Journal Article

Development and validation of a perinatal risk prediction model for recurrent respiratory tract infections in moderate-to-late preterm infants: a retrospective cohort study

Zhang, Ximeng, Fu, Linlin, Wang, Yuqi, Yang, Hongli, Zhi, Congcong

2025

Overview

Background Despite significant advancements in neonatal care, mid to late preterm infants (32–36 weeks’ gestation) remain at high risk for recurrent respiratory tract infections (RRTIs). Current prevention strategies are limited by the absence of individualized risk assessment tools. This study aimed to identify critical perinatal risk factors and to develop a robust, clinically applicable prediction model for RRTI in this vulnerable population. Methods A retrospective cohort study was conducted at a tertiary care hospital, enrolling 288 preterm infants born between April 2023 and April 2024. Comprehensive maternal, perinatal, and postnatal data were extracted from electronic medical records and supplemented by structured caregiver interviews. A multivariable logistic regression analysis using a stepwise selection method (entry criterion: P  < 0.05; exit criterion: P  > 0.10) was performed to determine independent predictors of RRTI. The derived model was externally validated in a temporally distinct cohort ( n  = 100) from the same center. Model performance was assessed by the area under the receiver operating characteristic curve (AUC), sensitivity, and specificity. Results Seven independent predictors were retained in the final model: small-for-gestational-age (OR = 3.53, 95% CI: 1.41–11.61), intrauterine infection (OR = 4.22, 95% CI: 1.81–9.83), mechanical ventilation > 72 h (OR = 3.00, 95% CI: 1.27–7.14), prolonged antibiotic use (> 30 days/year; OR = 2.23, 95% CI: 1.01–5.05), maternal passive smoking (OR = 2.91, 95% CI: 1.19–7.14), history of RSV infection (OR = 5.61, 95% CI: 2.24–14.08), and vaginal delivery as a protective factor (OR = 0.24, 95% CI: 0.08–0.71). The prediction model demonstrated excellent discriminatory performance with an AUC of 0.935 in the training cohort and 0.927 in the validation cohort. Overall accuracy was 75.3% for the training set and 82.0% for the validation set. Conclusions This study presents a novel risk stratification tool that effectively identifies high-risk moderate-to-late preterm infants and facilitates targeted interventions, such as RSV prophylaxis and enhanced immune monitoring. This advancement enables tailored RSV immunoprophylaxis planning in low-resource Asian NICUs. Nonetheless, further multi-center validation studies are warranted to confirm the model’s generalizability and to refine its predictive accuracy for broader clinical application.

• Publisher: BioMed Central,BioMed Central Ltd,Springer Nature B.V,BMC
• Subject: Accuracy / Allergies / Anemia / Antibiotics / Birth weight / Childbirth / Childbirth & labor / Chronic diseases / Clinical medicine / Cohort analysis / Complications and side effects / Family medical history / Female / Gender / Gestational Age / Health risk assessment / Hospitals / Humans / Infant, Newborn / Infant, Premature / Infant, Premature, Diseases - epidemiology / Infant, Premature, Diseases - etiology / Infections / Internal Medicine / Logistic regression analysis / Male / Maternal health services / Maternal health's impact on neonatal development / Medicine / Medicine & Public Health / Methods / Neonatal nursing / Newborn babies / Pediatrics / Pregnancy / Premature babies / Premature birth / Premature infant / Prevention / Questionnaires / Recall / Recurrence / Recurrent respiratory tract infections / Regression analysis / Respiratory diseases / Respiratory syncytial virus / Respiratory Tract Infections - epidemiology / Respiratory Tract Infections - etiology / Respiratory Tract Infections - prevention & control / Retrospective cohort study / Retrospective Studies / Risk Assessment - methods / Risk Factors / Risk prediction model / Statistical analysis / Vagina / Ventilators