Plasma-sprayed coatings of tetracalciumphosphate, hydroxyl-apatite, and a-TCP on titanium alloy: An interface study Christel P. A.T. Klein,*rtt' P. Patka: H. B. M. van der Lubber*J.G.C. Wolke,' and K. de Groot*ft *Department of Biomaterials, ACTA, Free University, Amsterdam, 'Biomaterials Research Group, Dept. of Biomaterials, Leiden University, School of Medicine, Leiden, and §Department of Surgery, Academic Hospital, ,Free University, Amsterdam In order to study the interaction of calcium phosphate coatings with bone tissue, coated titanium cylinders with a standard size were implanted in dog femora. Coatings were made by plasma spraying powders of hydroxylapatite, P-whitlockite, and tetracalciumphosphate particles. The plasma spraying process tuns P-whitlockite into a-TCP. Bone bonding and bone formation were evaluated by mechanical push-out tests and histological observations. I-lydroxylapa-
tite and tetracalciumphosphate coatings show an interface strength after 3 months of implantation of 34.3 f 6.5 MPa and 26.8 2 3.9 MPa, respectively, while a-TCP and blanco titanium lead to an interface strength of 10.0 2 3.5 MPa and 9.7 1.3 MPa, respectively. Histological examinations revealed that hydroxylapatite and tetracalciumphosphate give rise to an excellent bone formation, while a-TCP and blanco titanium evoked remodeling and less bone contact. +_
INTRODUCTION
As with all implanted devices the requirements of the materials used for their construction are varied, but can be broadly classified under the headings of biocompatibility, biofunctionality, and availability. Biocompatibility is concerned with the interactions between materials and the tissues of the body and is the most important factor involved with this material selection. In particular, because of the good biocompatibility, calcium phosphate ceramics have received wide attention. These compounds have a close chemical resemblance to bone mineral and they give rise to an intimate contact with the surrounding host bone.' 13iofunctionalityis concerned with those mechanical and physical properties that enable the implanted device to perform its function. Calcium phosphate ceramics can be made with properties simulating hard tissues to a large extent, dense sintered ceramics may have compressive strengths up to 5000 kg,'cm2. However, due to their brittle nature they are useful only in applications, where the mechanical forces are nonexistent or 'To whom correspondence should be addressed. Journal of Biomedical Materials Research, Vol. 25, 53-65 (1991) CCC 0021-9304/91/Ol0053-13$04.00 0 1991 John Wiley & Sons, I[nc.
KLEIN ET AL.
54
primarily compressive. The progress of bioengineering has resulted in a new technique, plasma-spraying, which can now overcome mechanical drawbacks of bioactive ceramics. Composites with a biocompatible surface of calcium phosphate ceramics and an inner structure of metal for obtaining sufficient mechanical strength are required, if load-bearing parts of the skeleton are to be replaced. Implant coating seems to be an attractive alternative in obtaining a composite material that combines the inner strength of a metal with the biological surface characteristics of calcium phosphates. Metals in use for coating substrate, particularly titanium, have a long clinical history of implantation and need no special attention.’ Recently, calcium phosphate particles are applied in the plasma-spraying technique as coating on metallic prostheses. These coatings are usually very thin, 50 pm or less. Therefore, even slight solubilities might lead to dissolution of those coatings. Previous biodegradation behavior studies of hydroxylapatite and j3whitlockite revealed that P-whitlockite evoked bone remodeling and porous bone formation, probably due to its bioresorption. From in vitro studies it is known that the solubilities of various calcium-phosphates depend on the buffer solvent, as well as on pH-value and the specific calcium phosphate salt.3 It is very interesting to relate the in vivo biodegradation behavior of different coatings with respect to their bonding properties and their interaction to bone formation. In this paper we present a mechanical push-out study and histological examinations of the interface.
MATERIALS A N D METHODS
Cylindrical rods of standard Ti-6A1-4V titanium alloy were prepared. A coating of hydroxylapatite (HA), tetracalciumphosphate (TETRA) or a-TCP, 50 pm thickness, was applied using the plasma-spray technique as described ear lie^.^ Noncoated titanium (Ti) plug implants of the same dimensions as the coated plugs were used as controls (4.5 mm diameter, 6 mm length). The surface roughness value (Ra) of the different coated titanium plugs were respectively:Ha: Ra 11.26 2 1.90 pm, a-TCP: Ra 9.52 2 0.51 pm, TETRA: Ra 11.18 k 0.97 pm and blanco titanium: Ra 3.88 ? 0.26 pm. The roughness was measured with a Tayler-Hobson, Surtronic 10. Using sterile surgical techniques, holes of 4.7 mm diameter were drilled through the periosteum and cortex, into the medulla, employed a round stainless-steel burr; the rotation speed of the burr was 200 rounds/min). The distance between the holes was 2 cm. During the drilling process the preparation site was continuously cooled with physiological saline. Bone debris was carefully removed by rinsing the holes. The plugs were inserted press-fit into these holes (depth 12-14 mm) in the lateral cortex of the femora of adult beagles. To study exclusively the bone-bonding properties of the materials, the holes were slightly oversized. A total of 18 plugs were inserted into the femora of 3 dogs (3 plugs per femur). The distribution of the plugs was performed rand~mly.~
PLASMA-SPRAYED COATINGS ON TITANIUM ALLOY
55
There were no surgical complications. After a follow-up period of 3 months the dogs were sacrificed using a lethal dose of thiopental. After explant radiography, the femora were used for mechanical testing and histological examination. Each femur was sectioned into three pieces, each one with one plug. After removal of excess soft tissue samples of each testing material were used for mechanical testing and following histology. Samples of each testing material were used directly within 1 h, without any fixation procedure, for mechanical testing. Afterward these samples were fixed for histological procedures6 For the mechanical testing the bone pieces were positioned in a testing jig to allow accurate lining of the loading axis of the test machine with the long axis of the plug (Fig. I). The plugs were pushed out from the surrounding bone, using a Zwick test machine with a crosshead speed of 0.5 mm/min. The maximum force to loosen the implant was calculated from the loaddisplacement curves. To calculate a "shear strength" of the interface, the push-out force was divided by the total plug/cortical bone contact area; contact area = 3.14 x diameter plug x average cortex thickness the cortex was measured on both sides of the plug, and the average of these two measurements was taken as "average cortex thickness." For the histological observations the samples were fixed in formaldehyde buffer, followed by alcohol dehydration and embedding in polymethylmethacrylate. Undecalcified saw sections of 10 pm were prepared using a diamond cutting wheel with cooling.6 Tissue staining was performed with alcain-blue or Masson trichrome. Histology was reviewed with the following objectives: character of coating-bone interface, condition of the coating, response with cortical and trabecular bone, quality and quantity of new bone formation, and response of bone marrow. RESULTS
The x-ray diffraction pattern of the HA coating was comparable with the file #24-33 of the joint (Committee on Powder Diffraction Standards7 By
Diagram of Test Apparatus Figure 1. A schematic drawing of the test apparatus for push-out tests.
KLEIN ET AL.
56
plasma-spraying of P-whitlockite the crystallographic structure is changed from a pure hexagonal into a monoclinic a-tricalciumphosphate (according to JCPDS #29, 359). The diffractogram of TETRA was consistent with JCPDS #25-1137, referring to CaO(P04)2tetracalcium monoxide diphosphate (see Fig. 2):
Radiography All coated implants displayed proliferation of bone along the protruding periosteal and endosteal part of the implant (Fig. 3). Around the implant a proliferation and densification of trabecular bone can be seen at 3 months of implantation independent of the coating type.
Mechanical testing The push-out data are summarized in Table I. HA showed the highest push-out strength, followed closely by TETRA. a-TCP developed a bonding, similar to the uncoated titanium plug. The failure occurred at the coating bone interface, within the coating itself and at the interface coating-substrate. Failure occurred more or less evenly divided over these three levels.
I
440
~
3+5
h
30
~
23
~
208
Figure 2. The x-ray diffraction pattern of hydroxylapatite (HA), tetracalciumphosphate (TetraCP) and a-TCP estimated after the plasma-spraying procedure.
PLASMA-SPRAYED COATINGS ON TITANIUM ALLOY
57
Tetra
TC P
i
-A__-
40
35 3'0 Figure 2. (continued)
1
2.5
2013
Histology Histological sections of the HA-coated titanium plugs show clearly that defects in the periosteum and bone around the implant are filled with newly formed bone. Periosteal and endosteal bone proliferates along the protruding ends of the plugs in and out the femur. The cortical bone area is in very close contact with the implant (circa 90%), without any interposition of fibrous tissue. The cortex shows remodeling with osteoblasts and osteoclasts along the lacunae. Not only mature osteocytes, but also numerous young osteoblasts and complete haversian systems can be seen in direct contact with the HA
KLEIN ET AL.
58
Figure 3. Radiograph of a femur implanted with three plugs coated with hydroxylapatite, tetracalciumphosphate, or a-TCP. After 3 months trabecular bone growth can be observed around the implants.
coating. The coated plug was predominantly enclosed with both cortical and trabecular bone. The last one had been developed from the endosteum in the marrow cavity and in close contact with the implant (Figs. 4 and 5). After 3 months of implantation the hydroxylapatite coating was still distinctable, the thickness and density showed some variety. There was some fragmentation; phagocytic activity could be observed at some spots, but resorption could not be clearly seen. From the sections made from samples which were used first in the mechanical tests, it can be seen that the coating was incorporated in the newly formed bone and loose of the plug. The histological sections of TETRA coating revealed a bone reaction similar to that with HA coatings. Also a very intimate bone contact (over 290% of the implant surface) could be observed. The bone formation and architecture along the TETRA coating is the same as along HA (Figs. 6 and 7). After TABLE I Average Plug-Bone Interface Strength at 3 Months (n = 5) Load ( N ) HA TETRA
CY-TCP Ti
132 f 111 2 37 k 44 f
9.4 15.6 9.3 10.5
Shear Strength (MPa)
Cortex (mm)
34.3 f 6.5 26.8 & 3.9 10.0 k 3.5 9.7 & 1.3
2.6 f 0.2 2.7 ? 0.4 2.4 & 0.3 2.7 f 0.2
PLASMA-SPRAYED COATINGS ON TITANIUM ALLOY
Figure 4. Hydroxylapatite coating showed a very intimate mature bone contact. The coating is incorporated in the new formed bone. Implantation period of 3 months (staining of alcain-blue, original magnification X250).
Figure 5. Hydroxylapatite coating showed a decrease of density. Bone remodeling (osteoid, osteoclasts and osteoblasts) occurred, bone cells were present against the metal surface. Implantation period of 3 months (alcainblue staining, original magnification X250).
59
60
KLEIN ET AL.
Figure 6. Tetracalciumphosphate coating showed a decrease of density. An intimate mature bone contact can be seen. Osteoblasts activity can be observed against the coating surface (alcain-blue staining, original magnification X250).
Figure 7. Tetracalciumphosphate coating after 3 months of implantation showed a decrease of density and thickness. Osteoclastic cells can be seen (alcain-blue staining, original magnification X250).
PLASMA-SPRAYED COATINGS ON TITANIUM ALLOY
61
3 months the coating is still present. Osteoclasts or multinuclear cells can be seen near the coating material. Alpha-TCP (tricalciumphosphate) coatings give rise to bone formation from periosteum, cortex, and endosteum. However, the bone consisted of many remodeling lacunae, which also can be seen along the implant surface. So, this coating showed significantly less bone contact ( C50%) than hydroxylapatite and tetracalciumphosphate coatings. In the lacunae osteoid, osteoblasts, osteoclasts, and plasma cells can be seen. Especially near the coating osteoclasts or multinuclear cells were present (Fig. 8). The coating is still detectable in the sections. It was mostly anchored in the newly formed bone and, but in contrast to the other ones, resorption of the coating can be clearly seen. The blanco titanium plugs evoked a similar bone response as a-TCP coating. Hence, little bony contact ( ~ 5 0 %with ) the implant surface and many remodeling lacunae in the cortex and along the implant. Connective tissue against the implant surface can also be seen (Figs. 9 and 10). DISCUSSION
Dependent of the type of bone fracture, place and many other factors the time to repair in long bone fractures varies between 6 and 12 weeks in case of the cortex.’ Because one can expect a total bone healing after 3 months, the first observations of the in vivo study of interface bone-calcium phosphate
Figure 8. a-TCP coating after 3 months of implantation gives rise to bone remodeling with many osteoclasts. Also plasma cells can be observed in the bone lacunae. Coating density and thickness was decreased (alcain-blue staining, original magnification X250).
62
KLEIN ET AL.
Figure 9. Uncoated titanium plugs showed a little bone contact and mainly bone remodeling lacunae around the implant surface (Masson trichrome staining, original magnification X160).
coatings on titanium plugs were made after 3 months of implantation. Also, earlier studies with calciumphosphate ceramics were described after this implantation period. Long-term implantation studies of 6 months, 1 and 2 years are started and will be published later. To get some insight into the in vivo behavior of different calcium phosphates after an implantation period, a dynamic-functionally loaded model was chosen. In addition, with this model results can be compared with other studies." The x-ray diffraction patterns of the different calcium phosphate coatings after plasma-spraying were compared with the files of the Joint Committee on Powder Diffractions Standards.' It can be noticed that tricalciumphosphate powders with a P-whitlockite crystallographic structure change after the plasma-spraying procedure in an a-tricalciumphosphate crystallographic structure. Sintering of tricalciumphosphate powders does not lead to a change of the P-whitlockite structure.* Due to the fact that the implants were not inserted in one geometrical plane, the x-ray pictures (Fig. 3) show different apparent dimensions of the cylindrical implants. The density around the implants in the marrow cavity indicates bone growth against the implants, as confirmed with the histological observations. It is known from previous studies that the time delay and/or preparation with a fixation medium before performing the mechanical testing can lead to other push-out data." Values for bone-plug interface shear strength measured directly after sacrificing the animal were approximately one half lower as compared to those obtained after a period of formaline fixation. To get realistic information about the bonding strength, the mechanical
PLASMA-SPRAYED COATINGS ON TITANIUM ALLOY
63
Figure 10. Uncoated titanium plug give rise to connective tissue formation against the implant surface. (Masson trichrome staining, original magnification x160).
tests were made with untreated ("fresh") samples. Calculations for interface stress are very questionable and very dependent on experimental set-up and assumptions for bony contact area. To allow comparisons with the literature data, both the absolute values for push-out load as well as the relative interface shear strength values are provided in Table I. The bone-plug interface shear strength of hydroxylapatite coatings were in accordance with the results on fresh samples (29 MPa) described by Geesink et al.'' The shear strength of tetracalciumphosphate (26.8 5 3.9 MPa) is, within experimental error, comparable with hydroxylapatite, while a-TCP and blanco titanium plugs gave much lower values (10.0 MPa and 9.7 MPa, respectively) (Table I). One should realize that, due to differences in elastic modulus within the cortical bone area contacting the implant surface leading to stress concentration at the most dense parts, the found values have little or no relation to the "true" shear strength of the interface." BlackI3 suggested, that only under standardized conditions, a push-out test can be useful for comparison only. Dhert et al.I5found that the choice of bone type and animal species can also influence mechanical test results. The surface roughness values (Ra) of the different coatings were approximately the same for all coatings (circa 10 pm), which make it possible to compare the bone bonding. The surface roughness of blanco titanium is lower, but the push-out strength is similar to a-TCP. One can conclude that both a high degradable calcium phosphate coating and metal plugs with a low roughness leads to a low shear strength.
KLEIN ET AL.
64
Comparison of the push-out data with the histological findings shows that a high push-out strength is related to good quality and quantity of the newly formed bone against the implant. An intimate and extensive bone contact leads obviously to a strong bonding at the interface of bone and coating. In the case of a-TCP or blanco titanium the presence of many remodeling lacunae and/or connective tissue is probably responsible for a weaker bonding at the interface. Comparing the in vivo data with the in vitro dissolubility results; shows that the in vitro dissolubility degree is not an important factor related to the in vivo degradation. TETRA ceramic particles had (in all buffers) the highest solubility, whereas in the in vivo studies TETRA showed a very good bone bonding and an apparent shear strength similar to hydroxylapatite. In terms of chemical reaction the dissolution may be described according to Driessen et al.I4as:
-
Ca3(P04)2-+ 3Ca2++ 2P0432P043-
+ 4Hf
CaO . Ca3(POa)z+ HzO .-+ 2P043and a reprecipitation as:
+
4Ca3(P04)z HzO
+ 4Hf
--
3[CaO . Ca3(P04)2]+ HzO
(1)
2HP042-
4Ca2++ 2(P04)3-+ 20H-
(2)
2HP042-
Ca10(P04)6(OH)2 + 2Ca2+ + 2HP042Calo(P04)6(0H)2 + 2Ca2++ 40H-
(3)
(4)
Equations (1)and (2) show that by dissolution an increase in pH would be result, while Eqs. (3) and (4) show that when reprecipitation would proceed accordingly and would be complete tricalcium phosphate result in a more acidic environment and tetracalcium phosphate in a more basic environment. It is possible that these reactions caused a good bone bonding and bone formation. CONCLUSION
This mechanical and histological study documented the bone bonding response to different calcium phosphate coatings on titanium alloy implants in femoral bone after 3 months of implantation. A number of tentative conclusions can be drawn from the findings. First, push-out tests show a higher shear-strength with HA and TETRA coatings (+30 MPa) than with a-TCP or blanco titanium (+lo MPa). The second conclusion is that histological observations revealed an excellent bone contact with HA or TETRA, while a-TCP and blanco titanium give rise to more bone remodelling lacunae and/or connective tissue along the implant surface. Furthermore, coatings reprecipitation of HA on TETRA may be the reason that TETRA shows good bone contact.
PLASMA-SPRAYED COATINGS ON TITANIUM ALLOY
65
B. J. H. van der Water, Department of Experimental Medicine, is acknowledged for laboratory animals assistance.
References 1. 2. 3. 4.
5.
6.
7. 8. 9.
10. 11. 12. 13. 14.
M. Jarcho, ”Biomaterial aspects of calcium phosphates,” Reconstr. Implants Surg. Implant Prostkodontics, 30, 25-47 (1986). D. F. Williams, Biocompatibility of Clinical Implant Materials, CRC Press, Boca Raton, 1981, Chap. 2. C. P. A.T. Klein, J. M. A. de Blieck-Hogervorst, J.G.C. Wolke, and K. de Groot, “Studies of the solubilities of different calcium phosphate ceramic particles in vitro,” Biomaterials, in press. K. de Groot, R. G.T. Geesink, C. P. A.T. Klein, and P. Serekian, ”Plasmasprayed coatings of hydroxylapatite,“ J. Biomed. Muter. Res., 21, 13751381 (1987). W. J. A. Dhert, C. P. A.T. Klein, J.G.C. Wolke, E. A. van der Velde, K. de Groot, and P. M. Rozing, ”Fluorapatite-, magnesiumwhitlockite-, and hydroxyapatite- coated titanium plugs: mechanical bonding and the effect of different sites,” 7‘h Cimtec. World Ceramic Congress, June 24-30, 1990 Montecatini Terme, Italy. H. B. M. van der Lubbe, C. P. A.T. Klein, and K. de Groot, “A simple method for preparing thin (10 pm) histological sections of undecalcified plastic embedded bone with implants,” Stain & Tecknol., 63, 171176 (1988). Joint Committee on Powder Diffraction Standards (JCPDS)-International Centre for Diffraction Data, Swarthmore, PA, USA. (Distribution centre for X-ray powder diffraction data). B. Koch, J.G.C. Wolke, and K. de Groot, ”X-ray diffraction studies on plasma-sprayed calcium phosphate coated implants,” (1989,J. Biomed. Muter. Res., in press. D. J. Simmons, “Fracture healing,” in M. R. Urist (ed.), Fundamentals and Clinical Bone Physiology, Lippincott, Philadelphia, 1980. R.G.T. Geesink, K. de Groot, and C. P. A.T. Klein, “Bone bonding to apatite coated implants,” J. Bone Joint Surg., 70B, 17-22 (1988). P. Adam, A.O. Nebelung, and M. Vogt, ”Loslichkeit und Umwandlungen biokeramischer Schichten,” Z . Zahnarztl. Implantol., IV, 15-21 (1988). K. J. Anusavice, P. H. Dehoff, and C.W. Fairhurst, “Comparative evaluation of ceramic-metal bond tests using finite element stress analysis,” J. Dent. Res., 59(3), 608-613 (1980). J. Black, ”Push-out tests,” J. Biomed. Muter. Res., 23, 1243-1245 (1989). E C. M. Driessens, Mineral Aspects in Dentistry, Karger, Basel, 1982.
Received October 24, 1989 Accepted June 21, 1990
tite and tetracalciumphosphate coatings show an interface strength after 3 months of implantation of 34.3 f 6.5 MPa and 26.8 2 3.9 MPa, respectively, while a-TCP and blanco titanium lead to an interface strength of 10.0 2 3.5 MPa and 9.7 1.3 MPa, respectively. Histological examinations revealed that hydroxylapatite and tetracalciumphosphate give rise to an excellent bone formation, while a-TCP and blanco titanium evoked remodeling and less bone contact. +_
INTRODUCTION
As with all implanted devices the requirements of the materials used for their construction are varied, but can be broadly classified under the headings of biocompatibility, biofunctionality, and availability. Biocompatibility is concerned with the interactions between materials and the tissues of the body and is the most important factor involved with this material selection. In particular, because of the good biocompatibility, calcium phosphate ceramics have received wide attention. These compounds have a close chemical resemblance to bone mineral and they give rise to an intimate contact with the surrounding host bone.' 13iofunctionalityis concerned with those mechanical and physical properties that enable the implanted device to perform its function. Calcium phosphate ceramics can be made with properties simulating hard tissues to a large extent, dense sintered ceramics may have compressive strengths up to 5000 kg,'cm2. However, due to their brittle nature they are useful only in applications, where the mechanical forces are nonexistent or 'To whom correspondence should be addressed. Journal of Biomedical Materials Research, Vol. 25, 53-65 (1991) CCC 0021-9304/91/Ol0053-13$04.00 0 1991 John Wiley & Sons, I[nc.
KLEIN ET AL.
54
primarily compressive. The progress of bioengineering has resulted in a new technique, plasma-spraying, which can now overcome mechanical drawbacks of bioactive ceramics. Composites with a biocompatible surface of calcium phosphate ceramics and an inner structure of metal for obtaining sufficient mechanical strength are required, if load-bearing parts of the skeleton are to be replaced. Implant coating seems to be an attractive alternative in obtaining a composite material that combines the inner strength of a metal with the biological surface characteristics of calcium phosphates. Metals in use for coating substrate, particularly titanium, have a long clinical history of implantation and need no special attention.’ Recently, calcium phosphate particles are applied in the plasma-spraying technique as coating on metallic prostheses. These coatings are usually very thin, 50 pm or less. Therefore, even slight solubilities might lead to dissolution of those coatings. Previous biodegradation behavior studies of hydroxylapatite and j3whitlockite revealed that P-whitlockite evoked bone remodeling and porous bone formation, probably due to its bioresorption. From in vitro studies it is known that the solubilities of various calcium-phosphates depend on the buffer solvent, as well as on pH-value and the specific calcium phosphate salt.3 It is very interesting to relate the in vivo biodegradation behavior of different coatings with respect to their bonding properties and their interaction to bone formation. In this paper we present a mechanical push-out study and histological examinations of the interface.
MATERIALS A N D METHODS
Cylindrical rods of standard Ti-6A1-4V titanium alloy were prepared. A coating of hydroxylapatite (HA), tetracalciumphosphate (TETRA) or a-TCP, 50 pm thickness, was applied using the plasma-spray technique as described ear lie^.^ Noncoated titanium (Ti) plug implants of the same dimensions as the coated plugs were used as controls (4.5 mm diameter, 6 mm length). The surface roughness value (Ra) of the different coated titanium plugs were respectively:Ha: Ra 11.26 2 1.90 pm, a-TCP: Ra 9.52 2 0.51 pm, TETRA: Ra 11.18 k 0.97 pm and blanco titanium: Ra 3.88 ? 0.26 pm. The roughness was measured with a Tayler-Hobson, Surtronic 10. Using sterile surgical techniques, holes of 4.7 mm diameter were drilled through the periosteum and cortex, into the medulla, employed a round stainless-steel burr; the rotation speed of the burr was 200 rounds/min). The distance between the holes was 2 cm. During the drilling process the preparation site was continuously cooled with physiological saline. Bone debris was carefully removed by rinsing the holes. The plugs were inserted press-fit into these holes (depth 12-14 mm) in the lateral cortex of the femora of adult beagles. To study exclusively the bone-bonding properties of the materials, the holes were slightly oversized. A total of 18 plugs were inserted into the femora of 3 dogs (3 plugs per femur). The distribution of the plugs was performed rand~mly.~
PLASMA-SPRAYED COATINGS ON TITANIUM ALLOY
55
There were no surgical complications. After a follow-up period of 3 months the dogs were sacrificed using a lethal dose of thiopental. After explant radiography, the femora were used for mechanical testing and histological examination. Each femur was sectioned into three pieces, each one with one plug. After removal of excess soft tissue samples of each testing material were used for mechanical testing and following histology. Samples of each testing material were used directly within 1 h, without any fixation procedure, for mechanical testing. Afterward these samples were fixed for histological procedures6 For the mechanical testing the bone pieces were positioned in a testing jig to allow accurate lining of the loading axis of the test machine with the long axis of the plug (Fig. I). The plugs were pushed out from the surrounding bone, using a Zwick test machine with a crosshead speed of 0.5 mm/min. The maximum force to loosen the implant was calculated from the loaddisplacement curves. To calculate a "shear strength" of the interface, the push-out force was divided by the total plug/cortical bone contact area; contact area = 3.14 x diameter plug x average cortex thickness the cortex was measured on both sides of the plug, and the average of these two measurements was taken as "average cortex thickness." For the histological observations the samples were fixed in formaldehyde buffer, followed by alcohol dehydration and embedding in polymethylmethacrylate. Undecalcified saw sections of 10 pm were prepared using a diamond cutting wheel with cooling.6 Tissue staining was performed with alcain-blue or Masson trichrome. Histology was reviewed with the following objectives: character of coating-bone interface, condition of the coating, response with cortical and trabecular bone, quality and quantity of new bone formation, and response of bone marrow. RESULTS
The x-ray diffraction pattern of the HA coating was comparable with the file #24-33 of the joint (Committee on Powder Diffraction Standards7 By
Diagram of Test Apparatus Figure 1. A schematic drawing of the test apparatus for push-out tests.
KLEIN ET AL.
56
plasma-spraying of P-whitlockite the crystallographic structure is changed from a pure hexagonal into a monoclinic a-tricalciumphosphate (according to JCPDS #29, 359). The diffractogram of TETRA was consistent with JCPDS #25-1137, referring to CaO(P04)2tetracalcium monoxide diphosphate (see Fig. 2):
Radiography All coated implants displayed proliferation of bone along the protruding periosteal and endosteal part of the implant (Fig. 3). Around the implant a proliferation and densification of trabecular bone can be seen at 3 months of implantation independent of the coating type.
Mechanical testing The push-out data are summarized in Table I. HA showed the highest push-out strength, followed closely by TETRA. a-TCP developed a bonding, similar to the uncoated titanium plug. The failure occurred at the coating bone interface, within the coating itself and at the interface coating-substrate. Failure occurred more or less evenly divided over these three levels.
I
440
~
3+5
h
30
~
23
~
208
Figure 2. The x-ray diffraction pattern of hydroxylapatite (HA), tetracalciumphosphate (TetraCP) and a-TCP estimated after the plasma-spraying procedure.
PLASMA-SPRAYED COATINGS ON TITANIUM ALLOY
57
Tetra
TC P
i
-A__-
40
35 3'0 Figure 2. (continued)
1
2.5
2013
Histology Histological sections of the HA-coated titanium plugs show clearly that defects in the periosteum and bone around the implant are filled with newly formed bone. Periosteal and endosteal bone proliferates along the protruding ends of the plugs in and out the femur. The cortical bone area is in very close contact with the implant (circa 90%), without any interposition of fibrous tissue. The cortex shows remodeling with osteoblasts and osteoclasts along the lacunae. Not only mature osteocytes, but also numerous young osteoblasts and complete haversian systems can be seen in direct contact with the HA
KLEIN ET AL.
58
Figure 3. Radiograph of a femur implanted with three plugs coated with hydroxylapatite, tetracalciumphosphate, or a-TCP. After 3 months trabecular bone growth can be observed around the implants.
coating. The coated plug was predominantly enclosed with both cortical and trabecular bone. The last one had been developed from the endosteum in the marrow cavity and in close contact with the implant (Figs. 4 and 5). After 3 months of implantation the hydroxylapatite coating was still distinctable, the thickness and density showed some variety. There was some fragmentation; phagocytic activity could be observed at some spots, but resorption could not be clearly seen. From the sections made from samples which were used first in the mechanical tests, it can be seen that the coating was incorporated in the newly formed bone and loose of the plug. The histological sections of TETRA coating revealed a bone reaction similar to that with HA coatings. Also a very intimate bone contact (over 290% of the implant surface) could be observed. The bone formation and architecture along the TETRA coating is the same as along HA (Figs. 6 and 7). After TABLE I Average Plug-Bone Interface Strength at 3 Months (n = 5) Load ( N ) HA TETRA
CY-TCP Ti
132 f 111 2 37 k 44 f
9.4 15.6 9.3 10.5
Shear Strength (MPa)
Cortex (mm)
34.3 f 6.5 26.8 & 3.9 10.0 k 3.5 9.7 & 1.3
2.6 f 0.2 2.7 ? 0.4 2.4 & 0.3 2.7 f 0.2
PLASMA-SPRAYED COATINGS ON TITANIUM ALLOY
Figure 4. Hydroxylapatite coating showed a very intimate mature bone contact. The coating is incorporated in the new formed bone. Implantation period of 3 months (staining of alcain-blue, original magnification X250).
Figure 5. Hydroxylapatite coating showed a decrease of density. Bone remodeling (osteoid, osteoclasts and osteoblasts) occurred, bone cells were present against the metal surface. Implantation period of 3 months (alcainblue staining, original magnification X250).
59
60
KLEIN ET AL.
Figure 6. Tetracalciumphosphate coating showed a decrease of density. An intimate mature bone contact can be seen. Osteoblasts activity can be observed against the coating surface (alcain-blue staining, original magnification X250).
Figure 7. Tetracalciumphosphate coating after 3 months of implantation showed a decrease of density and thickness. Osteoclastic cells can be seen (alcain-blue staining, original magnification X250).
PLASMA-SPRAYED COATINGS ON TITANIUM ALLOY
61
3 months the coating is still present. Osteoclasts or multinuclear cells can be seen near the coating material. Alpha-TCP (tricalciumphosphate) coatings give rise to bone formation from periosteum, cortex, and endosteum. However, the bone consisted of many remodeling lacunae, which also can be seen along the implant surface. So, this coating showed significantly less bone contact ( C50%) than hydroxylapatite and tetracalciumphosphate coatings. In the lacunae osteoid, osteoblasts, osteoclasts, and plasma cells can be seen. Especially near the coating osteoclasts or multinuclear cells were present (Fig. 8). The coating is still detectable in the sections. It was mostly anchored in the newly formed bone and, but in contrast to the other ones, resorption of the coating can be clearly seen. The blanco titanium plugs evoked a similar bone response as a-TCP coating. Hence, little bony contact ( ~ 5 0 %with ) the implant surface and many remodeling lacunae in the cortex and along the implant. Connective tissue against the implant surface can also be seen (Figs. 9 and 10). DISCUSSION
Dependent of the type of bone fracture, place and many other factors the time to repair in long bone fractures varies between 6 and 12 weeks in case of the cortex.’ Because one can expect a total bone healing after 3 months, the first observations of the in vivo study of interface bone-calcium phosphate
Figure 8. a-TCP coating after 3 months of implantation gives rise to bone remodeling with many osteoclasts. Also plasma cells can be observed in the bone lacunae. Coating density and thickness was decreased (alcain-blue staining, original magnification X250).
62
KLEIN ET AL.
Figure 9. Uncoated titanium plugs showed a little bone contact and mainly bone remodeling lacunae around the implant surface (Masson trichrome staining, original magnification X160).
coatings on titanium plugs were made after 3 months of implantation. Also, earlier studies with calciumphosphate ceramics were described after this implantation period. Long-term implantation studies of 6 months, 1 and 2 years are started and will be published later. To get some insight into the in vivo behavior of different calcium phosphates after an implantation period, a dynamic-functionally loaded model was chosen. In addition, with this model results can be compared with other studies." The x-ray diffraction patterns of the different calcium phosphate coatings after plasma-spraying were compared with the files of the Joint Committee on Powder Diffractions Standards.' It can be noticed that tricalciumphosphate powders with a P-whitlockite crystallographic structure change after the plasma-spraying procedure in an a-tricalciumphosphate crystallographic structure. Sintering of tricalciumphosphate powders does not lead to a change of the P-whitlockite structure.* Due to the fact that the implants were not inserted in one geometrical plane, the x-ray pictures (Fig. 3) show different apparent dimensions of the cylindrical implants. The density around the implants in the marrow cavity indicates bone growth against the implants, as confirmed with the histological observations. It is known from previous studies that the time delay and/or preparation with a fixation medium before performing the mechanical testing can lead to other push-out data." Values for bone-plug interface shear strength measured directly after sacrificing the animal were approximately one half lower as compared to those obtained after a period of formaline fixation. To get realistic information about the bonding strength, the mechanical
PLASMA-SPRAYED COATINGS ON TITANIUM ALLOY
63
Figure 10. Uncoated titanium plug give rise to connective tissue formation against the implant surface. (Masson trichrome staining, original magnification x160).
tests were made with untreated ("fresh") samples. Calculations for interface stress are very questionable and very dependent on experimental set-up and assumptions for bony contact area. To allow comparisons with the literature data, both the absolute values for push-out load as well as the relative interface shear strength values are provided in Table I. The bone-plug interface shear strength of hydroxylapatite coatings were in accordance with the results on fresh samples (29 MPa) described by Geesink et al.'' The shear strength of tetracalciumphosphate (26.8 5 3.9 MPa) is, within experimental error, comparable with hydroxylapatite, while a-TCP and blanco titanium plugs gave much lower values (10.0 MPa and 9.7 MPa, respectively) (Table I). One should realize that, due to differences in elastic modulus within the cortical bone area contacting the implant surface leading to stress concentration at the most dense parts, the found values have little or no relation to the "true" shear strength of the interface." BlackI3 suggested, that only under standardized conditions, a push-out test can be useful for comparison only. Dhert et al.I5found that the choice of bone type and animal species can also influence mechanical test results. The surface roughness values (Ra) of the different coatings were approximately the same for all coatings (circa 10 pm), which make it possible to compare the bone bonding. The surface roughness of blanco titanium is lower, but the push-out strength is similar to a-TCP. One can conclude that both a high degradable calcium phosphate coating and metal plugs with a low roughness leads to a low shear strength.
KLEIN ET AL.
64
Comparison of the push-out data with the histological findings shows that a high push-out strength is related to good quality and quantity of the newly formed bone against the implant. An intimate and extensive bone contact leads obviously to a strong bonding at the interface of bone and coating. In the case of a-TCP or blanco titanium the presence of many remodeling lacunae and/or connective tissue is probably responsible for a weaker bonding at the interface. Comparing the in vivo data with the in vitro dissolubility results; shows that the in vitro dissolubility degree is not an important factor related to the in vivo degradation. TETRA ceramic particles had (in all buffers) the highest solubility, whereas in the in vivo studies TETRA showed a very good bone bonding and an apparent shear strength similar to hydroxylapatite. In terms of chemical reaction the dissolution may be described according to Driessen et al.I4as:
-
Ca3(P04)2-+ 3Ca2++ 2P0432P043-
+ 4Hf
CaO . Ca3(POa)z+ HzO .-+ 2P043and a reprecipitation as:
+
4Ca3(P04)z HzO
+ 4Hf
--
3[CaO . Ca3(P04)2]+ HzO
(1)
2HP042-
4Ca2++ 2(P04)3-+ 20H-
(2)
2HP042-
Ca10(P04)6(OH)2 + 2Ca2+ + 2HP042Calo(P04)6(0H)2 + 2Ca2++ 40H-
(3)
(4)
Equations (1)and (2) show that by dissolution an increase in pH would be result, while Eqs. (3) and (4) show that when reprecipitation would proceed accordingly and would be complete tricalcium phosphate result in a more acidic environment and tetracalcium phosphate in a more basic environment. It is possible that these reactions caused a good bone bonding and bone formation. CONCLUSION
This mechanical and histological study documented the bone bonding response to different calcium phosphate coatings on titanium alloy implants in femoral bone after 3 months of implantation. A number of tentative conclusions can be drawn from the findings. First, push-out tests show a higher shear-strength with HA and TETRA coatings (+30 MPa) than with a-TCP or blanco titanium (+lo MPa). The second conclusion is that histological observations revealed an excellent bone contact with HA or TETRA, while a-TCP and blanco titanium give rise to more bone remodelling lacunae and/or connective tissue along the implant surface. Furthermore, coatings reprecipitation of HA on TETRA may be the reason that TETRA shows good bone contact.
PLASMA-SPRAYED COATINGS ON TITANIUM ALLOY
65
B. J. H. van der Water, Department of Experimental Medicine, is acknowledged for laboratory animals assistance.
References 1. 2. 3. 4.
5.
6.
7. 8. 9.
10. 11. 12. 13. 14.
M. Jarcho, ”Biomaterial aspects of calcium phosphates,” Reconstr. Implants Surg. Implant Prostkodontics, 30, 25-47 (1986). D. F. Williams, Biocompatibility of Clinical Implant Materials, CRC Press, Boca Raton, 1981, Chap. 2. C. P. A.T. Klein, J. M. A. de Blieck-Hogervorst, J.G.C. Wolke, and K. de Groot, “Studies of the solubilities of different calcium phosphate ceramic particles in vitro,” Biomaterials, in press. K. de Groot, R. G.T. Geesink, C. P. A.T. Klein, and P. Serekian, ”Plasmasprayed coatings of hydroxylapatite,“ J. Biomed. Muter. Res., 21, 13751381 (1987). W. J. A. Dhert, C. P. A.T. Klein, J.G.C. Wolke, E. A. van der Velde, K. de Groot, and P. M. Rozing, ”Fluorapatite-, magnesiumwhitlockite-, and hydroxyapatite- coated titanium plugs: mechanical bonding and the effect of different sites,” 7‘h Cimtec. World Ceramic Congress, June 24-30, 1990 Montecatini Terme, Italy. H. B. M. van der Lubbe, C. P. A.T. Klein, and K. de Groot, “A simple method for preparing thin (10 pm) histological sections of undecalcified plastic embedded bone with implants,” Stain & Tecknol., 63, 171176 (1988). Joint Committee on Powder Diffraction Standards (JCPDS)-International Centre for Diffraction Data, Swarthmore, PA, USA. (Distribution centre for X-ray powder diffraction data). B. Koch, J.G.C. Wolke, and K. de Groot, ”X-ray diffraction studies on plasma-sprayed calcium phosphate coated implants,” (1989,J. Biomed. Muter. Res., in press. D. J. Simmons, “Fracture healing,” in M. R. Urist (ed.), Fundamentals and Clinical Bone Physiology, Lippincott, Philadelphia, 1980. R.G.T. Geesink, K. de Groot, and C. P. A.T. Klein, “Bone bonding to apatite coated implants,” J. Bone Joint Surg., 70B, 17-22 (1988). P. Adam, A.O. Nebelung, and M. Vogt, ”Loslichkeit und Umwandlungen biokeramischer Schichten,” Z . Zahnarztl. Implantol., IV, 15-21 (1988). K. J. Anusavice, P. H. Dehoff, and C.W. Fairhurst, “Comparative evaluation of ceramic-metal bond tests using finite element stress analysis,” J. Dent. Res., 59(3), 608-613 (1980). J. Black, ”Push-out tests,” J. Biomed. Muter. Res., 23, 1243-1245 (1989). E C. M. Driessens, Mineral Aspects in Dentistry, Karger, Basel, 1982.
Received October 24, 1989 Accepted June 21, 1990
