C-853-07. Under Pressure: Burn Specific Antibiogram, Ecology, and Antibiotic Use Compared to Surgical-trauma and the Institution

Abstract Introduction Burn sepsis guidelines suggest a unit specific antibiogram for empiric antibiotic choices. Despite clinical observations regarding the ecology of burn infections, additional data are needed to justify unit specific antibiograms. Our objective was to create a burn (BICU) specific Gram-negative (GN) antibiogram, compare the relative frequency of pathogens to a surgical trauma ICU (STICU) and the institutional antibiogram. We further assessed extended spectrum beta-lactam (Ex-BL) use mix in BICU vs STICU and the institution overall. Methods Unit specific antibiograms and monthly antibiotic use patterns were collected at an ABA verified burn center over a 21 month period for BICU, STICU, and the institution overall. Escherichia spp, Klebsiella spp, Enterobacter spp, and P. aeruginosa counts were divided by overall GN counts to calculate rates of occurrence. Piperacillin-tazobactam (TZP), cefepime (FEP), and meropenem (MEM) are the primary Ex-BLs used with TZP being institutionally preferred. FEP and MEM were situationally restricted. Ex-BL mix was calculated by monthly antibiotic days / 1000 patient days with TZP as reference to normalize an intent to cover pathogens of interest. Difference in bacteria rate was analyzed by chi-squared and Wilcoxon-rank sum used to assess Ex-BL mix rates between BICU and STICU or institution. Results GN bacteria differed between BICU, STICU, and the institution, respectively: Enterobacter spp 28%, 27%, and 10%; P. aeruginosa 20%, 18%, and 12%; Klebsiella spp 19%, 26%, and 20% and Escherichia spp 16%, 29%, and 52%; p.01. This was driven by the difference between BICU and the institution (p.01) as opposed to STICU, p=.08. Median antibiotic use was different between BICU and STICU for TZP (118 vs 200, p.01), but not FEP (58 vs 37, p=.2) or MEM (40 vs 50, p=.07). FEP:TZP mix was different for BICU (45%) vs STICU (19%), p.01, not the institution (43%). MEM:TZP mix was not different between BICU, STICU, or institution; 27%, 25%, 30%; respectively, p=.5. The overall probability of empiric susceptibility given a GN result for BICU (FEP = 87%, TZP = 84%, MEM = 94%) closely resembled the overall institution (FEP = 84%, TZP = 88%, MEM = 96%) than STICU (FEP = 75%, TZP = 77%, MEM = 96%) due to resistance patterns despite differences in bacteria mix between BICU vs institution and similarities between BICU vs STICU. Conclusions BICU specific antibiogram confirmed striking differences in GN bacteria mix vs the institution. Yet empiric probabilities for empiric Ex-BL susceptibility in BICU resembled the institution, not STICU. The causal relationship between Ex-BL mix use and resistance requires further study. Applicability of Research to Practice A BICU specific antibiogram revealed the inadequacy of an institutional antibiogram. It is a useful tool that may improve empiric antibiotic choices and allow for surveillance of trends over time. Funding for the study N/A.