A retrospective study of sepsis-associated encephalopathy: epidemiology, clinical features and adverse outcomes

Sepsis-associated encephalopathy (SAE) is a common complication of patients with sepsis and can manifest as mild disturbance of consciousness, disorientation, cognitive impairment, convulsion or deep coma [1, 2]. Importantly, SAE can result in dramatically poorer outcomes in patients with sepsis, with the mortality increasing with a rising SAE severity to a maximum of 70% [3]. Surviving SAE patients are likely to exhibit prolonged or permanent side effects, including altered behaviour, cognitive impairment, reduced quality of life, or premature death [4]. To date, the diagnostic criteria and potential risk factors for SAE remain incompletely understood. Early identification, timely diagnosis and effective management may be important for the disease control in sepsis patients. The present study was therefore designed to explore the epidemiology, clinical characteristics and risk factors for SAE, as well as its adverse outcomes, via a retrospective study of 291 patients with sepsis.

The occurrence of SAE is one of the main manifestations of organ dysfunction caused by sepsis, excluding clinical or laboratory evidence of a central nervous system infection, a structural abnormality, or another encephalopathy (such as hepatic encephalopathy or uraemic encephalopathy); SAE refers to a diffuse brain dysfunction resulting from sepsis and is mainly exhibited as delirium, cognitive impairment, decreased learning and memory ability, coma, twitch and so on [3, 6]. The mechanism may involve the dysfunction of cerebral microvascular cells, the loss of blood-brain barrier integrity, mitochondria dysfunction, the activation of microglia and astrocytes, and neuronal death [7, 8]. Currently, the diagnostic criteria and potential risk factors for SAE remain incompletely understood, with no reliable means of clinically evaluating sepsis-associated neurological dysfunction [9]. In the present study, subjective evaluations (performed by clinical staff) and objective indicators (GCS < 15, CAM-ICU for delirium) were combined as the inclusion criteria of SAE patients, and we applied comprehensive exclusion criteria to exclude disturbances due to central nervous system infection, cerebrovascular disease, metabolic disease, sedative-related cognitive effects, toxicosis, etc. More objective indexes should be used in future studies to identify SAE accurately, including craniocerebral ultrasound and bedside electroencephalogram (EEG) monitoring, to detect SAE as early as possible by evaluating cerebral blood flow velocity and changes in EEG activity, especially to predict delirium in critical sepsis patients [4]. The typical changes of EEG in SAE patients include an excessive θ rhythm, an increased δ rhythm, a three-phase wave and burst suppression [10]. In addition, SAE patients exhibit increased brain injury biomarkers (i.e., neuron-specific enolase, S-100 beta-protein) and neuroradiological abnormalities [11‐13].

In recent decades, sepsis and SAE have been the focus of intensive medicine research, with high morbidity and mortality. The recently reported 28-day mortality and 180-day mortality rates of SAE were 45.95 and 55.41%, respectively [2]. In the present study, 43.6% of patients with sepsis were identified as having SAE, exhibiting a 42.5% of 28-day mortality, which was significantly higher than that of non-SAE patients. However, there was no significant difference in the hospitalization time, suggesting that encephalopathy resulting from sepsis may have a greater influence on long-term outcomes, including long-term cognitive dysfunction and functional disability, which seriously decrease the quality of life and bring a great burden to patients, their families and even society [14]. In addition, it should be noted that the choice of hospital LOS as an outcome indicator may result in statistical bias. The shortening of LOS may be due to the critical condition of patients and who succumbed during early stages/initial phase of resuscitation, rather than the result of effective treatment.

Elderly patients exhibit a higher risk and mortality from sepsis [15]. In the present study, we also found the age was an independent risk factor for 28-day mortality of SAE patients. In addition, severe patients with underlying diseases often progress more rapidly and have poor prognosis. Especially when the underlying diseases are hypertension, diabetes, acute renal injury, and chronic obstructive pulmonary disease (COPD), such patients might be more likely to develop central nervous system complications [2, 16‐18]. We found that patients with sepsis who had underlying hypertension had a higher risk of progression to SAE. Moreover, we found that the SOFA score and APACHE II score were not only independent risk factors for SAE in sepsis patients but also independent risk factors for 28-day mortality in SAE patients. This further confirms the advantages of SOFA scores and APACHE II scores in assessing the severity and prognosis of critical patients [19, 20] and suggests that the underlying health status and severity of sepsis patients are closely related to the occurrence of encephalopathy and adverse outcomes. A significantly decreased platelet count in SAE patients was also found in the present study, which suggested that platelets participate in immune and inflammatory responses against various pathogens, apart from their important role in the coagulation process [21].

It is interesting in the present study that the incidence of gastrointestinal infection and the rate of detection of Enterococcus were increased significantly in sepsis patients with encephalopathy, which suggests that the gut or gut microbiota might participate in the pathogenesis of SAE. The reported studies have shown that changes in gut homeostasis and the gut microbiome could influence the physiology, behaviour and cognitive function related to the brain, and the interaction between gut microbiota and the brain has gradually become a hot topic in neuroscience; this interaction may ensure the maintenance of gut homeostasis but also exhibit multiple effects on emotion, motivation and cognitive function [22, 23]. This complex interaction between the gut and brain has been named the “gut-brain axis (GBA)”, which plays an important role in monitoring and integrating intestinal function and linking the emotional and cognitive centres of the brain with peripheral intestinal functions and mechanisms [24]. It has been confirmed that the gut microbiome is involved in the pathophysiological development process of neurodegenerative diseases, stroke, emotional and affective disorders and other central nervous system diseases or encephalopathy [25‐27]. Patients with sepsis that is directly influenced by gastrointestinal infection or antibiotic administration [28] may develop disturbances of the gut microbiota. Animal experiments have found that faecal bacteria transplantation can change the gut microbiota of septic mice, suggesting that faecal bacteria transplantation and vagal nerve block are potential therapeutic targets in SAE [29]. The effects of gut microbiota disturbance and the GBA on SAE development deserve further study, especially with respect to upstream and downstream initiation mechanisms, which may provide more ideas to explore the pathogenesis and early diagnosis of SAE.

In the present retrospective study, we found that nearly half of patients with sepsis developed SAE, which was closely related to poor outcomes. The identified risk factors for SAE included older age, underlying hypertension, gastrointestinal infection, enterococcal infection, and decreased platelet count. Both the SOFA score and APACHE II score were independent risk factors for predicting the occurrence and adverse outcome of SAE. Future studies should be focused on investigating the long-term prognosis and cognitive function of SAE patients and the upstream initiation mechanism of SAE.