Abstract
Background: The mechanisms leading to aseptic loosening of a total
hip replacement are not fully understood. A fibrous tissue interface can be
present around the implant. Hypothetically, component micromovements can
compress this interface and cause increased fluid pressure according to
biphasic models. We tested the hypothesis that compression of a fibrous
membrane with or without the presence of high-density polyethylene particles
leads to bone degradation.
Methods: A titanium implant was inserted in forty-five rabbit
tibiae, and, after osseous integration was achieved, a fibrous tissue
interface was generated. The animals were randomized to undergo a sham
operation, treatment with compression of the fibrous membrane, treatment with
high-density polyethylene particles, or treatment with both compression and
particles. Morphometric analysis of the surrounding bone was performed on
cryostat sections after Giemsa staining and staining of tartrate-resistant
acid phosphatase activity.
Results: Forty specimens were available for analysis; five tibiae
with an infection were excluded. After nine weeks, the controls showed vital
bone, whereas the specimens treated with compression showed necrosis of bone
and replacement of bone by cartilage in a discontinuous layer (p < 0.05 for
both) but not fibrous tissue. Treatment with high-density polyethylene
particles caused replacement of bone by fibrous tissue (p < 0.05) but not
necrosis or cartilage formation. Compression combined with the presence of
high-density polyethylene particles caused bone necrosis and loss of bone with
replacement by cartilage and fibrous tissue (p < 0.05).
Conclusions: In this in vivo study in rabbits, fibrous membrane
compression led to bone necrosis and cartilage formation, possibly because of
fluid pressure or fluid flow, whereas the presence of high-density
polyethylene particles led to the loss of bone with replacement of bone by
fibrous tissue. Cartilage formation may be a protective response to fluid
pressure and/or fluid flow. Fibrous membrane compression may play an important
role in the early stages of loosening of a total hip replacement.
Clinical Relevance: The findings of this study suggest that
implantation techniques that prevent the formation of a fibrous tissue
interface (which may act as source of fluid pressure and/or fluid flow) may be
beneficial in reducing implant loosening.
The most frequent, and often devastating, long-term complication
after total hip replacement is aseptic loosening. Despite vigorous research,
the precise mechanisms leading to aseptic loosening are not fully
understood.
First, insufficient initial fixation or early loss of fixation has been
suggested to lead to early
loosening1. This may
be caused by an implant of improper size or design, an improper cementing
technique, or inferior bone
quality2,3
inducing micromovements and subsequent detachment of the component at the
interface.
Second, wear particles (in particular polyethylene) originating from
components4 and/or
cement have been considered to cause bone
resorption3. This
particle-induced bone resorption has been reproduced in in vitro
studies5-7,
but the results of in vivo studies have not been consistent. In some in vivo
studies, particle-induced osteoclastic bone resorption has been
observed8,9,
whereas in other studies only a decrease in bone formation was
seen10,11.
Third, high fluid pressure has been proposed to be, at least in part,
responsible for the process of
loosening12.
Pressures as high as 0.1 MPa (780 mm Hg) have been measured in the pseudojoint
cavity of total hip replacements during physiological
activities13. High
intracapsular pressures are often found in loose total hip
replacements13,14,
causing capsular distension as identified by
ultrasound14.
Capsular distension was less in hips that did not show clinical loosening,
indicating that pressure was lower in those
cases14. Peak
pressures of 0.07 MPa (500 mm Hg) have been measured in the presence of pelvic
osteolysis at the time of revision
surgery15. Pressure
differences may induce a flow of joint fluid in the effective joint space,
affecting periprosthetic
bone3,16.
In experiments on rabbits, an exogenously derived (oscillating) fluid
pressure was shown to cause bone
resorption17,18.
Furthermore, Van der Vis et
al.19 were the
first, as far as we know, to apply endogenous fluid pressure to a fibrous
membrane in rabbits and they found bone resorption. Similar results have been
obtained in rats20.
In a study in which polymethylmethacrylate particles and endogenously derived
fluid pressure were administered to rats, the fluid pressure appeared to cause
distinct bone resorption, whereas polymethylmethacrylate led to minimal
resorption
only21.
Radiostereometric analysis showed that the probability of late clinical
loosening is increased when prosthetic components migrate soon (within two
years) after
implantation22,23.
Migration of a prosthetic component should be possible only when the component
is surrounded by a membrane of fibrous tissue. Consequently, there are
correlations between the presence of a fibrous membrane, migration of a
prosthetic component, and late clinical
loosening1,12.
It has been speculated that the aforementioned fibrous membrane acts as an
interstitial fluid
compartment12,19.
It has been postulated that micromovements of the prosthesis cause repetitive
compression of this membrane and thereby increase local fluid pressure that
may affect the integrity of the surrounding bone. Resorption and subsequent
loosening may be the result. Indeed, compression of a fibrous tissue membrane
induced osteoclastic bone resorption in recent experiments on rabbits and
rats19,20.
In the present experimental study with rabbits, an endogenously derived
fibrous tissue membrane interposed between the vital tibia and a titanium
surface was compressed. Such a membrane can be thought of as a biphasic model
consisting of a solid matrix and an interstitial fluid compartment with
physiologic properties. In this study, it was hypothesized that compression
would lead to a combination of stress in the solid matrix and elevation of
fluid pressure in the fluid
compartment24,25.
Furthermore, the effect of the presence of high-density polyethylene particles
was also studied, especially as to whether their presence might lead to
modulation of the response to interface compression.
Animals
Forty-five skeletally mature New Zealand White rabbits (BMI,
Helmond, The Netherlands), with a mean weight (and standard deviation) of
38.20 ± 0.19 N, were used. The protocol for the animal experiments was
approved by the animal ethical committee of the Faculty of Medicine,
University of Amsterdam and all animal handling was performed according to
Dutch laws for the treatment of research animals.
The Implant
The model used in this study is a modification of the model introduced by
Van der Vis et
al.19. It consists
of a cubical titanium implant that is inserted into the rabbit tibia. A
fibrous tissue interface between the implant and vital bone is created, and
the effect of compression of this fibrous membrane on bone is evaluated. Our
modification allows the possibility of administering particles at the
interface.
The implant is a cube (7.2 mm3) with a cylindrical canal
(diameter, 4.50 mm) and a groove (7.2 × 2 × 2 mm) on one end. It
is made of commercially pure titanium with a surface roughness of 1.8 µm. A
nonmoveable "static" cylinder with a diameter of 4.50 mm and a
groove of 4.45 × 2 × 2 mm fits in the canal of the implant
(Fig. 1, A).
When the device is implanted in bone (see section entitled Surgical
Procedures), the roof of the groove faces the cortical bone surface and
both sides of the groove are embedded into the bone. In this way, a bone
bridge that is 7.2 mm long, 2 mm wide, and 2 mm high is created, and it is
surrounded by the titanium implant on all sides except for the endosteal
surface toward the bone marrow cavity. In this study, the construct was
allowed to integrate into bone for five weeks. Then, the "static"
cylinder was exchanged for a "dynamic" cylinder
(Fig. 1, B), which had
a groove on one end of 4.45 × 2 × 2.2 mm; thus, its width was 200
µm in excess of the width of the groove of the "static"
cylinder. The "dynamic" cylinder also had a biconcave handle on
top that could be rotated to each side with a maximum amplitude of 100 µm.
A stop screw prevented further rotation. Thus, a space of 100 µm, in which
particles could be administered, was created between vital bone and the
titanium surface at either side of the bone bridge (see section entitled
Surgical Procedures). During the following two weeks, a
100-µm-thick fibrous layer was allowed to form in this space at either side
of the standardized bone bridge in the presence or absence of particles.
After the two weeks of fibrous tissue growth, the biconcave handle of the
"dynamic" cylinder was grasped through the intact skin and
manually rotated alternately clockwise and counterclockwise, thus rotating the
cylinder inside the implant, intermittently compressing the fibrous membrane
on both sides of the bone bridge (see section entitled Test Phase).
By means of this movement with controlled amplitude, only the fibrous membrane
was compressed without direct mechanical contact with the underlying bone.
Particulate Materials
The high-density polyethylene particles that we used were donated by Smith
and Nephew Richards (Memphis, Tennessee) and were produced by Shamrock
Technologies (Newark, New Jersey). They were small enough (mean diameter, 4.6
µm; range, 0.4 to 8.0 µm) to be phagocytosed by
macrophages26. The
particles were polymerized as high-density polyethylene and then were ground
with a proprietary milling process into a smaller particle size. They were
reported to be 100% high-density polyethylene particles and highly
crystalline, with a specific gravity of 0.95. The size of the particles was
measured with use of a scanning electron microscope interfaced with a
morphometric image analysis system (Beckman Coulter, Fullerton, California).
The particles were washed two times in alcohol followed by two additional
washes in sterile water, and, after centrifugation, the supernatant was
removed and the particles were air-dried in a sterile environment. The
sterility of the particles was verified in anaerobic and aerobic cultures.
Surgical Procedures
All animals had two operations under aseptic conditions. Anesthesia was
induced with 10 mg/kg of xylazine (Bayer, Leverkusen, Germany) and 50 mg/kg of
ketamine (Aesculaap, Boxtel, The Netherlands) intramuscularly and was
maintained with inhalation anesthesia with a mixture of isoflurane, nitrous
oxide, and oxygen. Antibiotic prophylaxis with use of 10 mg/kg of enrofloxacin
(Bayer) administered subcutaneously was started one day before the operation
and was continued until one day after the operation.
In each animal, the operations were performed on the right tibia. In the
first operation, a medial parapatellar approach was used to expose the
metaphyseal bone surface of the tibia. Onto this relatively flat area, a metal
template with two longitudinal slots was fixed with two cortical screws
(diameter, 1.45 mm; Mathys, Bettlach, Switzerland). The dimensions of the
slots corresponded with the bars next to the groove of the implant. Cortical
bone was removed with a watercooled burr through the open slots. In that way,
a bone bridge was created with the dimensions of the groove of the implant.
Next, the template was removed and the implant was inserted over the bone
bridge and was fixed with two reinserted cortical screws with use of the same
holes as the ones used for the template. Then the "static"
cylinder was inserted through the cylindrical canal and was fixed to the
implant with a screw. Finally, the skin was closed over the entire device with
interrupted mattress sutures (Vicryl 3-0; Ethicon, Norderstedt, Germany). The
animals received 0.05 mg/kg of buprenorphine subcutaneously twice a day for
two days as postoperative analgesia.
As noted previously, the implants were allowed to integrate into bone for
five weeks. During this period, the animals did not seem to be hampered by the
implant.
The animals were randomized into five groups of nine animals each. Group I
consisted of the interface controls (to identify fibrous membrane formation);
group II, the controls that received no treatment; group III, those treated
with compression only; group IV, those that received particles only; and group
V, those treated with compression and particles.
After five weeks, a second operation was performed on all animals. First,
the screw was removed and then the "static" cylinder, with great
care, was lifted in a vertical direction only, avoiding rotational movement
that could damage the bone bridge. Then, the "dynamic" cylinder
with the wider groove was inserted into the cylindrical canal of the implant,
thus creating the space of 100 µm on either side of the standardized vital
bone bridge and allowing the formation of a 100-µm-thick fibrous membrane
during the following two weeks. In animals in group IV (particles only) and
group V (compression and particles), approximately 2.5 mg of high-density
polyethylene particles (the equivalent of 0.5 × 108
particles) were administered at either side of the bone bridge. This load of
particles was chosen to ensure that the bone tissue was exposed to an adequate
number of particles, in light of the findings in a previous report that
described the loads of particulate wear debris found in periprosthetic tissue
retrieved from loosened total hip
replacements27. A
stop screw secured the cylinder to the implant, restricting the cylinder from
rotating.
Test Phase
The application of membrane compression by movement in group III
(compression only) and group V (compression and particles) was started at
seven weeks after the first operation. The biconcave handle that protruded
underneath the intact skin was grasped between thumb and index finger, and the
rabbit leg was held in place with the other hand. The handle was rotated until
firm resistance was encountered and the stop screw prevented further rotation;
then the direction of rotation was reversed. Movement was applied twice a day,
sixty times during two minutes, for fourteen days
(Table I). During the
application of movement, sedation of the animals was not necessary.
Processing of Specimens
The animals in group I (the interface controls) were killed with an
overdose of pentobarbital (60 mg/kg) at seven weeks after the first operation,
and the animals in the other groups were killed in the same manner at nine
weeks.
The implants in the animals in group III (compression only) and group V
(compression and particles), which had undergone movement, were checked to
determine whether the cylinder had rotated properly. After removal of all soft
tissues that covered the implant, the implant was cleaned of any bone
overgrowth and the "dynamic" cylinder was removed. Next, the
implant was carefully separated from the bone with the use of a chisel.
Finally, the entire proximal metaphysis was removed with an oscillating
saw.
The specimens were immediately embedded in 8% gelatin white (Sigma, St.
Louis, Missouri) and were slowly frozen in liquid
nitrogen28.
Undecalcified cryostat sections (8 µm thick) were cut parallel to the
surface of the bone bridges with the use of adhesive tape (Scotch Tape 800;
3M, St. Paul, Minnesota) to maintain the integrity of the
sections28. The
sections were cut with the use of a tungsten-carbide-tipped knife (Spikker,
Zevenaar, The Netherlands) and were stained with Giemsa (Merck, Darmstadt,
Germany) for morphological orientation. Then, the pieces of tape to which the
sections were adherent were cut out and mounted between two layers of glycerol
jelly28. For each
bone bridge, three sections at different levels were selected for analysis:
one was as close to the periosteal surface of the bone bridge as possible;
another, as close to the endosteal surface as possible; and one, at
approximately the middle of the cortex. Adjacent to the latter section, an
additional section was used for localization of tartrate-resistant acid
phosphatase (TRAP) activity to stain osteoclasts selectively and to establish
the presence or absence of necrotic bone. Necrotic bone was characterized by
the absence of cells containing TRAP activity in (bone) lacunae.
Spatially calibrated digital images were obtained by scanning an entire
section with a 35-mm slide scanner (Coolscan 1000; Nikon, Tokyo, Japan).
Interactively, tissue components, i.e., fibrous tissue, cartilage, and areas
with absent TRAP activity, were delineated and segmented with use of the
image-processing program
Object-Image29.
Morphometric Analysis
All nine bone bridges from each of the five treatment groups were examined.
TRAP activity was classified as present when cells stained brightly red and as
absent when staining was hardly detectable or when cells did not stain at all.
The regions of the bone bridges that contained cells without activity were
qualitatively determined and labeled as necrotic. Thereafter, these areas were
determined quantitatively with use of image analysis. Areas of cartilage and
fibrous tissue were determined quantitatively with use of image analysis of
the Giemsa-stained sections. Areas of necrosis, cartilage, and fibrous tissue
were expressed as a percentage of the surface area of each bone bridge. The
surface area was defined as the area of the bone bridge that had been
subjected to the effect of membrane compression and/or particles, i.e., the
area enclosed by both sides of the groove of the "dynamic"
cylinder. Similar to the shape of the entire bone bridge, the shape of the
surface area was rectangular. The median surface area for the bone bridges
that were tested was 8.91 mm2 (range, 8.85 to 9.04 mm2),
corresponding with the length (4.45 mm) and the width (2 mm) of the groove of
the "static" cylinder.
The median values of the percentage of the tissue areas were calculated.
The median values between groups were compared and analyzed. The nonparametric
Mann-Whitney test (release 11.0; SPSS, Chicago, Illinois) was used for
statistical analysis; the level of significance was set at p < 0.05.
Morphology
After five weeks, when the "static" cylinder was
exchanged during the second operation, macroscopic inspection of the inner
surface of the groove of the "static" cylinder never showed
adhesion of (fibrous) tissue.
At the time that the animals were killed and the implants were removed,
forty of the forty-five specimens showed macroscopic apposition of bone around
the implant. After removal of this tissue, it appeared that the area between
the outer surface of the implant and the metaphyseal bone was tightly sealed
and there were no signs of inflammation. Bone-marrow cavities were sealed off
from the bone bridge and interface area by the formation of new bone on the
groove side of the implant, thereby providing a closed system in all cases
(see Fig. 1, B,
inset). Sections of these bone bridges showed vital bone, and the bridge edges
were sharp and regular. These observations indicated that osseointegration had
occurred and inflammation was absent. Therefore, these forty specimens were
included in the study (Table
II). The other five animals had an abscess develop over the
implant because of a bacterial infection, as proved on culture, and these
specimens were excluded.
All specimens in group I (the interface controls) showed a vital bone
bridge as described above, with a thin fibrous membrane on both sides
containing cells and extracellular matrix. Occasionally, cartilage was
present. The thickness of the fibrous membranes was 100 µm, indicating that
the space created by the broader groove at each side of the bone bridge had
been completely filled with a membrane of fibrous tissue
(Fig. 2, A).
Group II (controls) showed vital bone in the bone bridges as well, but,
instead of an outer fibrous membrane, thin lamellae of new bone with
osteocytes and areas of osteoid, with a thickness of 100 µm, were observed
to have formed (Fig. 2,
B).
Only small amounts of fibrous tissue within the bone bridge were observed
in group III, in which the fibrous membrane had been compressed. Conversely,
larger areas of bone loss, with replacement by cartilage, were seen
(Fig. 2, C). These
changes were most obvious at all four corners of the rectangular surface areas
where the amplitude of the movement of the "dynamic" cylinder was
maximal and compression of the fibrous layer had been highest because of the
design of the implant (Fig. 2,
C).
Underneath the fibrous membrane, again especially at all four corners of
the rectangular surface areas where membrane compression was highest, there
were regions of bone tissue, including bone lacunae, that lacked cells with
TRAP activity, indicating that these regions of bone tissue were not vital. In
the areas of bone bridges where compression was lowest (in the middle of the
rectangular surface areas), the numbers of cells showing TRAP activity in bone
and lacunae were normal, indicating that bone was vital
(Fig. 2, D). This
pattern was consistent in all sections investigated. Areas of vital and
nonvital bone were also sharply demarcated in all sections.
After membrane compression, sections of bone taken at the level of the
endosteal surface demonstrated the formation of new lamellar bone.
Morphometric analysis showed that bone bridges were qualitatively thicker in
compressed specimens than in noncompressed specimens (approximately 1 mm
compared with 0.5 mm) because of this new bone formation. However, the new
bone formation occurred outside the area of analysis and therefore was not
included in the study.
In group IV, in which high-density polyethylene particles had been
introduced, fibrous tissue was not only present in the 100-µm-thick space
on both sides of the bone bridges but it also replaced bone more centrally in
the bone bridges. Fibrous tissue was more or less evenly distributed along the
edges of the bone bridges, and it contained particles. Cartilage was sparsely
present (Fig. 2, E).
The presence of high-density polyethylene particles was confirmed by their
typical birefringent appearance. Particles were present as individual
entities, in clusters, or intracellularly. Giant cells were not observed
(Fig. 2, F).
Qualitatively, the appearance of the bone tissue was normal, and TRAP activity
and numbers of osteoclasts were present in amounts similar to that observed in
the bone tissue in group I (the interface controls) and group II (controls).
TRAP activity around lacunae filled with high-density polyethylene particles
was not qualitatively increased compared with group I and group II. Necrotic
areas of bone lacking TRAP activity were hardly present
(Fig. 2, G).
After the combined administration of compression and high-density
polyethylene particles in group V, both cartilage and fibrous tissue were
observed to have formed in the bone bridge. High-density polyethylene
particles were present in a similar way as in group IV (particles only). Giant
cells again were not present (Fig. 2,
H). Large areas of the bone bridges lacked TRAP activity,
indicating necrosis. In these areas, we never observed particles. When lacunae
were filled with particles, cells surrounding lacunae showed normal TRAP
activity (Fig. 2,
I).
Histomorphometry
As noted above, membrane compression in the absence of particles in group
III (compression only) caused bone necrosis and loss of bone with replacement
by cartilage. The areas of necrosis and cartilage were a median of 14% and 3%,
respectively, of the surface area of the bone bridges
(Table II,
Fig. 3). The addition of
high-density polyethylene particles without membrane compression in group IV
(particles only) led to loss of bone with replacement by fibrous tissue.
Fibrous tissue was found on a median of 6% of the surface area of the bone
bridges. The combined application of compression and particles in group V
induced necrosis, loss of bone with replacement by cartilage, and formation of
fibrous tissue with median surface areas of 15%, 3%, and 4%, respectively.
Compared with groups I and II (the controls), these changes were significant
(p < 0.05). When we compared the effect of the combined treatment with the
effect of membrane compression alone or high-density polyethylene particles
alone on the formation of necrosis, cartilage, or fibrous tissue, the changes
were not significant.
Our model is intended to resemble the clinical situation of a
prosthesis surrounded by a thin fibrous-tissue membrane producing
micromovements upon weight-bearing, thereby compressing the fibrous tissue
membrane. When components undergo early migration, the bone-prosthesis
interface is unstable and interposition of fibrous tissue must be present.
Successive component micromovements with compression of this fibrous tissue
interface may generate a fluid pressure or fluid flow, leading to necrosis and
subsequent loss of underlying bone. This may further impair prosthesis
fixation and may signify the onset of clinical loosening. This proposed
pathological mechanism corresponds with radiostereometric analysis studies
that have shown that early migrating prostheses have a higher prevalence of
loosening22,23.
Fluid pressure leading to loss of bone has been hypothesized by Landells as
early as 195330.
Indeed, recent animal experiments have shown that fluid pressure, either
exogenously applied or through compression of a fibrous membrane, induces bone
resorption17-21,31
and
necrosis17,19,20.
On the basis of in vitro
studies5-7,
high-density polyethylene particles and polymethylmethacrylate particles have
been shown to induce osteoclastic bone resorption by mediating inflammatory
reactions. However, some in vivo studies have described a decrease in bone
formation due to the presence of particles, without an increase in
resorption10,11
or even loss of
bone21. Therefore,
it is still not clear whether and which particles cause loss of bone and, if
so, whether this is through increased resorption by increased osteoclast
activity or through decreased bone formation by inhibition of osteoblasts, or
both.
In the present in vivo study in rabbits with a vital bone-prosthesis
interface, we studied the effect of compression of a fibrous tissue interface,
with or without high-density polyethylene particles, on bone and whether the
combination of both stimuli has a synergistic effect on bone loss.
High-density polyethylene particles were used because acetabular components of
almost all total hip replacements, cemented and uncemented alike, are made of
polyethylene and therefore seem to be the predominant particle type involved
in the loosening process.
In all specimens in group I (the interface controls that had five weeks for
healing after the first operation followed by two weeks without compression or
particles), vital bone and a fibrous membrane of 100 µm on both sides of
the bone bridge were present. In specimens in group II (the controls that had
five weeks for healing followed by four weeks without compression or
particles), the fibrous interfaces were replaced by thin lamellae of new bone
containing osteocytes and osteoid, thereby proving the vitality of the bone
bridges and the tendency for bone formation under stable circumstances.
Compression of the fibrous membrane interposed between the implant and
vital bone, by movement of the "dynamic" cylinder, led to areas of
necrotic bone in the bone bridges with empty lacunae, indicating early stages
of
necrosis32,33.
This fibrous membrane consisted of cells and proteins and interstitial body
fluid with physiologic properties. If it is assumed that the behavior of such
a fibrous membrane is biphasic, compression led to a combination of stress in
the solid matrix and high fluid pressure in the fluid
compartment24,25.
Necrotic areas were situated underneath the fibrous tissue membranes at all
four corners of the rectangular surface areas of the bone bridges, whereas
bone tissue in the middle of the bone bridges was not altered. Since the
amplitude of the movement of the cylinder was at its maximum at the corners of
the rectangular surface areas, membrane compression and the resulting fluid
pressure or fluid flow was likely maximum in these areas. This may explain why
the greatest changes to the underlying bone occurred at the corners.
Van der Vis et
al.19, using a
similar model, compressed a fibrous membrane interposed between the implant
and vital rabbit bone with the same amount of pressure and for the same
duration and found necrosis as well. They discussed the possibility that
pressurization of interstitial fluid leads to lethal disruption of the
canalicular processes of osteocytes. Moreover, induction of an interstitial
fluid flow may affect the interstitial balance in the extracellular matrix of
bone with subsequent osteocyte
death34.
In the present experiment, newly formed bone always sealed off the implant
and fibrous interface area from the medullary space; thus, we assume that this
was a closed system. Furthermore, the implant we used was similar to the
implant used by Van der Vis et
al.17,18,
which has been shown to be watertight. Although our model can be considered as
a closed system, reduction of the closed volume cannot be determined. We were
thus unable to measure the pressure applied. We propose that by rotating the
"dynamic" cylinder, the tissue fluid shifts lead to fluid flow as
well as to a temporary local increase in fluid pressure. Therefore, the
observed effects after membrane compression are likely to be due, at least in
part, to the effects of fluid flow or fluid pressure, although the exact
volume ratio between the solid matrix and fluid compartment in this membrane
is not known. The advantage of the current model over the model of Van der Vis
et
al.17,18
is that the compression of the fibrous interface better resembles the clinical
situation (compression of a periprosthetic membrane during
weight-bearing).
Compression alone led to cartilage formation. This cartilage formation was
observed mainly at the four corners of the surface areas. Mesenchymal cells
can differentiate into cartilage or fibrocartilage cells when hydrostatic
pressure or hypoxia are
applied35. In
vitro, chondrocytes produce matrix when hydrostatic pressures of 3 MPa are
applied36. In an
experiment with a rat bone chamber with application of hydrostatic stress of 2
MPa, chondrocytes were formed, usually in combination with
necrosis37.
Interestingly, necrosis was absent when chondrocytes had formed a continuous
layer in between the site where the pressure had originated and the underlying
bone37. Those
authors hypothesized that cartilage formation was a protective response of
bone to fluid flow and/or fluid pressure and prevented necrosis. According to
this hypothesis, the finding of necrosis in the compressed specimens in our
experiment was to be expected because in none of these specimens was cartilage
generated as a continuous layer; thus, it could not act as a protective
barrier.
The presence of high-density polyethylene particles led to bone loss and
replacement by fibrous tissue with no necrosis or cartilage formation. In
other in vivo studies, high-density polyethylene particles led to loss of bone
because of both decreased formation and increased resorption of
bone38 or because
of increased resorption of bone
only9, and bone
resorption was associated with increased numbers of
osteoclasts9,38.
In our experiment, bone loss in the presence of high-density polyethylene
particles was not associated with an increase in the numbers of osteoclasts.
However, loss of bone can be the result of increased functional activity of
osteoclasts without an increased number of
cells39. In
addition, particles can suppress the function of
osteoblasts40 and
inhibit the proliferation and
differentiation41
of osteoblasts. Therefore, loss of bone could have been caused by increased
osteoclastic activity or by decreased osteoblastic activity, or both. However,
fibroblasts exposed to particles respond with proliferation, possibly
explaining the abundance of fibroblasts and the subsequent formation of
fibrous tissue after particles were added in our experiment.
Polymethylmethacrylate particles are more potent in causing bone resorption
than is high-density polyethylene, at least in
vitro42. However,
Skoglund and
Aspenberg21 found
formation of bone after applying polymethylmethacrylate particles to a rat
bone surface. They hypothesized that the particles had been inactivated by
opsonization. Another explanation may be that the particle size in that study
ranged between 5 and 10 µm, which is at the upper limit for macrophage
resorption26,
whereas the particles that we used were smaller (mean size, 4.8 µm),
causing a stronger cellular
response43.
The combination of membrane compression and high-density polyethylene
particles did not lead to a significant increase in the amounts of cartilage,
necrosis, and fibrous tissue compared with the effect of one stimulus only.
This indicates that compression of the fibrous membrane leads to necrosis and
cartilage but not fibrous tissue formation, and treatment with high-density
polyethylene particles leads to the formation of fibrous tissue but not to
necrosis or cartilage. In other words, the different stimuli induced different
effects on bone, and therefore a synergistic effect between both compression
and high-density polyethylene particles was not observed.
Skoglund and
Aspenberg21 found
bone resorption after combined compression of a fibrous membrane with the
administration of polymethylmethacrylate particles, but resorption was not
greater than when pressure alone was applied. A synergism between pressure and
particles on bone degradation could not be concluded.
Particle release is a process that takes
time3. By adding a
surplus of particles, we attempted to simulate a situation that is present
only after a total hip replacement has been in situ for long period. Several
retrieval studies have shown results that are similar to our findings:
interface tissue retrieved from clinically stable but migrating prostheses
showed extensive cartilage formation in the presence of a fibrous tissue
interface
membrane44,45.
Necrosis has been observed in large amounts in fibrous membranes from total
hip replacements retrieved because of aseptic
loosening46-48,
as well as in a series of retrieved interfaces that were being analyzed in our
department at the time of writing (unpublished data).
The results of the present study suggest that compression of a
periprosthetic fibrous interface, which induces fluid pressure and/or fluid
flow, can be an important cause of early loosening of total hip replacements.
Our results also suggest that polyethylene particles are involved in the
loosening process as well, but probably only in later stages, when substantial
amounts of these particles have been formed in the joint
space3. The
formation of cartilage may be a protective response of bone to the necrotic
effect of fluid pressure and/or fluid flow, but further studies are needed to
clarify this observation. ?
In support of their research, the authors received noncommercial funds from
the Stichting Wetenschappelijk Onderzoek Orthopaedische Chirurgie and the
Stichting ter Bevordering Wetenschappelijk Onderzoek Orthopaedie en
Traumatologie Ziekenhuis Hilversum. They did not receive payments or other
benefits or a commitment or agreement to provide such benefits from a
commercial entity. No commercial entity paid or directed, or agreed to pay or
direct, any benefits to any research fund, foundation, educational
institution, or other charitable or nonprofit organization with which the
authors are affiliated or associated.
Mjoberg B. Theories of wear and
loosening in hip prostheses. Wear-induced loosening vs. loosening-induced
wear—a review. Acta Orthop Scand.1994;65:
361-71.65361Â
1994Â
[PubMed][CrossRef] Â
Mjoberg B. The theory of early loosening
of hip prostheses. Orthopedics.1997;20:
1169-75.201169Â
1997Â
[PubMed] Â
Schmalzried TP, Callaghan JJ. Wear in
total hip and knee replacements. J Bone Joint Surg Am.1999;81:
115-36.81115Â
1999Â
[PubMed] Â
Hirakawa K, Bauer TW, Stulberg BN, Wilde
AH, Secic M. Characterization and comparison of wear debris from failed total
hip implants of different types. J Bone Joint Surg Am.1996;78:
1235-43.781235Â
1996Â
[PubMed] Â
Murray DW, Rushton N. Macrophages
stimulate bone resorption when they phagocytose particles. J Bone Joint
Surg Br.1990;72:
988-92.72988Â
1990Â
Â
Wang W, Ferguson DJ, Quinn JM, Simpson
AH, Athanasou NA. Biomaterial particle phagocytosis by bone-resorbing
osteoclasts. J Bone Joint Surg Br.1997;79:
849-56.79849Â
1997Â
[PubMed][CrossRef] Â
Glant TT, Jacobs JJ, Molnar G, Shanbhag
AS, Valyon M, Galante JO. Bone resorption activity of particulate-stimulated
macrophages. J Bone Miner Res.1993;8:
1071-9.81071Â
1993Â
[PubMed][CrossRef] Â
Howie DW, Manthey B, Hay S,
Vernon-Roberts B. The synovial response to intra-articular injection in rats
of polyethylene wear particles. Clin Orthop Relat Res. 1993;292:352-7.
Erratum in: Clin Orthop.1994;298:
249.298249Â
1994Â
Â
Kim KJ, Kobayashi Y, Itoh T. Osteolysis
model with continuous infusion of polyethylene particles. Clin Orthop
Relat Res.1998;352:
46-52.35246Â
1998Â
Â
Aspenberg P, Herbertsson P.
Periprosthetic bone resorption. Particles versus movement. J Bone Joint
Surg Br.1996;78:
641-6.78641Â
1996Â
Â
Van Der Vis HM, Marti RK, Tigchelaar W,
Schuller HM, Van Noorden CJ. Benign cellular responses in rats to different
wear particles in intra-articular and intramedullary environments. J
Bone Joint Surg Br.1997;79:
837-43.79837Â
1997Â
[CrossRef] Â
Aspenberg P, Van der Vis H. Migration,
particles, and fluid pressure. A discussion of causes of prosthetic loosening.
Clin Orthop Relat Res.1998;352:
75-80.35275Â
1998Â
[PubMed] Â
Hendrix RW, Wixson RL, Rana NA, Rogers
LF. Arthrography after total hip arthroplasty: a modified technique used in
the diagnosis of pain. Radiology.1983;148:
647-52.148647Â
1983Â
[PubMed] Â
Robertsson O, Wingstrand H, Kesteris U,
Jonsson K, Onnerfalt R. Intracapsular pressure and loosening of hip
prostheses. Preoperative measurements in 18 hips. Acta Orthop
Scand.1997;68:
231-4.68231Â
1997Â
[CrossRef] Â
Schmalzried TP, Guttmann D, Grecula M,
Amstutz HC. The relationship between the design, position, and articular wear
of acetabular components inserted without cement and the development of pelvic
osteolysis. J Bone Joint Surg Am.1994;76:
677-88.76677Â
1994Â
[PubMed] Â
Schmalzried TP, Jasty M, Harris WH.
Periprosthetic bone loss in total hip arthroplasty. Polyethylene wear debris
and the concept of the effective joint space. J Bone Joint Surg
Am.1992;74:
849-63.74849Â
1992Â
Â
Van der Vis HM, Aspenberg P, de Kleine
R, Tigchelaar W, van Noorden CJ. Short periods of oscillating fluid pressure
directed at a titanium-bone interface in rabbits lead to bone lysis.
Acta Orthop Scand.1998;69:
5-10.695Â
1998Â
[PubMed][CrossRef] Â
Van der Vis HM, Aspenberg P, Marti RK,
Tigchelaar W, Van Noorden CJ. Fluid pressure causes bone resorption in a
rabbit model of prosthetic loosening. Clin Orthop Relat Res.1998;350:
201-8.350201Â
1998Â
[PubMed] Â
Van der Vis HM, Aspenberg P, Tigchelaar
W, Van Noorden CJ. Mechanical compression of a fibrous membrane surrounding
bone causes bone resorption. Acta Histochem.1999;101:
203-12.101203Â
1999Â
[PubMed] Â
Skripitz R, Aspenberg P.
Pressure-induced periprosthetic osteolysis: a rat model. J Orthop Res.
2000;18:481-4. Erratum in: J Orthop Res.2002;20:
397.20397Â
2002Â
[CrossRef] Â
Skoglund B, Aspenberg P. PMMA particles
and pressure—a study of the osteolytic properties of two agents proposed
to cause prosthetic loosening. J Orthop Res.2003;21:
196-201.21196Â
2003Â
[PubMed][CrossRef] Â
Karrholm J, Borssen B, Lowenhielm G,
Snorrason F. Does early micromotion of femoral stem prostheses matter?
4-7-year stereoradiographic follow-up of 84 cemented prostheses. J Bone
Joint Surg Br.1994;76:
912-7.76912Â
1994Â
Â
Karrholm J, Herberts P, Hultmark P,
Malchau H, Nivbrant B, Thanner J. Radio-stereometry of hip prostheses. Review
of methodology and clinical results. Clin Orthop Relat Res.1997;344:
94-110.34494Â
1997Â
[PubMed] Â
Prendergast PJ, Huiskes R, Soballe K.
Biophysical stimuli on cells during tissue differentiation at implant
interfaces. J Biomech.1997;30:
539-48.30539Â
1997Â
[PubMed][CrossRef] Â
Mow VC, Kuei SC, Lai WM, Armstrong CG.
Biphasic creep and stress relaxation of articular cartilage in compression?
Theory and experiments. J Biomech Eng.1980;102:
73-84.10273Â
1980Â
[PubMed][CrossRef] Â
Horowitz SM, Doty SB, Lane JM, Burstein
AH. Studies of the mechanism by which the mechanical failure of
polymethylmethacrylate leads to bone resorption. J Bone Joint Surg
Am.1993;75:
802-13.75802Â
1993Â
Â
Maloney WJ, Smith RL, Schmalzried TP,
Chiba J, Huene D, Rubash H. 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:
1301-10.771301Â
1995Â
[PubMed] Â
Van Noorden CJ, Vogels IM. Enzyme
histochemical reactions in unfixed and undecalcified cryostat sections of
mouse knee joints with special reference to arthritic lesions.
Histochemistry.1986;86:
127-33.86127Â
1986Â
[PubMed][CrossRef] Â
Landells JW. The bone cysts of
osteoarthritis. J Bone Joint Surg Br.1953;35:
643-9.35643Â
1953Â
[PubMed] Â
Astrand J, Skripitz R, Skoglund B,
Aspenberg P. A rat model for testing pharmacologic treatments of
pressure-related bone loss. Clin Orthop Relat Res.2003;409:
296-305.409296Â
2003Â
[PubMed][CrossRef] Â
Dambrain R, Dhem A. Bone vitality in
induced radionecrosis of the mandible estimated by osteocytic population
counting. Strahlentherapie.1984;160:
39-44.16039Â
1984Â
[PubMed] Â
James J, Steijn-Myagkaya GL. Death of
osteocytes. Electron microscopy after in vitro ischaemia. J Bone Joint
Surg Br.1986;68:
620-4.68620Â
1986Â
Â
Kaysinger KK, Ramp WK. Extracellular pH
modulates the activity of cultured human osteoblasts. J Cell
Biochem.1998;68:
83-9.6883Â
1998Â
[CrossRef] Â
Carter DR, Blenman PR, Beaupre GS.
Correlations between mechanical stress history and tissue differentiation in
initial fracture healing. J Orthop Res.1988;6:
736-48.6736Â
1988Â
[PubMed][CrossRef] Â
Lippiello L, Kaye C, Neumata T, Mankin
HJ. In vitro metabolic response of articular cartilage segments to low levels
of hydrostatic pressure. Connect Tissue Res.1985;13:
99-107.1399Â
1985Â
[PubMed][CrossRef] Â
de Rooij PP, Siebrecht MA, Tagil M,
Aspenberg P. The fate of mechanically induced cartilage in an unloaded
environment. J Biomech.2001;34:
961-6.34961Â
2001Â
[PubMed][CrossRef] Â
Goodman S, Aspenberg P, Song Y, Knoblich
G, Huie P, Regula D, Lidgren L. Tissue ingrowth and differentiation in the
bone-harvest chamber in the presence of cobalt-chromium-alloy and
high-density-polyethylene particles. J Bone Joint Surg Am.1995;77:
1025-35.771025Â
1995Â
[PubMed] Â
Athanasou NA. Cellular biology of
bone-resorbing cells. J Bone Joint Surg Am.1996;78:
1096-112.781096Â
1996Â
[PubMed] Â
Yao J, Cs-Szabo G, Jacobs JJ, Kuettner
KE, Glant TT. Suppression of osteoblast function by titanium particles.
J Bone Joint Surg Am.1997;79:
107-12.79107Â
1997Â
[PubMed] Â
Dean DD, Schwartz Z, Blanchard CR, Liu
Y, Agrawal CM, Lohmann CH, Sylvia VL, Boyan BD. Ultrahigh molecular weight
polyethylene particles have direct effects on proliferation, differentiation,
and local factor production of MG63 osteoblast-like cells. J Orthop
Res.1999;17:
9-17.179Â
1999Â
[CrossRef] Â
Pandey R, Quinn J, Joyner C, Murray DW,
Triffitt JT, Athanasou NA. Arthroplasty implant biomaterial particle
associated macrophages differentiate into lacunar bone resorbing cells.
Ann Rheum Dis.1996;55:
388-95.55388Â
1996Â
[PubMed][CrossRef] Â
Green TR, Fisher J, Stone M, Wroblewski
BM, Ingham E. Polyethylene particles of a `critical size' are necessary for
the induction of cytokines by macrophages in vitro.
Biomaterials.1998;19:
2297-302.192297Â
1998Â
[PubMed][CrossRef] Â
Ryd L, Linder L. On the correlation
between micromotion and histology of the bone-cement interface. Report of
three cases of knee arthroplasty followed by roentgen stereophotogrammetric
analysis. J Arthroplasty.1989;4:
303-9.4303Â
1989Â
[PubMed][CrossRef] Â
Hirakawa K, Jacobs JJ, Urban R, Saito T.
Mechanisms of failure of total hip replacements: lessons learned from
retrieval studies. Clin Orthop Relat Res.2004;420:
10-7.42010Â
2004Â
[PubMed][CrossRef] Â
Schuller HM, Scholten PE, Lettinga K,
Marti RK, Van Noorden CJ. High cathepsin B activity in arthroplasty interface
membranes. A histochemical study of 9 loose cemented total hip prostheses.
Acta Orthop Scand.1993;64:
613-8.64613Â
1993Â
[PubMed][CrossRef] Â
Johanson NA, Bullough PG, Wilson PD Jr,
Salvati EA, Ranawat CS. The microscopic anatomy of the bone-cement interface
in failed total hip arthroplasties. Clin Orthop Relat Res.1987;218:
123-35.218123Â
1987Â
[PubMed] Â
Hatton A, Nevelos JE, Nevelos AA, Banks
RE, Fisher J, Ingham E. Alumina-alumina artificial hip joints. Part I: a
histological analysis and characterisation of wear debris by laser capture
microdissection of tissues retrieved at revision. Biomaterials.2002;23:
3429-40.233429Â
2002Â
[PubMed][CrossRef] Â