The purpose of this study was to evaluate experimental peri-implant breakdown microbiologically, radiographically and histologically. Hydroxyapatite-coated, titanium plasma-sprayed, and titanium alloy surfaces were investigated. Eighty-four implants were placed in 14 beagle dogs. Standardized radiographs and microbiologic samples (DNA) were obtained. Dogs were sacrificed at 3 and 6 months. Undecalcified histologic sections were prepared. Thickness of hydroxyapatite coating, changes in crestal bone height, and marginal changes in osseointegration were measured. Vertical bone loss was computed. Radiographs were analyzed using computer-assisted densitometric image analysis (CADIA). Microbial analysis (DNA) did not clearly favor any of the examined surfaces. CADIA did not show differences among implant surfaces. No significant differences among the three implants were noted for histometry, except the experimental titanium plasma-sprayed surface showed an increase in vertical bone loss 6 months (P < .05). Thickness of hydroxyapatite was decreased in active peri-implantitis sites (P < .05). Clinical attachment level was shown to be the most sensitive clinical parameter for detecting histologic changes. All implants were equally susceptible to peri-implantitis.