Abstract
BACKGROUND: Early and convenient prediction of the severity of acute pancreatitis (AP) is crucial for its treatment and prognosis. This study aimed to develop and validate a nomogram model for assessing the risk of severe acute pancreatitis (SAP) based on the theory of Persistent Inflammation, Immunosuppression, and Catabolism Syndrome (PICS). METHODS: A total of 4,930 AP patients from the MIMIC-IV database were selected as the derivation cohort, which was divided into the SAP group (n = 975) and non-severe acute pancreatitis (NSAP) group (n = 3,955) according to the 2012 Atlanta classification criteria. The 9 hematological indicators collected at the earliest time point within 48-72 h of admission were subjected to logistic regression analysis, and the statistically significant indicators selected were used to establish the model. A validation cohort consisting of 233 AP patients (34 in the SAP group and 199 in the NSAP group) admitted to the Department of General Surgery, Anhui No.2 Provincial People's Hospital from January 2016 to October 2024 was used to verify the model's performance. RESULTS: Multivariate Logistic regression showed that neutrophil-lymphocyte ratio (NLR), systemic immune-inflammation index (SII), white blood cell count (WBC), hemoglobin (Hb), and red blood cell distribution width (RDW) were independent predictors of SAP (P < 0.05). The nomogram model equation was constructed as follows: logit(P) = ln(2.37)⋅ log(NLR) + ln(0.45)⋅ log(SII) + ln(2.60)⋅ log(WBC) + ln(0.85)⋅ Hb + ln(1.14)⋅ RDW. The area under the receiver operating characteristic curve (AUC) of the derivation cohort was 0.730 (95% CI: 0.708-0.743), with a Hosmer-Lemeshow test P-value of 0.333. The AUC of the validation cohort was 0.795 (95% CI: 0.703-0.886). CONCLUSION: The nomogram model based on NLR, SII, WBC, Hb, and RDW has good predictive value for SAP and can provide a convenient tool for early clinical identification of SAP.