The number of revision hip arthroplasties is on the
rise1, and a
particularly complex and challenging problem in such procedures is femoral
bone loss. Numerous factors may give rise to loss of femoral bone stock,
including osteolysis secondary to particle debris, stress-shielding with
adaptive bone-remodeling, prior infection, and periprosthetic
fracture2-5.
Multiple previous failed reconstructive procedures with insertion and removal
of implants also compromise the integrity of the bone
stock6,7.
The options available for dealing with severe femoral bone loss include the
use of a long cemented or press-fit stem, impaction allografting, resection
arthroplasty, allograft-prosthetic replacement, and modular
replacement3,8-12.
The effectiveness of modular prosthetic replacements (so-called
megaprostheses) has been investigated in patients with proximal femoral bone
loss secondary to neoplastic
disease13-21.
There is a paucity of reports regarding the use of a proximal femoral
replacement, particularly the modular type, for nonneoplastic
conditions22-24.
Because of concerns about possible early failure, the use of this implant is
typically reserved for elderly and sedentary patients as implantation can be
done in a more expeditious manner than other, more complex reconstructive
procedures10,22,23.
The purpose of this study was to review the outcomes of modular femoral
replacement in patients who had proximal femoral bone loss secondary to
non-neoplastic conditions.
Demographic Data
We identified forty-eight patients in whom severe proximal femoral bone
loss due to non-neoplastic conditions had been treated with modular femoral
replacement, with or without allografting, at our two institutions between
1998 and 2003. Institutional review board approval was obtained prior to
initiation of the study. During the period of the study, a
prosthesis-allograft composite was utilized in eighteen other patients.
The study cohort comprised thirty-two women and sixteen men who had a mean
age of 73.8 years (range, forty-two to ninety-seven years), a mean height of
162 cm (range, 135 to 185 cm), a mean weight of 77 kg (range, 44 to 118 kg),
and a mean body mass index of 28 kg/m2 (range, 18 to 46
kg/m2) at the time of the index procedure. Twenty-seven patients
were seen at one institution, and twenty-one were seen at the other. The
patients had undergone an average of 2.7 (range, zero to eight) previous hip
operations. One patient with no previous surgery had sustained a hip fracture
secondary to radiation necrosis. The mean duration between the primary total
hip arthroplasty and the proximal femoral replacement was 17.5 years (range,
one to thirty-seven years). The mean duration between the last surgery and the
index megaprosthesis procedure was 40.8 months (range, two to 180 months).
The indications for use of the modular megaprosthesis included a Vancouver
type-B3 periprosthetic fracture (associated with loose stem and poor proximal
bone stock)25 in
twenty patients (Figs. 1-A and
1-B, 1-C and 1-D),
reimplantation in thirteen patients who had had a previous resection
arthroplasty because of periprosthetic infection, a failed arthroplasty in
thirteen patients (Figs. 2-A
and 2-B), a nonunion of an
intertrochanteric fracture with one patient, and radiation-induced
osteonecrosis with a subtrochanteric fracture in one patient.
Follow-up
The clinical and radiographic information on all patients was retrieved
retrospectively. Harris hip scores were recorded preoperatively and at the
time of follow-up. During the period of this study, there were six deaths
unrelated to the index procedure, three of which occurred less than
twenty-four months following the index surgery. Despite our best efforts, two
patients were lost to follow-up. The remaining forty-three patients were
followed both clinically and radiographically for at least twenty-four months.
The mean duration of follow-up for those patients was 36.5 months (range,
twenty-four to seventy-nine months). Kaplan-Meier survivorship analysis was
performed on the entire cohort of forty-eight patients.
Radiographic Evaluation
Serial anteroposterior and lateral radiographs of the treated joint were
reviewed in detail. Preoperative radiographs were reviewed to assess bone
stock, the position and status of the stem, and the morphology of the
fracture. The femur was divided into seven zones, as described by Gruen et
al.26, and the
acetabulum was divided into three zones, as described by DeLee and
Charnley27, to
evaluate the location of bone loss and the presence of lucent lines. The
integration of the porous portion of the prosthesis into the residual proximal
part of the femur, when present, was assessed on the basis of the presence or
absence of radiolucent lines and so-called
spot-welding28.
Radiographs were also evaluated for the presence of heterotopic ossification,
which was graded according to the Brooker et
al.29
classification whenever present.
All periprosthetic fractures were classified as Vancouver type B3
(associated with a loose stem and severe proximal bone
deficiency)25. The
femoral bone loss in the other patients was categorized with use of the
classification system of Paprosky et
al.30. All patients
had type III-B defects (metaphyseal and diaphyseal damage with <4 cm of
diaphyseal bone available for fixation).
Surgical Data
The surgical procedure was performed through a direct lateral approach in
thirteen patients, a posterolateral approach in three, and an anterolateral
approach in twenty-seven. The mean operative time was 342 minutes (range,
sixty-seven to 420 minutes). A constrained acetabular socket liner was
implanted in fifteen patients. Intraoperative instability and/or a severely
deficient abductor mechanism constituted the indications for the use of a
constrained liner. Twenty-one patients were treated with allograft bone, to
augment host bone, along with the megaprosthesis. Strut allograft bone was
utilized in physiologically young patients. The allograft was placed against
the native bone, when present, and spanned the body of the prosthesis and the
junction where the stem was cemented into the diaphysis. The Modular
Replacement System (Stryker Orthopaedics, Mahwah, New Jersey) was used in all
patients. This generation of the megaprosthesis is modular, allowing various
resection lengths, and has a proximally porous-coated titanium portion to
promote bone ingrowth and extracortical bone-bridging. The distal part of the
prosthesis, made of cobalt-chromium, was cemented in place with use of Simplex
cement (Stryker) or Palacos cement (Biomet, Warsaw, Indiana). The femoral
canal was prepared by broaching, with preservation of the cancellous bone,
when present, for better cement interdigitation. A cement restrictor was
inserted distal to the isthmus to allow pressurization and optimization of the
cement mantle. Cement was placed in such a manner that the porous-coated
portion of the stem was placed directly and firmly against the diaphyseal bone
with no interpositioning cement.
Statistical Analysis
Descriptive statistics and the Fisher exact test were used for analysis of
categorical data. Continuous data were analyzed with descriptive statistics
and a nonpaired t test. Simple survival analysis was done with the
Kaplan-Meier method. For all other tests, p < 0.05 was considered
significant. All analyses were performed with use of SPSS software (version
13; Chicago, Illinois) and SAS software (version 9.1; Cary, North
Carolina).
Functional Outcome
The mean Harris hip
score31 for the
entire cohort significantly improved (p < 0.05) from 37.1 points (range, 15
to 61 points) preoperatively to 64.9 points (range, 13 to 91 points) at the
time of the latest follow-up. The outcome was considered excellent (a hip
score of >90 points, no use of a walking aid, and no pain in the hip) in
eight hips (19%), good (a hip score between 80 to 90 points, mild pain, and
occasional use of a walking aid) in fourteen hips (33%), fair (a hip score
between 50 to 79 points, moderate pain, and frequent use of walking aids) in
ten hips (23%), and poor (a hip score of <50 points, substantial pain,
difficulty walking even with use of an aid) in eleven (26%). Of the eleven
poor scores, six were affected by multiple joint involvement with symptomatic
degenerative joint disease. The patient with a postoperative score of 13
points had multiple joint involvement that compromised walking despite
painless function of the index hip at the time of a thirty-six-month
follow-up.
Survivorship
With revision as the end point, Kaplan-Meier analysis of the entire cohort
of forty-eight patients revealed a prosthetic survival rate of 87% at one year
and 73% at five years (Fig.
3).
Radiographic Findings
The immediate postoperative radiographs revealed the components to be well
positioned. The acetabular components were in 15° to 25° of
anteversion and 40° to 50° of abduction. The varus-valgus alignment of
the femoral components was within 5° of neutral. At the latest follow-up
evaluation, a nonprogressive radiolucency measuring <2 mm in thickness was
noted around the femoral stem in six hips and a radiolucency measuring >2
mm in thickness was noted in two. A progressive radiolucency was not seen in
any hip. A strut allograft had been utilized in twenty-one hips and was noted
to have definitely incorporated in nine of them, probably incorporated in ten,
and resorbed or not integrated in two.
Complications and Reoperations
Thirteen (30%) of the forty-three patients had a complication after the
index surgery. The major complications were instability (eight patients),
failure of the acetabular component (four), and infection (one). Of the eight
patients with instability, six required a reoperation because of dislocation
and two, who had subluxation, required no further intervention. Five of the
six dislocations occurred in hips without a constrained liner, whereas the
sixth hip had dissociation of a constrained liner. The revision was done with
a constrained liner in all six hips. The acetabular component failed in four
patients, all of whom had a constrained liner.
Ten patients had a total of eighteen additional surgical procedures. These
included revision because of instability (six cases), revision of the
acetabular component because of loosening (three), resection arthroplasty
because of a non-reconstructable failure of a constrained acetabular component
(one), and irrigation and débridement with retention of the components
in one patient with a periprosthetic infection. The remaining reoperations
consisted of six closed reductions and one open reduction of a
dislocation.
Despite recent advances in device manufacturing and surgical techniques,
the management of proximal femoral bone loss continues to be challenging.
Proximal femoral replacement and use of an allograft-prosthesis composite are
two valuable options for treatment of these patients. Although revision
arthroplasty with use of a modular fluted stem may also have a role in the
treatment of this condition, the long-term results of such reconstruction are
not yet
available32. An
allograft-prosthesis composite potentially increases bone stock in the
proximal part of the femur and provides sites for soft-tissue attachment,
including the abductor
muscles23,24,33,34.
A proximal femoral replacement has the advantage of probably being more
available to most surgeons than a whole femur or a proximal femoral allograft,
and it is relatively less technically demanding to implant.
A previous study on the outcomes of the use of allograft-prosthesis
composites demonstrated moderate improvement in function and an excellent
success rate at an average of nine years
postoperatively35.
The use of an allograft-prosthesis composite is not, however, without its
problems. Infection, junctional nonunion, dislocation, and aseptic loosening
are some of the reported
complications33,36.
In a previous comparative study, the use of an allograft-prosthesis composite
was found to be marginally better than proximal femoral replacement for
reconstruction following tumor
resection13.
On the basis of available reports on the use of proximal femoral
replacement for non-neoplastic
conditions10,13,17,24,
the most common complication seems to be dislocation, which occurred in three
of eighteen hips in one
series13 and in two
of four in
another17. In
another retrospective study, at a mean of eleven years following thirty-three
revision hip arthroplasties with a prosthetic femoral replacement in
thirty-two patients with non-neoplastic conditions, four femoral components
and seven acetabular components had been revised because of aseptic
loosening24. With
revision as the end point, the overall prosthetic survival rate in the
aforementioned series was 64% at twelve years.
Despite its effectiveness in restoring function, proximal femoral
replacement can result in a relatively large number of complications. The most
important and common complication in our study was instability, which was
encountered in eight patients (19%). This instability rate is higher than that
usually seen after revision arthroplasty with use of conventional
prostheses37,38,
but it is comparable with that following the use of an allograft-prosthesis
composite35.
We believe that the etiology of instability following proximal femoral
replacement is multifactorial. First, and perhaps foremost, the reason for the
higher instability rate may be related to the poor quality of the soft tissues
(the abductor mechanism in particular) that is frequently seen in patients who
have undergone multiple previous operations. Moreover, as a result of proximal
bone loss and the inability to achieve a secure repair of the residual soft
tissues to the metal prosthesis, these patients are predisposed to
instability39,40.
Additionally, most of the patients undergoing proximal femoral replacement,
including those in our study, are relatively old and have other comorbidities
that can also adversely influence the prevalence of instability.
There have been a number of changes in recent years that may positively
influence the general outcome and, in particular, the prevalence of
instability following proximal femoral replacement. First, the introduction of
modular components has provided surgeons with a better ability to restore limb
length and appropriate soft-tissue tension, as was the case in all patients in
our study. Unlike the first generation of megaprostheses, which had a
monolithic stem designed to be fixed with
cement24, the
modular proximal femoral replacement permits more flexibility in determining
the resection length of the bone. Furthermore, the new generation of implants
with a porous-coated proximal surface promotes osseointegration and an
improved surface for soft-tissue reattachment. Hence, allograft bone placed
against the porous surface has the potential for incorporation and
osseointegration23.
Strut allograft was used in nearly one-half of the patients in this study,
with excellent incorporation.
Another advance in the prevention of instability after complex revision
arthroplasties is related to the availability of constrained acetabular
liners. Constrained acetabular components have a high success rate in the
treatment of recurrent
instability41. The
constrained liner can be either snap-fit or cemented into the shell, depending
on the type of acetabular
component24.
However, there can be problems with the use of constrained liners, such as
aseptic loosening of the acetabular component and catastrophic failure of the
socket in patients with poor bone stock; this occurred in one patient in our
series. With the availability of larger femoral heads, constrained liners may
be used less frequently.
We have implemented a number of strategies that we believe are likely to
have a positive influence on the outcomes of proximal femoral replacement.
First, we maintain a low threshold for the use of a constrained liner in
patients in whom intraoperative stability of the proximal femoral replacement
cannot be obtained with certainty. We defer the decision regarding the type of
acetabular liner to be used until we have completed the reconstruction of the
femur and obtained an impression about the stability of the
hip21. The
acetabulum is first exposed and examined carefully at the beginning of the
operation. If a previous acetabular component is appropriately positioned and
is stable, it is left in place and the liner is exchanged after femoral
reconstruction. If no previous component is in place or the fixation and/or
position is found to be suboptimal, a new component is usually inserted in a
press-fit manner with screw
fixation9. In an
effort to minimize instability and postoperative limping, the proximal femoral
bone, however poor in quality, and the soft tissue attached to it are always
retained at all
costs9,19.
Structural bone graft to enhance bone stock and provide more surface for
soft-tissue
attachment42 is
also used whenever appropriate.
We believe that a prerequisite for the success of a proximal femoral
replacement is that the length of the distal part of the femur be adequate to
obtain secure fixation of the femoral
stem24. When distal
bone is severely deficient, the use of total femoral replacement may be
considered9.
Successful results also require scrupulous attention to detail in preoperative
planning. Most patients undergoing reconstruction with a megaprosthesis have
had multiple previous procedures. Therefore, problems with removal of existing
implants, specific requirements for acetabular reconstruction, and the
possibility of a low-grade infection should be anticipated and addressed
appropriately. Preoperative templating to select the appropriate stem length
and diameter is also essential.
This study had some shortcomings. Perhaps the most important one is its
retrospective design with the inherent deficiency of variability in data
collection. Additionally, the relative short clinical and radiographic
follow-up, the nonconsecutive nature of the cohort, and the varied
preoperative diagnoses limit the conclusions that can be drawn from this
study. The study was intentionally limited to patients treated with a modular
prosthesis that has been in recent use. Despite the aforementioned
limitations, we found the megaprosthesis to be a valuable resource in the
armamentarium of the reconstructive hip surgeon who treats patients with
extensive bone loss for whom other available reconstructive procedures cannot
be utilized. ?