[Sepsis associated encephalopathy is an independently risk factor for nosocomial coma in patients with supratentorial intracerebral hemorrhage: a retrospective cohort study of 261 patients].

Objective To investigate whether the presence of sepsis associated encephalopathy (SAE) would predict nosocomial coma (NC) and poor outcome in patients with supratentorial intracerebral hemorrhage (SICH). Methods A retrospective cohort study was conducted. The adult acute SICH patients with or without coma admitted to intensive care unit (ICU) of Shuyang People' Hospital Affiliated to Xuzhou Medical University from December 2012 to December 2015 were enrolled. Brain computed tomography (CT) scans were analyzed and the patients were divided into pre-hospital coma (PC) and NC groups. The clinical data and the incidence of SAE of patients in two groups were compared, and the 30-day prognosis was followed up. Univariate and Cox regression analyses were performed to analyze whether SAE would predict NC and poor outcome in patients with SICH. Results A total of 330 patients with acute SICH and coma were enrolled, excluding 60 cases of infratentorial cerebral hemorrhage, 3 cases of primary intraventricular hemorrhage, and 6 cases of unknown volume hematoma. Finally, 261 patients were included, with 111 patients of NC events, and 150 patients of PC events. 69 (62.2%) SAE in SICH with NC and 33 (22.2%) SAE in SICH with PC was diagnosed, and the incidence of SAE between two groups was statistically significant (P < 0.01). Compared with PC group, SICH patients in the NC group had lower incidence of hypertension (81.1% vs. 96.0%), longer time from onset to NC [days: 2.3 (23.9) vs. 0 (0.5)] and length of ICU stay [days: 5.0 (34.0) vs. 3.0 (12.0)], higher initial Glasgow coma score (GCS, 10.2±1.5 vs. 6.6±1.6) and sequential organ failure assessment (SOFA) score [4.0 (6.0) vs. 3.0 (3.0)], lower initial National Institutes of Health Stroke Scale (NIHSS) score (19.4±6.6 vs. 30.2±6.8), as well as more frequent sepsis (78.4% vs. 38.0%), vegetative state (24.3% vs. 14.0%), acute respiratory failure (24.3% vs. 10.0%), pneumonia (37.8% vs. 24.0%), septic shock (8.1% vs. 0), acute liver failure (5.4% vs. 0), hypernatremia (8.1% vs. 0), CT indicating that more frequent vasogenic edema (64.9% vs. 16.0%) and white matter lesion (13.5% vs. 2.0%), and less mannitol usage (94.6% vs. 100.0%), and less brain midline shift (32.4% vs. 68.0%) and hematoma enlargement (8.1% vs. 30.0%), less hematoma volume (mL: 28.0±18.8 vs. 38.3±24.4) in CT, and higher 30-day mortality (54.1% vs. 26.0%) with statistical differences (all P < 0.05). It was shown by Cox regression analyses that SAE [hazard ratio (HR) = 3.5, 95% confidence interval (95%CI) = 1.346-6.765, P = 0.000] and SOFA score (HR = 1.8, 95%CI = 1.073-1.756, P = 0.008) were independent risk factors of death of SICH patients with NC, and hematoma enlargement was independent risk factor of death of SICH patients with PC (HR = 3.0, 95%CI = 1.313-5.814, P = 0.000). Conclusion SAE is the independent factor of inducing NC event and poor prognosis in SICH patients. Key words: Sepsis associated encephalopathy; Supratentorial intracerebral hemorrhage; Prehospital coma Nosocomial coma; Systemic inflammatory response syndrome; Prognosis