Bioceramics: From
Concept
to
Clinic
Larry
L.
Hench*
Department
of
Materials Science and Engineering,
University of
Florida,
Gainesville,
Florida
3261
1
Ceramics used for the repair and re-
construction of diseased or damaged
parts of the musculo-skeletal sys-
tem, termed bioceramics, may be bio-
inert (alumina, zirconia), resorbable
(tricalcium phosphate), bioactive (hy-
droxyapatite, bioactive glasses, and
glass-ceramics), or porous for tissue
ingrowth (hydroxyapatite-coated met-
als, alumina). Applications include re-
placements for hips, knees, teeth,
tendons, and ligaments and repair for
periodontal disease, maxillofacial re-
construction, augmentation and stabi-
lization of the jaw bone, spinal fusion,
and bone fillers after tumor surgery.
Carbon coatings are thromboresistant
and are used for prosthetic heart
valves. The mechanisms of tissue
bonding to bioactive ceramics are be-
ginning to be understood, which can
result in the molecular design of bio-
ceramics for interfacial bonding with
hard and soft tissues. Composites are
being developed with high toughness
and elastic modulus match with bone.
Therapeutic treatment of cancer has
been achieved by localized delivery of
radioactive isotopes via glass beads.
Development of standard test methods
for prediction of long-term (20-year)
R
E
Newnham-contributing editor
Manuscript No. 196865 Received March
3,
1991;
approved April
23, 1991.
Supported by U.S. Air Force Office
of
Scientific
Research under Contract No F496290-88-C-0073.
The conducted research was reviewed and ap
proved by the University
of
Florida Institutional
Animal Care and Use Committee to ensure that
the research procedures adhered to the standards
set forth in the Guide for the Care and Use
of
Laboratory Animals (NIH Publication 85-23) as
promulgated by the Committee on Care and Use
of Laboratory Animals of the Institute
of
Labora-
tory Animal Resources. Commission
on
Life
Sci-
ences, National Research Council.
%rnber, American Ceramic Society.
mechanical reliability under load is
still needed. [Key words: bioceramics,
structure, dental ceramics, interfaces,
mechanics.]
1.
Introduction
MANY
millennia ago, the discovery of
human kind that
fire
would irreversibly
transform clay into ceramic pottery led
eventually to an agrarian society and
an enormous improvement in the qual-
ity and length of
life.
Within the last
four decades another revolution has
occurred in the use of ceramics to im-
prove the quality of
life.
This revolution
is the innovative use of specially de-
signed ceramics for the repair and re-
construction of diseased or damaged
parts of the body. Ceramics used for
this purpose are termed bioceramics.
Bioceramics can be single crystals
(sapphire), polycrystalline (alumina or
hydroxyapatite (HA)), glass (Bio-
glass@*), glass-ceramics (Ceravital@+
or A/W glass-ceramic), or composites
(stainless-steel-fiber-reinforced Bio-
glass@ or
polyethylene-hydroxyapatite
Ceramics and glasses have been
used for a long time in the health-care
industry for eye glasses, diagnostic in-
struments, chemical ware, thermome-
ters, tissue culture flasks, and fiber
optics for endoscopy. Insoluble porous
glasses have been used as carriers for
enzymes, antibodies. and antigens
since they have several advantages,
notably resistance
to
microbial attack,
pH changes, solvent conditions, tem-
perature, and packing under the high
pressure which is required for rapid
(PE-HA)).
*University
of
Florida, Gainesville, FL
'Leitz Gmbh, Wetrlar. FRG
1487
1488
Journal
of
the American Ceramic Society
-
Hench Vol.
74,
No.
7
Table
I.
Types of Implant-Tissue
Response
If the material
is
toxic, the surrounding
tissue dies.
If the material
is
nontoxic and bio-
logically inactive (nearly inert), a
fibrous tissue of variable thickness
forms.
flow.' Ceramics are also widely used in
dentistry as restorative materials, gold
porcelain crowns, glass-filled ionomer
cements, dentures, etc. In these appli-
cations they are called dental ceram-
ics as discussed by Preston.'
This review
is
devoted
to
the use of
bioceramics as implants
to
repair
parts of the body, usually the hard
tis-
sues of the musculo-skeletal system,
such as bones or teeth, although a
brief review of the use
of
carbon coat-
ings for replacement of heart valves
is
also included. Dozens of ceramic com-
positions have been tested;',3 however,
few have achieved human clinical ap-
plication. It
is
now known that clinical
success requires the simultaneous
achievement of a stable interface with
connective tissue and a match of the
mechanical behavior of the implant with
the tissue
to
be replaced.
If
the material
is
nontoxic and bio-
II.
Twes of Bioceramics-Tissue
-.
logically active (bioactive), an inter-
facial bond forms.
Attachment
The mechanism of tissue attachment
the surrounding tissue replaces it.
response at the implant interface.'
No
material implanted in living tissues
is
inert; all materials elicit a response
from living tissues. The four types of re-
sponse (Table
I)
allow different means
of achieving attachment
of
prostheses
to
the musculo-skeletal system. Table
II
summarizes the attachment mecha-
nisms, with examples.
A
comparison
of
the relative chemi-
cal activity of these different types of
bioceramics
is
given in Fig. 1. The
relative reactivity shown in Fig. I(a)
correlates very closely with the rate of
Table
II.
Types of Bioceramics
-Tissue Attachment and Bioceramic Classification
Type
of
bioceramic Type
of
attachment
Example
1
Dense, nonporous, nearly inert
Al2O3
(single crystal and
ceramics attach by bone polycrystalline)
growth into surface
irregularities by cementing
the device into the tissues,
or by press fitting into a
defect (termed morphologi-
cal fixation).
2
For porous inert implants
A1203
(porous polycrystalline)
bone ingrowth occurs, Hydroxyapatite-coated porous
which mechanically metals
attaches the bone to the
material (termed biological
fixation).
reactive ceramics, glasses, Bioactive glass-ceramics
and glass-ceramics attach Hydroxyapatite
directly by chemical
bonding with the bone
(termed bioactive fixation).
4
Dense, nonporous (or porous), Calcium sulfate (plaster of Paris)
resorbable ceramics are Tricalcium phosphate
designed
to
be slowly
replaced by bone.
3
Dense, nonporous, surface- Bioactive glasses
Calcium phosphate salts
formation of an interfacial bond of im-
plants with bone (Fig. l(b)).4 Figure
l(b)
will be discussed in more detail in Sec-
tion
VIII.
The relative level of reactivity of an
implant influences the thickness of the
interfacial zone
or
layer between the
material and tissue. Analysis of failure
of implant materials during the last
20
years generally shows failure origi-
nating from the biomaterial-tissue in-
terfa~e.~,~ When a biomaterial is nearly
inert (type 1 in Table
I1
and Fig. 1) and
the interface
is
not chemically or bio-
logically bonded, there is relative
movement and progressive develop-
ment of a nonadherent fibrous capsule
in both soft and hard tissues. Move-
ment at the biomaterial-tissue inter-
face eventually leads
to
deterioration in
function of the implant or the tissue at
the interface or both. The thickness of
the nonadherent capsule varies greatly
depending upon both material (Fig.
2)
and extent of relative motion.
The fibrous tissue at the interface
with dense, medical-grade alumina im-
plants can be very Consequently,
as discussed later, if alumina implants
are implanted with a very tight me-
chanical fit and are loaded primarily
in
compression, they are successful clini-
cally. In contrast, if a type
1,
nearly
inert, implant
is
loaded such that inter-
facial movement can occur, the fibrous
capsule can become several hundred
micrometers thick and the implant
loosens quickly. Loosening invariably
leads to clinical failure, for a variety
of
reasons, including fracture of the im-
plant or the bone adjacent
to
the
implant. Bone at an interface with
a type
1,
nearly inert, implant is very
often structurally weak because of
disease, localized death of bone (es-
pecially if so-called bone cement,
poly(methy1 methacrylate (PMMA)
is
used), or stress shielding when the
higher elastic modulus of the implant
prevents the bone from being loaded
properly.
The concept behind nearly inert,
microporous bioceramics (type
2
in
Table
II
and Fig. 1)
is
the ingrowth of
tissue into pores on the surface or
throughout the implant, as originated by
Hulbert et
aL3
many years ago. The in-
creased interfacial area between the
implant and the tissues results in an
increased inertial resistance
to
move-
ment of the device in the tissue. The
interface
is
established by the living
tissue in the pores. Figure
3
shows liv-
ing bone grown into the pores of an
alumina bioceramic. This method of
attachment is often termed biological
fixation. It
is
capable of withstanding
more complex stress states than type 1
implants, which achieve only morpho-
logical fixation." The limitation associ-
ated with type
2
porous implants,
however,
is
that, for the tissue
to
remain
July
1991
viable and healthy,
it
is
necessary for
the pores
to
be greater than
100
to
150
pm in diameter (Fig.
2).
The large
interfacial area required for the porosity
is due
to
the need to provide
a
blood
supply
to
the ingrown connective tissue.
Vascular tissue does not appear in
pores which measure less than
100
pm.
If
micromovement occurs at the inter-
face of a porous implant, tissue is
damaged, the blood supply may be cut
off, tissues die, inflammation ensues,
and the interfacial stability can be de-
stroyed. When the material is a metal,
the large increase in surface area can
provide a focus for corrosion of the im-
plant and
loss
of metal ions into the
tissues, which may cause a variety
of
medic a
I
pro b
I
em
s
These pot en
t
i a
I
problems can be diminished by using
a bioactive ceramic material such as
HA as
a
coating on the porous metal,
as first shown by Ducheyne et
a/.’
The
HA coating also speeds the rate of
bone formation in the pores. However,
the fraction of large porosity required
for bone growth in any material de-
grades the strength of the material.
Consequently, this approach to solving
interfacial stability
is
best when used
as porous coatings or when used as
unloaded space fillers in tissues.
Resorbable biomaterials (type
4
in
Table
II
and Fig.
1)
are designed
to
de-
grade gradually over a period of time
and be replaced by the natural host
tis~ue.’~-’~ This leads
to
a
very thin
or nonexistent interfacial thickness
(Fig.
2).
This
is
the optimal solution
to
the problem of biomaterials if the re-
quirements
of
strength and short-term
performance can be met. Natural tis-
sues can repair themselves and are
gradually replaced throughout life by a
continual turnover of cell populations.
As
we grow older, the replacement of
cells and tissues is slower and less ef-
ficient, which is why parts “wear out,”
unfortunately some faster than others.
Thus, resorbable biomaterials are
based on the same principles of repair
Interfacial thickness
(pm)
Initial
Bioceramics: From Concept to Clinic
which have evolved over millions of
years. Complications in the develop-
ment of resorbable bioceramics are
(1)
maintenance of strength and the
stability of the interface during the
degradation period and replace-
ment by the natural host tissue and
(2)
matching resorption rates
to
the
repair rates
of
body tissues (Fig. I(a))>
which themselves vary enor mousl y.
Some dissolve
too
rapidly and some
too
slowly. Because large quantities of
material may be replaced, it is also es-
sential that a resorbable biomaterial
consists only
of
metabolically accept-
able substances. This criterion im-
poses considerable limitations on the
compositional design of resorbable
biomaterials. Successful examples
are resorbable polymers such as
1489
Material interface
1000
100
10
I
0
1
inert)
1
r
I
I
I
III
I
I
I
11
I I
I
T
I
I
I
111
3
10
100
1000
Implantation time
(d)
Fig.
1.
Bioactivity spectrum for various biocerarnic implants
(a)
relative rate of bioreactivity and (b) time dependence of forma-
tion of bone bonding at an implant interface
((A)
45S5
Bioglass@
(6)
KGS CeravitaP (C)
5584
3
Bioglass@
(D)
A/W
glass-ceramic
(E)
HA
(F)
KGX Ceravitala’, and
(G)
AI2O3-Si3N4)
Bone
10
100
1000
I
1000
100 10 1
0
1
10
100
1000
Fig.
2.
Comparison
of
interfacial thickness
of
reaction layer of
bioactive implants
or
fibrous tissue
of
inactive bioceramics in bone
Fig.
3.
lngrowth of bone within >100-pm pores of alumina bioce-
ramic (Photograph courtesy of
S
Hulbert
)
1490
Journal
of
the
American Ceramic Society
-
Hench
Vvl.
74,
No.
7
poly(lactic acid)-poly(glyco1ic acid)
(PLA-PGA) used for sutures, which are
metabolized
to
carbon dioxide and
water and therefore are able to func-
tion for a period and then dissolve and
disappear. Porous or particulate Cal-
cium phosphate ceramic materials
such as tricalcium phosphate (TCP)
are successful materials for resorbable
hard-tissue replacements when only
low mechanical strength is required,
such as in some repairs of the jaw or
head.
Another approach
to
the solution of
the problems of interfacial attachment
is the use of bioactive materials (type 3
in Table
II
and Fig.
1).
The concept of
bioactive materials is intermediate be-
tween resorbable and bi~inert.’,~’~
A
bioactive material is one that elicits a
specific biological response at the inter-
face of the material which results
in
the
formation
of
a bond between the tissues
and the material (shown first in 1969).’4
This concept has now been expanded
to include a large number of bioactive
materials with a wide range of rates of
bonding and thickness of interfacial
bonding layers (Figs.
1
and 2). They in-
~Iude~.~ bioactive glasses such as Bio-
glass? bioactive glass-ceramics such
as Ceravitalo, A/W glass-ceramic, or
machineable glass-ceramics; dense HA
such as durapatite or Calcitite@*, or
bioactive composites such as Palavi-
tat@§, stainless-steel-fiber-reinforced
BioglassB; and PE-HA mixtures.
All
of
these bioactive materials form an inter-
facial bond with adjacent tissue. How-
ever, the time dependence of bonding,
the strength of the bond, the mecha-
nism of bonding, and the thickness of
the bonding zone differ for the various
materials.
It
is important
to
recognize that rela-
tively small changes in the composition
of
a biomaterial can affect dramatically
whether it
is
bioinert, resorbable, or
bioactive. These compositional effects
on surface reactions are discussed in
Section
V.
*Calatek, lnc
,
San
Diego,
CA
‘Leitz Gmbh
111.
Nearly Inert Crystalline
Bioceramics
High -de nsi
t
y, hi gh-pu
r
i t
y
(
>
9 9.5%)
alumina (a-Al2O3) was the first bioce-
ramic widely used clinically. It
is
used
in load-bearing hip prostheses and
dental implants because of its combi-
nation of excellent corrosion resistance,
good biocompatibility, high wear re-
sistance, and high ~trength.’.~~’~-‘~
Although some dental implants are
single-crystal ~apphire,’~ most alumina
devices are very-fine-grained polycrys-
talline
a-Al,O,.
A very small amount of
magnesia (<0.5%) is used as an aid
to
sintering and
to
limit grain growth dur-
ing sintering.
Strength, fatigue resistance, and
fracture toughness of polycrystalline
a-Al2O3 are a function of grain size and
percentage of sintering aid, i.e., purity.
Alumina with an average grain size of
<4
prn
and >99.7% purity exhibits
good flexural strength and excellent
compressive strength. These and other
physical properties are summarized in
Table
Ill6
with the International Stand-
ards Organization
(ISO)
requirements
for alumina implants. Extensive testing
has shown that alumina implants which
meet or exceed
IS0
standards have
excellent resistance
to
dynamic and
impact fatigue and also resist subcriti-
cal crack growth.20 An increase in the
average grain size
to
>7
pm can de-
crease mechanical properties by about
20%. High concentration of sintering
aids must be avoided because they re-
main in the grain boundaries and de-
grade fatigue resistance, especially in
a corrosive physiological environment.’
Methods exist for lifetime predictions
and statistical design of proof tests for
load-bearing ceramics. Applications of
these techniques show that specific
prosthesis load
limits
can be set for an
alumina device based upon the flexural
strength of the material and its use en-
vironment.*’ Load-bearing lifetimes of
30 years at 12000-N loads, similar
to
those expected in hip
joints,
have been
predicted.6 Results from aging and fa-
tigue studies show that it is essential
that alumina implants be produced at
the highest possible standards of qual-
Table
111.
Physical Characteristics
of
Alumina and Partially Stabilized Zirconia (PSZ) Bioceramics
IS0
alumina Cortical Cancellpus
High alumina ceramics* standard
6474
PSZ*
bone+ bone
Content (percent by weight)
A1203
>
998
A1203
2
9950
Zr02
>
97
Density (g/cm3)
>
393
z
390
5
6-6 12
1
6-2
1
Average grain size (pm)
3-6
c7
1
Compressive strength (MPa) 4500
2-12
Surface roughness,
R,
(pm)
0
02
0
008
Hardness (Vickers),
HV
2300
>
2000 1300
Bendina strenath IMPa)
550
400
1200
50-150
(aftertesting in Ringer’s solution)
Young’s Modulus (GPa) 380
200
7-25 0.05-0.5
Fracture toughness,
Kc
(MPa. m-1’2) 5-6 15 2-12
Slow crack growth,
n
(unitless) 30-52 65
*Reference
6
‘References 120 and 121 *Reference 122
July
1991
ity assurance, especially
if
they are
to
be used as orthopedic prostheses in
younger patients
(-40
years old).
Alumina has been used in orthope-
dic surgery for nearly
20
years, moti-
vated largely by
(1)
its excellent type
1
biocompatibility and very thin capsule
formation (Fig.
2)
which permits ce-
rnentless fixation
of
prostheses3 and
(2)
its exceptionally low coefficients of
friction and wear
rate^.^,^
The superb tribiologic properties
(friction and wear) of alumina occur only
when the grains are very small
(14
pm)
and have a very narrow size distribu-
tion. These conditions lead
to
very low
surface roughness values
(R,50.02
pm,
Table
Ill).
If large grains are present,
they can pull out and lead
to
very rapid
wear because of local dry friction and
abrasion caused by the alumina grains
in the joint-bearing surfaces.6
Alumina on alumina load-bearing
wearing surfaces, such as in hip pros-
theses, must have a very high degree
of sphericity produced by grinding and
polishing the two mating surfaces to-
gether. The alumina ball and socket in
a hip prosthesis are polished together
and used as a pair. The long-term coef-
ficient of friction of an alumina-alumina
joint decreases with time and ap-
proaches the values of a normal joint.
This leads
to
wear of alumina on alu-
mina articulating surfaces that are
nearly
10
times lower than metal-PE
surfaces (Fig,
4).
Low wear rates have
led
to
wide-
spread use in Europe of alumina non-
cemented cups press fitted into the
acetabulum (socket) of the hip. The
cups are stabilized by bone growth into
grooves or around pegs. The mating
femoral ball surface is also of alumina
which
is
bonded to a metallic stem.
Long-term results in general are excel-
lent, especially for younger patients.
However, Christel
et
a/.6
caution that
stress shielding of the bone can occur.
This is due
to
the high Young's modu-
lus
of alumina (Table
Ill),
which pre-
vents the bone from being loaded, a
requirement for bone
to
remain healthy
and strong. The Young's modulus of
cortical bone ranges between
7
and
25
GPa (as discussed in Section
X)
which
is
10
to
50
times lower than alu-
Bioceramics:
From Concept
to
Clinic
mina. Christel et
a/.6
report that stress
shielding may be responsible for can-
cellous bone atrophy and loosening of
the acetabular cup
in
older patients
with senile osteoporosis or rheumatoid
arthritis. Consequently, it is essential
that the age of the patient, nature of the
disease of the joint, and biomechanics
of the repair
be
considered carefully be-
fore any prosthesis is used, including
those made from alumina ceramics. In
the United States, the primary use of
alumina is for the ball of the hip joint
(Fig.
5),
with the acetabular component
being ultrahigh-molecular-weight
PE.
Other clinical applications of alumina
prostheses, reviewed by Hulbert et
a/.,3
include knee prostheses, bone screws,
alveolar ridge (jaw bone) and maxillo-
facial reconstruction, ossicular (middle
ear) bone substitutes, keratoprosthe-
ses (corneal replacements), segmental
bone replacements, and blade and
screw and post-type dental implants.
IV.
Porous Ceramics
The potential advantage offered by a
porous ceramic implant (type
2,
Table
II,
Figs.
1
and
2)
is its inertness
combined with the mechanical stability
of the highly convoluted interface devel-
oped when bone grows into the pores
of the ceramic. Mechanical require-
ments of prostheses, however, severely
restrict the use of low-strength porous
ceramics
to
low-load- or non-load-
bearing applications. Studies show
that, when load bearing is not a pri-
mary requirement, nearly inert porous
ceramics can provide a functional im-
plant,1.3.22-24 When pore sizes exceed
100
pm, bone will grow within the inter-
connecting pore channels near the sur-
face and maintain its vascularity and
long-term viability (Fig.
3).
In this man-
ner the implant serves
as
a structural
bridge and model or scaffold for bone
formation. The microstructures of cer-
tain corals make an almost ideal in-
vestment material for the casting of
1491
g
0'15
[
-
Friction
I
.-
r:
lb
100
id00
1
&oo
Testing time
(h)
Fig.
4.
Time dependence of
(-)
coefficient of friction and
(---)
index of wear of alumina-alumina versus metal-PE hip joint (in vitro
testing)
Fig.
5.
Medical-grade alumina used as femoral balls in total hip
replacement
Note three alternative types of metallic stems used for
morphological fixation
(Photograph courtesy of
J
Parr)