Abstract
BACKGROUND: The Endothelial Activation and Stress Index (EASIX) is an emerging biomarker that serves as a straightforward and objective measure of systemic endothelial dysfunction and critical illness severity. This study aims to evaluate the prognostic value of EASIX for 28-day mortality in patients with pulmonary sepsis. MATERIALS AND METHODS: This retrospective study utilised a two-cohort design. The internal cohort was derived from MIMIC-IV; an external cohort was derived from a tertiary hospital (2022-2025). The association between the EASIX and 28-day mortality was evaluated using multivariable Cox regression, restricted cubic spline (RCS) analysis, and Kaplan-Meier survival curves. An ensemble machine-learning approach (Boruta, LASSO-COX, XGBoost, and SVM) was employed for feature selection. Significant predictors were incorporated into a multivariate Cox model to construct a prognostic nomogram. The model's discriminative performance was assessed using receiver operating characteristic (ROC) curves and the area under the curve (AUC), and compared against conventional severity scores. RESULTS: A total of 5,416 patients were analyzed. In multivariable adjusted models, the EASIX emerged as an independent predictor of short term mortality. Each unit increase in EASIX was associated with a 7% higher risk of 28-day ICU death (HR 1.07, 95% CI 1.05-1.11, p < 0.001). A clear dose-response relationship was observed across EASIX quartiles, with mortality rising from 13.29% (Q1) to 27.92% (Q4); patients in Q4 had nearly twice the mortality risk of those in Q1 (HR 1.99, 95% CI 1.60-2.46). RCS analysis revealed a nonlinear relationship. Machine-learning feature selection consistently identified EASIX as a core variable. The final prognostic model, integrating EASIX with five other clinical features, demonstrated stable and superior discriminative ability (AUC 0.67-0.73) compared to traditional severity scores in both internal and external validation. CONCLUSION: EASIX is a potent and independent predictor of short-term mortality in pulmonary sepsis. Its integration into a pragmatic prognostic model enhances early risk stratification, highlighting its potential as a readily available clinical tool.