Objective To examine the effect of therapeutic hypothermia on MR biomarkers and neurodevelopmental outcomes in babies with mild hypoxic-ischaemic encephalopathy (HIE).

Design Non-randomised cohort study.

Setting Eight tertiary neonatal units in the UK and the USA.

Patients 47 babies with mild HIE on NICHD neurological examination performed within 6 hours after birth.

Interventions Whole-body cooling for 72 hours (n=32) or usual care (n=15; of these 5 were cooled for <12 hours).

Main outcome measures MRI and MR spectroscopy (MRS) within 2 weeks after birth, and a neurodevelopmental outcome assessment at 2 years.

Results The baseline characteristics in both groups were similar except for lower 10 min Apgar scores (p=0.02) in the cooled babies. Despite this, the mean (SD) thalamic NAA/Cr (1.4 (0.1) vs 1.6 (0.2); p<0.001) and NAA/Cho (0.67 (0.08) vs 0.89 (0.11); p<0.001) ratios from MRS were significantly higher in the cooled group. Cooled babies had lower white matter injury scores than non-cooled babies (p=0.02). Four (27%) non-cooled babies with mild HIE developed seizures after 6 hours of age, while none of the cooled babies developed seizures (p=0.008). Neurodevelopmental outcomes at 2 years were available in 40 (85%) of the babies. Adverse outcomes were seen in 2 (14.3%) non-cooled babies, and none of the cooled babies (p=0.09).

Conclusions Therapeutic hypothermia may have a neuroprotective effect in babies with mild HIE, as demonstrated by improved MRS biomarkers and reduced white matter injury on MRI. This may warrant further evaluation in adequately powered randomised controlled trials.