Objective: The purpose of this study was to develop and validate an empirical scoring system to predict the evolution of neonatal respiratory distress syndrome (RDS) into chronic lung disease (CLD) in preterm infants, by comparing it with a more complicated logistic regression model.
Design: Clinical study.
Setting: Neonatal intensive care unit.
Patients: The retrospective analysis of a 3-year experience showed that a gestational age (GA) of less than 30 weeks, a birth weight (BW) of less than 1000 g, the diagnosis of hyaline membrane disease (HMD) and pulmonary interstitial emphysema (PIE) during the first 72 h of life, the peak inspiratory pressure (PIP) and the fraction of inspired oxygen (FIO2) were the highest relative risk factors correlated with the evolution of CLD. On this basis an empirical and a statistical scoring system were defined and prospectively applied at 3 and 5 days of life to 228 neonates with BW less than 1,250 g. The results obtained with both scoring systems were then compared.
Results: Of the 149 infants surviving at 28 days of life, 67 (GA: 29.9 +/- 2.3 weeks; BW: 1,058 +/- 143 g) were normal and 82 (GA: 27.5 +/- 3.9 weeks: BW: 838 +/- 200 g) had CLD. Using a cut-off value of 4.0, the empirical scoring system showed a specificity of 97.0% and a sensitivity of 92.7% on the 3rd day of life; on the 5th day of life the specificity was still 95.5%, while sensitivity remained 92.7%. The areas under the ROC curves plotted with both scoring systems tested were similar.
Conclusions: The proposed empirical scoring system is easy to use and is highly reliable. The application of this scoring system provides the opportunity to direct aggressive treatment for CLD toward only very high risk patients between the 3rd and 5th days of life.