This study aimed to compare the accuracy of four neonatal illness severity scores for predicting mortality in persistent pulmonary hypertension of the newborn (PPHN). This retrospective study included neonates diagnosed with PPHN between 2013 and 2022. The illness severity scores of four commonly used tools were completed for each infant: the Clinical Risk Index for Babies-II (CRIB-II), the Score for Neonatal Acute Physiology-Perinatal Extension version II (SNAPPE-II) in the first 12 h after admission and maximum oxygenation index (OI) and Vasoactive-Inotropic score (VIS) during the first 24 h (OI24max and VIS24max), 48 h (OI48max and VIS48max), and 72 h (OI72max and VIS72max) after admission. We constructed a receiver operating characteristic (ROC) curve to assess the discrimination and accuracy of the scores and determine the cutoff values for predicting mortality. We enrolled 146 neonates (131 survivors and 15 nonsurvivors). The CRIB-II, SNAPPE-II, maximum OI, and VIS were significantly higher in nonsurvivors than in survivors. An OI72max score of 41 showed the highest accuracy in predicting mortality (area under the ROC curve [AUC] of 0.88) with an OI48max score of 31 (AUC: 0.86) and VIS72max score of 430 (AUC: 0.80) showing good accuracy. The best CRIB-II and SNAPPE-II cutoff scores for predicting mortality were 4 (AUC: 0.74) and 32 (AUC: 0.84), respectively. The most accurate illness severity score for predicting mortality was OI72max score of 41. However, the OI48max, SNAPPE-II, and VIS72max scores also showed good accuracy. Mortality prediction using these scores can guide early management and close monitoring.
Read full abstract