Addition of SNAP to perinatal risk factors improves the prediction of bronchopulmonary dysplasia or death in critically ill preterm infants

BACKGROUND Bronchopulmonary dysplasia (BPD) is the most common serious pulmonary morbidity in premature infants. The score for neonatal acute physiology (SNAP) is a physiologic severity index for neonatal intensive care and correlates well with neonatal mortality and clinical outcomes. The prognostic value of the SNAP score for BPD in preterm infants remains to be clarified. The aim of the study was to determine whether SNAP can predict the development of BPD or death, and to investigate the contribution of SNAP to the predictive accuracy of other potential perinatal risk factors for the adverse outcome in critically ill preterm infants. METHODS We conducted a study in 160 critically ill preterm infants with less than 33 gestational weeks. The original SNAP score was prospectively calculated based on 28 items collected during the first 24 hours of admission. The potential perinatal risk factors were assessed during the first 72 hours of life. Major outcome measures were BPD and mortality before the time of BPD screening. RESULTS Of the 160 infants, 17 died and 41 developed BPD. The SNAP score was significantly associated with BPD or death (odds ratio [OR] =1.28; 95% confidence interval [CI], 1.16-1.41; p <0.001), even after adjustment for gestational age (OR =1.27; 95% CI, 1.13-1.41; p <0.001). High SNAP score was an independent predictor of BPD or death (area under the curve [AUC] =0.78; 95% CI, 0.70-0.85; p <0.001), with additional predictive value when combined with other perinatal risk factors. Multivariate regression analysis resulted in a final model, including SNAP, gestational age, apnea of prematurity, patent ductus arteriosus, and surfactant use as independent risk factors, with a higher predictive accuracy compared with individual components (AUC =0.92; 95% CI, 0.87-0.96; p <0.001). CONCLUSIONS SNAP is associated with adverse outcome of BPD or death. High SNAP scores are predictive of BPD or death in critically ill preterm infants, and add prognostic value to other perinatal risk factors.