Rapid Diagnosis of Intracerebral Hemorrhage in Patients With Acute Stroke by Measuring Prehospital GFAP Levels on a Point-of-Care Device (DETECT)

Abstract

Background and objectives: The rapid identification of intracerebral hemorrhage (ICH) in patients with symptoms of acute stroke is decisive for prehospital triage and initiation of targeted therapies. Glial fibrillary acidic protein (GFAP) is a highly promising blood-biomarker indicating ICH. In this study, we investigated the potential of a new GFAP test run on a point-of-care platform for distinguishing ICH from ischemic stroke (IS) and stroke mimics (SM) in the prehospital phase.

Methods: This prospective diagnostic accuracy study was conducted at the RKH Klinikum Ludwigsburg, a tertiary care hospital in Baden-Württemberg, Germany. Patients with symptoms of acute stroke admitted within 6 hours of symptom onset were enrolled. Blood samples were collected in the prehospital phase. Plasma GFAP measurements were performed on the i-STAT-Alinity device (duration: 15 minutes) in-hospital. The gold standard was the final diagnosis categorized ICH, IS, or SM.

Results: A total of 353 patients were enrolled (mean age 74.6 ± 13.4 years; 46.7% female). GFAP concentrations were elevated in patients with ICH (n = 76; median 208 pg/mL [interquartile range 60-5,907]) compared with IS (n = 258; 30 pg/mL [29-51]) and SM (n = 19; 48 pg/mL [29-97]; p < 0.001). The optimal GFAP cutoff point to differentiate ICH from IS and SM was 55 pg/mL (area under the curve of 0.880, 95% CI 0.834-0.925, p < 0.001). IS and SM GFAP levels slightly increased in parallel with increasing age. Hence, different GFAP cutoff points were determined to identify ICH across 3 age groups with moderate to high positive predictive values (PPVs) (90.0%-95.5%; minimum lower CI 55.0%, maximum upper CI 99.3%) (sensitivity values 56.3%-72.4%, specificity values 98.9%-99.0% and negative predictive values [NPVs] 87.5%-92.7%). Vice versa, in patients with a moderate to severe neurologic deficit (NIH Stroke Scale >6), GFAP values <30 pg/mL ruled out ICH with a NPV of 100.0%.

Discussion: Laboratory GFAP measurements on a point-of-care platform in blood samples collected from patients with symptoms of acute stroke in the prehospital phase can help to identify ICH with moderate to high PPV. Following confirmation in larger independent cohorts using optimized eligibility criteria and validated age-specific cutoff values, GFAP testing could facilitate optimized triage and the initiation of blood pressure-lowering therapy and anticoagulation reversal in earlier time frames.

Classification of evidence: This study provides Class II evidence that elevated plasma GFAP levels within 6 hours of symptom onset accurately distinguish patients with ICH from IS and SM.