Evaluating Targeted Therapeutic Response With Predictive Blood-Based Biomarkers in Patients With Chronic Mild Traumatic Brain Injury

Chronic consequences of mild traumatic brain injury (mTBI) are heterogeneous, but may be treatable with targeted medical and rehabilitation interventions. A biological signature for the likelihood of response to therapy (i.e., "predictive" biomarkers) would empower personalized medicine post-mTBI. The purpose of this study was to correlate pre-intervention blood biomarker levels and the likelihood of response to targeted interventions for patients with chronic issues attributable to mTBI. Patients with chronic symptoms and/or disorders secondary to mTBI >3 months previous (104 days to 15 years; n = 74) were enrolled. Participants completed pre-intervention assessments of symptom burden, comprehensive clinical evaluation, and blood-based biomarker measurements. Multi-domain targeted interventions for specific symptoms and impairments across a 6-month treatment period were prescribed. Participants completed a follow-up testing after the treatment period. An all-possible model's backward logistic regression was built to identify predictors of improvement in relation to blood biomarker levels before intervention. The minimum clinically important difference (MCID) of the change score (post-intervention subtracted from pre-intervention) for the Post-Concussion Symptom Scale (PCSS) to identify treatment responders from non-responders was the primary outcome. The MCID for total PCSS score was 10. The model to predict change in PCSS score over the 6-month intervention was significant (R² = 0.09; p = 0.01) and identified ubiquitin C-terminal hydrolase L1 (odds ratio [OR] = 2.53; 95% confidence interval [CI], 1.18-5.46; p = 0.02) and hyperphosphorylated tau (p-tau; OR = 0.70; 95% CI, 0.51-0.96; p = 0.03) as significant predictors of symptom improvement beyond the PCSS MCID. In this cohort of chronic TBI subjects, blood biomarkers before rehabilitation intervention predicted the likelihood of response to targeted therapy for chronic disorders post-TBI.

Association of Plasma Biomarkers with Sleep Outcomes and Treatment Response After Mild Traumatic Brain Injury

Sleep disturbances occur in up to 70% of patients with mild traumatic brain injury (mTBI). Modern mTBI management recommends targeted treatment for the patient's unique clinical manifestations (i.e., obstructive sleep apnea, insomnia). The purpose of this study was to evaluate the association of plasma biomarkers with symptom reports, overnight sleep evaluations, and response to treatment for sleep disturbances secondary to mTBI. This study is a secondary analysis of a prospective multiple interventional trial of patients with chronic issues related to mTBI. Pre- and post-intervention assessments were conducted, including overnight sleep apnea evaluation, the Pittsburgh Sleep Quality Index (PSQI), and blinded analysis of blood biomarkers. Bivariate Spearman correlations were conducted for pre-intervention plasma biomarker concentrations and 1) PSQI change scores and 2) pre-intervention sleep apnea outcomes (i.e., oxygen saturation measures). A backward logistic regression model was built to evaluate the association of pre-intervention plasma biomarkers with improvement in PSQI over the treatment period (p < 0.05). Participants were 36.3 ± 8.6 years old and 6.1 ± 3.8 years from their index mTBI. Participants reported subjective improvements (PSQI = -3.7 ± 3.8), whereas 39.3% (n = 11) had improved PSQI scores beyond the minimum clinically important difference (MCID). PSQI change scores correlated with von Willebrand factor (vWF; ρ = -0.50; p = 0.02) and tau (ρ = -0.53; p = 0.01). Hyperphosphorylated tau correlated with average saturation (ρ = -0.29; p = 0.03), lowest desaturation (ρ = -0.27; p = 0.048), and baseline saturation (ρ = -0.31; p = 0.02). The multi-variate model (R ² = 0.33; p = 0.001) retained only pre-intervention vWF as a predictor (odds ratio = 3.41; 95% confidence interval, 1.44-8.08; p = 0.005) of improving PSQI scores beyond the MCID. vWF had good discrimination (area under the curve = 0.83; p = 0.01), with an overall accuracy of 77%, sensitivity of 46.2%, and specificity of 90.0%. Validation of vWF as a potential predictive biomarker of sleep improvement post-mTBI could optimize personalized management and healthcare utilization.