Materials Science and Engineering C 37 (2014) 60–67

Contents lists available at ScienceDirect

Materials Science and Engineering C journal homepage: www.elsevier.com/locate/msec

Structure, properties and animal study of a calcium phosphate/calcium sulfate composite cement Wei-Luen Chen a, Chang-Keng Chen a, Jing-Wei Lee b, Yu-Ling Lee a, Chien-Ping Ju a, Jiin-Huey Chern Lin a,⁎ a b

Department of Materials Science and Engineering, National Cheng-Kung University, 70101 Tainan City, Taiwan, ROC Section of Plastic Surgery, Department of Surgery, National Cheng-Kung University Medical College and Hospital, 70403 Tainan, Taiwan, ROC

a r t i c l e

i n f o

Article history: Received 24 September 2013 Received in revised form 19 December 2013 Accepted 24 December 2013 Available online 5 January 2014 Keywords: Animal study Bone substitutes Biphasic calcium based ceramic Osteoconduction

a b s t r a c t In-vitro and in-vivo studies have been conducted on an in-house-developed tetracalcium phosphate (TTCP)/ dicalcium phosphate anhydrous (DCPA)/calcium sulfate hemihydrate (CSH)-derived composite cement. Unlike most commercial calcium-based cement pastes, the investigated cement paste can be directly injected into water and harden without dispersion. The viability value of cells incubated with a conditioned medium of cement extraction is N 90% that of Al2O3 control and N80% that of blank medium. Histological examination reveals excellent bonding between host bone and cement without interposition of fibrous tissues. At 12 weeks-post implantation, significant remodeling activities are found and a new bone network is developed within the femoral defect. The 26-week samples show that the newly formed bone becomes more mature, while the interface between residual cement and the new bone appears less identifiable. Image analysis indicates that the resorption rate of the present cement is much higher than that of TTCP or TTCP/DCPA-derived cement under similar implantation conditions. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Bioresorbable bioceramic has become one of the most promising bone substitute materials today. Calcium phosphate and calcium sulfate are two typical bioresorbable materials of such category [1–6]. It is known that when the resorption rate of a bioresorbable implant material is adjusted to be similar to the growth rate of natural bone, the implanted material can be gradually replaced by a new bone [7]. Furthermore, when appropriate kinds and amounts of calcium-based powder and setting solution are mixed, an injectable calcium-based cement paste may be formed which can be used in bonding, filling and repairing damaged natural bone for orthopedic, dental, maxillofacial and other applications via minimally invasive procedures [8–10]. Despite their many advantages, such as being highly biocompatible, osteoconductive, non-exothermic and X-ray detectable, most currently-used calcium phosphate and calcium sulfate bone substitute materials have their respective disadvantages. For example, calcium phosphates demonstrate bioresorption rates

⁎ Corresponding author at: Department of Materials Science and Engineering, National Cheng-Kung University, No. 1 University Road, 70101 Tainan City, Taiwan, ROC. Tel./fax: +886 6 274 8086. E-mail address: [email protected] (J.-H.C. Lin). 0928-4931/$ – see front matter © 2014 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.msec.2013.12.034

which are often clinically too low [11,12], leading to a long time (even longer than a year [13,14]) for the resorption process to be completed. On the other hand, calcium sulfates often show dissolution rates which are too high to allow new bone to grow into bone cavities in the most effective way (“one-to-one” resorption) [6,7]. It seems logical to combine calcium phosphate and calcium sulfate into a composite formula and, if appropriately processed, expect to see the inherent benefits from each component of the composite. One major advantage for making such a composite cement implant may be its easy adjustment in resorption rate by adjusting the weight or volume ratio of the two components. An adjustable resorption rate is desirable considering the fact that the resorption rate of a calcium-based implant is usually related to many material factors, such as chemical composition, phase content, degree of crystallization, and porosity, and that different resorption rates are often recommended for different applications/implantation sites [15–17]. Following this philosophy, a series of tetracalcium phosphate (TTCP)/dicalcium phosphate anhydrous (DCPA)/calcium sulfate hemihydrate (CSH)-derived composite cements have recently been developed in the present authors' laboratory [18]. Preliminary tests indicated that one of the most promising candidates of this series of calcium-based composite cements is comprised of a TTCP/DCPA/CSH mixed powder with a weight ratio of 2.69:1:4.51. Reported in the present study are experimental results of this particular calciumbased cement, including crystal structure/phases, morphology, some

W.-L. Chen et al. / Materials Science and Engineering C 37 (2014) 60–67

physical and mechanical properties, along with cytotoxicity and rabbit implantation study.

61

weight loss measurement were dried in anhydrous ethanol at 50 °C for 1 day after being removed from the Hanks' solution. The weight loss ratios were obtained from the equation,

2. Materials and methods Weight loss ratio ð% Þ ¼ ðW0 −Wt Þ=W0  100; 2.1. Cement preparation and working time/setting time measurement The TTCP powder used for the study was fabricated in-house using a method suggested by Brown and Epstein [19]. To prepare the composite cement for the study, this TTCP powder, a commercial DCPA powder and a commercial CSH powder were uniformly mixed with a weight ratio of 2.69:1:4.51. The composite powder was then mixed uniformly with a 0.6 M (NH4)2HPO4 setting solution at a liquid/powder (L/P) ratio of 0.38 ml/g to form a cement paste. Working time of the cement paste was determined by the time after that the cement paste was no longer workable, while the setting time was measured according to the standard method set forth in ISO 1566 for dental zinc phosphate cements. The cement was considered set when a 400 g weight loaded onto a Vicat needle with a 1 mm dia. tip failed to make a perceptible circular indentation on the surface of the cement. In the course of the test, the cement was kept in a 50–70% relative humidity environment at 24 °C. Each average working time or setting time value was obtained from 4 measurements taken under the same testing conditions. To evaluate the cement dispersion behavior in water, the cement paste after being mixed for 1 min was quickly loaded into a 5 ml syringe with needle removed and directly injected into 37 °C deionized water. In so doing, whether or when the cement was dispersed in water could be easily determined by the naked eye. 2.2. pH measurement The pH values of the Hanks' solution [20], a widely used SBF [35,36], was used wherein hardened composite cement samples were immersed for 1, 3, 7, 14, 28 and 42 days, were measured using a pH meter (Suntex Instruments SP-2300, Taipei, Taiwan). After the powder and setting solution were mixed for 5 min, 2 g of cement was taken and immersed in 20 ml of daily-refreshed Hanks' solution with a pH value of 7.4. Each average pH value was obtained from 4 measurements taken under the same testing conditions. 2.3. X-ray diffraction and microstructural examination To study crystal structure/phases of the hardened cement immersed in Hanks' solution for the designated series of time, X-ray diffraction (XRD) was conducted using a Rigaku D-MAX B X-ray diffractometer (Tokyo, Japan) with Ni-filtered CuKα radiation operated at 30 kV and 20 mA at a scanning speed of 1°/min. The various phases of the samples being analyzed were identified by matching their characteristic peaks with data compiled in the JCPDS files. The compression-fractured surface morphology of the hardened cement was examined using a FEI Quanta 400 F environmental scanning electron microscope (ESEM) operated at 5 kV in secondary electron mode. Energy dispersive spectroscopy (EDS) point analysis was conducted on selected samples using an EDS system (INCA x-act, Oxford instrument, UK) operated at 10 kV with a spot size of 1 μm. To eliminate the charging effect, the surface for examination was coated with a thin layer of gold using an ion sputtering system (JFC-1100, JEOL, Japan) to facilitate electric conduction of the sample. 2.4. Porosity, weight loss and compressive strength measurement Porosity values of the hardened cement immersed in Hanks' solution for different periods of time were determined according to the ASTM C830-00 (2006) method. The weight loss values were measured using an electronic balance with an accuracy of 0.001 g. Samples for the

where W0 is the weight of the cement before immersion and Wt is the cement weight after immersion. Each average porosity or weight loss value was obtained from 6 measurements taken under the same testing conditions. The compressive strength (CS) of the hardened cement was determined according to the ASTM 451-99a method using a desk-top mechanical tester (Shimadzu AG-10KNX, Tokyo, Japan) at a crosshead speed of 1.0 mm/min. To conduct the test, the cement paste after being mixed for 1 min was packed into a 6 mm diameter, 12 mm deep (ASTM 451-99a) cylindrical stainless steel mold under a pressure of 1.4 MPa for 30 min. After being removed from the mold, the hardened cement was immersed in Hanks' solution at 37 °C for various periods of time. After immersion, the samples were removed from the solution for CS measurement while they were still wet. Each average CS value was obtained from 6 measurements taken under the same testing conditions. 2.5. Cytotoxicity test The cytotoxicity test was performed according to ISO 10993-5 methods, wherein an extraction method was used for this study. NIH/ 3 T3 fibroblasts, which are frequently used for the cytotoxicity test of orthopedic implant materials [39,40] with a seeding density of 5000 per well were pre-cultured for 24 h in Dulbecco's modified essential medium (DMEM) supplemented with bovine serum (10%) and PSF (1%). The extract was prepared by immersing hardened cement in the culture medium at a ratio of 0.2 g/ml at 37 °C for 24 h, followed by the collection of the liquid by centrifugation. Preparation of the extracts of negative control (Al2O3 powder) and positive control (0.3% phenol solution) followed the same extraction procedure. The extracts were added into a 96 well microplate (100 μl per well) incubated in a 5% CO2 humidified atmosphere at 37 °C. After 24 h, the extract was removed from the microplate and a mixture of the culture medium (100 μl) and WST-1 (10 μl) was added to the wells and incubated for 1 h at 37 °C. Cell viability was measured using the WST-1 assay, which is a colorimetric assay of mitochondrial dehydrogenase activity where the absorbance at 450 nm is proportional to the amount of dehydrogenase activity in the cell. After 1 h incubation, the mixture of medium and WST-1 was transferred to a 96 well microplate and the absorbance at 450 nm was measured using an ELISA reader. Four bars were tested for each sample (n = 4). 2.6. Animal implantation and histological examination Animal study was performed at National Cheng-Kung University Medical College Animal Center, Tainan, Taiwan. Adult (weighing 2.8– 3.5 kg), healthy, male New Zealand white rabbits were used as the experimental animals. The rabbits were housed individually in stainless steel cages with free access to food and water. An acclimation period of a minimum of 7 days was allowed between receipt of the animals and the start of the study. Injection sites were shaved and cleansed with 70% ethanol and Betadine™ (povidone iodine 10%). All animals were operated under general anesthesia. Zoletil 50 (0.05 ml/100 g, Virbac, France) was used as general anesthesia, while xylocaine (AstraZeneca, England) was used as local anesthesia. To implant cement paste in the medial condyle of femur, a longitudinal incision was made on the anterior surface of the femur. The inner side of the knee joint was cut to expose the femur. After exposure of the femur, the periosteum was reflected and a 2 mm pilot hole was drilled. The hole was sequentially widened with drills of increasing size until a final diameter of 5 mm was reached.

62

W.-L. Chen et al. / Materials Science and Engineering C 37 (2014) 60–67

A special 5 mm diameter drill burr was used and a ring was inserted at a depth of 10 mm to ensure appropriate length (10 mm) of the drill hole. (Fig. 1a). The cement powder used to prepare the cement implant for animal study was γ-ray-sterilized at China Biotech Co. (Taichung, Taiwan) with a dosage of 25 kGy. The setting solution was autoclave-sterilized for 30 min at 121 °C, according to the method suggested by Massey [21]. The cement paste for implantation was mixed in a bowl with a spatula and then transferred into a 5 ml syringe for injection into the prepared bone cavity. The injection was conducted carefully in a retrograde manner from the bottom to the surface of the bone defect. After injection of cement, subcutaneous tissues and skin were closed up layer by layer with nylon threads. To reduce the risk of perioperative infection, the rabbits were treated with antibiotics injected subcutaneously at a dose of 2 mg/kg. The animals were sacrificed at 4, 12 and 26 weeks postoperation. After the animals were sacrificed, the femur portions were excised immediately and excess tissues were removed. The retrieved bone was sectioned using a low-speed diamond blade (No. 11-4245, Buehler, USA). For consistency, samples for histological examination were always taken from the section located between 2.4 and 3.6 mm from the top of the implant, as indicated in Fig. 1b. The sectioned samples were fixed in 10% paraformaldehyde diluted in a 0.1 M phosphate buffer (pH 7.4) for 3 days, dehydrated in increasing grades of ethanol, defatted and cleaned with xylene, then embedded in EpoFix resin (Struers, Denmark). The embedded samples were ground using silicon carbide grit paper, followed by wet-cloth polishing with 1.0 and 0.3 μm Al2O3 powder. For histological examination, all these nondecalcified specimens were stained with toluidine blue and examined using a polarized light microscope (DM2500P, Leica Co., Germany). In order to enhance resolution and at the same time to have an overall picture of implant resorption behavior, more than a hundred highermagnification polarized light micrographs were taken sequentially on each section being examined. These 100 + micrographs were then superimposed to form a large composite picture covering the entire implant/bone cross section. The total areas of implant residues appearing on these composite pictures were determined using an image analysis system (Image-Pro Plus Version 6.0, Media Cybernetics Inc., MD, USA). To determine the implant residual ratio, an imaginary circle with a diameter of 5 mm (the original implant diameter) was superimposed on the composite micrograph being examined. The contour of the bone and residual implant material were traced based on the contrast in the image, and the areas of implant residuals (white areas) were measured using the aforementioned image analyzer. The sectional implant residual ratios were determined according to the equation,

3. Results and discussion 3.1. Working time, setting time and dispersion behavior of cement paste The cement paste prepared according to the method described in “Materials and Methods” exhibited an average working time and setting time of 10.1 min and 11.1 min, respectively, which are generally acceptable due to the general needs for the surgeries in the use of bone cement [37,38]. As mentioned earlier, in order to assess the dispersion behavior of the cement paste in water, the cement paste after being mixed for 1 min was loaded into a 5 ml syringe with needle removed and injected into 37 °C Hanks' solution. The observation indicated that the cement paste was able to harden in water without dispersion. One hour after injection, the hardened cement in the water still remained in its original shape and was not dispersed even after shaking the water container. The non-dispersive nature of the present cement, which may reduce the possibility of clinical complications such as cement embolism [22], is a good indication of its safety for clinical use, especially in minimally invasive procedures. 3.2. In-vitro changes in phases, morphology and pH value of hardened cement immersed in Hanks' solution Fig. 2 demonstrates typical XRD patterns of the various starting powders and hardened cement immersed in Hanks' solution for different periods of time. The XRD patterns of monolithic TTCP, DCPA, CSH, calcium sulfate dihydrate (CSD) and hydroxyapatite (HA) phases from JCPDS databases were also given as references. As indicated in the XRD patterns, the phase transitions involving the hydration of CSH and formation of CSD were most significant in the first two weeks. After immersion for 14 days, CSD peaks started to decrease in intensity indicating a gradual dissolution of CSD, while CSH peaks became almost invisible. The XRD patterns also indicated that both TTCP and DCPA peaks had largely diminished in the first few days' samples, indicating a quick dissolution process of TTCP and DCPA in the solution, in agreement with an earlier study of a biphasic TTCP/DCPA-derived cement [23]. After 14 days, the phase transition from TTCP/DCPA to HA was substantially completed, while HA and CSD became two major phases. It should be mentioned that all the calcium-based phases of the present cement, whether in the starting powders or in reaction/phase transformation products, are highly biocompatible and considered safe for use as an implant material [1,6,24,25]. Fig. 3 represents typical scanning electron micrographs of the compression-fractured surfaces of hardened cement samples immersed in Hanks' solution for different periods of time. The micrographs

Implant residual ratioð% Þ ¼ Implant residual area=Original implant area:

Fig. 1. Schematic drawings of the implantation site (a) and implant/bone section for histological examination (b) for the study.

Fig. 2. XRD patterns of starting powders and hardened cement immersed in Hanks' solution for different periods of time.

W.-L. Chen et al. / Materials Science and Engineering C 37 (2014) 60–67

63

Fig. 3. Scanning electron fractographs of hardened cement immersed in Hanks' solution for different periods of time.

indicated that both macropores and micropores were present throughout the fracture surfaces of all cement samples. Fine, needle-shaped crystals were readily observed in the 1-day sample. After immersion for 7 days, more and thicker faceted crystals were present on the fracture surface. These crystals are believed to be CSD crystals due to their morphological similarities to those observed in other studies [26–28],

and also due to their atomic-ratio similarities to that of CSD (Ca:S: O = 1:1:6). The EDS point analysis results indicate that the average (n = 10) Ca:S:O ratios of the faceted crystals in 1-day sample and 7day sample are 1.0:1.0:6.0 and 1.0:1.2:6.1, respectively, with negligible phosphorus signals. After 14 days, such faceted crystals were no longer observed, although the XRD pattern of the 14-day sample still showed a

Fig. 4. Changes in pH value of Hanks' solution wherein hardened cement is immersed for different periods of time.

Fig. 5. Weight loss and porosity values of hardened cement immersed in Hanks' solution for different periods of time.

64

W.-L. Chen et al. / Materials Science and Engineering C 37 (2014) 60–67

[23], although XRD patterns did not show significant intensities of these two phases at this immersion stage. After 7 days, the average pH value became 6.6 and remained stable until 28 days, when it slightly dropped again. After 42 days, it went down to 6.3. This slight decrease in pH value (from 28 days to 42 days) of the solution was probably a result of the combined effect of the formation of HA [23] and the gradual dissolution of CSD [6], which is consistent with the present XRD results. 3.3. In-vitro changes in weight loss, porosity and compressive strength of hardened cement immersed in Hanks' solution

Fig. 6. Changes in compressive strength of hardened cement immersed in Hanks' solution for different periods of time.

As indicated in Fig. 5, both weight loss and porosity values of the hardened cement continued to increase with immersion time, as expected. After being immersed for 1day, the hardened cement had readily lost 18.6% of its weight. The weight loss and porosity values of the hardened cement increased respectively to 40.2% and 55.0% after 28 days, and respectively to 45.2% and 59.7% after 42 days. Although the dissolution of phosphates could contribute to the increases in porosity and weight loss to a certain degree, the dissolution of CSH and CSD should be the most dominant cause to the observed immersioninduced increases in porosity and weight loss due to their high dissolution rates in the solution [6]. This interpretation is also supported by the present XRD results. Fig. 6 demonstrates changes in CS value of the hardened cement immersed in Hanks' solution for different periods of time. As shown in the figure, the CS value of the cement continued to decrease with immersion time. For example, after immersion for 14 days, the average CS value had decreased to 13.6 MPa from its 1-day CS value of 18.0 MPa. After 42 days, the CS value dropped to 2.7 MPa. Although many factors have been suggested to affect the strength of calciumbased cement, such as apatite content [29–31], porosity level [5,32] and cement morphology [33], the large immersion-induced increase in porosity was probably the most important factor that caused the significant long-term degradation in CS of the present cement. This in-vitro long-term decay in strength is actually one inherent and necessary feature of a bioresorbable material due to its gradual dissolution and replacement by the surrounding bone. A one-way ANOVA method was used to evaluate the statistical significance of compressive strength, pH value, weight loss and porosity. In all cases, the results were considered statistically different with p b 0.05. 3.4. Cytotoxicity test and animal study

Fig. 7. Viability values of cells incubated for 24 h in conditioned mediums adulterated with cement extraction, blank medium, Al2O3 powder (negative control) and 0.3% phenol (positive control).

strong intensity of CSD, suggesting that CSD at this stage might have a different or finer morphology which became indistinguishable from other phases under SEM. After 42 days, a very porous morphology was developed, in agreement with the following porosity data. This porous structure could be one primary reason for the excellent implant resorption/bone ingrowth behavior observed in the histological examination to be discussed later. Fig. 4 shows the long-term changes in pH value of the dailyrefreshed Hanks' solution wherein the cement was immersed for different periods of time. As shown in the figure, all average pH values were in the weakly acidic range from 6.2 to 6.6 during the entire immersion process. The average pH value of the Hanks' solution decreased from its 1-day average value of 6.5 to its 3-day average value of 6.2 due to dissolution of CSH and formation of CSD [6], in agreement with the present XRD results. After 3 days, the average pH value of the solution started to increase probably due to the continual dissolution of TTCP and DCPA

Fig. 7 demonstrates the viability values of cells incubated for 24 h in conditioned mediums adulterated with 24-hour cement extraction, blank medium, Al2O3 powder (negative control) and 0.3% phenol (positive control). As indicated in the figure, the cells incubated with conditioned medium of Al2O3 powder and cement extraction exhibited similar viability values (0.68 and 0.65, respectively). Although the viability value of cells incubated with conditioned medium of cement extraction (0.65) was lower than blank medium (0.80), it was within an acceptable range, i.e., 96% that of Al2O3 control and 81% that of blank medium, and considered non-toxic. (Note: According to ISO 10993 method, the test material has a cytotoxic potential if the cell viability value is less than 70% that of the blank medium). Surgical application of the present cement was relatively easy and met the surgeon's needs. The setting of the cement was never disturbed due to capillary bleeding from the marrow cavity. After setting, the overlying soft tissue could be closed without interrupting the cement surface. Overfilled cement was removed with a scalpel blade prior to setting. All drilled holes were found to be completely filled with cement. Neither infection nor setting-induced heat effect on the surrounding

Fig. 8. Histological examination of toluidine blue-stained sections (a–c) and their corresponding contrast images for residual implant analysis (d–f). IC: implanted cement; OB: old bone; NB: new bone.

W.-L. Chen et al. / Materials Science and Engineering C 37 (2014) 60–67

65

66

W.-L. Chen et al. / Materials Science and Engineering C 37 (2014) 60–67

tissue, which is a common concern for the implantation of PMMA [34], was observed in the present study. Furthermore, no bone formation within muscles or outside the original drill hole were observed after any of the implantation periods. The residual cement appeared to be nicely incorporated with surrounding bone and did not elicit an obvious inflammatory reaction, necrosis or fibrous encapsulation in the adjacent tissues. As mentioned earlier, for comparison in a consistent way, the samples for histological examination were always taken from similarly located sections of the retrieved femoral condyle (between 2.4 and 3.6 mm from implant top surface). It was also mentioned that, on each section, 100+ higher-magnification micrographs were taken and superimposed forming a large composite picture which covered the entire implant/bone cross section being examined. One major advantage in assessing implant resorption and new bone growth behavior from this composite micrograph over observation merely from a single, low magnification photograph is its much higher capability (resolution) of differentiating tissues of various kinds in any local area. With size reduced, the composite picture conveniently demonstrates an overall implant/bone morphology. A series of typical histological morphologies of toluidine blue-stained sections at different implantation times are shown in Fig. 8. This figure also provides an example demonstrating such reduced-sized composite micrographs, along with locally magnified micrographs in a few selected areas. In general, excellent bonding between host bone and implanted cement without interposition of fibrous tissues was observed in all regions of inspection in the present study. Gross examination of the retrieved femoral condyle revealed that, at 4 weeks post-implantation, the cement implant appeared as a dense mass with noticeable cracks throughout cement, as shown in its composite picture (Fig. 8a). These cracks could accelerate the resorption process at a later stage by dividing up the implant mass into separate islands and allowing blood/fluid to penetrate through these cracks into the interior of the implant. Highermagnification micrographs at selected bone-implant interfacial areas revealed that typically the trabecular bone at bone-implant interface was in intimate contact with the implant surface. Higher remodeling activities were found in the 12-week samples. The composite picture (Fig. 8b) indicates that the implanted cement was readily separated into isolated islands due to a combined effect of implant dissolution and new bone formation. A new bone network was developed within the femoral defect, while the implanted cement was completely incorporated into the newly formed bone structure. At this stage, crevices filled with lamellar new bone structure were clearly observed. Direct and intimate contact between new bone and implanted cement and extensive remodeling of the bone at implantbone interface were observed. More extensive remodeling activities were observed in the 26-week samples. At this stage, the newly formed bone became more mature and the interface between residual cement and new bone became less identifiable. Isolated/detached cement residues were observed throughout the new lamellar bone structure (Fig. 8c). The residual ratios of the implants were determined using the method described in Materials and Methods. Histomorphometric measurements were performed on the ground sections, wherein bone matrix appeared black and the remaining cement appeared white, as seen in the processed contrast images (Fig. 8d–f). Analysis based on these images indicated that the cement residual ratios of the 4, 12 and 26week samples were 86.4, 42.5 and 13.8%, respectively. Earlier rabbit implantation studies of the present laboratory on single-phase TTCPderived and biphasic TTCP/DCPA-derived calcium phosphate cements indicated that the two implants had 24-week resorption rates of 58 and 45%, respectively, at similar implant location and implantation time [7,16]. Compared to these calcium phosphate cements, the resorption rate of the present calcium phosphate/calcium sulfate composite cement is apparently much higher. This desired higher resorption rate is believed to be attributed to the presence of calcium sulfate in

the cement which generally has a higher dissolution rate than calcium phosphate [6,11–14], as mentioned in the Introduction. The faster dissolution of calcium sulfate, in turn, created the observed pores and cracks throughout the cement implant which is believed to have helped conduct new bone structure to grow within, thereby accelerating the resorption process. 4. Conclusions 1. The present TTCP/DCPA/CSH-derived cement paste was able to completely harden in water without dispersion, and demonstrated an average working time and setting time of 10.1 min and 11.1 min, respectively. 2. The XRD results indicated that the hydration of CSH and formation of CSD in the hardened cement immersed in Hanks' solution were most significant in the first two weeks. After immersion for 14 days, CSD peaks started to decrease, while CSH peaks became almost invisible. Both TTCP and DCPA peaks had largely diminished in the first few days' samples. After 14 days, HA and CSD became two primary phases. 3. The SEM micrographs indicated that macropores and micropores were present throughout the compression-fractured surfaces of all cement samples. Fine, needle-shaped crystals were observed in the 1-day sample. After immersion for 7 days, more and thicker faceted crystals were present, which were no longer observed after 14 days. After 42 days, a very porous morphology was developed. 4. The average pH values of the Hanks' solution wherein the cement was immersed were in the weakly acidic range from 6.2 to 6.6 during the entire immersion process. 5. Both weight loss and porosity values of the hardened cement increased with immersion time. The average weight loss and porosity values of the hardened cement increased respectively to 40.2% and 55.0% after 28 days, and respectively to 45.2% and 59.7% after 42 days. 6. The CS value of the cement decreased with immersion time. After immersion for 14 days, the CS value decreased to 13.6 MPa from its 1-day CS value of 18.0 MPa. After 42 days, the CS value went down to 2.7 MPa. 7. The viability value of cells incubated with conditioned medium of cement extraction was 96% that of Al2O3 control and 81% that of blank medium, which is considered non-toxic. 8. Surgical application of the present cement was relatively easy. The setting of the cement was never disturbed due to capillary bleeding from the marrow cavity. The residual cement appeared to be nicely incorporated with surrounding bone and did not elicit an obvious inflammatory reaction. Excellent bonding between host bone and implanted cement without interposition of fibrous tissues was observed. At 12 weeks post-implantation, significant remodeling activities were found and the implanted cement was readily separated into isolated islands. A new bone network was developed within the femoral defect. At 26 weeks post-implantation, more extensive remodeling activities were observed and the newly formed bone became more mature and the interface between residual cement and new bone became less identifiable. 9. Analysis based on the processed contrast images indicated that the implant residual ratios of the 4, 12 and 26-week samples were 86.4, 42.5 and 13.8%, respectively, which are significantly lower than those revealed in the single-phase TTCP-derived and biphasic TTCP/DCPA-derived calcium phosphate cements under similar implantation conditions. Acknowledgement The authors would like to acknowledge the support for this research by the Southern Taiwan Science Park (Kaohsiung Science Park), Taiwan, ROC under the Research Grant # BZ-07-18-43-98.

W.-L. Chen et al. / Materials Science and Engineering C 37 (2014) 60–67

References [1] S.V. Dorozhkin, Bioceramics of calcium orthophosphates, Biomaterials 31 (2010) 1465–1485. [2] V.A. Dubok, Bioceramics—Yesterday, Today, Tomorrow, Powder Metall. Met. Ceram. 39 (2000) 381–394. [3] H. Schliephake, R. Gruber, M. Dard, R. Wenz, S. Scholz, Repair of calvarial defects in rats by prefabricated hydroxyapatite cement implants, J. Biomed. Mater. Res. A 69 (2004) 382–390. [4] P.F. Heini, U. Berlemann, Bone substitutes in vertebroplasty, Eur. Spine J. 10 (Suppl. 2) (2001) S 205–S 213. [5] M. Bohner, Calcium orthophosphates in medicine: from ceramics to calcium phosphate cements, Injury 31 (Suppl. 4) (2000) 37–47. [6] M.V. Thomas, D.A. Puleo, Calcium sulfate: properties and clinical applications, J. Biomed. Mater. Res. B Appl. Biomater. 88 (2009) 597–610. [7] W.C. Chen, C.P. Ju, Y.C. Tien, J.H. Chern Lin, In vivo testing of nanoparticle-treated TTCP/DCPA-based ceramic surfaces, Acta Biomater. 5 (2009) 1767–1774. [8] H. Oda, K. Nakamura, T. Matsushita, S. Yamamoto, H. Ishibashi, T. Yamazaki, S. Morimoto, Clinical use of a newly developed calcium phosphate cement (XSB-671D), J. Orthop. Sci. 11 (2006) 167–174. [9] G.D. Brown, B.L. Mealey, P.V. Nummikoski, S.L. Bifano, T.C. Waldrop, Hydroxyapatite cement implant for regeneration of periodontal osseous defects in humans, J. Periodontol. 69 (1998) 146–157. [10] L. Comuzzi, E. Ooms, J.A. Jansen, Injectable calcium phosphate cement as a filler for bone defects around oral implants: an experimental study in goats, Clin. Oral Implants Res. 13 (2002) 304–311. [11] M.T. Fulmer, I.C. Ison, C.R. Hankermayer, B.R. Constantz, J. Ross, Measurements of the solubilities and dissolution rates of several hydroxyapatites, Biomaterials 23 (2002) 751–755. [12] F. Braye, J.L. Irigaray, E. Jallot, H. Oudadesse, G. Weber, N. Deschamps, C. Deschamps, P. Frayssinet, P. Tourenne, H. Tixier, S. Terver, J. Lefaivre, A. Amirabadi, Resorption kinetics of osseous substitute: natural coral and synthetic hydroxyapatite, Biomaterials 17 (1996) 1345–1350. [13] P. Proussaefs, J. Lozada, G. Valencia, M.D. Rohrer, Histologic evaluation of a hydroxyapatite onlay bone graft retrieved after 9 years: a clinical report, J. Prosthet. Dent. 87 (2002) 481–484. [14] C.P. Klein, A.A. Driessen, K. de Groot, A. van den Hooff, Biodegradation behavior of various calcium phosphate materials in bone tissue, J. Biomed. Mater. Res. 17 (1983) 769–784. [15] B.R. Constantz, B.M. Barr, I.C. Ison, M.T. Fulmer, J. Baker, L. McKinney, S.B. Goodman, S. Gunasekaren, D.C. Delaney, J. Ross, R.D. Poser, Histological, chemical, and crystallographic analysis of four calcium phosphate cements in different rabbit osseous sites, J. Biomed. Mater. Res. 43 (1998) 451–461. [16] C.H. Tsai, R.M. Lin, C.P. Ju, J.H. Chern Lin, Bioresorption behavior of tetracalcium phosphate-derived calcium phosphate cement implanted in femur of rabbits, Biomaterials 29 (2008) 984–993. [17] S. Yamada, D. Heymann, J.M. Bouler, G. Daculsi, Osteoclastic resorption of calcium phosphate ceramics with different hydroxyapatite/beta-tricalcium phosphate ratios, Biomaterials 18 (1997) 1037–1041. [18] J.H. Chern Lin, S.H. Hung, W.L. Chen, C.K. Chen, J.W. Lee, R.M. Lin, C.P. Ju, Properties of TTCP/DCPA/CSH cement immersed in Hanks' solution, J. Med. Biol. Eng. 32 (2012) 201–204. [19] W.E. Brown, E.F. Epstein, Crystallography of tetracalcium phosphate, J. Res. Natl. Bur. Stand. Sect. A: Phys. Chem. 69 (1965) 547–551.

67

[20] D.C. Mears, Metals in medicine and surgery, Int. Met. Rev. 22 (1977) 119–155. [21] L.K. Massey, The Effect of Sterilization Methods on Plastics and Elastomers, 2nd ed. William Andrew, New York, 2006. [22] G. Geraci, G. Lo Iacono, C. Lo Nigro, F. Cannizzaro, M. Cajozzo, G. Modica, Asymptomatic bone cement pulmonary embolism after vertebroplasty: case report and literature review, Case Rep. Surg. (2013), http://dx.doi.org/10.1155/2013/ 591432. [23] W.C. Chen, C.P. Ju, J.H. Chern Lin, Variation in structure and properties of a non-dispersive TTCP/DCPA-derived CPC immersed in Hanks' solution, J. Oral Rehabil. 34 (2007) 541–551. [24] L.L. Hench, Bioactive materials: the potential for tissue regeneration, J. Biomed. Mater. Res. 41 (1998) 511–518. [25] A.J. Ambard, L. Mueninghoff, Calcium phosphate cement: review of mechanical and biological properties, J. Prosthodont. 15 (2006) 321–328. [26] M. Nilsson, L. Wielanek, J.S. Wang, K.E. Tanner, L. Lidgren, Factors influencing the compressive strength of an injectable calcium sulfate-hydroxyapatite cement, J. Mater. Sci. Mater. Med. 14 (2003) 399–404. [27] H.Y. Song, A.H. Esfakur Rahman, B.T. Lee, Fabrication of calcium phosphate-calcium sulfate injectable bone substitute using chitosan and citric acid, J. Mater. Sci. Mater. Med. 20 (2009) 935–941. [28] N.B. Singh, B. Middendorf, Calcium sulphate hemihydrate hydration leading to gypsum crystallization, Prog. Cryst. Growth Charact. Mater. 53 (2007) 57–77. [29] K. Ishikawa, S. Takagi, L.C. Chow, Y. Ishikawa, Properties and Mechanisms of Fast-setting Calcium Phosphate Cements, J. Mater. Sci. Mater. Med. 6 (1995) 528–533. [30] C.K. Chen, C.P. Ju, J.W. Lee, J.H. Chern Lin, Effect of gamma radiation on properties of a calcium phosphate-calcium sulfate composite cement, Mater. Trans. 54 (2013) 1160–1165. [31] Y. Miyamoto, K. Ishikawa, H. Fukao, M. Sawada, M. Nagayama, M. Kon, K. Asaoka, In vivo setting behaviour of fast-setting calcium phosphate cement, Biomaterials 16 (1995) 855–860. [32] C.H. Tsai, C.P. Ju, J.H. Chern Lin, Morphology and mechanical behavior of TTCP-derived calcium phosphate cement subcutaneously implanted in rats, J. Mater. Sci. Mater. Med. 19 (2008) 2407–2415. [33] L.E. Carey, H.H. Xu, C.G. Simon, S. Takagi, L.C. Chow, Premixed rapid-setting calcium phosphate composites for bone repair, Biomaterials 26 (2005) 5002–5014. [34] S. Harving, K. Soballe, C. Bunger, A method for bone cement interface thermometry, Acta Orthop. Scand. 62 (1991) 546–548. [35] W.L. Chen, N.Y. Fan, J.W. Lee, R.M. Lin, C.P. Ju, J.H. Chern Lin, Immersion-induced changes in structure and properties of a calcium sulfate hemihydrate cement, Mater. Trans. 53 (2012) 1154–1158. [36] C.K. Chen, C.P. Ju, J.H. Chern Lin, Setting solution concentration effect on properties of a TTCP/DCPA-derived calcium phosphate cement, J. Mater. Sci. Mater. Med. 23 (2012) 2109–2114. [37] K.S. Oh, Y.K. Jeong, J.P. Yu, S.K. Chae, H.Y. Kim, H.Y. Lee, S.S. Jeun, Preparation and In Vivo Studies of β-TCP Based Bone Cement Containing Polyphosphate, Key Eng. Mater. 284–286 (2005) 93–96. [38] P.L. Lai, C.L. Tai, I.M. Chu, T.S. Fu, L.H. Chen, W.J. Chen, Hypothermic manipulation of bone cement can extend the handling time during vertebroplasty, BMC Musculoskelet. Disord. 13 (2012) 198. [39] E.R. Takamori, E.A. Figueira, Evaluation of the Cytocompatibility of Mixed Bovine Bone, Braz. Dent. J. 18 (3) (2007) 179–184. [40] B. Balasaheb, P. Chandanshive, Rai, Synthesis of hydroxyapatite nanotubes for biomedical applications, Mater. Sci. Eng. C 33 (2013) 2981–2986.