Serial Measurements of Serum Glial Fibrillary Acidic Protein in Moderate-Severe Traumatic Brain Injury: Potential Utility in Providing Insights into Secondary Insults and Long-Term Outcome.

In patients with traumatic brain injury (TBI), serum biomarkers may have utility in assessing the evolution of secondary brain injury. A panel of nine brain-injury- associated biomarkers was measured in archived serum samples over 10 days post-injury from 100 patients with moderate-severe TBI. Among the biomarkers evaluated, serum glial fibrillary acidic protein (GFAP) had the strongest associations with summary measures of acute pathophysiology, including intracranial pressure (ICP), cerebral perfusion pressure (CPP), and brain tissue pO2 (PbtO2). Group based trajectory (TRAJ) analysis was used to identify three distinct GFAP subgroups. The low TRAJ group (n = 23) had peak levels of 9.4 + 1.2 ng/mL that declined rapidly. The middle TRAJ group (n = 48) had higher peak values (31.5 + 5.0 ng/mL) and a slower decline over time. The high TRAJ group (n = 26) had very high, sustained peak values (59.6 + 12.5 ng/mL) that even rose among some patients over 10 days. Patients in the high TRAJ group had significantly higher mortality rate than patients in low and middle TRAJ groups (26.9% vs. 7.0%, p = 0.028). The frequency of poor neurological outcome (Glasgow Outcome Score Extended [GOS-E] 1-4) was 88.5% in the high TRAJ group, 54.2% in the middle TRAJ group, and 30.4% in the low TRAJ group (p < 0.001). ICP was highest in the high TRAJ group (median 17.6 mm Hg), compared with 14.4 mmHg in the low and 15.9 mm Hg in middle TRAJ groups (p = 0.002). High TRAJ patients spent the longest time with ICP >25 mm Hg, median 23 h, compared with 2 and 6 h in the low and middle TRAJ groups (p = 0.006), and the longest time with ICP >30 mm Hg, median 5 h, compared with 0 and 1 h in the low and middle TRAJ groups, respectively (p = 0.013). High TRAJ group patients more commonly required tier 2 or 3 treatment to control ICP. The high TRAJ group had the longest duration when CPP was <50 mm Hg (p = 0.007), and PbtO2 was <10 mm Hg (p = 0.002). Logistical regression was used to study the relationship between temporal serum GFAP patterns and 6-month GOS-E. Here, the low and middle TRAJ groups were combined to form a low-risk group, and the high TRAJ group was designated the high-risk group. High TRAJ group patients had a greater chance of a poor 6-month GOS-E (p < 0.0001). When adjusting for baseline injury characteristics, GFAP TRAJ group membership remained associated with GOS-E (p = 0.003). When an intensive care unit (ICU) injury burden score, developed to quantify physiological derangements, was added to the model, GFAP TRAJ group membership remained associated with GOS-E (p = 0.014). Mediation analysis suggested that ICU burden scores were in the causal pathway between TRAJ group and 6-month mortality or GOS-E. Our results suggest that GFAP may be useful to monitor serially in moderate-severe TBI patients. Future studies in larger cohorts are needed to confirm these results.