Identifying a Biological Signature of Trauma-Related Neurodegeneration Following Repeated Traumatic Brain Injuries Compared with Healthy Controls.

Abstract

The objective of this study was to compare participants at-risk for trauma-related neurodegeneration to a healthy control group on outcomes associated with Alzheimer's disease (AD), such as subjective symptoms, neurocognitive performance, plasma biomarkers, volumetrics, amyloid-beta (Aβ) positron emission tomography (PET), and tau PET. Participants completed a comprehensive assessment protocol for neurodegenerative disease, including magnetic resonance imaging (MRI), PET scans for tau and Aβ, blood draw, subjective symptom reports related to neurodegenerative disease, and objective neurocognitive assessment. Surveys included the Neurobehavioral Symptom Inventory (NSI), Insomnia Severity Index (ISI), Epworth Sleepiness Severity (ESS), PTSD Checklist for DSM-5 (PCL-5), Brief Symptom Inventory-18 (BSI-18), Satisfaction with Life Scale (SWLS), Barratt Impulsivity Scale (BIS), and Buss Perry Aggression Questionnaire (BPAQ). PET scans were read by a neuroradiologist and rated positive or negative based upon established cutoffs. General linear models compared participants with TBI history with controls on outcomes. Age, years of education, military status, biological sex, race/ethnicity, and total self-reported TBIs were included as covariates in all models with Bonferroni corrections. Forward stepwise linear regression models were built to associate neuroimaging outcomes with symptom domains; inclusion in the linear regression required a p value <0.1. The average age for both groups was ∼40 years. The TBI group reported an average of five TBIs; the control group reported an average of one TBI. Across seven regions of interest, only one TBI participant met established PET cutoffs for neuropathology in one cortical region. After controlling for age, sex, race/ethnicity, years of education, military status, and TBI history, there were no statistically significant differences between groups in any neurocognitive outcome (p = 0.06-0.95), Aβ or tau PET (p = 0.05-0.70), MRI volumetrics (p = 0.06-0.98), or plasma biomarkers (p = 0.06-0.85). The TBI group had higher NSI, PCL-5, BSI-18, BPAQ, ESS, and ISI scores compared with the controls (p < 0.001-0.042). Within the TBI group, amygdala normative percentile and/or amygdala asymmetry index were included in the final models for NSI, SWLS, PCL5, BIS, BPAQ, and ISI. Only two models included a statistically significant PET outcome in the final model. In this sample with a mean age of 40 and a history of 5+ TBIs, core diagnostic biomarkers for AD were not different from controls despite significantly higher symptom burden. Volumetrics in critical brain regions were associated with several symptom domains in the TBI group, indicating that cortical volumetrics (especially in the amygdala) may be a more viable early biomarker of chronic symptom burden in this population than PET scans.