Validation of a neuropathology score using quantitative methods to evaluate brain injury in a pig model of hypoxia ischaemia.

ABSTRACT

BACKGROUND: Neuropathological examination is the classic outcome measure in experimental studies of newborn brain injury to evaluate novel therapies. We have used a graded neuropathology score in an established global model of perinatal hypoxic-ischaemic (HI) injury. We wished to validate the score using cell counting in our model. NEW METHOD: 32 newborn pigs underwent a 45 min global HI insult then maintained at normothermia (NT, rectal temperature, Trectal 38.5 °C) for 72 h or mild total body hypothermia (HT, Trectal 37.0 °C) combined with selective head cooling for 48 h and subsequently maintained at NT for 24h before brain perfusion fixation. A perinatal pathologist scored haematoxylin and eosin stained 6 µm histological sections for injury in the hippocampus and basal ganglia on a 9-step scale (0.0=no injury, 4.0=>75% injury). We counted the number of healthy neurons in the hippocampus CA1 region and putamen using morphological criteria in eight random, non-overlapping fields from representative sections. RESULTS: Healthy neuronal cell density correlated with neuropathology score in the hippocampus CA1 (r = -0.74) and in the putamen (r = -0.75) and both measures detected a difference between groups. The correlation coefficients were better for the NT compared to the HT group in both the hippocampus (r = -0.87 vs. -0.53) and putamen (r = -0.77 vs. -0.54). COMPARISON WITH EXISTING METHOD: We have validated a histological neuropathological scoring system in our model of perinatal HI by showing correlation between neuronal cell count and estimated injury. CONCLUSIONS: Our neuropathology score is a valid method to assess brain injury with good reproducibility and sensitivity.