Longitudinal changes in brain metabolites following pediatric concussion

Abstract

Concussion is commonly characterized by a cascade of neurometabolic changes following injury. Magnetic Resonance Spectroscopy (MRS) can be used to quantify neurometabolites non-invasively. Longitudinal changes in neurometabolites have rarely been studied in pediatric concussion, and fewer studies consider symptoms. This study examines longitudinal changes of neurometabolites in pediatric concussion and associations between neurometabolites and symptom burden. Participants who presented with concussion or orthopedic injury (OI, comparison group) were recruited. The first timepoint for MRS data collection was at a mean of 12 days post-injury (n = 545). Participants were then randomized to 3 (n = 243) or 6 (n = 215) months for MRS follow-up. Parents completed symptom questionnaires to quantify somatic and cognitive symptoms at multiple timepoints following injury. There were no significant changes in neurometabolites over time in the concussion group and neurometabolite trajectories did not differ between asymptomatic concussion, symptomatic concussion, and OI groups. Cross-sectionally, Choline was significantly lower in those with persistent somatic symptoms compared to OI controls at 3 months post-injury. Lower Choline was also significantly associated with higher somatic symptoms. Although overall neurometabolites do not change over time, choline differences that appear at 3 months and is related to somatic symptoms.

Figure 1

Example spectrum (black) with LCModel fit (red), and voxel placement in the left dorsal lateral prefrontal cortex (L-DLPFC).

Figure 2

Demographics of concussion and orthopedic injury (OI) participants at each timepoint. Sex and age were not significantly different at the 12 day (p = 0.142, 0.201) and 3 month timepoints (p = 0.234, 0.225). Sex was significantly different between concussion and OI groups at 6 months = 0.0001. Age did not differ significantly at 6 months p = 0.925. Chi square was used to compare sex and t-tests for age.

Figure 3

Concentration of tNAA, tCr, tCho, Glx, and mI at 12 days, 3 months, and 6 months post-concussion. All concussion participants had data collected at 12 days post-injury and were randomized to either the 3- or 6 month data collection. Linear mixed models for each metabolite controlling for age, sex, site, and vendor showed no significant effect of time for any metabolites. The black lines in the boxes represent the median and the box represents the interquartile range, with the 1.5 times the interquartile range values represented by the whiskers. The circles are statistical outliers (greater than 1.5 time the interquartile range from the median) but there were no data quality issues or reasons to exclude these spectra.

Figure 4

Line plots showing metabolite concentration (tNAA, tCr, tCho, Glx, and mI) for the three groups (symptomatic, asymptomatic, and OI) at the three timepoints studied (12 days, 3 months, and 6 months). Symptom status was determined by the HBI reliable change of cognitive or somatic summary scores at 4 weeks (1 month). Metabolite trajectories were not significantly different over time and were not significantly different between groups.

Figure 5

Mean tCho concentrations at 3 months separated into participants with concussion and somatic symptoms, those with concussion but without persisting somatic symptoms, and orthopedic injury (OI). Symptomatic and asymptomatic were defined at 3 months. The black lines in the boxes represent the median and the box represents the interquartile range, with the 1.5 times the interquartile range values represented by the whiskers. The asterisk indicates a significant group difference (p = 0.047) between OI and symptomatic groups.

Figure 6

Scatterplot of the relationship of the metabolite concentrations that were found to have significant associations with HBI symptoms (Cognitive/Somatic) at 12 days, 3 months, and 6 months. (A) 12 day tCr levels and cognitive symptoms at 3 months, (B) 3 month tCho levels and somatic symptoms at 3 months, (C) 12 day mI levels and somatic symptoms at 6 months, (D) 6 month tNAA levels and 6 month cognitive symptoms, (E) 6 month Glx levels and 6 month cognitive symptoms.