Abstract
The knowledge of how nanostructures might affect early bone healing and osseointegration is limited. The aim of this study was to investigate if nanometer thick coatings of hydroxyapatite nanocrystals applied on a moderately rough surface might enhance early bone healing on screw-shaped dental implants and to evaluate if the thickness of the coat influences healing. Sandblasted and acid etched titanium implants coated with two different thicknesses of hydroxyapatite (test implants) and sandblasted and acid etched titanium implants (control implants), were inserted in rabbit tibia. After a healing time of 2, 4 and 9 weeks, a removal torque analysis and a histological evaluation were performed. The results from the removal torque analysis showed a tendency for higher values for the double coated hydroxyapatite after 4 weeks and for both the coated surfaces after 9 weeks of healing. The histological evaluations indicated slightly more new bone formation with the coated implants compared with the control; the differences did not reach statistical significance. The present study could not support the importance of nanometer thick coatings of hydroxyapatite nanocrystals in early bone healing, at least not when applied on a blasted and etched surface and placed in a cortical bone.
Dental implants have been used for almost four decades. Several factors have been identified as important for achieving osseointegration. Surface topography is one of these factors . A mean surface roughness of approximately 1.5 μm has been shown to initiate a stronger bone response compared with smoother and rougher surfaces on the micrometer level . Dental implant treatment is a reliable method with good clinical results , but intense research is underway to find an optimal implant surface to achieve faster implant integration and optimal healing in implant sites with poor bone quality and quantity. Recently, the focus has been on bioactive materials and the presence of nanostructures on the implant surface. Knowledge about the importance of nanostructures in early bone healing and osseointegration is limited. Previous experimental studies on a micrometer level have shown that plasma-sprayed hydroxyapatite (HA) coated implants have a stronger initial bone response compared with conventional titanium implants . The plasma spraying method creates coats with a thickness of 50–200 μm, but with poor adhesion between the coat and the underlying metallic implant so that long-term clinical results have been less favourable and associated with failure .
It has not been clarified whether the positive bone response was due to the proposed bioactivity of HA, to possible alterations in surface topography or to a greater press fit of the thicker HA-coated implants when screwed home in the same size defects as the controls. Thinner HA coats, in the nanometer range, have been developed to improve the coat and thus minimize potential problems with coat loosening. Several in vitro studies have shown an increased cellular response to different nanostructures . An in vivo study in a rabbit model showed increased bone formation after 4 weeks of healing on electropolished cylindrical titanium implants coated with nanometer sized particles of HA compared with uncoated controls . A study on screw-shaped implants placed in rabbit tibia indicated that surfaces with nanostructures have more newly formed bone on the apical part of the implant compared with a control surface . In a human study, dual acid etched (DAE) implants with nanometer-scale CaP crystals were placed in the posterior maxilla and DAE implants were used as controls. Results from this study showed significantly higher bone to implant contact (BIC) for the test implant compared with the control after 4 and 8 weeks of healing . Blasted and etched screw-shaped implants have shown good clinical results . Whether nanometer CaP promotes bone healing when applied on surfaces other than a double etched surface has not been investigated.
The aim of this study was to investigate if nanometer thick coatings of HA nanocrystals could enhance early bone healing on screw-shaped, blasted and etched dental implants and to evaluate if the thickness of the coat could influence bone healing.
Materials and methods
The implants used were screw-shaped, sandblasted and acid etched titanium implants coated with two different thicknesses of HA nanocrystals. HA nanocrystals with a crystal size of about 5 nm, were deposited onto the implants using a technique described by K jellin and Andersson . The thickness of the coatings was estimated using SEM. The single coat had a thickness of about 20 nm and the double coat about 40 nm. Sandblasted and acid etched titanium implants were used as controls. Powder X-ray diffraction (XRD) was used to determine the presence of crystalline HA structures. XRD was performed using a Bruker XRD D8 Advance (Bruker AXS, Karlsruhe, Germany) and monochromatic Cu radiation. A SEM analysis was performed on the implant surfaces using a Leo Ultra 55 FEG high resolution scanning electron microscope (Carl Zeiss SMT Inc., North America), operating at an acceleration voltage of 5–10 kV. The magnification used was 60,000× and the micrographs were recorded at randomly chosen areas of the implants using secondary electrons. Surface roughness was examined using a white light interferometer (MicroXAM™, Phaseshift, Arizona, USA), which represent a technique that, according to W ennerberg and A lbrektsson , is best suited to evaluate threaded implant surfaces. A 50× objective and a zoom factor of 0.62 was used. The size of the area measured was 264 μm × 200 μm and the vertical measuring range was 100 μm. The maximal resolution of the technique is 0.3 μm horizontally and 0.05 nm vertically. In order to be able to describe the surface topography, the roughness, the waviness and shape must be taken into consideration. The standard filter used to separate micrometer roughness from waviness and shape is a high-pass Gaussian filter and a filter size of 50 μm × 50 μm is optimal for threaded implants . To evaluate the height deviation at the nanometer level, a filter size of 1 μm × 1 μm was used, as suggested by S vanborg et al. . Surfascan software (Somicronic Instrument, Lyon, France) was used for filtration and evaluation. This equipment provides numerical descriptions and images of the surface topography. 3 valleys on each implant were measured and evaluated.
For numerical description of the surface topography, four parameters were used: Sa, the arithmetic mean of the roughness area from the mean plane; Sds, density of summits (i.e. number of peaks per area unit); Sdr, the ratio between the developed surface area and a flat reference area; Sci, core fluid retention index.
The parameters used represent one amplitude (Sa), one spatial (Sds), one hybrid (Sdr) and one functional (Sci) value. The functional parameter, core fluid retention index (Sci), is related to the bone biological ranking, based on earlier studies on a micrometer level of resolution. A low value may be related to a positive biological outcome for bone anchored implants . Mathematical formulas for the parameters can be found in the literature . X-ray photoelectron spectroscopy (XPS) was used for characterization of the surface chemical compositions. XPS survey spectra were obtained using a PHI 5000C ESCA System (Perkin–Elmer Wellesley, USA). An α excitation source was used at 250 W with an operating angle of 45°.
A rabbit model was chosen in this study; it has been used in several studies to evaluate different implant surfaces . 27 adult female New Zealand rabbits were divided into 3 groups, 9 animals in each. Group 1 had a healing time of 2 weeks, group 2 a healing time of 4 weeks and group 3 a healing time of 9 weeks. Before surgery, the animals were anaesthetized with an intramuscular injection of fentanyl 0.3 mg/ml and fluanisone 10 mg/ml (Hypnorm Vet, Janssen, Pharmaucetica, Beerse, Belgium) at a dose of 0.5 ml per kg body weight and an intraperitoneal injection of diazepam (Stesolid Novum, Alpharma, Denmark) at a dose of 2.5 mg per animal. 1 ml of lidocaine (Xylocain, Astra, Sweden) was administered subcutaneously at the surgical site as an analgesic and the operation was performed under aseptic conditions. Two implants were inserted into each left and right tibia in the rabbits. The implant sites were prepared using a drill, under copious saline irrigation. The implants were inserted in the bone under saline irrigation. A single dose of prophylactic antibiotic sulfadoxin 200 mg/ml and trimethoprim 40 mg/ml (Borgal, Intervet, Boxmeer, Netherlands) at a dose of 0.5 ml/kg and 0.5 ml analgesic buprenorphine 0.3 mg/ml (Temgesic, Schering-Plough, Belgium) were administered immediately after surgery. Immediately after surgery the rabbits were kept in separate cages to control wound healing. They had free access to tap water and were fed with pellets and hay. After initial healing, the rabbits were allowed to run freely in a specially designed room. The three groups of animals were killed after 2, 4 and 9 weeks of healing with a 10 ml overdose of pentobarbital 60 mg/ml (Pentobarbitalnatrium, Apoteksbolaget, Sweden).
Removal torque analysis was performed on each implant with an electrically controlled removal torque unit. The implants were subjected to the torque (Ncm), only to interrupt osseointegration but were not screwed out from the bone any further.
After torque analysis, each implant was removed in a block with the surrounding bone and fixed in 4% neutral buffered formaldehyde. The samples were dehydrated in alcohol solution and embedded in light curing resin (Technovit 7200 VLC, Kultzer & Co, Germany). The cutting and grinding were performed as described by D onath . The final sections were about 20 μm thick and stained with toluidine-blue. Histological evaluations were performed with a light microscope and image analysis software (Image analysis 2000, Sweden). The evaluation included measurement of the amount of new bone formation and bone area. New bone formation was calculated from the total amount of bone on each side of the implant, minus the amount of old bone. The measurements were performed at a magnification of 4× ( Fig. 1 ). The bone area was evaluated in each thread on each implant using, 10× magnification and presented as the mean value of all threads on the entire implant. An additional numerical calculation was made to try to normalize for any individual difference between the animals; the torque value was divided by the total bone length for each implant. Such values would disregard any differences between the amounts of bone present at the individual implant sites.
Statistical analysis was performed using SPSS (statistical package for the social studies). The interferometer surface characterization was analysed using one-way ANOVA. Differences were considered significant at p ≤ 0.05. Torque results, bone area and new bone were evaluated using Kruskal–Wallis and analysis of variance (ANOVA). The assumption of the ANOVA was tested with Levene’s test of equal variances, and if needed the data were log 10 transformed. Differences were considered significant at p ≤ 0.05.
Results
The postoperative period was uncomplicated for most animals, except for five rabbits that experienced tibia fracture, two in each group with 2 and 9 weeks of healing time and one in the group with 4 weeks of healing time. These rabbits had to be killed before the planned healing time and were not included in the results. No sign of infection or other deviation from normal was observed at the time of implant retrieval.
Implant surface characterization
The SEM images show that all the surfaces had nanostructures of various sizes ( Fig. 2 ). The results obtained in the interferometry characterization are presented in Table 1 and Fig. 3 shows images of the surface topography at micro and nanometer level. The mean average height deviation (Sa), on the micrometer level, showed that the double coated implant had the smoothest (0.77 μm) while the control had the roughest (1.08 μm) surface, although the difference was not significant ( p = 0.102). Both coated implants and the control had a similar Sa value on the nanometer level ( p = 0.256). All the implants had similar peaks per area unit (Sds) and the surface enlargement was greatest for the control implant (Sdr 142.5) on the micrometer, but there were no significant differences ( p = 0.61–0.99). The core volume index was highest for the double coated HA implant, which differed significantly ( p = 0.03) from the control implant on the micrometer level. XPS analysis showed the presence of calcium and phosphorus on the surface of both the coated implants while the control did not. XRD demonstrated the presence of crystalline HA ( Fig. 4 ).
| Gausfilter 50 μm × 50 μm | Gausfilter 1 μm × 1 μm | |||||||
|---|---|---|---|---|---|---|---|---|
| Sa (μm) | Sds (/mm 2 ) | Sdr (%) | Sci | Sa (nm) | Sds (/mm 2 ) | Sdr (%) | Sci | |
| Control | 1.08 (0.41) | 1 184807 | 142.5 | 1.21 | 114 (11.1) | 2 055650 | 74.9 | 0.95 |
| HA single coat | 0.91 (0.20) | 1 222269 | 117.4 | 1.23 | 111 (7.9) | 2 068575 | 73.1 | 0.86 |
| HA double coat | 0.77 (0.19) | 1 282006 | 119.3 | 1.43 | 118 (6.2) | 2 176006 | 84.1 | 0.79 |
Removal torque analysis
The removal torque value for the double coated surfaces was higher than both the single coated and the control after 2 and 4 weeks. The torque value after 9 weeks was higher for both coated surfaces compared with the control ( Fig. 5 ), but the numerical differences did not reach statistical significance ( p = 0.30–0.67). Two of the implants, one in the single HA group and one in the double coated group, reached the lateral wall and the second cortical layer and therefore ( Fig. 6 ) rendered a very high torque value, those implants were excluded from the results. This position could not be detected during surgery and it is therefore good to be able to verify the torque results with histological results if such outliers occur.
