Objective:

To examine: 1) whether microRNAs are differentially expressed in whole blood in individuals with traumatic brain injury (TBI) as compared to healthy controls and 2) whether they predict outcomes using the Glascow Outcome Scale – Extended.

Background:

Reliable biomarkers of injury severity and prognosis in TBI are critical for patient selection in future treatment trials and to inform clinical decision making.

Design/Methods:

Whole blood was obtained from 130 patients with TBI and 133 age- and sex-matched healthy controls in PaxGene tubes. RNA was isolated with PAXgene blood miRNA kit (Qiagen, Germantown, MD) and then sequenced with NanoString nCounter miRNA expression panels in three batches. Differential expression was assessed using DEseq2 and including sequencing batch, age, sex, and sequencing batch within the analysis model. Samples from within 1 day of TBI were compared to healthy controls. Stepwise backward regression based on the Akaike information criterion was used to determine which of the differentially expressed microRNAs predicted combined to predict the maximum GOSE score. P-values and multiple testing corrected p-values <0.05 were considered significant.

Results:

Forty-one (5.1%) microRNAs were elevated in TBI as compared to control, and thirty-six (4.5%) were significantly decreased in the same comparison. Of these, four – miR-29a-3p, miR-4516, miR-20a-5p/miR-20b-5p, miR-223-3p - significantly predicted the maximum GOSE score in two of three backward regression models constructed using scaled, normalized, and batch corrected data using three separate methods – remove unwanted variance, ComBat, and Limma. 

Conclusions:

MicroRNAs are ideal candidate biomarkers given their inherent stability in a variety of body fluids. Future studies will seek to validate the use of four miRNAs identified in this study as prognostic biomarkers for TBI, using quantitative PCR in larger validation cohorts.