Purpose: To evaluate the effects of implant geometry and collar macrostructure and microstructure on crestal bone height, mobility values, and attachment levels using two different types of radiographic analysis. Materials and Methods: Subjects included in this study were those who had single-tooth implants restored in the maxillary esthetic zone. Bone levels were measured from the implant-abutment (IA) interface on standardized radiographs at baseline and after 5 years using digital and standard methods. The digital method converted the radiographs into three-dimensional models based on radiodensity so that bone levels could be measured. Clinical attachment levels and Periotest values were recorded and all data were analyzed statistically. Results: The bone position from the IA interface for the rough-collar implant group was a mean of -0.61 ± 0.08 mm and it was -1.55 ± 0.10 mm for the smooth-collar implant group (-1.96 ± 0.16 mm for implants with a stepped collar and -0.85 ± 0.19 mm for implants with a straight collar). The mean bone level change over the 5-year period for the rough implants was -0.19 ± 0.09 mm and it was -0.36 ± 0.06 mm for the smooth implants. The crestal bone position relative to the implant at the time of surgery influenced mean bone level changes significantly. There were statistically significant differences in clinical attachment levels depending on implant location and on implant group and length. Conclusions: Implants with straight collars had less bone loss at the 5-year interval than implants with stepped collars. Bone level changes were greater when the implant was placed further subcrestal. The digitally converted radiographic method was significantly more precise for evaluating changes in bone levels. These findings suggest that implant geometry and implant surface roughness may play a significant role in bone level changes in the esthetic zone. Keywords: bone loss, dental implants, implant geometry, radiographic analysis, single tooth, surface roughness