Osteonecrosis of the femoral head is a devastating disease that often leads
to destruction of the hip joint in young adults who have the disease.
Historically, the results of standard total hip arthroplasty in patients with
osteonecrosis have been less than optimal when compared with the results in
young patients treated for other disorders, such as rheumatoid arthritis or
primary
osteoarthritis1-3.
More recently, total hip arthroplasty has been associated with better outcomes
in young adults with osteonecrosis of the femoral
head4-7;
however, most reports still show somewhat inferior results in patients who
have osteonecrosis compared with patients who have other
diagnoses8-11.
The ideal bearing surface for a standard total hip arthroplasty has not
been determined. Metal-on-polyethylene bearings remain the preferred
combination for older patients. However, osteolysis resulting from excessive
polyethylene wear debris is a common cause of implant failure in young and
active patients. Herberts and
Malchau12 analyzed
the results of the Swedish total hip replacement registry and found that
uncemented implant designs had a revision rate of 13%. Most of these revisions
were done in young patients who demonstrated early evidence of aseptic
loosening because of excessive polyethylene wear. Despite the introduction of
ultra-high molecular weight polyethylene and its improved wear
characteristics, complications related to wear debris have not been fully
eliminated, and hard bearings (metal-on-metal and aluminaon-alumina) might
help to extend the longevity of current implant
designs13.
The first use of alumina-on-alumina bearing interfaces in standard total
hip arthroplasty was reported in
197114. The
theoretical advantages of this hard-on-hard bearing include scratch resistance
and improved lubrication that contributes to low friction. Over the past
decade, the clinical use of alumina-on-alumina interfaces has gained
popularity and short-term results have demonstrated favorable survivorship
rates15-18.
D'Antonio and
colleagues19
reported the outcomes of 328 alumina-on-alumina bearings in a prospective,
randomized United States Investigational Device Exemption study with an
implant survival probability at five years of 94.1%. Thus, young and more
active patients with osteonecrosis may be excellent candidates for
alumina-on-alumina bearings as most such patients will require an implant with
an extended survivorship.
The primary purpose of this randomized clinical study was to evaluate the
safety and efficacy of total hip arthroplasty with use of an
alumina-on-alumina bearing in patients with osteonecrosis in comparison with
patients with osteoarthritis. A secondary goal of the study was to compare
these results to those of patients who had total hip arthroplasty performed
with a cobalt-chromium-on-polyethylene bearing.
Study Design and Patient Cohort
In October 1996, a United States Investigational Device Exemption
multi-institutional study was initiated as a prospective randomized clinical
trial to compare aluminaon-alumina bearings with
cobalt-chromium-on-polyethylene bearings in primary total hip arthroplasty.
Patients were randomly assigned to one of three study groups, each of which
had a different combination of acetabular components and bearing surfaces. In
September 1999, a fourth study group with a different cup design was added.
Three combinations had an alumina-on-alumina bearing surface, whereas the
remaining combination was a cobalt-chromium-on-polyethylene bearing surface.
In summary, each patient had a one-in-four chance of receiving any one of the
four acetabular cup designs and a three-in-four chance of receiving an
alumina-on-alumina bearing surface. A total of twenty-two investigators at
sixteen centers participated in this study (see Appendix). The safety of the
alumina-on-alumina bearing was assessed by documenting the nature and
incidence of postoperative complications, whereas the efficacy was assessed by
evaluating postoperative pain and functional and radiographic outcomes. The
study protocol was approved by the United States Food and Drug Administration
before initiation of the study, and informed patient consent was obtained
prior to enrollment. Each center was required to submit to their institutional
review board the study protocol with the consent form, and the protocol
approval was then renewed annually.
There were a total of 558 alumina-on-alumina bearings implanted in 512
patients in the investigational device exemption study. We identified seventy
patients (seventy-nine hips) with osteonecrosis of the femoral head who were
treated with an alumina-on-alumina bearing total hip arthroplasty. There were
fifty-four men and sixteen women with a mean age of 45.2 years (range,
twenty-one to sixty-seven years). The mean follow-up period for the
aluminaon-alumina osteonecrosis group was 4.2 years (range, 0.7 to 7.7 years).
These seventy-nine hips were directly matched to osteoarthritic hips from the
cohort of 439 hips that received alumina-on-alumina bearings during the same
time frame. The matching was done for gender, preoperative Harris hip scores,
and age at surgery. In this osteoarthritis matching group, there were
fifty-nine men and seventeen women who had a mean age of 46.5 years (range,
thirty to sixty-seven years) and a mean follow-up of 4.9 years (range, one to
eight years).
There were a total of 126 cobalt-chromium-on-polyethylene bearings
implanted in the investigational device exemption study, from which we
identified another matching subgroup of twenty-six hips (in twenty-five
patients) with a diagnosis of osteonecrosis. This subgroup included nineteen
men and six women who had a mean age of forty-four years (range, twenty-four
to seventy-five years) and who had a mean follow-up of 5.1 years (range, 0.8
to eight years). These twenty-six hips were directly compared with twenty-six
hips in patients with a diagnosis of osteoarthritis by matching for gender,
preoperative Harris hip scores, and age at surgery. In this matching group,
there were nineteen men and six women with a mean age of 44.8 years (range,
twenty-eight to seventy-four years) who had a mean follow-up of 4.1 years
(range, 0.1 to 7.2 years). A summary of patient demographics for all groups
can be found in Table I.
Implant Designs
The four cementless cup designs implanted in the study included three
alumina-on-alumina bearings (ABC System I, ABC System II, and Trident; Stryker
Orthopaedics, Mahwah, New Jersey), and one cobalt-chromium-on-polyethylene
bearing (ABC System III; Stryker Orthopaedics). The ABC System I consisted of
a porous-coated titanium shell with an alumina bearing couple. The ABC System
II had an arc-deposited titanium shell with a hydroxyapatite coating and an
alumina bearing insert. The third alumina-on-alumina system, the Trident
system, had a hydroxyapatite-coated cup with a metal sleeve backing and an
alumina-ceramic acetabular insert. The cobalt-chromium bearing system, the ABC
System III, had a titanium porous-coated acetabular shell with a polyethylene
acetabular insert against a cobalt-chromium femoral head. All patients
received a cementless Omnifit hydroxyapatite-coated femoral stem (Omnifit;
Stryker Orthopaedics).
Clinical Evaluation
The data coordinators at each participating center collected clinical and
radiographic data preoperatively, early postoperatively (four to six weeks),
at six months, at one year, and at subsequent annual evaluations. All data
coordinators and participating surgeons had training in the research protocol
to ensure consistency and reliability of data collection at all centers.
Documented data included patient demographics such as age, gender, body mass
index, and information concerning surgical approach, component size, use of
bone screws, time of follow-up, operative site-related complications,
revisions and/or component removal, preoperative bone type, and postoperative
radiographic findings. Pain, function, and deformity were evaluated with use
of a Harris hip-rating
system20 prior to
surgery and at each postoperative clinical visit. A Harris hip score of 90
points or more was defined as an excellent outcome; 80 to 89 points, a good
outcome; 70 to 79 points, a fair outcome; and less than 70 points, a poor
outcome.
Radiographic Evaluation
For radiographic evaluations, standard anteroposterior and lateral
radiographs of the pelvis were made at each clinical visit. Preoperative
radiographs were assessed to determine the bone quality of the proximal
portion of the femur in accordance with the ABC criteria described by Dorr et
al.21. The
postoperative radiographs were evaluated for radiolucencies, implant fixation,
implant migration, and osteolysis. Acetabular radiolucencies were evaluated
according to the zones described by DeLee and
Charnley22. The
femoral components were evaluated according to the zones described by Gruen et
al.23.
Radiolucencies were defined as a radiolucent zone anywhere around the
prosthesis, covering at least 30% of the zone length. All radiolucent lines
>2 mm were recorded. A cyst was defined as a scalloped erosion >2 mm in
diameter at the bone-prosthesis interface. Cortical thickening referred to an
increase in the outer diameter of the cortex secondary to hypertrophy. Bone
resorption was defined as a visible loss of trabecular bone or cortical
thinning. Heterotopic bone formation was graded with use of the Brooker
classification
system24. The
postoperative radiographs were evaluated by one of the senior authors (P.M.B.)
who did not participate in the investigative group.
Data Analysis
The data were compiled and tabulated with use of Microsoft Excel
spreadsheets (Microsoft, Redmond, Washington). All statistical analyses were
done with use of an SAS statistical software program (version 8.2; SAS
Institute, Cary, North Carolina). Descriptive statistics were calculated.
Survival probabilities from surgery to failure were determined and graphed
according to the Kaplan-Meier
method25. The end
point of survival was defined as revision (removal or exchange of one or more
components) for any reason. The log-rank test and the Wilcoxon signed-rank
test were used to compare survival probabilities between the osteonecrosis and
osteoarthritis groups. The collected Harris hip score data were summarized and
then analyzed according to continuous and categorical data. For continuous
data, a repeated-measures analysis of variance was used to characterize the
Harris hip score profile over time. This model incorporated the repeated
measures taken over the evaluation times, the groups (osteonecrosis compared
with osteoarthritis), and the interaction of evaluation time and the groups.
The categorical results of the Harris hip scores were stratified into two
groups: fair or poor results referred to Harris hip scores of <80 points,
and good or excellent results were defined as Harris hip scores of =80
points. Differences between the groups with regard to categorical data and
evaluation time were determined with use of the Fisher exact test, and
differences in continuous data were determined with use of the Student t test.
A p value of <0.05 was considered significant.
Clinical Results
The mean preoperative Harris hip scores (and standard deviation) in the
osteonecrosis and the osteoarthritis alumina-on-alumina groups were 45.8
± 12.3 points (range, 19.3 to 73.8 points) and 49.7 ± 12.3
points (range, 23.2 to 81.7 points), respectively
(Fig. 1-A). At the most recent
follow-up (minimum three years; mean 4.2 years) in the alumina-on-alumina
osteonecrosis group, the Harris hip score increased to 96.0 ± 8.8
points (range, 48.6 to 100 points). This was a mean improvement of 50.0
± 15.1 points (range, 13.3 to 80.8 points, paired t test, p <
0.0001) (Table II). For the
matched alumina-on-alumina osteoarthritis group, the Harris hip score
increased to 95.7 ± 7.8 points (range, 70.9 to 100 points) at latest
follow-up (minimum three years; mean 4.9 years). This was an average increase
of 45.6 ± 14.2 points (range, 16.1 to 76.8 points, paired t test, p
< 0.0001) (Table II). Thus,
after a comparable follow-up period, the mean improvements in Harris hip
scores were similar for alumina-on-alumina bearing couples in the
osteonecrosis and the osteoarthritis groups (50.0 points vs 45.6 points;
two-sample t test, p = 0.0884). A repeated-measures analysis demonstrated
that, with the numbers available, there was no statistical difference in
Harris hip scores over time between the groups (p = 0.2322)
(Fig. 1-A).
In the group with cobalt-chromium-on-polyethylene bearing couples, the mean
preoperative Harris hip score for the osteonecrosis and the osteoarthritis
groups was 42.2 ± 13.9 points (range, 21.4 to 77.8 points) and 48.8
± 13.3 points (range, 22.0 to 69.1 points), respectively
(Fig. 1-B). In the
osteonecrosis group, the mean Harris hip score improved to 96.4 ± 6.8
points (range, 71 to 100 points) at the time of the most recent follow-up
(minimum three years; mean 5.1 years). This was a mean improvement of 51.1
± 12.7 points (range, 22.3 to 66.5 points, two-sample t test, p <
0.0001) (Table II). Similar
results were found in the matched osteoarthritis group
(Table II). The mean Harris hip
score increased to 97.0 ± 5.2 points (range, 82.9 to 100 points) at the
time of the most recent follow-up (minimum three years; mean 4.1 years). This
was a mean 45.3-point gain (range, 20.4-77.9 points, two-sample t test, p <
0.0001). This improvement was comparable with that in the osteonecrosis group
(51.1 points versus 45.3 points; two-sample t test, p = 0.1870), and the
patients in both groups had similar clinical outcomes in terms of Harris hip
scores. With the numbers available, a repeated-measures analysis of the Harris
hip scores over time demonstrated no statistical difference between the
groups, p = 0.8536 (Table II)
(Fig. 1-B).
Implant Survival
With revision for any reason as the end point, the probability of implant
survival in the alumina-on-alumina osteonecrosis group was 95.5% (95%
confidence interval: 86.5% to 98.6%) at the time of both the five-year and the
seven-year follow-up. For the matched alumina-on-alumina osteo-arthritis
group, the survival probability was 100% at the time of the five-year
follow-up and 89.4% (95% confidence interval: 69.7 to 96.5%) at the time of
the seven-year follow-up. There was no difference in survivorship between the
groups (log-rank test: p = 0.7445, Wilcoxon test: p = 0.2032). Kaplan-Meier
survival curves for the alumina-on-alumina bearing groups are shown in
Figure 2-A.
For the cobalt-chromium-on-polyethylene groups, the survivorship at the
time of both the five-year follow-up and the seven-year follow-up for patients
with osteonecrosis was 92.3% (95% confidence interval: 72.6% to 96.0%). In the
matched osteoarthritis group, implant survival for the
cobalt-chromium-on-polyethylene bearing was 92.9% (95% confidence interval:
59.1% to 99.0%) at the time of both the five-year follow-up and the seven-year
follow-up. A comparison of the survival distributions of the two groups with
use of a log-rank test and a Wilcoxon test did not reveal any difference in
implant survivorship (p = 0.5897 and p = 0.3701, respectively, with the
numbers studied). The survival curves for the cobalt-chromium-on-polyethylene
bearings are shown in Figure
2-B.
With the numbers of patients with osteonecrosis that were studied, the
survival probability as compared at seven years between the alumina-on-alumina
bearing group and the cobalt-chromium-on-polyethylene bearing group did not
reveal a difference (log-rank test: p = 0.4004; Wilcoxon test: p = 0.3056)
(Fig. 3-A). With the numbers
studied of patients with osteoarthritis, there was also no difference at seven
years between the two bearing types (log-rank test: p = 0.6784, Wilcoxon test:
p = 0.2507) (Fig. 3-B).
Radiographic Results
At the most recent radiographic evaluation, progressive radiolucent lines
or evidence of osteolysis was absent in any zone around the femoral component
in all study groups. No acetabular cups with an alumina-on-alumina bearing
couple demonstrated a radiolucent line in any DeLee or Charnley zones.
However, there was one acetabular cup that migrated in the alumina-on-alumina
osteoarthritis group. In the cobalt-chromium-on-polyethylene bearing groups,
there was one hip that demonstrated radiolucent lines in all DeLee and
Charnley zones and one hip that showed migration of the acetabular cup.
Complications
The most common complication in the alumina-on-alumina bearing groups
included bursitis (7%), noninfected wound-related problems (6.9%),
intraoperative femoral cracks (5.7%), heterotopic bone formation (4.5%),
intraoperative ceramic insert chip (2.6%), and dislocation (2%). A statistical
analysis that was performed between the osteonecrosis group and the
osteoarthritis group did not reveal any differences in complication rates
between the two groups (two-sided Fisher exact test, p > 0.05 in all
comparisons). In the cobalt-chromium-on-polyethylene bearing groups, common
complications were superficial wound infections (5.8%), bursitis (3.9%),
soft-tissue trauma (3.9%), dislocation (3.8%), intra-operative femoral crack
(3.8%), and heterotopic bone formation (3.8%). Similar to the
alumina-on-alumina bearing groups, with the numbers studied, there were no
differences between the osteonecrosis and the osteoarthritis group with regard
to the complication rates (two-sided Fisher exact test, p > 0.05 in all
comparisons).
Revisions
Three patients (three hips, 3.8%) in the alumina-on-alumina osteonecrosis
group had one or more components revised. A fifty-five-year-old man required a
revision of all components due to excessive hip pain at three years
postoperatively. The revision was performed by a surgeon who was not
affiliated with this study, and no further information was available for this
patient. A thirty-nine-year-old woman had the femoral stem and head revised as
a result of a postoperative femoral fracture that occurred at the seventh
postoperative week. The patient underwent subsequent hardware removal two
years after the revision surgery. At three years after revision, she had a
Harris hip score of 100 points. The third revision was performed in a
fifty-two-year-old woman at twenty-two months postoperatively. The acetabular
insert and head were revised as a result of recurrent subluxations. The
patient was then lost to follow-up, and no further information was available.
Similarly, there were three patients (three hips, 3.8%) in the
alumina-onalumina osteoarthritis group who required revision surgery. A
fifty-year-old man had revision of all components as a result of sepsis at
five years postoperatively. The revision was performed by a surgeon who was
not affiliated with this study, and the patient was lost to follow-up. The
second patient was a fifty-three-year-old woman who underwent revision of the
insert and femoral head at five years after the index operation. The reason
for revision surgery was joint instability secondary to recurrent dislocations
following excision of heterotopic bone formation. At the time of the six-year
follow-up, she had a Harris hip score of 99 points. The third patient was a
forty-seven-year-old man who had the acetabular insert and head revised at
five years after the index operation due to persistent groin pain and
tendinitis that had first been identified three years after the index
procedure. As of the time of writing, the patient had not returned after the
revision surgery and was consequently listed as lost to follow-up.
In the group of patients with osteonecrosis treated with a
cobalt-chromium-on-polyethylene bearing, there were two patients (7.6%) who
required revision surgery. One patient was a seventy-five-year-old man who had
the femoral component revised two months after the index surgery as a result
of a femoral fracture after a fall. He was last seen at the time of the
seven-year follow-up and had a Harris hip score of 100 points. The second
revision was performed in a thirty-one-year-old man who had the acetabular
insert revised two days postoperatively as a result of a failed closed
reduction of a dislocation. Five years following the revision surgery, he had
a Harris hip score of 100 points. In the matched group of patients with
osteoarthritis, the one patient who required revision surgery was a
twenty-eight-year-old woman who underwent a revision of the acetabular shell,
insert, and head at forty-five months postoperatively to treat acetabular
loosening. After that, the patient was lost to follow-up.
The results of total hip arthroplasty for patients with a diagnosis of
osteonecrosis of the femoral head have not always been optimal. Revision rates
between 20% and 50% at approximately five years after surgery have been
reported for first-generation
devices3,8,24,26-29.
More recently, with use of new prosthetic designs and techniques, there have
been improved
outcomes4-7,30,31.
There is still interest in optimizing the results of modern-day total hip
arthroplasty in young patient populations.
There are only a few specific studies on alumina-onalumina hip arthroplasty
in patients with osteonecrosis. Bizot and
colleagues32
reviewed sixteen patients (twenty-seven hips) with bilateral osteonecrosis of
the femoral head after bone-marrow transplantation. The mean age at the index
surgery was thirty-one years (range, eighteen to forty-four years). In their
study population, there were no revisions for aseptic loosening or osteolysis
at a mean follow-up of fifty-nine months (range, twenty-four to 156 months).
The authors concluded that in young patients who are expected to outlive the
mean life-span of the prosthesis, total hip arthroplasty with an
alumina-on-alumina bearing surface is a safe and effective treatment
option.
Nich et al.33
reported on fifty consecutive total hip arthroplasties with an
alumina-on-alumina bearing surface performed in forty-one patients with
osteonecrosis. In all patients, a cemented femoral stem with a 32-mm alumina
head was used. At a mean follow-up time of sixteen years, sixteen hips had
been revised. With use of survivorship analysis with revision for any cause as
the end point, the authors reported a cumulative survival rate of 84.5%
(range, 73.7% to 95.2%) at ten years, and a survival rate of 65% (range, 49.7%
to 80.8%) at sixteen years. Despite the large number of failures, the authors
did not find any evidence of osteolysis, even in the hips that were followed
for as many as twenty-four years. Their high failure rate may be explained,
however, because they cemented an acetabular component that had a high
loosening rate in the first thirty-nine hips. Later in the series, they
changed to a press-fit metal-backed design. In the present study, the
cementless acetabular cup designs demonstrated excellent survivorship with
only two instances of loosening. In addition, several other studies that made
use of this same prosthesis have had low loosening rates on the acetabular
side19,34,35.
Various studies describing excellent results of aluminaon-alumina bearings
in young patients undergoing total hip arthroplasty have frequently included
patients with
osteo-necrosis36-40.
In contrast to the present study, in which we specifically analyzed patients
with osteonecrosis, it has been difficult to substratify the outcomes by
diagnosis in these studies. Fenollosa and
coworkers36
analyzed ceramic bearings in seventy-four young patients (ninety-four hips)
who had a mean age of thirty-eight years. Osteonecrosis was the most frequent
diagnosis in their series (twenty-three patients [twenty-five hips], 27%). At
a mean follow-up of 177 months, the survival for cemented prostheses was 80%,
whereas, for uncemented prostheses, the survival rate was higher at 95.74%.
Bizot et al.37
reported on sixty-two consecutive patients (seventy-one hips) who were managed
with hybrid alumina-on-alumina total hip arthroplasties and who had a mean age
of forty-six years (range, twenty-one to fifty-four years). Twenty-two hips
(31%) had a preoperative diagnosis of osteonecrosis. With revision for aseptic
loosening as the end point, the overall nine-year survival rate was 98.3%.
D'Antonio and
colleagues35
reported the results of a prospective, randomized study of 334 patients (349
hips) with alumina-on-alumina bearings. The study population had a young mean
age of fifty-three years, with osteonecrosis being the diagnosis in
fifty-three patients (fifty-six hips, 16%). At a mean follow-up of thirty-five
months, the implant survival rate was 98.8%. All of the above studies have
demonstrated similar survival rates for the alumina-on-alumina bearing as
compared with the 97.6% seven-year survival rate in the present study.
This prospective, randomized, multicenter study demonstrated an excellent
overall implant survivorship of 97.6% at seven years for the
alumina-on-alumina bearings. We are encouraged by the survival rate and the
fact that we did not detect any signs of osteolysis in this patient
population. Advances in alumina-bearing designs appear to have overcome
earlier problems with this bearing option and appear to be a safe and
effective option for young patients with osteonecrosis of the femoral
head.
A table listing the study investigators and co-investigators is available
with the electronic versions of this article, on our web site at
(go to
the article citation and click on "Supplementary Material") and on
our quarterly CD-ROM (call our subscription department, at 781-449-9780, to
order the CD-ROM). ?