Elsevier

The Journal of Pediatrics

Volume 148, Issue 4, April 2006, Pages 438-444.e1
The Journal of Pediatrics

Original article
Early prediction of poor outcome in extremely low birth weight infants by classification tree analysis

Presented in part at the Pediatric Academic Societies’ Meeting, San Francisco, Calif, May 1–4, 2004.
https://doi.org/10.1016/j.jpeds.2005.11.042Get rights and content

Objective

To predict death or neurodevelopmental impairment (NDI) in extremely low birth weight infants by classification trees with recursive partitioning and automatic selection of optimal cut points of variables.

Study design

Data from the Trial of Indomethacin Prophylaxis in Preterms were randomly divided into development (n = 784) and validation sets (n = 262). Three models were developed for the combined outcome of death (8 days to 18 months) or NDI (cerebral palsy, cognitive delay, deafness, or blindness at 18 months corrected age): antenatal: antenatal data; early neonatal: antenatal + first 3 days data; and first week: antenatal, first 3 days, and 4th to 8th days data. Decision trees were tested on the validation set.

Results

Variables associated with death/NDI in each model were: Antenatal: Gestation ≤25.5 weeks and antenatal steroids <7 days. Early neonatal: Birth weight ≤787 g and fluid intake >101 mL/kg/d. First week: Birth weight ≤787 g: transfusion >3 mL/kg/d. Birth weight >787 g: cranial echodense intraparenchymal lesion and transfusion >11 mL/kg/d. Correct classification rates were 61% to 62% for all models.

Conclusions

The ability to predict long-term morbidity/death in extremely low birth weight infants does not improve significantly over the first week of life. Effects of different variables depend on age.

Section snippets

Methods

Infants enrolled in the TIPP trial4 who survived beyond 8 days were included for analysis. In the TIPP trial, infants 500 to 999 g birth weight were enrolled from January 1996 to March 1998 at 32 participating centers in Canada, United States, Australia, New Zealand, and Hong Kong. The Institutional Review Board at each center approved the protocol, and informed consent was obtained from parents or legal guardians before entry into the study. NDI was defined as documentation in survivors of 1

Results

Of the 1202 infants enrolled in the TIPP trial, 97 died in the fist week, 13 were lost to follow-up, and 46 did not have all the outcome data necessary to rule out NDI. Therefore 1046 infants were included for analysis. Fifty percent of the infants were male, the mean gestational age was 26.1 weeks (SD 1.9 w; 26), and mean birth weight was 787g (SD 130 g; 795) (Table II). Four hundred thirty-five of the 1046 infants (41.6%) had development of the primary outcome (death 8 days to 18 months or

Discussion

The accumulation of data over the first week of life does not significantly improve the ability of classification trees to predict long-term morbidity/death in ELBW infants who survive beyond the first week of life, irrespective of model complexity. In contrast to the high accuracy of models for predicting early death,12, 13 our study demonstrates the difficulty of predicting long-term morbidity or death in ELBW infants, because correct outcome was predicted in only about 60% of the infants.

References (24)

  • ACOG Practice Bulletin: Clinical Management Guidelines for Obstetrician-Gynecologists: Number 38, September 2002. Perinatal care at the threshold of viability

    Obstet Gynecol

    (2002)
  • Perinatal care at the threshold of viability

    Pediatrics

    (2002)
  • Cited by (49)

    • Serum Cystatin C as a predictor of acute kidney injury in neonates: a meta-analysis

      2022, Jornal de Pediatria
      Citation Excerpt :

      Acute kidney injury (AKI) is an independent risk factor for neonatal death.1

    • Changes in long-term prognosis with increasing postnatal survival and the occurrence of postnatal morbidities in extremely preterm infants offered intensive care: a prospective observational study

      2018, The Lancet Child and Adolescent Health
      Citation Excerpt :

      Others have reported predictors of disability in extremely preterm infants, with inconsistent conclusions, probably reflecting the heterogeneity of cohort selection (geographical cohorts vs cohorts derived from randomised controlled trials), and differences in gestational ages, ages at follow-up, and definitions of outcomes. Secondary analyses from two large randomised controlled trials20,26 identified three major morbidities—bronchopulmonary dysplasia, serious brain injury on cranial ultrasound, and severe retinopathy of prematurity—that were independently associated with death at 1·5 years or neurodevelopmental impairment at 5 years. The likelihood of death or neurodevelopmental impairment increased with increasing counts of those morbidities.20,26

    View all citing articles on Scopus

    Supported by the Medical Research Council of Canada MT-13288. US centers were supported in part by NICHD U10 HD21364, U10 HD27851, U10 HD21373, U10 HD27881; M01 RR 00997, U10 HD27880; M01 RR 00070, U10 HD21385, U10 HD27904, and U10 HD34216.

    TIPP Investigators listed in the Appendix available at www.jpeds.com.

    View full text