Antibiogram for periprosthetic infections a tool for better informed selection of empiric antibiotics for surgical site infections

Introduction: Periprosthetic infection after breast reconstruction is not an uncommon complication, with incidence up to 24%. These infections are often treated empirically without knowing the causative bacteria or its sensitivities to various antibiotics. Even if cultures are obtained, results may not be available for several days. Methods: A retrospective chart review of 553 patients at a single institution between January 2009 and July 2014 was performed, identifying patients who (1) underwent implant-based breast reconstruction and subsequently suffered an infection and (2) had cultures available with sensitivities. We reviewed patient demographics, implant characteristics, prophylactic intravenous antibiotics, oral antibiotic maintenance used, microbiologic details, and outcomes. The goal was to identify the most common causative bacteria, as well as their sensitivities to commonly used antibiotics, to help guide antibiotic decision-making. Results: Of the 553 patients who underwent implant-based reconstruction, 114 (20.6%) patients suffered periprosthetic infections. Of these patients, 32 (28.1%) patients (56 reconstructions, with 33 tissue expanders and 23 implants) had cultures performed revealing 43 bacterial species, with the most common being Staphylococcus aureus (23.2%) and Pseudomonas aeruginosa (26.8%). Ceftazi-dime and piperacillin/tazobactam were equally effective covering 100% of Pseudomonas, enteric, and atypical organisms (P = 1), whereas vancomycin covered 100% of gram-positive organisms. Trimethoprim/sulfamethoxazole covered 100% of S. aureus, whereas clindamycin only covered 71% of S. aureus (P = 0.03). Additionally, trimethoprim/sulfamethoxazole was better able to cover atypical and enteric organisms. Ciprofloxacin covered 71% of Pseudomonas compared with 56% for levofloxacin (P = 0.14). Interestingly, cephalexin, a common choice for perioperative prophylaxis, was highly ineffective for gram-positive species in patients who later returned with infections. Conclusions: This study supports the efficacy of current intravenous antibiotics protocols but questions the efficacy of both clindamycin and levofloxacin in empirically treating periprosthetic infections and cephalexin in providing effective perioperative prophylaxis against skin flora. Because bacterial sensitivities vary by location and patient population, this study encourages other centers to develop their own antibiogram specifically tailored to periprosthetic infections to improve antimicrobial decision making and potentially improve implant salvage.