P54 Effects of universal early antimicrobial treatment in acute liver failure patients at a large volume transplant centre in the United Kingdom

<h3></h3> We assessed clinical consequences of universal early antimicrobial treatment in critically ill acute liver failure (ALF) patients. A retrospective observational study covering 23 months of ALF admissions. Patients admitted to ICU with ALF had their electronic patient records reviewed for microbiologic, radiologic and blood biomarker evidence of infection and resistance. Systemic inflammatory response syndrome (SIRS), non-neurologic sequential organ failure (SOFA) scores, clinical history and antimicrobial prescriptions were reviewed, including documented reasons for prescription. Antimicrobial duration as a proportion of ICU length of stay (LoS) was calculated. Data are expressed as median (interquartile range). There were 37 ALF admissions with an ICU LoS of 9 (6–16) days, for review. In line with local guidelines, all patients received antimicrobials on admission and patients were prescribed fluconazole (100%) alongside either Piperacillin/tazobactam (35/37; 94.6%) or meropenem (2/37; 5.4%). On admission, 10/37 (27%) patients already had some evidence of infection (9 radiologic and 1urine culture) – however, pre-ICU LoS was not significantly different between infected and non-infected groups. SOFA scores were significantly higher in infected patients (12 (8.5 – 14) vs 8 (7 – 10)) but C-reactive protein (11 (6–25) vs 11 (3–14)), white cell count (14.5 (9.5–19.5) vs 11.5 (7.5–15)) and SIRS scores (2 (1–2) vs 1 (1–2)) were not. Admission blood cultures were negative in all but were only taken in 20% (2/10) infected patients and 30% (8/27) non-infected. Despite universal antimicrobial treatment from the outset, 18/37 (49%) patients required further antimicrobial escalation/extension for presumed infection during their stay and median ICU antibiotic-free days indexed to ICU LoS was 0% (0–26.5). 3 resistant organisms were found in 2/37 (5%) patients (multi-drug resistant (MDR) pseudomonas, MDR staphylococcus haemolyticus and carbapenemase producing enterococci). Mortality rate (6/37 (16%)) and transplantation rate (19/37 (51%)) didn’t differ according to presence/absence of admission infections, but 2 patients meeting Kings College Criteria with no contra-indications to transplantation succumbed to uncontrolled sepsis. Admission infections were found in nearly 3 in every 10 ALF admissions. Although SOFA scores were significantly higher in patients with admission infections, overlap was high and the poor discrimination of traditional infection markers such as CRP, WCC and SIRS scores demonstrates the difficulty of identifying sepsis in this group. Sepsis remains a feared driver of avoidable ALF deaths, though the association of universal antimicrobial treatment with few antibiotic-free days, antimicrobial escalation in over half of admissions and the development of resistant organisms highlights the potential harms of over-treatment.