Background and aims Late-preterm infants have higher rates of respiratory distress compared to term infants. The aim of our study was to investigate perinatal and maternal risk factors for respiratory failure (RF) in late-preterm newborns with respiratory distress (RD) and to compare the ability of respiratory failure indices, such as alveolar-arterial oxygen difference (A-aDO₂), arterial/alveolar oxygen ratio (a/A ratio) and partial oxygen tension/fractional inspired oxygen (PaO₂/FiO₂), to predict RF in these infants.

Methods The medical records of newborns with gestational age 34–37 weeks, admitted in our NICU for RD between January 2006 and June 2007, were reviewed. Gender, gestational age, birthweight, delivery mode, antenatal steroids, premature rupture of membranes and maternal age, diabetes mellitus and hypertension were recorded. Arterial blood gas data within the first 12 hours of life (or until intubation), were also evaluated.

Results Ninety infants were studied and 30 (33.3%) of them developed RF. Logistic regression analysis identified the lack of antenatal steroid administration as the only significant risk factor for RF (p = 0.02). The maximum A-aDO₂ had the best predictive ability (AUC = 0.966). An A-aDO₂ of >200 mmHg had 96.7% sensitivity, 95% specificity and 19.3 likelihood ratio in identifying newborns at risk for RF.

Conclusions The lack of antenatal steroid administration may be related to the development of RF in late-preterm newborns with RD. The maximum A-aDO₂ within the first 12 hours of life may be considered as an early predictor of severe respiratory impairment in these infants.