Development and validation of a nomogram to predict severe influenza

Abstract

Background

Influenza is an acute respiratory disease posing significant harm to human health. Early prediction and intervention in patients at risk of developing severe influenza can significantly decrease mortality.

Method

A comprehensive analysis of 146 patients with influenza was conducted using the Gene Expression Omnibus (GEO) database. We assessed the relationship between severe influenza and patients' clinical information and molecular characteristics. First, the variables of differentially expressed genes were selected using R software. Least absolute shrinkage and selection operator (LASSO) and multivariate logistic regression analysis were performed to investigate the association between clinical information and molecular characteristics and severe influenza. A nomogram was developed to predict the presence of severe influenza. At the same time, the concordance index (C-index) is adopted area under the receiver operating characteristic (ROC), area under the curve (AUC), decision curve analysis (DCA), and calibration curve to evaluate the predictive ability of the model and its clinical application.

Results

Severe influenza was identified in 47 of 146 patients (32.20%) and was significantly related to age and duration of illness. Multivariate logistic regression demonstrated significant correlations between severe influenza and myloperoxidase (MPO) level, haptoglobin (HP) level, and duration of illness. A nomogram was formulated based on MPO level, HP level, and duration of illness. This model produced a C-index of 0.904 and AUC of 0.904.

Conclusions

A nomogram based on the expression levels of MPO, HP, and duration of illness is an efficient model for the early identification of patients with severe influenza. These results will be useful in guiding prevention and treatment for severe influenza disease.