Dislocation is one of the most common complications of total hip
arthroplasty and can cause major distress for the patient and the treating
surgeon1.
A study of 331 patients showed that nonoperative treatment is successful in
about two-thirds of patients who have instability of the hip, with the
remaining one-third requiring some form of surgical
intervention2. The
outcome of a reoperation for the treatment of instability without a clearly
identifiable cause is
unpredictable3-7.
Even when a correctable cause such as component malpositioning is found and
corrected, the success rate is approximately two of
three3. Surgical
options for treatment include revision to correct the position of the
components3,6,8,
trochanteric
advancement9,
improvement of soft-tissue tensioning by increasing
offset10, removal
of impinging
tissue6,11,
augmentation of the acetabular
lining12-14,
bipolar
arthroplasty15,16,
and insertion of a constrained cup
mechanism17,18.
Recurrent instability in patients in whom prior surgical procedures have
failed to restore stability presents a particularly challenging problem. These
patients may have extensive soft-tissue deficiency as an important
contributing cause of the instability. Revision arthroplasty with use of a
constrained cup has been proposed for these more challenging
situations17,18.
Theoretical concerns related to these implants include a possibly increased
risk of component loosening due to increased interfacial stresses and possible
adverse effects on polyethylene wear due to the thinner polyethylene and
increased number of articulating surfaces with some constrained designs.
At our institution, constrained acetabular components have been used for
the salvage of unstable hip replacements in three situations: (1) cases in
which no identifiable cause of instability could be found and corrected, (2)
cases in which prior surgical attempts at stabilization have failed, and (3)
cases in which there is marked deficiency of the soft-tissue attachments
around the hip, particularly involving the hip abductors.
The purpose of the present study was to evaluate the clinical and
radiographic outcome associated with the use of a single design of constrained
acetabular component for the treatment of instability, or potential
instability, at the site of a total hip arthroplasty.
Demographic Characteristics
All patients who had undergone hip arthroplasty with use of a constrained
acetabular component between 1985 (when we first began to use this implant)
and 1998 were identified. During the years of the study, 29,341 primary and
revision total hip arthroplasties were performed at our institution, with 111
arthroplasties in 110 patients involving the use of a constrained acetabular
cup. During this period, 212 other procedures were performed because of
instability at the site of a hip arthroplasty; these procedures included 135
cup revisions, nineteen stem revisions, twenty-five cup and stem revisions,
and thirty-three procedures performed because of a failed acetabular
insert.
Of the 110 patients who were identified, one had been lost to follow-up
before two years, leaving 109 patients (110 hips) available for study. The
patients included fifty-six women and fifty-three men who had had a mean age
of sixty-six years (range, twenty-one to eighty-nine years) at the time of the
index arthroplasty. The mean weight and height of the patients had been 72 kg
(range, 58 to 154 kg) and 169 cm (range, 152 to 186 cm), respectively.
Fifty-two arthroplasties had been performed on the left hip, and fifty-eight
had been performed on the right hip.
Follow-up
Clinical and radiographic data on all patients undergoing arthroplasty were
collected prospectively. Patients were contacted at a minimum of two months,
one year, two years, five years, and every five years thereafter on a routine
basis. The average duration of clinical follow-up was 3.2 years (range, two to
eight years), and the average duration of radiographic follow-up was 2.9 years
(range, two to eight years). All patients were followed clinically for a
minimum of two years, until failure of the hip prosthesis, or until death. One
patient was lost to follow-up, as mentioned previously. Of the 102 patients
who were alive at the end of the study period, ninety had been evaluated most
recently with a physical and radiographic examination and an interview in the
physician's office, ten had returned a detailed questionnaire by mail, and
two, who were unable to travel, had been interviewed by telephone and had had
radiographs made locally and sent for review. All patients who had died had
been evaluated with physical examinations at postoperative intervals, with
stability of the hip documented at the time of those visits.
Radiographic Evaluation
Serial anteroposterior and lateral radiographs of the involved joint were
reviewed to assess the position of the prosthesis and to look for signs of
loosening or wear of the implant. The femur was divided into seven
zones19 and the
acetabulum was divided into three
zones20, in keeping
with previous studies, in order to evaluate the location of radiolucent lines.
Definite loosening was defined as the presence of a complete radiolucent line
in all zones on any radiograph, femoral component subsidence of =2 mm, or
acetabular component migration or
tilt21.
Additionally, any cemented prosthesis with a cement fracture or any uncemented
prosthesis with progressive bead-shedding was defined as being
loose22. Possible
loosening was defined as the presence of a radiolucent line occupying >50%
but <100% of the bone-cement interface on any radiograph or the presence of
a progressive radiolucent line.
Preoperative Data
The underlying diagnosis at the time of the initial, primary hip
arthroplasty had been osteoarthritis in fifty-five hips, posttraumatic
arthritis in twenty-three, rheumatoid arthritis in fifteen, avascular necrosis
in eight, developmental dysplasia of the hip in seven, slipped capital femoral
epiphysis in one, and ankylosing spondylitis in one. Including the initial
arthroplasty, all patients had undergone at least two (average, 3.4; range,
two to nineteen) reconstructive operations on the hip prior to insertion of
the constrained cup. Seventy-nine patients had undergone at least one revision
specifically aimed at correcting the unstable hip.
The constrained acetabular components used in the present series consisted
of a bipolar mechanism captured with a polyethylene insert designed for
modular assembly to the PSL acetabular shell (Howmedica, Rutherford, New
Jersey) (100 hips) or cemented into a well-fixed cup (eleven hips).
Surgical Data
In seventy-nine hips the constrained acetabular cup was implanted for the
treatment of recurrent instability (dislocation), and in thirty-one hips it
was implanted because of absent or grossly deficient soft-tissue attachments
that were believed to be associated with a high risk for subsequent
instability.
The surgical approach for the index operation was anterolateral in
eighty-three hips, posterolateral in fourteen, transtrochanteric in five, and
combined (anterolateral with a trochanteric osteotomy) in eight. At the time
of the index operation, the existing acetabular and femoral components were
tested for fixation. Thirty-seven femoral components were found to be loose
and were revised. Of the 110 acetabular components, forty-one were well fixed
and well positioned and hence were not revised. Malpositioning was defined as
10° of variation from optimal positioning of the cup at 45° of
abduction and 10° of anteversion.
Of the forty-one well-fixed shells, thirty-nine had been inserted in a
press-fit fashion at a mean of 6.9 years (range, one to thirteen years)
previously and two had been cemented in place nine and twelve years
previously. The constrained acetabular insert was secured with cement into
eleven of these well-fixed shells and was snap-fit into a matching shell made
by the same manufacturer in the rest. In the cases in which the insert was
secured with cement, a burr was used to create striations on the back of the
liner for better cement interdigitation. Cement was then introduced into the
acetabular component, and the liner was inserted while steady pressure was
applied until the cement set (Figs.
1-A and
1-B).
Sixty-nine acetabular components were revised: twenty-five were revised
because of loosening, twenty were revised to accommodate the constrained
liner, thirteen were revised because of infection, six were revised because of
component malpositioning with concurrent soft-tissue deficiency, four were
revised because of the failure of a bipolar prosthesis, and one was revised
because of massive osteolysis. All sixty-nine hips in which the acetabular
component was revised received an uncemented acetabular component and a
matching constrained polyethylene insert.
Direction of Prior Instability
The most likely direction of prior dislocation or instability was
determined on the basis of the history and the preoperative and/or
intraoperative radiographic and clinical examinations. The direction of
dislocation was posterior in sixty-six hips (60%), anterior in thirty (27%),
and indeterminate in fourteen (13%).
Postoperative Protection
An abduction pillow was used for immediate postoperative immobilization in
all cases. Additional immobilization in the form of an abduction brace
(sixteen hips) or a hip-spica cast (five hips) was also used at the discretion
of the treating surgeon.
Statistical Analysis
The changes in the Harris hip score were evaluated with use of the Wilcoxon
signed-rank test. Significance was determined with use of a 95% confidence
level.
Functional Evaluation
At the time of the most recent follow-up, the constrained component had
restored stability (that is, there had been no dislocation or subluxation) in
108 hips (98%). No episodes of component disassembly or recurrent dislocation
were observed. Two patients (2%) had instability of the hip in the form of
perceived sensations of subluxation. Both of these patients had an adequate
range of motion, with no clinical evidence of impingement. One of these
patients had a 22-mm femoral head, and the other had a 28-mm femoral head.
Neither patient had a so-called skirted femoral head or neck, and both
patients had an uncemented hemispherical cup. Both patients were treated
nonoperatively.
The mean Harris hip score improved significantly, from 62.7 points (range,
24 to 95 points) preoperatively to 76.4 points (range, 38 to 95 points) at the
time of the latest follow-up (p < 0.0001). The outcome was considered to be
excellent or good (defined as a hip score of >80 points, no use of walking
aids, and no pain) for seventy-eight hips, fair (with one of the above
criteria not being met) in twenty-one, and poor (with at least two of the
above criteria not being met) for eleven. Of the twenty-one hips with a fair
outcome, eighteen were in patients in whom the hip score had been adversely
affected by the presence of symptomatic arthritis affecting other joints,
including the contralateral hip (nine patients), one or both knees (six), and
one or both ankles (three). Of the eleven hips with a poor outcome, eight were
in patients who had pain and used a walking aid, two were in patients who had
pain and a low hip score, and one was in a patient who had a low hip score and
used a walking aid.
Radiographic Findings
Nonprogressive radiolucent lines around the femoral component were
identified in zones 3, 4, and 7 in eight hips and in zones 1 and 7 in another
three hips. The femoral component was noted to be in slight varus in six hips
and in slight valgus in two. No acetabular component malpositioning (defined
as 10° of variation from optimal positioning of the cup at 45° of
abduction and 10° of anteversion) was detected in any of the 110 hips.
Wear of the acetabular cup was observed in six hips at the time of the latest
follow-up, but the presence of a metallic bipolar-type mechanism inside the
uncemented metallic shell made reproducible measurement of wear difficult.
Radiolucent lines around the acetabular component were seen in a total of
fifteen hips (14%); specifically, radiolucent lines were noted in all three
zones in six hips, in zones I and II in four hips, in zones II and III in four
hips, and in zones I and III in one hip. Radiolucent lines were progressive in
ten hips (9%). Four patients with gross loosening and migration of the
acetabular cup required revision surgery, and a fifth patient with definite
loosening and migration of the cup had a revision procedure pending at the
time of the most recent follow-up. Of the fifteen hips with progressive
acetabular radiolucency, eight were in patients in whom the acetabular
component had not been revised at the time of the index operation and the
remaining seven were in patients with a newly inserted cup. Radiolucent lines
were noted at a mean of six years (range, two to seven years) after the index
operation.
Complications
There were nine intraoperative fractures, including six femoral fractures
and three acetabular fractures. One postoperative periprosthetic fracture of
the femur occurred three years after surgery and required open reduction and
internal fixation. Additional complications included superficial infection
(two hips), hematoma formation (two), sciatic nerve palsy (one), deep venous
thrombosis (three), pulmonary embolus (one), and urinary tract infection
(seven). Deep infection developed in seven hips. Six of these seven infections
occurred in patients who had had an infection at the site of a primary
prosthesis that had been previously revised with use of a two-stage exchange
method; these six infections, therefore, may have represented failure due to
reactivation of the previous infection. Of the seven infections that developed
after the index procedure, one was treated successfully with early
débridement and antibiotic suppressive therapy, four were treated
successfully with two-stage exchange arthroplasty, and two were treated with
resection arthroplasty.
Reoperations and Revisions
There was a total of nine revisions. The revisions were performed for deep
infection in six hips, aseptic loosening of the acetabular component in two,
and periprosthetic fracture in one.
Instability after total hip replacement remains a common and difficult
problem. Recurrent instability is often multifactorial and frequently is not
associated with a clearly identifiable
cause3,17,18,23,24.
This is unfortunate as the best outcomes following the surgical treatment of
instability are seen when a discrete cause of the instability can be
identified and
corrected3. The
present study was performed to evaluate a technique that was developed for the
treatment of recurrent instability after total hip arthroplasty in patients in
whom no clearly identifiable cause of instability is found, those in whom
prior surgical attempts at stabilization have failed, and those who have a
major soft-tissue deficiency about the hip.
Our use of constrained liners was extremely successful in achieving the
main goal of preventing any further instability of the hip as no subsequent
dislocations or disassemblies were seen. We also noted a concomitant
substantial improvement in hip function, which demonstrates that eliminating
instability can help to restore function for these patients. Our results are
in agreement with those reported by Goetz et
al.18, who showed
that fifty-four (96%) of fifty-six hips were stable after insertion of the
same constrained acetabular device as was used in the present study. Anderson
et al.17 reported
that stability was achieved in fifteen (71%) of twenty-one patients who were
treated with a constrained acetabular cup. Despite these successful results in
terms of the restoration of hip stability, the use of any constrained device
is associated with real concerns related to the decreased range of motion,
potential for impingement, and increased interfacial stresses that may
predispose these hips to increased risks of wear, osteolysis, and
loosening.
In this relatively short-term study (mean duration of radiographic
follow-up, 2.9 years), a high (14%) rate of radiolucent lines around the
acetabular component was seen. In the study by Goetz et
al.18, two (5%) of
the thirty-eight hips with adequate radiographic follow-up were thought to be
definitely loose and seven (18%) of the thirty-eight hips had progressive
radiolucent lines around the acetabular component. Anderson et
al.17 observed
radiolucent lines around three of nineteen cups after thirty-one months of
follow-up; however, they believed that the radiolucencies were nonprogressive
and did not require revision.
A weakness of the present study is that many of the hips that had had
multiple operations frequently had associated acetabular bone deficiencies.
Some of the radiolucencies that were observed may have been a consequence of
this factor rather than a result of any effect caused by the constrained
acetabular insert. We believe, nonetheless, that our findings support the
previously raised concerns regarding the possibility of accelerated wear and
potential loosening of a constrained acetabular component. The great advantage
that these implants provide by eliminating instability must be balanced
against these concerns. Thus, the routine use of these components for any and
all instances of recurrent dislocation should be discouraged.
In conclusion, revision hip arthroplasty with use of this constrained
acetabular component is a reasonable and reliable method for restoring
stability at the site of a complex unstable hip replacement. The technique is
particularly valuable when substantial soft-tissue deficits are present. This
approach should be considered when a correctable cause of instability cannot
be identified or when instability cannot be corrected at the time of the
reoperation. In our current practice, this operation is considered to be an
excellent salvage procedure for the unstable hip for which multiple prior
surgical attempts at stabilization have failed.