Unraveling the influencing factors and predictive model for high - level gastrointestinal injury complicating moderate - severe acute pancreatitis

Abstract Objective To investigate the influencing factors for Moderately Severe Acute Pancreatitis (MSAP) complicated with high-grade Acute Gastrointestinal Injury (AGI ≥ grade 3), and to establish and evaluate a predictive model for this condition. Methods A retrospective analysis was conducted on 228 patients with MSAP admitted to Dongyang People’s Hospital, Wenzhou Medical University, China. between July1, 2022 and July 1,2025. Among them, 85 patients with high-grade AGI constituted the observation group, and 143 patients without high-grade AGI served as the control group. A three-stage variable screening strategy based on LASSO regression was employed to identify core predictors. A predictive model was then constructed using binary logistic regression and internally validated via the bootstrap method. Results Hyperlipidemia was the most common etiology of MSAP. Multivariate analysis identified hyperlipidemic etiology, plasma D-dimer level, and length of hospital stay as independent risk factors for high-grade AGI, while age at diagnosis, serum potassium (K+), and serum calcium (Ca2+) were protective factors. The final model demonstrated an area under the receiver operating characteristic curve (AUC) of 0.877, which increased to 0.888 after bootstrap validation. The model's sensitivity was 80.3% and specificity was 81.5%. Conclusion Hyperlipidemic etiology, plasma D-dimer, and length of hospital stay are independent risk factors for high-grade AGI complicating MSAP, whereas K⁺, Ca²⁺, and age at diagnosis are protective factors for high-grade AGI complicating MSAP.