Italian Multicenter Study on Very Low Birth Weight Babies.

The Italian Multicenter Study on Very Low Birth Weight babies (IMS-VLBW) is the first collaborative investigation performed in Italy on the health status of newborns weighing less than 1500 g at birth. Eight Neonatal Intensive Care Units (NICUs) participated in the study: Cagliari, Napoli, Padova, Palermo, Roma, Sassari, Trieste, Udine. Data were analyzed in the Laboratorio di Epidemiologia e Biostatistica of the Istituto Superiore di Sanità. The objectives of the study were established in the following: a) to collect accurate descriptive data on neonatal morbidity, mortality and long term outcome of VLBW babies admitted to NICUs; b) to analyze the risk factors of unfavourable outcome (death or handicap) and to analyze, with respect to outcome, the relationships between risk factors, neonatal diseases and therapeutical procedures; c) to test the feasibility of a multicenter follow-up programme based on the use in all participating Centers of the same diagnostic criteria (the results of follow-up will be presented in a forthcoming paper). In the years 1987 and 1988, 634 newborns weighing 500-1499 g at birth were enrolled in the study. In-hospital mortality for the whole group was 33.1% (65.1% in the 500-999 g birth weight class and 19.2% in the 1000-1499 g class). Mortality was not different for inborn vs outborn babies. A high incidence of unfavourable perinatal conditions was observed in these babies, namely birth asphyxia, sub-optimal care during transport, poor clinical conditions on arrival to the NICU. Neonatal diseases, like respiratory distress syndrome and peri-intra ventricular hemorrhage were also frequent and severe. A logistic regression analysis of pre-admission risk factors of in-hospital death identified eight statistically significant variables: birth weight; gestational age; sex; antenatal steroid stimulation of lung maturity; first minute Apgar score; absence of spontaneous respiration, body temperature and pH on arrival to the NICU. Using the equation derived from the logistic regression analysis a theoretical mortality rate, predicted on the basis of the local incidence of pre-admission risk factors, was calculated for each Center. In no case the predicted mortality was statistically different from the observed one, suggesting that in our study differences in observed mortality rates from one Center to another are largely influenced by pre-admission risk factors.