Scoring systems and biomarkers for early prediction of severe acute pancreatitis: a retrospective study

Objective. To determine prognostic value of laboratory tests and clinical scoring systems at admission to ICU in early stage of disease for predicting severity of acute pancreatitis. Material and methods. A single-center retrospective study included ICU patients with moderate-to-severe acute pancreatitis in 2022—2024. Patients were stratified into groups depending on disease severity. We evaluated APACHE II, BISAP, and SOFA scores, as well as laboratory biomarkers and composite indices. Analysis included ROC curve with AUC, multivariate logistic regression analysis and identification of predictors of severe disease. Results. Final analysis included data on 142 eligible patients divided into the moderate pancreatitis group (n=58) and severe pancreatitis group (n=84). In multivariate analysis, SOFA and BISAP scores, along with Red Cell Distribution Width — Coefficient of Variation (RDW-CV), were identified as independent risk factors of severe disease. Combination of SOFA and RDW-CV demonstrated the highest predictive value (AUC=0.844, sensitivity 71.4%, specificity 93.1%). Conclusion. SOFA score and RDW-CV may be useful for early identification of patients with high risk of severe acute pancreatitis and optimization of treatment.