A routine blood parameter-based nomogram to predict prolonged fever in children with primary EBV infection

This study aimed to develop a clinical nomogram using routine blood parameters to identify the risk of prolonged fever (>1 week) in children with acute primary Epstein-Barr virus (EBV) infection.

This retrospective study involved 791 hospitalized children with primary EBV infection. Patients were divided into two groups: fever duration ≤1 week and >1 week. Candidate routine blood parameters were screened through univariate analysis. Least Absolute Shrinkage and Selection Operator (LASSO) regression was used for variable selection to mitigate overfitting. A multivariate logistic regression model was then constructed using the optimal predictors and visualized as a nomogram. Model performance was rigorously quantified through the area under the receiver operating characteristic curve (AUC), calibration plots, and decision curve analysis (DCA), with internal validation performed using 1,000 bootstrap resamples.

Prolonged fever affected 27.4% of the cohort, and antiviral therapy was administered to all these patients (100%), significantly higher than the 56.1% in the short fever group (P < 0.001). LASSO regression identified white blood cell count (WBC), neutrophil count (NEU), platelet count (PLT), and atypical lymphocyte proportion (AL) as optimal predictors. The nomogram including these four parameters predicted prolonged fever with an AUC of 0.728 (95% CI: 0.689-0.768). Calibration and DCA confirmed the nomogram's clinical utility.

A nomogram based on four routine blood parameters (WBC, NEU, PLT and AL) provides a practical method for the early risk stratification of children at high risk for prolonged EBV-related fever. This model serves as an initial reference, though its clinical value requires validation through multicenter prospective studies.