To The Editor:
I read with great interest the article "Osteolysis in Association
with a Total Hip Arthroplasty with Ceramic Bearing Surfaces" (80-A:
1459-1468, Oct. 1998), by Yoon et al. Although this is not the first
article to demonstrate the potentially reactive nature of ceramic
particles in and around total hip replacements, it represents the first
thorough examination of a series of patients with a ceramic-on-ceramic
articulation demonstrating such a high degree of osteolysis.
My main issue is with the editorial board of The Journal for
accepting this paper in its current form, as it does not place proper
perspective on ceramic-on-ceramic articulations. There is no question that
the scientific work presented here is outstanding and that if ceramic
particles reach, or are available in, the proper size, they too
can activate the immune system and generate the inflammatory response
leading to osteolysis, which is so commonly attributed to polyethylene.
The problem is that this study is based on a device that could not
obtain Food and Drug Administration approval in its current form
and that was abandoned in the United States nearly a decade ago
because of its high rate of failure. Not only does the acetabular
component migrate frequently but that migration, coupled with a
skirted ceramic ball head in many circumstances, frequently leads
to ceramic-on-ceramic impingement. This ceramic-on-ceramic impingement
was unique to this system. It led to the generation of ceramic debris
at the point of impingement. Many of those particles found their way
into the articulation where severe body wear was the end result
and, I might add, not an unexpected result given the circumstances. Once
in the articulation, small ceramic fragments can convert this stable
articulation into a debris-production machine. The authors note that
osteolysis of the pelvis did not occur in patients without acetabular
migration. Although different types of compounds elicit immunological
responses in varying degrees, particle size may be the most important
variable. Any foreign material - ceramic, plastic, or metal - that
is of the proper size will create osteolysis in a substantial percentage
of patients. This type of osteolysis has not been reported in association with
ceramic prosthetic components that have more favorable design characteristics14, but it has been reported when ceramic
particles, whether they are from an articulating surface or are
used as an opacifying agent in cement, are present in the proper
size and in sufficient numbers9.
In the Discussion, the authors suggest that the diagnosis of
avascular necrosis was a reason for the early failures and cite
several articles from over a decade ago, when inferior techniques (first-generation
cementing techniques) were used and age-matched controls were scant.
Although they cite one recent paper, by Piston et al.15, another, more modern series of
total hip arthroplasties performed with cement has suggested that
patients with avascular necrosis can expect a rate of survival that
is similar to that of age-matched controls with osteoarthritis5. The failures in the series of Yoon
et al. were overwhelmingly due to component design, not diagnosis.
I am disappointed in The Journal's editorial
staff for choosing to publish this article, given the known poor
track record of this device. I would say that although the results
are interesting, they are simply not relevant to any of the ceramic
systems currently being developed for use in this country, in which
ceramic-on-ceramic impingement is not possible. I am disappointed
that the authors did not go beyond the clinical relevance relative
to a decade ago and point out how many of the design shortcomings
of the prostheses contributed to the osteolysis in their study as
well as how they have been addressed in more modern designs. For
these reasons, there is a strong danger that the readership of The
Journal will fail to appreciate these design flaws and will
attribute the debris production and subsequent osteolysis to the
ceramic-on-ceramic articulation alone.
Jonathan P. Garino, M.D.
Department of Orthopaedic Surgery
Hospital of the University of Pennsylvania
3400 Spruce Street, Silverstein Two
Philadelphia, Pennsylvania 19104-4283
To The Editor:
We were very interested in the article "Osteolysis in Association
with a Total Hip Arthroplasty with Ceramic Bearing Surfaces" (80-A:
1459-1468, Oct. 1998), by Yoon et al. As we implant ceramic-on-ceramic
material every day in young and active people with hip disease,
we are very aware of the details and conclusions that the authors
describe.
Their study contained some excellent material, but we also believe
that it included many overstatements and much speculation. They
described a series of 169 total hips implanted in 143 patients between
1983 and 1989. By the time of the study, the Mittelmeier prosthesis, Autophor,
had already been shown in Europe to be associated with too many
failures related to mechanical aseptic loosening of both the cementless
socket and the stem. These failures were attributed in part to impingement
of the neck of the prosthesis on the ceramic and to the threaded
shape of the cups inserted without cement. The fact that Yoon et
al. observed many instances of socket loosening and many radiolucent
lines at this level could easily have been predicted. Of the 169
hips, only 103 were followed for a five-year period. The authors
did not present their material regarding survivorship analysis,
which considers patients who drop out and allows comparison with
other studies. All of the results were presented in percentages,
which is far from acceptable when so many cases had dropped out.
The authors considered radiolucent lines or bone-implant demarcation
as osteolysis. We suggest that this is an overstatement because
the majority of the cases that were not revised were not examined
histologically. They showed slow migration of an alumina cementless
socket surrounded by a fibrocytic material without any sign of osteoclastic
activity, which typically defines an osteolytic lesion. In effect,
what they described in the few histological sections that were examined
was more of a fibrocytic reaction than an aggressive foreign-body
reaction. No osteoclastic activity was recorded, nor were trap-sensitive
cells quantified. Some macrophagic activity and a few giant cells
should not be considered as very aggressive. It could also have been
related, as we have already demonstrated, to metallic debris resulting
from alumina-stem impingement.
The stem of the prosthesis was made of a porous-coated cobalt-chromium
alloy. But in contrast to others4 who
used metal-on-polyethylene sliding material, Yoon et al. recorded
no osteolysis in the diaphysis. Only metaphyseal bone destruction
was noted, and it could have been related to stress-shielding or
eventually to a high quantity of debris related to the mushroom-shaped
ceramic head, as shown in Figs. 3-A and 3-B (page 1462), which was
abandoned by Mittelmeier many decades ago. It is not surprising
that some expansile osteolysis was noticed at the diaphyseal level
in only two cases, both of which had this head configuration. We
believe that only these two cases could be called osteolysis. The
others were radiolucent lines in the acetabulum or at the proximal
level of the femur.
Finally, the study of ten acetabular membranes obtained after
mechanical failure of the screw-in design and three membranes obtained
from proximal tissues surrounding the calcar region demonstrated
that considerable ceramic debris was found in conjunction with metallic
debris. This is not a surprise, and some years ago many investigators
demonstrated that the surrounding tissues removed at revision operations
involving ceramic hips contained some amount of alumina debris9. What was also demonstrated by us
and many other investigators was that the biological responses to
this debris were relatively benign and did not give rise to osteolysis
except in some cases in which this debris was massively produced,
as in these three samples. Moreover, in clinical practice, many
authors from different parts of the world have made the same conclusions2. Failures of the ceramic components
were mechanical, and osteolysis was encountered only in very specific
cases. In our series, for example, a socket had been loose for one
decade with titanium-on-ceramic impingement. A massive foreign-body
reaction was encountered in this case. In Yoon et al.'s study of
the retrieved material, we do not know how long the implants had
been loose. It is not a surprise that a mushroom-shaped head design
and a screw-in cup resulted in early loosening. This combination
of implant designs often creates some degree of impingement and
gives rise to accelerated wear of the ceramic components due to
high contact stresses. All of this information has already been
very well documented. In addition, it has also been demonstrated
that stable, well fixed ceramic material does not give rise to osteolysis.
This is not the case with other sliding materials16.
Overall, the article appears to be a negative endorsement of
ceramic-on-ceramic material. We suggest that the authors were aware
of this because their conclusions are similar to ours: the etiology
of acetabular component loosening was mechanical, and no osteolysis
was encountered at the diaphyseal level. Then why did they focus on
linear osteolysis, which we consider to be an overstatement of the
diagnosis of osteolysis? The reader must realize that even if some
exceptional cases of osteolysis were described in conjunction with
a ceramic-on-ceramic material, this pales in comparison with the
great number described with metal-on-ultra high molecular weight
polyethylene.
Alumina-on-alumina sliding components remain, as far as we know,
the best answer for active and young people because of the material's remarkable
biological tolerance at very long-term follow-up10,17.
It is not a magic material, and surgical technique, component orientation,
and design play a major role and are perhaps less forgiving than those
associated with other implanted materials. This might have been
the conclusion of their paper. We believe that the title would have
been acceptable if it had been something like "Another Study of
a Mittelmeier-Type Prosthesis."
Laurent Sedel, M.D.
Remy Nizard, M.D.
Pascal Bizot, M.D.
Corresponding author: Laurent Sedel, M.D.
H󯨴al Lariboisi籥
2, rue Ambroise-Parç??75475 Paris, Cedex 10, France
To The Editor:
Yoon et al., in their article "Osteolysis in Association with
a Total Hip Arthroplasty with Ceramic Bearing Surfaces" (80-A: 1459-1468, Oct.
1998), report on periprosthetic osteolysis in stable as well as
in unstable uncemented ceramic-on-ceramic hip prostheses and suggest
that their study supports the prevalent dogma of wear-particle-induced
osteolysis. Although their study contains interesting information,
their theoretical framework may be deceptive. Some confusion seems
to depend on the fact that, for years, loose prosthetic components
may appear to be stable on conventional radiography.
It seems justified to call attention to an alternative explanation12,13, in which prosthetic loosening
is postulated to be initiated at an early stage, either by insufficient
initial fixation or by early loss of fixation. The initially slight
loosening of the prosthetic component, resulting from either cause,
is then influenced by the degree of stress during normal human activity,
which is variable in different patients (because of differing body
weight and level of physical activity) and for different components
(because of differing prosthetic design, positioning, friction,
and wear). Micromovements of the slightly loose prosthetic component
cause joint fluid (and wear particles) to be pumped in and out of
the bone-implant interface (where the smallest wear particles would penetrate
farthest to the leading edge of bone resorption). The resulting
pressure waves (up to 200 millimeters of mercury1)
may devitalize a layer of bone1,18.
The devitalized bone is resorbed. (Indeed, macrophages and giant
cells in the interface membrane around failed cemented prostheses
are found mainly in areas adjacent to the bone surface as opposed
to the cement surface8.) The bone
resorption may expand locally and cause focal osteolysis. The ejected
joint fluid will be partially resorbed, and the inspissated wear
debris may be invaded by granulation tissue. This explains the high
concentration of wear debris found in these granulomatous lesions6.
It is interesting to note that an analogous mechanism has been
described for the development of subchondral bone cysts in osteoarthritis7, in which joint fluid and pieces
of cartilage are forced through a small fissure into the subchondral
bone, forming a bone cyst. Granulation tissue subsequently invades
the cyst - without participation of any foreign wear particles.
Indeed, the periprosthetic accumulation of the wear debris in the
tissue around loose prosthetic components may essentially be an
epiphenomenon.
Bengt Mjöberg, M.D., Ph.D.
Department of Orthopedics
Uppsala University Hospital
SE-751 85 Uppsala, Sweden
T. R. Yoon, S. M. Rowe, S. T. Jung, K. J. Seon, and W.
J. Maloney reply:
We are responding to the three letters concerning our article.
As Dr. Garino noted, the implant system that was used and reported
on in this study was abandoned in the United States approximately
one decade ago. However, The Journal of Bone and Joint Surgery is
an international journal, and this implant system is still in use
in some parts of the world. It was used in South Korea until a few
years ago and has now been abandoned there as well. Although the
majority of published results with this implant system are poor,
this is not universally true11.
In addition, the purpose of this study was not to report a survivorship
analysis with this implant system.
The purpose of this study was, as Dr. Garino noted, to report
on the histological response associated with osteolysis secondary
to ceramic wear particles. This system provided us with a unique
opportunity to do so since no polyethylene and no bone cement were
used in these implants. In addition, the vast majority of the extensively
porous-coated stems were stable. Energy-dispersive analysis of radiographs
demonstrated little or no metallic debris, and thus we were comfortable
attributing the periprosthetic osteolysis to the ceramic debris.
The purpose of this study was not to impugn ceramic articulations.
We agree, as stated in our Discussion, that the high rate of aseptic
loosening of the socket was related to the design. We also agree
that the diagnosis, young age, and high activity level of these
particular patients, although contributory to the high failure rate, were
not the primary reasons for failure.
Finally, Dr. Garino states that ceramic-on-ceramic systems in
which ceramic-on-ceramic impingement is not possible are currently
being developed in the United States. There is no doubt that both
the manufacturing techniques and the implant design of ceramic components are
improving. However, surgical implantation with respect to the angle
of inclination, edge-loading, and neck-socket impingement (metal on
ceramic) will continue to make the use of such devices somewhat
more technically demanding. In addition, only long-term results
will prove to the orthopaedic community that these design changes
represent an improvement.
With regard to the comments of Professor Sedel et al., we have
already addressed some of them above. Again, the purpose of our
study was to report on the radiographic and biological response
to a failed ceramic implant, not to present a survivorship analysis,
which has already been done by several other authors. As such, the
drop-out rate in this study is meaningless.
Professor Sedel et al. dispute our definition of osteolysis.
On that point, we are quite comfortable. The pattern of osteolysis
(linear compared with expansile) is determined by a variety of factors,
the most important of which is the bone-remodeling pattern around
a given implant system. This then determines access and the path
of least resistance for joint fluid and wear particles from the
articulation. Although the bias of Professor Sedel et al. toward
what is and what is not osteolysis is widely held in orthopaedic
surgery, we believe that it is incorrect. In all revised cases,
there was granulation tissue and histologically demonstrated foreign-body
reaction in the regions described radiographically as being osteolytic.
They are incorrect in stating that we described "a fibrocytic material
without any sign of osteoclastic activity." Furthermore, in the samples
examined with energy-dispersive x-ray analysis, metallic debris
was not significant. Histologically, metallic debris was noted in
one out of three femoral membranes and zero out of ten acetabular
membranes. We also disagree that "some macrophagic activity and
a few giant cells should not be considered as very aggressive."
This statement is contrary to a large body of work that has been
done over the past decade. In the absence of wear particles, tissue
macrophages are uncommon at the implant-bone interface. Their presence
is synonymous with the presence of foreign-body particles. The presence
of giant cells is more related to particle size and is not a factor
in determining whether osteolysis will or will not develop. In these
specimens, it was rare to find ceramic particles greater than ten
micrometers in size, which we feel explains the paucity of giant
cells seen in this tissue.
The metaphyseal bone destruction here was not related to stress-shielding
but to particle-induced bone resorption. As an aside, it should
be noted that even though there were some suggestions that stress-shielding
makes the femur more prone to osteolysis, it is not universally
true3. However, when osteolysis
does develop with extensively and circumferentially coated femoral
components, it is almost always limited to the metaphysis, provided
that the implant is stabilized by bone ingrowth.
Again, we agree that design certainly played a role in the failure
rate of this particular implant. Professor Sedel et al., in their
letter, noted that the mushroom-shaped head design was likely contributory.
However, only three of the implants reported in this series had
that particular design.
Finally, we appreciate the extensive experience of Professor
Sedel et al. with ceramic articulations and respect their opinion
when they state that "alumina-on-alumina sliding components remain,
as far as we know, the best answer for active and young people."
They may well be right; however, it is not currently the most widely
practiced technique.
Dr. Mjöberg suggests a potential alternative explanation for
the development of osteolysis in these patients. His concept is
suggested by the title of his recent article12,
"Theories of Wear and Loosening in Hip Prostheses. Wear-Induced
Loosening vs. Loosening-Induced Wear - a Review." We agree that
both loosening and wear are probably playing a role, as it is our
opinion that, with this socket design and fixation surface, osseointegration
is not likely to occur in a high percentage of cases. When an implant
is not osseointegrated, motion at the implant-bone interface with
physiological load can result in the production of wear debris.
This in turn can lead to periprosthetic bone resorption that, along
with motion-induced bone resorption, can lead to further destabilization
of the implant. We do not feel that this mechanism played an important
role in the osteolysis in the proximal part of the femur, as most
of the femoral components were stable.
The histology of pressure-induced osteolysis is quite different
from that of particle-induced osteolysis. With pressure-induced
osteolysis, there is bone resorption in association with necrosis. In
contrast, particle-induced osteolysis is characterized by a foreign-body
granuloma with activated osteoclasts. Bone necrosis is not a prominent
feature. We believe that these represent two different mechanisms
that have similar end points but are pathophysiologically quite different.
In terms of the mechanism for the development of a subchondral
bone cyst in osteoarthritis, we would make two points. First, the
role of fluid pressure in osteoclast activation in bone resorption
is a hypothesis and not a proven fact. Second, it is important to
note that with osteoarthritis there is likely to be cartilage debris in
the joint. Cartilage debris is a potent stimulator of the inflammatory
pathway and can of itself lead to secretion of bone-resorbing enzymes and
cytokines.
We would like to express our thanks again to Dr. Garino, Professor
Sedel et al., and Dr. Mjöberg for their interest and comments.
Taek Rim Yoon, M.D.
Sung Man Rowe, M.D.
Sung Taek Jung, M.D.
Kwang Jin Seon, M.D.
William J. Maloney, M.D.
Corresponding author: Taek Rim Yoon, M.D.
Department of Orthopaedics
Chonnam University Hospital
8 Hakdong
Kwangju 501-757, Korea
E-mail address: tryoon@chonnam.ac.kr
Editor's note:
In response to Dr. Garino's comments impugning the editors of
The Journal for publishing this manuscript, I believe that Dr. Yoon's
comments, as well as some of the points stated in his own letter,
are appropriate. First of all, as Dr. Yoon correctly points out,
The Journal of Bone and Joint Surgery is an international journal,
and this implant is still in use in some parts of the world. More
importantly, there is valuable information contained and conveyed
in this work; as Dr. Garino acknowledged, "there is no question
that the scientific work presented here is outstanding." Further,
The Journal's editorial staff does not choose to publish articles
based on the good or poor track record of a particular device. Rather,
articles are chosen on the basis of scientific merit through a rigorous
process of peer review, and, indeed, I wish to assure Dr. Garino
that this was the case with this manuscript. Lastly, I believe that
the readership of The Journal is enlightened enough to appreciate
the design features of this particular device and their relationship
to debris production. As Dr. Yoon succinctly points out in his reply,
the purpose of this study was to report on the histological response, "not
to impugn ceramic articulations."
Charles R. Clark, M.D.
Deputy Editor for Adult Reconstruction
The Journal of Bone and Joint Surgery
20 Pickering Street
Needham, Massachusetts 02492-3157