Abstract 15508: Incidence and Impact of Atrial Fibrillation in Critically Ill Septic Patients

IntroductionAtrial fibrillation (AF) can be triggered by several etiologies including an acute illness. The short and long-term implications of AF in this setting are poorly studied. Our aim was to identify the incidence and impact of Atrial Fibrillation (AF) in critically ill septic patients.MethodsWe conducted a single-center retrospective cohort study of patients admitted with sepsis, septicemia, severe sepsis, and septic shock (ICD-9 codes9959.1, 038, 9959.2, 785.0) between 2001 and 2012 utilizing the Medical Information Mart for Intensive Care (MIMIC) database. After reviewing clinical documentation, patients were divided into three groupsseptic patients who did not have concomitant AF (control), septic patients with new onset AF and no history of AF (new onset), and patients with AF and known history of AF (recurrent). In-hospital mortality, length of stay (LOS), and stroke incidence were the primary outcomes of interest. Univariate and multivariate analyses (i.e. logistic regression for mortality endpoint and Cox proportional hazards regression) were performed after adjusting for baseline and clinical characteristics.ResultsThere was a total of 5,037 ICU admissions for sepsis, of which 978 (19%) had AF (recurrent AF865, new onset AF113). The mean age of all patients was 69 (±13) of which 41% were females. Patients with new onset AF had a significantly longer average length of stay—13.1 days relative to 11.4 days and 11.9 days in recurrent AF and control patients, respectively (p=0.03). In-hospital mortality was 22.1% for septic patients without AF, 21.2% in septic patients with new AF and 18.7% in patients with recurrent AF (p=0.08). In a survival analysis (of in-hospital mortality), there was no significant difference between atrial fibrillation status (log rank p=0.43, see figure 1). Multivariate analyses of both mortality and survival revealed similar findings. The rate of embolic stroke during admission was