Influence of a Machined Collar on Crestal Bone Changes Around Titanium Implants: A Histometric Study in the Canine Mandible

Background: It has been shown that peri-implant crestal bone reactions are influenced by both a rough-smooth implant border in one-piece, non-submerged, as well as an interface (microgap [MG] between implant/abutment) in two-piece butt-joint, submerged and non-submerged implants being placed at different levels in relation to the crest of the bone. According to standard surgical procedures, the rough-smooth implant border for implants with a smooth collar should be aligned with the crest of the bone exhibiting a smooth collar adjacent to peri-implant soft tissues. No data, however, are available for implants exhibiting a sandblasted, large-grit and acid-etched (SLA) surface all the way to the top of a non-submerged implant. Thus, the purpose of this study is to histometrically examine crestal bone changes around machined versus SLA-surfaced implant collars in a side-by-side comparison. Methods: A total of 60 titanium implants (30 machined collars and 30 SLA collars) were randomly placed in edentulous mandibular areas of five foxhounds forming six different subgroups (implant subgroups A to F). The implants in subgroups A to C had a machined collar (control), whereas the implants in subgroups D to F were SLA-treated all the way to the top (MG level; test). Furthermore, the MGs of the implants were placed at different levels in relation to the crest of the bone: the implants in subgroups A and E were 2 mm above the crest, in subgroups C and D 1 mm above, in subgroup B 3 mm above, and in subgroup F at the bone crest level. For all implants, abutment healing screws were connected the day of surgery. These caps were loosened and immediately retightened monthly. At 6 months, animals were sacrificed and non-decalcified histology was analyzed by evaluating peri-implant crestal bone levels. Results: For implants in subgroup A, the estimated mean crestal bone loss (+/- SD) was -0.52 +/- 0.40 mm; in subgroup B, +0.16 +/- 0.40 mm (bone gain); in subgroup C, -1.28 +/- 0.21 mm; in subgroup D, -0.43 +/- 0.43 mm; in subgroup E, -0.03 +/- 0.48 mm; and in subgroup F, -1.11 +/- 0.27 mm. Mean bone loss for subgroup A was significantly greater than for subgroup E (P=0.034) and bone loss for subgroup C was significantly greater than for subgroup D (P <0.001). Conclusions: Choosing a completely SLA-surfaced non-submerged implant can reduce the amount of peri-implant crestal bone loss and reduce the distance from the MG to the first bone-implant contact around unloaded implants compared to implants with a machined collar. Furthermore, a slightly exposed SLA surface during implant placement does not seem to compromise the overall hard and soft tissue integration and, in some cases, results in coronal bone formation in this canine model. J Periodontol 2011;82:1329-1338.