This study was approved by our institutional review board, and all patients
provided informed consent.
We evaluated the cases of all patients in the original
cohort1 (140 hips in
124 patients) as well as of forty-six additional patients (forty-eight hips)
from the same surgeon's consecutive series of cases treated with acetabular
revision at least ten years prior to the time of the review. Thus, there was a
total of 170 patients (188 hips). The indications for the 188 revisions
included aseptic loosening of the acetabular component alone in eighty-nine
hips, aseptic loosening of both components in eighty-four hips, incidental
revision of the acetabular component during revision of the femoral component
in nine, recurrent dislocation in three, reconstruction following a previous
resection arthroplasty in two, and major osteolysis of the acetabulum in
one.
The polyethylene used in the Harris-Galante Porous cups varied according to
the year in which the component had been manufactured. In the early 1980s, GUR
1050 polyethylene was compression molded and sterilized with ethylene oxide.
During the late 1980s, the polyethylene was ram-extruded from GUR 4150 stock
with calcium stearate. The manufacturer later resumed using the GUR 1050
compression-molded polyethylene. In this review, we did not attempt to
identify lot numbers to determine the year of manufacture or the type of
polyethylene of each cup.
The 188 hips were followed for an average of 109 months (range, two to 201
months). Thirty-one patients (thirty-six hips) had died less than ten years
after the operation, and twenty-nine patients (thirty hips) had been lost to
or refused to return for follow-up. This left 110 patients (122 hips) who had
been followed for a minimum of ten years or had had a revision during the
minimum ten-year follow-up period, and this group is the subject of this
report. The study group of 122 hips was followed for an average of 144 months
(range, three to 201 months). When the eighteen hips in which the component
was revised were excluded, the duration of follow-up averaged 149 months
(range, 119 to 201 months).
The Harris-Galante-I or II Porous shell was used in all 122 hips. The shell
is a hemispherical titanium-alloy component with a sintered porous coating of
titanium fiber mesh. The Harris-Galante-I cup, which was used in fifty-nine
hips, has fixation holes to accommodate 5.0-mm screws and a locking mechanism
consisting of a series of tines. The Harris-Galante-II cup, which was used in
the remaining sixty-three hips, had been modified to include additional
locking tines in the larger sizes for improved fixation of the liner as well
as a thicker shell; also, the screws had been enlarged to 6.5 mm to improve
the initial fixation of the shell.
All procedures were done through a posterolateral approach. A trochanteric
osteotomy was performed in 107 hips (88%), and fifty-six (52%) of the 107
osteotomies were repeat osteotomies. A combination of vertical and horizontal
wires, with or without Vitallium mesh augmentation, was used for trochanteric
fixation. Both components were revised in seventy-one hips, whereas only the
acetabular component was revised in fifty-one. A cup of =65 mm in diameter
(a so-called jumbo
cup)2 was implanted
in four patients. A high hip center (at least 35 mm proximal to the
interteardrop
line)3,4
was used in fifty-eight hips. An acetabular structural allograft was used in
three hips, and a structural autograft was used in one. Particulate graft was
used in eighty-six hips. No cages were used. The mean outer diameter of the
acetabular component was 56 mm, and the average polyethylene thickness was 9.6
mm. All femoral heads were made of cobalt-chromium. The femoral head was 22 mm
in twenty-four hips, 26 mm in sixty-six hips, 28 mm in seventeen, and 32 mm in
twelve. The head size and wear rate were not recorded for three hips that
required early repeat revision.
The Harris hip
score5, which has
been validated6, and
radiographic analysis were used to determine the outcome. A Harris hip score
was calculated for all patients preoperatively, postoperatively, and at the
time of the latest follow-up. Patients who were unable to return for follow-up
completed a validated, self-administered
questionnaire7, from
which a Harris hip score could be calculated, and also submitted current
radiographs. Preoperative radiographs and intraoperative findings were used to
judge the extent of the acetabular defects present at the time of the
reconstruction according to the classification systems of the American Academy
of Orthopaedic
Surgeons8, Paprosky
et al.9, and Tanzer
et al.1. Cup
loosening and acetabular lysis were assessed on anteroposterior and frog-leg
lateral radiographs. Linear head penetration was measured with the method of
Livermore et
al.10.
A mean of two (range, one to five) previous hip operations had been
performed in these patients. Fifteen patients had undergone one previous
acetabular revision, three patients had undergone two, and one had undergone
three.
Interviews and examinations were carried out at the time of this review to
evaluate seventy-three of the 122 hips. The remaining forty patients
(forty-nine hips) returned questionnaires or completed them by telephone. All
110 patients (122 hips) had current radiographs available for review. There
were twenty-nine men and eighty-one women with an average age of fifty-four
years (range, twenty-five to eighty-one years), an average height of 162 cm
(range, 137 to 193 cm), and an average weight of 65.5 kg (range, 42 to 100 kg)
at the time of the index procedure. Activity ratings were not routinely
recorded.
Survivorship analyses, with revision for any reason, revision for aseptic
loosening of the acetabular component, or definite radiographic loosening of
the acetabular component as the end points, were performed with use of the
Kaplan-Meier
method11.
One hundred and four (85%) of the 122 hips followed for more than ten years
had the original revision acetabular component in place and intact at the time
of the latest follow-up. The rate of survival of the shell, with aseptic
loosening as the end point, was 98% (95% confidence limits, 0.96 to 1.0) at
ten years and 96% (95% confidence limits, 0.92 to 0.99) at twelve years. With
revision for any reason as the end point, the survival rate was 91% at ten
years and 88% at twelve years. The acetabular shell had been revised in
eighteen hips, and femoral revision with the acetabular component left intact
had been performed in seventeen other hips. An additional eleven hips had
another operation without revision of either component. These operations
included three procedures for hardware removal, three femoral bone-grafting
procedures, one trochanteric advancement, and four procedures for open
reduction and internal fixation of a femoral fracture. Two other patients had
closed reductions of dislocations.
Five (4%) of the 122 hips had a revision of the acetabular shell because of
aseptic loosening. An additional eight acetabular components had migrated, for
a total rate of aseptic loosening of 11% (thirteen of 122). Five of the 104
hips in which the index shell was in place at the time of follow-up had a
continuous radiolucency at the prosthesis-bone interface in the acetabulum,
but these shells were not considered to be loose because they caused no
symptoms and had not migrated. The most common cause for an acetabular
reoperation by twelve years was deep infection (eight of 122 hips; 7%). Three
other acetabular revisions (2%) were done because of recurrent dislocation.
Two additional well-fixed shells were revised incidentally at the time of a
femoral revision by another surgeon, who wished to implant a different total
hip system. In total, forty-six (38%) of the hips (in forty-two patients) had
a repeat revision or reoperation performed for any reason; however, only
eighteen of the revisions involved the index acetabular shell. No patient had
dislodgment of the liner.
Disarticulation was necessary to control infection in two patients, one of
whom had required a circumferential proximal femoral allograft to replace 15
cm of the proximal part of the femur during a complex reconstruction of both
the femur and the acetabulum following multiple failures of prior total hip
arthroplasties. The Harris-Galante-I Porous acetabular component was used in
this operation. The allograft became massively infected after another surgeon
attempted to add bone graft to the junction of the allograft with the host
bone, and disarticulation was ultimately required. The other patient, who had
previously had tuberculosis affecting the hip as well as a prior cup
arthroplasty and a prior revision total hip arthroplasty, required a jumbo cup
and a femoral revision at the age of eighty-three years. Deep sepsis developed
five years later and, at the age of ninety-six, she required disarticulation
because the chronic infection could not be controlled. Four other patients
with infection were treated with resection arthroplasty, and one had a staged
reimplantation.
There was no difference, with the numbers available, in the rate of aseptic
loosening (including the hips revised because of aseptic loosening of the
shell, the hips with migration of the shell, and those with a continuous
radiolucent zone) between the fifty-eight hips in which the reconstruction was
done with a high hip center and the hips in which it was done with an anatomic
hip center (p > 0.10).
Seven hips had grade-III heterotopic
ossification12, and
the one hip with grade-IV heterotopic ossification was the only one with a
limited range of motion (a 30° arc of flexion).
Eighty-one (76%) of the 107 trochanteric osteotomies were followed by
osseous union. Of the fifty-six hips that had had a prior osteotomy of the
trochanter, thirty-six had trochanteric union, four had nonunion without
migration, seven had migration, and nine had resorption of the trochanter; the
64% rate of osseous union after repeat osteotomy was significantly lower than
the rate in the hips with only one trochanteric osteotomy (p = 0.01).
Periacetabular osteolysis was identified in eleven (9%) of the 122 hips. On
the average, the largest diameter of the osteolytic lesions was 5 mm. Wear of
the acetabular liners averaged 0.11 mm/yr (range, 0.01 to 0.33 mm/yr).
The preoperative Harris hip score averaged 49 points (range, 26 to 93
points) for the 104 hips in which the index acetabular shell was in place at
the time of final follow-up. At the time of final follow-up of those 104 hips,
the Harris hip score had improved to an average of 78 points (range, 31 to 99
points). Forty-nine percent (fifty-one) of the 104 hips had a good or
excellent result (a Harris hip score of >79 points). Preoperatively 88%
(ninety-one) of the 104 hips had been moderately or severely painful, but at
the time of final follow-up 88% (ninety-one) were not or only mildly painful.
Sixty-three (65%) of the ninety-seven patients (66% [sixty-nine] of the 104
hips) used support most or all of the time preoperatively, whereas 62% of the
patients (63% [sixty-five] of the hips) used no assistive device or used a
cane occasionally for walking long distances at the time of the latest
follow-up.
The twenty-three hips with a Harris hip score of <70 points (and thus a
less than satisfactory overall result) at the time of the latest follow-up had
an average improvement of 13 points in the pain score alone. Five of these
hips were in patients who had pain and functional limitations related to
lumbar spine disease, despite having an intact hip reconstruction that
appeared to function well. Two additional hips had trochanteric nonunion,
another had trochanteric resorption, and two others had painful hardware that
contributed to a limp. Furthermore, the status of the femoral stem contributed
to the overall outcome of the hip surgery. Ten of these hips had a loose
femoral component, and eight others had undergone revision because of aseptic
loosening of the femoral component before the latest follow-up evaluation.
No significant difference between the results associated with the two cup
designs could be identified with the numbers available. Ten of the fifty-nine
hips with the Harris-Galante-I implant underwent repeat revision of the shell.
Five of these hips had an infection, two had aseptic loosening, two were
revised incidentally at the time of a femoral revision, and one had recurrent
dislocations. In comparison, eight of the sixty-three hips with the
Harris-Galante-II shell had a repeat revision, which was done because of
aseptic loosening in three of them, because of recurrent dislocations in two,
and because of infection in three.
The major finding of this study of 122 hips followed for an average of 12.5
years after an acetabular revision with the Harris-Galante Porous acetabular
component was that, in general, the procedures produced durable and successful
results. Repeat revision of the shell because of aseptic loosening was
necessary in 4% of the hips, and the total rate of aseptic loosening was 11%.
Pelvic osteolysis was identified in 9% of the hips. Eighty-eight percent of
the 104 hips in which the acetabular shell was in place at the time of
follow-up were not or only mildly painful, and 62% of the patients (63% of the
hips) used no assistive device or used a cane occasionally. Fifty-one (49%) of
the 104 hips had a good or excellent result.
In the initial
report1 on a subset
of these hips (original cohort, 140 hips in 124 patients) followed for
forty-one months, the shell fixation had failed in two hips and a complete
circumferential radiolucency was seen at the shell-bone interface in another
five.
In a series of 109 revisions with this shell design, Leopold et
al.13 reported a 2%
prevalence of radiographic evidence of aseptic loosening at an average of 10.5
years postoperatively; they performed no repeat revisions of the shell because
of aseptic loosening. One of their two patients with a loose shell was
asymptomatic, and the other was believed to be medically unfit for revision.
Similarly, Templeton et
al.14 reported
that, of thirty-two hips followed for an average of 12.9 years after
acetabular revision with the Harris-Galante Porous component because of
aseptic loosening, none had a repeat revision of the shell because of aseptic
loosening and only two had radiographic evidence of loosening of the shell.
Pelvic osteolysis was identified on plain radiographs of four hips (13%), and
the average maximum dimension of the lytic lesions was 5 mm.
This design of acetabular component is known to have two specific
limitations. First, the porous surface separated from the acetabular shell in
one hip in the present series, although without consequence. Leopold et
al.13 reported
separation of the porous mesh associated with osteolysis in 8% of the hips in
their series, which had been followed for an average of 10.5 years. However,
mesh separation has not been known to occur with any component manufactured
since improvements were made in the method of attaching the mesh to the
substrate. Second, dislodgment of the liner from the Harris-Galante Porous
socket has been reported
previously15,
although it did not occur in our series.
The major difference between the designs of the Harris-Galante Porous-I and
II components was the increase in the thickness of the shell, which allowed
the use of larger screws with larger heads. The type of locking mechanism
remained the same. We considered these design changes to be minor and found no
difference in the outcomes between the two versions. We therefore believe that
it was reasonable to combine the data on the two versions in this report.
We think that, in many revisions, screw fixation is of value to enhance the
initial stability of the implant, since bone stock is often compromised and
the intrinsic component stability provided by the press-fit technique may be
less dependable in some revisions than it is in most primary reconstructions.
Although no complications related to screw placement occurred in this series,
careful placement is
essential16,17.
Our rate of trochanteric union of 76% is lower than that reported by Bal et
al.18, who noted an
87% trochanteric union rate after 106 revisions. Also, in that earlier series,
repeat osteotomy of the trochanter did not appear to affect the rate of union,
whereas, in the present series, the union rate was significantly lower after
repeat osteotomy of the trochanter (64%; p = 0.01). One major difference
between the two series was that, in the previous
study18, bed rest
with the limb in abduction in balance suspension was maintained for three
weeks after trochanteric osteotomy, before the patient resumed walking.
Periacetabular osteolysis was identified in 9% of the hips in our study.
(It should be noted that we did not quantitate patient activity levels in this
investigation.) Osteolysis developed in 17% of the hips in the study by
Leopold et al.13.
The lower rate of lysis in our series may be a reflection of the average rate
of femoral head penetration (0.11 mm/yr). Dowd et
al.19 found a
higher rate of osteolysis and
Sochart20 found a
higher rate of loosening of the socket in hips that had an average rate of
linear penetration by the femoral head of =0.2 mm/yr.