Early GFAP and UCH-L1 point-of-care biomarker measurements for the prediction of traumatic brain injury and progression in patients with polytrauma and hemorrhagic shock

Abstract

Objective: Traumatic brain injury (TBI) and hemorrhage are responsible for the largest proportion of all trauma-related deaths. In polytrauma patients at risk of hemorrhage and TBI, the diagnosis, prognosis, and management of TBI remain poorly characterized. The authors sought to characterize the predictive capabilities of glial fibrillary acidic protein (GFAP) and ubiquitin C-terminal hydrolase L1 (UCH-L1) measurements in patients with hemorrhagic shock with and without concomitant TBI.

Methods: The authors performed a secondary analysis on serial blood samples derived from a prospective observational cohort study that focused on comparing early whole-blood and component resuscitation. A convenience sample of patients was used in which samples were collected at three time points and the presence of TBI or no TBI via CT imaging was documented. GFAP and UCH-L1 measurements were performed on plasma samples using the i-STAT Alinity point-of-care platform. Using classification tree recursive partitioning, the authors determined the measurement cut points for each biomarker to maximize the abilities for predicting the diagnosis of TBI, Rotterdam CT imaging scores, and 6-month Glasgow Outcome Scale-Extended (GOSE) scores.

Results: Biomarker comparisons demonstrated that GFAP and UCH-L1 measurements were associated with the presence of TBI at all time points. Classification tree analyses demonstrated that a GFAP level > 286 pg/ml for the sample taken upon the patient's arrival had an area under the receiver operating characteristic curve of 0.77 for predicting the presence of TBI. The classification tree results demonstrated that a cut point of 3094 pg/ml for the arrival GFAP measurement was the most predictive for an elevated Rotterdam score on the initial and second CT scans and for TBI progression between scans. No significant associations between any of the most predictive cut points for UCH-L1 and Rotterdam CT scores or TBI progression were found. The predictive capabilities of UCH-L1 were limited by the range allowed by the point-of-care platform. Arrival GFAP cut points remained strong independent predictors after controlling for all potential polytrauma confounders, including injury characteristics, shock severity, and resuscitation.

Conclusions: Early measurements of GFAP and UCH-L1 on a point-of-care device are significantly associated with CT-diagnosed TBI in patients with polytrauma and shock. Early elevated GFAP measurements are associated with worse head CT scan Rotterdam scores, TBI progression, and worse GOSE scores, and these associations are independent of other injury attributes, shock severity, and early resuscitation characteristics.

Keywords: point of care; polytrauma; shock; trauma; traumatic brain injury.

FIG. 1.

ROC curves for arrival GFAP 3094-pg/ml cut point (A), 4-hour GFAP 1661-pg/ml cut point (B), and 24-hour GFAP 1434-pg/ml cut point (C) for predicting high Rotterdam scores on initial head CT. Figure is available in color online only.

FIG. 2.

Forest plot for bivariate confounder adjustment for the GFAP 286-pg/ml cut point prediction of TBI on initial head CT. BP = blood pressure.

FIG. 3.

Forest plots for bivariate confounder adjustment for the arrival GFAP 3094-pg/ml cut point prediction of high Rotterdam score on the first (A) and second (B) head CT scans.