In clinical practice, cerebral MRI in newborns is highly predictive of neurodevelopmental outcome after therapeutic hypothermia.

OBJECTIVE: In the trials, a substantial proportion of newborns who underwent therapeutic hypothermia (TH) had an adverse outcome after hypoxic-ischaemic encephalopathy (HIE). Cooled babies were noted to have fewer cerebral lesions on MRI but when present lesions were predictive of adverse outcome. We investigate the predictive value of cerebral MRI in babies who undergo cooling in the clinical setting outside of the clinical trials in a prospective UK cohort. RESULTS: Of 75 babies recruited from four centres, neurodevelopment was available for 69 (92%) with 29% (20/69) being abnormal. The unfavourable MRI group (n = 22) had significantly lower motor (p < 0.001), language (p < 0.001) and cognition (p < 0.001) scores on Bayley-III assessment, compared to the favourable MRI group (n = 47). On multiple regression there was a significant relationship between basal ganglia and thalami abnormality and motor (p = 0.002), cognition (p = 0.011) and language (p = 0.013) outcomes. Half of the babies who had an MRI predictive of adverse outcome (11/22) had highest grade cerebral palsy. Cerebral MRI had 95% sensitivity, 94% specificity, 91% PPV and 98% NPV in predicting neurodevelopment. CONCLUSIONS: In this clinical cohort, fewer children had adverse neurodevelopment after TH compared to the TH trials. However, half the children who had an MRI predictive of adverse ND outcome had the most severe form of cerebral palsy. In this cohort, cerebral MRI was found to be highly predictive of neurodevelopmental outcome.