* Corresponding author. Present address: Department of Orthopedic

Surgery, Kobe City General Hospital, Minatojimanakamachi 4-6,
Chuou-ku, Kobe City 650-0046, Japan. Tel.: #[phone];
fax: #[phone].

E-mail address: shigeru

}[e-mail]

(S. Nishiguchi).

Biomaterials 22 (2001) 2525}2533

Titanium metals form direct bonding to bone after alkali

and heat treatments

Shigeru Nishiguchi

*, Hirofumi Kato , Hiroshi Fujita , Masanori Oka, Hyun-Min Kim,

Tadashi Kokubo

, Takashi Nakamura

Department of Orthopaedic Surgery, Faculty of Medicine, Kyoto University, Shougoin-kawaharacho 54, Sakyo-ku, Kyoto 606-8507, Japan

Institute for Frontier Medical Sciences, Kyoto University, Shougoin-kawaharacho 53, Sakyo-ku, Kyoto 606-8507, Japan

Department of Material Chemistry, Faculty of Engineering, Kyoto University, Yoshida-honmachi, Sakyo-ku, Kyoto 606-8501, Japan

Received 19 July 2000; accepted 18 December 2000

Abstract

In this article we evaluated the bone-bonding strengths of titanium and titanium alloy implants with and without alkali and heat

treatments using the conventional canine femur push-out model. Four kinds of smooth cylindrical implants, made of pure titanium or
three titanium alloys, were prepared with and without alkali and heat treatments. The implants were inserted hemitranscortically into
canine femora. The bone-bonding shear strengths of the implants were measured using push-out test. At 4 weeks all types of the alkali-
and heat-treated implants showed signi"cantly higher bonding strength (2.4}4.5 MPa) than their untreated counterparts
(0.3}0.6 MPa). At 12 weeks the bonding strengths of the treated implants showed no further increase, while those of the untreated
implants had increased to 0.6}1.2 MPa. Histologically, alkali- and heat-treated implants showed direct bonding to bony tissue
without intervening "brous tissue. On the other hand, untreated implants usually had intervening "brous tissue at the interface
between bone and the implant. The early and strong bonding to bone of alkali- and heat-treated titanium and its alloys without
intervening "brous tissue may be useful in establishing cementless stable "xation of orthopedic implants.

2001 Elsevier Science

Ltd. All rights reserved.

Keywords: Titanium; Titanium alloy; Alkali and heat treatment; Cementless "xation

1. Introduction

Titanium and its alloys have excellent mechanical

properties and biocompatibility [1,2]. In the "eld of
dentistry, smooth-surfaced titanium implants inserted
into the facial bones perform very well [3]. After
implantation, well-"xed dental implants, which are not
subjected to weight loading during some period after
implantation, exhibit direct contact with bone without
intervening "brous tissue in light microscopic level. This
is termed osteointegration. However, in skeletal bones,
smooth titanium implants usually tend to be encap-
sulated by "brous tissue and show only weak bonding to

bone in animal experiments even under unloaded
conditions [4}6]. Thus, recently, "xation of titanium
implants to bone usually depends on biological "xation
of its porous surfaces. It is shown that, in human cement-
less total hip arthroplasty (THA), osteointegration is
not always achieved on smooth-surfaced titanium
metals [7,8]. If smooth titanium metals have an ability
to bond to bone strongly without forming intervening

"brous

tissue, cementless THA implants may be able

to achieve long-term stability even without porous
coatings.

In order to combine the mechanical properties of tita-

nium with bone-bonding abilities of bioactive materials,
methods of coating the titanium surface with several
bioactive materials have been developed [4,9}17].
Hydroxyapatite (HA) plasma spray coating is one of
the most widely investigated and used methods for
orthopedic implants [18]. However, while HA-plasma-
spray-coated THA implants have produced encouraging
short- and middle-term results, concerns regarding re-
sorption or delamination of coating layer, in#ammatory

0142-9612/01/$ - see front matter

2001 Elsevier Science Ltd. All rights reserved.

PII: S 0 1 4 2 - 9 6 1 2 ( 0 0 ) 0 0 4 4 3 - 9

Table 1
Compositions of the titanium-based metals used in this study (%)

H

O

N

Fe

Al

Zr

Nb

Mo

Ta

V

Ti

max

max

max

max

Ti

0.001

0.15

0.03

0.15

bal.

Ti6Al4V

0.0125

0.20

0.05

0.30

5.50}6.75

3.50}4.50

bal.

Ti6Al2Nb1Ta

0.0125

0.1

0.03

0.25

5.5}6.5

1.5}2.5

0.5}1.0

0.5}1.5

bal.

Ti15Mo5Zr3Al

0.020

0.20

0.05

0.35

2.5}3.5

4.5}5.5

14.0}16.0

bal.

According to the speci"cation of Kobe Steel KS 50.

According to the speci

"cation of Kobe Steel KS 6-4.

According to the speci

"cation of MIL T-9046J Code A-3 Carbon content at the maximum of 0.05%.

According to the speci

"cation of Kobe Steel KS 15-5-3.

reactions and third-body wear of joint due to migration
of delaminated HA particles, have been raised [4,19}23].

Another approach to enhance the strength of titanium

and titanium alloy bonding to bone has recently been
developed. After alkali and heat treatments, titanium-
based metals form bone-like apatite in simulated body

#uid (SBF) [24], which has ion concentrations nearly

equal to human body #uid. This phenomenon also oc-
curs on the surfaces of bioactive glass and glass-ceramics.
Apatite formation on the surface of alkali- and heat-
treated titanium metals appears to occur in vivo, and
leads to bonding to living bone [25,26].

We have previously investigated the e!ect of alkali and

heat treatments on the bone-bonding abilities of titanium
and titanium alloys using our original detaching test
[27,28]. This test is simple to perform and is appropriate
for evaluating the chemical bonding between bone and
biomaterials since it eliminates the in#uence of mechan-
ical interlocking. However, it cannot measure the true
tensile strength because the bonding area cannot be mea-
sured, and therefore it is not widely used [29,30].

In the present study, we compared the bone-bonding

strength of alkali- and heat-treated titanium and titanium
alloys with that of untreated ones using conventional
push-out test. Smooth implants were used, and our results
will therefore be useful for comparisons with experiments
on the implants with porous surface, since porous mate-
rials are often used in clinical settings to enhance the
bonding strength during the early postimplantation peri-
od. The purposes of this study were to compare the bone-
bonding strength of four titanium-metal-based implants
using conventional push-out test and to determine which
of these metals would be the most suitable bioactive ma-
terial after alkali and heat treatments.

2. Materials and methods

2.1. Implant preparation

Cylindrical implants (diameter 6 mm, length 13 mm)

were prepared from pure titanium (Ti) or three titanium

alloys: Ti6Al4V, Ti6Al2Nb1Ta, and Ti15Mo5Zr3Al
(Kobe Steel Co., Japan). The compositions of these meta-
ls are summarized in Table 1. The surfaces of the im-
plants were abraded with

C400 abrasive paper. Half of

each implant was used untreated, as a control. The other
half was soaked in sodium hydroxide (NaOH) solution at
603C for 24 h and washed with distilled water and dried
at room temperature for 24 h. The concentrations of the
NaOH solutions were 5 mol/L for Ti and 10 mol/L for
the alloys. After alkali treatment, the implants were
heated to 6003C at a rate of 53C/min and maintained at
this temperature for 1 h and allowed to be cooled in the
furnace to room temperature. The implants were then
sterilized with ethylene oxide gas.

2.2. Implant evaluation

The diameters and surface roughness of the four types

of implants (both untreated and alkali- and heat-treated)
were measured because these parameters strongly a!ect
the results of the push-out test. The measurements of the
surface roughness were carried out for 1.25 mm along the
longitudinal axis twice for each implant (Surftest 501,
Mitsutoyo Co., Tokyo, Japan). The surfaces of the im-
plants were observed with a scanning electron micro-
scope (S-800, Hitachi Co., Tokyo, Japan).

2.3. Surgical procedure

The implant materials were evaluated in vivo using the

hemi-transcortical cylindrical model in dogs. Fourteen
adult female beagle dogs weighing 10}12 kg were used.
The animals were anesthetized by intramuscular admin-
istration of ketamine hydrochloride (50 mg/kg) followed
by diazepam (5 mg) and atropine sulfate (0.5 mg). Just
before the operation, an additional dose of 30 mg/kg
ketamine hydrochloride was injected intramuscularly.
The operations were performed in the usual sterilized
manner. The dogs were placed in the right decubitus
position and a lateral skin incision was made on the left
thigh and the middiaphyseal region of the left femur was
exposed. The implantation sites were prepared slowly

2526

S. Nishiguchi et al. / Biomaterials 22 (2001) 2525}2533

Fig. 1. Implantation method: seven implants were implanted into each
femur, through the lateral cortex only.

Fig. 2. (a,b) Sample preparation: the extracted femora were cut into
pieces, perpendicular to the axis, with each piece containing one im-
plant. (c) The push-out test: the support jig was arranged so that the
pushing load was applied perpendicular to the end of each implant.

and deliberately using a surgical electronic drill. Initially,
a H2.0 mm pilot hole was drilled on the lateral cortex
only. This was gradually enlarged using H4.0, 5.5, 6.0,
and 6.1 mm drill bits. During drilling, the hole was con-
tinuously cooled with saline. Just before insertion of the
implants, the hole was irrigated with saline containing
ipasemycin sulfate to remove any shards of bone. An
untreated implant was then inserted through the unilat-
eral cortex and seated into place as tightly as possible
using "nger pressure. The four types of untreated im-
plants were inserted in a randomized manner in order to
avoid any position-related di!erence. Each implant was
inserted with its center 1.4 cm away from that of the next
implant. Seven implants were inserted into each femur.
The fascia and subcutaneous layers were then closed with
silk sutures and the skin was closed with skin staples. The
dog was then placed in the left decubitus position and the
procedure was repeated to insert alkali- and heat-treated
implants into the right femur. The dogs were sacri"ced at
either four or 12 weeks after operations with pan-
curonium bromide and an overdose of phentobarbital
sodium and both femora were retrieved. The guidelines
of the care and use of laboratory animals of Institute of
Laboratory Animals, Faculty of Medicine, Kyoto Uni-
versity were observed throughout this study.

2.4. Mechanical tests

The implants were numbered from 1 (proximal) to

7 (distal) (Fig. 1). Each extracted femur was cut into
pieces, perpendicular to its axis, with each piece contain-
ing one implant. The bones containing implants nos. 2}6
were then split longitudinally (Fig. 2a and b). The bones
containing implants nos. 1 and 7 were prepared for
histological examination. All the implant-containing
bone specimens were preserved in a freezer at !203C
until required for mechanical testing, when they were
thawed at room temperature. For the push-out test, each
implant-containing bone specimen was mounted on
a special metal platform with a central circular opening,
which supported the bone to within 1 mm of the
bone}implant interface. This jig was designed so as to
keep the pushing load parallel to the long axis of the
implant (Fig. 2c). The pushing load was applied to the
implant end using AG-10 TB apparatus (Shimazdu,
Kyoto, Japan) at a cross-head speed of 0.5 mm/min until
the peak load was obtained.

The thickness of the cortical bone in contact with the

implant was measured at two sites for each push-out
sample. The average thickness was calculated and used to
determine the contact area according to the following
formula:

interface

area"3.14

;implant diameter

(6 mm)

;average cortical thickness. The shear strength at

the interface was calculated by dividing the load at failure
by the interfacial area. All data were expressed as
mean$standard deviation (SD) and statistically ana-

lyzed according to the kind of metals, postoperative time,
and with and without alkali and heat treatments using
a one-way ANOVA with Fisher's PLSD method as
a post-hoc test. We set 16 groups according to the kinds
of metals, with or without alkali and heat treatments, and
postoperative time. Di!erences with p(0.05 were con-
sidered to be statistically signi"cant.

2.5. Histological examination

Specimens for histology, i.e., those containing implants

nos. 1 and 7, were "xed in 10% phosphate-bu!ered
formalin for 7 days and dehydrated in serial concentra-
tions of ethanol (70, 80, 90, 99, 100, and 100% v/v;
increasing every three days). They were then embedded in
polyester resin. Sections of 500

m thickness were cut

with a band saw (BS-3000, EXACT cutting system, Nor-
derstedt, Germany), perpendicular to the axis of the im-
plant. For Giemsa surface staining, the sections were
bound to a transparent acrylic plate with a cyanoacrylic
adhesive, and ground to a thickness of about 80

m using

a grinding-sliding machine (Microgrinding MG-4000,
EXAKT,

Norderstedt,

Germany).

Several

500

m

S. Nishiguchi et al. / Biomaterials 22 (2001) 2525}2533

2527

Fig. 3. Scanning electron microscope examination of the Ti implant surface. (a) Alkali- and heat-treated: these implants have micro-porous structure.
(b) Untreated: each untreated metal implant shows scratch marks due to abrasion during manufacture.

Table 2
Mean surface roughness of the implants R (R ) (m) (n"2)

Untreated

Alkali- and heat-treated

Ti

1.16 (9.00)

1.09 (7.93)

Ti6Al4V

0.20 (1.62)

0.22 (2.21)

Ti6Al2Nb1Ta

0.49 (3.18)

0.46 (3.17)

Ti15Mo5Zr3Al

0.51 (3.52)

0.42 (3.24)

sections were polished with diamond paper and coated
with a thin layer of carbon for scanning electron micro-
scope observations (S-800, Hitachi Co., Tokyo, Japan).
The observations were performed mainly at the im-
plant}bone interfaces.

3. Results

3.1. Implant evaluation

The four kinds of implants had similar diameters,

which did not decrease after alkali and heat treatments.
The surface roughness of the implants was almost identi-
cal before and after alkali and heat treatments; however,
the pure titanium implants had slightly rougher surfaces
than those prepared from the titanium alloys at all as-
sessments (Table 2). Scanning electron microscopy
showed that the surfaces of the alkali- and heat-treated
implants had a very "nely irregular structure (Fig. 3a)
and that the surfaces of the untreated implants had
abrasive marks (Fig. 3b).

3.2. In vivo performance of the implants

All 14 dogs tolerated the implant operation well. With-

in 48 h, each dog could bear weight on its rear legs.
Although none of the femurs were found to be broken at
the time of sacri"ce, some of the implants had been
displaced. These displacements seemed to have occurred
very soon after implantation, because the insertion holes
were "lled with callus. Displacement occurred mainly
with untreated implants, especially those placed in no.
3 position (Table 3). This seemed to be due to the implant
being too long for the intramedullary canal of the prox-
imal femur, and therefore protruding from the bone. In
addition, the femur is at its narrowest at the point corre-
sponding to no. 3 position. The implants thus seemed to
be too long to be accommodated in the medullar cavity
of the femur of beagle dogs weighing 10}12 kg, and this
may account for instability leading to displacement.

Alkali- and heat-treated implants were found to have

been displaced less frequently than their untreated
counterparts. This provides additional support to the
theory that the alkali and heat treatments would prevent
implant displacement during the early postimplantation
period, although we did not expect this result before
starting this study. The displaced implants were excluded
from further tests. The implant numbers included in this
study are summarized in Table 4.

3.3. Push-out test

The results of the push-out test are summarized in

Figs. 4 and 5. At 4 weeks, the untreated implants showed

2528

S. Nishiguchi et al. / Biomaterials 22 (2001) 2525}2533

Table 4
Numbers of samples included in the study

4 weeks

12 weeks

Mechanical

Histo-
logical

Mechanical

Histo-
logical

Ti!

7

2

8

2

Ti#

9

3

9

2

Ti6Al4V!

7

4

8

3

Ti6Al4V#

8

4

10

3

Ti6Al2Nb1Ta!

7

2

8

2

Ti6Al2Nb1Ta#

8

3

8

2

Ti15Mo5Zr3Al!

8

3

8

2

Ti15Mo5Zr3Al#

9

3

9

2

Note: (!) untreated; (#) alkali- and heat-treated.

Fig. 5. Results of push-out tests at 12 weeks: (!) untreated; (#) alkali-
and heat-treated.

Table 3
Numbers of dislocated implants according to their positions along the
femur

Position

Untreated

Alkali- and heat-treated

1

1

0

2

2

0

3

7

1

4

4

0

5

0

0

6

0

0

7

0

0

Fig. 4. Results of push-out tests at 4 weeks: (!) untreated; (#) alkali-
and heat-treated.

low shear strengths (0.47$0.42 MPa for Ti, 0.56$
0.68 MPa

for

Ti6Al4V,

0.32$0.29 MPa

for

Ti6Al2Nb1Ta, and 0.50$1.06 MPa for Ti15Mo5Zr3Al).
All of the alkali- and heat-treated implants showed signi"-
cantly higher shear strengths than the untreated implants
(p"0.006 for Ti, p"0.02 for Ti6Al4V, p"0.03 for
Ti6Al2Nb1Ta, and p(0.0001 for Ti15Mo5Zr3Al). The
shear strengths of alkali- and heat-treated implants were
3.12$1.79 MPa for Ti, 2.76$1.85 MPa for Ti6Al4V,
2.35$2.39 MPa for Ti6Al2Nb1Ta, and 4.52$2.85 MPa
for Ti15Mo5Zr3Al. Among the treated implants, Ti15Mo-
5Zr3Al exhibited signi"cantly greater shear strength than
Ti6Al4V (p"0.04) or Ti6Al2Nb1Ta (p"0.01).

At 12 weeks, the bonding strengths of all the untreated

implants had increased compared with the 4-week assess-
ment, but did not reach the values exhibited by the alkali-
and heat-treated ones. The values for the untreated
implants were 1.24$1.05 MPa for Ti, 0.57$0.38 MPa
for Ti6Al4V, 0.78$0.65 MPa for Ti6Al2Nb1Ta, and
0.68$0.60 MPa for Ti15Mo5Zr3Al. The values for the
alkali- and heat-treated implants were 3.24$2.11 MPa
for Ti, 1.12$0.65 MPa for Ti6Al4V, 2.60$1.79 MPa
for Ti6Al2Nb1Ta, and 4.48$4.12 MPa for Ti15Mo-
5Zr3Al. Thus, all treated implants showed signi"cantly
higher strength than the untreated counterparts except

Ti6Al4V (p"0.02 for Ti, p"0.51 for Ti6Al4V, p"0.04
for Ti6Al2Nb1Ta, and p(0.0001 for Ti15Mo5Zr3Al). In
addition, all of the treated implants produced almost
identical bonding strength values at 12 weeks as at
4 weeks, except Ti6Al4V, which showed a decrease. The
di!erence was not, however, statistically signi"cant with
the number available (p"0.06). At 12 weeks, among the
treated implants, the shear strength of pure titanium was
signi"cantly higher than that of Ti6Al4V (p"0.01),
and that of Ti15Mo5Zr3Al was signi"cantly higher
than those of Ti6Al4V (p"0.0001) and Ti6Al2Nb1Ta
(p"0.03).

3.4. Histological examination

No foreign body or in#ammatory reaction was noted

with either the untreated or the alkali- and heat-treated
implants throughout the study, regardless of the kind of
metal used. Furthermore, none of the four metals pro-
duced any substantial di!erence in the histological "nd-
ings. With the treated implants, new bone formed in the
gap created at the implantation site within 4 weeks,
and the new bone was in direct contact with the implant
(Fig. 6a). On the other hand, with the untreated implants,
only an intervening "brous layer or a small amount of
bone was in direct contact (Fig. 6b).

At 12 weeks, the treated implants had almost the same

amount or a little more bone in contact with them than
at 4 weeks (Fig. 7a). In contrast, the untreated samples

S. Nishiguchi et al. / Biomaterials 22 (2001) 2525}2533

2529

Fig. 6. Giemsa surface staining of the implants at 4 weeks. (a) Alkali-
and heat-treated Ti15Mo5Zr3Al implant: new bone was formed at the
interface between the original bone and the implant, and showed direct
bonding to the implant. The interface between the newly formed bone
and the preexisting bone is clearly marked because of the di!erence in
staining and structure between the two portions, (B) bone. (b) Un-
treated Ti15Mo5Zr3Al implant: a "brous tissue layer can be seen
between the bone and the implant, (B) bone, (I) "brous tissue.

Fig. 7. Giemsa surface staining of the implants at 12 weeks. (a) Alkali-
and heat-treated titanium implant: the bone was in direct contact with
the implant. The interface between the newly formed bone and the
original bone became obscure, (B) bone. (b) Untreated titanium im-
plant: the "brous tissue layer between the bone and the implant was
much thicker than that at 4 weeks, (B) bone, (I) "brous tissue.

still showed intervening "brous tissue or a limited
amount of bone in direct contact with the implants
(Fig. 7b). In some untreated samples, the "brous tissue at
the interface between the bone and the implant was
much thicker than at 4 weeks (Fig. 7b). Although this
study was conducted without applying a constant load
to the implants, some stress may be applied to the im-
plant}bone interface due to muscular contraction
around the extruded portions of some implants. This
may have enhanced "brous tissue formation at the
bone}implant interface.

Scanning electron microscopy revealed almost the

same "ndings as those noted in the Giemsa-stained sam-
ples (Fig. 8a and b).

4. Discussion

Untreated titanium metal implants have often been

used as control in previous studies on the bone-bonding

behavior of HA-plasma-spray-coated titanium metal im-
plants using push-out test model [9,31]. In both studies,
histological examinations revealed that HA-coated im-
plants were in direct contact with the bone at 3 or
4 weeks after the implantation whereas untreated tita-
nium metal produced few areas of direct bone}implant
apposition or intervening "brous tissue at the interface
between bone and titanium implants even at 10 or 12
weeks. These histological "ndings regarding untreated
titanium metal implants are in accordance with our own

"ndings. In contrast, the alkali- and heat-treated tita-

nium alloys used in the present study showed direct
bone}implant apposition and osteoconductive charac-
teristics more similar to those displayed by HA-coated
implants than those by untreated titanium metals. These
results indicate that quite simple alkali and heat treat-
ments can induce osteoconductive properties in implants
manufactured from titanium and its alloys.

This study demonstrated that alkali- and heat-treated

titanium metal implants show signi"cantly higher

2530

S. Nishiguchi et al. / Biomaterials 22 (2001) 2525}2533

Fig. 8. Scanning electron microscope images (back-scattered mode) of Ti15Mo5Zr3Al implants at 4 weeks. (a) Alkali- and heat-treated implant: direct
bonding between the bone and the implant was observed. (b) Untreated implant: there was a gap between the implant and the bone.

bonding strength than untreated implants as early as
4 weeks after insertion. In the push-out test, implants
with a rougher surface show higher bonding strength
than those with a smoother surface when the implants
are made from the same material. However, surface
roughness measurements revealed that the surface
roughness of the implants used in this study did not
change even after alkali and heat treatments. As shown in
SEM examination (Fig. 3a), alkali- and heat-treated tita-
nium has a submicron-level trabecular and irregular
structure on its surface. This surface topology may play
a role in the bone-bonding strength of this material.
However, the average roughness of any kind of implant is
less than 1

m and only weak mechanical interlocking

occurs if at all. Thus it is supposed that chemical bonding
between alkali- and heat-treated titanium metals and
bone via an apatite layer plays a major role in bone-
bonding behavior.

Histological examinations also showed direct bone

apposition with the alkali- and heat-treated titanium
implants. In contrast, no or only scarce bone apposition
was observed on the surfaces of the untreated titanium
implants at both four and 12 weeks after implantation,
and an intervening "brous tissue layer was observed, in
accordance with previous reports [9,31]. These histologi-
cal "ndings support the results of the push-out tests.

During the study, some implants were displaced from

their original positions in vivo. The reasons for this initial
instability appear to be that the insertion hole was
0.1 mm larger than the diameter of the implants, and the
implant was longer than the intramedullary canal of the
canine femur. However, the incidence of the displace-

ments was much higher with the untreated implants than
with the alkali- and heat-treated implants. Both types of
implants had almost identical diameters and surface
roughness, and were implanted using the same technique.
This unexpected "nding indicates that alkali and heat
treatments increase the initial stability of the implants.

Both the implant itself and the living bone play a role

in the strong bonding between bone and alkali- and
heat-treated titanium metal implants. Alkali- and heat-
treated titanium implants have a thin reactive layer
formed as a result of the alkali and heat treatments, on
their surfaces. This layer can form apatite in SBF, like
bioactive glasses and glass-ceramics and this is also
thought to occur in vivo [24,32]. Apatite formation on
the surface of the material is considered a prerequisite for
direct bone bonding. Details of the mechanism underly-
ing apatite formation on the surface of alkali- and heat-
treated titanium metals have been described by Kim et al.
[25]. Brie#y, amorphous sodium titanate generated by
alkali and heat treatments exchanges Na

> for HO> and

forms hydrated titania in the body #uid on the implant
surface. The Ti}OH groups in this hydrated titania then
induce the apatite nucleation.

However, the early bone "lling and direct bone apposi-

tion cannot be explained by this mechanism alone.
Living bone must also play a major role in producing
strong bonding. The way in which living bone reacts with
alkali- and heat-treated titanium implants has, however,
not yet been elucidated. Osteogenic cells may attach to
the surfaces of alkali- and heat-treated titanium implants
themselves or to the apatite formed on their surfaces, and
this may enhance growth and di!erentiation. Thus, the

S. Nishiguchi et al. / Biomaterials 22 (2001) 2525}2533

2531

gap initially created between the bone and the implants
at the implantation procedure may be "lled with new
bone faster with alkali- and heat-treated implants than
with untreated implants. Once new bone forms, tight
bonding between the bone and the alkali- and heat-
treated titanium implant develops via the biological
apatite in the bone and the bone-like apatite on the
surface of the implant. This might explain the strong
bonding between the bone and the alkali- and heat-
treated titanium implants.

In the present study we examined four types of tita-

nium metals. Ti and Ti6Al4V are now widely used for
orthopedic implants. Ti6Al2Nb1Ta and Ti15Mo5Zr3Al
are new titanium alloys which do not contain vanadium
and may be used in the clinical setting in the future. Pure
titanium showed a slightly rougher surface than the three
alloys both before and after treatments. This may explain
why untreated and treated pure titanium implants pro-
duced a higher bonding strength than the titanium alloy
implants. The three alloys had almost the same surface
roughness both before and after treatments. Among
the treated alloy implants, Ti15Mo5Zr3Al showed the
highest bonding strength. Thus, pure titanium and
Ti15Mo5Zr3Al seem to be the most suitable substances
for manufacturing alkali- and heat-treated titanium alloy
implants. The factors responsible for the di!erences in
bonding strength between these four types of implants
are not clear at present.

The most serious problem in the long-term results of

total hip replacements is now osteolysis [33]. Osteolysis
is considered to be due to the in#ammatory reaction of
living body to the particulate debris generated at the
articulating surface. It is thought to develop along the
bone}implant interface because particulate debris gener-
ated at the articulating surface works its way through the

"brous

tissue between the bone and the stem of the

implant to the distal part of the implants [34,35]. Thus, if
early direct chemical bonding between bone and implant
metals could be achieved, the access of the debris to the
implant surfaces and subsequent osteolysis could be
prevented and might reduce such failures in cementless
hip replacement implants.

As this study was carried out under non-weight-bear-

ing conditions, the results cannot be directly applied to
the clinical situation, for example, cementless total re-
placement surgery. However, the results suggest that
further studies under weight-bearing conditions would be
encouraging, and that alkali- and heat-treated titanium-
based metals would be useful in alternative methods of
cementless "xation. The potential clinical applications of
these metals include spinal instruments and external

"xator pins for use under loading conditions or for long

periods of time, as well as cementless hip and knee
replacement implants.

In summary, alkali and heat treatments improved the

bone-bonding abilities of pure titanium and titanium

alloy implants as early as 4 weeks after implantation,
compared with their untreated counterparts. Among the
four

types

of

metals

tested,

pure

titanium

and

Ti15Mo5Zr3Al seemed most suitable for the manufac-
ture of alkali- and heat-treated implants, although fur-
ther studies will be needed to determine the optimal
conditions for the alkali and heat treatments. This tech-
nique has the potential to become an accepted method of
enhancing the bone-bonding abilities of titanium-based
metal orthopedic implants.

Acknowledgements

This study was supported by the Grant-in-Aid for

Scienti"c Research (No. 09358019), the Ministry of Edu-
cation, Science, Sports and Culture, Japan.

The authors would like to thank Yoshio Sasaki (Kobe

Steel Co.) for his assistance in the manufacture of the
implants, Kiyoyuki Okunaga (Nippon Electrolic Glass
Co.) for his assistance with the mechanical testing, and
Shino Nishiguchi for her assistance in the animal experi-
ments.

References

[1] Black J, Hastings G, editors. Handbook of biomaterial properties.

London: Chapman & Hall, 1998. p. 179}200.

[2] Boyer R, Welsch G, Collings EW, editors. Materials properties

handbook: titanium alloys. Ohio: ASM international, 1994.
p. 165}7.

[3] Albrektsson T, Albrektsson B. Osseointegration of bone implants:

a review of an alternative mode of "xation. Acta Orthop Scand
1987;58:567}77.

[4] Kanagasniemi IMO, Verheyen CCPM, van der Velde EA,

de Groot K. In vivo tensile testing of #uorapatite and hy-
droxylapatite plasma-splayed coatings. J Biomed Mater Res 1994;
28:563}72.

[5] Spivak JM, Ricci JL, Blumenthal NC, Alexander H. A new canine

model to evaluate the biological response of intramedullary bone
to implant materials and surfaces. J Biomed Mater Res 1990;
24:1121}49.

[6] Takatsuka K, Yamamuro T, Nakamura T, Kokubo T. Bone-

bonding behavior of titanium alloy evaluated mechanically with
detaching failure load. J Biomed Mater Res 1995;29:157}63.

[7] Urban RM, Jacobs JJ, Sumner DR, Peters CL, Voss FR, Galante

JO. The bone}implant interface of femoral stems with non-cir-
cumferential porous coating. J Bone Jt Surg (Am) 1996;
78:1068}81.

[8] Bobyn JD, Jacobs JJ, Tanzer M, Urban RM, Arbindi R, Sumner

DR, Turner TM, Brooks CE. The susceptibility of smooth im-
plant surfaces to periimplant "brosis and migration of polyethy-
lene wear debris. Clin Orthop 1995;311:21}39.

[9] Cook SD, Thomas KA, Kay JF, Jarcho M. Hydroxyapatite-

coated titanium for orthopedic implant applications. Clin Orthop
1988;232:225}43.

[10] Duchenne P, Cuckler JM. Bioactive ceramic prosthetic coatings.

Clin Orthop 1992;276:102}14.

[11] Geesink RG, de Groot K, Klein CP. Chemical implant "xation

using hydroxyl-apatite coatings. The development of a human

2532

S. Nishiguchi et al. / Biomaterials 22 (2001) 2525}2533

total hip prosthesis for chemical "xation to bone using hydroxyl-
apatite coatings on titanium substrates. Clin Orthop 1987;
225:147}70.

[12] Geesink RG, de Groot K, Klein CP. Bonding of bone to apatite-

coated implants. J Bone Jt Surg (Br) 1988;70:17}22.

[13] Ishizawa H, Ogino M. Characterization of thin hydroxyapatite

layers formed on anodic titanium oxide "lms containing Ca and
P

by

hydrothermal

treatment.

J

Biomed

Mater

Res

1995;29:1071}9.

[14] Kitsugi T, Nakamura T, Oka M, Senaha Y, Goto T, Shibuya T.

Bone-bonding behavior of plasma-sprayed coatings of Bioglass,
AW-glass ceramic, and tricalcium phosphate on titanium alloy.
J Biomed Mater Res 1996;30:261}9.

[15] Klein CP, Patka P, van-der-Lubbe HB, Wolke JG, de Groot K.

Plasma-sprayed coatings of tetracalciumphosphate, hydroxyl-
apatite, and alpha-TCP on titanium alloy: an interface study.
J Biomed Mater Res 1991;25:53}65.

[16] Lee J, Aoki H. Hydroxyapatite coating on Ti plate by a dipping

method. Biomed Mater Engng 1995;5:49}58.

[17] Takatsuka K, yamamuro T, Kitsugi T, Nakamura T, Shibuya T,

Goto T. A new bioactive glass-ceramic as a coating material on
titanium alloy. J Appl Biomater 1993;4:317}29.

[18] Geesink RG, Hoefnagels NH. Six-year results of hydroxyapatite-

coated total hip replacement. J Bone Jt Surg (Br) 1995;77:534}47.

[19] Bloebaum RD, Dupont JA. Osteolysis from a press-"t hy-

droxyapatite-coated implant. A case study. J Arthroplasty
1993;8:195}202.

[20] Collier JP, Surprenant VA, Mayor MB, Wrona M, Jensen RE,

Surprenant HP. Loss of hydroxyapatite coating on retrieved,
total hip components. J Arthroplasty 1993;8:389}93.

[21] Nilsson KG, Cajander S, Karrholm J. Early failure of hy-

droxyapatite-coating in total knee arthroplasty. A case report.
Acta Orthop Scand 1994;65:212}4.

[22] Nimb L, Gotfredsen K, Steen-Jensen J. Mechanical failure of

hydroxyapatite-coated titanium and cobalt}chromium}molyb-
denum alloy implants. An animal study. Acta Orthop Belg
1993;59:333}8.

[23] Overgaard S, Soballe K, Lind M, Bunger C. Resorption of hy-

droxyapatite and #uorapatite coatings in man. An experimental
study in trabecular bone. J Bone Jt Surg (Br) 1997;79:654}9.

[24] Kokubo T, Miyaji F, Kim H-M, Nakamura T. Spontaneous

apatite formation on chemically surface treated Ti. J Am Ceram
Soc 1996;79:1127}9.

[25] Kim H-M, Miyaji F, Kokubo T, Nakamura T. Preparation of

bioactive Ti and its alloys via simple chemical surface treatment.
J Biomed Mater Res 1996;32:409}17.

[26] Kim H-M, Miyaji F, Kokubo T, Nakamura T. Bonding strength

of bonelike apatite layer to Ti metal substrate. J Biomed Mater
Res 1997;38:121}7.

[27] Nishiguchi S, Nakamura T, Kobayashi M, Kim H-M, Miyaji F,

Kokubo T. The e!ect of heat treatment on bone-bonding ability
of alkali-treated titanium. Biomaterials 1999;20:491}500.

[28] Yan W-Q, Nakamura T, Kobayashi M, Kim H-M, Kokubo T.

Bonding of chemically treated titanium implants to bone. J Bio-
med Mater Res 1997;37:267}75.

[29] Hong L, Hengchang X. Tensile strength of the interface

between hydrokyapatite and bone. J Biomed Mater Res 1992;
26:7}18.

[30] Nakamura T, Yamarumo T, Higashi S, Kokubo T, Itoo S. A

new glass-ceramic for bone replacements: evaluation of its
bonding ability to bone tissue. J Biomed Mater Res 1985;
23:631}48.

[31] Hayashi K, Inadome T, Mashima T, Sugioka Y. Comparison of

bone}implant interface shear strength of solid hydroxyapatite
and hydroxyapatite-coated titanium implants. J Biomed Mater
Res 1993;27:557}63.

[32] Kitsugi T, Yamamuro T, Nakamura T, Kokubo T, Takagi M,

Shibuya T. SEM-EPMA observation of three types of apatite-
containing glass-ceramics implanted in bone: the variance of
a Ca}P rich layer. J Biomed Mater Res 1987;21:259}73.

[33] Harris

WH.

The

problem

is

osteolysis.

Clin

Orthop

1995;311:46}53.

[34] Jasty M, Bragdon C, Jiranek W, Chandler H, Maloney W, Harris

WH. Etiology of osteolysis around porous-coated cementless
total hip arthroplasties. Clin Orthop 1994;308:111}26.

[35] Maloney WJ, Woolson ST. Increasing incidence of femoral

osteolysis in association with uncemented Harris-Galante total
hip arthroplasty. A follow-up report. J Arthroplasty 1996;
11:130}4.

S. Nishiguchi et al. / Biomaterials 22 (2001) 2525}2533

2533