Abstract 149

Aim: To document the main neonatal outcome and contributing risk factors of very preterm infants treated in our level III unit. Methods: We evaluated the records from 1984-94 retrospectively and from 1995-96 prospectively. Inclusion criteria were: inborn, ≤32 weeks of gestation, no primary lethal condition. Mortality was defined as death during hospital stay, and intracranial hemorrhage (ICH) according to Papile. Statistics were performed by SAS at the Institute for Medical Information Processing*. Multivariate Logistic Regression included 12 potential risk variables. Table Mortality significantly decreased during the 13 years period from 18% to less than 5% (p<0.05), despite a significant increase in the rate of very preterm infants of < 26 weeks of gestation. The incidence and severity of ICH, however, were nearly identical. Logistic Regression for 1984-94 (599/630 cases, 95%) showed a predicting probability of 94.8% between death and the following risk variables(p<0.05): severe RDS, air-leak-syndrome, low Apgar (5 min ≤6), no prenatal steroids and decreasing birth weight. For the overall incidence of ICH the only significant predictive variable was severe RDS (p<0.01) with a predicting probability of 73.4%.

Table 1
Full size table

Conclusion: Within 13 years mortality decreased significantly, the rate of ICH, however, was unchanged. Our risk analysis had a high predicting probability for mortality, but not for ICH.