Development and external validation of a simple nomogram for predicting apnea in children hospitalized with bronchiolitis.

BACKGROUND

Apnea is one of the most life-threatening complications of bronchiolitis in children. This study aimed to determine early predictors of apnea in children hospitalized with bronchiolitis and develop a simple nomogram to identify patients at risk of apnea.

METHODS

This retrospective, observational study included children hospitalized with bronchiolitis in two hospitals in China. Demographic and clinical characteristics, laboratory results, pathogens, and pulmonary iconography results were recorded. A training cohort of 759 patients (one hospital) was used to identify early predictors of apnea during hospitalization. The least absolute shrinkage and selection operator (LASSO) regression analysis method was used to optimize variable selection. The nomogram was developed visually based on the variables selected by multivariable logistic regression analysis. Discrimination (concordance index, C-index), calibration, and decision curve analysis (DCA) were used to assess the model performance and clinical effectiveness.

RESULTS

A total of 1,372 children hospitalized with bronchiolitis were retrospectively evaluated, 133 (9.69%) of whom had apnea. Apnea was observed in 80 of the 759 patients with bronchiolitis in the training cohort and 53 of the 613 patients in the external validation cohort. Underlying diseases, feeding difficulties, tachypnea, retractions and pulmonary atelectasis in the training cohort were independent risk factors for apnea and were assembled into the nomogram. The nomogram exhibited good discrimination with a C-index of 0.883 (95% CI: 0.839-0.927) and good calibration. The DCA showed that the nomogram was clinically useful in estimating the net benefit to patients.

CONCLUSION

We developed a nomogram that is convenient to use and able to identify the individualized prediction of apnea risk in patients with bronchiolitis. These patients might benefit from early triage and more intensive monitoring.