A nomogram for early prediction of pertussis-associated pneumonia in children: a retrospective clinical study.

Abstract

Background: Pertussis, an acute respiratory infection caused by Bordetella pertussis, has recently shown a global resurgence. Pneumonia represents one of its most frequent and severe complications. The present study aimed to identify risk factors for pertussis-associated pneumonia and to develop a predictive model for early identification of high-risk pediatric patients.

Methods: A retrospective analysis was performed on 123 children hospitalized with pertussis at the Children's Hospital Affiliated to Zhengzhou University between 1 June 2024 and 31 December 2024(training cohort). Using univariate analysis, least absolute shrinkage and selection operator (LASSO) and multivariate logistic regression analysis, identified the predictors for pertussis-associated pneumonia. An independent cohort enrolled between 1 January 2025 and 31 July 2025, was used for external validation cohort. Receiving operating characteristic curve (ROC), calibration plots, decision curve analysis (DCA), and clinical impact curves (CIC) were used to evaluate the performance of the model.

Results: We identified five independent predictors of pertussis-associated pneumonia: white blood cell count (WBC; OR = 1.112, 95% CI: 1.012-1.221), interleukin-6 (IL-6; OR = 1.153, 95% CI: 1.049-1.267), lymphocyte percentage (Lym%; OR = 1.045, 95% CI: 1.013-1.078), interleukin-17a (IL-17a; OR = 1.209, 95% CI: 1.024-1.427), and Co-infection (OR = 3.096, 95% CI: 1.260-7.610). The area under the receiver operating characteristic curve (AUC) was 0.824 (95% CI: 0.747-0.900) in the training cohort and 0.792 (95% CI: 0.718-0.865) in the external validation cohort. The Hosmer-Lemeshow test indicated good calibration (P = 0.590 in the training cohort; P = 0.735 in the external validation cohort). Decision curve analysis and clinical impact curves confirmed clinical utility.

Conclusion: A clinically useful nomogram model was developed for early prediction of pertussis-associated pneumonia in children. Application of this model may assist clinicians in promptly identifying high-risk patients and implementing timely interventions to improve outcomes.Notably, this study had a small sample size (n = 123) relative to the five predictors, and the external validation cohort was from the same center, which may limit the model's generalizability to heterogeneous clinical settings.