A risk prediction model based on immune-inflammatory-nutritional indicators for predicting 28-day mortality in sepsis patients with acute respiratory distress syndrome

Qi Xin^*Xingbo DangGongliang DuYanlong YangHaitao Jing

• Shaanxi Provincial People's Hospital, Xi'an, China

Background: Sepsis is a life-threatening condition often complicated by organ dysfunction and is associated with a high mortality rate. The dysregulation of immune response, inflammation, and nutritional status are critical factors contributing to its pathogenesis. This study aimed to develop a nomogram that integrates prognostic immune-inflammatory-nutritional indicators with other clinical information to predict 28-day mortality in sepsis patients with acute respiratory distress syndrome (ARDS). Methods: Clinical data from 635 adult sepsis patients with ARDS were obtained from Shaanxi Provincial People's Hospital and randomly divided into a training set (n=477) and a validation set (n=158). To identify predictors of 28-day mortality in sepsis patients with ARDS, univariate analysis and least absolute shrinkage and selection operator (LASSO) regression were utilized. Subsequently, a multivariate logistic regression model was constructed to identify independent predictors. A nomogram was then developed by integrating the selected indicators. The model's performance was assessed with respect to discrimination, calibration, and clinical utility through the use of the AUC, calibration plots, and decision curve analysis (DCA). Results: The independent predictors utilized for the construction of the nomogram included the albumin-alkaline phosphatase ratio (AAPR), albumin-bilirubin grade (ALBI), neutrophil– lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), prognostic nutritional index (PNI), systemic immune-inflammation index (SII), and lactate-albumin ratio (LAR). Notably, the nomogram exhibited superior predictive performance, with an AUC of 0.873 in the training set and 0.837 in the validation set, as compared to the SOFA score, which showed an AUC of 0.689 in the training set and 0.684 in the validation set, for predicting 28-day mortality in sepsis patients with ARDS. The calibration plots demonstrated excellent consistency. DCA confirmed the model's clinical utility, showing a positive net benefit across a wide range of clinically relevant threshold probabilities (approximately 10% to 70%), which supports its potential to guide clinical decision-making. Conclusion: We have successfully developed and validated a robust nomogram that integrates seven readily accessible immune-inflammatory-nutritional indicators. This model serves as an individualized and precise tool for predicting the 28-day mortality risk in sepsis patients with acute respiratory distress syndrome (ARDS), thereby potentially enhancing early risk stratification and informing clinical decision-making.