Bone 42 (2008) 365 – 371

www.elsevier.com/locate/bone

Stainless steel screws coated with bisphosphonates gave stronger fixation and
more surrounding bone. Histomorphometry in rats
K. Wermelin a,⁎, F. Suska b , P. Tengvall c , P. Thomsen b , P. Aspenberg a
a
Section for Orthopaedics, Inst. of Clinical and Experimental Medicine, Faculty of Health Sciences, Linköping University, SE-581 85 Linköping, Sweden
b
Department of Biomaterials, Inst. of Surgical Sciences, Sahlgrenska Academy at Göteborg University, Box 412, SE-405 30 Göteborg, Sweden
c
Division of Applied Physics, Department of Physics and Measurement Technology, Biology and Chemistry, Linköping University, SE-581 83 Linköping, Sweden
Received 16 July 2007; revised 5 October 2007; accepted 10 October 2007
Available online 11 December 2007

Abstract

Coating of stainless steel screws with bisphosphonate in a fibrinogen matrix leads to an enhancement of the pullout strength 2 weeks after
insertion in rat tibiae. This effect then increases over time until at least 8 weeks. The pullout force reflects the mechanical properties of the bone
within the threads, which acts as a screw nut. The aim of the present study was to find descriptive and morphometric histological correlates to the
increased pullout strength. Because the bisphosphonates are applied via the implant surface, we also measured bone to implant contact and how far
away from the surface any effects could be seen.
Stainless steel screws underwent one of three treatments: uncoated control, controls coated with a layer of cross-linked fibrinogen, or screws
further modified with bisphosphonates covalently linked and physically adsorbed to the fibrinogen layer. At 1 (n = 33) and 8 (n = 27) weeks, bone
to implant contact and bone area density in the threads were measured, as well as bone area density at 250 and 500 μm from the outer edge of the
threads. Additionally, removal torque for each screw treatment was measured at 2 weeks (n = 28).
At 8 weeks, the part of the bisphosphonate screw that was located in the marrow cavity had become surrounded with bone, whereas there was
almost no bone surrounding the controls. The bone area density in the threads along the entire bisphosphonate screw was increased by 40%
compared with uncoated controls, and at 250 μm distance it was more than doubled. At 1 week, coated screws had less implant–bone contact, but
at 8 weeks there was no difference between uncoated and bisphosphonate-coated screws. The bisphosphonate screws had 50% increased removal
torque at 2 weeks compared to uncoated screws. Howship's lacunae and osteoclasts were found near the screws with bisphosphonates at 8 weeks,
suggesting that some bone remodeling took place near the implant, in spite of the presence of bisphosphonates.
© 2007 Elsevier Inc. All rights reserved.

Keywords: Histomorphometry; Stainless steel; Torque removal; Bisphosphonate; Implant fixation

Introduction Implant insertion in cancellous bone leads to an inflamma-

Early rigid fixation of an implant to bone is important for
its long-term results. Hip or knee replacements that show
increased early migration relative to the bone during the first
one or two years have a dramatically increased risk of
loosening, which often becomes apparent first several years
later [1–3].
⁎ Corresponding author. Section for Orthopaedics, Inst. of Clinical and
Experimental Medicine, Faculty of Health Sciences, Linköping University, SE-
581 85 Linköping, Sweden. Fax: +46 13 224434.
E-mail address: karwe@inr.liu.se (K. Wermelin).
8756-3282/$ - see front matter © 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.bone.2007.10.013
tory response, followed by release of cytokines and growth
factors which eventually stimulate differentiation of bone
forming cells [4]. These cells produce primitive woven bone,
which is then remodeled to lamellar bone. However, it appears
that resorption and formation during this remodeling are not
always coupled. Consider fracture repair: not only can a bone
callus grow as a result of predominant formation, but it can also
diminish as a result of predominant resorption. This can be
observed in almost all fractures. Similarly, the timing of the
dominance of formation or resorption around an implant may
partly determine how much bone will surround it. Thus, if
resorption can be reduced or delayed, formation may still be
increased, leading to a net increase in the amount of bone.
366 K. Wermelin et al. / Bone 42 (2008) 365–371

Indeed, the early migration of knee prostheses can be reduced if
bisphosphonates are given either systemically or applied di-
rectly to the bone surface prior to insertion of the prostheses
[5,6].
We have previously applied bisphosphonates to the surface
of stainless steel screws, implanted them in rat bone and found
that the pullout strength gradually increased for up to 8 weeks
[7,8]. The pullout force reflects the macromechanical properties
of the bone within the threads, which can be said to act as a
screw nut. The mechanical results call for a histological exam-
ination to clarify their structural background. We hypothesized
that the previously observed gradual increase in fixation was
due to an increasing amount of bone forming around the screw.
Secondarily, we wanted to describe how far away from the
implant bisphosphonate-induced changes could be seen and if
there was an effect on bone implant contact. Finally, we studied
if the bisphosphonates influenced removal torque, as a me-
chanical counterpart to bone implant contact.
Materials and methods
Surgical equipment was sterilized in an autoclave. Sterile gloves, theatre
caps, gowns and surgical masks were used. The rats were anesthetized with
isoflourane gas. Each rat received a subcutaneous injection of 7 mg oxy-
tetracycline as infection prophylaxis and 0.05 mg of buprenorphine for
postoperative pain relief. One of the legs was shaved and cleaned with chloro-
hexidine alcohol. The rat was put in a sterile surgical glove, a hole was cut out of
the glove, through which the shaved leg was pulled out. Sterile tape was wrapped
around the paw and the leg cleaned once more with chloro-hexidine alcohol.
A 5–6 mm longitudinal incision was made along the medial aspect of the rat
tibia. The periosteum was reflected dorsally to the physis. A hole, 1.3 mm in
diameter, was drilled with an air driven drill, during continuous supply of sa-
line solution, at a distance from the proximal tibial metaphysis of approximately
9–16 mm in the Morphometric study and 2–5 mm in the Removal torque study.
The more proximal site was chosen to ascertain that the tip of the screw would not
engage the opposite cortex. Tapping was not required. A screw was inserted and
the skin was closed over the screw.
In the Removal torque study a silicone membrane gasket (Silastic®, Dow
Corning, Midland, Michigan, USA) was placed between the under-side of the
screw head and the bone, to minimize friction from parts other than the threads
during torque testing.
Treatment was randomized by lottery. The treatments were randomly named
A, B and C. Each rat was assigned to a treatment by lot, drawn at surgery.
Because all screws looked alike, there was no possibility to discern the dif-
ferently treated screws from each other.

The fixation of stainless steel screws was studied in 2 experiments; Mor- Histological preparation and evaluation
phophometric study and Removal torque study. For both of these studies, the
preparation of the bisphosphonate-coated screws was executed according to the
Immediately after the rats were killed (after 1 or 8 weeks), the proximal
same protocol. A multilayer of fibrinogen was covalently bound to the screws
section of the tibia with the screw was removed en bloc, fixed in formaldehyde
followed by the binding of two types of bisphosphonates, one covalently linked, (4%) and dehydrated in a graded series of ethanol.
and one adsorbed. In order to estimate the coated film thickness on the screws, All persons handling or evaluating the specimens were blinded for treatment
flat silicon surfaces were treated according to the same protocol (except for
until all data were collected and finalized for statistical analysis.
hydrogen fluoride etching) and organic film thicknesses measured by null The specimens were then embedded in a plastic resin (LR White, The
ellipsometry. The surgical procedure was also the same, except for the insertion London Resin Co Ltd, Hampshire, UK). The calcified bone with the remaining
site as determined by the distance from the physis. screw was divided by sawing (Exact cutting and grinding equipment, Exact
Apparate Bau, Nordenstedt, Germany) using a diamond blade into sections
Stainless steel screw preparation parallel to the long axis of the tibia and along the axis of the screw. Ground
sections approximately 15–20 μm thick were prepared and stained with 1%
Stainless steel screws, measuring 1.7 mm in diameter (type M 1.7) and 3 mm
in length, were used. The screws were designed with a hole in the head, so that
they could be fastened to a hook in a materials testing machine. This type of
screw has been used in several previous experiments [7–9]. The screws were
etched in 40% hydrogen fluoride and cleaned as previously described [7].
Ten layers of fibrinogen (human plasminogen free fibrinogen, 10 mg/ml
dissolved in distilled water, Haemchrom Diagnostica, Sweden) were coated by
EDS/NHS coupling techniques, followed by chemical binding of pamidronate
(pamidronate dissolved in distilled water, 1 mg/ml, Aredia®, Novartis, Sweden)
into the cross-linked fibrinogen film by the same technique. Finally, ibandronate
was weakly bound onto the screws during an over night incubation (ibandronate
dissolved in water, 50 μg/ml, Bondronate®, Roche, Switzerland). This pro-
cedure was used to maximize the bisphosphonate uptake. Fibrinogen control
screws were put in distilled water, during incubations in bisphosphonate solu-
tions, until the preparation of all screws was finished.
The thickness of the coating was measured with an ellipsometer as pre-
viously described [7]. Measurements were performed on four silica chips
prepared in parallel with the screws, and data were collected from at least 10
spots on each sample. The amount of pamidronate and ibandronate was ap-
proximately 530 ng/cm2 (44 Å) and 380 ng/cm2 (32 Å).

Animal experiments

The number of rats was 69 in the Morphometric study (35 for 1 week and 34
for 8 weeks) and 30 in the Removal torque study. Male Sprague–Dawley (SD)
rats with a mean body weight of 415 g (ranging from 314 g to 491 g) were used
(age approximately 10–11 weeks). The animals were kept two per cage and Fig. 1. Areas for measurement of bone area density. a. Triangles for mea-
given free access to rat-chow and water. Institutional guidelines for the care and surement of bone density in screw thread. b. Rectangles for measurement of
treatment of laboratory animals were followed and the study had been approved bone density at 250 μm distance. c. Rectangles for density at 500 μm distance.
by the regional ethics board. All measurements were made on both sides of the screw.
 K. Wermelin et al. / Bone 42 (2008) 365–371
Fig. 2. Primitive bone or osteoid (O) and a fragment of old bone (B) in the
marrow cavity at 1 week around a bisphosphonate screw.
Toluidine blue. The interface and the surrounding bone tissue were studied in
light microscopy (Nikon Eclipse E600, Nikon Instruments INC, Melville, USA)
equipped with Image Analysis System (IE 2000, TeknoOptik AB, Sweden).
Bone to implant contact (percent) was measured as the relation of the
perimeter of direct bone to screw contact to total perimeter of the screw in the
bone compartment. Four threads on each side of the screw were used for this
measurement (Fig. 1).
Bone area density was measured within the triangles formed by the threads.
Four triangles were measured on each side of the screw. The bone tissue outside
the threads was divided in rectangles. One side in each rectangle was defined by
Fig. 3. Uncoated control (left) and bisphosphonate screw (right). A sleeve of new bone can be seen around the bisphosphonate screw in the marrow cavity (dotted line).
367
Fig. 4. A multinucleated cell in a resorption lacune (a), suggestive on on-going
bone resorption adjacent to a bisphosphonate-coated screw. There is also a
multinucleated cell with lobulated nuclei (b), probably an apoptotic osteoclast.
the line between two adjacent thread edges, which measured approximately
280 μm. The other side of the rectangle was 250 μm. Four such rectangles were
measured on each side of the screw. A second series of rectangles was then
placed outside of the first (extending from 250 μm to 500 μm). Bone area
density (percent) was measured as amount of bone area per total area inside the
threads and the rectangles respectively (Fig. 1). Bone fragments produced by the
drilling were included in the bone area.
Measurements were made on the first four threads, counted from the screw
head, and on each side of the implant. Hence, there were 8 measurements from
within the thread triangles, 8 from the first line of rectangles, and 8 from the
second line. The mean of the 8 areas was calculated to produce a single value for
statistical analysis.
After histomorphometry was finalized and statistically analyzed, another
blinded investigator (PA) tried to identify the 9 specimens from the 8 weeks time
point that were treated with bisphosphonates, based on the reported quantitative
differences. Qualitative histology was done after that.
Biomechanical evaluation
In the Removal torque study rats were killed after 2 weeks and the tibiae
harvested. The screws in the tibiae were subjected to a torque moment test in a
computerized materials testing machine [10] (100 R, DDL Inc., Eden Praire,
368 K. Wermelin et al. / Bone 42 (2008) 365–371

Table 1B
P-values for morphometry at 1 week
Bone Bone area Bone area Bone area
contact, density, density, density,
thread (%) thread (%) 250 μm (%) 500 μm (%)
1-way ANOVA 0.00 0.25 0.92 0.86
Uncoated vs fibrinogen 0.01 1.0 0.92 0.99
Uncoated vs 0.00 0.31 0.96 0.91
bisphosphonate
Fibrinogen vs 0.69 0.32 0.99 0.86
bisphosphonate
ANOVA, followed by Tukey's post hoc test.

At 8 weeks, both uncoated and fibrinogen coated control

Fig. 5. Osteoclast (black arrow) resorbing newly synthesized bone at 8 weeks
adjacent to a bisphosphonate screw in the medullary cavity.
MN, USA), at a speed of 0.1 mm/s [9]. Measuring was done within 20 min after
the animals were killed. All torque/time curves showed a gradual, almost linear
increase in torque until it reached a characteristic peak, after which the torque
moment decreased. This maximum value is regarded as torque moment at
failure. The energy is based on the area under the curve from sampling starting at
3.2 N mm until the torque had dropped to 10% of the maximal value or the screw
had rotated 90°.
Statistical analysis
Results from 1 and 8 weeks were analyzed separately using ANOVA with
treatment as an independent variable, followed by Tukey's post hoc test, by the
use of SPSS version 14.00.
Results
Animals
Seven rats died during anesthesia. In the Morphometric study,
two samples were destroyed in the polishing machine. In the
Removal torque study, one screw was misimplanted and the rat was
excluded at surgery. One screw was accidentally loosened before
measuring. The number of remaining rats is given in Table 3A.
Qualitative results
About one third of the length of the screw was located in
cortical bone, and the remaining part in the marrow cavity. At
1 week, scattered areas of primitive bone or osteoid were seen in
the marrow cavity, together with fragments of bone from the
drilling procedure (Fig. 2). This was similar for all implant types.
screws appeared denuded of bone in the marrow cavity, or were
partially covered with a thin layer of lamellar bone. The primitive
bone or osteoid seen in 1-week specimens was now absent in the
controls. In contrast, the bisphosphonate screws were surrounded
by a thick sleeve of bone, with a rather primitive, woven and
cancellous appearance (Fig. 3). Large multinucleated cells were
seen adjacent to Howship's lacunes in the cortex adjacent to the
screw (Fig. 4). This was also seen occasionally around the
bisphosphonate screws.
In a few bisphosphonate specimens, large multinucleated
cells were seen adjacent to Howship's lacunes in the new bone
in the marrow cavity. Some of these cells appeared to be re-
sorbing bone, whereas others were detached and had multi-
lobulated nuclei, suggestive of apoptosis (Fig. 4). Some areas of
lamellar bone suggest that remodeling had occurred also in the
marrow cavity around the bisphosphonate screws (Fig. 5).
The blinded investigator was able to pick out 8 of the 9
bisphosphonate specimens from the 8-week time point, based on
the information that they had a higher bone density around them.
Quantitative
Bone area density in screw thread
At 1 week there was no significant difference between the
three screw types regarding bone area density in the threads
(Tables 1A and 1B). In contrast, at 8 weeks the bone area
density in the threads was 40% higher for the bisphosphonate
screws compared to uncoated and fibrinogen coated ones
(Tables 2A and 2B).
Bone area density at 250 μm distance
At 1 week there was no significant difference between the
three screw types regarding bone area density at 0–250 μm
distance from the thread top (Tables 1A and 1B). At 8 weeks the

Table 1A
Morphometric results at 1 week
Bone contact, thread (%) Bone area density, thread (%) Bone area density, 250 μm (%) Bone area density, 500 μm (%)
n Mean SD n Mean SD n Mean SD n Mean SD
Uncoated 10 54 27 10 34 13 10 34 13 10 19 11
Fibrinogen 11 25 15 11 34 10 11 36 12 11 20 11
Bisphosphonate 12 18 17 12 27 10 12 36 12 12 18 6
For a definition of the areas of measurement, see Fig. 1.
 K. Wermelin et al. / Bone 42 (2008) 365–371 369

Table 2A Table 3A
Morphometric results at 8 weeks Removal torque at 2 weeks
Bone contact, Bone area Bone area Bone area Torque moment at Energy (N mm)
thread (%) density, density, density, failure (N mm)
thread (%) 250 μm (%) 500 μm (%)
n Mean SD n Mean SD
n Mean SD n Mean SD n Mean SD n Mean SD
Uncoated 10 67 20 10 21 8.1
Uncoated 9 69 15 8 45 12 8 25 16 8 17 16 Fibrinogen 9 44 20 9 19 8.1
Fibrinogen 9 35 11 9 45 10 9 34 14 9 23 13 Bisphosphonate 9 100 20 9 56 10
Bisphosphonate 9 65 13 9 63 8 9 55 20 9 35 18
For a definition of the areas of measurement, see Fig. 1.
The energy uptake until failure was 167% larger for bi-
sphosphonate screws compared to uncoated controls, and 180%
bone area density outside the bisphosphonate screws was larger than fibrinogen coated screws (Tables 3A and 3B).
more than doubled in comparison to the uncoated screws
(120% higher) and 62% higher than fibrinogen coated screws Discussion
(Tables 2A and 2B).
Several studies have shown an improved fixation of
Bone area density at 500 μm distance bisphosphonate-coated implants in bone [11–17]. Also an
At 1 and 8 weeks there was no significant difference between increased bone density at some distance from the implant has
the three screw types regarding bone area density at 250–500 μm been described [11,13,15,16]. The novelty of this study lies in
from the thread top (Tables 1A and 1B). However, at 8 weeks the method used for applying the bisphosphonates and our
there was a trend towards an increased bone area density for the finding of a sleeve of new bone around the bisphosphonate-
bisphosphonate screws compared to controls (Tables 2A and 2B). coated screws in the marrow cavity. This was morphometrically
confirmed by an increase in average bone area density sur-
Bone to implant contact rounding the bisphosphonate-coated screws. The increase in
At 1 week, the fibrinogen and bisphosphonate screws had bone area density extended at least 250 μm away from the
less bone contact in the thread compared to the uncoated screws thread top. Because bisphosphonates do not induce de novo
(Tables 1A and 1B). For the bisphosphonate screws, the bone bone formation, the bone sleeve likely had a precursor in all
contact was 66% less compared to the uncoated screws. The groups, which had been quickly resorbed in the absence of a
fibrinogen coated screws showed 54% less bone contact than bisphosphonate, but was preserved and served as a scaffold for
uncoated ones. more bone formation in the bisphosphonate group. This pre-
At 8 weeks, fibrinogen coated screws showed 49% less bone cursor was seen at 1 week as minute scattered primitive bone or
contact than uncoated stainless steel controls, but bispho- osteoid in the marrow cavity in all groups.
sphonate screws and uncoated controls did not differ (Tables 2A However, it cannot be excluded that the bisphosphonates
and 2B). also exerted an anti-apoptotic effect on the osteoblasts residing
close to the bisphosphonate-coated screws [18,19].
Removal torque The histomorphometric data include bone surrounding the
The removal torque increased until failure and was followed part of the screw, which was situated in dense cortical bone.
by lowered friction during the continued turning of the screws. Because the density of this part is very high, and likely did not
The removal torque at failure was 50% larger for bispho- change, the changes in the marrow cavity were in fact more
sphonate screws compared to uncoated controls and more than dramatic than the numbers show. This is demonstrated by the
doubled compared to the fibrinogen coated controls (Tables 3A ease at which it was possible for a blinded person to pick out the
and 3B). bisphosphonate specimens based on morphological criteria only.
Although the new-formed bone had a somewhat primitive
appearance, there were signs of at least some remodeling,
suggesting that the presence of bisphosphonate in the bone does
Table 2B not totally preclude osteoclastic activity. We have previously
P-values for morphometry at 8 weeks
Bone Bone area Bone area Bone area
Table 3B
contact, density, density, density,
P-values for removal torque
thread (%) thread (%) 250 μm (%) 500 μm (%)
Torque moment at failure Energy
1-way ANOVA 0.00 0.00 0.00 0.08
(N mm) (N mm)
Uncoated vs fibrinogen 0.00 1.0 0.50 0.75
Uncoated vs 0.83 0.01 0.00 0.08 1-way ANOVA 0.00 0.00
bisphosphonate Uncoated vs fibrinogen 0.05 0.85
Fibrinogen vs 0.00 0.00 0.04 0.26 Uncoated vs bisphosphonate 0.00 0.00
bisphosphonate Fibrinogen vs bisphosphonate 0.00 0.00
ANOVA, followed by Tukey's post hoc test. ANOVA, followed by Tukey's post hoc test.
370 K. Wermelin et al. / Bone 42 (2008) 365–371
shown that in pathological situations, the resorptive response
may be stronger than can be inhibited by moderate concentra-
tions of bisphosphonate. [20] A few apparent osteoclasts were
seen adjacent to the bisphosphonate-coated screws. Although
they looked apoptotic, they had lived long enough to make
a resorption lacuna, suggesting that bisphosphonates at the
amounts used in this study shorten osteoclast survival rather
than preclude their activity completely (Fig. 4).
Removal torque was measured at 2 weeks, to allow for
comparison with previous studies of pullout force [7,8]. These
previous studies had shown a greater pullout force for the
bisphosphonate screws, suggesting enhanced bone formation
or a stronger bone matrix around the screws. The increased
removal torque in the bisphosphonate group of the present study
indicates a higher friction at the interface, and hence an in-
creased bone contact.
The present results suggest that most of the previously found
improvement in mechanical fixation is due to the increase in the
amount of bone in the marrow cavity. This is not what we had
expected when the first experiments with this model were
planned. In cortical bone, implant fixation leads to resorption near
the implant surface, and thereby a decreasing early bone to
implant contact [21]. We originally planned to eliminate this
decrease by the use of bisphosphonates. Instead, we found an
increase in pullout force over several weeks [7], which likely can
be explained by the formation of a bone sleeve, as seen in the
present study. The finding of a net anabolic effect is yet another
indication that bone formation and resorption are not always
coupled. Moreover, this observation may open for an extended
clinical utility. For example, bisphosphonate-coated screws used
for fracture fixation in osteoporotic cancellous bone might
decrease the risk of cut out, due to the formation of a mechanically
stabilizing bone “sleeve”. One major limitation of this study,
however, is the small size of the screws, which means that an
improved bone quality less than a mm away from the implant
could improve fixation dramatically. For screw fixation to be
improved, it is likely that the bone quality must be enhanced at a
distance corresponding to the depth of the thread. With a larger
screw, therefore, there must be an effect of the bisphosphonate
coating at a larger distance. This may be difficult to achieve.
The fibrinogen coating on the stainless steel had a negative
effect on bone to implant contact. This was seen at 1 week re-
gardless of whether or not the screw also carried the bispho-
sphonates. At 8 weeks, however, it seems that the bisphosphonates
had compensated for this effect and no difference to uncoated
controls remained, whereas screws with only fibrinogen still
displayed less bone contact. However, the bone contact percentage
of the present study was remarkably high compared to other
similar studies [22]. It is possible that the surface etching pro-
cedure had a contact-promoting effect per se. If so, the effect of the
fibrinogen coating may not be the inhibition of processes neces-
sary for contact, but rather a masking of this beneficial surface
structure. Also, it remains to be found whether similar apparently
negative effects are exerted by other drug carrier coatings such as
polymers or other proteins. In an earlier study the release kinetics
of 14C-alendronate immobilized with our fibrinogen technique
showed a 60% release within 8 h in buffer [23].
The present study differs from previous studies regarding
bisphosphonate-coated implants [11–17] in two aspects. Previous
studies have used a calcium phosphate coating (mostly hydro-
xyapatite) for immobilization of the bisphosphonate. Hydro-
xyapatite by itself leads to a very strong attachment of the bone to
the implant surface. This does not necessarily imply better
fixation, because the fixation of a screw is dependent upon the
strength of the surrounding bone at a certain distance from the
surface (the bone “nut”), rather than on surface friction. However,
the high friction with a hydroxyapatite coating may preclude
screw removal. The other aspect, not previously reported is the
finding of a sleeve of new bone in the marrow cavity.
Acknowledgments
The authors thank Bibbi Mård for technical assistance. This
investigation was supported by the local Strategic research
projects Materials in Medicine (Östergötlands läns landsting,
Linköpings universitet), and the Swedish Research Council
(project 2031).
References
[1] Mjöberg B. The theory of early loosening of hip prostheses. Orthopedics
1997;20(12):1169–75.
[2] Ryd L. Roentgen stereophotogrammetric analysis of prosthetic fixation in
the hip and knee joint. Clin Orthop Relat Res 1992;276:56–65.
[3] Kärrholm 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 Jt Surg Br 1994;76(6):912–7.
[4] Einhorn TA. The cell and molecular biology of fracture healing. Clin
Orthop Relat Res 1998;355S:7–21.
[5] Hilding M, Ryd L, Toksvig-Larsen S, Aspenberg P. Clodronate prevents
prosthetic migration: a randomized radiostereometric study of 50 total knee
patients. Acta Orthop Scand 2000;71(6):553–7.
[6] Hilding M, Aspenberg P. Postoperative clodronate decreases prosthetic
migration: 4-year follow-up of a randomized radiostereometric study of 50
total knee patients. Acta Orthop 2006;77(6):912–6.
[7] Wermelin K, Tengvall P, Aspenberg P. Surface bound bisphosphonate
enhance screw fixation in rats-increasing effect up to 8 weeks after insertion.
Acta Orthop 2007;78(3):385–92.
[8] Tengvall P, Skoglund B, Askendal A, Aspenberg A. Surface immobilized
bisphosphonate improves stainless-steel screw fixation in rats. Biomater-
ials 2004;25(11):2133–8.
[9] Skoglund B, Holmertz J, Aspenberg P. Systemic and local ibandronate
enhance screw fixation. J Orthop Res 2004;22(5):1108–13.
[10] Skripitz R, Aspenberg P. Implant fixation enhanced by intermittent treatment
with parathyroid hormone. J Bone Jt Surg Br Vol 2001;83(3):437–40.
[11] Peter B, Pioletti DP, Laïb S, Bujoli B, Pilet P, Janvier P, et al. Calcium
phosphate drug delivery system: influence of local zoledronate release on
bone implant osteointegration. Bone 2005;36(1):52–60.
[12] Yoshinari M, Oda Y, Inoue T, Matsuzaka K, Shimono M. Bone response to
calcium phosphate-coated and bisphosphonate-immobilized titanium
implants. Biomaterials 2002;23(14):2870–85.
[13] Tanzer M, Karabasz D, Krygier JJ, Cohen R, Bobyn JD. Bone
augmentation around and within porous implants by local bisphosphonate
elution. Clin Orthop Relat Res 2005 Dec;441:30–9.
[14] Meraw SJ, Reeve CM, Wollan PC. Use of alendronate in peri-implant
defect regeneration. J Periodontol 1999;70(2):151–8.
[15] Meraw SJ, Reeve CM. Qualitative analysis of peripheral peri-implant bone
and influence of alendronate sodium on early bone regeneration. J Periodontol
1999;70(10):1228–33.
[16] Peter B, Gauthier O, Laib S, Bujoli B, Guicheux J, Janvier P, et al. Local
delivery of bisphosphonate from coated orthopedic implants increases
 K. Wermelin et al. / Bone 42 (2008) 365–371
implants mechanical stability in osteoporotic rats. J Biomed Mater Res A
2006;76(1):133–43.
[17] Kajiwara H, Yamaza T, Yoshinari M, Goto T, Iyama S, Atsuta I, et al. The
bisphosphonate pamidronate on the surface of titanium stimulates bone
formation around tibial implants in rats. Biomaterials 2005;26(6):581–7.
[18] Plotkin LI, Aquirre JI, Kousteni S, Manolagas SC, Bellido T. Bispho-
sphonates and estrogens inhibit osteocyte apoptosis via distinct molecular
mechanisms downstream of extracellular signal-regulated kinase activa-
tion. J Biol Chem 2005;280(8):7317–25.
[19] Plotkin LI, Weinstein RS, Parfitt AM, Roberson PK, Manolagas SC,
Bellido T. Prevention of osteocyte and osteoblast apoptosis by bispho-
sphonate and calcitonin. J Clin Invest 1999;104(10):1363–74.
371
[20] Åstrand J, Aspenberg P. Reduction of instability-induced bone resorption
using bisphosphonates: high doses are needed in rats. Acta Orthop Scand
2002;73(1):24–30.
[21] Dhert WJ, Thomsen P, Blomgren AK, Esposito M, Ericson LE, Verbout
AJ. Integration of press-fit implants in cortical bone: a study on interface
kinetics. J Biomed Mater Res 1998;41(4):574–83.
[22] Brånemark R, Öhrnell LO, Nilsson P, Thomsen P. Biomechanical
characterization of osseointegration during healing: an experimental in
vivo study in the rat. Biomaterials 1997;18(14):969–78.
[23] Wermelin K, Aspenberg P, Linderbäck P, Tengvall P, Bisphosphonate
coating on titanium screws increases the mechanical fixation in rat tibia
after two weeks. J Biomed Mater Res A. Article in press.