Abstract
OBJECTIVE: To develop and internally validate an individualized nomogram integrating intestinal barrier-specific biomarkers and systemic clinical indicators to help assess intestinal barrier function and provide a reference for prognosis prediction in patients with severe sepsis. METHODS: Three hundred fifty-two patients with severe sepsis admitted between January 2022 and December 2024 were continuously enrolled and randomly divided into training (n = 246) and validation (n = 106) sets. Plasma samples and clinical data-including demographics, injury assessments, and initial laboratory indicators-were collected. Prognosis-related variables were identified via univariate analysis. LASSO regression was used for variable selection, and multivariate logistic regression identified independent predictors of poor prognosis. Model performance was evaluated using receiver operating characteristic (ROC) curves, calibration plots, and decision curve analysis (DCA) in both training and validation sets. RESULTS: Baseline characteristics did not differ significantly between sets (all P > 0.05). Multivariate analysis identified admission SOFA score, intestinal fatty acid-binding protein, D-lactate, procalcitonin, and blood lactate as independent risk factors for poor prognosis (all P < 0.05). The nomogram demonstrated good calibration and fit, with C-indexes of 0.771 (training) and 0.641 (validation), mean absolute errors of 0.026 and 0.043, and non-significant Hosmer-Lemeshow test results (P = 0.423 and P = 0.496, respectively). The AUCs were 0.771 (95% CI: 0.698-0.845) and 0.641 (95% CI: 0.512-0.770), with sensitivities of 0.672 and 0.462, and specificities of 0.804 and 0.800. CONCLUSION: The constructed nomogram, incorporating intestinal barrier biomarkers and systemic clinical indicators, can help assess intestinal barrier-related risk and provide a reference for predicting adverse outcomes in severe sepsis. It offers a valuable decision-support tool for early goal-directed intervention and demonstrates significant clinical translational potential.