P003: Door-to-antibiotics and mortality for emergency department patients presenting with septic shock

Introduction: We examined our local sepsis patient population, and specifically our most vulnerable patients - those presenting to the emergency department (ED) in septic shock - for variables predictive of survival to hospital discharge. We applied the familiar ED paradigm of, “Door to,” to calculate the impact of time to antibiotics against patient survival to hospital discharge. Methods: Retrospective chart review of patients aged > = 18 years, presenting to tertiary care ED between 01 Nov 2014 and 31 Oct 2015. Patients determined to have sepsis if A) > = 2 SIRS criteria and ED suspicion of infection (ED acquisition of blood/urine cultures or antibiotic administration) and/or B) received ED or Hospital discharge diagnosis of sepsis (ICD-10 diagnostic codes A4xx and R65). Patients sub-classified with septic shock if A) triage SBP < = 90mmHg, B) triage MAP < = 65mmHg or C) serum lactate > = 4mmol/L. “Door Time” was defined as the earliest time recorded for the patient encounter, either the time the patient registered in the Emergency Department, or the triage time. A generalized linear model was performed with a binomial distribution using survival to discharge as the response variable. Age, sex, ED arrival method, time to antibiotics, ED serum lactate and ED serum glucose level were the predictor variables. Results: 13506 patient encounters met inclusion criteria (10980 unique patients). Linear regression of time to antibiotics against survival to hospital discharge failed to achieve statistical significance. Linear regression of the secondary outcome variables achieved statistical significance for age and serum lactate level. Per the model, as age increased by 1 year, the odds of dying prior to hospital discharge increased by 3.8% and as serum lactate increased by 1 mmol/L, odds of dying prior to hospital discharge increased by 11.1%. Conclusion: We found no association between time to antibiotic treatment and mortality. Causal relationships require randomized controlled trials, and this analysis contributes to clinical equipoise.