When the FDA cleared Abbott’s TBI blood test for the Alinity i lab system in 2023, the agency wrote down exactly what a negative result means: it “is associated with the absence of acute intracranial lesions visualized on a head CT scan.” That sentence is about bleeding and other damage a CT scanner can see. It says nothing about concussion, and that gap matters if you were hurt in a Tampa Bay crash and your emergency room chart shows a normal brain injury blood test.
These tests measure two proteins, GFAP and UCH-L1, that leak into the blood when brain cells are damaged. The FDA’s decision summaries for three of these tests, and the published abstracts of two 2026 studies, line up on both what the tests do well and where they stop.
What the FDA Cleared These Tests to Do
The first test of this kind, the Banyan Brain Trauma Indicator, reached the market through the FDA’s de novo process, the route for first-of-its-kind devices, in February 2018. Its stated use is to help evaluate adults 18 and older with suspected traumatic brain injury and a Glasgow Coma Scale score of 13 to 15, the range doctors call mild, using blood drawn within 12 hours of the injury. The later Abbott tests carry nearly the same language. The Alinity i version uses plasma or serum drawn within 12 hours. The whole-blood i-STAT cartridge, cleared in 2024, uses blood drawn within 24 hours.
Each one says the result is used “in conjunction with other clinical information” and “to assist in determining the need for a CT” of the head. The FDA’s rules for this class of device require the labeling to state that it isn’t meant to be used as a stand-alone device. In plain terms, the question these tests answer is whether a patient needs a CT scan. They were never cleared to answer whether a patient has a concussion.
The Test Was Graded Against the CT Scan
Every study the FDA relied on measured the blood test against one thing: what the head CT showed. In the Banyan study, a panel of three neuroradiologists read each CT, and “positive” meant an acute intracranial lesion, such as a subdural hematoma or a contusion. A patient whose CT was clean counted as negative, whatever symptoms they had.
A test measured against CT findings can’t tell you whether a negative result rules out injuries a CT can’t show, and as my traumatic brain injury and concussion page explains, a standard CT is built to find bleeding and fractures, not the microscopic stretching of nerve fibers in a concussion. None of these studies set out to test whether a negative result rules out concussion. The reference standard was the scan.
What the Test Does Well
These tests are good at what they were built for. In the Banyan study, 1,947 patients were evaluable and 120 had a positive CT. The test was positive in 117 of those 120, a sensitivity of 97.5% (it flagged 117 of 120), and a negative result was right 99.6% of the time. None of the five patients with a lesion that needed surgery got a false negative. Abbott’s Alinity i study reported 96.7% sensitivity and a 99.4% negative predictive value, and the i-STAT study reported 96.5% sensitivity with 14 surgical lesions, all flagged.
The 2026 research points the same way. Menacho-Roman and colleagues, in the Journal of Applied Laboratory Medicine, looked back at 2,932 adults seen within 12 hours of a mild head injury at one tertiary hospital. Their median age was 83. CT found intracranial injury in 230 of them, or 7.8%, and the dual GFAP and UCH-L1 test showed 97.4% sensitivity and a 98.6% negative predictive value. In that study’s patients, a negative result had a 98.6% negative predictive value for intracranial injury the CT could see.
What a Negative Result Leaves Out
A negative result is a narrow finding. It doesn’t rule out a concussion, which a CT usually can’t see either. It doesn’t speak to whiplash, a torn ligament or a herniated disc. And the tests do miss some injuries a CT can see. In the i-STAT study, 10 of the 283 patients with a positive CT tested “Not Elevated.” None of those 10 needed surgery, but each had an injury the scan could see.
Timing matters too. A sample drawn after the cleared window, 12 hours for some versions and 24 for the whole-blood cartridge, falls outside the use the FDA cleared, so I want the draw time and the injury time from the chart before anyone leans on the number.
A Positive Result Isn’t Proof Either
The weakness runs the other way as well, and it can hurt an injury client who over-reads an elevated result. Specificity, the share of people with a clean CT who also test negative, was about 36% in the Banyan study and about 40% in both Abbott studies. A study by Pandey and colleagues in the October 2026 Annals of Emergency Medicine tested 104 adults at a Level I trauma center. Of those, 83 had elevated levels and 11 had injury on CT. Specificity was 0.22 and the positive predictive value (how often an elevated result was right) was 0.12, and false-positive patients were older, a median age of 72 against 58.5 for true negatives. The authors called it a possible “initial rule-out tool” and said larger multicenter studies are needed.
So an elevated GFAP or UCH-L1 reading supports considering a head CT, alongside everything else the doctor knows about the patient. Standing alone, it doesn’t prove a brain injury.
How an Adjuster or Defense Doctor May Use the Result
A negative brain injury blood test can be cited the way a clean CT already is: as a lab value that seems to close the question. An adjuster may write that objective testing was normal. A doctor hired by the defense may testify that biomarkers showed no brain injury. The response comes straight from the FDA’s own words. The test was cleared to help decide whether a CT is needed, it was measured against CT findings, and a negative result is tied to the absence of lesions visible on that scan. A concussion claim rests on the symptom history, the treating doctors’ records and, where it fits, neuropsychological testing, and a protein level drawn in the first hours after a crash doesn’t replace any of that.
When Your ER Chart Shows the Test
If your emergency room records show a GFAP and UCH-L1 test, I ask for the actual values, the cut-offs the lab used, the test brand and sample type, the time of the draw and the time of the crash. Then I compare all of it against your symptoms in the days and weeks that followed, because the blood test was only ever built to answer the CT question. Keep telling your doctors about headaches, memory trouble, light sensitivity or mood changes, even when a test came back normal, so the record shows what the test couldn’t measure.
My traumatic brain injury and concussion page covers how a concussion claim is built, and my page on the insurance company playbook after a Florida crash walks through the arguments carriers raise.
Further Reading
The sources behind this post, so you can read what I read.
- FDA De Novo decision memorandum, Banyan Brain Trauma Indicator, DEN170045 (2018)
- FDA 510(k) decision summary, Abbott TBI test on the Alinity i system, K223602 (2023)
- FDA 510(k) decision summary, i-STAT TBI Cartridge, K234143 (2024)
- Menacho-Roman et al., “Diagnostic Accuracy and Prognostic Value of the Dual GFAP/UCH-L1 Test in Patients Presenting to the Emergency Department with Mild Traumatic Brain Injury,” Journal of Applied Laboratory Medicine 11(6):1378-1391 (2026)
- Pandey et al., “Associations Between Diagnostic Blood-Based Biomarkers and Neuroimaging Findings in Mild Traumatic Brain Injury,” Annals of Emergency Medicine (Oct. 2026)
This post is general information about the law, not legal advice, and it does not create an attorney-client relationship. Past results do not guarantee a similar outcome. The FDA records discussed are the De Novo decision memorandum for the Banyan Brain Trauma Indicator, DEN170045 (2018), and the 510(k) decision summaries for Abbott’s TBI test on the Alinity i system, K223602 (2023), and the i-STAT TBI Cartridge, K234143 (2024). The studies discussed are Menacho-Roman et al., J. Appl. Lab. Med. (2026), doi:10.1093/jalm/jfag110, and Pandey et al., Ann. Emerg. Med. (Oct. 2026), doi:10.1016/j.annemergmed.2026.08.731.

