A study funded by the National Institutes of Health found that biomarkers present in the blood on the day of a traumatic brain injury (TBI) can accurately predict a patient’s risk of death or severe disability six months later.
Measuring these biomarkers may enable a more accurate assessment of patient prognosis following TBI, according to results published in Lancet Neurology.
Researchers with the Transforming Research and Clinical Knowledge in TBI (TRACK-TBI(link is external)) study examined levels of glial fibrillary acidic protein (GFAP) and ubiquitin carboxy-terminal hydrolase L1 (UCH-L1)—proteins found in glial cells and neurons, respectively—in nearly 1,700 patients with TBI. TRACK-TBI is an observational study aimed at improving understanding and diagnosis of TBIs to develop successful treatments.
The study team measured the biomarkers in blood samples taken from patients with TBI on the day of their injury and then evaluated their recovery six months later. Participants were recruited from 18 high-level trauma centers across the United States. More than half (57%) had suffered TBI as the result of a road traffic accident.
The study showed that GFAP and UCH-L1 levels on the day of injury were strong predictors of death and unfavorable outcomes, such as vegetative state or severe disability requiring daily assistance to function. Those with biomarker levels among the highest fifth were at greatest risk of death in the six months post-TBI, with most occurring within the first month.
GFAP and UCH-1 are currently used to aid in the detection of TBI. Elevated levels in the blood on the day of the TBI are linked to brain injury visible with neuroimaging. In 2018, the U.S. Food and Drug Administration approved use of these biomarkers to help clinicians decide whether to order a head CT scan to examine the brain after mild TBI.
The new study suggests that GFAP and UCH-L1 may also help to predict recovery, particularly among patients with moderate to severe TBI. The biomarkers improved the accuracy of current prognostic models.