Interinstitutional variation in prediction of death by SNAP-II and SNAPPE-II among extremely preterm infants

BACKGROUND: Illness severity scores predict death among infants admitted to NICUs. We know of no study limited to a population defined by an extremely low gestational age. METHODS: A total of 1467 infants born before the 28th postmenstrual week at 14 institutions were given Score for Neonatal Acute Physiology II (SNAP-II) and Score for Neonatal Acute Physiology Perinatal Extension II (SNAPPE-II) values based on data collected within the first 12 postnatal hours. All deaths in the intensive care nursery were identified. RESULTS: The rate of death before postnatal day 28 was 13% (interinstitutional range: 7%-20%), whereas the overall mortality rate was 18% (8%-31%). SNAP-II values, SNAPPE-II values, and mortality rates tended to decrease with increasing gestational age. Even within gestational age strata, however, the risk of death decreased with decreasing SNAP-II and SNAPPE-II values. The positive predictive values of most SNAP-II and SNAPPE-II cutoff levels were close to 30%. In general, institutions\u27 mortality rates increased with the proportions of infants whose SNAP-II values were \u3e/=30. CONCLUSION: The physiologic instability in the first 12 postnatal hours that is identified by illness severity scores conveys information about the risk of death among infants at the lowest gestational ages