1137 Initial Risk Score Predicts Survival without Severe Intraventricularhemorrhage (IVH) and/or Periventricular-Leukomalacia (PVL) in Very-Low-Birthweight (VLBW) or Verylow-Gestational-Age (VLGA) Infants

Abstract

Background: We showed that a predictive model with eight perinatal risk factors correctly predicts pre-discharge mortality (ESPR 2009). We aim to predict neonatal survival without severe IVH and/ or PVL.

Methods: In 2006-07, EuroNeoNet collected data from 5,270 VLBW/ VLGA infants (< 32 weeks), admitted before day 28 to 65 NICUs from 14 European countries. A combined outcome for postdischarge survival without grades_3,4 IVH and cystic PVL was used. Regression models to predict survival and combined outcome were developed adjusted for perinatal factors by step-wise selection. Calibration by Hosmer-Lemeshow and discrimination by area under ROC curve was used, latter validated on a different sample. Models to predict combined outcome adjusted by birthweight and/or gestation were also performed. Comparisons between models were based on non-parametric tests over ROC curves.

Results: Mean(±SD) gestation was 28.6(2.7) weeks and birthweight 1,178(382) g. Adding the need for resuscitation at birth (more than just O₂ therapy) to previously developed model for survival (gestation, birthweight, gender, prenatal-steroiduse, delivery-mode, 1 and 5-min Apgar scores and major congenital anomalies) correctly predicted survival without severe IVH and PVL (AUC: 0.825; 95%;CI 0.81-0.84; H&L: 0.489; validation sample: H&L:0.061, AUC 0.829;95%CI:0.807-0.85). This model performed better than those based on birthweight and/or gestation.

Conclusion: Score based on nine initial risk factors adjusted for neonatal survival without severe IVH and PVL can be used as a short-term surrogate for risk of adverse developmental outcome at followup. The score could be used to measure quality improvement initiatives.

Log in or create a free account to read this content

Gain free access to this article, as well as selected content from this journal and more on nature.com

Access through your institution

or

Sign in or create an account

Continue with Google

Continue with ORCiD