Models predict adverse outcomes after neonatal hypoxic-ischemic encephalopathy

  • Pediatrics
  • 15 Jan 2021

  • curated by Susan London
  • Clinical Essentials
Access to the full content of this site is available only to registered healthcare professionals. Access to the full content of this site is available only to registered healthcare professionals.

Takeaway

  • Among infants with neonatal hypoxic-ischemic encephalopathy (HIE), 2 models enabled good prediction of death or neurodevelopmental impairment as early as the first few hours of life.

Why this matters

  • Reliable prediction may help target early therapies for those at highest risk.

Key results

  • 49.8% of infants had the primary outcome of death or neurodevelopmental impairment:
    • 37.0% died.
    • 12.8% had neurodevelopmental impairment.
  • Model components:
    • Early model (data obtained around neonatal ICU admission):
      • HIE severity;
      • Epinephrine administration in delivery room;
      • Respiratory support at admission; and
      • Fraction of inspired oxygen of 0.21 at admission.
    • Cumulative model (data obtained throughout hospitalization) included above with modifications, additions:
      • Combined severity variable (EEG findings at 24 hours plus HIE severity);
      • Steroids for BP management; and
      • Significant brain injury on MRI.
  • Discovery models had high areas under the curve:
    • 0.852 for early model.
    • 0.861 for cumulative model.
  • In validation cohort, goodness-of-fit χ2:
    • P=.236 for early model.
    • P=.060 for cumulative model.

Study design

  • US multicenter, retrospective cohort study (Children’s Hospitals Neonatal Consortium Database, 2010-2016).
  • 486 infants born ≥35 weeks of gestation, treated with therapeutic hypothermia for HIE.
  • Main outcome: death or neurodevelopmental impairment (Bayley Scales of Infant Development score, deafness, blindness, gross motor delay) at ≥11 months.
  • Funding: None.

Limitations

  • Composite outcome.
  • Unmeasured, residual confounding.
  • Infants lacking neurodevelopmental follow-up were excluded.