Bloodstream Infection in Moderately Severe and Severe Acute Pancreatitis: Microbiological Features and a Prediction Model

Abstract

Background: Prediction of BSI would contribute to the management of MSAP and SAP. This study aimed to investigate pathogens distribution of bloodstream infection (BSI) in moderately severe acute pancreatitis (MSAP) and severe acute pancreatitis (SAP) patients and develop a prediction model for BSI development. Methods: MSAP and SAP patients admit to West China Hospital from January 2012 to December 2018 were collected. Include patients were divided into BSI group and non-bloodstream infection (NBSI) group. Isolated pathogens and antibiotic resistance were recorded. The least absolute shrinkage and selection operator (LASSO) regression was used to select the risk factors for BSI. Logistic regression was used to develop a nomogram. Random forest was used to validate the stable ability of prediction factors. The areas under receiver operating characteristic (AUC) and calibration curves were plotted to evaluate the prediction performance. Findings: 399 BSI patients and 1155 NBSI patients were enrolled in this study. Of the 408 isolated pathogens, 226 (55.39%) were gram-negative bacteria. Seven predictors (intensive care unit stay, cardiovascular failure, percentage of neutrophil, percentage of lymphocyte, triglyceride, low-density lipoprotein, and albumin) were identified for BSI. The nomogram prediction model has 0.913 and 0.888 AUCs in training and validation sets. The calibration curves show the nomogram has a good consistency. Interpretation: Gram-negative bacteria were the major BSI pathogens. A new nomogram with a high predictive value based on selected factors was built. It can be conveniently used to predict BSI development and help clinicians rationally use antibiotics in MSAP and SAP patients. Funding: This work was supported by the 1.3.5 project for disciplines of excellence, West China Hospital, Sichuan University (ZYJC18027); Sichuan Science and Technology Program (2020YFS0263); and Sichuan Science and Technology Program (2021YFS0234). Declaration of Interest: The authors declare no potential conflicts of interest. Ethical Approval: The Ethics Committee of West China Hospital approved the study (No. 2021-675)