Biomaterials 21 (2000) 1291}1298

Osteoconduction at porous hydroxyapatite with various

pore con"gurations
Bong-Soon Chang , Choon-Ki Lee *, Kug-Sun Hong
, Hyuk-Joon Youn
,
Hyun-Seung Ryu
, Sung-Soo Chung, Kun-Woo Park
Department of Orthopedic Surgery, Seoul National University College of Medicine, Clinical Research Institute, Seoul National University Hospital,
Yungun-dong 28, Chongno-Ku, Seoul 110-744, South Korea

College of Material Science and Engineering, Seoul National University, Shinrim-dong, San 56-1, Kwanack-Ku, Seoul 151-742, South Korea
Department of Orthopedic Surgery, Samsung Medical Center, Sung Kyun Kwan University, Ilwon-dong 50, Kangnam-Ku, Seoul 135-230, South Korea
Received 7 September 1999; accepted 17 January 2000

Abstract

To assess the histological response and the reinforcing e!ects of bone ingrowth within porous hydroxyapatite (HA) implants
depending on pore geometry, four kinds of cylindrical-type with parallel linear pores (
50, 100, 300, 500 lm), one kind of sponge-type
with irregular interconnecting pores (
250 lm) and one cross-type with crossing linear pores (
100;120 lm) of porous HA were
prepared. Eighty-four rabbits were divided into six groups, and a 5;5;7 mm sized porous HA block was inserted through the
medial cortical window of the proximal tibia. Histomorphological changes were examined using light and scanning electron
microscopy. A biomechanical compression test was performed using material test machines. After implantation, the implants showed
di!erent histological changes depending on pore geometry. Active osteoconduction was also found in the
50 lm sized cylindrical-
type porous HA. Evidence of remodeling of new bone and bone marrow formation within porous HA was found in the larger
cylindrical-types (
300, 500 lm), and the sponge- and cross-types. The biomechanical test showed that the ultimate compressive
strength increased signi"cantly in the
300 lm sized cylindrical-type, and in the sponge- and cross-types eight weeks after
implantation. Porous HA with cylindrical pores could be a useful graft material due to its strength, osteoconductivity and the ease
with which its pore geometry can be controlled.  2000 Elsevier Science Ltd. All rights reserved.

Keywords: Hydroxyapatite; Pore; Geometry; Osteoconduction

1. Introduction
100 lm was necessary for bone ingrowth into the
Hydroxyapatite (Ca (PO ) (OH) ), with its high bio-
  
compatibility and good bioa$nity, stimulates osteocon-
duction and is slowly replaced by the host bone after
implantation [1}11]. The material factors a!ecting the
biological response to the implant in vivo include the
property of HA powder, sintering property, structural
con"guration and the presence of secondary phases
[12}16]. The structural property of porous HA is more
resorbable and more osteoconductive than dense HA,
and it has been used as arti"cial bone graft material in
many experimental and clinical trials [6,9,17]. In these
studies, it was claimed that a minimum pore size of
* Corresponding author. Tel.: 82-2-760-2336; fax: 82-2-764-2718.
E-mail address: choonki@plaza.snu.ac.kr (C.K. Lee).
porous implant materials. However, most of these
conclusions were based on studies using implants of
random pore geometry. Porous HA with random pore
geometry (coral, naphthalene, polymer bead and other
synthetic HA) has wide ranges of porosity and pores
of various sizes with much smaller interconnecting
fenestration [4,18,19]. It is suspected that the size of
interconnection is the main limiting factor of osteocon-
duction rather than the size of the pores themselves.
Therefore, there has been no consensus regarding the
optimal conditions for osteoconduction, such as pore
size, shape, interconnection and the arrangement of pores
[2,15,16,18].
In this study, we developed a porous HA with
parallel cylindrical pores of various sizes without
interconnecting fenestration between adjacent pores,
in order to evaluate the actual e!ect of pore size on

0142-9612/00/$ - see front matter  2000 Elsevier Science Ltd. All rights reserved.
PII: S 0 1 4 2 - 9 6 1 2 ( 0 0 ) 0 0 0 3 0 - 2
1292 B.-S. Chang et al. / Biomaterials 21 (2000) 1291}1298
osteoconduction. We also developed three types of por-
ous HAs (cylindrical-, sponge- and cross-types) in order
to evaluate the e!ects of interconnection and the pore
arrangement.
2. Materials and methods
2.1. HA powder
Stoichiometric HA powder (Ca/P ratio 1.67) was syn-
thesized by solid state reaction using CaHPO ) 2H O
 
(EP, Junsei Chemical, Tokyo, Japan) and CaCO (EP,

Junsei Chemical, Tokyo, Japan). CaHPO ) 2H O was
 
ignited at 11003C to be converted to Ca P O . The
  
Ca/P ratio 1.67 of Ca P O and CaCO was mixed
   
by ball-milling in water using zirconia media and cal-
cined at 11003C. Its composition was determined by
X-ray di!raction phase analysis (Model M18XHF-
SRA, Mac Sci., Japan). It was con"rmed to be
stoichiometric HA.
2.2. Porous HA block
Four kinds of cylindrical-type porous HA were pre-
pared. After coating with HA slurry, sheets of polyester
"bers (
50, 100, 300, 500 lm sized) were compacted into
block shape for porous HA with parallel cylindrical
pores without interconnecting fenestration in between
them. They were then burned out at 8003C for 24 h and
sintered at 13003C for 2 h. Using polyurethane foam with
340 lm sized pores, sponge-type porous HA was syn-
thesized. For the cross-type porous HA with intermedi-
ate cylindrical- and sponge-type interconnections,
a polypropylene net (
100;120 lm) coated with HA toxylin and eosin. The specimen refrigerated in 2.5%
slurry was used. The actual pore size of the prepared
porous HA was analyzed using light microscopy and
scanning electron microscopy (SEM, Model XL-20, Phi-
lips, The Netherlands) of a cross-section of the block. The
density of the sintered porous HA block was calculated
from measured weight and dimensions and the relative
density was determined by the ratio of the measured
density over the theoretical density of HA (3.156 g/cm).
The total porosity of porous HA is de"ned by the follow-
ing equation:
Total porosity (%)"100 (%)!relative density (%).
The macroporosity of cylindrical-type porous HA was
evaluated from the ratio of areas of pores, produced by
the burned polyester "bers, to total area in light micro-
graphs made on at least 10 "elds of cross-section. The
microporosity is
Microporosity (%)
"Total porosity (%)!Macroporosity (%).
Rectangular blocks (5;5;7 mm) of porous HA were
prepared and sterilized conventionally with ethylene ox-
ide gas (553C, 12 h).
2.3. Surgical model
Eighty-four New Zealand white rabbits weighing
about 3.0 kg each were divided into six groups according
to the type of porous HA implants used. The study was
approved by the Standing Committee on Ethics at the
Laboratory for Animal Research of the Clinical Research
Institute, Seoul National University Hospital. A 2 cm
longitudinal incision was made on the left leg just distal
to the tibial tuberosity. The periosteum was retracted,
and the medial aspect of the proximal metaphysis was
exposed. A cortical window (5;5 mm) was made on the
proximal tibia. A 5;5;7 mm sized porous block was
inserted into the corticocancellous defect. Four rabbits
from each group were sacri"ced for light microscopic
examination at four weeks after implantation, and the
rest ("ve for biomechanical testing, four for light micro-
scopic examination, and one for SEM examination) were
sacri"ced at eight weeks.
2.4. Histologic study
The specimens were "xed in 70% alcohol, and then
embedded in methyl methacrylate. Sections 10 lm in
thickness parallel to the longitudinal axis, were stained
with Goldner's modi"ed Masson trichrome. For the
cross-sectional analysis of bone ingrowth and bonding
status within pores, another specimen was "xed in for-
malin, decalci"ed and embedded in para$n. Transverse
sections 4 lm in thickness, were stained with hema-
glutaraldehyde, was washed in an ultrasonic cleaner,
dehydrated in 95% alcohol and examined using a scann-
ing electron microscope.
2.5. Biomechanical study
Five specimens from each group, harvested eight
weeks post-implantation, were stored in a deep-freezer
(!703C) for later biomechanical study. The deep-frozen
tibiae were thawed over 4 h at room temperature. After
removing bone tissue from the surfaces of the HA block,
the compressive strength was measured by the material
test machine (Model 5585, 5565, Instron, Canton, Massa-
chusetts, USA) according to ASTM C 773-4. After hy-
dration of porous HA specimens with distilled water,
compressive force was applied at a cross-head speed of
0.1 mm/min. In cylindrical-type porous HA samples, the
direction of force was parallel to the channel of the pores,
for the implants in vivo were compressed in that direc-
tion. Comparisons were made using the Wilcox rank sum
test and the median test.
 B.-S. Chang et al. / Biomaterials 21 (2000) 1291}1298
3. Results
3.1. Evaluation of porous HA implants
Prepared porous implants were examined by light
microscopy and SEM (Fig. 1). To estimate the actual
pore size of the synthesized HA blocks, the maximum
and minimum diameters of a pore were measured using
photos of the surface in several of the areas of each block.
The average pore size was compared with the size of the
"ber. There were no signi"cant di!erences between the
measured pore sizes and the "ber size in cylindrical-type
porous HA (Table 1). The estimated total porosity of
cylindrical-type porous HA was approximately 45%, and
the estimated macroporosity was 13}33%.
3.2. Gross xndings after implantation
The implant inserted into the medial cortex of the
tibial tuberosity was tightly united with the host bone in
all cases, without any sign of movement at the interface.
300 lm sized cylindrical-type porous HA. At eight
The portion of the porous HA exposed outside of the
Fig. 1. SEM photos of porous implants. Transverse cut (A) of
100 lm sized cylindrical-type ;120, sponge-type ;50 (B) and longitudinal cut of
cross-type porous HA ;50 (C).
1293
cortex was partially covered with new bone at four and
eight weeks post-implantation. The brittleness of im-
plants noticed before implantation was no longer ob-
served when they were harvested. Thus, the strength of
the implants seemed to improve over time.
3.3. Light microscopic xndings
HA implants appear to have no early adverse e!ects,
such as in#ammation and foreign body reaction. Os-
teoconduction through the pores was found in all six
implant groups and new bone was found on the surface
of pores with no histologically demonstrable intervening
non-osseous tissue. At four weeks after implantation, the
implants showed di!erent histological changes depend-
ing on the pore geometry. In cylindrical-type porous HA,
a new vessel was arranged in a longitudinal pattern along
the pore (Fig. 2A). In the sponge- and cross-types, an
irregular woven bone pattern was found (Fig. 2B). New
bone formation through the pores was most evident with
the
weeks after implantation, active osteoconduction was
1294 B.-S. Chang et al. / Biomaterials 21 (2000) 1291}1298

Table 1
Pore size & porosity of the porous HA implants

Implant type Measured pore Total porosity Macroporosity

(lm) size (lm) (%) (%)

Cylinder, 50 49.3$1.63 42.3$0.47 12.9$0.79

Cylinder, 100 94.1$1.28 49.9$2.72 20.3$3.96
Cylinder, 300 272.4$5.4 42.1$2.83 23.6$1.63
Cylinder, 500 472.2$8.2 47.3$1.97 32.8$1.81
Sponge 250$5.0 83.9$2.31 *
Cross 116.7$5.9 67.6$3.61 24.9$4.51

Fig. 3. Light microscopic "ndings eight weeks after implantation.

300 lm sized cylindrical-type (A), sponge-type porous HAs (B). Longi-
tudinal section, Goldner's modi"ed Massen trichrome, ;200. Evidence
of remodeling and bone marrow formation within pores was found.

intermediate form, between the cylindrical- and sponge-

types. Although the total amount of new bone formation
had not increased compared to the fourth week, evidence
of remodeling and bone marrow formation within the
pores was found with porous HA of the large cylindrical
types (
300, 500 lm), and the sponge- and cross-types
Fig. 2. Light microscopic "ndings four weeks after implantation.
(Fig. 3). In transverse sections, the new bone grown

300 lm sized cylindrical-type (A) and sponge-type porous HAs (B).
Longitudinal section, Goldner's modi"ed Massen trichrome, ;200.
50 and
A new vessel was arranged in a longitudinal pattern along the pore in
cylindrical-type porous HA. In the sponge-type, an irregular woven
300 and 500 lm cylindrical-type HAs
bone pattern was found.
through the pores showed an osteon structure forming
concentric lamellae around a single vessel in the
100 lm cylindrical-type HAs, and around multiple
vessels in the
(Fig. 4).

3.4. SEM xndings

also found even with the
50 lm sized cylindrical-type
porous HA. Better yet, in the
100}500 lm groups, the At eight weeks after implantation, the pores were "lled
bone ingrowth patterns resembled Haversian canal with bone and connective tissue. In the cylindrical-type
structures instead of woven bone structures. In the porous HA, the direction of new bone formation was
sponge-type HA, the new bone showed an irregular parallel to the long axis of pores around several central
woven bone structure. The cross-type HA showed an vessels. The discrimination of new bone from connective
 B.-S. Chang et al. / Biomaterials 21 (2000) 1291}1298 1295

Fig. 4. Light microscopic "ndings eight weeks after implantation.

50 (A) 100 (B), 500 lm sized cylindrical-type porous HAs (C). Transverse section,
hematoxylin and eosin, ;200. The new bone grown through the pores showed an osteon structure forming concentric lamellae around a single vessel
in smaller pores and multiple vessels in larger pores.

tissue was not clear. Nevertheless, the interface between
new bone and HA showed a close union without any
gaps. Especially in the
300 lm sized cylindrical-type overall strength of implants improved after implantation
HA, the pores were gradually tapered and replaced with
bone-like osteon (Fig. 5A and B). There was space only
for nutrient vessels at the opening of the pores. The
300 lm sized cylindrical-type
vessels, which were interconnected with their neighbor-
hoods through the cracks in the HA, were similar to the
Volkmann's canal in appearance. In the sponge-type HA,
the bone ingrowth pattern had no rule and was irregular
300 lm sized cylindrical-, sponge- and
along the surface of the pores (Fig. 5C). In the cross-type
HA, bone ingrowth showed the patterns of both the
cylindrical- and sponge-type HA at the same time
(Fig. 5D). In cylindrical-type pattern, the direction of
bone ingrowth was parallel to the long axis of the pores.
At the area of intersection between two cylindrical pores,
the space became enlarged and bone ingrowth showed an
irregular pattern similar to sponge-type HA.
types (6.12}12.7 MPa), and lower in the sponge-type
(0.24 MPa) and the cross-type (0.19 MPa) (Table 2). The
in all groups, with rather wide standard deviations. The
maximum compressive strength after implantation
was highest in the
(15.8 MPa), followed by the 50 lm (14.3 MPa), the
500 lm (11.5 MPa) and the 100 lm (8.65 MPa) sized cy-
lindrical-, the cross-type (7.83 MPa) and the sponge-type
(4.53 MPa). The
cross-type HAs also showed statistically signi"cant in-
creases in strength compared to pre-implantation (Wil-
cox rank sum test, median test, P(0.05). The curve
pattern of the ultimate compressive strength also became
smoother. Thus the implants, especially the sponge and
cross-types, became more #exible and ductile after im-
plantation.

3.5. Biomechanical study 4. Discussion

The maximum compressive strength of porous The ideal arti"cial bone demands good biocompatibil-
HA before implantation was high in cylindrical- ity without the possibility of in#ammation or foreign
1296 B.-S. Chang et al. / Biomaterials 21 (2000) 1291}1298

Fig. 5. SEM photos eight weeks after implantation ;120. Transverse cut (A) and longitudinal cut (B) of
300 lm sized cylindrical- and sponge-type
(C) and longitudinal cut of cross-type porous HA (D). In the cylindrical-type, the pore was tapered by the ingrown bone. In the sponge-type HA, the
bone ingrowth pattern was irregular along the surface of the pores. In the cross-type HA, bone ingrowth showed two patterns at the same time, the
cylindical- and the sponge-type.

Table 2
Increase in maximum compressive strength of the porous HA implants

Implant type (lm) Pre-implant (n) Post-implant 8 week (n) Increment Increment/
macroporosity

Cylinder, 50 12.7$1.36(3) 14.5$7.33(3) 1.8 0.14

Cylinder, 100 6.12$2.02(5) 8.65$4.87(3) 2.5 0.12
Cylinder, 300 7.98$1.66(4) 15.8$5.73(3) 7.8
0.33
Cylinder, 500 10.3$3.58(4) 11.5$4.52(4) 1.2 0.04
Sponge 0.24$0.05(6) 4.53$2.14(5) 4.3
*
Cross 0.19$0.07(3) 7.83$3.26(5) 7.6
0.31

Unit: MPa.

Statistical signi"cance, P(0.05.

body/toxic reactions. Strong bonding with the host bone,
active bone ingrowth into the graft, and bioabsorbability
are also required [16,20]. Strength su$cient to resist the
mechanical load in the implanted bone is also needed.
However, none of the biomaterials that have been de-
veloped in previous studies meet all of these criteria. HA,
which is one of the major inorganic materials in normal
bone, has good biocompatibility and osteoconductivity.
However, its fragility is a drawback, like other ceramic
materials. Therefore, it can be used alone in areas that do
not require good mechanical strength. It can also be used
with supplementary metal "xation in areas which bear
 B.-S. Chang et al. / Biomaterials 21 (2000) 1291}1298
large amounts of the mechanical load [1,12]. The struc-
ture of the dense sintered body is stronger and more able
to bond rapidly with host bone, but its use is limited due
to its high level of brittleness and low osteoconductivity
and absorbability. Porous HA, although its initial
strength is weak, is considered a good substitute, because
it shows good osteoconduction and is replaced by the
host bone [3,17,21].
Previous researchers have failed to control the accu-
rate size of the pores, or the interconnections between
pores and porosity, in preparation of the porous HA that
was used in their studies. The optimal geometry for
osteoconduction has not yet been determined. Hubert
et al. proposed the
100 lm sized pore as the smallest even in
50 lm sized pores, contrary to previous
pore than can be used for osteoconduction, and the
150 lm sized pore as the optimal pore [20]. Flatley
et al. reported that the
500 lm sized pore was the as osteon with a central
20}30 lm capillary running
most compatible for osteoconduction [19]. However,
most of these studies were based on the use of implants
50 lm sized pores could occur because the size of the
of random pore geometry. Porous HA with random
pore geometry has wide ranges of porosity and pores
20 lm) between the pores of HA made with
100 lm
of various sizes with much smaller interconnecting fenes-
tration. Therefore, the size of interconnection rather
than the pore size may determine the degree of osteocon-
duction.
In this study, to assess the di!erences in osteoconduc-
tion according to the pore sizes, we developed cylindrical
porous HA sized from
50 to 500 lm without intercon- The results of SEM examinations were similar to the
nection between the pores. In order to evaluate the e!ects
of pore geometry and interconnection on osteoconduc-
tion, we also developed three kinds of porous HA, i.e.,
cylindrical-, sponge- and cross-type. These were similar
in pore size, but di!erent in pore shape and interconnec-
tion pattern. The "nal pore size of porous HA was
smaller by 0}10% than the initial size of the inserted
polyester "ber, because of the densi"cation of the HA
granules during sintering. In the sponge-type HA, it was
only possible to measure the size of interconnection
(
250 lm). However, the largest pore was actually over
300 lm in diameter. In the cross-type HA, the pore size
was
120 lm, but the area of interconnection was ap-
proximately
150 lm or larger. showed smoother curve patterns and had biomechani-
Environmental e!ects are greater in cases of porous
material than in dense material, because the more pores
there are, the wider the surface of the porous material
becomes, and its mechanical strength weakens inversely
proportional to its porosity [19]. To provide the same
environment for osteoconduction in each group, we at-
tempted to synthesize the porous HA with 60}70% por-
osity. However, we were unable to completely achieve
this goal. The compressive strength of porous HA before
50 lm sized pore was enough for
implantation was generally inversely proportional to the
300 lm sized pore was
porosity. But, some di!erences in compressive strength
between the groups were found due to the di!erent pore
geometry and preparation methods in each group as
described in Section 2.2.
1297
In cylindrical-type HAs, we suggest that the implants
increased in strength early due to their directional prop-
erty. This is because their patterns of osteoconduction
initially showed osteon structures in which new bone was
formed followed by a cutting cone, similar to the re-
modeling process of fracture healing. In the sponge- and
cross-type HA, vigorous active osteoconduction was also
observed, but a quantitative comparison between the
groups was not performed. The pattern of bone ingrowth
was similar to the irregular woven bone structure gener-
ally found in fracture healing process in sponge- and
cross-type HA.
Active osteoconduction was observed in our study
reports in which osteoconduction hardly occurred in
such small pores. The new bone had the same structure
through the pore. We suggest that osteoconduction in
pore was greater than that of the interconnection
(
sized naphthalene or polymer beads. The larger pores
that contained extra space, in addition to the space
needed for bone ingrowth, showed bone marrow forma-
tion without in#ammation or foreign body reaction. This
suggests high bioa$nity and biocompatibility of HA
implants.
light microscopic "ndings. The newly ingrown bone
bonded directly with the HA and had a smoother inter-
face than it did for pre-implantation. As Hemmerle et al.
had previously reported [22], this suggests absorption of
some of the HA grain, and the formation of new apatite
in the body #uid and in the bone tissue.
The cylindrical-type HA appeared clinically appli-
cable, because its compressive strength was nearly the
same as or slightly superior to that of normal cancellous
bone; 7}8 MPa before implantation and much improved
after implantation [6]. Not only did the compressive
strength increase, but the change in the curve pattern was
also meaningful. The post-implant HAs in all groups
cally increased in ductility, strength and #exibility. This
was especially apparent in the sponge- and cross-
type HAs which were weaker in strength due to their
high porosity, but better in #exibility and ductility be-
cause of the high proportion of bony replacement after
implantation.
In conclusion, patterns of osteoconduction at porous
HAs were di!erent with pore con"guration. In cylin-
drical-type HAs, the
osteoconduction, and the
optimum. Porous HA with cylindrical pores could be
a useful graft material due to its strength, osteoconduc-
tivity and the ease with which its pore geometry can be
controlled.
1298 B.-S. Chang et al. / Biomaterials 21 (2000) 1291}1298
Acknowledgements
This study was supported by a grant (CHMP-96-E-4-
1009) from the '97 Good Health R & D Project, Korean
Ministry of Health & Welfare.
References
[1] Abdel-Fattah WI, Osiris WG, Mohamed SS, Khalil MR. Recon-
struction of resected mandibles using a hydroxyapatite veterinary
bone graft. Biomaterials 1994;15:609}14.
[2] Bucholz RW, Carlton A, Holmes RE. Hydroxyapatite and trical-
cium phosphate bone graft substitutes. Orthop Clin North Am
1987;18:323}34.
[3] Hench LL. Bioceramics: from concept to clinic. J Am Ceram Soc
1991;74:1487}510.
[4] Holmes RE, Bucholz RW, Monney V. Porous hydroxyapatite as
a bone-graft substitute in metaphyseal defects. J Bone Jt Surg
1986;68A:904}11.
[5] Ishihara K, Arai H, Nakabayashi N, Morita S, Furuya KI.
Adhesive bone cement containing hydroxyapatite particle
as bone compatible "ller. J Biomed Mater Res 1992;26:
937}45.
[6] Jarcho M. Calcium phosphate ceramics as hard tissue prosthetics.
Clin Orthop 1981;157:259}78.
[7] Ohgushi H, Goldberg VM, Caplan AI. Repair of bone defects
with marrow cells and porous ceramic. Acta Orthop Scand
1989;60:334}9.
[8] Radin SR, Ducheyne P. E!ect of bioactive ceramic composition
and structure on in vitro behavior. III. Porous versus dense
ceramics. J Biomed Mater Res 1994;28:1303}9.
[9] Tsuruga E, Takita H, Itoh H, Wakisaka Y, Kuboki Y. Pore size of
porous hydroxyapatite as the cell-substratum controls BMP-
induced osteogenesis. J Biochem 1997;121:317}24.
[10] Yamamura K, Iwata H, Yotsuyanagi T. Synthesis of antibiotic-
loaded hydroxyapatite beads and in vitro drug release testing.
J Biomed Mater Res 1992;26:1053}64.
[11] Zyman Z, Ivanov I, Glushko V, Dedukh N, Malyshkina S. Inor-
ganic phase composition of remineralisation in porous CaP cer-
amics. Biomaterials 1998;19:1269}73.
[12] Gauthier O, Bouler J, Aguado E, Pilet P, Daculsi G. Macropor-
ous biphasic calcium phosphate ceramics: in#uence of macropore
diameter and macroporosity percentage on bone ingrowth. Bio-
materials 1998;19:133}9.
[13] Johnson KD, Frierson KE, Keller TS, Cook C, Scheinberg R,
Zerwekh J, Meyers L, Sciadini MF. Porous ceramics as bone graft
substitutes in long bone defects: a biomechanical, histological,
and radiographic analysis. J Orthop Res 1996;14:351}69.
[14] Le Huec JC, Schaeverbeke T, Clement D, Faber J, Le Rebeller A.
In#uence of porosity on the mechanical resistance of hydroxyapa-
tite ceramics under compressive stress. Biomaterials 1995;16:113}8.
[15] Magan A, Ripamonti U. Geometry of porous hydroxyapatite
implants in#uences osteogenesis in baboons (Papio ursinus).
J Craniofac Surg 1996;7:71}8.
[16] Verheyen CC, de Wijn JR, van Blitterswijk CA, de Groot K,
Rozing PM. Hydoxylapatite/poly(L-lactide) composites: an ani-
mal study on push-out strengths and interface histology. J Bio-
med Mater Res 1993;27:433}44.
[17] Bagambisa FB, Joos U, Schilli W. Mechanisms and structure of
the bond between bone and hydroxyapatite ceramics. J Biomed
Mater Res 1993;27:1047}55.
[18] Eggli PS, Muller W, Schenk RK. Porous hydroxyapatite and
tricalcium phosphate cylinders with two di!erent pore size ranges
implanted in the cancellous bone of rabbits. Clin Orthop
1988;232:127}38.
[19] Flatley TJ, Lynch KL, Benson M. Tissue response to implants of
calcium phosphate ceramic in the rabbit spine. Clin Orthop
1983;179:246}52.
[20] Hulbert SF, Morrison SJ, Klawitter JJ. Tissue reaction to three
ceramics of porous and non-porous structures. J Biomed Mater
Res 1972;6:347}74.
[21] Vuola J, Taurio R, GoK ransson H, Asko-Seljavaara S. Compressive
strength of calcium carbonate and hydroxyapatite implants after
bone-marrow-induced osteogenesis. Biomaterials 1998;19:223}7.
[22] Hemmerle J, Oncag A, Erturk S. Ultrastructural features of the
bone response to a plasma-sprayed hydroxyapatite coating in
sheep. J Biomed Mater Res 1997;36:418}25.