Supplementary Material")
and on our quarterly CD-ROM (call our subscription department, at
781-449-9780, to order the CD-ROM). ?
Note: The authors acknowledge the assistance of Dr. Damian
Rispoli in the preparation of this paper.
Shafer BL, Fehringer EV, Boorman RS,
Churchill RS, Matsen FA 3rd. Ulnar component fracture after revision total
elbow arthroplasty with proximal ulnar bone loss: a report of 2 cases.
J Shoulder Elbow Surg.
2003;12:
297-301.12297
2003
[PubMed][CrossRef]
Gill DR, Morrey BF. The Coonrad-Morrey
total elbow arthroplasty in patients who have rheumatoid arthritis. A ten to
fifteen-year follow-up study. J Bone Joint Surg Am.
1998;80:
1327-35.801327
1998
[PubMed]
Bryan RS, Morrey BF. Extensive posterior
exposure of the elbow. A tricepssparing approach. Clin Orthop Relat
Res. 1982;166:
188-92.166188
1982
Gschwend N, Scheier NH, Baehler AR.
Long-term results of the GSB III elbow arthroplasty. J Bone Joint Surg
Br. 1999;81:
1005-12.811005
1999
[CrossRef]
Weiland AJ, Weiss AP, Wills RP, Moore
JR. Capitellocondylar total elbow replacement. A long-term follow-up study.
J Bone Joint Surg Am.
1989;71:
217-22.71217
1989
[PubMed]
Greenwald AS, Narten NC, Wilde AH.
Points in the technique of recementing in the revision of an implant
arthroplasty. J Bone Joint Surg Br.
1978;60:
107-10.60107
1978
[PubMed]
Li PL, Ingle PJ, Dowell JK.
Cement-within-cement revision hip arthroplasty, should it be done? J
Bone Joint Surg Br. 1996;78:
809-11.78809
1996
Lieberman JR, Moeckel BH, Evans BG,
Salvati EA, Ranawat CS. Cement-withincement revision hip arthroplasty.
J Bone Joint Surg Br.
1993;75:
869-71.75869
1993
[PubMed]
McGrory BJ, Morrey BF, Rand JA, Ilstrup
DM. Correlation of patient questionnaire responses and physician history in
grading clinical outcome following hip and knee arthroplasty. A prospective
study of 201 joint arthroplasties. J Arthroplasty.
1996;11:
47-57.1147
1996
[PubMed][CrossRef]
Berry DJ, Kessler M, Morrey BF.
Maintaining a hip registry for 25 years. Mayo Clinic experience. Clin
Orthop Relat Res. 1997;344:
61-8.34461
1997
[CrossRef]
Morrey BF, Adams RA. Semiconstrained
arthroplasty for the treatment of rheumatoid arthritis of the elbow. J
Bone Joint Surg Am. 1992;74:
479-90.74479
1992
King GJ, Adams RA, Morrey BF. Total
elbow arthroplasty: revision with use of a non-custom semiconstrained
prosthesis. J Bone Joint Surg Am.
1997;79:
394-400.79394
1997
[PubMed][CrossRef]
Morrey BF, Adams RA, Bryan RS. Total
replacement for post-traumatic arthritis of the elbow. J Bone Joint
Surg Br. 1991;73:
607-12.73607
1991
Charnley J. Fracture of femoral
prostheses in total hip replacement. A clinical study. Clin Orthop
Relat Res. 1975;111:
105-20.111105
1975
[CrossRef]
Woolson ST, Milbauer JP, Bobyn JD, Yue
S, Maloney WJ. Fatigue fracture of a forged cobalt-chromium-molybdenum femoral
component inserted with cement. A report of ten cases. J Bone Joint
Surg Am. 1997;79:
1842-8.791842
1997
Crowninshield RD, Maloney WJ, Wentz DH,
Levine DL. The role of proximal femoral support in stress development within
hip prostheses. Clin Orthop Relat Res.
2004;420:
176-80.420176
2004
[PubMed][CrossRef]
Zuckerman JD, Shapiro JA, Moghtaderi S,
Kummer FJ. Fatigue failure of a shoulder hemiarthroplasty stem: a case report.
J Shoulder Elbow Surg.
2003;12:
635-6.12635
2003
[PubMed][CrossRef]
Schneeberger AG, Adams R, Morrey BF.
Semiconstrained total elbow replacement for the treatment of post-traumatic
osteoarthrosis. J Bone Joint Surg Am.
1997;79:
1211-22.791211
1997
[PubMed]
Callister WD Jr. Materials
science and engineering: an introduction. 3rd ed. New York: Wiley;
1994.
1994
Long M, Rack HJ. Titanium alloys in
total joint replacement—a materials science perspective.
Biomaterials. 1998;19:
1621-39.191621
1998
[PubMed][CrossRef]
Kamineni S, Morrey BF. Distal humeral
fractures treated with noncustom total elbow replacement. J Bone Joint
Surg Am. 2004;86:
940-7.86940
2004
McCallum JD 3rd, Hozack WJ. Recementing
a femoral component into a stable cement mantle using ultrasonic tools.
Clin Orthop Relat Res.
1995;319:
232-7.319232
1995
[PubMed]
Sanchez-Sotelo J, O'Driscoll S, Morrey
BF. Periprosthetic humeral fractures after total elbow arthroplasty: treatment
with implant revision and strut allograft augmentation. J Bone Joint
Surg Am. 2002;84:
1642-50.841642
2002
[CrossRef]
Kamineni S, Morrey BF. Proximal ulnar
reconstruction with strut allograft in revision total elbow arthroplasty.
J Bone Joint Surg Am.
2004;86:
1223-9.861223
2004
[PubMed]
Mansat P, Adams RA, Morrey BF.
Allograft-prosthesis composite for revision of catastrophic failure of total
elbow arthroplasty. J Bone Joint Surg Am.
2004;86:
724-35.86724
2004
[PubMed]
Redfern DR, Dunkley AB, Trail IA,
Stanley JK. Revision total elbow replacement using the Souter-Strathclyde
prosthesis. J Bone Joint Surg Br.
2001;83:
635-9.83635
2001
[PubMed][CrossRef]