Objective To investigate the independent predictors of in-hospital mortality in patients with severe acute pancreatitis (SAP) admitted to the intensive care unit (ICU), and to develop and validate a nomogram prediction model for in-hospital mortality risk. Methods Clinical data of 184 SAP patients admitted to the ICU of Shenzhen People's Hospital from January 2014 to December 2023 were retrospectively collected. Univariate logistic regression model was used to screen variables associated with in-hospital mortality in ICU SAP patients, and multivariate logistic regression model was applied to identify independent predictors of their in-hospital mortality, based on which a nomogram prediction model was constructed. The discrimination of the model was evaluated using the receiver operating characteristic (ROC) curve, The calibration curve was used to assess the agreement between predicted and observed probabilities. Decision curve analysis and the clinical impact curve were employed to evaluate the model's clinical utility. Results Of the 184 ICU SAP patients included, 27 died during hospitalization and 157 survived, yielding an in-hospital mortality rate of 14.7%. Univariate logistic regression analysis results showed that body temperature, respiratory rate, heart rate, use of vasoactive drugs, mechanical ventilation, platelet count, red blood cell distribution width, antithrombin Ⅲ activity, and levels of hemoglobin, serum creatinine, blood urea nitrogen, total bilirubin, alanine aminotransferase, aspartate aminotransferase, interleukin‑6, procalcitonin, lactate, serum calcium, serum phosphorus, high‑density lipoprotein cholesterol, and total cholesterol were all associated with in‑hospital mortality in ICU SAP patients (all P<0.05). Multivariate logistic regression analysis results revealed that respiratory rate (OR=1.130, 95%CI: 1.023-1.247), hemoglobin level (OR=0.983, 95%CI: 0.966-0.999), and blood urea nitrogen level (OR=1.109, 95%CI: 1.048-1.174) were independent predictors of in‑hospital mortality in ICU SAP patients (all P<0.05). The nomogram prediction model constructed on the basis of the above factors had an area under the ROC curve of 0.781 (95%CI: 0.693-0.869), a sensitivity of 66.7%, and a specificity of 81.1%. The calibration curve showed good agreement between the predicted and observed probabilities (mean absolute error=0.077). Decision curve analysis indicated that the model provided a high clinical net benefit across a wide range of threshold probabilities. The clinical impact curve further confirmed the model's high efficacy in identifying high risk patients. Conclusion In-hospital mortality in ICU SAP patients is closely related to increased respiratory rate, elevated blood urea nitrogen level, and decreased hemoglobin level. The nomogram prediction model constructed based on the above indicators shows favorable predictive performance and clinical application value, and can provide a reference for the early clinical identification of high‑risk patients.

无