A Clinical Score to Predict CPAP Weaning Success in Late Preterm Neonates: Development and Validation

Background

Late preterm neonates (34-36 weeks) account for approximately 71% of all preterm births and represent the largest subgroup of preterm infants. Respiratory distress affects 25-35% of this population, primarily due to surfactant deficiency and delayed lung fluid clearance. Nasal continuous positive airway pressure (nCPAP) is the primary non-invasive respiratory support modality, with early initiation. However, CPAP weaning decisions remain based on subjective clinical judgment, with significant practice variability.

Gap in Knowledge

No validated scoring system exists to guide CPAP weaning in late preterm neonates, despite this population representing the majority of CPAP users. Premature weaning attempts result in respiratory failure and prolonged NICU stay, while delayed weaning causes nasal trauma and increased complications.

Objective

We aimed to develop and validate the Late Preterm Weaning (LPW) Score to predict successful CPAP weaning in neonates born at 34-36 weeks gestation, incorporating readily available bedside parameters into a practical clinical decision tool.

Methods

This was a two-phase study with a retrospective derivation cohort (September 2020–August 2022, n=189) and a prospective validation cohort (September 2022–December 2023, n=127). Candidate predictors included corrected gestational age, 12-hour median respiratory rate, FiO₂, work of breathing, duration of stable settings, and hemodynamic stability. Multivariate logistic regression identified independent predictors and β-coefficients created a 0-12 point-scoring system. The primary outcome was successful weaning (remaining off CPAP ≥72 hours without re-escalation).

Results

The LPW Score contained six independent predictors with excellent discriminative ability (AUC 0.94, 95% CI 0.90-0.98). At cutoff ≥8 points, sensitivity was 94.7%, specificity was 92.1%, the positive predictive value was 93.5%, and the negative predictive value was 93.8%. Performance remained robust in the validation cohort (AUC 0.92, p=0.45). Risk stratification demonstrated weaning success rates: 96.2% (score 9-12), 72.4% (score 6-8), and 18.9% (score 0-5).

Conclusion

The LPW Score provides objective, standardized guidance for CPAP weaning in late preterm neonates, with potential to reduce practice variability and optimize clinical outcomes. The score’s bedside applicability and strong predictive performance support implementation across NICU settings to standardize evidence-based weaning practices.