Isolated revision of an acetabular component with retention of the femoral
component can lead to specific technical challenges. In addition to the usual
complexities of acetabular revision surgery, the presence of the retained
femoral component can dictate differences in exposure, alter the choices of
head size, and limit the options available for restoration of limb length,
offset, and soft-tissue tension.
Several authors have noted a high rate of dislocation after isolated
acetabular revisions. Jones and Lachiewicz reported a dislocation rate of 18%
(twelve of sixty-eight) after isolated acetabular revisions compared with 8%
(eleven of 142) after revisions of both components (p =
0.03)1. They found
no difference between these groups with regard to survival of the acetabular
implant if a dislocation occurred. Similarly, Paprosky and Weeden reported a
19% dislocation rate (thirteen of sixty-eight) after isolated acetabular
revisions, despite the use of a constrained liner in 15% of the
hips2. Seven of the
thirteen hips that dislocated required a reoperation because of recurrent
dislocation. It should be noted that a high percentage (18%) of the index
acetabular revisions in their series were performed because of recurrent
dislocation.
We evaluated the clinical and radiographic results of a consecutive series
of isolated revisions of the acetabular component with use of the
Harris-Galante Porous acetabular component (Zimmer, Warsaw, Indiana). Our
purpose was to report the results of these revisions and to ascertain the
complication rate, particularly the dislocation rate. Furthermore, we sought
to determine if the use of a trochanteric osteotomy was related to the
dislocation rate.
Between July 1984 and November 1996, the senior author (W.H.H.) performed
ninety-five isolated acetabular revisions in ninety-three patients using the
Harris-Galante Porous I or II acetabular component. All ninety-five hips were
followed for a minimum of twenty-four months. A subset of sixty-three
acetabular revisions in sixty-one patients who had been followed for at least
sixty months was the central focus of this report. The other thirty-two hips
were excluded from the detailed analysis but were included in the survivorship
analysis (Fig. 1).
Twenty Harris-Galante Porous I and forty-three Harris-Galante Porous II
components were used in the series of sixty-three hips. The Harris-Galante
Porous I component differs from the Harris-Galante Porous II component in that
it has a thinner shell and smaller-diameter cancellous bone screws are used
(5.1 mm compared with 6.5 mm for the Harris-Galante Porous II component).
The average duration of follow-up of the sixty-three hips was 130 months
(range, sixty to 208 months). The group consisted of thirty-nine women and
twenty-two men with an average weight of 65.8 kg (range, 42 to 105 kg), an
average height of 1.67 m (range, 1.37 to 1.90 m), and an average age of 56.5
years (range, twenty-five to eighty-four years). Twenty-six hips were
classified as Charnley class A; thirty-one, as Charnley class B; and six, as
Charnley class C3.
Fifty-nine of the sixty-one patients were alive at the time of the last
follow-up for this study.
Prior to the index operation, the sixty-one patients had undergone a total
of 100 hip operations, including sixty-three primary arthroplasties
(fifty-eight with cement, three without cement, and two hybrid procedures)
(Figs. 2-A and
2-B) and nine revision
arthroplasties. The index acetabular revision was the first revision in
fifty-five hips, the second acetabular revision in seven hips, and the third
acetabular revision in one hip. The average interval between the index
acetabular revision and the previous arthroplasty was 121 months (range, two
to 214 months).
A trochanteric osteotomy was performed in fifty-seven hips (fifty-five
patients). Thirty of these procedures were performed after a previous
osteotomy of the greater trochanter. Thirty-five of the trochanteric
osteotomies were conventional osteotomies, and twenty-two were vertical in
orientation4. The
conventional osteotomy was performed from approximately 1 cm distal to the
vastus tubercle and was angled proximally and medially toward the interval
between the tendon of the gluteus minimus and the superior hip capsule. A
vertical trochanteric osteotomy was required in cases in which the anatomy had
been altered by a previous trochanteric advancement and the greater trochanter
was attached to the lateral femoral cortex. Sixteen (73%) of the vertical
osteotomies and fourteen (40%) of the conventional osteotomies were repeat
osteotomies. Of the fifty-seven trochanteric osteotomies, nine were repaired
with chromium-cobalt wires alone and forty-eight were repaired with
chromiumcobalt wires as well as chromium-cobalt
mesh5. Acetabular
screws were used in all of the acetabular revisions, with the number of screws
averaging 3.4 (range, two to seven). Thirty-five hips were reconstructed with
a so-called high hip center—i.e., the center of the head was at least
3.5 cm proximal to the interteardrop
line6. A so-called
jumbo cup, defined as a cup with an outer diameter of =65 mm, was placed in
four hips.
All sixty-three hips were evaluated with use of anteroposterior hip and
pelvic, frog-leg lateral, true lateral, and two oblique pelvic radiographs
(Judet views) at each time-period. Acetabular deficiencies were graded, on the
basis of the operative notes and the preoperative radiographs, with use of
three different systems: that described by Paprosky et
al.7, the AAOS
(American Academy of Orthopaedic Surgeons)
classification8, and
the classification described by Tanzer et
al.9. The grading of
the bone deficiencies at the time of the index operation revealed extensive
acetabular bone loss (Table I).
Thirty-three hips had preoperative osteolysis in zone I (twenty-eight linear
and five balloon lesions); thirty, in zone II (twenty-seven linear and three
balloon lesions); and twenty-four, in zone III (twenty-one linear and three
balloon lesions).
Penetration of the femoral head into the polyethylene was measured with use
of the technique of Livermore et
al.10. Only hips
for which comparable radiographs were available according to the criteria of
Russotti and Harris6
and that had had no additional revisions of the shell, femoral head, or
acetabular liner were included in this analysis.
The chi-square test was used to compare dichotomous data. The Fisher exact
test was used when indicated. Dichotomous data were stratified with use of the
Mantel-Haenszel
test11. The
significance level was set at p = 0.05. Survival analysis was performed with
the method of Kaplan and
Meier12.
Nine (9%) of the ninety-five hips with an isolated acetabular revision
required a rerevision of the shell. Five of those hips failed less than sixty
months postoperatively. Of the five hips, three underwent a repeat isolated
revision of the index acetabular component—two because of recurrent
dislocation and the third because the shell had migrated into the
pelvis—within the first three months after the operation, one hip had
recurrent subluxations and ultimately required revision of both components,
and the fifth hip presented with a dislodged polyethylene liner one year after
the index revision. Overall, of the nine shell rerevisions in the ninety-five
hips, four were due to recurrent dislocations; four, to aseptic loosening; and
one, to dissociation of the liner.
At an average of 130 months postoperatively, fifty-nine (94%) of the
sixty-three acetabular shells had been retained. Of the four rerevised shells
in this group, three had loosened and were revised at 9.3, 11.3, and 12.1
years postoperatively (Figs. 3-A,
3-B, 3-C, and 3-D).
The fourth hip had a well-fixed shell that was revised because of late
recurrent dislocation at 9.5 years. There was no difference, with the numbers
available, in the rerevision rate between the Harris-Galante Porous I and II
shells (p = 0.59, Fisher exact test). In addition to the four hips with a
rerevision of the shell, eight hips underwent exchange of the polyethylene
liner at the time of an operation on the femoral component. Thus, twelve (19%)
of the sixty-three hips that had been followed for a minimum of sixty months
underwent some form of additional surgery on the acetabular reconstruction
after the index operation.
With revision of the shell for any reason as the end point, Kaplan-Meier
analysis demonstrated a survival rate of 90.5% (95% confidence interval, 83.5%
to 97.6%) at 120 months for the entire series of ninety-five isolated
acetabular revisions. With rerevision of the shell because of aseptic
loosening as the end point, the 120-month survival rate was 96.8% (95%
confidence interval, 92.3% to 100%).
Of the fifty-one intact index acetabular components (shells in place
without a liner exchange), two were associated with balloon osteolysis at the
time of the latest follow-up, one was associated with a circumferential
radiolucency that was seen on the immediate postoperative radiographs but not
on the latest radiographs, and six were associated with a circumferential
radiolucency at the time of the latest follow-up. Comparable radiographs
suitable for assessment of shell migration and femoral head penetration were
available for thirty-three of the fifty-one hips. No component had migrated,
and the average rate of linear femoral head penetration was 0.098 mm/yr
(range, 0 to 0.29 mm/yr).
The Harris hip
score13,14
was available for forty-nine of the fifty-one hips that had not had additional
acetabular surgery. The average Harris hip score improved from 55 points
(range, 22 to 93 points) preoperatively to 81 points (range, 31 to 100 points)
at the time of the latest follow-up; this mean improvement of 26 points was
significant (p < 0.0001). Pain ratings and use of walking supports are
shown in Tables II and
III.
Thirty-seven of the fifty-seven trochanteric osteotomy sites united and
nine had fibrous union without displacement, a combined rate of 81%. There was
a fibrous union with migration of <1 cm in three additional hips. Eight
trochanteric osteotomy sites (14%) failed because of either migration or
resorption. The rate of union was 74% (twenty-six of thirty-five hips) after
the conventional trochanteric osteotomies and 50% (eleven of twenty-two hips)
after the vertical osteotomies. This difference was not significant.
Of the sixty-three hips, five (8%) dislocated postoperatively. The average
interval between the index operation and the first dislocation was
ninety-three days (range, twelve to 227 days). Four of the five hips were
successfully treated with closed reduction and a brace. The fifth hip, which
had undergone the index revision because of recurrent dislocation, continued
to dislocate postoperatively. The patient underwent an exchange of the femoral
head to a larger size and an exchange of the liner to an extended-lip liner.
Although this procedure was successful for eleven years, the dislocations
recurred, ultimately requiring a constrained liner. One other hip had repeated
episodes of subluxation; the shell eventually loosened and was rerevised at
nine years.
The rate of dislocation was 5% (three of fifty-seven) in the hips with a
trochanteric osteotomy compared with two of six in those without an osteotomy.
This difference was not significant (p = 0.07). There were no dislocations in
the seven hips with a modular femoral head-neck junction, whereas 9% (five) of
the fifty-six hips without such modularity had dislocated; again, the
difference was not significant. When the dislocation rates of the hips with
and without a trochanteric osteotomy were analyzed with the modularity of the
femoral component accounted for, a significantly higher dislocation rate was
found in the hips without a trochanteric osteotomy (p = 0.04, Mantel-Haenszel
test). Of the eight hips with either resorption or >1 cm of proximal
migration of the trochanter, one dislocated.
Other complications associated with the index operation, in addition to the
trochanteric issues and dislocations discussed above, occurred in 49%
(thirty-one) of the sixty-three hips, with a total of thirty-seven
complications (Table IV). Eight
(13%) of the sixty-three hips had a nerve palsy, with six of the palsies
resolving fully or nearly so. With the numbers available, there was no
association between the development of a nerve palsy and trochanteric
osteotomy (p = 0.57).
Acentral issue common to all acetabular revisions is obtaining stable
fixation of the acetabular shell. In our overall series of ninety-five
isolated revisions of the acetabular component, the rate of shell rerevision
for any reason was 10% (nine rerevisions) and the 120-month survival rate with
rerevision of the shell for any reason as the end point was 90.5%.
The dislocation rate in this series was 8% (eight) of the ninety-five hips
and 8% (five) of the sixty-three hips followed for a minimum of sixty months.
This rate of dislocation is lower than that in the series of isolated
acetabular component revisions reported by Jones and
Lachiewicz1 and by
Paprosky and Weeden2
(18% and 19%, respectively). The use of a trochanteric osteotomy in our series
was associated with a decreased rate of dislocation, although this difference
was not significant. However, when the effect of trochanteric osteotomy on the
dislocation rate was analyzed with the modularity of the femoral head-neck
junction accounted for, there was a significantly higher dislocation rate in
hips without an osteotomy.
The 13% rate of nerve injury in this series was higher than the rates
reported after revision total hip
arthroplasty15,16.
Possible reasons for this high rate are the increased dissection required for
exposure in isolated acetabular revisions, increased tension on the sciatic
nerve during femoral exposure, retractor positioning, or impingement on the
sciatic nerve by a monoblock femoral head.
Limitations of this study include the retrospective acquisition of data
from our patient database as well as the availability of only sixty-three of
the ninety-five hips at sixty months. The remaining hips were excluded from
the detailed analysis because of a duration of follow-up of less than sixty
months, revision at less than sixty months, or death of the patient or severe
illness.
Although isolated acetabular revision continues to pose challenges related
to bone stock, dislocation, and increased risks of neurovascular and other
complications, the findings in this series confirm the utility of the
Harris-Galante Porous acetabular component. In the group of sixty-three hips
that had been followed for a minimum of sixty months, the rate of aseptic
loosening was 5% at an average of 130 months. There were no deep infections,
and the rate of osteolysis was 4%. The dislocation rate of 8% suggests a need
for the use of additional techniques such as repair of the posterior capsule
and short external rotators, postoperative bracing, use of larger head
diameters, and use of a constrained socket in selected cases. The rate of
nerve palsies in this series was high despite the common use of a trochanteric
osteotomy. The rate of dislocation, the requirement for walking aids by some
of the patients, and the prevalence of other complications (49%) emphasize the
difficulties encountered with isolated revisions of the acetabular component
of a total hip arthroplasty.