Hydroxyapatite coatings are biocompatible and osteoconductive,
and they enhance bone apposition to metal substrates. Hydroxyapatite-coated
implants in humans have shown good bone apposition1,2 and have provided excellent clinical
results after durations of follow-up of nearly ten years3-6. Rapid bone apposition is probably
an important mechanism whereby early fixation of hydroxyapatite-coated
implants is achieved, and long-term fixation is probably influenced
by bone-remodeling around these implants.
Alendronate (4-amino-1-hydroxybutylidene-1,1-bisphosphonate sodium)
is a relatively new bisphosphonate that is a potent inhibitor of
bone resorption. Although the detailed mechanism of action of alendronate
is still unclear, it has been shown to inhibit osteoclasts’ ruffled
border formation, and it decreases osteoclastic resorption without
destroying osteoclasts7-10. In
addition to inhibiting osteoclasts, alendronate may promote the
deposition of bone matrix protein, osteocalcin, and collagen by
osteoblasts11.
Most total joint implants function very well, but bone resorption,
apparently initiated by particles of wear debris, may lead to aseptic
loosening or regions of osteolysis. Since implant-related bone loss
appears to be the result of osteoclastic resorption, stimulated
either directly or indirectly by macrophages, it has been hypothesized
that osteolysis might be controlled, in part, by the administration
of alendronate. A limited study of total hip arthroplasty in canines
recently provided evidence that peri-implant osteolysis induced
by orthopaedic wear particles was minimized by oral alendronate therapy12,13.
It is anticipated that many patients with osteoporosis, in whom
total joint replacement may be indicated, may be treated with alendronate.
Additionally, some surgeons may want to treat some patients prophylactically
with alendronate to reduce the risk of osteolysis. For this reason,
it is important to understand the influence of alendronate on bone
apposition and bone-remodeling around hydroxyapatite-coated total
joint implants. The purpose of this study was to determine the influence
of alendronate on early bone apposition and on remodeling around
hydroxyapatite-coated total hip implants in canines.
Experimental Design
The study included twelve skeletally mature adult dogs weighing
approximately 30 kg (range, 25.2 to 31.5 kg). Upon arrival, all
animals were examined to ensure that they had a normal health status.
The animals were also screened, during a one-week quarantine period,
to exclude any that had an acute or chronic medical condition. Each
dog underwent a staged bilateral hip arthroplasty with twenty weeks
between the two procedures. The animals were killed four weeks after the
second operation, so the implants were in vivo for four and twenty-four
weeks. Six of the dogs received oral therapy with alendronate (Fosamax;
Merck, Rahway, New Jersey) on an empty stomach from the date of
surgery until they were killed. A dose of 10 mg of alendronate was
given daily, which was similar to the dose given in previous studies12,13. The other six dogs were untreated
controls. One animal was killed during the surgery because of an
intraoperative femoral fracture. Three animals that had a femoral
fracture, dislocation of the hip, or sciatic nerve palsy also were
excluded and were replaced. The study design and experimental procedures were
approved by our institution’s Animal Care and Use Committee.
Description of Device
The hydroxyapatite-coated total hip prostheses for the dogs were
specially designed, manufactured, and provided by Stryker Howmedica
Osteonics (Allendale, New Jersey). They were composed of a titanium-alloy
(Ti-6Al-4V) stem, a cobalt-chromium-alloy modular head, and an ultra-high molecular
weight polyethylene cup. The proximal half of the stem had a macrotextured
surface consisting of arc-deposited CP-titanium with a plasma-sprayed
coating of highly crystalline hydroxyapatite (nominal thickness,
50 m). The final surface roughness was approximately 31 m Ra. The
canine femoral stems were available in standard and large sizes. There
was a Morse-tapered neck for the connection of a femoral head, and
the femoral heads were available with three offsets (+0, +3,
and +6 mm). Polyethylene acetabular components with an
inner diameter of 17 mm were available in outer diameters of 25,
27, and 29 mm.
Surgical Procedure
Anesthesia was induced with 200 mg of Telazol (a combination
of tiletamine and zolazepam) and 0.3 mg of buprenorphine, both given
intramuscularly, and was maintained with halothane, nitrous oxide,
and oxygen. Surgery was performed with aseptic technique. The total
hip arthroplasty was performed through a craniolateral approach,
as previously described14. A 15-cm
skin incision was centered at the level of the greater trochanter.
After the fascia was divided in line with the skin incision, the
biceps femoris muscle was retracted to expose the greater trochanter.
The middle gluteal muscle was retracted, and the cranial portion
of the deep gluteal muscle was cut near the attachment to the greater
trochanter; then the capsule of the hip joint was exposed and resected.
After cutting of the teres ligament, the femoral head was dislocated. Then
the femoral neck was cut according to the osteotomy line determined
with use of a femoral neck resection template. The acetabular cartilage
was curetted, and the acetabulum was reamed until bleeding from
subchondral bone was seen. After three anchor holes, 3.5 mm in diameter,
were drilled, the acetabulum was cleaned and the cup was implanted
with bone cement. The medullary canal of the femur was drilled to
a diameter determined by review of the preoperative anteroposterior
radiograph and then was reamed manually with tapered reamers and
broaches. After reaming, the canal was irrigated with saline solution.
The femoral stem was inserted in approximately 10° of retroversion.
After trial reduction of the femoral head, the size of the femoral
head was chosen and the head was placed onto the femoral neck with
gentle but secure impacts. The hip was reduced, and the stability
of the joint was assessed with the lower limb in all positions.
The deep gluteal muscle was reattached close to its origin, and
the other muscles were placed back in position. The fascia and skin were
closed in layers.
Postoperative Care and Follow-up
Animals were given ampicillin sodium (1 g intravenously) perioperatively
and ampicillin and clavulanate (500 mg orally) twice daily for five
days postoperatively. Buprenorphine (0.3 mg orally) was given every
six to ten hours for two to four days postoperatively for pain relief.
All dogs were allowed to recover from the anesthesia in a controlled
setting and were allowed immediate weight-bearing and walking in their
cages. All animals were examined daily, with particular attention
given to the surgical sites.
Radiographs
Anteroposterior and lateral radiographs of both hips were made
before and immediately after surgery and immediately after the animal
was killed. The postoperative radiographs were evaluated for evidence
of osseous changes around the stems.
Gross and Histological Examination
At the end of the study period, the dogs were killed with an overdose
of intravenous pentobarbital (70 mg/kg). The hip joint
was opened immediately after the animal was killed, and the soft
tissues around the implant were inspected for gross evidence of
inflammation or infection. The surface of the esophagus and the
gastroduodenal wall were carefully observed for evidence of inflammation,
ulcers, or other changes recognized as possible complications of
alendronate therapy15. The femur
was harvested en bloc and cleaned of the soft tissues. Each implant
and the surrounding bone were placed in 70% ethanol within
one hour after death and were fixed for at least two weeks. The
specimens were dehydrated in a graded series of ethanols and embedded
in Spurr plastic. After polymerization, the embedded femora were
radiographed, and the specimen radiographs were used to define three
matched levels in the region of the hydroxyapatite coating (proximal,
middle, and distal, all perpendicular to the axis of the femur)
of each specimen. Each section was isolated by transverse sectioning
with use of a slow-speed diamond saw (Isomet 2000; Buehler, Lake
Bluff, Illinois) under continuous water flow, resulting in sections
of 1 to 2 mm in thickness. Each section was glued to a Plexiglas
slide and then ground with use of sandpaper and a grinding table
(Polimet and Handimet; Buehler). Sections were hand-polished with
a Buehler Ecomet IV polisher to a final thickness of approximately
35 to 50 m and stained with Giemsa stain.
Histomorphometry
Interface
Each microscope slide was observed with use of a light microscope
with transmitted light. The image was transmitted to an interactive
image analysis computer system (Bioquant 95, Nashville, Tennessee),
and automated image-processing was used to help to define the interface
between the implant and the adjacent tissue. The fractional linear
extents of bone apposition were measured by a combination of automatic edge
detection and operator selection with use of image analysis software
and were expressed as a percentage of the circumference of the implant
at each section level. The linear extent of the hydroxyapatite coating
on the implant was similarly quantified and expressed as a percentage.
The thickness of the layer of hydroxyapatite, when it was satisfactorily
visualized, was also measured in at least ten areas around the implant
and was expressed in micrometers. These measurements were used to
calculate the mean hydroxyapatite thickness in each section.
Bone Areas
In order to evaluate quantitative changes of bone due to stress-shielding
or alendronate, we quantified the total amount of bone, both cortical
and trabecular, in each histologic section. We used a method similar
to one previously described16-18.
A dissecting microscope fitted with a video camera was interfaced
with a computer with image analysis software (Bioquant). Each microscope
slide was visualized at a uniform low magnification with the dissecting
microscope, and the image was captured by the digitizing computer
hardware and software. The area of bone was isolated with use of
automated image processing and was expressed as square micrometers.
Each section was divided into quadrants (anterior, posterior, medial,
and lateral), and the cortical and trabecular bone areas were digitized
and expressed as a percentage of total area.
Exclusions
When possible, measurements were obtained from all three sections
of each femur. As described below, small cracks were present in
the proximal-medial cortex of some femora. If there was evidence
of a fracture that affected the normal osteosynthesis around the
implant, that area was excluded from analysis. One femur had a medial
cortical crack that paralleled the length of the implant and was
associated with abundant callus formation and fibrosis. This femur
was excluded from analysis.
Statistical Analyses
Comparisons of the linear extents of bone apposition and hydroxyapatite
coating, the thickness of the hydroxyapatite coating, and the areas
of cortical and trabecular bone between the two groups were performed
with use of the Student t test or the Mann-Whitney U test with 95% confidence
intervals. Analysis of variance with the Fisher PLSD (projected
least significant difference) test was used to assess the influence
of the location of the section (proximal, middle, or distal) on
the measurements of bone apposition and the hydroxyapatite coating
or on the average bone area among the four quadrants. The chi-square
test was also used to compare the frequency of small cracks between
the two groups. P values of <0.05 were considered significant.
General Condition of Experimental Animals
Although four dogs were killed for the reasons described above,
the remaining dogs tolerated the surgical procedure well. The animals
appeared to walk normally, and there was no evidence of infection
or other adverse reaction during the experimental period. No lesions
of esophagitis, recognized as one complication of alendronate therapy15,19, were observed. Evidence of chronic
gastritis was found in a few dogs at the time that they were killed,
but no serious complications of the gastrointestinal tract were
identified.
Radiographic Findings
None of the four-week specimens in either group showed any specific
radiographic changes around the implant. Some femora in both groups
showed evidence of trabecular bone condensation and increased femoral
cortical bone diameter at twenty-four weeks (Figs. 1-A and 1-B), but no obvious
differences between the two groups were observed.
Histologic Findings
Small cracks, as described above, were present in ten femora in
the alendronate treatment group and in three femora in the control
group (p < 0.05, chi-square test). Most of the cracks occurred
early in the course of the study; with additional surgical experience,
their prevalence was reduced. (The surgery schedule was not randomized
with respect to alendronate treatment.) The areas in which these
cracks were found showed endosteal and periosteal new-bone formation,
minimal adjacent apposition of bone to the stem, and a localized increase
in peri-implant fibrous tissue.
Areas away from the cortical cracks showed features typical of
mechanically stable implants, including extensive bone apposition
and the absence of fibrous membranes (Figs. 2-A, 2-B, 2-C, 2-D, 2-E, and 2-F). In four-week specimens, direct
apposition of bone to the hydroxyapatite-coated surface with a lack
of fibrous tissue between the bone and stem was observed in both
groups (Figs. 3-A and 3-B). The hydroxyapatite
coatings showed relatively uniform density and thickness in both
groups. Occasional osteoclasts associated with bone-remodeling were
identified in the femora in both groups. In some areas of all specimens,
there was evidence of coating dissolution, including particles of hydroxyapatite
and macrophages (Fig. 4). In the twenty-four-week specimens,
apposition of bone to the hydroxyapatite coating was maintained
(Fig. 5).
There was no obvious qualitative difference in trabecular or cortical
bone architecture between the two groups or the two time-periods.
Histomorphometric Findings
To test for intraobserver and interobserver variability of the histomorphometric
findings, two operators measured the same specimens on different
days. Both the intraobserver and the interobserver variability was
within 5%.
There was no consistent difference with respect to the hydroxyapatite
coating or bone apposition based on section location (proximal,
middle, or distal). Similarly, there was no difference in the average
cortical or trabecular bone area among the quadrants with respect
to section location. Therefore, the data were summed for each case
and then averaged for each group (Table I).
Bone Apposition
The linear extent of bone apposition was 67% at four
weeks and 64% at twenty-four weeks in the alendronate treatment group
and 67% at four weeks and 69% at twenty-four weeks in
the control group. The extent of bone apposition was not significantly
different between the two groups at either time-period, and it did
not change significantly during the study period (Table I).
Extent of Hydroxyapatite Coating
The linear extent of the hydroxyapatite coating was 60% at four
weeks and 46% at twenty-four weeks in the alendronate treatment
group and 61% at four weeks and 49% at twenty-four
weeks in the control group. The extent of the hydroxyapatite coating
was not significantly different between the two groups at either
time-period, but it significantly decreased with time in both groups
(Table I, p < 0.01
for both groups, unpaired t test).
Thickness of Hydroxyapatite Coating
The thickness of the hydroxyapatite coating was 43 m at four weeks
and 41 m at twenty-four weeks in the alendronate treatment group
and 42 m at four weeks and 40 m at twenty-four weeks in the control
group. The thickness of the hydroxyapatite coating did not differ
significantly on the basis of treatment, and it did not change significantly
during the study (Table I). The hydroxyapatite coatings appeared
to be of relatively uniform thickness in both groups.
Cortical Bone Area
The average percentage of the total area occupied by cortical bone
in both groups was close to those reported in humans20, dogs21,
minipigs22, and baboons23,24. The average bone area in the
cortex was 97% at four weeks and 95% at twenty-four
weeks in the alendronate treatment group and 95% at four
weeks and 99% at twenty-four weeks in the control group.
The average cortical bone area was not significantly different between
the treatment groups, and it did not change significantly during
the study (Table I).
Trabecular Bone Area
The average trabecular bone area was 39% at four weeks
and 37% at twenty-four weeks in the alendronate treatment
group and 39% at four weeks and 36% at twenty-four
weeks in the control group. The average trabecular bone area was
not significantly different on the basis of treatment, and it did
not change significantly during the study (Table I).
Bisphosphonates are pyrophosphate analogs characterized by a
phosphorous-carbon-phosphorous bond25.
They show limited absorption in the gastrointestinal tract and limited
penetration into cells, but they rapidly bind to bone mineral, especially
in regions of bone resorption beneath osteoclasts9,25.
By altering side-chain substitutions, synthetic bisphosphonates
can inhibit osteoclastic resorption, with potencies varying by as
much as 10,000-fold among compounds26,27.
An early bisphosphonate, disodium etidronate, has been effective
in reducing bone loss in Paget disease28,
hypercalcemia of malignancy29,
metastatic carcinoma30, and osteoporosis31, but it was sometimes associated
with spontaneous fracture32, delayed
fracture-healing7, and osteomalacia26,33,34. These findings suggest that,
in addition to inhibiting osteoclasts, disodium etidronate may either
directly or indirectly influence more complicated aspects of bone
turnover.
Alendronate (4-amino-1-hydroxybutylidene-1,1-bisphosphonate sodium),
a newer-generation bisphosphonate, is reported to be an up to 1000
times more potent inhibitor of bone resorption and have a 1000-fold
higher safety margin with respect to inhibition of mineralization
when compared with disodium etidronate8,35,36.
Alendronate has been shown to inhibit ruffled border formation and
to decrease net osteoclastic bone resorption without destroying
the osteoclasts themselves7,9,10.
In addition to inhibiting osteoclasts, alendronate may influence
bone formation by directly inhibiting crystal growth, by directly
inhibiting osteoblasts, or by indirectly inhibiting bone resorption,
the likely first step in coupled new-bone formation11. Alendronate has been approved by
the United States Food and Drug Administration for the treatment
of diseases characterized by excessive osteoclastic bone resorption,
such as hypercalcemia of malignancy, Paget disease, metastatic bone disease,
and sometimes osteoporosis. An experimental study suggested that
alendronate also may be of benefit in the treatment of osteolysis
induced by particles of orthopaedic wear debris12.
The mechanisms responsible for the osteoconductive properties
of hydroxyapatite are unclear. One may involve dissolution of a
small amount of calcium and phosphate from the amorphous phase of
the synthetic coating shortly after implantation, followed by precipitation
of a calcium phosphate apatite of low crystalline order and higher
carbonate content, similar to normal bone mineral37.
This crystal precipitation, and possible subsequent crystal growth,
may influence the extent or distribution of bone apposition to bioactive
coatings. It has been suggested that alendronate may rapidly cover
the surface of endogenous bone mineral and may reduce the rate of
endogenous apatite crystal growth and dissolution9.
If alendronate were to interfere with the kinetics of ion dissolution
and precipitation from hydroxyapatite coatings, then it might inhibit
bone apposition. In spite of these concerns, our results showed
no inhibition of bone apposition to hydroxyapatite coatings in the
presence of alendronate at either four or twenty-four weeks postoperatively.
Once initial bone apposition has occurred, long-term fixation of
total joint implants is influenced by regional bone-remodeling.
Because new-bone formation is thought to be coupled with bone resorption,
it is possible that the normal bone-remodeling around implants induced
by changing mechanical loads might be altered by alendronate, potentially
compromising long-term fixation. While alendronate has been shown
to influence skeletal remodeling in disorders with abnormal bone
turnover, including ovariectomized animals and those with hyperthyroidism23,24,38-41, most studies have suggested
a minimal influence on normal bone. For example, Balena et al.21 and Peter et al.42 found
that alendronate administration had no substantial long-term influence
on canine cortical or cancellous bone, and Wang et al.13 reported that alendronate therapy
had no substantial influence on the biomechanical properties of
bone around a canine total hip replacement at twenty-four weeks.
Lafage et al.22 found no substantial
influence of alendronate therapy on cancellous bone volume in minipigs,
but they reported a significant decrease in the mean area of cortical
Haversian cavities (p = 0.02). Although our study period
was only twenty-four weeks, our results showed no detectable influence
of alendronate on the quantity or qualitative pattern of peri-implant bone-remodeling.
More specifically, alendronate therapy did not significantly influence
the extent of bone apposition to the implant, the trabecular bone
area in the surrounding metaphysis, or the adjacent metaphyseal
and diaphyseal cortical bone area.
Also of interest in this study is the higher rate of small proximal
cortical cracks seen in the calcar region in canines treated with
alendronate when compared with those who did not receive alendronate
therapy. Our implantation schedule was not randomized with respect
to alendronate treatment, however, and most of the alendronate-treated
animals were operated on early in the course of the study. Therefore,
we suspect that the difference in frequency of cortical cracks primarily reflects
increasing experience gained by the surgical team over the course
of the study rather than a change in the mechanical properties of
the cortical bone due to alendronate therapy.
The processes of bone-remodeling that participate in calcium homeostasis
and skeletal remodeling in response to changing mechanical loads
involve the interaction of many different cells, but the first step
is often described as involving osteoclasts. These same processes
are necessary to achieve and maintain adequate mechanical stability
of hydroxyapatite-coated total joint implants. Although alendronate
has not been found to have detrimental effects on the skeleton of
normal laboratory animals, it is appropriate to be concerned about
its safety in patients receiving hydroxyapatite-coated implants. Nevertheless,
our results suggest that, at doses believed to be therapeutic, alendronate
is unlikely to cause changes in bone apposition or remodeling that
might compromise early fixation of total joint prostheses.