Predictors of neurological outcome in cooled neonates

Background We define clinical predictors of neurological outcome in neonates with hypoxic–ischemic encephalopathy undergoing hypothermia therapy. Methods Twenty‐one neonates who underwent selective head cooling between 2004 and 2010 and were followed neurologically for ≥24 months were investigated retrospectively. Patients were divided according to the neurological outcome at 2 years of age into group A (n = 11), patients with normal neurological function, and group B (n = 10), patients with neurological disabilities (n = 9) or those who died (n = 1). Predictors were determined by χ2 and Mann–Whitney U‐tests, anova, Spearman rank correlations and receiver–operator curves. Results Group B showed higher average blood lactate levels during the first day, particularly at 24 h of life; lower day‐3 cerebral blood flow resistance index; higher maximum dobutamine dose used; higher rate of thiamylal sodium used; more severe background electroencephalogram suppression during the first week (group A: 11/11 cases ≤ grade 3; group B: 7/9 cases at grade 4–5) and higher rate of cerebral lesions on magnetic resonance imaging in the second week (group A: 1/11 case; group B: 9/10 cases) than group A. The most useful predictor of poor prognosis was cerebral parenchymal lesions on magnetic resonance imaging with 90%, 90% and 90% of sensitivity, specificity and accuracy, followed by week‐1 background electroencephalogram ≥ grade 4 with 70%, 100% and 85% and day‐3 cerebral blood flow resistance index < 0.46 with 71%, 88% and 80%, respectively. Conclusions Prediction of post‐cooling neurological outcome could be improved substantially by evaluating multiple factors.