A novel in vivo model of prosthetic joint infection: application to the evaluation of treatment efficacy and of antibiotic pharmacokinetic

Background: Periprosthetic joint infections (PJI) are catastrophic complications of joint replacement surgeries. Current treatment options are unsatisfactory due to poor functional results following staged-replacements and high failure rates after debridement, implant retention and antibiotic therapy (DAIR). New treatment options have been explored in-vitro, and require the development of specific in-vivo models mimicking PJI. Here we describe a novel model of infected knee arthroplasty in New-Zealand White (NZW) rabbits and examined it suitability to evaluate antibiotic pharmacokinetics or treatment efficacy, focusing on a standard DAIR procedure. Materials: 15 NZW rabbits were assigned to three groups: vancomycin pharmacokinetics (A), infection (B), and DAIR (C). All rabbits received a unicompartimental tibial arthroplasty using a custom 3D-printed TI-6Al-4V implant. The joint of groups B and C rabbits were inoculated with 5 log10 CFU of ATCC33591 MRSA injected in the knee at the end of the surgery. All rabbits benefitted from a post-operative analgesia with transcutaneous fentanyl. One week after the initial surgery, the joint space of group C rabbits were surgically debrided. Rabbits of groups A and C received i.v. injections of vancomycin 50mg/kg b.i.d. for one week. Serum concentrations of vancomycin were measured in group A rabbits after the first and fifth doses. Rabbits were euthanized two weeks after the initial surgery. Bacterial counts from joint tissues and the implant were obtained in groups B and C rabbits. Results: Vancomycin pharmacokinetics: The average AUC0-12 was of 213.0mg*h/L for the first dose FIG.I) and 207.8mg*h/L for the fifth dose (FIG.II), with complete elimination of the drug in 12h. As the MIC for vancomycin of ATCC33591 is 1mg/L, the AUC0-24/MIC ratio is above 400, i.e. the usual pharmacokinetic/pharmacodynamic target against staphylococci. Infection and DAIR (FIG.III): All animals of the B and C groups were infected. CFU counts were homogeneously distributed in all tissues in group B and bacterial loads were significantly reduced in all samples except for the articular capsule in group C. Conclusion: We developed a reproducible model of infected knee arthroplasty treated with debridement and vancomycin at therapeutic concentrations that is suitable to evaluate new adjuvant therapies to DAIR