644

Introduction: Brain dysfunction is common during sepsis and is associated with increased mortality. These observations have been limited by potential confounding factors and the lack of evaluation over time. Hypothesis: To investigate the prognostic value of brain dysfunction as well as changes in neurological status over time in septic patients. Methods: Analysis of the Sepsis Occurrence in Acutely Ill Patients (SOAP) database. All data were collected until death, hospital discharge or for 60 days. Central nervous system (CNS) status was assessed using the neurological SOFA sub-score (nSOFA): CNS failure was defined as nSOFA > 2. Early failure was considered if occurred within 2 days since admission. Changes in nSOFA (?nSOFA) were calculated as the difference in nSOFA on day 3 (or day 2 if data on day 3 were not available) and nSOFA on admission. Brain improvement (BI) was defined as?nSOFA < 0 for initial nSOFA of 1 and 2, and nSOFA < 3 on day 3 for those with initial nSOFA 3 and 4. Patients without brain dysfunction had nSOFA of 0 during the first 3 ICU days. All the others were considered as having brain deterioration (BD). Results: After exclusion of 776 patients who presented with primary brain injury, we analysed the data of 2371 patients (904 with sepsis). In septic patients, ICU mortality was associated with a higher nSOFA score and early CNS failure; however, in a multivariable analysis, these factors were not independently predictors of poor outcome. When analyzing septic and non-septic patients, sepsis had an increased odds ratio (OR) for mortality of 1.21 [1.18-1.25], CNS failure of 1.45 [1.24-1.68] and sepsis with CNS failure of 1.76 [1.46-2.12]. nSOFA changes yielded 213 patients with BI and 278 with BD. After adjustment for confounding factors, sepsis had an increased odds ratio (OR) for ICU mortality of 1.42 [1.34-1.59], BD of 1.98 [1.79-2.18]. The presence of sepsis and BD increased the OR for ICU mortality to 2.80 [2.56-3.06]. Conclusions: Brain dysfunction is not independently associated with mortality in septic patients, but it has a greater impact on outcome in septic than in non-septic patients, even after correction for confounding factors.