Ital J Pediatr. 2025 Sep 24;51(1):271. doi: 10.1186/s13052-025-02118-0.
ABSTRACT
BACKGROUND: Due to the non-specific symptoms of pulmonary tuberculosis (PTB) in children, the diagnosis of PTB in children is a major challenge for clinicians in the absence of microbiological confirmation. This study aims to construct a simple clinical prediction model for empiric diagnosis of PTB through careful clinical symptoms and medical history.
METHODS: Retrospective analysis of clinical data and laboratory data of children with PTB and community acquired pneumonia (CAP) diagnosed at Tianjin Children’s Hospital from January 2018 to October 2023. All patients were randomly divided into a 7:3 ratio into a modeling group and a validation group. The modeling group was used to perform logistic analysis to identify independent risk factors and construct a clinical prediction model for PTB in children. The validation group was used to further assess the clinical efficacy of the model.
RESULTS: A total of 434 children were included in this study. The modeling group included 305 patients (125 with PTB, 180 with CAP) and validation group included 129 patients (53 with PTB, 76 with CAP). Four variables including basic disease, tuberculosis contact history, maximum body temperature and weight loss were identified as potential predictors used for developing a nomogram. The nomogram showed a good diagnostic performance in the modeling group [area under the curve (AUC) (95% confidence interval (CI)), 0.810(0.759 ~ 0.860)]. The decision curve analysis (DCA) and calibration curve indicated that the clinical prediction model for pediatric PTB has good clinical practicality and accuracy. The validation group also showed good clinical efficacy [AUC (95%CI), 0.864(0.794 ~ 0.934)], indicating that the model is feasible and reproducible.
CONCLUSIONS: This study developed and validated a nomogram for predicting PTB in children. This nomogram represents good clinical performance and might be utilized clinically in the empirical diagnosis of PTB in children.
PMID:40993784 | DOI:10.1186/s13052-025-02118-0
