Purpose: To evaluate the influence of the collar surface topography on peri-implant marginal bone preservation. Materials and Methods: A total of 156 patients who received at least one cylindrical implant of regular diameter with an external-hexagon platform in the posterior region of the maxilla or mandible were recalled to the office for a retrospective evaluation. Implantation sites and implant length information were recorded, and implants were divided according to the collar surface topography: machined (M) or rough (R) surface. Each implant was assessed by digital periapical radiography, using a sensor holder for the paralleling technique. The marginal bone remodeling was determined as the distance from the implant platform to the first bone-to-implant contact, and the known implant length was used to calibrate the images in the computer software. The follow-up measurements were compared with those obtained from the radiograph taken at the time of prosthetic loading to determine the late bone remodeling. An independent t test was used to compare data. Results: From 138 patients who attended the recall, 242 implants (M = 126; R = 116) were evaluated with a mean follow-up period of 4.6 ± 0.9 years. Similar success rates were found in both groups (M = 95.0%; R = 95.9%). Late bone remodeling in the maxilla was not influenced by the implant collar (P = .504); however, lower marginal bone loss was observed in the R group (1.20 ± 0.52 mm) compared with the M group (1.58 ± 0.73 mm) in the mandible (P = .007). Conclusion: Although the implant collar design did not influence the success rate of dental implants, the rough collar design reduced the late marginal bone remodeling around external-hexagon implants in the mandible. The maxilla was not benefited by the rough collar design. Keywords: bone loss, dental implants, osseointegration, surface