Biomaterials 23 (2002) 3673–3680

A new method to produce macropores in calcium phosphate cements

R.P. del Reala, J.G.C. Wolkeb, M. Vallet-Reg!ıa, J.A. Jansenb,*
a
!
Facultad de Farmacia, Departamento de Qu!ımica Inorganica !
y Bioinorganica, UCM 28040, Madrid, Spain
b
Department of Biomaterials, College of Dental Science, University Medical Center ‘Nijmegen’, P.O. Box 9101,
NL-6500 HB Nijmegen, The Netherlands
Received 5 October 2001; accepted 12 March 2002

Abstract

A new way to create macropores in calcium phosphate cements has been developed. The method consists in adding NaHCO3 to
the starting cement powder (Biocement D) and using two different liquids: first a basic liquid to form the paste and later an acid
liquid to obtain CO2 bubbles. Mercury intrusion measurements showed a dramatic increase both in macropores with an average size
of 100 mm and in the total porosity (even higher than 50% with respect to the Biocement D). This method does not change in
any significant way the final reaction products of the starting material after being soaked 3 days in Ringer solution. Only, due
to the increase of the porosity, the compressive strength of the porous cement decreases significantly. r 2002 Elsevier Science Ltd.
All rights reserved.

Keywords: Porosity; Calcium phosphate cement; Resorbable bone substitute; Tricalcium phosphate

1. Introduction passive one. The osteoclastic cells produce a pH close to

5.5, which in turn increases the dissolution rate of the
Calcium phosphate cements (CPCs) show several implant. Usually this type of resorption only occurs on
advantages with respect to other materials which are the cement surface because the pores as present in the
used for bone repair. For example, they are injectable, cements do not allow the penetration of cells or blood
easy to shape and can be maintained locally. Therefore, vessels in the material [2–4].
they are very effective to fill bone defects with an Already several attempts have been made to
irregular shape. Furthermore, CPCs are very bone improve the resorption behaviour of CPCs e.g. by
compatible and osteoconductive. On the other hand, increasing the porosity of the material. This can be
they have poor mechanical properties. Currently, this achieved by using the so-called calcium phosphate
prevents their use in loaded conditions. This last emulsion technique [5,6]. So far, the most extensive
problem is even enhanced by the fact that the in vivo experiments have been done by mixing the CPC with
resorption of CPC is very slow. crystals of the right dimensions of highly soluble and
In view of this, two types of resorption can be non-toxic compounds, such as sucrose [7] or mannitol
distinguished, i.e., passive and active. Passive resorption [8]. After the complete hardening of the material the
is due to the dissolution rate of the material into the macrocrystals are removed just by soaking the samples
body fluids and it depends on the final components of in water.
the set cement. Further, this type of resorption is Although both methods have their merits, the
determined by the porosity of the samples, ionic disadvantage is that the porosity cannot be created
substitutions, crystallinity and pH of the cement–tissue during setting of the cement in the in vivo environment.
interface [1]. For the emulsion technique, an additional heat treat-
Active resorption is due to cellular activity (like ment is required to achieve porosity. Addition of sucrose
osteoclast-like cells) and is, in this way, related with the or mannitol requires dissolution of these components
after application and setting of the cement in the bone
*Corresponding author. Tel.: +31-24-3614006; fax: +31-24-
defect in order to get macroporosity.
3614657. Therefore, the aim of this work has been the
E-mail address: [email protected] (J.A. Jansen). development of another method to increase the

0142-9612/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved.
PII: S 0 1 4 2 - 9 6 1 2 ( 0 2 ) 0 0 1 0 1 - 1
3674 R.P. del Real et al. / Biomaterials 23 (2002) 3673–3680

macroporosity of CPCs, which is based on the creation height) in a retrograde fashion. Four different prepared
of gas bubbles during the fabrication process. samples were used in this study (see Table 1).

2.3. Analysis methods

2. Materials and methods
First, we determined initial setting time (the max-
2.1. Starting calcium phosphate cement imum time for the paste to be deformed without
damaging the structure of the solidifying cement) of
For our studies we used the so-called Biocement D [9], the various cements. Therefore, a bronze block was used
and this was kindly provided by Merck Biomaterials. as a mould containing six holes with the earlier
The powder component of this material consists of a mentioned dimensions. The bronze block was placed
mixture of the following compounds: 62.5% a-trical- in a water bath and kept at 371C. Subsequently, a
cium phosphate (a-TCP), 26.8% dicalcium phosphate standard Gillmore needle was used for this purpose. The
anhydrous (DCPA), 3.9% CaCO3 and 1.8% hydro- measurements were statistically evaluated using a one-
xyapatite (HAp). An aqueous dissolution of Na2HPO4 way analysis of variance (ANOVA) and a multiple
(4% in weight) is used as the liquid component. The comparison procedure (Tukey’s test).
ideal liquid/powder ratio for clinical applications of the In order to evaluate the porosity of the samples, we
cement has been shown to be 0.35 ml/g [10]. A used a mercury intrusion (Micromeritics Autopore III)
commercially available syringe (Sherwood medical and N2 adsorption (Micromeritics ASAP 2010C)
monoject 2 ml) is applied for delivery of the cement technique. With the mercury intrusion, we measured
[11]. The syringe is closed at the tip with a plastic the pore size distribution in the range 300–0.003 mm and
stopper and filled with 1 g of the standard cement with N2 adsorption pore sizes ranging from 300 to
powder. After adding 0.35 ml of Na2HPO4 (4%), the 1.7 nm. The porosity can be expressed by the following
syringe is closed with the injection plunger and placed in equation:
a mixing apparatus (Silamat). After mixing during 15 s, P ¼ rbulk VHg ;
the syringe is taken out of the mixing apparatus, the
stopper removed, and the cement can be injected. where P; rbulk and VHg are the porosity, bulk density of
the material and mercury intrusion volume per gram,
2.2. Preparation of porous calcium phosphate cement respectively. VHg can be split into those in the three
regions of interest, i.e.,
To create porous cement, NaHCO3 was added to the P ¼ rbulk ðV30010 þ V100:1 þ V0:10:003 Þ;
above-mentioned solid components of the starting
powder. All components (0.5 g of standard powder plus where V30010 ; V100:1 and V0:10:003 are the mercury
0.05 g of NaHCO3) were placed into the syringe with intrusion volumes per gram in the specific pores-size
Na2HPO4 (2%) aqueous dissolution or pure water and, intervals (expressed in mm). We can, therefore, assign a
after a mixing time of 15 s, the plunger was removed and porosity for each interval. The sum of the three intervals
NaH2PO4 (8%) aqueous dissolution (pH 4.15) was is the total porosity.
added. Thereafter, the syringe was shaken again using In order to measure the compressive strength, six
different shaking times. This second step, addition of samples of every composition were removed from the
NaH2PO4, accelerates the hardening of the cement and moulds after 1 day setting and then soaked in Ringer
also provides the decrease of pH necessary to favour the solution during 3 days at 371C. Thereafter, the non-
production of H2CO3 (which is decomposed in H2O and dried samples were placed in an Instron testing bench
CO2) from NaHCO3. Finally, the material was injected and compressive strength was measured at 1 mm/min
into a cylindrical mould (6 mm in diameter, 12 mm in crosshead speed.

Table 1
Composition of the samples and method to prepare them

Sample 1 Sample 2 Sample 3 Sample 4

Standard (g) 0.5 0.5 0.5 0.5

NaHCO3 (g) 0.05 0.05 0.05 0.05
Basic liquid (ml) Na2HPO4 (2%) 0.12 Na2HPO4 (2%) 0.12 Water 0.12 Water 0.12
First shaking time (s) 15 15 15 15
Acid liquid (ml) NaH2PO4 (8%) 0.12 NaH2PO4 (8%) 0.15 NaH2PO4 (8%) 0.12 NaH2PO4 (8%) 0.15
Second shaking time (s) 2 0.5 2 0.5
 R.P. del Real et al. / Biomaterials 23 (2002) 3673–3680 3675

In order to obtain X-ray diffraction patterns (Philips The samples made using water as first liquid showed a
PW 1830), the samples were removed from the moulds significantly lower setting time ðPo0:05Þ:
after 1 day and soaked in Ringer solution for 3 days.
Then the samples were placed into an oven at 501C
3.2. Porosity
during 2 days to dry them and later they were triturated
using an agate mortar. The measurements were made in
3.2.1. Mercury intrusion porosity
y  2y conditions using Cu Ka radiation.
The final porosity of all samples as well as the increase
Finally, specimens were examined using a scanning
(%) of the porosity with respect to the standard is given
electron microscope (JEOL 6400-LINK AN 10000 at
in Table 3. Fig. 1 shows the mercury intrusion volume
20 kV) to check their porosity as well as to obtain
per gram for a standard sample and sample 2. The
information about their structural morphology.
curves of samples 1, 3 and 4 are qualitatively similar to
sample 2. The mercury intrusion for standard sample
shows a bimodal distribution of pore sizes, centered at
0.02 and 0.5 mm (Fig. 2). We have to notice that no pores
3. Results
with a diameter larger than 1 mm are present. Sample 2
shows three important contributions to the pore size
3.1. Initial setting times
distribution, centered at 0.02, 0.6 and 100 mm (Fig. 3).
This is the same distribution as the standard sample
The initial setting times of the samples are listed in
considering pores o1 mm. On the other hand, in this
Table 2. The statistical analysis showed no differences
sample also pores with diameter around 100 mm are
between samples made using Na2HPO4 as first liquid.
present.
The porosity data of the different pore size intervals
Table 2 as well as the total porosity of the various samples are
Initial setting time listed in Table 4. For the standard sample, the interval
with the highest porosity is 10–0.003 mm. For the porous
Sample 1 Sample 2 Sample 3 Sample 4
cements (samples 1–4), the porosity due to large pores
Setting time (min) (S.T.) 4.170.4 3.970.2 5.170.2 6.470.7 (10–300 mm) ranges between 13.2 and 20.2.
The mercury-intrusion volume for pores below 1 mm is
larger for sample 2 than for the standard. If we suppose
that the pore size distribution and intrusion volume per
Table 3
Total porosity of the samples and percentage of increase with respect
gram below 10 mm are the same for all the samples (not
to the standard extra pores in the range 10–0.003 mm but only large
pores would be created with this method), the difference
Standard Sample 1 Sample 2 Sample 3 Sample 4
in the intrusion volume per gram in this range (0.003–
Porosity (%) 39.4 51.8 59.8 55.6 56.1 10 mm), between porous samples and sample standard,
DP  100=PðstÞ 0 31.5 51.8 41.1 42.4 could be assigned to macropores whose unique penetra-
DP is the porosity difference between samples 1–4 and the standard tion channels for the mercury are smaller than 10 mm.
P(st). Further, the intrusion volume above 10 mm in sample 2

Fig. 1. Mercury intrusion volume for sample standard and sample 2.

3676 R.P. del Real et al. / Biomaterials 23 (2002) 3673–3680

Fig. 2. N2 adsorption isotherm curve for sample standard.

Fig. 3. Pore size distribution and SEM micrographs (magnification  2000 for the left and  200 for the right micrograph) for the standard sample.

Table 4
shows the volume of pores reached by the mercury
Porosity assigned to different pore size intervals through channels of at least this diameter. Therefore,
the total mercury intrusion related to the range
Standard Sample 1 Sample 2 Sample 3 Sample 4
300–10 mm would be the sum of both contributions,
P300210 0.7 13.2 20.2 19.1 16.6 i.e., the volume difference between the standard and the
P1020:1 19 21.7 24.6 21.8 25.5 porous sample in the range 10–0.003 mm and the volume
P0:120:003 19.7 16.9 15 14.7 14
intrusion shown by the porous sample above 10 mm.
Total porosity 39.4 51.8 59.8 55.6 56.1
Under this assumption, the percentage of large pores
 R.P. del Real et al. / Biomaterials 23 (2002) 3673–3680 3677

reached by the mercury through large channels micrograph on the left shows the pores with diameter
(>10 mm) would be 59, 60, 69 and 57 for samples 1, 2, around 1 mm. Different aspects of the samples can be
3 and 4, respectively. noticed. The standard sample shows larger particles
whereas the particle size of sample 2 is more homo-
3.2.2. Nitrogen adsorption geneous.
Fig. 4 shows the N2 adsorption isotherm for the
standard sample. The curve can be identified as a type 3.5. X-ray diffraction patterns
III isotherm corresponding to a non-porous solid with a
nearly continuous distribution of pores. The curves for Fig. 5 shows the X-ray pattern of the standard sample
samples 1–4 are similar to this one. The plot of and sample 2. The patterns of the other samples (1, 3
dV =d log D vs. pore diameter shows a monomodal and 4) are completely similar to the pattern of sample 2.
distribution centered on 0.004 mm in all specimens, The standard sample shows HAp as the predominant
which corresponds with the non-completely defined phase as well as some amounts of DCPA and a-TCP.
maximum around 0.003 mm in the mercury intrusion For samples 1–4, DCPA disappears and the amount of
curves. Table 5 shows the BET specific area for the a-TCP decreases whereas the low crystallinity remains.
samples as well as the most probable pore size measured Therefore, no new phases appear due to the process of
by gas desorption. creating pores inside the material.

3.3. Compressive strength

4. Discussion
The compressive strengths of the samples are listed in
Table 6. A significant decrease with respect to the Biocement D, as used in the current experiment, is an
standard sample can be noticed in all the cases ðPo0:05Þ: already extensively studied CPC. This material has been

3.4. Scanning electron microscopy Table 5

Specific area and most probable pore size measured by N2 adsorption
Figs. 2 and 3 show SEM micrographs at several Standard Sample 1 Sample 2 Sample 3 Sample 4
magnifications together with the pore size distribution of
the sample standard and sample 2, respectively. Samples Specific area (m2/g) 15.1 11.4 12 11.7 12
Pore size (nm) 3.8 4 3.8 3.8 3.8
1, 3 and 4 show results similar to sample 2. The

Fig. 4. Pore size distribution and SEM micrographs (magnification  2000 for the left and  30 for the right micrograph) for sample 2.

Table 6
Compressive strength

Standard Sample 1 Sample 2 Sample 3 Sample 4

Compressive strength (MPa) 1573 2.170.3 1.370.6 1.670.5 1.270.5

3678 R.P. del Real et al. / Biomaterials 23 (2002) 3673–3680

Fig. 5. X-ray diffraction patterns for sample standard and sample 2.

proven to possess optimal clinical parameters such as A pilot study was done to determine the optimal
cohesion time (minimum time required to avoid the concentration and the amount of liquid. We noticed that
disintegration due to the body fluids after the implanta- the best results were obtained with 8% NaH2PO4
tion), initial setting time (the maximum time for the instead of 4%. The use of Na2HPO4 (4%) as basic
paste to be deformed without damaging the structure of liquid, together with NaH2PO4 (8%) as acid liquid
the solidifying cement) and final setting time (the reduced dramatically the injectability. The final pH of
maximum time for the material to be touched without the reaction among the basic liquid, NaHCO3 and the
scratching it) [9,10,12–14], as well as a nearly total acid liquid was 7.3, which means that the addition of an
injectability. Besides, it shows an adequate compressive acid liquid does not affect the final pH of the cement.
strength (B50 MPa) [12], which enables it to be used As presented in our results, the percentage of
in compression-loaded applications (for human trabe- macropores accessible by large channels ranges for the
cular bone the maximum compressive strength is porous cements between 57 and 69%. Theoretically, this
30 MPa). percentage of macropores will be suitable for the
Further, Biocement D is a biocompatible and penetration of blood vessels and cells inside the material.
osteoconductive material [15] similar to other CPCs Besides, the increase of the total porosity will contribute
[16,17]. Animal studies have proven that it is already to the increase of the passive resorption with respect to
covered by new deposited bone after two weeks of the standard cement.
implantation in trabecular bone in goats. The material is The use of less acid liquid will not produce an optimal
degraded following the natural bone remodelling amount of bubbles and the use of more acid liquid will
process [15]. Additional in vivo experiments showed impoverish dramatically the initial setting time due to
that the passive resorption for this material is very slow the increase of the liquid/powder ratio. Therefore, our
[18], despite the low crystallinity (see Fig. 5) and the pilot studies indicated that the optimal liquid/powder
presence of carbonateapatite. This confirms that the ratio to create porosity seems to be 0.49 ml/g. This is still
absorption rate of the material can only be improved by larger than used for the standard cement [10]. This
enhancing the osteoclast activity [11,19,20]. Therefore, larger ratio implies that some increase of the initial
we suggested that interconnected pores larger than setting time will occur with respect to the standard
100 mm have to be created in the cement in order to cement [9].
allow the penetration of cells and blood vessels [2–4]. Further, we used two different liquid components to
We decided to add NaHCO3 to achieve this. NaHCO3 create the porosity. In this procedure, it is essential that
has been used before in CPCs, mainly to increase the the powder and the first liquid are completely mixed. We
carbonateapatite content aiming on an increase of the noticed that the use of water instead of Na2HPO4 (2%)
passive resorption [21,22]. However, the calcium carbo- did not affect in a decisive way the total porosity or the
nate content of the original standard cement already compressive strength of the various cements. Especially,
performs this task. Therefore, the main reason that we the compressive strength appeared not to be affected by
used NaHCO3 was to create macropores inside the the use of two liquids or the increase of the liquid/
material by means of CO2 bubbles in order to improve powder ratio (Table 6). The major problem for a clinical
the active resorption. application point of view will be the effect of the
 R.P. del Real et al. / Biomaterials 23 (2002) 3673–3680 3679

porosity increase on the final compressive strength [23]. which will also increase the final solubility of the
The achieved low values make these cements non- cement.
suitable for weight bearing situations. They can only be
applied in situations where a low stress is present and a
fast bone ingrowth and cement resorption is required.
5. Conclusions
We also have to notice that the value of the compressive
strength for the standard cement does not correspond
In order to increase the active as well as the passive
with earlier reported values [12]. This is due to the fact
resorption of injectable CPCs, we have developed a new
that the surface of our samples was not specially
method to create macropores. On the basis of our data,
prepared for the measurements as was done for the
we conclude that the addition of NaHCO3 to the cement
previous measurements (samples 1–4 are too brittle to
powder result in an increase of porosity of more than
be grained).
50%. The final biological feasibility and effect of our
Our data also showed that the use of water increases
method has still to be proven by in vivo experiments.
the initial setting time with respect to the samples made
with Na2HPO4 (Table 2). The way of measuring the
final setting time by means of the Gillmore needle
appeared to be inappropriate for our porous materials. Acknowledgements
The compressive stress as exerted during these measure-
ments is around 4 MPa. This is already twice the This research was supported by the Dutch Technol-
compressive strength of the samples (Table 6) as ogy Foundation STW and by CICYT, Spain, through
achieved by final setting. Therefore, we determined only research project MAT99-0466. R. P. del Real is grateful
the initial setting time of our samples. to Consejer!ıa de Educacion
! y Cultura (CAM.). We also
To determine the porosity of our cement, we used thank A. Rodr!ıguez (Electron Microscopy Center,
mercury intrusion as well as N2 adsorption method. It Complutense University) for the valuable technical
can be suggested that we are only able to measure open and professional assistance.
connected macropores on the surface of the cement and
not closed macropores inside the cement sample. We
consider both methods as complimentary. The mercury References
technique provides information about large pores as
present at the outside of the cement. However, mercury [1] LeGeros RZ, Parsons JR, Daculsi G, Driessens F, Lee D, Liu ST,
will also penetrate into the cement via channels running Metsger S, Peterson D, Walker M. Significance of the porosity
from the outside to the inner surface. From a physical and physical chemistry of calcium phosphate ceramics. Biode-
point of view, there is no difference between channels gradation–bioresorption. Ann N Y Acad Sci 1988;523:268–71.
[2] Holmes R. Bone regeneration within a coralline hydroxyapatite
and macropores. The N2 adsorption does not provide implant. Plast Reconstr Surg 1979;63:626–33.
reliable information about macropores. These are [3] Shimazaki K, Mooney V. Comparative study of porous hydro-
considered with this technique as outer surface. There- xyapatite and tricalcium phosphate as bone substitute. J Orthop
fore, N2 adsorption is especially suited for the low range Res 1985;3:301–10.
.
[4] Eggli PS, Muller W, Schenk RK. Porous hydroxyapatite and
of pore diameters. Besides, a recent in vivo study
tricalcium phosphate cylinders with two different pore size ranges
confirmed that osteoblastic and osteoclastic cells are implanted in the cancellous bone of rabbits. Clin Orthop
able to penetrate into the cement porosity [24]. This 1988;232:127–38.
resulted in increased bone deposition as well as [5] Bohner M, Gasser B, Mathys R. Calcium phosphate emulsions.
resorption till deep inside the cement material. Conse- In: Proceedings of the World Biomaterials Congress, Hawaii,
quently, it can be supposed that the majority of the 2000, Abstract No. 20.
[6] Bohner M. Calcium phosphate emulsions: possible applications.
pores inside the cement are connected. In: Giannini S, Moroni A, editors. Bioceramics 13. Zuerich: Trans
The use of the acid liquid dissolves the DCPA and a- Tech Publications Ltd., 2001. p. 65–768.
TCP particles due to the decrease of the pH. This [7] Takagi S, Chow LC. Formation of macropores in calcium
provides a more homogeneous aspect to the cement as phosphate cement implants. J Mater Sci: Mater Med
can be seen in the SEM micrographs (the porous 2001;12:135–9.
[8] Markovic M, Takagi S, Chow LC. Formation of macropores in
material has a much smoother surface appearance). calcium phosphate cements through the use of mannitol crystals.
Finally, the X-ray patterns showed that all porous In: Giannini S, Moroni A, editors. Bioceramics 13. Zuerich: Trans
samples had a low crystallinity. Still, the amount of Tech Publications Ltd., 2001. p. 773–6.
HAp increased with respect to the standard. We know [9] Khairoun I, Boltong MG, Driessens FCM, Planell JA. Effect of
calcium carbonate on clinical compliance of apatitic calcium
that not all added NaHCO3 will react in the acid
phosphate bone cement. J Biomed Mater Res (Appl Biomater)
environment by creating CO2. This means that a part of 1997;38:356–60.
the NaHCO3 together with CaCO3 (already included in [10] Driessens FCM, Khairoun I, Boltong MG, Planell JA. Design of
the standard powders) will form carbonateapatite [21], a calcium phosphate bone cement suitable for the fixation of
3680 R.P. del Real et al. / Biomaterials 23 (2002) 3673–3680

metal endoprostheses. In: Sedel L, Rey Ch, editors. Bioceramics [18] Driessens FCM, Boltong MG, Zapatero MI, Verbeeck RMH,
11. Oxford: Elsevier Science Ltd., 1997. p. 279–82. Bonfield W, Bermudez O, Fernandez E, Ginebra MP, Planell JA.
[11] Ooms EM, Wolke JGC, van der Waerden JPCM, Jansen JA. In vivo behaviour of three calcium phosphate cements and a
Trabecular bone response to injectable calcium phosphate (Ca-P) magnesium phosphate cement. J Mater Sci: Mater Med
cement. J Biomed Mater Res, in press. 1995;6:272–8.
[12] Driessens FCM, Boltong MG, de Maeyer EAP, Vecruysse CWJ, [19] Frankenburg EP, Goldstein SA, Bauer TW, Harris SA, Poser
Wenz R, Verbeeck RMH. Comparative study of some experi- RD. Biomechanical and histological evaluation of a calcium
mental or commercial calcium phosphate bone cements. In: phosphate cement. J Bone Jt Surg Am 1998;80:1112–24.
LeGeros RZ, LeGeros JP, editors. Bioceramics 11. Singapore: [20] Yuan H, Li Y, de Brujin JD, de Groot K, Zang X. Tissue
World Scientific, 1998. p. 231–3. responses of calcium phosphate cement: a study in dogs.
[13] Khairoun I, Driessens FCM, Boltong MG, Planell JA, Wenz R. Biomaterials 2000;21:1283–90.
Addition of cohesion promotors to calcium phosphate cements. [21] Miyamoto Y, Toh T, Ishikawa K, Yuasa T, Nagayama M, Suzuki
Biomaterials 1999;20:393–8. K. Effect of added NaHCO3 on the basic properties of apatite
[14] Khairoun I, Boltong MG, Driessens FCM, Planell JA. Some cement. J Biomed Mater Res 2001;54:311–9.
factors controlling the injectability of calcium phosphate bone [22] Miyamoto Y, Toh T, Yuasa T, Takechi M, Momota Y,
cements. J Mater Sci: Mater Med 1998;9:425–8. Nagayama N, Ishikawa K, Suzuki K. Basic properties of apatite
[15] Wolke JGC, Ooms EM, Jansen JA. In vivo resorption behaviour cement containing carbonate apatite and its resorption by
of a high strength injectable calcium-phosphate cement. In: cultured osteoclasts. In: Giannini S, Moroni A, editors. Biocera-
Giannini S, Moroni A, editors. Bioceramics 13. Zuerich: Trans mics 13. Zuerich: Trans Tech Publications Ltd., 2001.
Tech Publications Ltd., 2001. p. 793–6. p. 829–32.
[16] Munting E, Mirtchi AA, LeMaitre J. Bone repair of defects filled [23] Ishikawa K, Osaoka K. Estimation of ideal mechanical strength
with a phosphocalcic hydraulic cement: an in vivo study. J Mater and critical porosity of calcium phosphate cement. J Biomed
Sci: Mater Med 1993;4:337–44. Mater Res 1995;29:1537–43.
[17] Jansen JA, De Ruijter JE, Schaeken HG, Van der Waerden [24] del Real R, Ooms E, Wolke J, Vallet-Regi M, Jansen J. In vivo
JPCM, Planell JA, Driessens FCM. Evaluation of tricalciumpho- bone response to porous calcium phosphate cement. J Biomed
sphate/hydroxyapatite cement for tooth replacement: an experi- Mater Res, submitted for publication.
mental animal study. J Mater Sci: Mater Med 1995;6:653–7.