Between June 1982 and December 2000, 204 total elbow prostheses were
inserted in 166 patients at our department. None of the patients had had a
previous total elbow arthroplasty of the same joint. Thirty-eight patients had
a bilateral procedure. The mean duration of follow-up was 6.4 years (range,
two to nineteen years). Forty-one elbows were followed for longer than ten
years.
The patients included 117 women and forty-nine men. The average age of the
patients at time of the operation was sixty-one years (range, twenty-four to
eighty-three years). The average age of the men was sixty years, and the
average age of the women was sixty-two years. The operation was performed in
119 right elbows and eighty-five left elbows. Two surgeons performed all of
the operations, although the senior author (P.M.R.) inserted most (173) of the
prostheses.
All patients had advanced rheumatoid arthritis. The diagnosis was
seropositive rheumatoid arthritis for 183 elbows, seronegative rheumatoid
arthritis for sixteen elbows, and juvenile rheumatoid arthritis for five
elbows. The indications for the insertion of a total elbow prosthesis were
pain in and restriction of motion of the elbow joint combined with grade-III,
IV, or V destruction of the elbow joint according to the radiographic criteria
of Larsen et al5.
Preoperatively, sixteen elbows (8%) had Larsen grade-III changes, ninety (44%)
had Larsen grade-IV changes, and ninety-eight (48%) had Larsen grade-V
changes. The mean time between the diagnosis of rheumatoid arthritis and the
insertion of the prosthesis was 19.1 years (range, one to fifty-nine years);
the mean interval was 15.3 years for men and 20.6 years for women (Student t
test, p < 0.01). Thirty-two elbows had been treated with synovectomy at a
mean of 3.9 years before insertion of the prosthesis; in fourteen of these
elbows, synovectomy had been performed in combination with excision of the
radial head.
Preoperatively, forty patients (forty-three elbows; 21%) were unable to
work, seventy-one patients (eighty-five elbows; 42%) could perform light
housework, and thirty-one patients (forty-seven elbows; 23%) could perform
light manual work. The remaining twenty-four patients (twenty-nine elbows;
14%) were able to perform most daily tasks.
Forty-five elbows (22%) had no pain at rest, sixty-nine (34%) had mild
pain, sixty-five (32%) had moderate pain, and twenty-five (12%) had severe
pain.
All patients were examined by the senior author and were assessed
clinically and radiographically before the operation, at one and two years
after the operation, and then at regular intervals thereafter with use of a
shortened version of the assessment developed by
Souter6. The data
for all patients were collected with MRDM software (Medical Research Data
Management,
).
This software was converted into SPSS (SPSS, Chicago, Illinois) for the data
analysis.
Forty-four patients (fifty-five elbows) died of unrelated causes at a mean
of 6.5 years after the operation. Two patients, who had the operation in 1993
and 1996, were followed for two and four years, respectively, but could not be
traced after that time. Twenty-four elbows were revised at a mean of 4.6 years
postoperatively because of loosening of one or two components (ten), loosening
after fracture (six), dislocation (four), infection (two), diaphyseal humeral
fracture (one), and restricted range of motion (one). Another nine patients
(nine elbows) had an excision arthroplasty.
The Prosthesis
The Souter-Strathclyde total elbow prosthesis was developed in 1973 and was
first inserted in 1977. The humeral component is made of Vitallium and has a
flat intramedullary stem for fixation in the epicondylar ridges, with flanges
for the capitellum and medial epicondyle of the humerus. The ulnar component
is made of ultra-high-molecular-weight polyethylene and has a keel and a small
stem. Both components were fixed with Palacos cement (Schering-Plough,
Kenilworth, New Jersey). Different sizes are available for both components. A
snap-fit metal-backed ulnar component was developed to increase stability in
elbows that may be unstable. We used sixty-two small, 102 medium, twenty-seven
large, and thirteen medium long-stemmed humeral components
(Fig. 1) and ninety-two small,
ninety-two medium, and twenty metal-backed ulnar components with a 5-cm stem.
Fifteen of the twenty metal-backed ulnar components had a snap-fit design. The
determination of the component sizes was made preoperatively with use of
templates.
Operation
The operation was carried out by one of two surgeons with use of the same
technique4,7.
Prophylactic intravenous antibiotics were routinely employed.
Until 1994, the fibrous arch that covers the ulnar sulcus was incised only
in patients who had preoperative complaints related to the ulnar nerve. Since
1994, this release has been performed routinely. In eighty-two elbows (40.2%),
a humeral bone plug was used and the cement was pressurized into the bone
marrow during insertion of the prosthesis. This procedure has been used
routinely since 1994. The mean duration of the operation was 165 minutes
(range, eighty-five to 275 minutes). An ipsilateral shoulder arthroplasty was
performed in eight patients, and an arthrodesis of the ipsilateral wrist was
performed in three patients at the same time.
Postoperative Treatment
The patients were treated with a compression dressing for five days
postoperatively, followed by collar-and-cuff immobilization. Passive and
active flexion and extension exercises were started on the fifth day under the
supervision of a physiotherapist. All patients were treated with a posterior
splint at night for six weeks postoperatively.
Intraoperative Complications
In five elbows, the medial epicondyle (four) or the lateral epicondyle
(one) was fractured during the trial insertion of the standard (short-stemmed)
humeral component. Because of the fractures, a long-stemmed humeral component
was used.
Statistical Analysis
The Cox model was used to investigate the effect of different component
sizes and the use of a bone plug on the rate of revision, with the level of
significance set at p < 0.05. Pain, function, and ulnar nerve dysfunction
after the operation were analyzed by means of a generalized linear mixed model
with multivariate normal random effects with use of penalized
quasi-likelihood, with the level of significance set at p <
0.058,9.
Housekeeping ability, an ordinal variable, was analyzed in a standard linear
mixed-effects model with use of maximum likelihood, with the level of
significance set at p < 0.05. The Pearson chi-square test for independence
was used first for the correlation between incision of the ulnar arch and the
presentation of postoperative complaints related to the ulnar nerve and second
for the correlation between the use of a humeral bone plug and humeral
loosening, with the level of significance set at p < 0.05. Finally, the
Kaplan-Meier method was used to analyze the overall survival of the
prosthesis, with removal of the prosthesis (with or without revision) as the
end point. All data were analyzed with use of S-PLUS software (Insightful,
Seattle, Washington).
The clinical results are summarized in
Table I.
Pain
At the time of the latest follow-up, 139 (68.1%) of the 204 elbows had no
pain, six (2.9%) had pain at rest, and fifty-nine (28.9%) had occasional pain
with loading or stress. One hundred and seventy-one elbows (83.8%) had
improved strength after the operation, eight (3.9%) had worse strength after
the operation, and twenty-five (12.3%) had no change in strength after the
operation. There was a trend toward an increase in pain with time since the
operation. All four measures of pain (pain at rest, pain at night, pain with
movement, and pain with loading or stress) increased during the follow-up
period (p < 0.05). The severity of pain also increased with time, although
this increase was not significant.
Function
Motion improved after the operation; specifically, the mean range of
flexion increased from 109° to 134°, the mean range of supination
increased from 34° to 57°, and the mean range of pronation increased
from 60° to 75°. The mean flexion deformity increased from 32° to
33°. None of the five elbow functions that were assessed changed
significantly during the follow-up period.
Ulnar Nerve
Preoperatively, thirty-one patients (thirty-six elbows) had paresthesias in
the distribution of the ulnar nerve. In twenty patients (twenty-two elbows),
this complaint resolved postoperatively. Ten patients (ten elbows) without
previous neurological symptoms had development of paresthesias in the
distribution of the ulnar nerve postoperatively; in eight patients (eight
elbows), the paresthesias developed within three months of follow-up. In three
patients (three elbows) with complete loss of sensitivity, release of the
ulnar nerve was performed within one week. The other seven patients (seven
elbows) with paresthesias or pain were treated nonoperatively. Of the ten
patients with new complaints, eight had no complaints related to the ulnar
nerve at the time of the latest follow-up and two had paresthesias of the
ulnar nerve sometimes but the complaints did not interfere with the activities
of daily life. The prevalence of postoperative complaints related to the ulnar
nerve decreased from twenty of eighty-six among elbows treated before 1994 to
eighteen of 118 among elbows treated after 1994, but this difference was not
significant (Pearson chi-square test for independence).
Size of Components
Despite the trend in the
literature10,11
toward the use of large and long-stemmed humeral components and metal-backed
snap-fit ulnar components, we still prefer standard components because of the
small amount of bone that has to be removed before insertion. In the present
study, no significant difference could be found between sizes of components
and the survival rate (p = 0.20 and p = 0.09), even after exclusion of the
long-stemmed components.
Use of Humeral Bone Plug
In the present series, no significant difference in the prevalence of
humeral loosening was observed between elbows treated with and without the use
of a humeral bone plug (Pearson chi-square test for independence).
Relationship Between Preoperative Status and Postoperative
Outcome
The relationship between preoperative clinical scores and outcome is of
importance. If a patient had pain at rest or at night before the operation,
there was a significant chance that pain at rest or at night would develop
after the operation. With regard to function, the preoperative ability to lift
a kettle or teacup with the ipsilateral hand seemed to enhance the chance that
the patient would have the same ability postoperatively (p = 0.03). For all
other activities tested, better function of the rheumatoid elbow before the
operation seemed to ensure better function after the operation.
Patients with paresthesias in the distribution of the ulnar nerve before
the operation had a high risk for the development of ulnar complaints after
the operation (p < 0.01). Release of the ulnar arch (routinely performed
after 1994) did not change this relationship.
Prognostic Factors for Humeral Loosening
With the numbers available, gender, age, Larsen classification, use of a
humeral bone plug, and the length of time between the diagnosis of rheumatoid
arthritis and the total elbow arthroplasty did not significantly increase the
chance of humeral loosening.
If pain at rest or at night was present during the follow-up assessments,
the chance of humeral loosening was increased (p < 0.001). Restriction of
function during the follow-up period also enhanced the chance of humeral
loosening; this relationship was significant for all functions measured
(Table I).
Complications
The complication rate was 29.9% (Table
II). The total elbow prosthesis was revised in twenty-four elbows
because of loosening of the humeral component (ten), fracture of the distal
part of the humerus or proximal part of the ulna with secondary loosening
(six), dislocation (four), infection (two), restricted range of motion (one),
and fracture of the humerus proximal to the prosthesis necessitating revision
with a long-stemmed humeral component (one). One prosthesis was removed
because of humeral loosening, and ten were removed because of deep infection
(Table III). In three elbows
the infection occurred early (one, three, or four months after surgery),
whereas in seven elbows the infection occurred at an average of seven years
(range, two to fourteen years) postoperatively. The infection was eradicated
in all ten elbows after resection arthroplasty and appropriate antibiotic
therapy. Eight elbows did not have a reoperation after the resection
arthroplasty. At the time of the most recent follow-up, all were free of pain
but had restriction of function because of instability. In three of the ten
elbows that were revised because of loosening of the humeral component, the
ulnar component was left in situ. Three prostheses that were revised with a
long-stemmed humeral component loosened again after an average of 4.1 years.
Re-revision was performed with a long-stemmed component of the same size
combined with bone-grafting. In another two elbows, the site of the revision
arthroplasty became infected after seventy-one and thirty-one months; both
elbows were treated with resection arthroplasty. None of the latter five
elbows were painful, and no complications were seen at the time of the latest
follow-up.
Nine patients had development of infected rheumatoid nodules in the
involved elbow years after the operation. Four patients (four elbows) that had
a superficial infection were treated with excision of the nodules and
antibiotics, with the prosthesis left in situ. Three of these patients were
asymptomatic at the time of the most recent follow-up; the fourth patient
still had purulent drainage from the nodules but removal of the prosthesis was
not indicated because of poor general health status. The other five patients
(five elbows) initially had a superficial infection, but by the time that they
were evaluated a deep infection had developed. In all five cases, the
prosthesis was removed as described above.
Seven elbows dislocated within the first few months after surgery. All
seven elbows initially were treated with closed reduction and two weeks of
immobilization. Three elbows had no additional dislocations. The other four
could not be sufficiently reduced, but the patients refused surgical treatment
because of the absence of clinical complaints. After a few years, however, all
four elbows were painful and were revised with use of a snap-fit ulnar
component. None of the four elbows had additional dislocations of the
prosthesis.
Three patients had release of the ulnar nerve within four weeks after the
original arthroplasty because of sensory and motor impairment. The ulnar nerve
complaints resolved in all three patients following surgery.
Radiographic Evaluation
Radiographs were made for all patients at the time of the latest follow-up.
Five patients (five elbows) had radiographic signs of loosening according to
the criteria of
Souter6 (a complete,
progressive radiolucent line of >1 mm around the component)
(Fig. 2). These patients did
not have clinical complaints, and a revision operation was not
recommended.
At the time of the latest follow-up, partial radiolucent lines were
detected around the proximal part of the humeral component in sixty-two elbows
(36.3%) and in the supracondylar area in fifty-two elbows (30.4%).
At the time of the last follow-up, radiolucent lines were seen along the
ulnar shaft in twenty-four elbows (14.0%) and around the olecranon in
fifty-four elbows (31.6%). No migration of any humeral or ulnar component was
noted at the time of the latest follow-up.
Survival
Kaplan-Meier analysis with removal of the prosthesis (with or without
revision) as the end point showed an overall survival rate of 90.1% (95%
confidence interval, 85.6% to 94.6%) after five years, 77.4% (95% confidence
interval, 68.7% to 86.0%) after ten years, and 65.2% (95% confidence interval,
24.3% to 100.0%) after eighteen years (Fig.
3). All elbows, including those in which the prosthesis was
revised or removed during the first few months, were taken into account.
Overall, 94.9% of all 166 patients were very satisfied with the prosthesis at
the time of the latest follow-up and stated that they would undergo this
operation again.
The present study confirms the results associated with the
Souter-Strathclyde prosthesis as reported in previous
studies12-17.
Overall, we found a removal rate of 16.2% and a complication rate of 29.9%.
These rates are in accordance with those associated with other total elbow
arthroplasties18-21.
Unlike previous investigators, we counted all elbows, even when the implant
had been removed or revised within a few months after the procedure.
Strikingly, none of the elbows with more than thirteen years of follow-up were
revised after this period. The overall rate of survival of the primary
Souter-Strathclyde elbow prosthesis in the present study was equal to those
reported in other recent
studies13,17.
Most studies in the literature have involved the use of standard
elbow-scoring systems developed by different authors, with the result being
rated on a scale from 0 to 100 points. A reliable comparison between these
studies is not possible because of the different weights for each of the
scored
items22,23.
Because pain is the most important indication for a total elbow prosthesis in
patients with rheumatoid arthritis, most patients are satisfied after the
operation and, remarkably, 94.9% of the patients in the present series stated
that they would undergo this operation again.
Trail et al. noted differences in the survival rate in association with the
use of different component
sizes11. The use of
a long-stemmed humeral component combined with a metal-backed ulnar component,
with or without a snap-fit, seems to be associated with a decreased rate of
early loosening when compared with the use of standard-sized components.
However, less bone stock is left when the larger components are used for
primary surgery, and this can complicate fixation when revision is required.
In the present series, we did not observe any differences in the survival rate
among components of different sizes. For this reason, we prefer to use the
short-stem standard ulnar component (small or medium) and the short-stem
standard humeral component (small, medium, or large) for primary total elbow
replacement in patients with normal amounts of bone loss.
In hip and knee replacement, the use of a femoral bone plug reduces blood
loss and improves the mantle of cement around the
prosthesis24-26.
However, we did not observe any improvement in terms of clinical or
radiographic outcome in association with the use of a bone plug in the
humerus.
For radiographic follow-up, we used only conventional radiographs in two
planes. In a previous study in which radiostereometry was used, eight (44.4%)
of eighteen elbows had signs of aseptic loosening of the humeral component in
the first two years after insertion of the Souter-Strathclyde
prosthesis27. In
contrast, with use of conventional radiographs, we found a loosening rate of
only 7.8%. This difference may be related to the fact that loosening that is
detectable with radiostereometry precedes loosening that is detectable with
conventional radiographs. The importance of the location and thickness of
radiolucent lines in relation to early loosening is still unclear. Other
factors that may be important in the development of early loosening are the
position of the prosthesis in the bone after insertion, the postoperative
biomechanical axis, and the quality of the bone
stock28.
A relationship was found between the presence of preoperative symptoms
(pain, restriction of function, and ulnar nerve complaints) and the chance
that the same symptoms would develop postoperatively. Only for ulnar
complaints has this relationship been described previously in the
literature19.
High rates of loosening of the humeral component have been described
previously in the literature, but the reason for this complication is still
not
known4,13,17,27.
In the series of Ikävalko et
al.10, long-stemmed
humeral components showed a decreased rate of loosening compared with standard
components.
Loosening of the prosthesis after fracture is of concern. In their series
of 525 Souter-Strathclyde elbow prostheses, Ikävalko et al. described
seven cases of loosening that were associated with
fractures13. In the
present study, six elbows had loosening of the prosthesis after a fracture; in
all six cases, the olecranon or the distal part of the humerus had fractured
as the result of a fall or severe loading. Poor bone stock combined with
excessive bone resection at the time of insertion of the prosthesis could play
a role in this complication.
Incising the fibrous ulnar arch has become a routine part of our surgical
technique, but routinely visualizing, releasing, and transposing the ulnar
nerve remains controversial in the
literature6,14.
A surgeon can damage the ulnar nerve unnecessarily in patients with few or no
complaints. However, we believe that incising the fibrous ulnar arch, with or
without transposition of the ulnar nerve, is mandatory for patients with ulnar
nerve compression symptoms before the operation. It is important to
discriminate between preexisting and new postoperative ulnar nerve
problems.
The dislocation rate in the present series (3.4%) was comparable with those
in other
reports18,29.
All dislocations of the prosthesis occurred in the first few months after
surgery. Although the Souter-Strathclyde prosthesis might be more constrained
than other nonconstrained elbow prostheses, the quality and tension of the
collateral ligaments after insertion of the prosthesis remain very important
for the prevention of
dislocation21. When
the ligaments cannot be balanced and instability remains at the time of trial
reduction, the surgeon might opt for the snap-fit ulnar component.
The reported rate of humeral loosening in elbows treated with the
Souter-Strathclyde implant has varied in the literature, but Ikävalko et
al. reported a loosening rate of 6% in their recent study of 525
Souter-Strathclyde
prostheses13. The
more frequent use of long-stemmed humeral components for primary surgery
probably will reduce the rates of early loosening in the future.
We conclude that the primary Souter-Strathclyde prosthesis is a
satisfactory solution for the treatment of severely symptomatic rheumatoid
arthritis of the elbow. The need for minimal bone resection before insertion
of the Souter-Strathclyde prosthesis provides a major advantage compared with
other elbow prostheses, especially when the potential for revision operations
is taken into account. Currently, the results associated with the
Souter-Strathclyde total elbow prosthesis are comparable with those associated
with other prostheses, but loosening of the humeral component remains a
concern.