A nomogram for predicting neonatal acute respiratory distress syndrome in patients with neonatal pneumonia after 34 weeks of gestation.

Abstract

Objective: To establish a prediction nomogram for early prediction of neonatal acute respiratory distress syndrome (NARDS).

Methods: This is a retrospective cross-sectional study conducted between January 2021 and December 2023. Clinical characteristics and laboratory results of cases with neonatal pneumonia were compared in terms of presence of NARDS diagnosis based on the Montreux Definition. The NARDS group and non-NARDS group were then compared to establish a prediction nomogram for early prediction of NARDS. The predictive accuracy and compliance of the model were evaluated using subject operating characteristic curves, area under the ROC curve, and calibration curves, and the model performance was estimated by self-lifting weight sampling. The Hosmer-Lemeshow test was used to assess the goodness of fit of the model.

Findings: NARDS group consisted of 104, non-NARDS group consisted of 238 newborns in our study. Gestational age, triple concave sign, blood glucose measurement after birth (Glu), Apgar score at the 5th minute (Apgar5), neutrophil count (ANC) and platelet count (PLT) are independent predictors of NARDS in late preterm and term newborns who present with progressive respiratory distress and require varying degrees of respiratory support within the first 24 h of life to minimize work of breathing and restore organismal oxygenation. The area under the ROC curve was 0.829 (95% CI = 0.785-0.873), indicating the model's strong predictive power. In addition, decision curve analysis showed that the model had significantly better net benefits.

Conclusion: In this study, a predictive column-line plot was constructed based on six clinically accessible conventional variables. Early application of this model has a better predictive effect on the early diagnosis of NARDS, thus facilitating more timely and effective interventions.