How often do blood biomarkers flag concussion?

In a study of 323 athletes with concussion, the brain protein GFAP was elevated, meaning it had changed reliably from each athlete’s own baseline, in 16.0% at 0 to 12 hours after injury and 19.0% at 12 to 36 hours. In 37% of athletes, no marker measured was elevated at 0 to 12 hours, so a single marker flagged only a subset.

What did the CARE Consortium biomarker study measure?

Croghan et al., Sports Medicine 2026 analysed baseline and post-injury blood samples from 521 participants: 323 athletes with concussion and 198 controls. The panel included GFAP, UCH-L1, tau, neurofilament light, p-tau181 and inflammatory markers.

The study used a reliable change index: the difference between a post-injury and a baseline value, divided by the standard error of that difference in uninjured controls. A marker counted as elevated when the index exceeded 1.65 (the upper 5%). That is a personal-baseline question, not a population-cutoff one, the distinction behind our post on lab ranges versus personal baselines.

At 0 to 12 hours, GFAP was elevated in 16.0% of athletes with concussion and IL-1RA in 16.4%. At 12 to 36 hours GFAP was elevated in 19.0%, and at 36 to 60 hours in 14.6%. The share with no elevated marker was 37%, 45% and 63% across those three windows.

Are blood biomarkers enough to identify concussion on their own?

In this study, not as a single measure. Loss of consciousness or post-traumatic amnesia was associated with elevated GFAP at 12 to 36 hours (45% versus 15%, p = 0.004).

A separate 2026 prospective study followed 68 professional rugby players and collected samples after 23 concussion episodes (Oris et al., Sports Medicine Open 2026). At 36 hours, GFAP alone had 100% specificity but 17% sensitivity. Combining GFAP with S100B raised sensitivity to 52%, with specificity still 100%. With so few episodes, these figures are imprecise.

In the CARE study, elevations appeared in single markers rather than as a global rise across markers, and the authors concluded that any single marker has limited ability to aid concussion diagnosis.

The bottom line

• In 323 athletes with concussion, GFAP was elevated in 16.0% at 0 to 12 hours and 19.0% at 12 to 36 hours, and 37% had no elevated marker at 0 to 12 hours.
• In a small rugby study (23 episodes), GFAP alone had 17% sensitivity at 100% specificity, and adding S100B raised sensitivity to 52%.
• GFAP alone flagged a minority of concussion cases in both studies (16.0% to 19.0% in CARE, 17% in rugby), in different populations and at different sampling times.