We studied systemic ceftriaxone, and free/local tobramycin and doxycycline in a controlled rat model representing a generic acute exogenous joint infection. We hypothesized that evidence of infection (quantitative colony forming units [CFU], qualitative scanning electron microscopy [SEM], histopathology) (1a) would be reduced with local versus systemic antibiotic, (1b) any antibiotic would be superior to control, (2) there would be a difference among antibiotics, and (3) antibiotic would not be detectable in serum at 4-week euthanasia. Study groups included infected and noninfected (1) control (no treatment), (2) systemic ceftriaxone (daily), (3) local tobramycin, and (4) local doxycycline (10 rats/group; power = 0.8). With IACUC approval, a reliable acute exogenous joint infection was created by slowly injecting 50- μl, 10(4) CFU Staphylococcus aureus, into the distal femoral medullary canal. The antibiotic formulation was introduced locally to the femoral canal and joint space. After 4 weeks, serum, pin, bone, and synovium were obtained. CFU/ml of bone and synovium were quantified using macrotiter method. SEM imaged biofilm on the surface of the pin, histopathology identified tissue response, liquid chromatography/mass spectrometry quantified plasma antibiotic. (1) Groups receiving any antibiotic reported lower CFU/ml in synovium compared with no treatment. (2) In the synovium, free/local tobramycin reduced CFU/ml to a greater extent than free/local doxycycline (p  < 0.05). (3) Antibiotic in plasma after the local application was nondetectable in all groups after 4 weeks. SEM revealed no difference in biofilm on pin among all groups.