White matter injury after neonatal encephalopathy is associated with thalamic metabolite perturbations

Abstract

Background: Although thalamic magnetic resonance (MR) spectroscopy (MRS) accurately predicts adverse outcomes after neonatal encephalopathy, its utility in infants without MR visible deep brain nuclei injury is not known. We examined thalamic MRS metabolite perturbations in encephalopathic infants with white matter (WM) injury with or without cortical injury and its associations with adverse outcomes.

Methods: We performed a subgroup analysis of all infants recruited to the MARBLE study with isolated WM or mixed WM/cortical injury, but no visible injury to the basal ganglia/thalamus (BGT) or posterior limb of the internal capsule (PLIC). We used binary logistic regression to examine the association of MRS biomarkers with three outcomes (i) WM injury score (1 vs. 2/3); (ii) cortical injury scores (0/1 vs. 2/3); and (iii) adverse outcomes (defined as death, moderate/severe disability) at two years (yes/no). We also assessed the accuracy of MRS for predicting adverse outcome.

Findings: Of the 107 infants included in the analysis, five had adverse outcome. Reduced thalamic N-acetylaspartate concentration [NAA] (odds ratio 0.4 (95% CI 0.18-0.93)) and elevated thalamic Lactate/NAA peak area ratio (odds ratio 3.37 (95% CI 1.45-7.82)) were significantly associated with higher WM injury scores, but not with cortical injury. Thalamic [NAA] (≤5.6 mmol/kg/wet weight) had the best accuracy for predicting adverse outcomes (sensitivity 1.00 (95% CI 0.16-1.00); specificity 0.95 (95% CI 0.84-0.99)).

Interpretation: Thalamic NAA is reduced in encephalopathic infants without MR visible deep brain nuclei injury and may be a useful predictor of adverse outcomes.

Funding: The National Institute for Health Research (NIHR).

Keywords: Biomarkers; Magnetic resonance imaging; Magnetic resonance spectroscopy; Neonatal encephalopathy; Therapeutic hypothermia.

Fig. 1

Boxplots of proton thalamic MRS [NAA] and Lactate/NAA values for neonates with WM injury with or without cortical injury compared to those with isolated MRI WM injury. The boxplots exhibit the spread of data points across the median and IQR for each injury score group. Medians are represented by horizontal lines; the boxes represent the upper and lower quartiles; and the whiskers indicate 1.5xIQR from upper and lower quartiles. Blue dots indicate normal outcome, red dots are patients with adverse outcomes, and the black dots are those with unknown outcomes.

Fig. 2

Forest plot demonstrating the sensitivity and specificity of thalamic [NAA], Lactate/NAA, NAA/Choline, NAA/Creatine, WM, and the cortex before and after excluding infants with BGT/PLIC injury.

Fig. 3

(a) Axial T₁-weighted MR image showing normal basal ganglia and thalamic area and normal myelination in the posterior limb of internal capsule (b) Axial T₂- weighted MR image showing injury in the frontal and occipital WM (arrowheads). The region of interest for the MR spectroscopy was in the left thalami (red rectangle). (c) Proton MR spectrum analysed by LC Model; [NAA] 4.65 mmol/kg/wet weight (low), Lac/tNAA 0.41, tNAA/Cr 1.19, tNAA/Cho 0.86. The infant had severe adverse outcome and cerebral palsy at two years.