A study on early prediction of infected pancreatic necrosis in patients with severe acute pancreatitis        LIU Xing*, HUANG Li, LONG Ling-xian, et al.*Department of Intensive Care Medicine, Hunan Provincial People’s Hospital (The First-Affiliated Hospital of Hunan Normal University), Changsha 410005, China
Corresponding author： HUANG Li,E-mail:331738525@qq.com
Abstract    Objective    To explore the early warning signs of infected pancreatic necrosis（IPN）secondary to severe acute pancreatitis（SAP）. Methods    Clinical data of SAP in Hunan Provincial People’s Hospital from 2012 to 2019 were retrospectively analyzed. The demographic characteristics , past medical history， etiology, organ failure, library data in 48 h, and CTSI score at 72 hours after onset were collected. The risk factors of IPN were analyzed by multivariate binary Logistic regression analysis, the ROC curve was used to compare the predictive power of the Logistic regression model, CTSI, BISAP, APACHE Ⅱ, and SOFA scores, and the cutoff value of each score for predicting IPN was analyzed. Results    There were 277 SAP patients,74 in the IPN group and 203 in the control group. Logistic regression analysis showed that delirium（OR=4.049,95%CI 1.058～15.500, P=0.041）was a risk factor for IPN, but hematocrit（Hct）（OR=0.939,95%CI 0.897～0.983, P=0.007）was a protective factor for IPN. The areas under the ROC curve of the Logistic regression model, CTSI, BISAP, APACHE Ⅱ, and SOFA were 0.719,0.702,0.622,0.593, and 0.561 respectively. The cutoff values of IPN predicted were CTSI ≥ 7 points, BISAP ≥ 3 points, APACHE Ⅱ≥ 14 points, and SOFA ≥ 5 points respectively. Conclusion    Early delirium and decreased Hct can predict IPN in the later stage of SAP, and it is recommended to transfer high-risk patients with IPN in time.