The effect of sol-gel-formed calcium phosphate coatings on bone ingrowth and osteoconductivity of porous-surfaced Ti alloy implants

H. Q. Nguyen^a, D. A. Deporter^{a, b}, R. M. Pilliar^{, , a, b}, N. Valiquette^a and R. Yakubovich<sup>b

a</sup> Faculty of Dentistry, University of Toronto, 124 Edward Street, Toronto, Canada M5G 1G6
^b Institute of Biomaterials and Biomedical Engineering, University of Toronto, Canada

Received 31 March 2003;  accepted 14 July 2003. ; Available online 23 September 2003.
Biomaterials
Volume 25, Issue 5 , February 2004, Pages 865-876

1. Abstract

Ti-6Al-4V implants formed with a sintered porous surface for implant fixation by bone ingrowth were prepared with or without the addition of a thin surface layer of calcium phosphate (Ca-P) formed using a sol-gel coating technique over the porous surface. The implants were placed transversely across the tibiae of 17 rabbits. Implanted sites were allowed to heal for 2 weeks, after which specimens were retrieved for morphometric assessment using backscattered scanning electron microscopy and quantitative image analysis. Bone formation along the porous-structured implant surface, was measured in relation to the medial and lateral cortices as an indication of implant surface osteoconductivity. The Absolute Contact Length measurements of endosteal bone growth along the porous-surfaced zone were greater with the Ca-P-coated implants compared to the non-Ca-P-coated implants. The Ca-P-coated implants also displayed a trend towards a significant increase in the area of bone ingrowth (Bone Ingrowth Fraction). Finally, there was significantly greater bone-to-implant contact within the sinter neck regions of the Ca-P-coated implants.

Author Keywords: Bone ingrowth; Calcium phosphate coating; Dental implant

2. Article Outline

1. Introduction

2. Materials and methods

2.1. Implants

2.2. Animal model

2.3. Specimen collection

2.4. Morphometric analysis

2.5. Freeze-fracture SEM analysis

2.6. Analysis of bone ingrowth in relation to the "sinter neck" regions of the porous surface

3. Results

3.1. Morphometric analysis

3.2. Freeze-fracture SEM observations

3.3. Bone ingrowth in relation to the sinter necks

4. Discussion

5. Conclusions

Acknowledgements

References

3. 1. Introduction

Bone-interfacing implants made from titanium alloy (Ti-6Al-4V) having a porous surface region formed by sintering a multilayer of atomized metallic powders over a machined core have been used successfully both in orthopaedics (joint replacements) and dentistry (endosseous dental implants) [1, 2, 3 and 4]. Necessary requirements for successful bone ingrowth include suitable initial implant stabilization and an acceptable porous geometry to allow uninhibited bone ingrowth (i.e. pore size >100 
m or so) [1]. Secure implant fixation is achieved through the resulting 3-dimensional mechanical interlocking of bone and implant. For endosteal dental implants, the rapid establishment of osseointegration is important since it both allows earlier functional loading of the implants and reduces the probability of inadvertent early loading and excessive movement of the implant relative to host bone inhibiting new bone formation.

A number of studies have suggested the effectiveness of calcium phosphate surface coatings, nominally hydroxyapatite (HA), to enhance the osteoconductivity of metallic implants [5, 6, 7, 8, 9 and 10]. These have demonstrated more rapid osseointegration of implants with calcium phosphate coatings, most often formed by plasma spraying, as evidenced by increased resistance to torsion and pull-out forces at early times [11, 12, 13 and 14]. The calcium phosphate plasma-sprayed coatings typically are 30-50 
m thick and consist of hydroxyapatite and other phases (amorphous calcium phosphate, tricalcium phosphate) [15] that form during the plasma spray coating operation. Adherence of the coating to the underlying metallic substrate is achieved primarily through mechanical interlocking with a pre-roughened metallic substrate with the integrity and reliability of the coating-to-substrate anchorage being determined by the processing conditions used during plasma spray coating. Delamination and fragmentation of plasma-sprayed `HA' coatings has been reported [16] either due to significant structural defects at the coating-substrate interface introduced during coating application, or as a result of in vivo dissolution of more soluble phases within the layer leading to undermining and release of more stable debris. Liberation of particulate debris can result in local inflammatory reactions leading to bone resorption and eventual implant loosening. Recognizing the potential advantage of calcium phosphate surfaces for promoting osteoconductivity, a number of alternative techniques have been investigated for making thin submicron, surface films of calcium phosphate on metallic substrates [17]. These approaches offer the potential advantage of applying a calcium phosphate surface film over a substrate while retaining desired geometric features of the underlying substrate. Thus, an osteoconductive calcium phosphate film can, for example, be applied over a plasma-sprayed Ti layer or sintered Ti alloy powders forming porous surface zones. This allows both a preferred surface chemistry and geometry in order to promote osteoconductivity and secure implant fixation through mechanical interlock of bone with implant surface elements. In addition, the inadvertent liberation of large (i.e. >1 
m) particulate debris is avoided. The use of physical vapour deposition (magnetron sputtering) or wet chemical methods (sol-gel coating) are methods for forming such thin films that appear to result in strong bonding of the calcium phosphate films to the underlying substrate thereby further reducing the risk of film delamination and consequent inflammatory reactions [9, 10, 11, 12, 13 and 18].

Sol-gel wet chemical processing uses relatively low temperature anneals to form calcium phosphate films from appropriate calcium and phosphate containing precursor solutions. One variant of this method involves dip coating an implant into a colloidal inorganic suspension, which is allowed to dry and then annealed to form the desired calcium phosphate film [19]. In the present study, this method was used to apply thin calcium phosphate films ~1 
m thickness over Ti-6Al-4V implants prepared with a sintered porous surface zone approximately 300 
m deep [20]. The thin sol-gel-formed Ca-P films deposited over the sintered Ti alloy particles had minimal effect on the 3-dimensional open-pored structure intended for primary and long-term implant fixation. It was the purpose of this study to determine whether the addition of this thin calcium phosphate layer would result in increased osteoconductivity as well as more complete bone ingrowth at early post-implantation periods compared with that already inherent with the porous surface structure itself [21].

4. 2. Materials and methods

2.1. Implants

For this investigation, 34 porous-surfaced cylindrical implants (3.5 mm diameter×8 mm length; Fig. 1) were custom-made (Innova Corporation, Toronto). Core implant forms with one end having an expanded flat head (to facilitate standardized implant placement with the underside of the flat head juxtaposing the periosteal bone surface) were machined from Ti-6Al-4V rod stock and subsequently prepared with a 0.3 mm thick (approximately) sintered porous surface region. As described elsewhere, this porous region is made by vacuum sintering atomized Ti-6Al-4V powders of 45-150 
m size range [22 and 23]. All implants were cleaned and passivated as previously described [24] after which 18 were further treated with a calcium phosphate sol-gel coating. The sol-gel coating procedure was based on that reported by Qui et al. [19] but modified to allow coating of the Ti-6Al-4V substrate as opposed to the glass and quartz substrates that Qui et al. treated. This modified treatment involved the use of a vacuum annealing treatment (10⁻³-10⁻⁴ Torr at temperature) to minimize oxidation of the Ti alloy substrate. This required careful monitoring of the heat-up rate and vacuum since burn-off of the organic constituents of the sol-gel coating during annealing resulted in the release of elements that caused bursts of pressure increase. The final annealing temperature was set at 760°C (±10°C) with a 15 min hold time. The samples were then furnace cooled in vacuum to room temperature.

(41K)

Fig. 1. Porous-surfaced Ti-6Al-4V implants used in this study.

The solution used for the sol-gel dip-coating process was of sufficiently low viscosity to allow it to completely permeate throughout the porous region. To determine the film structure and chemistry after annealing, 10 mm diameter discs (~2 mm thick) that provided flat surfaces for analysis were heat-treated using the same annealing process as the porous-surfaced implants. The disc samples were characterized using thin film X-ray diffraction (TFXRD) (to determine the crystallinity and phases present in the film), relective-mode Fourier transform infrared spectroscopy (RM-FTIR) (to identify the chemical groups present within the calcium phosphate layer), and X-ray photoelectron spectroscopy (XPS) (to characterize the surface chemistry and determine the Ca:P ratio of the sol-gel-formed film). The porous-surfaced implants were examined using scanning electron microscopy (SEM) (secondary electron imaging) in order to characterize the surface topography of the calcium phosphate-coated samples. 18 implants were sol-gel coated while the remaining 16 implants were left uncoated and served as controls. All implants were sterilized by gamma irradiation.

2.2. Animal model

Seventeen white male New Zealand rabbits (wt. ~4.5 kg) were utilized following procedures conforming to the standards of the Animal Care Act as administered by the Canadian Council on Animal Care. All implants were placed with the animals under general anesthesia. Three experiments were undertaken as follows:

Experiment I. Ten rabbits each received one control and one experimental (sol-gel coated) implant in the diaphysis of the right tibia, separated by a space ~15 mm. Bicortical stabilization in the medial and lateral cortices was attempted with both implants. In five of these, the experimental implant was placed in the more proximal position, and in the other five animals it was installed in the more distal position.

Experiment II. The control implants were placed in one tibia of each of six animals and the experimental implants installed in the contra-lateral tibia with both implants in approximately the same position and in contact with both cortices. Three of these rabbits received calcium phosphate-coated implants in the right tibia and the other three received calcium phosphate-coated implants in the left tibia.

Experiment III. One rabbit received one calcium phosphate-coated implant in each tibia. These two implants were used to examine the implant-to-bone interface using the techniques of freeze-fracturing and scanning electron microscopy.

For implant installation, using sterile technique a longitudinal incision was made through the shaved dermis on the medial surface in the proximal diaphyseal region of the tibia. Skin, muscle and periosteum were dissected to expose the tibial bone after which a series of surgical burs were used at speeds of 8000-10,000 rpm and under continuous saline irrigation to prepare the implant site(s). The intent was to penetrate both the medial and lateral cortices of the tibia so as to permit secure initial fixation of the implant in each case. Each prepared site was slightly undersized with respect to the implant diameter thereby ensuring a tight initial "press-fit" of the implant. The implants were hand-torqued to full depth in the prepared sites using a hemostat and grasping the flattened head present on one end of the implant. Afterwards, the reflected tissues were re-approximated and sutured in separate layers using 4-0 Vicryl suture material (Ethicon Sutures Ltd.), and the animals given Buprenorphine HCL (0.3 mg/kg IM) as required to manage post-operative pain.

2.3. Specimen collection

All animals were killed after 2 weeks of healing by injecting T-61 (Intervet, Whitby, Ont., Canada) into the marginal ear vein. Tibial specimens containing the implants were removed using a Stryker saw and immediately fixed in a mixture of formaldehyde, methanol and water, embedded in Osteobed (methylmethacrylate with dibutyl pthalate and catalyzed with benzyl peroxide; Polysciences Inc., Warrington, PA), and prepared as non-demineralized sections (~40 
m thickness) of bone and implant using a diamond wafering blade and petrographic grinding technique as described previously [22]. Eight sections were obtained from each implant according to the scheme depicted in Fig. 2. The first saw cut was in the cross-sectional plane in relation to the tibia and along the long axis of the implant so as to yield proximal and distal halves. The proximal half-block was then sectioned along the long axis of the tibia to yield proximal anterior and proximal posterior quarters. The distal half-block was similarly sectioned to produce distal anterior and distal posterior quarters. The cut surfaces of these quarter blocks were sectioned further so as finally to yield a proximal anterior cross-section (PAX), a proximal anterior longitudinal section (PAL), a proximal posterior cross-section (PPX), a proximal posterior longitudinal section (PPL), a distal anterior cross-section (DAX), a distal anterior longitudinal section (DAL), a distal posterior cross-section (DPX) and a distal posterior longitudinal section (DPL) for each implant.

(8K)

Fig. 2. (a) Schematic representation showing positioning of implants in rabbit tibial sites. (b) Schematic representation indicating the sections prepared for histomorphometric and back-scattered SEM assessment; proximal anterior cross-section (PAX), proximal anterior longitudinal section (PAL), proximal posterior longitudinal section (PPL), proximal posterior cross-section (PAX), distal anterior cross-section (DAX), distal anterior longitudinal section (DAL), distal posterior longitudinal section (DPL), and distal posterior cross-section (DPX).

2.4. Morphometric analysis

Each section was examined by backscattered electron microscopy and the resulting micrographs (Fig. 3) printed at a final magnification of 34×, and analyzed using quantitative image analysis (Bioquant Image Analyzer, R&M Biometrics, TN, USA) to determine the relative osteoconductivity of the two implant surface treatments. To do this (Fig. 4a), direct contact of bone and other (non-bone) tissues with the outermost layer of the sintered particles from each of the endosteal cortical bone surfaces (i.e. medial and lateral) to the farthest point of bone contact along the implant surface was determined. From these measurements the following histomorphometric parameters were determined:

1. Absolute contact length (ACL), i.e. the total length of bone contact with the outermost layer of particles;

2. Maximal contact length (MCL), i.e. the total length of non-bone contact with the outermost layer of particles was added to the ACL to give the total length of implant surface available for bone contact;

3. Straight line bone growth (SLBG) which was measured as a straight line from the endosteal cortical bone level to the furthest point of bone growth (Fig. 3). In addition to these length measurements, the following area determinations were made for both ends of each implant in each section within a rectangular area ( Fig. 4) defined by the endosteal surface of the associated cortex, the furthest extent of linear bone growth along the outermost aspect of the porous surface region, the solid implant core, and a line tangent to the outermost aspect of the porous surface layer and parallel to the solid implant core:

4. Bone ingrowth area (BIA), i.e. the area of bone tissue which had grown into the porous surface layer at each end of each implant and bounded by the defined rectangle; and

5. Maximum area available for bone ingrowth (MBIA) determined by subtracting the total area of the spherical particles of the porous surface layer within the defined rectangle from the total area of the rectangle.From these data the following parameters were determined [5 and 24]:

6. Contact Length Fraction (CLF)=(ACL/MCL)×100.

7. Bone Ingrowth Fraction (BIF)=(BIA/MBIA)×100.

(25K)

Fig. 3. Scanning electron micrograph of a sample implant (plane of SECTION=DPL) showing the typical relationship of implant to bone. A short segment (^*) of the flattened implant head (upper right corner) can be seen in contact with the medial cortex (M). Examples of the measurement, straight line linear bone growth (SLBG) in relation to the medial and lateral (L) cortices are also shown.

(29K)

Fig. 4. Schematic showing how absolute contact length (ACL), maximal contact length (MCL) and area measurements were determined. The outermost aspect of the porous surface layer was traced (i) to the furthest point of bone contact from the endosteal surface where it contacted bone (hatched shading) to give the ACL and (ii) from this point to where it contacted tissue other than bone. These two measured lengths were summed to provide the MCL. The area measurements were made within a defined rectangle the sides of which compromised the endosteal surface of the associated cortex (a-c), the furthest point of bone contact along the porous surface (b-d), the solid implant core (c-d), and a line drawn parallel to the solid implant core and tangent to the outermost layer of particles of the porous surface (a-b). The area (BIA) occupied by bone (hatched shading) and by the particles of the porous surface layer (PSA) was traced separately, while the maximal area available for bone ingrowth (MBIA) was calculated by subtracting the PSA from the total area of the defined rectangle.

ACL, CLF, SLBG, BIA and BIF from experiments I and II were submitted to multiple factor analysis of variance (ANOVA) for repeated measures using the general linear models procedure of SAS. The ANOVA tested main effects and interactions between factors (p<0.05). If significant main effects were found between factors tested (i.e. implant surface type, implant position, medial vs. lateral aspect of implant, implant type to position interaction, experiment I vs. experiment II), the Duncan multiple-range test was used for pair-wise comparisons of the means (p<0.05).

2.5. Freeze-fracture SEM analysis

The two calcium phosphate-coated implant-containing specimens from experiment III were fixed and dehydrated similarly to those from the other animals. The tissue blocks then were frozen in liquid nitrogen and freeze-fractured to permit separation of bone and implant surface using a knife and mallet. The two implants with whatever bone remained attached and the two separated bone fragments were all critical point dried with liquid CO₂ using standard methods, and subsequently mounted on aluminum studs and sputter-coated with 3 nm of platinum. Specimens were examined by secondary electron imaging (Hitachi S-2500 SEM operating at 10 kV).

2.6. Analysis of bone ingrowth in relation to the "sinter neck" regions of the porous surface

During the sintering process employed to create the porous-surfaced titanium alloy implants, sinter necks are formed at particle-particle and particle-implant core junctions (Fig. 5a). The sinter necks have a concavo-convex geometry with the radius of curvature of the concavity being small (equal to approximately 2-3 
m). Bone growth into such small cusp-like regions appears to be inhibited, at least at 2 weeks after implant placement, so that a very small region devoid of bone often results. Examination of the freeze-fractured specimens from experiment III gave the subjective impression of more complete bone filling into the calcium phosphate-coated sinter neck regions than for the porous-surfaced implants without a Ca-P coating. Therefore, all implants in experiments I and II were re-examined using backscattered SEM to assess the extent of bone ingrowth or fill within the sinter neck regions and, in particular, if the fill into this region was increased by the presence of the calcium phosphate film. To do this, two scanning electron micrographs (one of the implant surface in contact with each of the medial and lateral cortices) were prepared from each implant section at a final magnification of 230×. `Sinter neck regions' were defined arbitrarily as the area of porosity adjacent to a sinter neck and bounded by a line drawn tangent to the smaller of the two connected particles and parallel to a line through the centre of these particles (Fig. 5b). Using the micrographs, a non-parametric analysis was used to assess bone fill (ingrowth) into the sinter neck regions of all implants as follows. Bone ingrowth into any sinter neck region was assigned to one of three possible groups. A sinter neck region assigned to Type 1 had up to a third of its area occupied by bone, while sinter neck regions assigned to Types 2 and 3 had between one-third and two-thirds, and between two-thirds and 100% respectively of the designated sinter neck regions filled with bone. For each micrograph all sinter neck regions were categorized and the number of sinter neck regions in each of the three groups determined. The results ( Table 2) are presented as the percentage of each category found within the calcium phosphate-coated and uncoated sample sinter neck regions examined. These data were submitted to ANOVA using SAS at the 5% level of significance to determine if differences existed for the two different implant surface preparations.

(64K)

Fig. 5. (a) SEM micrograph of a porous-surfaced control implant (500×). Sinter necks and thermal etch lines can be seen. (b) The morphometric technique for determining the degree of bone ingrowth into the sinter neck regions. Each sinter neck region was defined by a line drawn tangent to the smaller of the two connected spherical particles forming the sinter neck, and a second line parallel to this first line and through the center of the particles.

5. 3. Results

Fig. 5 and Fig. 6 show the typical SEM appearances of the two types of implants used. The scanning electron micrographs of the calcium phosphate-coated implant (Figs. 6a and b) verify that the sol-gel coating process did not cause any significant change in the size or shape of the pores intended for bone ingrowth. Addition of the calcium phosphate film did, however, alter the topography of the sintered particles. SEM examination revealed a somewhat more irregular surface. High magnification SEM examination, ( Figs. 6c and d) revealed that the calcium phosphate film was thickest in the sinter neck regions (approaching 2 
m). This was verified by examination of freeze-fractured samples (see below).

(231K)

Fig. 6. Scanning electron micrographs depicting the typical appearance of calcium phosphate-coated implants (a=300×; B=500×). Higher magnification micrographs of the calcium phosphate coating in a sinter neck region are shown in (c) at 2000× and (d) 5000×. In these regions, the HA film appeared thicker and showed signs of delaminating in some regions.

The XPS analysis of the annealed film indicated that the Ca:P ratio was equal to approximately 1.7 (±0.2). This corresponds to the expected Ca:P ratio for calcium hydroxyapatite (Ca₁₀(PO₄)₆(OH)₂, HA). Qui et al. [19] reported the formation of an HA phase after annealing Ca-P sol-gel-coated glass or quartz discs below 1000°C using the same coating solutions as in the present study (but annealed in air rather than vacuum).

Thin film X-ray diffraction (TF-XRD) of our sol-gel-coated samples indicated the expected diffraction peaks corresponding to the Ti-6Al-4V substrate as well as additional peaks corresponding to calcium hydroxyapatite (Ref-JCPDS International Centre for Diffraction Data card #09-0432). The diffraction peaks corresponding to the calcium hydroxyapatite were broad suggesting a poorly crystallized HA. No other phases could be definitely detected. The RM-FTIR spectra indicated the presence of PO₄³⁻ groups.

In experiment I in which two implants (one control, one experimental) had been placed in the right tibia only of 10 experimental animals, one control and one calcium phosphate-coated implant placed in the more proximal position contacted the mesial cortex only (i.e. did not reach and engage the lateral cortex). The control implant in this situation did not become osseointegrated and was not included in the morphometric analyses. The one calcium phosphate-coated implant engaging only the mesial cortex did become osseointegrated to this cortex, and its implant-to-bone interface with this cortex was included in the morphometric analyses. In experiment II in which only one implant was placed in each tibia, all implants contacted both cortices, making this the more desirable experimental model.

3.1. Morphometric analysis

The morphometric data were submitted to ANOVA and this revealed that there were significant differences as outlined below in several of the parameters studied. There were no significant differences (p<0.05) between the medial and lateral measurements amongst the various sections for either the calcium phosphate-coated or non-calcium phosphate-coated implants. Likewise, when comparing implant surface (i.e. sections PAX, DAX, DPX, etc.), proximal versus distal implant position in experiment I, or the data from experiment I versus experiment II (p<0.05), there were no significant differences for either calcium phosphate-coated or non-coated implants.

In view of the finding of no statistically significant differences for the results of experiments I and II, the results were combined and are presented as one set of data in Table 1. Analysis of the data indicated that absolute contact length (ACL) was significantly higher (p<0.0001) for calcium phosphate-coated implants (1.18 mm) than for the non-coated implants (0.74 mm) as was the contact length fraction (CLF) (40.4% vs. 27.0%; p<0.0001) and straight line bone growth (SLBG) (1.19 mm vs. 1.04 mm; p<0.001). There were no differences in bone ingrowth area (BIA) between implant types, but the bone ingrowth fraction (BIF) was greater for the calcium phosphate-coated implants. The difference was marginally significant (34.7% vs. 31.0%; p=0.06).

Table 1. Combined morphometric data from experiments I & II

3.2. Freeze-fracture SEM observations

Examination of SEM micrographs of the sol-gel-coated implants recovered by freeze-fracture technique verified extensive bone growth from the endosteum along the implant surface. Numerous bone spicules were observed attached to the implant surface in the medullary region. In sample micrographs (Fig. 7), it can be seen that the calcium phosphate layer was well-adapted to the metal substrate (Fig. 7a), that bone was in direct contact with the calcium phosphate layer ( Figs. 7b-d), and that there were numerous osteocytes throughout the bone which had ingrown the surface porosity ( Figs. 7b-d). In the sinter neck regions, the calcium phosphate film appeared somewhat thicker than elsewhere, and to have delaminated in regions. That this delamination had occurred in vivo rather than in the course of sample preparation was indicated by the presence of bone that had formed between the delaminated calcium phosphate layer and substrate ( Fig. 7d).

(233K)

Fig. 7. Scanning electron micrographs of a freeze-fractured specimen showing that the calcium phosphate film is well-adapted to the Ti-6Al-4V particles (a, b). Bone is observed in direct contact with the HA film (b-d) with numerous osteocytes (b-d). In the sinter neck regions where calcium phosphate film delamination, bone had formed between the film and the underlying Ti alloy substrate (arrows—c, d).

3.3. Bone ingrowth in relation to the sinter necks

As indicated, bone ingrowth into the sinter neck regions observed by backscattered scanning electron microscopy was categorized semi-quantitatively into 3 groups (Types 1-3 as described previously) depending on the extent of bone formation within these regions. Sinter neck regions with the most ingrowth (at least 2/3 of the available area in the defined sinter neck) were classified as Type 3 and those with the least ingrowth (<1/3 of the area available for ingrowth) were allocated to Type 1, the remaining sinter neck regions being classified as Type 2. The data collected suggested more extensive ingrowth into the sinter neck regions of the calcium phosphate-coated implants than the control implants (Table 2). There were significantly more (p=0.0001) sinter neck regions of Types 2 and 3 with the calcium phosphate-coated implants than with the control implants, while the control implants showed significantly more (p=0.0001) Type 1 sinter neck regions than the calcium phosphate-coated implants. To get a weighted average of bone ingrowth into sinter neck regions with the two implant types, the number of sinter neck regions in each group were multiplied by the group number (i.e. 1, 2 or 3). The weighted average ingrowth for the calcium phosphate-coated implants was 2.01, while that for the non-coated implants was 1.47, and the difference was statistically significant (p=0.0001).

Table 2. Bone ingrowth into the neck regions

Group 1: no bone fill to less than 1/3 of the neck region examined.

Group 2: 1/3 to less than 2/3 of the neck region filled with bone.

Group 3: at least 2/3 of the neck region filled with bone.

6. 4. Discussion

Calcium phosphate coatings (nominally hydroxyapatite or mixed hydroxyapatite and tricalcium phosphate) for bone-interfacing devices used in orthopaedics and dentistry have been investigated since the early 1970s as a means of improving "bone bonding" and promoting more rapid osseointegration (i.e. enhance osteoconductivity) [8, 12, 25, 26, 27, 28, 29, 30, 31, 32 and 33]. Calcium phosphate coatings generally have been applied by plasma spraying resulting in rather thick layers (30-100 
m thickness) which may be susceptible to delamination and degradation [9, 10, 11, 12, 13 and 26]. These potential problems suggest that using calcium phosphate plasma-sprayed coatings as the primary means of dental implant fixation may not be advisable. Additionally, they suggest an advantage of thin film coatings (<1
m) that should be less likely to invoke unacceptable inflammatory host response even in the event of some film fragmentation. This has prompted a number of studies for implant surface modification using physical vapour (sputtering) or chemical (precipitation from solution, sol-gel) calcium phosphate deposition methods [15 and 17].

In the present investigation, such thin film surface layers were produced by a sol-gel coating process. Using this process, it was possible to deposit very thin calcium phosphate films onto the sintered porous surface region of Ti-6Al-4V implants. The calcium phosphate film was approximately 1 
m or so thick over most of the sintered particle surfaces with somewhat thicker regions developing (approximately 2 
m) within the sinter neck regions. The thicker deposit in this region is believed to be due to capillary effects drawing the coating solution into this concavo-convex region.

Film characterization indicated that the film resulting from the vacuum anneal process corresponded to a poorly crystalline hydroxyapatite. In general, the HA film appeared well adhered to the underlying Ti alloy substrate except in some sinter neck regions where delamination of some thicker portions was observed. This was believed to be a result of the greater residual stresses that developed within these thicker film regions. The RM-FTIR studies further indicated the presence of OH⁻ and PO₄³⁻ groups within the film. Most significant, however, was the finding that the film appeared to promote greater new bone formation at the implant surface within the 2-week in vivo study period.

The osteoconductive nature of `HA' coatings has been demonstrated by others [12, 28, 29 and 30]. This appears to be further supported by the results of the present investigation in which the extent of bone growth at 2 weeks, as assessed by contact length fraction (CLF) measurements were greater compared with calcium phosphate-coated implants. Furthermore, in the sinter neck regions, greater bone fill occurred by 2 weeks with the calcium phosphate-coated implants. While these differences were significantly different in our experiments, their clinical significance is unknown at this time. While greater bone fill and ingrowth occurred by 2 weeks with the calcium phosphate-coated implants, both designs (HA-coated and uncoated) had adequate mechanical interlocking of bone and the implant porous surface to provide the required fixation for the conditions of this study (healthy rabbits, tibial implant sites with good quality cortical bone, snug initial implant fit). The results do, however, suggest a possible advantage for calcium phosphate-coated, porous-surfaced Ti-6Al-4V implants in less favourable situations in which faster and more extensive bone ingrowth would be desirable.

Using similar animal models but different implant designs and plasma-sprayed `HA' coatings, others [12, 32 and 33] have reported that thicker `HA' layers applied by plasma spraying enhance osteoconduction at early times (out to 6 weeks of healing). However, significant differences in bone-to-implant contact were no longer noted by 12 weeks.

The finding of enhanced osteoconductivity and increased bone contact for implants with calcium phosphate surface coatings is not observed in all studies. Thus, it is noteworthy that Dhert et al. [34] reported that implant integration of press-fit Ti-6Al-4V implants placed in rabbit tibial sites was not affected by a hydroxyapatite plasma-sprayed coating. These studies involved implant placement for 3-28 days in similar sites to those used in the present study. The authors proposed that the most important parameter in initial implant healing was the bone itself rather than implant surface characteristics. It should be noted that the implant designs used in that study were different than those used in the present study. The effect of a 3-dimensional interlock of tissue and implant, as occurs with porous-surfaced implants, results in a significant difference in healing response as demonstrated in a study by Simmons et al. [35] in which a significant difference in rate of bone formation (osseointegration) and the resulting interface shear strength occurred for porous-surfaced versus plasma-sprayed implants at very early (4- and 8-day) post-implantation periods. This may explain the different response in terms of bone-implant integration (with or without a calcium phosphate layer) in our study compared with the Dhert et al. study. The importance of considering all factors, (implant- and tissue-related) in assessing potential interfacial reactions is indicated in view of such seemingly contradictory reports.

The mechanism(s) by which calcium phosphate coatings promote osteoconduction is not yet clear, although both implant topography and surface chemistry are believed to be important. Topography has been postulated as having an effect on fibrin clot retention and osteprogenitor cell migration [36] or on cellular formation of factors influencing cell differentiation [37]. Surface chemistry (calcium phosphate specifically) may result in preferential adsorption of proteins that promote osteoprogenitor cell adherence and activation. Another possibility is that local dissolution of calcium phosphate coatings may provide Ca²⁺ and PO₄³⁻ ions conducive to tissue mineralization and bone formation. The issue is unclear and requires further study. Plasma-sprayed calcium phosphate coatings, like other surface modifications such as those produced by sandblasting, acid etching or plasma spraying with titanium result in irregular or textured surfaces. These all promote osteoconduction to varying degrees [38 and 39]. The sol-gel-formed calcium phosphate film of the present study was observed by SEM to introduce a slight surface texture compared with the non-calcium phosphate-coated surfaces, and this may have contributed to the observed differences in osteoconduction between the two designs. Plasma-sprayed calcium phosphate coatings containing a high percentage of amorphous calcium phosphate (~40%) have been reported to result in increased bone-to-implant contact, an effect possibly related to greater rates of dissolution of the coating and local elevation of calcium and phosphate ions [40, 41 and 42] although it could equally be related to the development of greater surface roughness as a result of heterogeneous coating dissolution. While the poorly crystalline HA film that was formed in the present study would be less stable than stoichiometric HA, there was no indication of significant in vivo dissolution during the 2-week study period. In view of the relatively minor modification in surface topography resulting from the sol-gel-formed thin coating, we believe that preferential protein adsorption is the reason for the observed enhanced osteoconductivity of the calcium phosphate-coated implants. However, this issue requires confirmation through further study.

Whether the results reported here are clinically relevant in terms of reducing the periods of limited loading of porous-surfaced implants in order to allow bone ingrowth to occur remains to be determined, but is suggested by the results of another study using the same model system [43]. Similarly, the effect of calcium phosphate films on implant performance in non-ideal situations (i.e. low density cancellous bone) is not known although the results suggest a possible benefit. While previous studies have indicated that porous-surfaced implants per se are osteoconductive and become osseointegrated in shorter periods compared to other designs [35 and 44] and perform well in bone of low density [45], additional clinical benefits may result with calcium phosphate coating of such devices. The long-term stability of the sol-gel-formed calcium phosphate coatings utilized in this study also is currently under investigation. It is anticipated that the poorly crystalline HA film will degrade but without any negative effects as a result of its limited thickness. The primary purpose of the coating is to provide an early boost in the initial healing response (i.e. promote osteoconductivity), and since the calcium phosphate surface layer is not essential for long term implant stability with porous-surfaced implants, its loss should be of little clinical significance.

7. 5. Conclusions

The formation of a sol-gel-formed calcium phosphate film, identified as a poorly crystallized hydroxyapatite layer, over the sintered porous surface zone of Ti-6Al-4V implants significantly increases the rate of bone ingrowth into the porous region thereby resulting in faster implant fixation and osseointegration. Further, the Ca-P film promotes more complete bone fill of the pore regions. These findings suggest the effectiveness of such sol-gel-formed Ca-P films over porous-surfaced structures for enhancing osteoconductivity although the exact cause (modified surface chemistry or topography) remains uncertain.

8. Acknowledgements

The authors wish to thank Mr. Robert Chernecky and Mr. Lu Gan for valuable technical support, and Mrs. Caroline Chu for secretarial assistance. The work was financed by funds provided by the former Medical Research Council of Canada (now the Canadian Institute for Health Research).

9. References

1. R.M. Pilliar, Porous-surfaced metallic implants for orthopedic applications. J Biomed Mater Res 21 (1987), pp. 1-33. Abstract-MEDLINE | Abstract-EMBASE | Abstract-Compendex  

2. D.A. Deporter, P.A. Watson, R.M. Pilliar, M. Pharoah, D.C. Smith, M. Chipman, D. Locker and A. Rydall, A prospective clinical study in humans of a partially porous-coated, endosseous dental implant: three to four results. Int J Oral Maxillofac Impl 11 (1996), pp. 83-91.

3. D.A. Deporter, R. Todescan, P. Watson, M. Pharoah, D. Levy and K. Nardini, Use of the Endopore^® dental implant to restore single teeth in the maxilla. Protocol and early results. Int J Oral Maxillofac Impl 13 (1998), pp. 263-272. Abstract-MEDLINE  

4. D.A. Deporter, P.A. Watson, R.M. Pilliar, A.M. Melcher, J. Winslow, T.P. Howley, P. Hansel, C. Maniatopoloulos and A. Rodriguez, A histological assessment of the initial healing response adjacent to porous-surfaced titanium alloy dental implants in dogs. J Dent Res 65 (1986), pp. 1064-1070. Abstract-MEDLINE  

5. S.D. Cook, K.A. Thomas, J.F. Kay and M. Jarcho, Interface mechanics and histology of titanium and hydroxylapatite-coated titanium for dental implant applications. Int J Oral Maxillofac Impl 2 (1987), pp. 15-22. Abstract-MEDLINE  

6. K.A. Thomas, J.F. Kay, S.D. Cook and M. Jarcho, The effect of surface macrotexture and hydroxylapatite coating on the mechanical strengths and histologic profiles of titanium implant materials. J Biomed Mater Res 21 (1987), pp. 1395-1414. Abstract-MEDLINE | Abstract-EMBASE | Abstract-Compendex  

7. D.P. Rivero, J. Fox, A.K. Skipor, R.M. Urban and J.O. Galante, Calcium phosphate-coated porous titanium implants for enhanced skeletal fixation. J Biomed Mater Res 22 (1988), pp. 191-201. Abstract-EMBASE | Abstract-MEDLINE  

8. M.S. Block, A. Delgado and M.G. Fontenot, The effect of diameter and length of hydroxyapatite-coated dental implants on ultimate pull-out force in dog alveolar bone. J Oral Maxillofac Surg 48 (1990), pp. 174-178. Abstract-EMBASE | Abstract-MEDLINE  

9. M. Filiaggi, R.M. Pilliar and N.A. Coombs, Characterization of the interface in the plasma-sprayed HA coating/Ti-6Al-4V implant system. J Biomed Mater Res 25 (1991), pp. 1211-1229. Abstract-MEDLINE | Abstract-EMBASE | Abstract-Compendex  

10. Whitehead RY, Lacefield WR, Lucas LC, Lemons JE. Structural and bond strength evaluations of plasma-sprayed HA coatings. Proceedings of the Fourth World Biomaterials Congress, Berlin, 1992. p. 503.

11. J. Delecrin, S. Szmuckler-Moncler, G. Daculsi, J. Rieu and B. Duquet, Ultrastructural crystallographic and chemical analysis of different calcium phosphate plasma coatings. In: W. Bonfield, H. Hastings and K.E. Janner, Editors, Bioceramics vol. 4, Butterworth-Heinemann Ltd., London (1991), pp. 311-315.

12. K. De Groot, R.G.T. Geesink, C. Klein and P. Serekian, Plasma-sprayed coatings of hydroxyapatite. J Biomed Mater Res 21 (1987), pp. 1375-1381. Abstract-MEDLINE | Abstract-Compendex  

13. S.R. Radin and P. Ducheyne, Plasma spraying induced changes of calcium phosphate ceramic characteristics and the effects on in vitro stability. J Mater Sci Med 3 (1992), pp. 33-42.

14. C.P. Klein, A. Driessen and K. De Groot, Relationship between the degradation behaviour of calcium phosphate ceramics and their physical-chemical characteristics and ultrastructural geometry. Biomaterials 5 (1989), pp. 157-160.

15. W.R. Lacefield, Current status of ceramic coatings for dental implants. Impl Dent 7 (1998), pp. 305-314.

16. S.R. Radin and P. Ducheyne, Plasma spraying induced changes in calcium phosphate ceramic characteristics and the effect on in vitro stability. J Mater Sci Mater Med 3 (1992), pp. 33-42.

17. R.M. Pilliar and M.J. Filiaggi, New calcium phosphate coating methods. Bioceramics 6 (1993), pp. 165-171.

18. Gan L, Wang J, Pilliar RM. Calcium phosphate sol-gel-derived thin films on porous-surfaced implants for enhanced osteoconductivity: II. Evaluation of film to substrate interface strength, submitted.

19. Q. Qui, P. Vincent, B. Lowenberg, M. Sayer and J.E. Davies, Bone growth on sol-gel calcium phosphate thin films in vitro. Cells Mater 3 (1993), pp. 351-360.

20. R.M. Pilliar, Overview of surface reactivity of metallic endosseous dental implants: textured and porous surface structured designs. Impl Dent 7 (1998), pp. 305-314. Abstract-MEDLINE  

21. Dziedic DM, Davies JE. Effect of implant surface topography on osteoconduction. Proceedings of 20th Meeting of Society for Biomaterials, Boston, 1993. p. 298.

22. D.A. Deporter, P.A. Watson, R.M. Pilliar, M. Chipman and N. Valiquette, A histological comparison in the dog of porous-coated vs. threaded dental implants. J Dent Res 69 (1990), pp. 1138-1145. Abstract-MEDLINE  

23. R.M. Pilliar, Dental implants: materials and design. JCDA 56 (1990), pp. 857-861. Abstract-MEDLINE  

24. D.A. Deporter, B. Friedland, P.A. Watson, R.M. Pilliar, T.P. Howley, D. Abdulla, A.H. Melcher and D.S. Smith, A clinical and radiographic assessment of a porous-surfaced titanium alloy dental implant system in dogs. J Dent Res 65 (1986), pp. 1071-1077. Abstract-MEDLINE  

25. M. Jasty, H. Rubash, G.D. Paiement, C.R. Bragdon, J. Parr and W. Harris, Porous-surfaced uncemented components in experimental total hip arthroplasty in dogs. Effect of plasma-sprayed calcium phosphate coatings on bone ingrowth. Clin Orthoped Rel Res 280 (1992), pp. 300-309. Abstract-MEDLINE | Abstract-EMBASE  

26. De Groot K, Klein CP, Wolke JG, De Blieck-Hogervorst JM. Plasma-sprayed coatings of calcium phosphate. In: Yamamuro T, Hench LL, Wilson J, editors. CRC handbook of bioactive ceramics, vol. 2. Boca Raton: CRC Press; 1990. p. 133-42.

27. M.S. Block, I.M. Finger, M.G. Fontenot and J.N. Kent, Loaded hydroxylapatite-coated and grit-based titanium implants in dogs. Inter J Maxillofac Impl 4 (1989), pp. 219-225. Abstract-MEDLINE  

28. P. Ducheyne, L.L. Hench, A. Kogan, M. Martens, A. Bursens and J.C. Mulier, Effect of hydroxyapatite impregnation on skeletal bonding of porous-surfaced implants. J Biomed Mater Res 14 (1980), pp. 225-237. Abstract-Compendex | Abstract-EMBASE | Abstract-MEDLINE  

29. R.G.T. Geesink, K. De Groot and C.P.A. Klein, Chemical implant fixation using hydroxl-apatite coatings. Clin Orthoped Rel Res 225 (1987), pp. 147-170. Abstract-MEDLINE | Abstract-EMBASE  

30. J.A. Jansen, J.P. van der Waerden and J.G.C. Wolke, Histologic investigation of the biologic behaviour of different hydroxyapatite plasma-sprayed coatings in rabbits. J Biomed Mater Res 27 (1993), pp. 603-610. Abstract-MEDLINE | Abstract-EMBASE | Abstract-Compendex  

31. J.F. Osborn, Bonding osteogenesis under loaded conditions. The histological evaluation of a human autopsy specimen of a hydroxyapatite ceramic-coated stem of a titanium hip prosthesis. In: H. Ooislie, H. Aoki and R. Sawai, Editors, Bioceramics, Ishiyaka EuroAmerica Inc., Tokyo (1989), pp. 388-399.

32. M. Gottlander and T. Albrektsson, Histomorphometric studies of hydroxyapatite-coated and uncoated cp titanium threaded implants in bone. Int J Oral Maxillofac Impl 6 (1991), pp. 399-404. Abstract-MEDLINE  

33. S.D. Cook, G.C. Baffes, A.J. Palafox, M.W. Wolfe and A. Burgess, Torsional stability of HA-coated and grit-blasted titanium dental implants. J Oral Impl 28 (1992), pp. 359-365. Abstract-MEDLINE  

34. W.J.A. Dhert, P. Thomsen, A.K. Blomgren, M. Esposito, L.E. Ericson and A.J. Verbout, Integration of press-fit implants in cortical bone: a study on interface kinetics. J Biomed Mater Res 41 (1998), pp. 574-583. Abstract-MEDLINE | Abstract-Compendex | Abstract-EMBASE   | Full Text via CrossRef

35. C.A. Simmons, N. Valiquette and R.M. Pilliar, Oseointegration of sintered porous-surfaced and plasma-spray coated implants: an animal model study of early post-implantation healing response and mechanical stability. J Biomed Mater Res 47 (1999), pp. 127-138. Abstract-Compendex | Abstract-MEDLINE | Abstract-EMBASE   | Full Text via CrossRef

36. J.E. Davies, Mechanisms of endosseous integration. Int J Prosthodont 11 (1998), pp. 391-402.

37. K. Kieswetter, Z. Schwartz, T.W. Hummert, D.L. Cochran, J. Simpson and B.D. Boyan, Surface roughness modulates the local production of growth factors and cytokines by osteoblast-like MG-63 cells. J Biomed Mater Res 32 (1996), pp. 55-63. Abstract-MEDLINE   | Full Text via CrossRef

38. A. Schroeder, E. van der Zypen, H. Stich and F. Sutter, The reaction of bone, connective tissue and epithelium to endosteal implants with plasma-sprayed titanium surfaces. J Maxillofac Surg 9 (1981), pp. 15-25. Abstract-EMBASE | Abstract-MEDLINE  

39. D. Buser, R.K. Schenk, S. Steinemann, J.P. Fiorellini, C.H. Fox and H. Stich, Influence of surface characteristics on bone integration of titanium implants. A histomorphometric study in miniature pigs. J Biomed Mater Res 25 (1991), pp. 889-902. Abstract-EMBASE | Abstract-MEDLINE  

40. K. De Groot, J.G. Wolke and J.A. Jansen, State of the art: hydroxyapatite coatings for dental implants. J Oral Impl 20 (1994), pp. 232-234.

41. H. Caulier, J.P. van der Waerden, Y.C. Paquay, J.G. Wolke, W. Walk, I. Naert and J.A. Jansen, Effect of calcium phosphate (CaP) coatings on trabecular bone response: a histological study. J Biomed Mater Res 29 (1995), pp. 1061-1069. Abstract-MEDLINE | Abstract-EMBASE | Abstract-Compendex  

42. H. Caulier, S. Vercaigne, I. Naert, J.P. van der Waerden, J.G. Wolke, W. Walk and J.A. Jansen, The effect of CaP plasma-sprayed coatings on the initial bone healing of oral implants: an experimental study in the goat. J Biomed Mater Res 34 (1997), pp. 121-128. Abstract-MEDLINE | Abstract-EMBASE | Abstract-Compendex   | Full Text via CrossRef

43. Tache A, Gan L, Deporter DA, Pilliar RM. Effect of surface chemistry on the rate of osseointegration of sintered porous-surfaced Ti-6Al-4V implants. Submitted for publication.

44. D.A. Deporter, P.A. Watson, R.M. Pilliar, M. Pharoah, M. Chipman and D.C. Smith, A clinical trial of a partially porous-surfaced endosseous dental implant in humans: protocol and 6 month results. In: W.R. Laney and D.E. Tolman, Editors, Tissue integration in oral, orthopedic and maxillofacial reconstruction, Quintessence, Chicago (1992), pp. 250-258.

45. D.A. Deporter, R. Todescan and N. Riley, Porous-surfaced dental implants in the partially edentulous maxilla: assessment for subclinical mobility. Int J Perio Rest Dent 22 (2002), pp. 184-192. Abstract-MEDLINE  

Corresponding author. Faculty of Dentistry, University of Toronto, 124 Edward Street, , Toronto, , Canada M5G 1G6. Tel.: [phone]x4473

---