Peri-implant bone response to orthodontic loading: Part 2. Implant surface geometry and its effect on regional bone remodeling

ORIGINAL ARTICLE

Peri-implant bone response to orthodontic loading: Part 2. Implant surface geometry and its effect on regional bone remodeling Rodrigo Oyonarte,a Robert M. Pilliar,b Douglas Deporter,c and Donald G. Woodsided Toronto, Ontario Canada Introduction: Bone response to orthodontic loading was compared around 2 different types of osseointegrated implants (porous surfaced and machined threaded) to determine the effect of implant surface geometry on regional bone remodeling. Methods: Five beagles each received 3 implants of each design in contralateral mandibular extraction sites. After a 6-week initial healing period, abutments were placed, and, 1 week later, the 2 mesial implants on each side were orthodontically loaded for 22 weeks. All implants remained osseointegrated throughout orthodontic loading except for 1 threaded implant that loosened. Back-scattered scanning electron microscopy and fluorochrome bone labeling techniques were used to compare responses around the 2 types of implants. Results: The loaded, porous-surfaced implants had significantly higher marginal bone levels and greater bone-to-implant contact than did the machinedthreaded implants. Conclusions: Significant differences in peri-implant bone remodeling and bone formation in response to controlled orthodontic loading were observed for the 2 implant designs. Short, poroussurfaced implants might be more effective for orthodontic applications than machine-threaded implants. (Am J Orthod Dentofacial Orthop 2005;128:182-9)

O

sseointegrated dental implants as orthodontic anchorage units might offer advantages over more traditional treatment approaches that rely on teeth or extraoral devices for anchorage.1-3 Several authors have analyzed the morphological reactions of peri-implant bone to continuous horizontal loading of bone-interfacing implants by using threaded titanium implants with either a machined or a textured surface finish.1,4-11 Most of these studies have shown that maintenance of osseointegration during orthodontic loading of implants appears to be associated with increased periimplant bone density (in zones subjected to compressive stress) and bone remodeling activity.1,4-7,11 However, the optimal implant design for orthodontic applications remains to be determined. Orthodontically induced marginal bone remodeling From the University of Toronto. a Former Resident, Orthodontics Program, Faculty of Dentistry. b Professor of biomaterials, Faculty of Dentistry. c Professor of periodontology, Faculty of Dentistry and Institute of Biomaterials and Biomedical Engineering. d Professor emeritus, Faculty of Dentistry. Partially funded by the Canadian Institutes for Health Research and the American Association of Orthodontists Foundation (Biomedical Research Award 2001). Reprint requests to: Dr R. M. Pilliar, University of Toronto, Faculty of Dentistry, 124 Edward St, Toronto, Ontario, Canada M5G 1G6; e-mail, [email protected] Submitted, September 2003; revised and accepted, February 2004. 0889-5406/$30.00 Copyright © 2005 by the American Association of Orthodontists. doi:10.1016/j.ajodo.2004.02.024

182

depends on, among other factors, the magnitude of forces applied to an implant and the deformation of the loaded bone,4 as well as the nature of the bone itself.6 Wehrbein et al6 found that small forces of approximately 1 N did not result in added bone formation (compared with unloaded controls) around orthodontically loaded, threaded implants, although a force of 2 N applied to implants in the palatal bones of dogs did, but only on bone subjected to compressive forces. Melsen and Lang11 reported statistically significant differences in bone remodeling in monkey mandibles around threaded implants that were orthodontically loaded versus others that were not. These studies indicate that, within a typical orthodontic force range, the higher the force applied to threaded dental implants, the greater the subperiosteal bone formation on the compression side of the implants and, conversely, the greater the resorption of bone on the tension side.6,9,12 Although several studies have reported on the response of periimplant bone to orthodontic forces, the effects of variations in implant surface design on this load-related remodeling have not been reported. In an earlier article,13 we compared the morphological effects of orthodontic loading on peri-implant bone using sintered, porous-surfaced (PS) and machined-threaded (MTh) implants. These morphometric results showed that orthodontically loaded, PS implants retained marginal bone to a significantly greater height and developed

American Journal of Orthodontics and Dentofacial Orthopedics Volume 128, Number 2

greater bone-to-implant contact than similarly loaded MTh implants. In this article, we report the nature of bone remodeling around these 2 implant designs as determined by fluorochrome labeling techniques and back-scattered scanning electron microscopy. MATERIAL AND METHODS

The experimental protocol was approved by the University of Toronto Animal Care Committee, and the materials and methods were reported in detail previously.13 Briefly, 5 adult beagles had several mandibular teeth extracted. After a healing period, 6 custom-made implants (3 each of 2 types) were placed into the extraction sites. The implants on 1 side of the mandible were tapered (approximately 5° taper angle), press fit, PS Ti-6Al-4V; implants on the other side were conventional cylindrical specimens made of commercially pure titanium and MTh. The bone-interface surfaces of the PS implants were made by sintering Ti-6Al-4V alloy powders to a machined implant core.14 Both types were 5 mm long, including a 1-mm, smooth, machined coronal collar portion (ie, no sintered layer or threads). The 2 more mesial implants on each side of the mandible constituted the loaded test implants, and the distal-most implant on each side was left as a nonloaded control. After a healing period, the implants were uncovered, and the test implants were loaded with custom-made abutments and 100-g nickel-titanium springs (Sentalloy springs; GAC, Islandia, NY) for the first 5 weeks and thereafter with 300-g springs for 17 weeks. Control implants received only standard, temporary healing abutments, which kept the implants exposed to the oral cavity. After the 22-week experimental period, the animals were killed, and specimens were collected for histological assessment. Fluorochrome bone labeling was used to assess rates of bone remodeling. Each dog was given 2 doses of calcein green (intravenously, 10 mg/kg), the first at 17 days after orthodontic force application and the second 10 days later. Doses of oxytetracycline (intravenously, 20 mg/kg) were given to each dog at 14 days, and again 4 days before they were killed (intravenously, 12.5 mg/kg). All retrieved specimens were fixed by immersion in 70% ethanol, dehydrated, then embedded in Osteobed resin (Polysciences, Warrington, Pa) as described previously.13 Two ground sections per implant and surrounding tissues (approximate thickness, 40 ␮m) were prepared in the sagittal (mesiodistal) plane. One of these 2 sections was stained with toluidine blue and van Gieson’s picrofuchsin and then platinum-coated for back-scattered, scanning electron microscopic exami-

Oyonarte et al 183

nation; the second section was left unstained to allow examination of the fluorescent labeling. The stained sections were first examined by using back-scattered scanning electron microscopy (BSEM) (20kV accelerating voltage; Hitachi S-2500, Tokyo, Japan). Micrographs were collected at 30x magnification. On average, 7 micrographs were obtained for each implant, and these micrographs were spliced by using Sigmascan Pro 5.0 software (SPSS Science, Chicago, Ill). The spliced BSEM images were then quantitatively assessed by using computer-assisted image analysis to calculate percent bone area (%BA) between the implant surface and up to 1 mm from the surface region examined on the mesial and distal aspects of all implants (Fig 1). Percent BA measurements were made by using Optimas 6.5 software at 2x digital magnification (Media Cybernetics, Silver Spring, Md). In performing these area measurements, threshold values of grey levels in the digitized back-scattered electron micrographs were assigned for bone and soft tissues, and regions of interest were determined on the mesial and distal aspects of each implant in 3 different zones relative to the implant surface (Fig 1). One zone extended up to 100 ␮m from the implant surface (0/100 ␮m measurements). The other zones were located between 100 and 500 ␮m (100/500 ␮m measurements) and between 500 and 1000 ␮m from the implant surface (500/1000 ␮m measurements). A fourth zone was defined as the total region of up to 1000 ␮m from the implant surface (ie, including all 3 defined zones, [0/1000 ␮m measurements]). These measurements were performed for both the whole implant length and a standardized distance of 1643 ␮m from the apical border of the machined collar region (the coronal subregion). The latter measurement corresponds to the average distance encompassing the length of 3 coronal threads of the threaded implant design. Using this defined zone of measurement allowed comparison of percent area of bone for MTh versus PS implants in their respective coronal subregions only and permitted a clear differentiation of the effects of loading and surface geometry on peri-implant %BA— both close to and far from the implant for the total length of the implants and the defined coronal subregion. The unstained tissue sections were examined under fluorescence microscopy by using 3.5x and 16x objectives (Carl Zeiss III RS microscope, Oberkochen, Germany) to estimate bone mineral appositional rates (Fig 2). Digital pictures were obtained and analyzed by using Sigmascan Pro 5.0 and Optimas 6.5 software. Bone mineral appositional rate (MAR) was determined by measuring the distance between the centers of 2

184 Oyonarte et al

Fig 1. Histomorphometric measurements from BSEM micrographs. % bone area measurements. A, 100 ␮m; B, 500 ␮m; C, 1000 ␮m measurements.

fluorescent bands (labels) and dividing by the time between labeling (mm/day). The mean MAR was determined from all the active bone-forming surfaces.17 MAR was calculated by using the digital photos of 16x magnified images. Measurements were made at the coronal, middle, and apical thirds of each implant within about 500 ␮m from the implant surface. In 5 instances, no fluorescent-labeled bone sites were found at a given peri-implant region (coronal, middle, or apical third), and, in these situations, the average of the other 2 sites obtained for the same implant surface was assigned to this region. To quantify the overall remodeling activity of peri-implant bone within 500 ␮m from the implant surface, we made single and double labeled-surface (LS) measurements using spliced images obtained with a 3.5x objective lens (Fig 3). The cumulative lengths of clearly identifiable single and double oxytetracycline fluorescent labels were measured between the alveolar bone crest and the apex of the implants by using Sigmascan Pro 5.0 at 4x digital magnification. Statistical analyses

The study had a split-mouth configuration, and each mandible provided a pair of experimentally loaded implants of each type and a single nonloaded implant as the control. Each loaded implant had 1 surface with adjacent bone in compression (the region between the 2 orthodontically loaded implants) and the opposite surface with bone in tension. This allowed analysis of differences in bone subjected to compression versus tension both within implant types (either PS or MTh) and between implant types (ie, porous-surfaced implants were matched to contralateral threaded, orth-

American Journal of Orthodontics and Dentofacial Orthopedics August 2005

Fig 2. Fluorescence microscopy. A, 3.5x objective; B, 16x objective (same sample).

Fig 3. Spliced oxytetracycline-labeled images from paired PS and MTh, orthodontically loaded, peri-implant bone sites. A, PS implant and surrounding bone; B, MTh implant and surrounding bone.

odontically loaded implants). Similarly, control implant surfaces could be compared within implant type (control vs compression vs tension zones) as well as between implant types. Percent BA, MAR, and LS measurements were assessed with Student paired t tests and 1-way analysis of variance (ANOVA) statistical test. When significant differences were detected between parameters with the ANOVA test, Bonferroni multiple comparison tests were used as a further assessment. Statistical tests were 2-tailed and interpreted at the 5% significance level.

Oyonarte et al 185

American Journal of Orthodontics and Dentofacial Orthopedics Volume 128, Number 2

Table I.

Average percent bone area measurements between implant types Total compression

%BA 0/100 Av PS ⫾ SE Av Mth ⫾ SE Diff ⫾ SE P value %BA 100/500 Av PS ⫾ SE Av Mth ⫾ SE Diff ⫾ SE P value %BA 500/1000 Av PS ⫾ SE Av Mth ⫾ SE Diff ⫾ SE P value %BA 0/1000 Av PS ⫾ SE Av Mth ⫾ SE Diff ⫾ SE P value

58.1 ⫾ 4.57 68.16 ⫾ 3.61 ⫺10.06 ⫾ 6.5 .165

Coronal compression 71.48 ⫾ 5.05 55.75 ⫾ 7.7 15.73 ⫾ 10.2 .167

Total tension

Coronal tension

Total control

Coronal control

56.45 ⫾ 6.39 66.68 ⫾ 3.16 ⫺10.23 ⫾ 6.67 .169

70.36 ⫾ 4.9 50.25 ⫾ 8.8 20.11 ⫾ 9.69 .077

55.98 ⫾ 3.18 67.22 ⫾ 2.38 ⫺11.24 ⫾ 4.31 .028

71.88 ⫾ 2.16 67.23 ⫾ 3.7 4.65 ⫾ 4.5 .328

77.58 ⫾ 4.83 71.49 ⫾ 4.69 6.09 ⫾ 6.68 .392

83.39 ⫾ 6.53 54.31 ⫾ 9.65 29.08 ⫾ 11.18 .035

70.25 ⫾ 6.41 67.21 ⫾ 4.68 3.04 ⫾ 5.93 .624

88.04 ⫾ 2.54 47.94 ⫾ 10.3 40.1 ⫾ 8.85 .003

72.92 ⫾ 4.07 59.32 ⫾ 2.47 13.6 ⫾ 3.07 .002

85.09 ⫾ 1.83 56.98 ⫾ 5.21 28.11 ⫾ 4.68 .000

63.24 ⫾ 5.4 62.09 ⫾ 6.91 1.15 ⫾ 9.08 .903

72.54 ⫾ 3.22 49.28 ⫾ 10.04 23.26 ⫾ 11.87 .091

53.6 ⫾ 8.76 55.14 ⫾ 6.04 ⫺1.54 ⫾ 7.49 .843

75.55 ⫾ 5.98 45.58 ⫾ 10.83 29.98 ⫾ 9.69 .017

57.82 ⫾ 6.17 45.99 ⫾ 4.09 11.83 ⫾ 5.41 .057

74.33 ⫾ 4.0 51.78 ⫾ 6.38 22.55 ⫾ 8.12 .021

67.18 ⫾ 4.45 66.65 ⫾ 5.21 0.52 ⫾ 7.49 .946

76.11 ⫾ 3.95 52.4 ⫾ 8.77 23.71⫾10.13 .052

60.03 ⫾ 6.73 61.79 ⫾ 4.46 ⫺1.76 ⫾ 5.7 .766

78.83 ⫾ 3.7 47.38 ⫾ 9.4 31.45 ⫾ 8.0 .006

62.72 ⫾ 4.12 55.11 ⫾ 2.8 7.6 ⫾ 2.79 .023

77.58 ⫾ 1.7 56.84 ⫾ 4.79 20.74 ⫾ 5.07 .003

ANOVA: 100 ␮m, P ⫽ .005; 500 ␮m, P ⫽ .003; 1000 ␮m, P ⫽ .082; 0/1000 ␮m, P ⫽ .004.

RESULTS

All implants remained osseointegrated after the 22-week orthodontic loading period with the exception of 1 orthodontically loaded MTh implant that loosened after 16 weeks of continuous horizontal loading. The neighboring, previously orthodontically loaded MTh implant thereafter was excluded from analysis. This reduced the number of experimental pairs from 10 to 8, and these provided 8 compression-loaded and 8 tension-loaded peri-implant bone regions for each of the 2 implant designs. The original 10 control implants were, however, all available for assessment, because all remained integrated, allowing assessment of their mesial and distal aspects. Abutments were lost with some of the control implants at various times during the experiment, necessitating their replacement. This indicated some functional loading of the implants during the study. Both oxytetracycline labels could be clearly identified in the analyzed, unstained sections. However, the calcein green labels could not be clearly identified, so were not useful in determining mineral appositional rates. Percent BA measurements between implant types were used to quantify the amount and distribution of peri-implant bone and were expressed as percent area of mineralized tissue surrounding each implant within a 1-mm region from the implant surface (0/1000 ␮m measurements) and within zones and subregions previously described. The remaining space is filled with either nonmineralized tissue or portions of the implant

surface region (sintered particles or thread regions in the 0/100 ␮m zone), or it is simply empty space after section preparation. To test for overall differences in the percent area of peri-implant bone (%BA) between implant types for the different loading conditions (compression or tension forces due to orthodontic loading), and next to the mesial and distal aspects of the control implants, an ANOVA test was performed on mean differences in %BA between implant types. This showed statistically significant differences in %BA (P ⬍ .005), with PS implants having greater %BA (for all conditions and next to all surfaces). Further analysis applying a Bonferroni test indicated that these significant differences between implant types were found primarily next to the coronal implant subregions (P ⬍ .04). ANOVA and paired t tests were also used to analyze %BA in the 3 defined zones surrounding the coronal subregion of the implants (Table I). Significant differences were found between implant types for bone subjected to both tension and compression (P ⬍ .01). In the coronal subregion subjected to tension, significantly greater %BA (paired t test) was found for the PS implants in the 100/500 ␮m, 500/1000 ␮m, and 0/1000 ␮m zones. Significantly greater %BA values were also found for the PS implants in the coronal subregion subjected to compression in the 100/500 ␮m zone. All coronal zones of the control PS implants had greater %BA values compared with the MTh implants. In contrast to the coronal %BA measurements, no

186 Oyonarte et al

Table II.

American Journal of Orthodontics and Dentofacial Orthopedics August 2005

Average percent bone area between total and coronal implant regions PS implants

0/100 ␮m Total ⫾ SE Coronal ⫾ SE Mean diff ⫾ SE P value 100/500 ␮m Total ⫾ SE Coronal ⫾ SE Mean diff ⫾ SE P value 500/1000 ␮m Total ⫾ SE Coronal ⫾ SE Mean diff ⫾ SE P value 0/1000 ␮m Total ⫾ SE Coronal ⫾ SE Mean diff ⫾ SE P value

MTh implants

Compression (n ⫽ 10)

Tension (n ⫽ 10)

Control (n ⫽ 10)

56.93 ⫾ 3.93 70.01 ⫾ 4.36 ⫺13.08 ⫾ 3.81 .007

57.02 ⫾ 5.26 68.3 ⫾ 4.55 ⫺11.28 ⫾ 5.52 .072

55.98 ⫾ 3.18 71.88 ⫾ 2.16 ⫺15.9 ⫾ 3.92 .003

68.16 ⫾ 3.61 55.75 ⫾ 7.7 12.41 ⫾ 6.04 .079

66.68 ⫾ 3.16 50.25 ⫾ 8.79 16.43 ⫾ 6.29 .035

67.22 ⫾ 2.38 67.23 ⫾ 3.7 ⫺0.01 ⫾ 4.06 .998

74.73 ⫾ 4.74 80.59 ⫾ 5.56 ⫺5.86 ⫾ 4.74 .248

70.83 ⫾ 5.19 86.15 ⫾ 2.79 ⫺15.32 ⫾ 5.03 .014

72.92 ⫾ 4.07 85.09 ⫾ 1.83 ⫺12.17 ⫾ 4.86 .034

71.49 ⫾ 4.69 54.31 ⫾ 9.65 17.18 ⫾ 6.48 .033

67.22 ⫾ 4.68 47.94 ⫾ 10.3 19.28 ⫾ 6.85 .026

59.32 ⫾ 2.47 56.98 ⫾ 5.21 2.34 ⫾ 5.95 .703

61.94 ⫾ 4.41 72.72 ⫾ 2.74 ⫺10.78 ⫾ 6.26 .119

54.7 ⫾ 6.96 69.64 ⫾ 6.17 ⫺14.94 ⫾ 7.37 .073

57.82 ⫾ 6.17 74.33 ⫾ 4.0 ⫺16.51 ⫾ 5.29 .012

62.09 ⫾ 6.91 49.28 ⫾ 10.04 12.81 ⫾ 3.56 .009

55.14 ⫾ 6.04 45.58 ⫾ 10.83 9.56 ⫾ 7.64 .251

45.99 ⫾ 4.09 51.78 ⫾ 6.38 ⫺5.79 ⫾ 6.09 .367

65.36 ⫾ 3.93 74.91 ⫾ 3.4 ⫺9.54 ⫾ 4.43 .06

60.83 ⫾ 5.37 75.14 ⫾ 3.99 ⫺14.3 ⫾ 5.57 .03

62.72 ⫾ 4.12 77.58 ⫾ 1.7 ⫺14.86 ⫾ 4.18 .006

66.65 ⫾ 5.21 52.4 ⫾ 8.77 14.26 ⫾ 4.7 .019

61.79 ⫾ 4.46 47.38 ⫾ 9.4 14.4 ⫾ 6.77 .071

55.11 ⫾ 2.8 56.84 ⫾ 4.79 ⫺1.73 ⫾ 5.36 .754

significant differences were found in %BA for the orthodontically loaded implants of the 2 designs measured over the total implant lengths for either the tension or compression bone regions (Table I). For the control implants, however, the %BA (total length measurements) showed significant differences. At the 0/100 ␮m zone, MTh control implant surfaces had a higher %BA, whereas in the 100/500 ␮m and 0/1000 ␮m zones, PS implants had significantly higher %BA. The values of %BA for bone subjected to compression or tension (due to orthodontic traction) for orthodontically loaded and the peri-implant bone next to the respective control implants in different zones were compared within implant type by using ANOVA and paired t tests. No significant differences were found (P ⬎ .05). Comparing the %BA next to the total lengths and coronal subregions of the PS and MTh implants indicated significant differences in bone distribution patterns (P ⬍ .05; paired t tests) for both implant types. PS implants generally showed higher %BA next to the coronal subregions for the orthodontically loaded and control implants. In contrast, the orthodontically loaded MTh implants displayed a greater %BA for the total length of the implant compared with coronal subregions. For MTh implants, most of the osseous anchorage appeared to occur in the middle and apical portions of the implant, whereas for the PS implants (orthodontically loaded and controls), the major anchorage occurred in the coronal subregion. MTh control implants

Compression (n ⫽ 8)

Tension (n ⫽ 8)

Control (n ⫽ 10)

displayed no significant differences between the coronal subregion and total length %BA values (Table II). No significant differences were found for MAR between implant types for any of the compression, tension, or control zones (ANOVA; P ⬎ .05). Therefore, MAR values for orthodontically loaded and control implants of each type were grouped and compared for differences associated with implant type using the paired t test. As a result of this analysis, a statistically significant difference in MAR (P ⫽ .011) was detected with MTh implants having higher MAR than PS implants (Table III). Significant differences in fluorochrome labeling were found for bone next to PS and MTh implant surfaces. Orthodontically loaded bone associated with MTh implants presented significantly greater levels of labeling than bone next to PS implants for both compression and tension surfaces (P ⬍ .01; paired t test). Similarly, bone next to MTh control implants showed greater labeling activity than bone around PS control implants (P ⬍ .05; paired t test; Table III). No significant differences were found for singleand double-fluorochrome-labeled, surface-measurement comparisons within implant types in any region of bone next to either implant type. DISCUSSION

Osseointegrated dental implants have been recognized as having considerable value as anchorage units

Oyonarte et al 187

American Journal of Orthodontics and Dentofacial Orthopedics Volume 128, Number 2

Mineral appositional rates (MAR in ␮m per day) and single and double fluorochrome labeled surface (LS) measurements comparisons between implant types

Table III.

MAR-between Porous ⫾ SE Threaded ⫾ SE Mean diff ⫾ SE P value LS-between PS ⫾ SE MTh ⫾ SE Mean diff ⫾ SE P value

Compression (n ⫽ 8)

Tension (n ⫽ 8)

Control (n ⫽ 10)

Total (n ⫽ 26)

1.916 ⫾ .20 2.579 ⫾ .18 ⫺.66 ⫾ .30 .06

2.273 ⫾ .15 2.586 ⫾ .15 ⫺.31 ⫾ .18 .125

2.251 ⫾ .12 2.485 ⫾ .19 ⫺.23 ⫾ .25 .366

2.155 ⫾ .09 2.545 ⫾ .1 ⫺.39 ⫾ .14 . 011

1622.7 ⫾ 336 3132.8 ⫾ 322.4 ⫺1510.1 ⫾ 411 .008

1882.6 ⫾ 333.4 3358.5 ⫾ 330.1 ⫺1475.9 ⫾ 413.4 .009

2103.7 ⫾ 324.2 3177.4 ⫾ 468.6 ⫺1073.7 ⫾ 470.9 .049

1887.7 ⫾ 188.2 3219.4 ⫾ 221.6 ⫺1331.8 ⫾ 248.3 .000

ANOVA: P ⫽ .005 Bonferroni, MTh tension surfaces ⬎ PS compression surfaces, P ⬍ .05.

for complex orthodontic treatment cases, and generally in reducing the need for patient compliance-dependent mechanics. Moreover, their use promotes better treatment outcomes and reduced treatment times. To date, all reported experiences1,4-11,15,16 have been with implants of a threaded design, generally with a machined surface geometry. Other types of implant-surface geometry, including sintered porous surfaces, appear to offer advantages in achieving faster osseointegration because of enhanced osteoconduction and more secure implant-to-bone fixation.17 It is important, therefore, to determine whether there is a preferred surface geometry for implants intended for orthodontic anchorage applications. Continuous bone remodeling is important around dental implants to preserve osseous structural integrity18,19 through the repair of microdamage resulting from loading.20-24 This study was conducted to assess potential differences in peri-implant bone remodeling for implants of 2 widely different surface designs. Implants with a sintered porous surface geometry were compared with MTh implants. Percent BA measurements were used to evaluate the amount and distribution of peri-implant bone after osseointegration. These were determined both for bone next to the total implant length and next to the coronal subregion only. The latter measurements corresponding to the bone adjacent to the 3 most coronal threads of the threaded implants and a similar length for PS implants were specifically chosen for analysis, because it had been reported12 that, under dynamic loading conditions of MTh implants, bone resorption was localized primarily in coronal subregions. In addition, several authors have described variations in the amount of new mineralized bone formation at different distances from loaded dental implant surfaces.7,8,11,25,26 Therefore, the %BA measurements in

our study were assessed within the areas defined by 3 zones—all located within 1 mm of the implant surfaces under investigation. Most of the differences in %BA between implant types occurred in the coronal subregions, with PS implants displaying greater %BA in regions experiencing both tensile and compressive loading compared with MTh implants. Greater %BA in the coronal subregion was also observed in control PS implants. PS implants, whether orthodontically loaded or not, developed most of their osseous anchorage through bone ingrowth in the coronal subregions. In contrast, MTh implants relied primarily on the apical two-thirds length of the threaded implant for osseointegration. These findings likely relate to differences in the way in which force transfer occurs with the 2 designs. The boneimplant interface of machined surfaces cannot accommodate tensile-force transfer across this interface,12,27 in contrast to the sintered, porous surface design that, as a result of 3-dimensional mechanical interlocking of implant and bone, can transfer tensile (as well as shear and compression) forces across the interface.17 When %BA was analyzed comparing PS and MTh implants at the 3 zones next to the coronal peri-implant subregion, differences were observed primarily in the 100/500 ␮m subregions. Significantly lower %BA was found for MTh implants (P ⬍ .05 for bone experiencing compression; P ⬍ .01 for bone in tension and next to control implants). Chen et al,28 in a finite element analysis, predicted that bone strain due to orthodontic loading would be greatest around the tips of MTh implant threads and in the peri-implant region up to 0.7 mm (700 ␮m) away from the threaded implant surface (in the mandibular retromolar area, with axial loads of 100 N). A common finding in studies of continuous orthodontic loading of osseointegrated implants has been an

188 Oyonarte et al

increased amount of bone in regions next to these implants experiencing compression relative to nonloaded control implants.1,4-8,11,25 In our study, %BA for total implant length was greater next to orthodontically loaded MTh implants than to control MTh implants, although the differences were not statistically significant. Orthodontically loaded MTh implants showed lower mean %BA compared with control implants, especially in regions experiencing tension, but the differences were not statistically significant. For PS implants, most orthodontically loaded and control implants showed similar %BA both in regions of compression and tension due to orthodontic loading, or corresponding to no orthodontic force, as was found for the defined coronal subregions. In addition to %BA measurements, our study included estimates of MAR by using fluorochrome bone labeling. MAR was found to be significantly higher next to MTh implants, particularly for the bone compressed by orthodontic traction force. These findings are of particular interest because identical forces were applied, and they suggest that bone adjacent to MTh implant surfaces subjected to compression experienced greater strains. If bone is subjected to large stresses and strains,29 the MAR will increase and the bone will remodel (functionally adapt) to reestablish a preferred physiological state of stress.24,30 Although both PS and MTh implants can become osseointegrated under appropriate conditions, and in both cases osseointegration occurs through mechanical interlock of bone with the implant, the extent and nature of this mechanical interlock is very different for the 2 designs. Although MTh implants establish primarily linear contact with surrounding bone, PS implants become 3-dimensionally interlocked with bone as a result of bone ingrowth into the 3-dimensional, open-pored structure that characterizes the surface region of this design.17 This difference in bone-toimplant fixation results in very different stress distribution within the peri-implant bone. This, in turn, causes differences in MAR for the 2 designs—ie, lower peak stresses and a more uniform distribution of stresses for the PS design,31 resulting in slower remodeling. Within implant design, the only significant difference found in MAR was observed for bone between the orthodontically loaded implants (compression region) and bone next to the control implants of the MTh design, again with higher values observed in the compressively loaded bone. In contrast, with the PS implants, differences in MAR for bone between the orthodontically loaded implant compression regions and the opposing tension regions were not significant. Results of single and double labeling with fluoro-

American Journal of Orthodontics and Dentofacial Orthopedics August 2005

chromes further indicated that orthodontically loaded and control threaded implants resulted in significantly greater bone remodeling next to the threaded implants than the PS implants, supposedly as a result of higher maximum localized stresses in bone next to the MTh implants. Another interesting finding was that MAR was greater for both implant types in the coronal subregion, suggesting greater influence of applied horizontal forces in these regions. Roberts et al4 reported that MAR in dogs was about 1 ␮m per day in steady state. Our results show considerably greater mineral appositional rates for both implant designs. According to Garetto et al,18 a rapid acceleratory phenomenon is expected to persist in bone surrounding osseointegrated dental implants during the first year of implant function. Therefore, the MAR measurements reported here are probably not steady-state values but, rather, might represent the additive effects of osseous response to strain from orthodontic loading superimposed on a rapid acceleratory phenomenon after implantation. This might be preferred, because maintenance of functional rigid fixation of an osseointegrated implant depends on continuously elevated bone-remodeling activity18,19,29; this has been well described.1,4-6 Studies of bone implants subjected to ex vivo fatigue loading and bending,19,28 as well as other studies on bone remodeling around implants in different species including humans,18,32 indicate the occurrence of microfractures of bone surrounding loaded implants. To our knowledge, our study is the first to show in vivo that differences in implant surface design and in stress transfer at the bone-to-implant interface can significantly affect peri-implant bone remodeling. Our findings suggest that PS implants appear suitable for use as orthodontic anchorage units and might be a preferred design. PS implants appear to be better suited to resist loosening from horizontal force application and, based on our %BA, MAR, and oxytetracycline labeling measurements, require lower remodeling activity to sustain healthy osseointegration. Areas of high stress concentration, similar to those predicted for the peri-implant bone surrounding threaded implants,27,28 are less likely with PS implants, thereby resulting in a more physiologic biomechanical environment.31 In addition, because %BA for PS implants occurs primarily in the coronal implant region, shorter implant lengths are feasible for use in orthodontic anchorage. This makes implant placement practical in limited space. The surgical protocol to be used is less invasive, and offers promising alternatives for the dentofacial orthopedic treatment of growing patients.

American Journal of Orthodontics and Dentofacial Orthopedics Volume 128, Number 2

CONCLUSIONS ●

●

Implant surface design appears to be an important determinant of peri-implant bone remodeling with dental implants used as orthodontic anchorage units. Sintered, PS implants show more favorable patterns of bone remodeling under orthodontic loading, compared with MTh implants, suggesting that shorter implant lengths will still maintain osseointegration.

We thank Susan Carter, Nancy Valiquette, and Caroline Chu for their invaluable assistance. REFERENCES 1. Roberts WE, Marshal KJ, Gongloff RK. Rigid endosseous implants for orthodontic and orthopedic anchorage. Angle Orthod 1989;59:247-56. 2. Wehrbein H, Mertz BR. Aspects of the use of endosseous palatal implants in orthodontic therapy. J Esthet Dent 1998;10:315-24. 3. Costa A, Raffaini M, Melsen B. Miniscrews as orthodontic anchorage: a preliminary report. Int J Adult Orthod Orthognath Surg 1998;13:201-9. 4. Roberts WE, Smith RK, Zilberman Y, Mozsary PG, Smith RS. Osseous adaptation to continuous loading of rigid endosseous implants. Am J Orthod 1984;86:95-111. 5. Wehrbein H, Diedrich P. Endosseous titanium implants during and after orthodontic load: an experimental study in the dog. Clin Oral ImpI Res 1993;4:76-82. 6. Wehrbein H, Glatzmaier J, Yildrim M. Orthodontic anchorage capacity of short titanium screw implants in the maxilla: an experimental study in the dog. Clin Oral Impl Res 1997;8:13141. 7. Akin-Nergiz N, Nergiz I, Schulz A, Arpak N, Niedermeier W. Reactions of peri-implant tissues to continuous loading of osseointegrated implants. Am J Orthod Dentofacial Orthop 1998; 114:292-8. 8. Majzoub Z, Finotti M, Miotti F, Giardino R, Nicoli Aldini N, Cordioli G. Bone response to orthodontic loading of endosseous implants in the rabbit calvaria: early continuous distalizing forces. Eur J Orthod 1999;21:223-30. 9. De Pauw GAM, Dermaut L, De Bruyn H, Johansson C. Stability of implants as anchorage for orthopedic traction. Angle Orthod 1999;69:401-7. 10. Saito S, Sugimoto N, Morohashi T, Ozeki M, Kurabayashi H, Shimizu H, et al. Endosseous titanium implants as anchors for mesiodistal tooth movement in the beagle dog. Am J Orthod Dentofacial Orthop 2000;118:601-7. 11. Melsen B, Lang NP. Biological reactions of alveolar bone to orthodontic loading of oral implants. Clin Oral Impl Res 2001; 12:144-52. 12. Duyck J, Ronold HJ, Van Oosterwick H, Naert I, Vander Sloten J, Ellingsen JE. The influence of static and dynamic loading on marginal bone reactions around osseointegrated implants: an animal experimental study. Clin Oral Impl Res 2001;12:207-18. 13. Oyonarte R, Deporter D, Pilliar RM, Woodside DG. Peri-implant bone response to orthodontic loading: part I. A histomorphometric study of implant surface design. Am J Orthod Dentofacial Orthop 2005; in press. 14. Pilliar RM. Porous-surfaced metallic implants for orthopedic applications. J Biomed Mat Res Appl Biomat 1987;21:1-33.

Oyonarte et al 189

15. Smalley W, Shapiro PA, Hohl TH, Kokich VG, Branemark P. Osseointegrated titanium implants for maxillofacial protraction in monkeys. Am J Orthod Dentofacial Orthop 1988;94:285-95. 16. Linder-Aronson S, Nordenram A, Anneroth G. Titanium implant anchorage in orthodontic treatment: an experimental investigation in monkeys. Eur J Orthod 1990;12:414-9. 17. Pilliar RM. Overview of surface variability of metallic endosseous dental implants: textured and porous surface-structured designs. Implant Dent 1998;7:305-14. 18. Garetto LP, Chen J, Parr JA, Roberts WE. Remodeling dynamics of bone supporting rigidly fixed titanium implants: a histomorphometric comparison in four species including humans. Implant Dent 1995;4:235-43. 19. Huja SS, Katona TR, Burr DB, Garetto LP, Roberts WE. Microdamage adjacent to endosseous implants. Bone 1999;25: 217-22. 20. Burr DB, Martin RB, Schaffler MB, Radin EL. Bone remodeling in response to in vivo fatigue microdamage. J Biomech 1985;18: 189-200. 21. Mori S, Burr DB. Increased intracortical remodeling following fatigue damage. Bone 1993;14:103-9. 22. Frost HM. Vital biomechanics: proposed general concepts for skeletal adaptations to mechanical usage. Calc Tiss Int 1988;42: 145-55. 23. Frost HM. Transient-steady state phenomena in microdamage physiology: a proposed algorithm for lamellar bone. Calc Tiss Int 1989;44:367-81. 24. Frost, HM. Wolff’s law and bone’s structural adaptations to mechanical usage: an overview for clinicians. Angle Orthod 1994;64:175-88. 25. Gotfredsen K, Berglund T, Lindhe J. Bone reactions adjacent to titanium implants subjected to static load: a study in the dog. I. Clin Oral Impl Res 2001;12:1-8. 26. Gotfredsen K, Berglund T, Lindhe J. Bone reactions adjacent to titanium implants with different surface characteristics subjected to static load: a study in the dog. II. Clin Oral Impl Res 2001;12:196-201. 27. Van Oosterwyck H, Duyck J, Vander Sloten J, Van der Perre G, De Cooman M, Lievens S, et al. The influence of bone mechanical properties and implant fixation upon bone loading around oral implants. Clin Oral Impl Res 1998;9:407-18. 28. Chen J, Esterle M, Roberts WE. Mechanical response to functional loading around the threads of retromolar endosseous implants utilized for orthodontic anchorage: coordinated histomorphometric and finite element analysis. Int J Oral Maxillofac Impl 1999;14:282-9. 29. Burr DB, Schaffler MB, Yang KH, Lukoschek M, Sivaneri N, Blaha JD, et al. Skeletal response to elevated strain environments: Is woven bone a response to elevated strain? Bone 1989;10:223-33. 30. Roberts WE. Bone physiology, metabolism and biomechanics in orthodontic practice. In: Graber TM, Vanarsdall RL, editors. Orthodontics: current principles and techniques. 3rd ed. Saint Louis: Mosby; 2000. 31. Pilliar RM. Implant surface design for development and maintenance of osseointegration. In: Ellingsen JE, Lyngstadaas SP, editors. Bio-implant interface: improving biomaterials and tissue reactions. Boca Raton, Fla: CRC Press; 2003. p. 43-58. 32. Trisi P, Rebaudi A. Progressive bone adaptation of titanium implants during and after orthodontic load in humans. Int J Period Rest Dent 2002;22:31-43.