AUTHOR=Yan Jia , Yilin Huang , Di Wu , Jie Wang , Hanyue Wang , Ya Liu , Jie Peng TITLE=A nomogram for predicting the risk of mortality in patients with acute pancreatitis and Gram-negative bacilli infection JOURNAL=Frontiers in Cellular and Infection Microbiology VOLUME=12 YEAR=2022 URL=https://www.frontiersin.org/journals/cellular-and-infection-microbiology/articles/10.3389/fcimb.2022.1032375 DOI=10.3389/fcimb.2022.1032375 ISSN=2235-2988 ABSTRACT=Objective

Gram-negative bacilli (GNB) are common pathogens of infection in severe acute pancreatitis (SAP), and their occurrence increases the mortality of SAP. Early identification of SAP severity and prognosis is of great significance to SAP treatment. This study explored risk factors for mortality in patients with SAP and GNB infection and established a model for early prediction of the risk of death in GNB-infected SAP patients.

Methods

Patients diagnosed with SAP from January 1, 2016, to March 31, 2022, were included, and their baseline clinical characteristics were collected. Univariate logistic regression analysis was performed to screen for death related variables, and concurrently, a Boruta analysis was performed to identify potentially important clinical features associated with mortality. The intersection of the two results was taken for further multivariate logistic regression analysis. A logistic regression model was constructed according to the independent risk factor of death and then visualized with a nomogram. The performance of the model was further validated in the training and validation cohort.

Results

A total of 151 patients with SAP developed GNB infections. Univariate logistic regression analysis identified 11 variables associated with mortality. The Boruta analysis identified 11 clinical features, and 4 out of 9 clinical variables: platelet counts (odds ratio [OR] 0.99, 95% confidence interval [CI] 0.99–1.00; p = 0.007), hemoglobin (OR 0.96, 95% CI 0.92–1; p = 0.037), septic shock (OR 6.33, 95% CI 1.12–43.47; p = 0.044), and carbapenem resistance (OR 7.99, 95% CI 1.66–52.37; p = 0.016), shared by both analyses were further selected as independent risk factors by multivariate logistic regression analysis. A nomogram was used to visualize the model. The model demonstrated good performance in both training and validation cohorts with recognition sensitivity and specificity of 96% and 80% in the training cohort and 92.8% and 75% in the validation cohort, respectively.

Conclusion

The nomogram can accurately predict the mortality risk of patients with SAP and GNB infection. The clinical application of this model allows early identification of the severity and prognosis for patients with SAP and GNB infection and identification of patients requiring urgent management thus allowing rationalization of treatment options and improvements in clinical outcomes.