Abstract: Objective To analyze the prevalence and main risk factors of secondary pleural effusion (PE) in patients with severe acute pancreatitis (SAP),then construct a quantitative risk nomogram predictive model to guide clinical practice.Methods A retrospective summary was conducted on 382 SAP patients into our hospital from July 2017 to July 2024.They were randomly divided into model group (n=255) and validation group (n=127) at 2∶1.The patients in model group were divided into PE group of 58 cases and non-PE group of 197 cases according to whether PE occurred during hospitalization.The clinical datas and clinical laboratory tests between the two groups were compared,then LASSO regression and multivariate Logistic regression analysis were used to screen the risk factors of PE and construct the nomogram model.The ROC was used for evaluation model prediction effect.Results It showed that the level of fibrinogen (FIB),Creactive protein (CRP),and blood amylase in the PE group were significantly higher than those in the non-PE group,while albumin (ALB) and fasting blood glucose (FBG) were significantly lower(P<0.05).The Logistic regression model showed that the acute physiology and chronic health assessment (APACHE Ⅱ) score,the sequential organ failure assessment (SOFA) score,FIB and CRP were the risk factors for secondary PE in SAP patients,while albumin was the protective factor.R software was used to establish the nomogram model,with total score of 200 points.ROC showed that the area under the curve (AUC) of the model for predicting PE in model and validation groups were 0.878 (95%CI：0.823~0.934,P<0.001) and 0.843(95%CI 0.801~0.922,P＜0.001),indicating that the model had good diagnostic performance.The calibration curve overlapped with the standard curve without obvious deflection,indicating that the model has good consistency.The decision curve analysis suggested that the model has potential clinical application value.Conclusion Higher of admission APACHE Ⅱ and SOFA scores,serum FIB and CRP levels,and lower of albumin suggest an increasing risk of PE;constructing a user-friendly nomogram model has great potential for guiding early and accurate identification of high-risk PE patients in clinical practice.

Key words: severe acute pancreatitis, pleural effusion, nomogram model, acute physiology and chronic health assessment, sequential organ failure assessment