Biologic Width and Crestal Bone Remodeling with Sintered Porous-Surfaced Dental Implants: A Study in Dogs

Transcription

1 Biologic Width and Crestal Bone Remodeling with Sintered Porous-Surfaced Dental Implants: A Study in Dogs Douglas Deporter, DDS, PhD 1 /Arwa Al-Sayyed, DDS, MSc 2 /Robert M. Pilliar, BSc, PhD 1 /Nancy Valiquette, BSc 1 Purpose: The aim of this study was to obtain histometric measurements of bone and peri-implant mucosal tissue contact with implants of 2 sintered porous-surfaced designs. The short-collar design had a collar height (smooth coronal region) of 0.75 mm, while the long-collar model had a smooth coronal region of 1.8 mm. Materials and Methods: Implants (2 per side) were placed in healed mandibular extraction sites of 4 beagle dogs using a submerged technique. After 4 weeks of healing, they were uncovered and used to support fixed partial dentures for a 9-month period. After sacrifice, specimens were retrieved and nondemineralized sections were examined histometrically to determine the most coronal bone-to-implant contact (first BIC) using the microgap as a reference and standard mucosal parameters of biologic width. Results: Significant (P =.001) differences in first BIC were found between designs (1.97 mm for long-collar versus 1.16 mm for short-collar implants) for posteriorly located implants but not for anteriorly located ones (1.21 mm versus 1.38 mm; P =.40). If crestal bone loss involved sintered surface, fibrous connective tissue ingrowth was observed to replace lost bone. No significant differences in peri-implant mucosal measurements (total peri-implant mucosal thickness; length of the epithelial component of this mucosa, and thickness of the connective tissue component) were detected between implant designs. Conclusions: Results suggest that biologic width accommodation drives initial crestal bone loss with sintered porous-surfaced implants. Histometric data obtained for bone contact showed no significant differences between the long- and shortcollar implant designs. INT J ORAL MAXILLOFAC IMPLANTS 2008;23: Key words: biologic width, bone remodeling, collar height, sintered porous-surfaced dental implants Anumber of explanations have been presented as possible causes for early crestal bone loss with endosseous, root-form dental implants. These have included surgical trauma, 1 thin buccal cortical bone with compromised vascularity, 2 high localized 1 Professor, Department of Periodontics, University of Toronto, Faculty of Dentistry, Toronto, Ontario, Canada. 2 Consultant, Periodontics and Dental Implants, Riyadh Armed Forces Hospital, Riyadh, Saudi Arabia. 3 Professor Emeritus, Department of Biomaterials, University of Toronto, Faculty of Dentistry, Toronto, Ontario, Canada. 4 Research Assistant, Department of Biomaterials, University of Toronto, Faculty of Dentistry, Toronto, Ontario, Canada. Correspondence to: Dr Douglas A. Deporter, University of Toronto, Faculty of Dentistry, 124 Edward Street, Toronto, ON M5G 1G6, Canada. Fax: douglas.deporter@utoronto.ca stresses at thread apices in the case of threaded implants, 3 5 stress shielding adjacent to machined (smooth) coronal collar regions of implants, 6 8 the level and size of the microgap between the implant (root component) and prosthetic abutment, 9,10 and establishment of biologic width in the peri-implant mucosal tissues While there is not yet a fundamental understanding of why a biologic width invariably develops with natural teeth and implants, it appears that once an endosseous dental implant is exposed to the oral environment, a mucosal seal of minimal thickness forms around its neck apical to the abutment-implant junction, with a characteristic microgap. In animal studies with a variety of implant designs, this junctional seal has been reported to include a soft connective tissue component with a thickness of approximately 1.0 to 1.5 mm Volume 23, Number 3, 2008

2 Fig 1 The implant designs used. Arrows denote the microgap regions. The short-collar design (left) had a machined collar height of 0.75 mm, while the long-collar design (right) had a collar height of 1.8 mm. Fig 2 A composite drawing showing some of the histometric measurements collected. The left side represents the features of the long-collar (1.8-mm collar) design, while the right side shows the short-collar design. By the end of the experiment, bone levels (B) with the long collar ended at or just coronal to the machinedto-porous-surfaced junction (MP-jx), indicated by the arrow on the left side. With the short-collar design, loss of crestal bone (first BIC) had extended 0.5 mm beyond this junction. GM = gingival margin; aje = apical extension of the junctional epithelium; B = the most coronal point of bone contact; A/F = the microgap; f-bic = first bone-to-implant contact; PM = peri-implant mucosa; CTC = thickness of connective soft tissue. All previously reported animal investigations of biologic width and implants have involved the use of threaded implant devices. Sintered porous-surfaced implants generally are nonthreaded and become securely fixed (osseointegrated) by bone ingrowth into the 300-µm thick, sintered porous-surface layer that is formed over the intended bone-interfacing region. In initial animal studies done with sintered porous-surfaced implants, limited crestal bone loss was observed primarily next to the smooth (non porous-coated) coronal-most (collar) region of the implants. In an experiment in which implants with 2 different coronal collar heights were compared, radiographic data indicated that the shorter of the 2 resulted in less crestal bone loss over a 9- month functional period. 14 It appeared that once bone had resorbed to the level of the machined collar-to-sintered porous-surfaced junction (MP-jx) of the implants, resorption ceased, since force transfer across the mechanically interlocked porous surface was then possible. 6,15 In recognition of the limitations of radiographs in terms of accurate quantitative measurement of bone loss, for the present study bone loss was assessed physically by means of specimens retrieved from dogs. Quantitative microscopy was used to measure hard and soft tissue implant interfaces. MATERIALS AND METHODS Two sintered porous-surfaced dental implant designs (overall length, 5.4 mm) made to the investigators specifications by Innova Corporation (Toronto, ON, Canada) were used in the present study. They differed only in the height of their machined smooth coronal region: One model ( long collar ) had a smooth coronal region of 1.8 mm, and the other ( short collar ) had a smooth coronal region of 0.75 mm (Fig 1). Two implants of each type were placed in healed premolar extraction sites on opposite sides of the mandibles (ie, one type of implant per side) of each of 4 beagle dogs. Implants were installed with the whole implant root (including the smooth coronal region) submerged within bone, leaving only the healing caps unsubmerged. One month later, implants were uncovered and connected to transgingival abutments, and 1 week later they were loaded with implant-supported custom-made fixed partial dentures. All implant crown margins were supragingival and did not affect the condition of peri-implant mucosa. Teeth and implants were cleaned thrice weekly using topical chlorhexidine and toothbrushes supplemented with a waterirrigating device. Fixed partial dentures were left in function for 9 months, at which time animals were sacrificed and tissue specimens were collected. The International Journal of Oral & Maxillofacial Implants 545

3 Table 1 Mean Values for ACL and CLF ACL CLF Mean SD Mean SD Short collar Long collar P Table 2 Mean Values in mm for First BIC First BIC Posterior implants Anterior implants Long collar Short collar P ACL = absolute contact length; CLF = contact length fraction. Tissue blocks were processed and embedded in methylmethacrylate (Osteobed; Polysciences, Warrington, PA) to allow preparation of nondemineralized sections. 16 Blocks were sectioned in each instance as close to the mid-region of the implant as possible to provide 4 sections (2 buccolingual and 1 each in the mesial and distal orientations) of each implant and its surrounding tissues. Photographs were obtained of all sections (final magnification 40 ) and assessed using a Bioquant System IV software package (R & M Biometrics, Nashville, TN). The extent of contact between bone and porous implant surface in each section was expressed as absolute bone contact length (ACL) and as contact length fraction (CLF), as previously described. 14 CLF values for each section were determined by dividing ACL by the maximum length of the same implant surface available for bone contact, ie, CLF = ACL/maximal length available for contact with bone. Measurement of first bone-implant contact (BIC) was included to quantify crestal bone loss in each of the 4 sections of each implant. This parameter was calculated in each section as a straight line measurement from the top of the implant (ie, the level of the microgap) to the highest point of alveolar bone contact with implant surface, either the porous-surfaced or machined collar region (Fig 2). The peri-implant mucosa-to-implant interface was assessed using previously established histometric parameters, 12 as shown in Fig 2. The 4 parameters measured were total thickness of peri-implant mucosa (PM); length of the epithelial component of this mucosa (GM to aje); thickness of soft connective tissue (CTC) between the apical termination of the junctional epithelium and the most coronal point of bone contact with the implant (aje to B); and height of mucosa above the microgap (GM to A/F). Statistical Analysis Analysis of variance was used. The factors analyzed were animal, implant location (anterior or posterior implant), aspect of implant (buccal, lingual, mesial, and distal), and implant type (short collar versus long collar). P <.05 was considered the threshold of statistical significance. RESULTS Histometric data obtained for bone contact are shown in Table 1. There were no significant differences in absolute bone contact length or contact length fraction with respect to implant design. There were also no differences for these parameters for the variables dog, implant site, or aspect of implant. Crestal bone loss as first BIC (using the microgap as reference) is displayed in Table 2. ANOVA indicated a significant interaction (P =.002) between first BIC and implant site, ie, anterior versus posterior location. As a result, there were significant differences (P =.001) in first BIC between designs, but only for the posteriorly located implants (1.97 mm for the long-collar versus 1.16 mm for the short-collar implants). Differences in first BIC between the long- and short-collar implants in anterior locations were not significant (P =.40). There was also a significant effect (P =.02) of implant surface; lingual surfaces had mean first BIC values (1.01 mm) significantly less than buccal (1.58 mm), mesial (1.61 mm), or distal (1.27 mm) surfaces. It is useful here to relate mean bone loss data to machined collar (smooth coronal region) height. The height of the smooth coronal region for the long-collar design was 1.8 mm, and mean bone loss (first BIC) for this design indicated that the bone crest was coronal to the MP-jx in the case of anterior implants (first BIC = 1.21 mm) or slightly apical (mean of 0.17 mm) to this junction (first BIC = 1.97 mm) for posteriorly located implants. Two sample sections of longcollar implants are shown in Fig 3. For short-collar implants (smooth coronal region = 0.75 mm), bone loss (first BIC) was 1.16 mm for posteriorly located implants versus 1.38 mm for anteriorly located ones. Clearly, in both instances with short-collar implants, bone loss resulted in first BIC being located apical to MP-jx. Bone loss had proceeded beyond MP-jx by a mean of 0.41 mm for posteriorly located implants and of 0.63 mm for anteriorly located implants. However, in such instances, any exposed sintered surface had become ingrown with connective tissue fibers (Fig 4). 546 Volume 23, Number 3, 2008

4 Fig 3 Photomicrographs of long-collar implants. (a) A sample in which the crestal bone remains above the MP-jx (arrow). (b) A sample in which the crest has resorbed to this junction (arrow). a 150 µm b 200 µm Fig 4 Photomicrographs of a short-collar implant. (a) The bone crest (lower arrow) is apical to the MP-jx (upper arrow). Junctional epithelium (which separated from the collar during processing to create an artifactual space, indicated with an asterisk) extends to this junction. The short segment (about 0.5 mm) of sintered surface no longer embedded in bone is ingrown with connective tissue to form a "gingival attachment." The latter can be seen in (b), which shows the area of interest from 4a at a higher magnification. * b 100 µm a 150 µm With regard to peri-implant mucosal measurements (Table 3), there were no significant differences between implant designs. Total mucosal thickness (PM) was 4.01 mm for short-collar implants and 4.03 mm for long-collar implants (P =.48). Mean lengths of junctional epithelium (GM to aje) were 1.96 mm for short collars and 1.56 mm for long collars (P =.69), while the corresponding thicknesses of connective tissue were 2.20 mm (short collar) and 2.12 mm (long collar; P =.79). Height of mucosa using the microgap as a fixed reference (gingival margin to the microgap) was greater for the short-collar than for the long-collar design (2.79 mm versus 2.56 mm), although again, the difference was not significant (P =.17). The International Journal of Oral & Maxillofacial Implants 547

5 Table 3 Mean Values in mm for Peri-Implant Mucosal Measurements PM GM-aJE CTC GM-A/F Mean SD Mean SD Mean SD Mean SD Short collar Long collar PM = total thickness of peri-implant mucosa; GM-aJE = length of the epithelial component of the periimplant mucosa; CTC = connective tissue component of the peri-implant mucosa; GM-A/F = height of the gingival margin in relation to the microgap. DISCUSSION The present report provides histometric observations from an experiment with 2 sintered porous-surfaced dental implant designs for which radiographic data on crestal bone profile changes (ie, loss of crestal bone) have been reported. 14 The 2 designs were identical but for 1 feature. They both had a tapered truncated cone shape with the majority of implant length having a sintered surface geometry, but one design ( long collar ) had a 1.8-mm smooth coronal region, while the other ( short collar ) had a 0.75 mm smooth coronal region. Radiographic data collected over a 9-month functional period suggested crestal bone loss to be limited to the SCR for both designs and, therefore, to be less with the short collar. The loss of crestal bone was attributed to low local stresses acting on bone in the collar regions due to secure implant fixation more apically (through bone ingrowth) and, hence, inadequate force transfer from implant to bone in coronal regions. Bone disuse atrophy was hypothesized, and the process attributed to stress shielding of bone in this region by the securely fixed, highly stiff (relative to bone) titaniumalloy implant. 5,6,7,14 It was proposed that this would lead to remodeling of crestal bone and a net loss of bone. A similar effect had been proposed for other implant designs with smooth-collar regions. 8 The histometric data reported here, however, suggest that this earlier explanation attributing crestal bone loss primarily to a biomechanical cause was not correct. Histometric measurements were made for parameters including bone contact with sintered surface (absolute contact length and contact length fraction), crestal bone loss (first BIC), and various periimplant soft tissue dimensions. All implants were successfully integrated. Absolute contact length and contact length fraction values were similar for both designs (Table 1), although they were somewhat lower than those reported in earlier experiments This outcome was interpreted to indicate that the interface between bone and sintered surface was highly effective in creating and maintaining implant integration. The overall linear length of porous surface was 22% shorter for long-collar implants than for short-collar ones, and yet they both achieved similar degrees of bone contact in response to the applied loads. This would be in keeping with findings in humans with sintered porous-surfaced implants that longer lengths (9 or 12 mm) offer no benefit over shorter ones (7 mm). 18,19 First BIC was used to assess crestal bone loss and was meant to be the histometric counterpart of the radiographic measurements already reported. 14 First BIC was significantly different for the 2 designs but also was dependent on implant location. Posterior implants had mean first BIC of 1.97 mm for long-collar implants, and this was significantly greater than that for short-collar implants (mean first BIC of 1.16 mm). However, there was no significant difference for mean first BIC between designs when implants were located anteriorly (long collar = 1.21 mm; short collar = 1.38 mm). The reason for this location-dependent variation is not clear, but it may be related to the sample size. These patterns of loss could not be explained on the basis of stress shielding. For both long-collar (smooth-collar region of 1.8 mm), at least for distally located long-collar implants, and shortcollar designs, bone loss progressed beyond MP-jx (0.17 mm for the posteriorly located long-collar implants; 0.41 mm for posteriorly located short-collar implants and 0.63 mm for anteriorly located shortcollar implants). These results did not agree with the corresponding earlier radiographic findings, 14 which suggested that crestal bone loss was strictly limited to the machined collar segments of both implant designs. This difference in outcomes between radiographic and histometric measures can be explained by the fact that radiographic images underestimate marginal bone loss with dental implants. 20 In this latter investigation, bone levels measured histometrically around threaded implants placed in goat jaws were located 0.85 mm more apically than estimated in radiographs. 548 Volume 23, Number 3, 2008

6 The present findings with sintered porous-surfaced implants agree with those of Hermann and colleagues, 9,10,21 who used threaded implants with a sand-blasted and acid-etched surface (SLA; Straumann) and polished collars of varying heights. They concluded that a 2-piece SLA-treated implant suffered bone loss until the alveolar crest came to be at least 2 mm apical to the microgap. This compares well with the to 1.97-mm loss found in the present study with sintered porous-surfaced implants. Hermann et al 21 also reported that if the polished collar was less than 2 mm in height, bone loss extended beyond the machined-to-sla-surface junction, ie, to the SLA surface. It must be recognized as well that in the present study, abutments were removed and replaced monthly throughout the study to allow connection of the custom radiographic film holder to be connected to each implant, 14 and this may have contributed to crestal bone loss as well. 22 Histometric data collected on biologic width in the present study generally agreed with that reported by other investigators for threaded dental implants. With Brånemark-type implants (Nobel Biocare, Göteborg, Sweden) used in dogs, Berglundh et al 12 reported an overall peri-implant mucosal thickness of 3.8 mm, with 1.66 mm of connective tissue and a thickness of 2.14 mm for the junctional epithelium. Abrahamsson et al 11 examined biologic widths with 3 different threaded implant designs, 2 of which were placed with a traditional 2-stage surgical approach (Astra Tech [Molndal, Sweden] & Brånemark-type). The third (Bonefit ITI threaded implants; Straumann, Basel, Switzerland) was placed with a 1-stage approach. The authors reported that all 3 types developed similar interfaces, with an epithelial thickness of 1.64 to 2.35 mm and a connective tissue thickness of 1.15 to 1.47 mm. They also reported that where peri-implant mucosal tissue was thin, angular bone defects developed at the alveolar crest as a result of compensatory resorption. This outcome was further tested by Berglundh and Lindhe. 13 They placed Brånemark-type implants in dogs, and at re-entry, they purposefully reduced the connective tissue thickness around some implants using a split-thickness flap and tissue discard. It was reported that implants with thinned ( 2 mm) mucosa consistently showed crestal bone cratering, while controls did not. Earlier animal work in the laboratories of the present authors used sintered porous-surfaced implants with a 2-mm machined collar height. In those studies, presumably because a 2-mm smooth coronal region is sufficient for biologic width accommodation, crestal bone loss beyond MP-jx was not seen. In the present study, where both implant designs had somewhat shorter collar heights (1.8 mm and 0.75 mm), in all instances with the 0.75-mm smooth coronal region and some situations with the 1.8-mm smooth coronal region, crestal bone loss proceeded beyond MP-jx. Unlike other types of implant surface geometry (such as machined surfaces, plasmasprayed titanium or hydroxyapatite surfaces, or sandblasted, acid-etched surface treatments), where periimplant mucosal tissue fibers are oriented parallel to implant surfaces, sintered porous-surfaced implants are characterized by oblique fiber orientation relative to the implant surface, a consequence of connective tissue ingrowth into the interconnected pore network. 26,27 This soft tissue attachment seemed to provide an effective seal around the implant necks, as it was not associated with obvious signs of peri-implant inflammation, as has been reported for some other textured implant surfaces denuded of their original bone contact. 28 This interpretation is supported by recent results in human clinical trials with sintered porous-surfaced implants. 18,29 Implants used in these investigations had 1-mm-wide smooth coronal regions and as a result did experience some loss of crestal bone beyond MP-jx, presumably in order to establish appropriate biologic widths. Nevertheless, following resultant bone loss, crestal bone levels remained stable over functional periods of 5 or more years. In summary, early crestal bone loss in relation to sintered porous-surfaced implants appears to be driven by the need to establish a biologic width of periimplant mucosa, as is the case with other dental implant designs. If machined collar height is less than 2 mm, bone loss will continue beyond the junction between the machined surface and the sintered porous layer until biologic width is established. However, should the sintered region become denuded of bone as a result of crestal bone loss, it will become ingrown with mucosal collagen fibers, ensuring a stable implant-tissue interface. ACKNOWLEDGMENTS The authors wish to thank Dr Dennis Tarnow for his helpful critical comments on an earlier version of this paper. They also thank Lynda Woodcock for assistance in preparing the manuscript for submission. The International Journal of Oral & Maxillofacial Implants 549

7 REFERENCES 1. Gomez-Roman G. Influence of flap design on peri-implant inter-proximal crestal bone loss around single-tooth implants. Int J Oral Maxillofac Implants 2001;6: Spray JR, Black CG, Morris HF, Ochi S. Influence of bone thickness on facial marginal bone response: Stage 1 placement through stage 2 uncovering. Ann Periodontol 2000;5: Adell R, Lekholm U, Rockler B, et al. Marginal tissue reactions at osseointegrated titanium fixtures. A three-year longitudinal prospective study. Int J Oral Maxillofac Surg 1986;15: Rieger MR, Mayberry M, Brose MO. Finite element analysis of six endosseous implants. J Prosthet Dent 1990;63: Oyonarte R, Deporter DA, Pilliar RM, Woodside DG. Periimplant bone response to orthodontic loading. Part 2: Implant surface geometry and its effect on regional bone remodeling. Am J Orthod Dentofacial Orthop 2005;128: Pilliar RM, Deporter DA, Watson PA, Valiquette N. Dental implant design: Effect on bone remodeling. J Biomed Mater Res 1991;25: Vaillancourt H, Pilliar RM, McCammond D. Factors affecting crestal bone loss with dental implants partially covered with a porous coating. A finite element analysis. Int J Oral Maxillofac Implants 1996;11: Wiskott HW, Belser UC. Lack of integration of smooth titanium surfaces: A working hypothesis based on strains generated in the surrounding bone. Clin Oral Implants Res 1999;10: Hermann JS, Buser D, Schenk RK, Cochran DL. Crestal bone changes around titanium implants. A histometric evaluation of unloaded, non-submerged and submerged implants in the canine mandible. J Periodontol 2000;71: Hermann JS, Schoolfield JD, Schenk RK, Buser D. Influence of the size of the microgap on crestal bone changes around titanium implants. A histometric evaluation of unloaded nonsubmerged implants in the canine mandible. J Periodontol 2001;72: Abrahamsson I, Berglundh T, Wennstrom J, Lindhe J.The periimplant hard and soft tissues at different implant systems. A comparative study in the dog. Clin Oral Implants Res 1996; 7: Berglundh T, Lindhe J, Ericsson I, Marinello CP, Liljenberg B, Thomsen P.The soft tissue barrier at implants and teeth. Clin Oral Implants Res 1991;2: Berglundh T, Lindhe J. Dimension of the peri-implant mucosa. Biological width revisited. J Clin Periodontol 1996;23: Al-Sayyed A, Deporter DA, Pilliar RM, et al. Predictable crestal bone remodelling around two porous-coated titanium alloy dental implant designs. A radiographic study in dogs. Clin Oral Implants Res 1994;5: Deporter DA, Watson PA, Pilliar RM, Chipman M, Valiquette N. A histological comparison in the dog of porous-surfaced vs threaded dental implants. J Dent Res 1990;69: Deporter DA, Watson PA, Pilliar RM, et al. A histological assessment of the initial healing response adjacent to porous-surfaced, titanium alloy dental implants in dogs. J Dent Res 1986; 65: Pilliar RM, Deporter DA, Watson PA, et al.the effect of partial coating with hydroxyapatite on bone remodelling in relation to porous-surfaced titanium alloy dental implants in dogs. J Dent Res 1991;70: Rokni S,Todescan R, Watson P, Pharoah M, Adegbembo AO, Deporter D. An assessment of crown-to-implant root ratios with short sintered porous-surfaced implant-borne prostheses in partial edentulism. Int J Oral Maxillofac Implants 2005; 20: Deporter DA, Watson PA, Pharoah M,Todescan R,Tomlinson G. Ten-year results of a prospective study using porous-surfaced dental implants and a mandibular overdenture. Clin Implant Dent Relat Res 2002;4: Caulier H, Naert I, Kalk W, Jansen JA.The relationship of some histologic parameters, radiographic evaluations, and Periotest measurements of oral implants: An experimental animal study. Int J Oral Maxillofac Implants 1997;12: Hermann JS, Cochran D, Nummikoski P, Buser D. Crestal bone changes around titanium implants. A radiographic evaluation of unloaded nonsubmerged and submerged implants in the canine mandible. J Periodontol 1997;68: Abrahamsson I, Berglundh T, Lindhe J.The mucosal barrier following abutment dis/reconnection. An experimental study in dogs. J Clin Periodontol 1997;24: Comut AA, Weber HP, Shortkroff S, Cui FZ, Spector M. Connective tissue orientation around dental implants in a canine model. Clin Oral Implants Res 2001;12: Abrahamsson I, Zitzmann NU, Berglundh T, Linder E, Wennerberg A, Lindhe J.The mucosal attachment to titanium implants with different surface characteristics: An experimental study in dogs. J Clin Periodontol 2002;29: Schierano G, Ramieri G, Cortese MG, Aimetti M, Preti G. Organization of the connective tissue barrier around long-term loaded implant abutments in man. Clin Oral Implants Res 2002;13; Inoue T, Cox JE, Pilliar RM, Melcher AM. Effect of the surface geometry of smooth and porous-surfaced titanium alloy on the orientation of fibroblasts in vitro. J Biomed Mater Res 1987;21: Pilliar RM, Cameron HU, Binnington AG. Radiographic and morphologic studies of load-bearing porous surface-structured implants. Clin Orthop Relat Res 1981;156: Hermann JS, Buser D, Schenk RK, Scholfield JD, Cochran DL. Biologic width around one- and two-piece titanium implants. A histometric evaluation of unloaded non-submerged and submerged implants in the canine mandible. Clin Oral Implants Res 2001;12: MacDonald K, Pharoah M,Todescan R, Deporter DA. Use of sintered porous-surfaced dental implants to restore single teeth in the maxilla: A 7 to 9 year follow-up. Int J Periodontics Restorative Dent 2006 (in press). 550 Volume 23, Number 3, 2008