Magnetic Resonance Spectroscopy Predicts Outcomes for Children With Nonaccidental Trauma

Abstract

OBJECTIVE: We evaluated proton magnetic resonance spectroscopic imaging (MRSI) findings for children with traumatic brain injury attributable to nonaccidental trauma (NAT) early after injury, to determine whether brain metabolite changes predicted outcomes. METHODS: Proton MRSI (1.5 T) was performed (mean: 5 days after injury [range: 1–30 days]) through the level of the corpus callosum for 90 children with confirmed NAT. Regional N-acetylaspartate/total creatine, N-acetylaspartate/total choline, and choline/creatine ratios and the presence of lactate were measured. Data on long-term outcomes defined at ≥6 months were collected for 44 of 90 infants. We grouped patients into good (normal, mild disability, or moderate disability; n = 32) and poor (severe disability, vegetative state, or dead; n = 12) outcome groups. RESULTS: We found that N-acetylaspartate/creatine and N-acetylaspartate/choline ratios (mean total, corpus callosum, and frontal white matter) were significantly decreased in patients with poor outcomes (P < .001). A logistic regression model using age, initial Glasgow Coma Scale score, presence of retinal hemorrhage, lactate on MRSI scans, and mean total N-acetylaspartate/creatine ratio predicted outcomes accurately in 100% of cases. CONCLUSIONS: Reduced N-acetylaspartate levels (ie, neuronal loss/dysfunction) and elevated lactate levels (altered energy metabolism) correlated with poor neurologic outcomes for infants with NAT. Elevated lactate levels may reflect primary or secondary hypoxic-ischemic injury, which may occur with NAT. Our data suggest that MRSI performed early after injury can be used for long-term prognosis.