Eight skeletally mature adult male mongrel dogs weighing 27 to 35 kg (mean,
29.4 kg) were used in this study. Skeletal maturity was confirmed by
radiographic evidence of closure of the proximal epiphyseal plate of the
tibia. Each dog underwent same-day bilateral implantation of a femoral
segmental prosthesis. On the experimental side, 3.5 mg of rhOP-1 in a dry
powder form (Stryker Biotech, Hopkinton, Massachusetts) mixed with 1 g of a
bovine-derived type-I-collagen additive (Stryker Biotech) was applied under
and between the cortical bone strips. The bovine type-I-collagen additive was
a dry powder of carboxymethylcellulose sodium. At the time of surgery, the
contents of the collagen additive and the rhOP-1 powder were mixed and wetted
with 3.0 mL of sterile saline solution to form a paste-like material. On the
control side, autogenous cancellous bone chips and marrow were harvested from
the femoral canal to be used as osteoinductive factors. The animals were
examined daily for signs of bone fracture or infection until they were killed
twelve weeks after the surgery. The experimental study protocol was approved
by the Animal Use and Care Committee at Johns Hopkins University.
Prosthetic Design
The prosthetic implant developed for this study was a bistemmed segmental
diaphyseal bone-replacement prosthesis made of a cobalt-chromium alloy. The
surface of the segmental portion of the prosthesis was coated with two layers
of sintered cobalt-chromium-alloy beads (diameter, 600 to 800 µm). The
implant was modular with two components joined by a conical coupling; the body
of the prosthesis was 60 mm in length and 16 mm in diameter. The
intramedullary stems were 10 mm in diameter and fluted for secure cement
fixation. The proximal stem was 30 mm long, and the distal stem was 40 mm long
(Fig. 1).
Allograft Preparation
Previously harvested canine mid-femoral and mid-tibial diaphyseal segments
stored at —80°C were used as the allograft sources for this study.
The bones were formed as strips 30 mm long and approximately 2 × 2 mm in
cross section with use of a special cutting jig. The bone strips were then
washed in saline solution to remove bone marrow and debris. After defatting
with methanol, the allograft strips were stored at —80°C for three
days. Eighteen bone strips were bundled with use of nonabsorbable sutures at
intervals of 1 to 2 mm. Two weeks before the surgery, the bone strips were
sterilized with ethylene oxide gas and stored at room temperature for aeration
to remove residual ethylene oxide and its breakdown
products23.
Surgical Procedure
The dogs were given Telazol (tiletamine and zolazepam; 0.5 to 1.0 mg/kg,
intramuscularly) as an anesthetic premedication. General anesthesia was
induced by inhalation of halothane and O2 and NO2 after
intubation. Cefazolin (1 g) was injected intramuscularly. A lateral approach
to the femoral diaphysis was made through the fascia lata between the vastus
lateralis and biceps femoris muscles. The femoral periosteum was elevated
circumferentially, and the insertions of the adductor muscles were incised
along the linea aspera. A 6-cm segment of the diaphysis was resected with a
reciprocating saw. The field was irrigated continuously to minimize thermal
necrosis of bone. The prosthesis was secured by intramedullary cement
fixation.
On the control side, autogenous cancellous bone chips and marrow (13
cm3 in volume and 8.5 to 11.4 g by weight) were harvested from the
remaining proximal and distal femoral canals and were evenly distributed over
the bone-prosthesis junctions and adjacent femoral cortex. The allograft
strips, preassembled in a mesh form, were applied over the grafted areas, and
the remaining cancellous bone chips and marrow mixtures were inserted between
the allograft strips. Finally, the cortical bone strips were secured around
the grafted area with use of nonabsorbable suture ends. The periosteum and
soft tissue were carefully replaced over the reconstructed area and were
repaired in layers with use of absorbable sutures.
On the experimental side, an approximately 1-mm-thick layer of OP-1 putty
was applied with a custom-made applicator at the proximal and distal
prosthesis-bone junctions, which were to be covered by the allogenic bone
strips. The preassembled cortical allograft strips were tied around the
proximal and distal bone-prosthesis junctions over the OP-1-putty-covered
areas. Pressure was applied to allow the OP-1 putty to fill the gaps between
the cortical strips. The soft tissue and wound were closed and repaired in a
manner similar to that used on the control side (Figs.
2-A,
2-B, and
2-C). The wounds were not
drained on either the control or the experimental side.
Immediately after the surgery, the animals were closely monitored for any
complications until they recovered from the anesthesia. The animals were given
butorphanol (0.02 mg/kg, intramuscularly, for twelve hours) after the surgery
for pain relief. All animals were examined twice a day for the first week
after the surgery. Immediate weight-bearing and mobilization were allowed
postoperatively as tolerated. The animals were examined for signs of lameness,
incisional swelling, and limb edema.
At biweekly intervals, the animals were given the bone-labeling agents
xylenol orange (90 mg/kg, intravenously), calcein blue (30 mg/kg,
intravenously), alizarin complexone (30 mg/kg, intravenously), and
oxytetracycline (30 mg/kg, intramuscularly) so that mineral apposition rates
could be measured with use of a multiple fluorochromic labeling technique as a
histological evaluation after the animals were killed.
Load-Bearing-Gait Analysis
Load-bearing during normal gait was measured before the surgery and at
four, seven, ten, and twelve weeks postoperatively. Each dog was led by a
leash over a force-plate (model OR-6-6; Advanced Mechanical Technology,
Watertown, Massachusetts), with a minimal approach distance of 4 m, which
allowed the establishment of a normal walking speed. An observer recorded each
successful foot strike, defined by complete contact of either the left or the
right hindfoot within the margins of the force-plate surface. At least six
valid maximum vertical ground-reaction-force measurements were recorded, and
the mean was calculated for each hindlimb at each time-period.
Radiographic Analysis
Anteroposterior and lateral radiographs were made before the surgery to
exclude skeletal abnormalities and to ensure that the femoral medullary size
was appropriate to receive a prosthetic stem for cement fixation. Standardized
radiographs were made immediately after the surgery and at one, two, four,
six, nine, and twelve weeks postoperatively with the animal sedated with
Telazol (0.5 to 1.0 mg/kg, intramuscularly). The radiographs were digitized
with use of a light box and a charged coupled device camera (model DXC-151;
Sony, Tokyo, Japan). The mineralized area around the prosthesis was determined
by tracing the contour of the mineralized area on the digitized radiographic
images with use of image-analysis software (Bioquant System IV; R and M
Biometrics, Nashville, Tennessee) and a sonic digitizer (Summa Sketch II Plus;
Sammagraphics, Seymour, Connecticut). The mineralized areas at one week and
thereafter were expressed as a percentage of the measured area immediately
after the surgery (the area covered by the allograft on the experimental side
and that covered by the allograft or cancellous bone chips on the control
side).
Histological Analysis
The surrounding muscle and soft tissue were removed from the retrieved
femur. When the bridging bone covered the coupling joint, it was cut with a
reciprocating saw. The prosthesis was separated at the coupling joint. The
distal bone-prosthesis segment was used for histological analysis. The
specimens were fixed in a 70% ethanol solution, dehydrated in increasing
concentrations of ethanol, defatted in acetone, and embedded in
polymethylmethacrylate (Technovit 9100; Kulzer, Wehrheim, Germany). A section
was taken from each of the following regions: the distal junction between the
prosthesis and bone (shoulder region), 10 mm distal to the junction (host bone
region), and 10 mm proximal to the junction (prosthesis region). After they
were ground to a thickness of 100 µm, unstained sections were viewed under
ultraviolet light to identify labeled new bone formation in the allograft. The
distance between two adjacent labels was measured in at least four locations
with use of the image analysis software (Bioquant System IV), and the mean
distance was divided by the labeling interval time to determine the mineral
apposition rate.
Contact microradiographs of the same sections were made with use of
high-resolution x-ray film (Industrex SR; Kodak-Industrie, Challon sur Saune,
France) exposed with 35 kW and 3 mA, a target-to-specimen distance of 20 cm,
and an exposure time of forty-five seconds. The contact microradiographs were
developed for five minutes in a film developer (model D-19; Eastman Kodak,
Rochester, New York). The contact microradiographs were viewed with back
light, and the surface area (perimeter) of the prosthesis in contact with the
bone tissue of the extracortical bone envelope was measured with use of
image-analysis software (Bioquant System IV). The bone contact area was
expressed as a percentage of the entire circumferential porous surface of the
prosthesis. The porosity of the allograft was determined with use of the
digitized images of contact microradiographs and custom software on a
workstation (Iris Indigo Elan; Silicon Graphics, Mountain View,
California)24. The
void area was expressed as a percentage of the graft cross-sectional area.
Biomechanical Testing
The proximal part of the femur, including half of the segmental prosthesis
plus the stem, was used for torsional testing to quantify the strength of the
extracortical bone-bridging and the ingrowth fixation to the host bone. In
order to eliminate the confounding fixation strength of the bone cement, the
cement was removed prior to torsional testing, as previously
described13,25.
Briefly, the proximal part of the femur, including the trimmed femoral head
and the greater trochanter, was embedded in Wood's metal with a specially
designed alignment jig. The block of Wood's metal and the conical coupling of
the prosthesis were fixed to an instrument to remove the cement. A custom-made
trephine, 12.00 mm in outer diameter and 10.10 mm in inner diameter, was
advanced through the trimmed femoral head or the greater trochanter with water
irrigation for lubrication and removal of the cement debris. The trephine was
advanced to the set depth where the tip of the trephine reached the stem-body
junction of the proximal part of the prosthesis. Radiographs were used to
verify that the cement had been completely removed. After cement removal, the
bone-prosthesis specimens were mounted on a materials testing machine (MTS
Bionix 858; MTS, Eden Prairie, Minnesota) for torsional testing. The proximal
end of the test specimen (the Wood's metal block) was connected to a universal
joint to correct the malalignment of the test specimen in order to apply
accurate axial rotation. The distal end of the specimen was fixed to a
specially designed mounting jig to fit the conical coupling mechanism of the
proximal component of the prosthesis. The specimens were kept moist with
saline solution during testing to prevent alterations in the biomechanical
properties of bone and other connective tissues. A slow internal rotation at
15°/min up to a maximum of 40° was used. The slope of the initial
linear portion of the torque-rotational angle curve was used to determine the
torsional stiffness. Because of the complex yield behavior during torsional
testing, a failure point was not clear in the torque-rotational angle curve in
some specimens. Therefore, torsional strength was defined as the maximum
torque recorded during testing.
Statistical Analysis
A one-way analysis of variance with a Tukey-Kramer post hoc test was used
to analyze time-related differences in the load-bearing and radiographic
parameters for each side. Differences between the experimental and control
sides at each time point were analyzed with the two-sided paired t test.
Differences in mechanical and histomorphologic parameters between the
experimental and control sides were also analyzed with use of the two-sided
paired t test. Correlations between the mechanical parameters (torsional
stiffness and maximum torque) and the mineralized area at twelve weeks or the
histological parameters (new bone area and porosity of the allograft at three
regions and bone ingrowth) were analyzed with a simple regression model.
Differences were considered to be significant when the p value was <0.05.
The results are presented as the mean and standard deviation.
All dogs had a transient increase in body weight of approximately 5% during
the postoperative period. Red and white blood-cell counts, serum electrolyte
levels, and protein levels were within normal ranges. A fracture was observed
radiographically in the proximal part of the femur on the control side of one
dog twelve weeks after the surgery. The fracture probably was due to
overreaming of a small medullary canal. Although the fracture was apparent on
the radiograph at twelve weeks, which was just before the dogs were scheduled
to be killed, a retrospective evaluation of a sudden decease in load-bearing
on the control side between seven and ten weeks indicated that the fracture
may have occurred during that period. Therefore, that dog was excluded from
the analyses. Another dog exhibited lameness for two months after the surgery,
and distal stem loosening on the experimental side was diagnosed
radiographically at nine weeks after the surgery. Because there were no
remarkable deviations in the load-bearing data and the implant was rigidly
fixed to bone with no indication of loosening when the animal was killed, this
dog was included in the study.
Load-Bearing During Gait
All animals were able to stand unassisted on the first postoperative day
and bear weight on both operatively treated limbs within two days. An initial
decrease in load-bearing was observed on both sides four weeks after the
surgery, but these load-bearing values were not significantly different from
those at other time-periods (Table
I). There were no significant differences in load-bearing between
the experimental side and the control side throughout the experimental
period.
Radiographic Analysis
The periosteal mineralized area increased significantly between one and
four weeks and between one and six weeks after the surgery on the experimental
side (p < 0.01) and between one and four weeks after the surgery on the
control side (p < 0.05). On the experimental side, a significant decrease
in the periosteal mineralized area was observed between four and twelve weeks
(p < 0.05); however, the mineralized area at twelve weeks remained
increased (by 228%) relative to the original area (measured immediately after
the surgery). On the control side, the mineralized area decreased to 86% of
the original area during the same period (p < 0.01). The mineralized area
on the experimental side was significantly larger than that on the control
side at two weeks (a 1.9-fold increase; p < 0.04), four weeks (a 2.7-fold
increase; p < 0.01), and six weeks (a 2.4-fold increase; p < 0.03) after
the surgery (Fig. 3,
Table II).
Biomechanical Testing
No fractures in the host bone or extracortical bone envelope were observed
macroscopically during torsional testing. The torsional stiffness on the
experimental side was 2.3-fold greater than that on the control side (p <
0.03). The maximum torque on the experimental side was 2.2-fold greater than
that on the control side (p = 0.058) (Table
III).
Histological Analyses
The allograft strips were located evenly, as they had been placed during
the surgery, on both the experimental and the control sides. With the numbers
studied, no significant differences were detected between the experimental and
control sides in terms of new bone area in any of the regions or in terms of
bone ingrowth over the prosthesis (the prosthesis region)
(Table IV). In three animals,
the original allograft strips were completely resorbed at the prosthesis
region on the experimental side. The porosity in the remaining allograft on
the experimental side was significantly (3.8-fold) higher than that on the
control side at the bone-prosthesis junction (the shoulder region, p <
0.02). It was also significantly higher at 10 mm distal to that junction (the
host-bone region, p < 0.004). The average of the allograft porosity values
for the three regions was significantly higher on the experimental side (27.8%
± 13.0%) than on the control side (8.5% ± 5.7%, p < 0.007).
The allografts on the experimental side were surrounded by new bone and
integrated with the extracortical bone envelope
(Fig. 4-A), whereas the outline
of the allografts was still clear on the control side
(Fig. 4-B). New-bone labeling
was intensive around the allograft, but some labeling could be seen within the
remaining allograft and on the resorbed surface in the allograft on the
experimental side. Labeled Haversian systems were rarely seen in the allograft
on the experimental side. Haversian systems with well-demarcated fluorochromic
labels were seen sparsely in the allograft on the control side. The mean
mineral apposition rates in the allograft at different regions ranged from
0.61 to 1.10 µm/day; no significant differences between the experimental
side and the control side were detected with the numbers available.
Relationship Between Mechanical Parameters and Radiographic or
Histological Parameters
Torsional stiffness strongly correlated with the area of the periosteal
mineralization at twelve weeks (r = 0.739, p < 0.003), the new-bone area at
the prosthetic shoulder region (r = 0.564, p < 0.04), the bone ingrowth (r
= 0.669, p < 0.009), and the porosity of the allograft at the shoulder
region (r = —0.537, p < 0.003). The maximum torque also strongly
correlated with the periosteal mineralized area at twelve weeks (r = 0.913, p
< 0.0001), the new-bone area over the prosthesis region (r = 0.685, p <
0.007), the new-bone area at the prosthetic shoulder region (r = 0.704, p <
0.005), and the bone ingrowth (r = 0.739, p < 0.003).
We investigated whether OP-1 putty, when used in combination with an
allogenic cortical bone onlay graft, could serve as a substitute for
autogenous cancellous bone graft and marrow to enhance fixation of a segmental
bone-replacement prosthesis through extracortical bone-bridging. In our
previous study, in which we used the same bilateral prosthetic reconstruction
model with a different bone-grafting technique, autogenous onlay bone-grafting
significantly increased the torsional stiffness (p < 0.007) and strength (p
< 0.05) of the fixation of a segmental replacement prosthesis compared with
the stiffness and strength of prosthetic fixation without
bone-grafting13. In
the present study, the use of allograft bone strips augmented with OP-1 or
autogenous cancellous bone chips and marrow produced torsional fixation
stiffness and strength that were superior or equivalent to those associated
with fixation with an autogenous cortical onlay graft augmented with
autogenous cancellous bone and marrow. Therefore, we suggest that allogenic
cortical bone strips that have an osteoconductive property can serve as a
substitute for autogenous cortical bone strips as the onlay bone graft for
implant fixation through extracortical bone-bridging and ingrowth. Because
autogenous cortical bone strips are difficult to obtain and are often limited
in amount and quality, the application of osteoconductive allogenic cortical
strips augmented with rhOP-1 or autogenous cancellous bone chips and marrow
should achieve equivalent or better clinical results. In addition, use of the
preassembled allograft-strip mesh simplified the surgery and substantially
improved the quality of graft placement and fixation.
Serial radiographs showed that, in this model, the OP-1 putty resulted in
more new bone formation around the prosthesis at two, four, and six weeks
after the surgery compared with the new bone formation in the group treated
with the autogenous cancellous bone and marrow. Early new bone formation
following use of rhOP-1 has been reported in various settings, including
fracture-healing, allograft incorporation, and bone-defect
healing17-22.
In our previous study in which we used the same bilateral prosthetic
reconstruction model with a different bone-grafting technique, cancellous bone
chips that had been simply placed on the surface of the host bone and
prosthesis tended to scatter because of the movement of the surrounding soft
tissues during the early phase of extracortical
bone-bridging11.
Therefore, early callus formation stimulated by rhOP-1 and the allograft strip
mesh would be beneficial to secure the graft materials at the junctions
between the host bone and the prosthesis.
The torsional stiffness in the group treated with the combination of
allogenic cortical bone strips and OP-1 putty was superior to the torsional
stiffness in the group treated with the allogenic strips and autogenous
cancellous bone graft and marrow (p < 0.03). The torsional stiffness and
the maximum torque of the fixation with the allograft and the OP-1 putty (the
experimental side) were 2.8-fold and 2.3-fold greater than those of the
fixation with the autogenous cortical strips and autogenous cancellous bone
and marrow grafts in our previous study, in which we used the same bilateral
prosthetic reconstruction
model13.
Furthermore, the torsional stiffness and the maximum torque were 50.5-fold and
11.8-fold greater than those values following fixation with no graft in the
previous
study13.
In the present investigation, chronological changes in the mineral
apposition rate throughout the study period could not be evaluated with the
multiple bone-labeling technique because of insufficient numbers of
well-demarcated bone labels in the allograft. Although there were some
qualitative differences in the distribution pattern of the bone labels in the
allograft between the experimental and control sides, further bone
histomorphometric analysis will be required to investigate possible
differences in the remodeling process of the allograft.
Effects of bone morphogenetic proteins (BMPs) on bone resorption have been
reported by several
investigators18,19,21,22.
Sumner et al. reported the dose-dependent effects of BMP-2 on bone resorption
in the host bone adjacent to a BMP-2-loaded implant before new bone formed
around the
implant26. In the
present study, a high dose of rhOP-1 (rhBMP-7) was used, and the allograft
porosity was significantly higher on the experimental side than it was on the
autograft side. In previous studies in which a 4-cm intercalary allograft was
used in canine femora, increased bone resorption was also found in the
allograft treated with
rhOP-121,22.
In one of these studies, immersion of the allograft in rhOP-1 solution (1
mg/mL in 5.0% lactose buffer solution) for only one hour before implantation
increased porosity (by 10.4%) in the allograft at twelve weeks after the
surgery21. In the
other study, involving the same intercalary allograft model except for the
dose and method of application of the rhOP-1 (3.5 mg of rhOP-1 combined with
type-I-collagen putty), the porosity of the intercalary graft was 12.9% when
the OP-1 had been applied in the medullary space and 10.2% when it had been
applied in the extracortical
space22. Additional
studies are needed to test whether the effect of rhOP-1 on allograft
resorption is dose-dependent.
In the present study, strong correlations were found between the torsional
stiffness or maximum torque and the new-bone area or bone ingrowth of the
extracortical bone envelope. These results were unsurprising because the
extracortical bone envelope is the main structure that transmits the torque
between the prosthesis and the host bone through the porous coating surface of
the prosthesis after removal of the cement. Interestingly, a strong negative
correlation was found between the torsional stiffness and the porosity of the
allograft at the host bone-prosthesis junction. This result may be explained
by assuming that the allograft strips were still contributing as a structural
component of the extracortical bone envelope at twelve weeks and that the
reduction in the stiffness of the allograft due to increased bone porosity
affected the entire torsional stiffness. However, while allograft porosity was
significantly higher on the experimental side, torsional stiffness was also
significantly higher on the experimental side. Allograft integration with the
surrounding bone tissue, as observed on microradiographs, may have overcome
the reduction in the stiffness of the allograft on the experimental side.
Microstructural analysis, microscopic material testing, and stress analysis of
the extracortical bone envelope are required to investigate the contribution
of allograft integration to the mechanical characteristics of prosthetic
fixation.
Gamma irradiation is most frequently used for allograft sterilization.
However, ethylene oxide sterilization was used in the present study because of
a concern that gamma irradiation might reduce the mechanical strength of the
thin allogenic cortical strips. The ethylene oxide sterilization may have
further reduced osteoconductive and osteoinductive properties of the allograft
in the present
study27.
In the present study, the contralateral side was used as a control to
minimize intersubject variations and to allow comparisons between the
experimental and control sides in a paired fashion. However, locally delivered
recombinant human transforming growth factor-ß2 was reported to have a
systemic enhancing effect on bone regeneration at a remote site in a canine
bilateral proximal humeral
model28. Therefore,
there is a possibility that rhOP-1 applied to the experimental side in our
study systemically enhanced extracortical bone-bridging and ingrowth on the
contralateral side.
In this study, the periosteum was repaired at the fixation site, which may
have also enhanced extracortical bone-bridging and ingrowth on both the
experimental and the control sides. Because the periosteum is absent in some
clinical cases of massive bone defects, the possible contribution of the
periosteum seen in this study should be taken into consideration when its
conclusions are translated to the clinical setting.
In conclusion, in this animal model, we demonstrated that OP-1 putty
combined with allogenic cortical bone strips can be an effective bone-graft
substitute for fixation of a segmental bone-replacement prosthesis with use of
the extracortical bone-bridging principle. ?