Content uploaded by L'Hocine L.H" Yahia
Author content
All content in this area was uploaded by L'Hocine L.H" Yahia
Content may be subject to copyright.
VOL. 79-B, N
O
. 1, JANUARY 1997 135
S. Lerouge, BIng, MASc, Doctoral Candidate
L’H. Yahia, PhD, Professor and Director of the Group for Research in
Biomechanics and Biomaterials
Institute of Biomedical Engineering,
´
Ecole Polytechnique, CP Box 6079,
Succ. Centre Ville, Montreal, Canada H3C 3A7.
O. Huk, MD, Assistant Professor of Surgery
McGill University, Jewish General Hospital, 3755 Chemin de la Cˆote
Sainte-Catherine, Montreal, Canada H3T 1E2.
L. Sedel, MD, Senior Orthopaedic Surgeon
J. Witvoet, MD, Orthopaedic Surgeon in Chief
Hˆopital Lariboisi`ere, 2 rue Ambroise Par´e, 75475 Paris Cedex 10,
France.
Correspondence should be sent to Dr L’H. Yahia.
©1996 British Editorial Society of Bone and Joint Surgery
0301-620X/97/16621 $2.00
CERAMIC-CERAMIC AND METAL-POLYETHYLENE
TOTAL HIP REPLACEMENTS
COMPARISON OF PSEUDOMEMBRANES AFTER LOOSENING
S. LEROUGE, O. HUK, L’H. YAHIA, J. WITVOET, L. SEDEL
From the
´
Ecole Polytechnique, Montreal, Canada and the Hˆopital Saint-Louis, Paris, France
We made a semiquantitative study of the comparative
histology of pseudomembranes from 12 loose cemented
ceramic-ceramic and 18 metal-polyethylene total hip
replacements. We found no significant difference in
cellular reaction between the two groups, but there was
a major difference in the origin of the particulate
debris. In the metal-polyethylene group, polyethylene of
articular origin was predominant, while in the
ceramic-ceramic group the cellular reaction appeared to
be a response to zirconia ceramic particles used to
opacify cement used for fixation.
Isolation and characterisation of the debris showed
that the zirconia particles formed the greatest
proportion (76%) in ceramic-ceramic hips, while
alumina debris of articular origin formed only 12%.
Our study has indicated that aseptic loosening of
ceramic cups is not due to a response to debris
generated at the articular interface, but to mechanical
factors which lead to fragmentation of the cement.
J Bone Joint Surg [Br] 1997;79-B:135-9.
Received 15 February 1996; Accepted after revision 12 August 1996
It is now recognised that the biological response to wear
debris is one of the main mechanisms of aseptic loosening
of metal-polyethylene total hip replacements (THRs).
Study of the pseudomembranes from such cases has
shown that polyethylene particles generated by friction at
both articular and non-articular interfaces are the most
frequent component of this debris.
1,2
They are found in
considerable concentration in the periprosthetic tissues,
and are associated with intense cellular reactions. These
are characterised by the presence of stimulated macro-
phages which secrete mediators of bone resorption.
3-5
In response to the problems of polyethylene wear, Bou-
tin
6
introduced an alumina ceramic-ceramic load-bearing
couple. Clinical experience with this has shown an 89.4%
survivorship at ten years;
7
most of the failures were due to
aseptic loosening of the socket, but the role of the cellular
reaction to debris in this loosening has not been estab-
lished. In vivo the wear rate of ceramic-ceramic is 4000
times less than that of metal-polyethylene
8
and alumina
ceramic is known to be one of the most inert biomaterials.
We therefore hypothesised that the foreign-body inflam-
matory response around ceramic joints should be less
intense than that around metal-polyethylene prostheses.
Our present study was designed to provide a systematic
histological comparison of pseudomembranes from the
two joint couples which had been revised for aseptic
loosening. We also isolated and characterised the debris
from the ceramic pseudomembranes to establish para-
meters for the particles such as those already reported for
polyethylene and metal.
MATERIALS AND METHODS
From 1990 to 1994, we collected the pseudomembranes
from a series of 39 patients having revision operations for
aseptic loosening of the socket at the Hˆopital Saint-Louis
in Paris. To allow comparison between ceramic and metal-
polyethylene, we excluded the nine cementless sockets
since they were heterogeneous in design. Thirty cemented
sockets remained; 12 were ceramic-ceramic and 18 metal-
polyethylene. All the membrane specimens came from the
cup-bone interface.
The ceramic-ceramic prosthesis (Ceraver-Osteal, Rois-
sy, France) had a collared, cemented Ti4Al6V femoral
stem. The modular head and the socket were matched
pairs of dense alumina ceramic (Al
2
O
3
). The femoral
stems in the metal-polyethylene hips were of either cobalt-
chrome alloy or stainless steel. More details of the two
series are given in Table I which shows no significant
demographic variation between the groups, but two impor-
tant differences should be noted. First, all but one ceramic
THR failed by cup loosening alone, while loosening
involved both components in most of the metal-polyethyl-
ene THRs. Secondly, different cements were used to fix
the acetabular components. In the ceramic group, several
different brands of cement had been used, but all of them
contained zirconium dioxide (ZrO
2
) ceramic granules as
the radio-opacifying agent. In the metal-polyethylene
group, which had been implanted in other hospitals,
information on the type of cement used was not always
available, but in most of them barium sulphate (BaSO
4
)
had been used as the radio-opacifying agent.
Histological examination. We studied 5 m sections
stained with haematoxylin and eosin under transmitted
and polarised light. We graded each type of debris particle
and foreign-body inflammatory cell (macrophage, giant
cell) on a scale from 0 to 3 according to a modification
9
of
the method described by Willert and Semlitsch.
10
Ten
microscopic fields were examined on each of two different
slides for each specimen by a single independent observer
(SL) blinded to source and the grading allowed a semi-
quantitative comparison between ceramic-ceramic and
metal-polyethylene groups.
Debris characterisation. Debris from the ceramic pseu-
domembranes was isolated by tissue-digestion techniques
using collagenase and formic acid.
11
We then used SEM in
association with energy-dispersive X-ray analysis
(EDAX) and computer-assisted image analysis to study
the debris characteristics. The ratio of each type of debris
was calculated after the identification of 100 to 500 partic-
les per specimen, and particle-size distribution curves
were established by using the equivalent circle diameter
which extrapolates the surface area of a particle to a
circle.
We used non-parametric Spearman and Kolmogorov-
Smirnov statistical tests to study correlations and differ-
ences between groups respectively.
RESULTS
Grading of cellular reaction and wear debris. The mean
grade of macrophages, of giant cells and both cell types
together for each group is shown in Figure 1. We found no
difference in the mean grade of cellularity between
ceramic and metal-polyethylene membranes. The mean
grade of each type of debris particle for the two groups is
shown in Figure 2. In the polyethylene group of 18, we
identified polyethylene in all and metal in 16; there were
empty cement vacuoles corresponding to cement dissolved
during preparation in 13. Particles of radio-opacifying
136 S. LEROUGE, O. HUK, L’H. YAHIA, J. WITVOET, L. SEDEL
THE JOURNAL OF BONE AND JOINT SURGERY
Table I. Details of 30 patients having revision of cemented cups
Joint
Ceramic-ceramic Metal-polyethylene
Number of cemented cups revised 12 18
Mean age in years (±
SD
) 64 ± 14 61 ± 16
Male:female 1:11 5:13
Implantation time in years (±
SD
) 8.7 ± 3.4 10.3 ± 4.0
Cement radio-opacifying agent ZrO
2
BaSO
4
*
Loose component at revision
Acetabulum 11 3
Both 1 15
* in the MPE group, information on the type of cement was not always available,
but in most cases BaSO
4
was used as the radio-opacifying agent
Fig. 1
Mean grade of macrophage, giant cell, and of both
cell types in pseudomembranes collected from
ceramic-ceramic (CC) and metal-polyethylene
(MPE) THRs.
agent were found occasionally in the empty vacuoles.
Depending on the size of each particle, we saw typical
macrophage and giant-cell reactions to both metal and
polyethylene debris. Particles larger than 5 to 10 m were
usually found within giant cells. The mean grade of poly-
ethylene debris was higher than that of other types of
debris.
In the 12 ceramic-ceramic pseudomembranes, we iden-
tified particles of titanium in 7 and ceramic in 11, with
empty cement vacuoles in 10. Surprisingly, and in contrast
to the polyethylene group, the presence of the radio-
opacifying particles (ZrO
2
) in the ceramic-ceramic pseu-
domembranes were not limited to the empty cement
vacuoles. Numerous small ZrO
2
particles were found dis-
persed in the tissue, their size and colour being very
similar to those of alumina debris. Both types of ceramic
particle were yellow amber in colour with brown outlines,
and most were submicron in size both intra- and extrac-
ellularly. Light microscopy could not clearly distinguish
submicron Al
2
O
3
from submicron ZrO
2
ceramic debris,
and for histological grading, the two types of ceramic
particle were placed in one category. In five specimens,
grade-3 ceramic debris was found in association with
grade-2 or grade-3 macrophage reactions (Fig. 3). Giant
cells were very rare and associated only with cement
vacuoles.
We found no significant differences between the mean
grade of particulate debris in ceramic and polyethylene
groups.
Debris from ceramic pseudomembranes. EDAX analy-
sis allowed us to identify the debris in the ceramic-ceramic
pseudomembranes, confirming the presence of particles of
alumina ceramic (Al
2
O
3
), titanium alloy (TiAlV) and zir-
conia ceramic (ZrO
2
). These represented 12%, 12% and
76% of the total particle load, respectively. Of the ceramic
debris, 86% was ZrO
2
and only 14% Al
2
O
3
. In the five
hips showing grade-3 ceramic debris, the isolated particles
were ZrO
2
of cement origin and not Al
2
O
3
of prosthetic
origin. Size-distribution analysis showed that Al
2
O
3
and
ZrO
2
particles had overlapping size distributions, but that
the latter were on average smaller (0.28 ± 0.08 m) with a
smaller size distribution than Al
2
O
3
(0.44 ± 0.25 m) or
TiAlV (0.61 ± 0.31 m) (Fig. 4).
DISCUSSION
According to the mechanism of metal-polyethylene loos-
ening, polyethylene of articular origin migrates within the
effective joint space, and provokes an inflammatory action
leading to periprosthetic osteolysis.
12
Our results support
other studies which have identified polyethylene as the
most abundant debris generated in such joints (Shanbhag
137CERAMIC-CERAMIC AND METAL-POLYETHYLENE TOTAL HIP REPLACEMENTS
VOL. 79-B, N
O
. 1, JANUARY 1997
Fig. 2
Mean grade of each type of debris in pseudomem-
branes of loosened ceramic-ceramic (CC) and met-
al-polyethylene (MPE) THRs. Zirconia and alumina
ceramic were grouped together since they cannot be
distinguished by light microscopy.
Fig. 3
Section showing ZrO
2
submicron particles (grade 3) in a ceramic-ceramic pseudomembrane with grade-3
presence of macrophages. Note the absence of giant cells (haematoxylin and eosin, original magnification
1000).
et al 1994). In contrast, we have found that alumina wear
debris of articular origin represents a small proportion of
the particle load in ceramic-ceramic pseudomembranes.
This also agrees with the very low in vivo wear rate
measured by Dorlot et al
8
. Thus, the biological mechanism
of prosthetic loosening in metal-polyethylene THRs can-
not be applied to loosening in ceramic-ceramic THRs; a
different cause must be proposed.
The pseudomembranes from our cemented ceramic cups
were characterised by a large quantity of ZrO
2
particles
associated with a fairly intense macrophage reaction. This
predominance of ZrO
2
debris was unexpected. It was not
of prosthetic origin since the alumina ceramic used for
THRs is very pure (>99.5% alumina) and does not contain
zirconia.
The zirconia originated from the cement. It had been
added to the fixation cement as a radio-opacifying agent in
the form of 5 to 15 m particles. The ZrO
2
debris which
we identified had a much smaller diameter (0.28 ±
0.08 m), which corresponds to the size of the micro-
structural grains of the original ZrO
2
particles added to the
polymethylmethacrylate cement.
We suggest that loosening of the ceramic cups was
secondary to cement fragmentation at the bone-prosthesis
interface. This would liberate particles of ZrO
2
which then
disintegrated into submicron grains. These are small
enough to be phagocytosed by macrophages and this may
have contributed to loosening. Our five cases showing an
intense inflammatory reaction to grade-3 ZrO
2
debris indi-
cate that ZrO
2
ceramic is not inert when it is present in
large amounts.
The key event in loosening is cement fragmentation
which is related to mechanical rather than biological fac-
tors. Alumina has a very high Young’s modulus (380 GPa)
in comparison with human cortical bone (20 GPa) and
finite-element analysis has shown that this mismatch leads
to a very different pattern of stress-distribution than that
seen for a polyethylene cup or a natural joint.
13
The
relative rigidity of a ceramic socket causes stress shielding
of the adjacent acetabular bone leading to atrophy and
progressive lack of support for the cement mantle. It
seems likely that cement sandwiched between stiff
ceramic and compliant bone may then fracture.
14,15
An
additional factor is the low damping capacity of alumina
ceramic; this may also increase the risk of cement or bone
microfractures by excess transmission of impact loading.
The role of mechanical factors in the aseptic loosening of
ceramic joints is supported by our clinical observations.
16
We found better long-term results for ceramic hips in
young patients whose supporting acetabular bone is less
osteopenic and therefore less subject to microfractures.
We conclude that the cemented ceramic cups reported in
this series probably became loose as a result of cement
fragmentation at the ceramic bone interface leading to a
macrophage reaction to polymethylmethacrylate and also
to zirconia ceramic particles derived from the cement. Our
finding of small amounts of alumina debris of prosthetic
origin confirms its theoretical and reported wear proper-
ties, but this wear reduction can be used to clinical advan-
tage only when future designs focus on the reduction of
mechanical factors, such as the lack of damping properties
and the large mismatch in Young’s modulus between
ceramic and bone.
This study was supported by the Caisse R´egionale d’Assurance Maladie
d’Ile de France (CRAMIF, France) and the Canadian Orthopaedic Foun-
dation (CORE/ACORE, Canada).
No benefits in any form have been received or will be received from a
commercial party related directly or indirectly to the subject of the
article.
REFERENCES
1. Huk OL, Bansal M, Betts F, et al. Polyethylene and metal debris
generated by non-articulating surfaces of modular acetabular compo-
nents. J Bone Joint Surg [Br] 1994;76-B:568-74.
2. Shanbhag AS, Jacobs JJ, Glant TT, et al. Composition and
morphology of wear debris in failed uncemented total hip replace-
ment. J Bone Joint Surg [Br] 1994;76-B:60-7.
3. Goodman SB, Fornasier VL. Clinical and experimental studies in
the biology of aseptic loosening of joint arthroplasties and the role of
polymer particles. In: Particulate debris from medical implants.
Kenneth R St John, 1992:27-37.
4. Glant TT, Jacobs JJ, Molnar G, et al. Bone resorption activity of
particulate-stimulated macrophages. J Bone Miner Res 1993;8:
1071-9.
5. Maloney WJ, Smith RL, Schmalzried TP, et al. Isolation and
characterization of wear particles generated in patients who have had
failure of a hip arthroplasty without cement. J Bone Joint Surg [Am]
1995;77-A:1301-10.
6. Boutin P. Arthroplastie totale de la hanche par proth`ese en alumine
fritt´ee: etude experim´entale et premi`eres applications cliniques. Rev
Chir Orthop 1972;58:229-46.
7. Nizard RS, Sedel L, Christel P, et al. Ten-year survivorship of
cemented ceramic-ceramic total hip prosthesis. Clin Orthop 1992;
282:53-63.
8. Dorlot J-M, Christel P, Meunier A. Wear analysis of retrieved
alumina heads and sockets of hip prostheses. J Biomed Mater Res
1989;23(A3 Suppl):299-310.
138 S. LEROUGE, O. HUK, L’H. YAHIA, J. WITVOET, L. SEDEL
THE JOURNAL OF BONE AND JOINT SURGERY
Fig. 4
SEM micrograph of isolated ZrO
2
debris, showing their uniform size and
rounded morphology (original magnification 50 000).
9. Agins HJ, Alcock NW, Bansal M, et al. Metallic wear in failed
titanium-alloy total hip replacements: a histological and quantitative
analysis. J Bone Joint Surg [Am] 1988;70-A:347-56.
10. Willert HG, Semlitsch M. Reactions of the articular capsule to wear
products of artificial joint prostheses. J Biomed Mater Res 1977;
11:157-64.
11. Lerouge S, Huk O, Yahia L’H, Sedel L. Characterization of in vivo
wear debris from ceramic-ceramic total hip arthroplasties. J Biomed
Mater Res 1996:in press.
12. Schmalzried TP, Jasty M, Harris WH. Periprosthetic bone loss in
THA: polyethylene wear debris and the concept of the effective joint
space. J Bone Joint Surg [Am] 1992;74-A:849-63.
13. Crolet JM. L’ancrage du composant cotyloidien dans les prosth`eses
totales de hanche; simulations exploratoires. Thesis. Université de
Technologie de Compi`egne. France, 1989.
14. Boutin P, Christel P, Dorlot J-M, et al. The use of dense alumina-
alumina ceramic combination in total hip replacement. J Biomed
Mater Res 1988;22:1203-32.
15. Sedel L, Christel P, Meunier A, Bouotin P. Alumina/bone interface:
experimental and clinical data. In: Oonishi H, Aoki H, Sawai, eds.
Bioceramics 1. Proceedings of the 1st International Symposium on
Ceramics in Medicine, Kyoto, Japan, 1989:262-5.
16. Sedel L, Nizard S, Kerboull I, Witvoet J. Alumina-alumina hip
replacement in patients younger than 50 years. Clin Orthop 1994;
298:175-83.
139CERAMIC-CERAMIC AND METAL-POLYETHYLENE TOTAL HIP REPLACEMENTS
VOL. 79-B, N
O
. 1, JANUARY 1997
