The objective of this study was to set up anin vivogentamicin susceptibility test for biofilm prevention in bone tissue and on implants. Twenty-five pigs were allocated to six groups. Pigs in group A (n= 6) were inoculated with saline. Pigs in groups B (n= 6), C (n= 3), D (n= 3), E (n= 3), and F (n= 4) were inoculated with 10 μl saline containing 104CFU ofStaphylococcus aureus. Different concentrations based on the MIC of gentamicin for the specific strain were added to the 10-μl inoculum for groups C (160× MIC), D (1,600× MIC), E (16,000× MIC), and F (160,000× MIC). The inocula were injected into a predrilled tibial implant cavity, followed by insertion of a steel implant (2 by 15 mm). The pigs were euthanized after 5 days.In vitro, all the doses used were found to be bactericidal after up to 6 h. All implant cavities of pigs inoculated with bacteria and bacteria plus 160× MIC or 1,600× MIC of gentamicin were positive forS. aureus. In animals in each of groups E (16,000× MIC) and F (160,000× MIC), 2/3 and 1/4 of the implant cavities wereS. aureuspositive, respectively. By grouping groups C and D (<10,000× MIC) and groups E and F (>10,000× MIC), a significant decrease in the number of implant-attached bacteria was seen only between the high-MIC-value group and group B. Histologically, it was demonstrated that 1,600×, 16,000×, and 160,000× MIC resulted in a peri-implant tissue reaction comparable to that in saline-inoculated animals.In vivo, the antimicrobial tolerance of the inoculated planktonic bacteria was increased byin vivo-specific factors of acute inflammation. This resulted in bacterial aggregation and biofilm formation, which further increased the gentamicin tolerance. Thus, susceptibility patternsin vitromight not reflect the actualin vivosusceptibility locally within a developing infectious area.