Abstract
Sepsis-associated encephalopathy (SAE) is a significant clinical challenge in sepsis patients, contributing to prolonged hospitalization and increased mortality. The dysregulated immune response and neuroinflammatory processes have been implicated in its pathogenesis. This study aimed to develop a predictive model for SAE based on a combination of immunological markers. A retrospective study was conducted at a single center from January 2020 to December 2023, involving 98 sepsis patients. Clinical, laboratory, and immunological parameters were analyzed. Statistical analyses included chi-square tests, Wilcoxon rank-sum tests, logistic regression to identify independent risk factors, and receiver operating characteristic curve analysis to evaluate the predictive performance of the combined model. The encephalopathy group had a significantly higher prevalence of pulmonary disease, vasopressor use, and elevated body temperature. Key laboratory findings included significantly lower levels of interferon-gamma (IFN-γ) (P < .001), tumor necrosis factor-alpha (TNF-α) (P < .001), and a decreased CD4+/CD8 + ratio (P = .001) compared to the non-encephalopathy group. The logistic regression model confirmed these immunological markers as independent risk factors. The combined model of IFN-γ, TNF-α, and the CD4+/CD8 + ratio demonstrated a high predictive value with an AUC of 0.845. The combination of IFN-γ, TNF-α, and CD4+/CD8 + ratio provides a validated model for predicting SAE. These findings suggest that incorporating these immunological markers could improve risk stratification and early intervention. Future prospective, multicenter studies are recommended to validate this model for routine clinical practice.