Purpose: To examine the osseointegration of various implant surfaces after bacterial contamination and cleaning. Materials and Methods: Four types of implant surface were manufactured: machined (M); plasmaspray hydroxyapaptite (HA); sandblasted, large-grit, acid-etched (SA); and titanium anodic oxide (TAO) were manufactured. The surface characteristics of these implants were determined using a scanning electron microscope, an energy dispersive spectrometer, and a contact profilometer. Each surface was subdivided into control and test groups. Test implants were co-incubated with Prevotella intermedia for 2 weeks, then cleaned with cotton pellets, soaked in saline, and irrigated. Control implants underwent the same cleaning procedure, but without bacterial contamination. Four control or test implants with different surface types were randomly inserted into the tibia of 10 New Zealand white rabbits. After 6 weeks of healing, 5 rabbits were sacrificed for histomorphometry, and the rest for removal torque assay. Results: Bacterial contamination adversely influenced every implant surface in terms of bone-to-implant contact (BIC) ratio and required removal torque. The negative results reached significant levels for rougher surfaces (HA and SA). For both contaminated and uncontaminated samples, HA and SA implants required significantly higher removal torque than that required for M implants. Conclusion: Bacterial contamination jeopardized osseointegration on every tested implant surface. A more negative effect on BIC was found for implants with rougher surfaces. However, contaminated rough-surfaced implants showed more removal torque resistance than contaminated smooth implants. Palabras clave: histometry, osseointegration, peri-implantitis, Prevotella intermedia, rabbit, removal torque assay