Histologic and Biomechanical Evaluation of Alumina-Blasted/Acid-Etched and Resorbable Blasting Media Surfaces

The use of endosseous dental implants is a highly predictable treatment modality supported by a wealth of evidence reporting their safety and high survival rates over the long term.¹  Because the implant surface has been identified as one of the six important factors for implant anchorage in bone,²  there have been remarkable efforts to improve surface design, as evidenced by the extensive number of studies published since the 1980s. Despite the robust and positive clinical results for turned surfaces, which have been shown to last for up to 20 years,³  the purpose of changing implant surface topography and/or chemistry is to improve bone healing and reduce the waiting time between implant placement and its functional loading with a prosthesis.⁴ 

Basically, the engineering processes to modify an implant surface include chemical or physical alterations.⁵  Incorporating inorganic phases into the titanium oxide layer, such as calcium phosphate, has been shown to provide higher levels of early biomechanical fixations and bone-to-implant contact (BIC) percentage values compared with as-turned, or grit-blast/acid-etched, surfaces.^6,7  Specific to the addition of calcium- and phosphorus-based materials as coatings, some of the interest is due to the inherent presence of these elemental components in the natural bone. On the other hand, the rationale for making physical modifications is to create a rougher surface at the micrometer level that, up to a certain extent, increases bone anchorage and, at the nanometer scale, increases surface energy and improves osseointegration.^5–7  However, it is important to bear in mind that surface topography changes with different manufacturing processes; surface chemistry and physics may also be changed, although unintentionally.⁸ 

Compared with smooth surfaces (S_a < 0.5 μm), an increase in surface texture, as observed for moderately rough surfaces (S_a 1.0–2.0 μm), has shown improved bone biological response and mechanical properties.^9–11 Surface texturing can be achieved by a variety of methods, including acid-etching,^12,13  grit-blasting followed by acid-etching, anodization,^14,15  grit-blasting with alumina,¹⁶  titanium oxide,¹⁷  silica,^18,19  or resorbable biocompatible bioceramics (resorbable blasting media [RBM]).^20–22  The goal of the RBM technique is to produce a moderately rough surface by grit-blasting; the bioceramics reduce the likelihood of decreases in biocompatibility due to particle embedding or particle detachment from the surface (RBM is highly biocompatible).^22,23 

Because of past experiences of adhesive failures between the implant device and bioceramic thick coatings, such as plasma-sprayed hydroxiapatite (20–50 μm), and because of its partial dissolution/resorption in vivo,^24–28  smaller-scale bioceramic coatings have been developed. These coatings may involve thickness in the range of a few micrometers or even nanometers, such as the calcium phosphate (CaP) coatings created by discrete crystalline deposition.²⁹  Early on, the coating dissolution may be influenced in vivo by thickness, micro/nanoscale texturing achievable through any of these processes, and controlled composition. Concerning the residual CaP incorporation by the RBM process, surface topography and chemistry are influenced by blasting media composition; particle size; processing parameters, such as blasting pressure and distance, and subsequent acid-etching treatments.⁴  All these parameters may significantly influence the short- and long-term host-to-implant response.

Changes in surface involving the deposition of hydroxiapatite or other CaP compositions result in the surface modification currently investigated most often, altering both chemistry and topography.⁸  However, the available evidence supporting the potential value of surface chemical alterations alone, such as CaP, in improving bone response is inconclusive. One claimed contributing factor is the sparse assessment of surface topography and/or chemical alterations performed by appropriate analytical tools,⁵  in tandem with in vivo investigations to allow a sound correlation of the interplay between bone response and surface design.⁸  Hence, the present study aims to evaluate the effect of an RBM-treated surface, compared with an alumina-blasted/acid-etched (AB/AE) surface in the torque to interface failure and BIC. Surface characterization was performed by three different methods, including scanning electron microscopy (SEM), optical interferometry (IFM), and X-ray photoelectron spectroscopy (XPS). Our tested null hypothesis was that surface treatment would not affect torque or BIC values.

The implants used in this study were commercially available Ti-6Al-4V screw-type implants 4 mm in diameter and 13 mm long, which were provided by the manufacturer (ADIN Dental Implants Systems Ltd, Afula, Israel). A total of 56 implants was used and divided into 2 groups according to surface treatment: pure titanium surface treated with RBM as the experimental group (n = 24) and an AB/AE group as the control group. The remaining implants were used for surface characterization (4 per group).

The surface characterization was accomplished with three different methods. First, SEM (Philips XL 30, Eindhoven, The Netherlands) was performed at various magnifications under an acceleration voltage of 15 kV to observe the topography of both surfaces.

The second step was to determine the roughness parameters at the micrometer length scale by optical IFM (Phase View 2.5, Palaiseau, France). For each surface, 3 implants were evaluated, recording S_a (arithmetic average high deviation) and S_q (root mean square) at the flat region of the implant cutting edges (3 measurements per implant) using a filter size of 250 μm × 250 μm. After data normality verification, statistical analysis at the 95% level of significance was performed by 1-way analysis of variance (ANOVA).

The third procedure was the surface-specific chemical assessment performed by XPS. The implants were inserted in a vacuum transfer chamber and degassed to 10⁻⁷ torr. The samples were then transferred under vacuum to a Kratos Axis 165 multitechnique XPS spectrometer (Kratos Analytical, Chestnut Ridge, NY). Survey spectra were obtained using a concentric hemispherical analyzer with a 165-mm mean radius operated at a constant pass energy of 160 eV for survey and 80 eV for high-resolution scans. The take-off angle was 90°, and a spot size of 150 μm × 150 μm was used. The implant surfaces were evaluated at various locations.

With the approval of the Ethics Committee for Animal Research at Federal University of Santa Catarina, 12 mongrel dogs were acquired and remained for 2 weeks in the animal facility before the first surgical procedure.

For the surgery, 3 drugs were administered until general anesthesia was achieved by intramuscular injection. The drugs were atropine sulfate (0.044 mg/kg), xilazine chlorate (8 mg/kg), and ketamine chlorate (15 mg/kg). The implantation site was the distal part of the femur (n = 2 per limb); the right limb of each animal provided implants that remained for 4 weeks in vivo, and the left limb provided implants that remained 2 weeks in vivo. For implant placement, the surgical site was shaved with a razor blade, followed by application of antiseptic iodine solution. An incision of ≈5 cm through the skin and periosteum was performed, and the periosteum was elevated for bone exposure. Sequential drills from the manufacturer were used, following the sequence (pilot drill, 2.0 mm, 3.0 mm, and 3.5 mm) under abundant saline irrigation at 1.200 rpm. The implants were placed in an interpolated distribution to minimize bias from different implantation sites and surface type (sites 1 and 2 from proximal to distal) along the distal part of the femur for torque and histomorphometric evaluation. The first implant was placed 1 cm below the joint capsule at the central medial-lateral position of the distal part of the femur. The other devices were placed along the distal direction at distances of 1 cm from each other along the bone. After placement the healing caps were inserted and sutured in layers with vicryl 4-0 (Ethicon Johnson, Miami, Fla) for the periosteum and nylon 4-0 (Ethicon Johnson) for the skin. The animals stayed in the animal care facility and received antibiotics (benzyl penicillin benzatine 20 000 UI/kg) and anti-inflammatory (ketoprofen 1% 1 mL/5 kg) medication to control the pain and infection. Euthanasia was performed after 4 weeks by anesthesia overdose, and the limbs were retrieved by sharp dissection.

According to the initial protocol, half of the specimens of right and left limbs were nondecalcified processed to slides for histomorphologic and histomorphometric (percentage of BIC) evaluation. The other half was subjected to biomechanical testing (torque to interface failure).

At necropsy, the femurs were retrieved by sharp dissection, and surgical blades were used to remove soft tissue. The implants in bone were reduced to blocks and were then immersed in 10% buffered formalin solution for 24 hours. The blocks were then washed in running water for 24 hour and gradually dehydrated in a series of alcohol solutions ranging from 70% to 100% ethanol. After dehydration, the samples were embedded in a methacrylate-based resin (Technovit 9100, Heraeus Kulzer GmbH, Wehrheim, Germany) according to the manufacturer's instructions. The blocks were then cut into slices (∼300 μm thick), aiming the center of the implant along its long axis with a precision diamond saw (Isomet 2000, Buehler Ltd, Lake Bluff, Ill), and glued to acrylic plates with an acrylate-based cement. A 24-hour setting time was allowed before grinding and polishing. The sections were then reduced to a final thickness of ∼30 μm by means of a series of silicon carbide abrasive papers (400, 600, 800, 1200, and 2400) (Buehler Ltd) in a grinding/polishing machine (Metaserv 3000, Buehler Ltd) under water irrigation.³⁰  The sections were then stained with toluidine blue and referred to optical microscopy for histomorphologic evaluation.

The BIC was determined at 50×–200× magnification (Leica DM2500M, Leica Microsystems GmbH, Wetzlar, Germany) by means of a computer software program (Leica Application Suite, Leica Microsystems GmbH). The regions of BIC along the implant perimeter were subtracted from the total implant perimeter, and calculations were performed to determine the BIC.

Torque testing required the femur to be adapted to an electronic torque machine equipped with a 200 Ncm load cell (Test Resources, Minneapolis, Minn).

Custom machined tooling was adapted to the internal connection of each implant, and the bone block was carefully positioned to avoid specimen misalignment during testing. The implants were torqued in a counterclockwise direction at a rate of ∼0.196 radians/min, and torque versus displacement curve was recorded for each specimen.

Preliminary statistical analyses showed no effect of implant site (ie, there were no consistent effects of implant position along the femur) on all measurements. Therefore, the site was not considered further in the analysis. Further statistical evaluation was performed by ANOVA using implant surface and time in vivo as independent variables and torque and BIC as dependent variables. Statistical significance was indicated by P levels less than 5%, and post hoc testing used the Fisher least significant difference test.

The electron micrographs and representative 250 μm × 250 μm IFM three-dimensional reconstructions of the control and experimental implant surfaces are presented in Figures 1 and 2, respectively. Their respective micrometer length scale S_a and S_q values are presented in Figure 3. The surface texture observed at intermediate and high magnification levels did not show evidence of particle embedding in either surface (Figure 1). The micrometer length scale IFM measurements showed higher S_a and S_q values for the experimental surface than for the control surface (Figure 3). The survey XPS spectra detected the presence of carbon, oxygen, titanium, nitrogen, aluminum, and vanadium for the control surface, and those elements, plus calcium and phosphorus, for the experimental surface (Table 1). Statistical analysis indicated no significant differences between surface types in regards to torque to interface failure and BIC (P > .71 and P > .98, respectively). Although statistical analysis showed no difference between the times in vivo for torque (P > .71), there was statistical difference between BIC and time in vivo (P < .04) (Table 2).

Histomorphologic evaluation showed that, irrespective of implant surface, bone to implant response at cortical regions was characterized by woven bone formation and interfacial remodeling at 2 weeks (Figure 4a), followed by its initial replacement by lamellar bone and higher bone organization levels at 4 weeks (Figure 4b). At trabecular regions, a likewise bone to implant response was observed for both surfaces, where woven bone was observed in the vicinity of the implant surfaces at 2 weeks (Figure 5a), and the initial formation of lamellar bone surrounding primary osteonic structures in a higher degree of bone organization was observed at 4 weeks (Figure 5b).

The alleged chemical effects of CaP incorporations to an implant surface have long drawn attention as a positive modulator of bone healing. However, whether the subsequent topographical changes contribute alone or in combination with the chemical alterations is still unclear.⁸  In the present study, both experimental and control surfaces presented average roughness in the moderately rough range (S_a 1 – 2 μm).³¹  Although the roughness of the experimental surface was in the range of what has been shown to result in the strongest bone response (S_a of approximately 1.5 μm)^8,32–34  and was significantly higher than the control surface (S_a of approximately 0.85 μm), torque values were not significantly different at the 2- or 4-week evaluation points. Even though a rougher profile was expected for the control group because the alumina blasting medium was harder than the one used on the experimental group, several other parameters, such as pressure, distance, particle size, and subsequent acid etching, may result in a wide range of surface roughness measurements for control implants. As no information regarding the manufacturing process of both surfaces was provided by the manufacturer, further comments on how such surface characteristics were achieved would be speculative in nature. However, from a surface chemistry standpoint, both surfaces presented elemental composition similar to the composition presented in previous studies.^21,22,35 

Time in vivo significantly increased BIC values proportionally for both surfaces. The temporal increase for BIC values only and not for torque suggests that static histomorphometric parameters, such as BIC, present weak evidence of a better implant bone interface or biomechanical fixation. It has been previously emphasized that less BIC in a higher-magnitude mechanical property supporting bone may be preferred over a device with more BIC presenting bone with lower magnitude mechanical properties.⁴  Hence, BIC is a measurable parameter with important but limited significance when considering its implication in biomechanical testing such as torque.

Our imaging results showed similar surface-texture morphology for both surfaces at intermediate and high magnification levels in the SEM and IFM reconstructions. Although the micrometer length scale IFM measurements showed higher S_a and S_q values for the experimental surface than for the control, and no differences in torque and BIC were observed, an investigation where only nanometer roughness was compared with a micrometer scale control showed stronger bone response for the former surface, and in that study chemistry was ruled out as the explanatory factor.³⁶  When a nano-capped surface (approximately 20 nm) was compared to a polished titanium surface in a different experiment by the same group, stronger bone response was observed for the nano-surface.³⁷  However, whether surface chemistry modification, the presence of nano-roughness, or the combination of both factors fostered bone to implant response (and to what extent) is unknown.

The histomorphologic sections depicted bone in close contact with the implant surface in both trabecular and cortical bone for both experimental and control surfaces, indicating their biocompatible and osseoconductive potential. The healing pathway observed in the present study was similar to what has been described for screw-root form implants, where woven bone was observed around both surfaces at 2 weeks, followed by its initial replacement by lamellar bone at 4 weeks.^9,22,38 

Based on the results, surface treatment did not affect torque or BIC values at early implantation times, leading to the acceptance of our postulated null hypothesis. Both surfaces presented equivalent histomorphologic, histomorphometric, and mechanical testing outcomes despite significant differences in surface roughness and varied chemical composition. The fact that both surfaces presented different roughness and chemistry profiles is a limitation of the present study; thus, controlled experimental designs where both roughness and chemistry are varied in a controlled fashion, along with an appropriate systematic surface classification as proposed by Dohan Ehrenfest et al, would be desirable.⁵  The interplay between roughness and/or chemical alterations and their individual contribution to osseointegration is not yet understood.

AB/AE

alumina-blasted/acid-etched

ANOVA

analysis of variance

BIC

bone-to-implant contact

CaP

calcium phosphate

IFM

interferometry

RBM

resorbable blasting media

SEM

scanning electron microscopy

XPS

X-ray photoelectron spectroscopy