There has been growing interest in the use of methods for enhancing bone
repair in order to improve the treatment of recalcitrant fractures, to speed
the healing of osseous defects related to trauma or cancer, to accelerate
distraction osteogenesis, and to provide consistent implant fixation,
especially in patients managed with revision total joint replacement. A number
of physical and chemical modalities are being investigated, and some of these
new techniques are now in clinical
use1-6.
Intramembranous bone regeneration is critical following cementless joint
replacement in order to establish a mechanical connection between the implant
and the host
skeleton7. The
presence of variable bone ingrowth in implants retrieved at the time of
autopsy or revision
surgery8,9
and the relatively poor clinical results of revision of failed total joint
replacements10,11
have provided motivation to find anabolic agents for intramembranous bone
regeneration and implant fixation.
Important issues related to cementless implant fixation, particularly when
used for revision of failed joint replacements, include the presence of
osseous defects at the bone-implant
interface12. Gaps
of 0.5 to 3 mm have been shown to inhibit bone ingrowth, with larger gaps
causing more
inhibition13-15.
In a variety of gap models, isoforms of transforming growth factor-beta
(TGF-ß)14,16-21
and bone morphogenic protein
(BMP)22-31
have been shown to stimulate bone ingrowth, gap-healing, or implant fixation.
In general, the findings of those studies have been consistent with those of
reports on other growth factors in the context of implant
fixation32-34.
The potential interaction between exogenously delivered growth factors is
relatively unexplored, but there are data in the literature to support the
hypothesis that combined treatment may provide more secure implant fixation
than the use of a single growth factor. Some of the early work in the
periodontal field and work involving dental implants involved the use of
combinations of growth factors (platelet-derived growth factor [PDGF] and
insulin-like growth factor
[IGF])35-38.
PDGF has been shown to potentiate demineralized bone matrix-induced bone
formation in a rat subcutaneous
model39. Combined
IGF-1 and TGF-ß1 treatment has been shown to be synergistic for
fracture-healing in a rat
model40. An in
vitro study of fetal rat calvaria demonstrated that IGF-I, PDGF-BB, and
TGF-ß individually increased matrix apposition and that the simultaneous
addition of all three growth factors led to the greatest matrix
apposition41. A
material now known to contain BMP-2, BMP-3, and TGF-ß was used in an
experimental model for head and neck
surgery42.
TGF-ß1 was shown to have positive effects on osteoinduction when added to
implants of demineralized bone matrix (a preparation likely to contain BMP-2
and other BMPs) that were implanted on the cranial periosteum of
rabbits43.
TGF-ß2 also was shown to enhance the osteoinductive activity of a
preparation containing BMP-2 and BMP-3 in rat subcutaneous
tissue44.
TGF-ß1 was shown to act synergistically with BMP-7 in the induction of
endochondral bone formation in
baboons45, and the
combined use of TGF-ß3 and BMP-2 was more effective in inducing ectopic
bone formation in mice than was the use of these factors
alone46.
The present study was designed to test the hypothesis that rhTGF-ß2
and rhBMP-2 act in an additive or synergistic manner to enhance implant
fixation. We used a canine model in which the primary end points, including
the strength of fixation of the implant, interface stiffness, and energy to
failure, were mechanical. Secondary end points describing bone ingrowth, bone
volume, and trabecular architecture in areas near the implant were also
investigated and were helpful in accounting for the variance seen in the
primary end points.
Recombinant human TGF-ß2 and rhBMP-2 were investigated in an
Institutional Animal Care and Use Committee-approved study. The study involved
a canine model in which porous-coated rods were implanted bilaterally in the
proximal part of the humerus in the presence of 3-mm-wide defects between the
implant and host bone (Fig. 1).
The implants were 50 mm long and had a 4-mm-thick central rod made from the
titanium alloy Ti-6Al-4V and a 1.5-mm thick commercially pure titanium
fiber-metal surface with a porosity of 50% (a gift from Zimmer, Warsaw,
Indiana). The porous surface was 30 mm long and gave the implant an outer
diameter of 7 mm. Polyethylene spacers with a diameter of 13 mm were placed
proximally and distally to maintain a 3-mm gap at the interface. Twenty-eight
animals were separated into four groups of seven animals each. In Group 1, the
implant in the left humerus was loaded with 12 µg of rhTGF-ß2 and the
contralateral (control) implant was loaded with buffer only; in Group 2, the
implant in the left humerus was loaded with 25 µg of rhBMP-2 and the
contralateral (control) implant was loaded with buffer only; in Group 3, the
implant in the left humerus was loaded with 12 µg of rhTGF-ß2 as well
as 25 µg of rhBMP-2 and the contralateral (control) implant was loaded with
buffer only; and in Group 4, the gap surrounding the implant in the left
humerus was packed with autogenous bone graft and the gap surrounding the
contralateral implant was left empty. All animals were examined at four
weeks.
The porous surfaces of all implants in Groups 1, 2, and 3 were coated with
hydroxyapatite-tricalcium phosphate by plasma flame-spraying of a commercial
preparation of hydroxyapatite
(Ca10[PO4]6[OH]2). The
hydroxyapatite-tricalcium phosphate coating was applied to serve as a carrier
for the growth factor applications. The major phase of the ceramic after
passing through the plasma flame-spray was tricalcium phosphate
(Ca3[PO4]2), with some hydroxyapatite and
uncharacterized portions also
present47. The
implants in Group 4 were not coated with hydroxyapatite-tricalcium phosphate.
All implants were gamma-sterilized. For the implants in Groups 1, 2, and 3,
gamma sterilization occurred before application of the growth factor or buffer
and sterility was maintained throughout processing.
The target doses of the growth factors (rhTGF-ß2 [a gift from Genzyme,
Framingham, Massachusetts] and rhBMP-2 [a gift from Wyeth Research, Cambridge,
Massachusetts]) were evenly applied to the implant surfaces by means of a
pipette with use of an aqueous buffer (5-mM glutamate, 5-mM NaCl, 2.5%
glycine, 0.5% sucrose, 0.01% Tween 80; pH 4.5). The contralateral (control)
implants for Groups 1, 2, and 3 were prepared in a similar manner with use of
an equal volume of aqueous buffer. The implants were then air-dried. We chose
the 12-µg and 25-µg doses of TGF-ß2 and BMP-2, respectively, on the
basis of previous dose-response experiments performed in the same model with
the same hydroxyapatite-tricalcium phosphate coating of the implant
surface16,22
and unpublished data. Although not examined in the present study, previous
work with a similar application of TGF-ß has shown that 15% to 25% of the
applied growth factor is released in vitro within twenty-four to forty-eight
hours, with little subsequent
diffusion16,48.
The implantations were performed in skeletally mature adult male mongrel
dogs with a body mass ranging from 30 to 45 kg with use of the same surgical
procedure described
previously49. In
the autograft group (Group 4), approximately 6 cm3 of autogenous
bone particles that had been obtained during preparation of the implantation
sites were lightly packed within the 3-mm gap in the left humerus whereas the
gap in the right humerus was not treated. The animals recovered from the
surgery uneventfully and were fully able to walk within two days. No
complications occurred, and all wounds healed normally. The animals were
killed with the administration of a supersaturated solution of barbiturates
four weeks after surgery.
The retrieved humeri with implants in situ were radiographed, wrapped in
saline solution-soaked gauze, and frozen. A diamond-tipped blade on a cut-off
saw was used to make cuts of the intact specimens perpendicular to the long
axis of the implant so that 5-mm-thick wafers were obtained. Two wafers were
refrozen and were used for mechanical push-out testing, and two were placed in
10% neutral buffered formalin and were used for backscatter scanning electron
microscopy and histological analysis (Fig.
2). The methods of scanning electron microscopy and histological
analysis have been described in detail
previously14,16.
Briefly, the wafers were embedded in methylmethacrylate, and a surface was
prepared for scanning electron microscopy. For each wafer, the prepared
surface was carbon-coated, and four equally spaced and sized images of the
full thickness of the porous coating from each of two wafers were made at
30× magnification (model 840A [JEOL USA; Peabody, Massachusetts],
equipped with a backscatter electron detector linked to a frame grabber),
stored digitally, and analyzed with the aid of image-analysis software
(Image-Pro Plus; Media Cybernetics, Silver Spring, Maryland) to measure the
volume fraction of bone ingrowth—that is, the amount of void space
within the porous coating occupied by mineralized bone. Four equally spaced
images of the complete depth of the gap were used to determine the volume
fraction (BV/TV), trabecular plate thickness (Tb.Th), intertrabecular
separation (Tb.Sp), trabecular number (Tb.N), and trabecular bone-specific
surface (BS/BV) of mineralized bone in the 3-mm gap. These images were
collected at 20× magnification and were analyzed with custom
software50. In
addition, one image from an area located approximately 2 mm medial to the
implantation site, but not directly affected by the surgical procedure, was
analyzed with regard to the same five parameters that were used to describe
the morphology of bone within the gap. From these same specimens, 1-mm-thick
sections were cut, attached to plastic slides, ground to a nominal thickness
of 100 µm, stained with toluidine blue/basic fuchsin, and viewed with light
microscopy to provide a qualitative assessment of the interface.
For mechanical testing, the wafers were thawed, kept moist, and tested at
room temperature. The implants from the left and right humeri of each animal
were treated identically and were tested within minutes of each other. The
wafers were supported by a baseplate that had a 14-mm-diameter opening so that
failure during testing could occur at the bone-implant interface, within the
3-mm gap, or at the junction of the 3-mm gap and the surrounding host bone. A
plunger with a 5.5-mm diameter, thereby covering the solid part of the implant
and approximately one-half the thickness of the porous coating, was brought
into contact with the specimen. With use of a materials-testing machine (model
1321; Instron, Norwood, Massachusetts), the force that was required to push
the implant out of the bone was determined at a displacement rate of 0.25
mm/min. The force-displacement curves were recorded, and the ultimate force
was normalized to the nominal outer surface area of the implant to determine
the strength of fixation. The slope of the linear portion of the
load-displacement curve was calculated to determine the interface stiffness,
and the area under the curve to the point of failure was assessed to determine
the energy to failure.
The data were normally distributed. Analysis of variance was used to test
for differences in the contralateral (control) variables between groups to
determine if the presence of the hydroxyapatite-tricalcium phosphate coating
had effects on the end points. Repeated-measures analyses of variance with
side (treatment or contralateral control) as the within-subjects factor and
group as the between-subjects factor were performed, followed by post hoc
tests. When significant main effects (side or group) or interaction terms
(group-by-side) were found, paired and Student t tests were used to determine
which specific differences were significant. Pearson product-moment bivariate
correlations and stepwise multiple linear regressions were used to determine
which morphological end points explained variance in the mechanical end
points. For these analyses, the correlations were calculated with use of the
treated-side values from all four groups. For all tests, the level of
significance was set at p < 0.05.
Effect of Hydroxyapatite-Tricalcium Phosphate Coating
There were no differences among the contralateral (control) sites in the
three growth-factor-treatment groups (p > 0.05 for each variable; analysis
of variance). Therefore, these twenty-one implants were lumped together and
were compared with the contralateral (control) implants in the
autogenous-bone-graft group. No differences were found in any of the
mechanical or morphological end points (p > 0.05, Student t tests), except
for trabecular-specific surface in the region of interest located 2 mm medial
to the implantation site (p = 0.026). Given that eighteen variables were
compared, it is not surprising that one was significant at the 0.05 level.
These analyses indicate that the presence of the hydroxyapatite-tricalcium
phosphate coating did not affect implant fixation mechanics, bone ingrowth,
bone regeneration in the gap, or bone volume in the region of interest 2 mm
medial to the implantation site.
Effect of Treatments on Fixation Mechanics
Fixation strength on the treatment side was significantly greater than that
on the contralateral (control) side in each group
(Fig. 3, A). Analysis
of variance showed that side (p < 0.001), group (p = 0.009), and the
group-by-side interaction (p = 0.002) were all significant. Fixation strength
following combined treatment with both TGF-ß2 and BMP-2 was greater than
that following treatment with TGF-ß2 or BMP-2 alone (p < 0.05).
Specifically, combined growth factor treatment led to a 5.7-fold increase in
fixation strength compared with the value for the contralateral (control) side
whereas TGF-ß2 and BMP-2 alone led to 2.9-fold and 2.3-fold increases,
respectively. In addition, the 4.5-fold increase in fixation strength in the
autograft group was greater than the increase in the BMP-2 group (p <
0.05).
Interface stiffness was significantly greater when the treatment side was
compared with the contralateral (control) side in each group
(Fig. 3, B). Analysis
of variance showed that side, group, and the group-by-side interaction were
all significant (p < 0.001). Interface stiffness following combined
treatment with both TGF-ß2 and BMP-2 was greater than that following
treatment with TGF-ß2 or BMP-2 alone (p < 0.05). Specifically,
combined growth factor treatment led to a 5.7-fold increase in interface
stiffness compared with the value for the contralateral (control) side whereas
the use of TGF-ß2 and BMP-2 alone led to 3.1-fold and 3.2-fold increases,
respectively. In addition, the 6.4-fold increase in interface stiffness in the
autograft group was greater than increases in the BMP-2 group and the
TGF-ß2 group (p < 0.05).
Energy to failure was greater on the treated side than on the contralateral
(control) side when TGF-ß2 was used alone or in combination with BMP-2
and when autograft was used, but the use of BMP-2 alone was not stimulatory
(Fig. 3, C). Analysis
of variance showed that side was significant (p < 0.001), group was
marginally significant (p = 0.053), and the side-by-group interaction was
significant (p = 0.021) as there were large increases in only the TGF-ß
group and the combined growth factor group, with a more modest increase in the
autograft group and no increase in the BMP-2 group. Thus, although the
increase relative to the contralateral (control) side in the combined growth
factor group (4.5-fold) was greater than those in the TGF-ß2 group
(3.5-fold) and the autograft group (2.5-fold), these differences were not
significant, with the numbers studied.
Effect of Treatments on Bone Ingrowth
Bone ingrowth was significantly elevated in each group when the value for
the treatment side was compared with that for the contralateral (control) side
as well as when the value for each growth factor treatment was compared with
that for the autograft treatment (Fig. 4,
A). Analysis of variance indicated that side (p <
0.001) and group (p = 0.025) were significant but that the side-by-group
interaction was not significant (p = 0.113). Compared with the values for the
contralateral (control) side, the combined growth factor treatment led to a
3.6-fold increase whereas the use of TGF-ß2 and BMP-2 individually led to
2.3-fold and 1.7-fold increases, respectively.
All of the treated implants had bone ingrowth. Three implants on the
contralateral (control) side (including one in the TGF-ß group and two in
the autograft group) lacked bone ingrowth. On the contralateral (control)
side, bone ingrowth usually occurred over a restricted area of the implant
within the superficial aspect of the porous coating. In the growth factor and
autograft-treated implants, bone ingrowth was more consistent and usually
involved most of the circumference of the implant as well as the full depth of
the porous coating. This observation was most noticeable in the implants
treated with both TGF-ß2 and BMP-2. The bone took the form of woven or
lamellar trabeculae. There was no evidence of fibrocartilage or the existence
of an endochondral phase of bone formation.
Effect of Treatments on Bone Regeneration in the Gap
Bone volume in the gap (that is, the ratio of bone volume to total volume)
was greater on the treated side than on the contralateral (control) side in
the TGF-ß2, TGF-ß2 and BMP-2, and autograft groups
(Fig. 4, B). Analysis
of variance indicated that side was significant (p < 0.001), that group was
marginally significant (p = 0.079), and that the side-by-group interaction was
significant (p = 0.001) because of the relatively large effect in the
autograft group. For the autograft group, the measurement of bone volume
included remnant graft particles as well as newly formed bone. We estimate
that the actual amount of newly formed bone constituted approximately one-half
of the total bone volume and, therefore, would be slightly less than new-bone
formation in the gap in the growth factor groups. Compared with the values for
the contralateral (control) side, combined growth factor treatment led to a
2.1-fold increase in bone volume in the gap whereas the use of TGF-ß2 and
BMP-2 individually led to 1.7-fold and 1.2-fold increases, respectively.
Morphometric analysis of the trabecular bone within the 3-mm gap showed
elevations in trabecular number compared with the value for the contralateral
(control) side in the TGF-ß2 group, the combined growth factor group, and
the autograft group (Fig. 4,
C); increased trabecular thickness in all treatment
groups (Fig. 4, D);
decreased trabecular spacing in the combined growth factor and autograft
treatment groups (Fig. 4,
E); and decreased trabecular bone-specific surface in the
TGF-ß2 group, the combined growth factor group, and the autograft group
(Fig. 4, F). The only
difference among the groups for the gap trabecular architecture variables was
that trabecular number was greater in the autograft group than in the BMP-2
group (Fig. 4, C).
Bone in the gap and porous coating usually was in continuity in the treated
implants, particularly in the combined growth factor group, whereas there was
often incomplete bone growth across the gap in the controls
(Fig. 5). In the control
specimens, these areas of incomplete bone growth typically consisted of loose
connective tissue with the fibers oriented parallel to the interface. These
membranes were usually <0.5 mm thick, although they were occasionally as
much as 2 mm thick. In some instances, there was bone ingrowth immediately
deep to these interface membranes.
Effect of Treatments on Bone Volume in the Area Located 2 mm Medial
to the Test Site
Bone volume was elevated in the combined growth factor and autograft
treatment groups in the region of interest located 2 mm medial from the
surgical site compared with the values for the contralateral (control) side
(Fig. 6). Analysis of variance
showed significant side (p < 0.001), group (p = 0.011), and side-by-group
(p < 0.001) effects.
In all groups, there was histologic evidence of ongoing or recent bone
formation on existing bone surfaces in the area located 2 mm outside the
surgical margin (Fig. 7,
A). Evidence of bone resorption in these areas was not
common. In the combined growth factor treatment group, there were often also
bridging woven bone trabeculae between more mature, presumably preexisting
trabeculae (Fig. 7,
B). Although these bridging trabeculae occasionally were
found in association with some of the other treatments, they were not nearly
as common as in the combined growth factor treatment group.
Relationships Between Implant Fixation, Bone Ingrowth, Bone Volume,
and Architecture
Bivariate correlations showed that the single variable most strongly
correlated with fixation strength was bone volume in the 3-mm gap (r = 0.677,
p < 0.01) (see Appendix), that the single variable most strongly correlated
with interface stiffness was trabecular spacing in the region of interest
located 2 mm from the implantation site (r = -0.607, p < 0.01), and that
the single variable most strongly correlated with energy to failure was bone
volume in the region of interest located 2 mm from the implantation site (r =
0.412, p < 0.05). None of the mechanical parameters were correlated with
bone ingrowth. When fixation strength was plotted as a function of bone volume
in the gap, the data points for the combined growth factor treatment group
fell above the regression line, suggesting that for any given level of
new-bone formation, the combination treatment led to stronger fixation than
did treatments with the individual growth factors or with autograft
(Fig. 8). In the stepwise
multiple-regression analyses, the adjusted r2 values indicated that
the total variance explained was 63.3% for fixation strength, 48.0% for
interface stiffness, and 37.5% for energy to failure
(Table I). The variables that
entered the equation included parameters quantifying trabecular bone volume
and architecture in the gap and the adjacent region of interest, but not bone
ingrowth.
Use of either rhTGF-ß2 or rhBMP-2 in isolation stimulated implant
fixation, but the degree of stimulation for two of the three mechanical end
points was greater when the growth factors were used in combination than when
they were used alone. Specifically, local combined application of
rhTGF-ß2 and rhBMP-2 had additive positive effects on fixation strength
and interface stiffness: the combined treatment led to a 5.7-fold increase
compared with the values for the contralateral (control) side, whereas
individual treatments led to 2.3-fold to 3.2-fold increases. TGF-ß2,
alone and in combination with BMP-2, also enhanced energy to failure compared
with the values for the contralateral (control) side. Mechanical fixation in
the autograft group was also increased compared with the value for the
contralateral (control) side, but to a slightly lesser degree than was found
in association with the combination growth factor treatment. Thus, the
combined use of TGF-ß2 and BMP-2 at the early time-point studied
stimulated mechanical fixation of a porous-coated implant to a degree at least
equivalent to that achieved with autogenous bone graft.
A potentially confounding factor in the experimental design was that all of
the treated and contralateral (control) implants in the growth factor groups
were coated with hydroxyapatite-tricalcium phosphate whereas none of the
implants in the autogenous bone graft group were coated with
hydroxyapatite-tricalcium phosphate. The hydroxyapatite-tricalcium phosphate
coating was used as a delivery vehicle for the growth factor treatments.
Consistent with our previous
study16, we found
that the hydroxyapatite-tricalcium phosphate coating had no effect on any of
the mechanical end points as evidenced by comparison of the contralateral
(control) implants. As reported
previously16, it is
likely that the large gap size (3 mm) accounted for this lack of an effect
because studies by other
investigators13
have indicated that calcium phosphate coatings on implants are only effective
in situations in which the gap size is not >1 mm.
Mechanical fixation of the implant is an important consideration because it
is a marker of the functional competence of the regenerated bone. Increased
fixation strength in the twofold to threefold range has been previously
reported for
BMP-729,30
and TGF-ß120,
consistent with the current findings for TGF-ß2 and BMP-2 when used
alone. One study demonstrated a sixfold increase when the use of BMP-7
combined with hydroxyapatite granules was compared with the use of the
hydroxyapatite granules alone, although that same study demonstrated a slight
reduction when the use of BMP-7 combined with allograft was compared with the
use of allograft alone because the allograft was itself probably
stimulatory31. We
found a 5.7-fold increase in fixation strength when the combined use of
TGF-ß2 and BMP-2 was compared with no treatment. There also have been
reports in which local application of
BMP-729-31,51,
TGF-ß248,
BMP-227,
TGF-ß118,19,
and PDGF34 led to
no significant change in fixation strength.
The models in which growth factors have been tested to enhance implant
fixation have varied in terms of the species studied (rat, rabbit, sheep,
dog), the delivery method (direct adsorption onto the implant or delivery by
means of material placed in a gap surrounding the implant, either as a
collagen-based device or as an admixture with allograft or ceramic particles),
the nature of the bone-implant interface at the time of surgery, and the
choice of the control for comparison. All of these studies have been
short-term studies, with most involving a time-period of six weeks or less.
The studies that have shown positive effects on implant fixation have involved
models in which there was a substantial challenge (for example, a controlled
revision model in which gaps and a less-than-ideal osteogenic milieu can be
assumed30) or have
been studies such as the present one in which large initial interface gaps
were
used20,31.
In the one study in which motion was
present19, the
growth factor had no effect. Given the variability in the models and the
likelihood that other important considerations such as dose and the timing of
delivery may be critical but not easy to control, negative or neutral findings
associated with any particular growth factor need to be interpreted with
caution before it is concluded that a given growth factor lacks utility.
Consistent with the novel finding that rhTGF-ß2 and rhBMP-2 act
additively to enhance implant fixation, TGF-ß1 previously was shown to
act synergistically with BMP-7 in the induction of endochondral bone formation
in baboons45 and
the combined use of TGF-ß3 and BMP-2 was reported to lead to more ectopic
bone formation in an in vivo mouse
model46. The
mechanistic basis for these observations remains to be determined. TGF-ß
and BMP work through related but separate signal transduction pathways, and
TGF-ß is typically thought to control osteoprogenitor cell proliferation
whereas BMPs are more important in osteoblast
differentiation52,53.
It is likely that the stimulatory effects of these two growth factors are not
identical and that one may potentiate the activity of the other because of
possible interdependent regulation of
activity54. Another
possible mechanism of interaction is that BMP has been reported to stimulate
bone resorption22,
and TGF-ß, in addition to its anabolic effects, may also have
anticatabolic
effects55.
Previous studies involving the same canine model (and unpublished data)
have demonstrated that the optimum dose was approximately 12 µg for
TGF-ß2 and 25 µg for
BMP-216,22
when hydroxyapatite-tricalcium phosphate-treated porous-coated implants were
used in the presence of a 3-mm gap. What remains to be explored is whether
suboptimal doses used in combination would also be effective. It may be
important to use the lowest doses possible because high doses of TGF-ß
can inhibit
mineralization56-58
whereas high doses of BMP-2 have been associated with stimulation of bone
resorption22,59,60.
BMPs induce ectopic bone formation through an endochondral
pathway61 but, in
the present study, there was no evidence of chondrogenic activity in animals
that received rhBMP-2. This finding is consistent with those of our previous
studies16,22,
supporting the concept that the enhancement of implant fixation occurs in
association with the promotion of the intramembranous bone formation pathway.
Nevertheless, it is possible that, because of the timing of the observation
(four weeks), there may have been an endochondral phase that was missed. More
detailed time-course studies would be needed to definitively address this
issue.
In our previous studies with rhBMP-2 and rhTGF-ß2, we observed effects
on bone regeneration at the contralateral implantation site, especially in
terms of bone ingrowth and bone formation in the
gap16,22.
In the present study, that effect was not readily apparent for the mechanical
variables, and, at most, was only marginally present for bone ingrowth and
bone volume in the gap in association with use of BMP-2. In addition, it was
only in the combined growth factor treatment and autograft treatment groups
that we observed an increase in bone volume in the ipsilateral bone in the
region of interest located 2 mm medial to the implantation site.
While all treatments led to enhanced bone ingrowth compared with that on
the contralateral (control) side, each of the growth factor treatments led to
more bone ingrowth than autogenous bone-grafting did. Nevertheless, bone
ingrowth was not correlated with the mechanical end points, possibly because
the variation in bone ingrowth among the four groups was relatively small
compared with differences in some of the other morphometric end points.
For the combined growth factor treatment, the interface mechanical end
points tended to have higher values than those predicted by the independent
variables entered into the regression models. In addition, the presence of an
additive effect on interface mechanics in the absence of additive effects on
bone ingrowth and the quantity and morphology of the bone that formed in the
3-mm gap suggest that other qualities of the regenerated bone that were not
considered in the present study may help to explain the mechanical findings.
This implication is reinforced by the regression analyses, which showed that
no more than approximately 63% of the variance in interface mechanics was
accounted for by variance in the morphological end points, implying that
nearly 40% of the variance remains to be explained. Although some of this
unexplained variance is related to experimental error, other qualities of the
regenerated bone, such as the three-dimensional architecture (as opposed to
the two-dimensional architecture examined in the present study), the degree of
mineralization, mineral crystallinity, and collagen maturity merit
study56,62.
One of the issues requiring attention in revision total hip replacement is
loss of bone
stock11. Even with
modern cementing techniques, the failure rates for revision total hip
replacement are still relatively high—ranging from 10% to 25%—and
the results of repeat revisions with use of cement are
worse10,11.
These poor results of revision of failed total hip replacement with use of
cement have stimulated the use of cementless revision total hip replacement,
but there is often the need for bone
graft11. None of
the studies that have been reported to date have compared the findings of
growth factor treatment with those of autogenous bone-grafting. The present
study, therefore, provides the first evidence that growth factor treatment of
an implant in the presence of a large interface gap provides an early
mechanical result that is equivalent to that obtainable with autogenous
bone-grafting. The use of locally delivered growth factors, particularly
growth factor cocktails such as the one reported in the present study, may
represent a new method of enhancing bone regeneration in clinically
challenging situations such as those presented by revision joint
replacement.